Information on Hodgkin's for Mary Ann

from: Abeloff: Clinical Oncology, 2nd ed.,

Chapter 90 - Hodgkin's Disease

Dwight Kaufman

Dan L. Longo

Hodgkin's disease (HD) is a clonal malignancy of the lymphatic system with protean clinical manifestations. One of the first tumor types found to be curable at both localized and advanced stages of disease, it has been of surpassing importance in the evolution of the disciplines of both medical and radiation oncology. Carefully designed prospective clinical studies of Hodgkin's disease, with anatomic and pathologic correlations with clinical outcomes, have led to dramatic improvements in the outlook for patients with this disease and have served as the template for accurate staging systems and design of clinical trials leading to improved outlook for other types of cancer as well. Hodgkin's disease is now cured in about four fifths of the cases, a near inversion of the odds of mortality as compared with cases diagnosed before 1960. However, many aspects of the biology of Hodgkin's disease remain poorly understood, and considerable disagreement persists among oncologists regarding optimal therapy of both early and advanced disease. The debates relate to how best to achieve an even higher cure rate while minimizing the considerable treatment-related late side effects of both chemotherapy and radiation therapy.

ETIOLOGY AND EPIDEMIOLOGY

The etiology of Hodgkin's disease has not been determined. Environmental risk factors do not appear to play

2621

Earlier studies based on case control comparisons using a surrogate risk group in communities with a high incidence of acquired immunodeficiency syndrome (AIDS) suggested that HIV infection was not a risk factor for development of Hodgkin's disease, ^{[10] [11]} and indeed there has been no easily recognizable epidemic of Hodgkin's disease in the high-AIDS areas. However, a longitudinal study of 6,700 homosexual men with known HIV status demonstrated conclusively that the incidence of Hodgkin's disease in HIV-infected men is increased, with an age-adjusted standardized morbidity ratio of 5.0, resulting in 19 excess cases of Hodgkin's disease per 100,000 person-years of HIV infection. ^[12] In the same population, the relative risk for non-Hodgkin's lymphomas was 37.7 resulting in 225 excess cases per 100,000 person-years of HIV infection. Hodgkin's disease has not been classified by the Centers for Disease Control as an AIDS-defining illness, although most of the HIV-infected patients with Hodgkin's disease would be classified with AIDS on the basis of CD4+ T-cell count less than 200/mm³ .

Hodgkin's disease represents less than 0.7 percent of all new cases of cancer in the United States and 0.3 percent of cancer-related deaths. The age-adjusted incidence rate for Hodgkin's disease is 2.9 per 100,000; this rate has decreased by 15.5 percent over the past 21 years. ^[13] This decrease was most significant in the 65 and over age group; however, the incidence increased modestly in adolescents and young adults ages 15 through 34 years with nodular sclerosis (NSHD) subtype. The apparent decrease in incidence in the older population is due to improved diagnostic accuracy (cases of non-Hodgkin's lymphoma having been misclassified as Hodgkin's disease in the Surveillance, Epidemiology, and End Results [SEER] database from the early 1970s), whereas the increase in NSHD in young adults may be a real increase. ^[14] The incidence is higher in whites than in blacks, in males than in females, and in general, is higher in developed countries than in undeveloped countries. A bimodal distribution of age at the time of diagnosis is observed with peak incidences occurring in the 20- to 24- and the 80- to 84-year-old age groups ^[13] ; the earlier peak predominantly involves patients with diagnosis of NSHD. In the United States, 7,100 new cases of Hodgkin's disease are estimated for 1998 and 1,400 deaths are expected. ^[15] The country-wide age-adjusted death rate declined by 66 percent over the 30-year period ending in 1989, while the 5-year survival for patients with Hodgkin's disease increased from 40 percent in 1960-1963 to 78 percent in the 1983-1988 period. For the latest period in which data are available (1989-1993), 5-year survival is 86 percent

PATHOLOGY

The characteristic neoplastic cells in Hodgkin's disease generally comprise a small minority of the cellular population of the involved lymph nodes, existing in a background of normal lymphocytes and inflammatory cells. In fact, the four histologic subtypes of the Rye classification system, which has been used almost exclusively in North America since its proposal in 1966, including nodular sclerosis (NSHD), lymphocyte predominant (LPHD), mixed cellularity (MCHD), and lymphocyte depleted (LDHD), are based entirely on descriptions of the nonmalignant background of the involved node rather than on the characteristics of the malignant cells themselves. Recognizing the increasing role immunophenotyping now plays in the diagnosis of lymphomas, including Hodgkin's disease, as well as maturing information about the biology, immunophenotype, and clinical characteristics of the nodular subtype of lymphocyte predominant Hodgkin's disease (NLPHD), the International Lymphoma Study Group proposed a revised classification system for Hodgkin's disease in 1994, as part of the Revised European-American Lymphoma (REAL) classification. ^[16] The REAL classification system recognizes NLPHD as a distinctly different disease from classic Hodgkin's disease, which now comprises NSHD, MCHD, and LPHD. The new World Health Organization (WHO) classification is patterned on the REAL classification.

A diagnosis of Hodgkin's disease requires microscopic

2622

Reed-Sternberg Cells

The malignant cell in classic Hodgkin's disease is the Reed-Sternberg cell or one of its morphologic variants. Reed-Sternberg cells are giant cells with abundant eosinophilic cytoplasm, and multiple, large, deep blue-staining nucleoli. The Reed-Sternberg cell characteristically has two or more nuclei or lobes (it is not entirely clear whether Reed-Sternberg cells are multinucleated or have polylobated nuclei). The so-called owl's eye appearance is classic; it is generated by a central nucleolus within each nucleus (or nuclear lobe). Although nonclassic Reed-Sternberg variants may be diagnostic of Hodgkin's disease involvement in tissue specimens when the diagnosis has been previously established, or at relapse, most pathologists require the unequivocal demonstration of at least one characteristic Reed-Sternberg cell for the initial diagnosis of classic Hodgkin's disease. Particularly in LDHD, Reed-Sternberg cells may be quite difficult to distinguish from the cells of other types of lymphoma, particularly peripheral T-cell lymphoma ^{[17] [18]} or anaplastic large-cell CD30+ null/T-cell lymphoma. ^[19]

Lacunar Cell Variant

The lacunar cell variant is characteristic of NSHD. This cell is a giant cell with multiple nuclei, or a multilobated nucleus, typically with less conspicuous nucleoli than the classic Reed-Sternberg cell, and a pale, eosinophilic or clear cytoplasm, if the node is preserved in B5 fixative. Formalin fixation of the node results in retraction of the cytoplasm around the nucleus giving the artifactual appearance of a nucleus separated from adjacent cells by a clear space that is, a nucleus appearing to reside within a lacuna. These cells are usually quite numerous in NSHD.

A number of immunophenotypic markers are routinely expressed on Reed-Sternberg cells and its variants in biopsy specimens. However, their specificity in distinguishing Hodgkin's disease from benign lymphoid processes or non-Hodgkin's lymphoma has been disappointing. Two moderately specific antigens, CD30 (Ki-1) ^{[20] [21] [22] [23]} and CD15 (Leu-M1), ^[21] are usually expressed on both typical Reed-Sternberg cells and mononuclear variants found in all subtypes of classic Hodgkin's disease. CD30 (Ki-1) is also expressed in Ki-1-positive anaplastic large cell lymphoma (ALCL) and can be expressed on EBV-infected B cells and the Th2 subset of CD4+ T cells. In normal lymphoid tissue, its expression is restricted to a few extrafollicular activated T and B lymphoblasts and the B lymphoblasts located at the rim of germinal centers. ^{[20] [22]} CD15 (Leu-M1) is present on 85 to 95 percent of normal monocytes and has been shown to be rarely expressed in immunoblastic lymphomas ^[24] and in lymphomatoid papulosis, acute myelogenous leukemia (AML), chronic myelogenous leukemia (CML), and acute lymphoblastic leukemia (ALL). Other less specific markers include the low-affinity interleukin-2 (IL-2) receptor (CD25), transferrin receptor (CD71), and HLA-DR, which are common to activated lymphoid cells as well as monocytes. Most of the nonmalignant lymphocytes present in nodes involved with Hodgkin's disease are CD4+ T cells. The paucity of cytotoxic CD8+ T cells may be related to the absence of class I HLA expression by Reed-Sternberg cells.

In some cases (about 20 percent), the Reed-Sternberg cell expresses B-cell markers; in some (about 20 percent), T-cell markers. Rarely (about 3 percent of cases) the cells express both B- and T-cell markers. In most cases (about 57 percent), the cells express neither B- nor T-cell markers. Immunohistologic analysis is difficult because the malignant cells are surrounded by normal T and B cells and the localization of any particular marker can be difficult to ascertain with certainty. Examination of DNA extracted from Hodgkin's disease-involved tissues occasionally shows clonal immunoglobulin gene rearrangements, clonal T-cell receptor gene rearrangements, both, or neither. However, this technique is suboptimal for the analysis of genetic changes in rare cells within a tumor mass. Experiments involving the isolation by micromanipulation of single Reed-Sternberg cells, too, have yielded some conflicting results. However, it appears that Reed-Sternberg cells that express B-cell markers such as CD20 contain clonally rearranged immunoglobulin genes and, thus, are derived from B cells. Not all cases express B-cell markers, so it is not clear whether Reed-Sternberg cells can only be derived from B cells. Only in the LPHD subset do

2623

Lymphocytic-Histiocytic Cell

Lukes originally described a variant cell associated solely with LPHD, which he called the lymphocytic-histiocytic or L&H cell variant. ^[25] These are large cells with a variable amount of cytoplasm, a multilobated, puffy, twisted nucleus with a vesicular chromatin pattern, and multiple small, basophilic nucleoli. Because of their nuclear contour, they have been described as popcorn cells. L&H cells are usually scattered in faintly defined nodules composed of small, benign lymphocytes and histiocytes with small clusters of the malignant cells or may be more abundantly interspersed among diffuse infiltrates of lymphocytes and epithelioid histiocytes, which completely efface the lymph node architecture.

Immunophenotyping is key to accurate characterization of the cells of LPHD. L&H cells are negative for CD15 (Leu-M1), negative or only weakly positive for CD30 (Ki-1), and almost always positive, in optimally fixed and processed specimens, for pan-B monoclonal markers such as CD20. ^{[26] [27] [28] [29] [30] [31] [32]} LP popcorn cells also express the B-cell marker CD79a, and in most cases, they contain J chain, a B-cell component. They express epithelial membrane antigen (EMA), which is not expressed on any normal B cells. In addition, the non-neoplastic cells in LP have some interesting features. The T cells that cluster around the neoplastic cells often express CD57, a natural killer cell marker. Within the vague nodules of tumor is a meshwork of follicular dendritic cells that express CD21.

Histologic Subtypes

Classic Hodgkin's Disease

The histologic features of NSHD include the finding of typical Reed-Sternberg cells, lacunar cells, and interconnective broad sclerotic bands of collagenous connective tissue that divide the lymph node into cellular nodules. The lacunar cells may be diffusely scattered or clustered together in the nodules among a variety of normal cells, lymphocytes, epithelial histiocytes, eosinophils, and neutrophils. The finding of clusters or sheets of lacunar cells in an appropriate background in the absence of the sclerotic bands is considered diagnostic of NSHD in the cellular phase, if diagnostic binucleated or multinucleated Hodgkin's cells can also be identified. ^[33] NSHD disease may be subclassified further into NS lymphocyte predominant, NS mixed cellularity, and NS lymphocyte depleted, and a two-tiered grading system is widely used in Europe. However, prognosis among these subtypes is similar when aggressive treatment, particularly chemotherapy, is used ^{[34] [35]} ; thus, further subtyping of NSHD is of little clinical utility.

NSHD is the most common histologic type occurring in developed countries, accounting for 50 to 75 percent of all cases. The incidence of NSHD peaks in late adolescence and the early 20s, but remains the most common subtype at all ages. A modest female predominance has been reported in most series. NSHD most frequently involves the mediastinum and the supraclavicular areas. In the Stanford series reported in 1971, 69 of 85 NSHD cases (81 percent) had mediastinal involvement and 69 of 76 cases (91 percent) of Hodgkin's disease with mediastinal involvement had NSHD. ^[36]

In MCHD, lymph nodes are effaced by a pleiomorphic cellular infiltrate comprising a mixture of normal histiocytes, neutrophils, eosinophils, plasma cells, lymphocytes, and fibroblasts, the prototypic environment referred to as the "appropriate milieu" for the diagnosis of Hodgkin's disease. The criteria for diagnosis include abundant Reed-Sternberg cells and diagnostic variants with an absence of fibrotic bands. MCHD accounts for 15 to 30 percent of all cases of Hodgkin's disease, the second most frequent histology in non-HIV-infected patients. MCHD is the most frequent histology in HIV-infected patients, accounting for nearly 60 percent of the cases. There is peak age incidence between 30 and 45 years, and between ages 40 and 55, MCHD and NSHD have similar incidences. Male/female ratio is 1.5:1. MCHD is much less likely than is NSHD to be localized and has a propensity for retroperitoneal node involvement.

In LDHD, bizarre multinucleated Reed-Sternberg cells or "malignant-looking" Hodgkin's cells are found in a characteristic stroma profoundly depleted of lymphocytes and other reactive cells; a disorganized deposition of proteinaceous fibrillar matrix may be seen. With modern immunologic techniques, most of the cases previously classified as LDHD are now identified as belonging to the non-Hodgkin's lymphomas. Fewer than 5 percent of initial Hodgkin's disease diagnoses are classified as LDHD, most frequently in elderly men presenting with advanced stage disease. LDHD is found in 20 percent of HIV-infected patients with Hodgkin's disease. It is more frequently seen as an evolution from MCHD in relapsed, heavily pretreated cases. LDHD has a propensity for involvement of retroperitoneal nodes and extralymphatic sites. LDHD can be difficult to distinguish from the lymphocyte-depleted subtype of NSHD and from non-Hodgkin's lymphoma, particularly peripheral T-cell lymphoma ^{[17] [18]} and anaplastic large-cell lymphoma. ^[19] In the National Cancer Institute (NCI) series reported by Kant et al., of 39 cases originally diagnosed as LDHD, 11 were reclassified as lymphocyte-depleted variant of NSHD and 14 others were reclassified as non-Hodgkin's lymphoma. ^[17]

Lymphocyte-Predominant Hodgkin's Disease

It is now clear that LPHD is a B-cell lymphocytic malignancy, different both biologically and clinically from classical Hodgkin's disease. ^[37] The diagnosis of LPHD now must meet both histologic and immunophenotypic criteria. LPHD is characterized by pan B-cell+, CD15-, CD30- neoplastic L&H cells, which may be sparse or numerous, and abundant lymphocytic and/or histiocytic stroma with few inflammatory cells and essentially no necrosis. In most cases, faint nodules can be discerned at low power, and these cases have been

2624

Although clonal immunoglobulin JH gene rearrangements have not been found, the malignant cells of LPHD are of B-cell lineage. In situ immunostaining and hybridization techniques demonstrate immunoglobulin light chain protein and mRNA in the L&H cells of LPHD. ^[38] In 28 of 32 cases of LPHD reported by Stein et al., the L&H cells contain cytoplasmic J chains. ^[39] Surface IgM and IgD expression is seen, consistent with a follicular center origin of these cells; however, rearrangement and expression of the bcl-2 gene typical of follicular B-cell lymphomas is not detected in LPHD, nor is t(14;18) detected by PCR. ^[40] The T cells present in the LPHD-involved nodes are somewhat unusual in that they express both CD4 and CD57. ^[41]

LPHD has distinctive clinical features differing from those of classical Hodgkin's disease. ^[42] The incidence is evenly distributed across all ages from 20s to 60s, with a 2:1 male predominance. ^[43] Three fourths of cases present as stages I and II, ^{[42] [44] [45]} with disease found predominantly in cervical, axillary, or inguinal lymph nodes, and almost never involving the mediastinum. ^[46] There is a very low likelihood of finding intra-abdominal disease at laparotomy in patients with clinical stage I LPHD presenting in either the inguinofemoral, cervical, or axillary nodal regions. ^[44] This finding implies that these patients can be safely treated with radiation therapy to only one side of the diaphragm without surgical staging.

Similar to follicular lymphoma, LPHD has a tendency to run an indolent course, even when left untreated, with an 80 percent 10-year survival. ^[43] LPHD has a very high complete remission rate when treated with radiation therapy. An analysis of 71 patients with LPHD treated at the Joint Center for Radiation Therapy in Boston revealed 10-year freedom-from-relapse and 10-year overall survival rates of 80 percent and 93 percent, respectively. ^[42] A distinct tendency towards late relapses, which nevertheless can be effectively retreated and do not significantly affect overall survival, has been reported. In one study, initial chemotherapy did not improve the freedom from relapse, with relapses occurring at a slow but continued rate to at least 10 years after treatment. ^[45]

Non-Hodgkin's lymphomas have been reported subsequent to both treated and untreated LPHD. ^{[32] [43] [45] [47]} Sundeen et al. ^[47] described seven cases and Chittal et al. ^[32] characterized five additional cases of either evolution from LPHD to diffuse large cell lymphoma (DLCL) or composite LPHD and DLCL. Both groups considered the DLCL to have progressed from the LPHD rather than having arisen independently, although molecular evidence that the subsequent malignant lymphoma and the LPHD arose from the same cell is lacking. The cases reported by Sundeen et al. all responded well to therapy, with six of seven complete remissions and no relapses (with median follow-up of 22 months); the one partial remission remained stable off therapy for 1.5 years. ^[47]

Lymphocyte-rich classic Hodgkin's disease is a rare form of Hodgkin's disease that resembles LPHD, but the malignant cells look like classic Reed-Sternberg cells rather than popcorn cells. The background contains small lymphocytes and scattered eosinophils and plasma cells. Recognition of this form of Hodgkin's disease is felt to be important because this disease behaves much more like classic Hodgkin's disease than does LPHD

BIOLOGIC CHARACTERISTICS

Attempts to characterize unequivocally the ancestry of the Hodgkin's neoplastic cell using immunohistochemical or genetic marker techniques have been inconclusive. Clonal rearrangements of immunoglobulin genes or T-cell receptor-alpha or -beta genes have been found using Southern blotting of DNA from involved nodes ^{[48] [49] [50] [51]} ; however, in the heterogeneous population consisting of a small fraction of Reed-Sternberg cells in tissues involved with Hodgkin's disease, these findings are not readily interpretable. Analyses of the Hodgkin's cell lines show clonal immunoglobulin rearrangements in six lines, T-cell receptor rearrangements in four, and none of the cell lines is germline for both. ^{[52] [53] [54] [55] [56]} Since no Ig protein or T-cell receptors have been detected in any, the observed Ig and TCR gene rearrangements in cell lines may represent abortive rearrangements, arrested very early in the stepwise recombinatorial sequence. ^[55] These rearrangements are not specific for malignant cells, and since lineage infidelity may be a common feature of tumors derived from primitive hematopoietic cells, they may not be used to assign lineage. Thus, although the available data point to a lymphoid origin, we do not believe them sufficient yet to justify unequivocal assignment of lymphoid origin to all Hodgkin's cells, and in fact, a heterogeneous origin is possible.

A search for characteristic marker proto-oncogenes in Hodgkin's disease, particularly focusing on c- myc, ^[57] activating point mutations in c- ras, ^[58] and other oncogenes ^[59] has been inconclusive. Using the PCR technique, Stetler-Stevenson et al. reported DNA sequences carrying the fusion of the proto-oncogene bcl-2 with JH Ig sequences, the characteristic finding associated with the t(14;18) translocation typical of follicular non-Hodgkin's B-cell lymphoma, in 17 of 53 (32 percent) nodal tissues involved with Hodgkin's disease. ^[60] The bcl-2 rearrangement was confirmed in one study, with 4 of 21 patients showing evidence of the translocation ^[61] ; however, a similar study failed to demonstrate the bcl-2 rearrangement by PCR analysis in a study of 34 cases of Hodgkin's disease. ^[62] It is possible that the bcl-2 translocation may arise in submicroscopic follicular B-cell lymphoma (composite lymphoma) coexisting

2625

Genetic lesions affecting many chromosomes and regions have been identified in Reed-Sternberg cells. None of the many abnormalities qualify as recurring lesions, and the genes disrupted have not yet been identified to the extent that putative involvement in pathogenesis can be inferred. The aneuploidy results in variable numbers of individual chromosomes being present in the cells. In one study, Reed-Sternberg cells contained between two and eight copies of individual chromosomes. Mutations in p53 have been identified in some cases and, depending upon the technique used to examine the cells, 30 to 60 percent of cases contain evidence of EBV infection. When EBV is present, it is usually in the form of a clonal episome. Few viral antigens are expressed, however; LMP1 is the only viral gene product that has been consistently found in the cases containing EBV genomes.

Many of the characteristic clinical features of Hodgkin's disease can be explained by expression of cytokines and hematopoietic growth factors by the malignant cells. ^[63] These factors may act in a complex autocrine and paracrine loop, stimulating both the malignant cells and the nonmalignant stromal cells, including lymphoid and myeloid cells and fibroblasts, to proliferate and to secrete other growth factors. These factors could then act locally to stimulate the Hodgkin's cells or systemically to cause "B" symptoms (fever, night sweats, and weight loss), immunodeficiency, eosinophilia, thrombocytosis, and marrow fibrosis. Many known cytokines and growth factors have been found in the supernatants from various Hodgkin's-derived cell lines. ^[64] Expression of a known eosinophil growth factor, IL-5, in the cytoplasm of clearly identifiable Reed-Sternberg cells was demonstrated in 16 of 16 cases of Hodgkin's disease with eosinophilia, but in none lacking eosinophilia, by in situ hybridization of tissue specimens. ^[65] Similarly, the Reed-Sternberg cells in involved lymph nodes that were infiltrated with large numbers of eosinophils were found by immunostaining techniques to contain and secrete abundant amounts of IgE ^[66] ; elevated levels of circulating IgE have also been demonstrated in Hodgkin's disease patients with eosinophilia. Eosinophils express the cell surface Fc receptor for IgE, the CD23 antigen. The fibrotic bands characteristic of nodular sclerosis Hodgkin's disease have been postulated to be a response to stimulation by tumor growth factor-beta (TGF-beta), a cytokine capable of stimulating proliferation of, as well as collagen synthesis by, fibroblasts. TGF-beta was demonstrated in the cytoplasm of Reed-Sternberg cells in one case and on the cell surface of five other cases of NSHD, but in no other histologic subtypes, by immunohistochemical stain of tumor tissue ^[67] ; TGF-beta has been found in the urine of four patients with untreated NSHD and the protein disappeared following effective therapy. ^[68]

CLINICAL MANIFESTATIONS

The first manifestation of Hodgkin's disease in at least 90 percent of cases is enlarged lymph nodes, with cervical adenopathy the most frequent presenting site (Table 90-1) . ^[3] Although usually painless, pain and tenderness of the enlarged nodes in Hodgkin's disease may be experienced. Rarely, the pain may be brought on or exacerbated by alcohol ingestion. Incidental discovery of an asymptomatic mediastinal mass on routine chest

TABLE 90-1 -- CLINICAL MANIFESTATIONS OF HODGKIN'S DISEASE

Findings at presentation

Adenopathy

Mediastinal mass

Splenomegaly

Abdominal mass

Symptoms

Fever, weight loss, night sweats

Pruritis

Bone pain

Laboratory findings

Thrombocytosis

Leukocytosis

Eosinophilia

Elevated erythrocyte sedimentation rate

Elevated alkaline phosphatase

Paraneoplastic syndromes

Dermatologic

Nodular prurigo

Ichthyosis

Psoriasiform lesions

Erythema nodosum

Dermatomyositis

Linear IgA bullous dermatosis

Leukocytoclastic vasculitis

Toxic epidermal necrolysis

Renal and metabolic

Nephrotic syndrome

Hypercalcemia

Hypoglycemia

Lactic acidosis

Neurologic

Inflammatory brachial plexopathy

Guillain-Barre syndrome

Sensory ganglionitis

Acute cerebellar degeneraton

Stiff-man syndrome

Ophelia syndrome

2626

Constitutional symptoms including high fevers, weight loss, and drenching and debilitating night sweats have been associated with poor prognosis. To be defined as "B" symptoms in the Ann Arbor staging classification, weight loss must be unexplained and greater than 10 percent of the body weight during the 6 months before staging; fever must be unexplained, persistent, or recurrent temperatures greater than 38°C over the previous month; and night sweats must be recurrent and drenching over the previous month. The pattern of the fever has been noted to be intermittent in some patients, not occurring daily but in cycles of more or less continuous fever lasting 1 or 2 weeks separated by afebrile periods of similar duration. However, this classic manifestation of Hodgkin's disease called Pel-Ebstein fever is rare. It is more common for the fever to peak in the evening and break precipitously in the early morning hours leading to night sweats. The patient is often unaware of the fever until it breaks. These systemic symptoms are undoubtedly associated with the elaboration by the malignant cells of circulating cytokines, such as tumor necrosis factor (TNF) and IL-1, each of which has been identified in Hodgkin's cell lines in culture. ^[70] The diagnostic work-up for fever of unknown origin may lead to a diagnosis of Hodgkin's disease, most frequently with the discovery of an abdominal mass by CT scan or at exploratory laparotomy. ^[71] In women, night sweats must be distinguished from "hot flashes" associated with ovarian failure, particularly in patients previously treated with chemotherapy.

Laboratory values are frequently completely normal at the time of diagnosis; the most frequent abnormal findings are hematologic, including thrombocythemia, eosinophilia, granulocytosis even to the extent of a leukemoid reaction, elevated ESR, and less frequently, significant anemia. Elevated ESR is more frequent in more advanced stage disease and may have prognostic significance in patients treated with radiation therapy alone. The ESR can be used as a marker for disease recurrence in previously treated patients. ^[71] The serum alkaline phosphatase level is not infrequently elevated in patients presenting with advanced disease, and can be a signal of either liver involvement or bone or bone marrow involvement.

In addition to the characteristic systemic symptoms, a variety of paraneoplastic syndromes have been reported in patients with Hodgkin's disease, including hematologic manifestations and alterations of the skin, nervous system, and kidneys. With the exception of thrombocytosis and eosinophilia, these syndromes are unusual, and are most likely to be seen in relapsed patients with widespread disease. Eosinophilia occurs in approximately 15 percent of cases ^[3] and appears to be related to expression by Reed-Sternberg cells of IL-5, an eosinophil growth factor. ^[65] It can be a manifestation of general leukocytosis or a specific absolute elevation of eosinophils. A survival advantage has been reported for patients with selective eosinophilia. ^[73] Thrombocytosis is also common, and frequently heralds relapse in patients who had elevated platelet counts before initial treatment. Both autoimmune (Coombs'-positive) hemolytic anemia ^{[74] [75] [76]} and autoimmune thrombocytopenic purpura (ITP) ^[77] can be seen in Hodgkin's disease and, although more likely in advanced disease, have been reported as presenting manifestations.

Pruritis is a common manifestation of Hodgkin's disease and may be severe, debilitating, and associated with nodules and plaques. ^{[78] [79]} It has been reported to precede the diagnosis of Hodgkin's disease by several years, and the recrudescence of pruritus may be a harbinger of relapse. The intensity can increase as the disease progresses. A variety of nonspecific dermatologic disorders can be associated with active Hodgkin's disease, ^[80] including ichthyosis and hyperkeratosis, psoriasiform lesions, urticaria, erythema nodosum, leukocytoclastic vasculitis, ^[81] dermatomyositis, ^[82] linear IgA bullous dermatosis ^{[83] [84]} (manifested with widespread, severely pruritic hemorrhagic bullae), and even fatal toxic epidermal necrolysis. ^[85] With the advent of more effective therapy in recent years, it appears that these skin manifestations are less common than in the first half of the century. The skin is directly infiltrated with Hodgkin's disease in 0.5 to 3.4 percent of cases ^{[86] [87]} either by retrograde lymphatic spread, direct extension from underlying lymph nodes, or by hematogenous dissemination. Although usually seen in advanced disease, presentations in the skin as a first manifestation have been reported. ^{[87] [88]}

Renal and metabolic disorders in association with Hodgkin's disease have been described. A single case of primary renal Hodgkin's disease has been reported. ^[89] Nephrotic syndrome occurring in the absence of amyloidosis or renal vein thrombosis is an unusual but recognized paraneoplastic syndrome that has been most frequently reported in Hodgkin's disease, but also in non-Hodgkin's lymphomas, leukemias, and various carcinomas. ^[90] In Hodgkin's disease, the most frequent renal abnormality associated with nephrotic syndrome is a minimal change lesion or lipoid nephrosis, ^{[91] [92] [93]} but membranous and membranoproliferative glomerulonephritides have also been found. The pathophysiology of these lesions is not known. Secondary systemic amyloidosis with nephrotic syndrome associated with Hodgkin's disease has been reported. ^[94] Paraneoplastic hypercalcemia has been uncommonly reported in Hodgkin's disease ^{[95] [96]} ; humoral parathyroid hormone (PTH)-like substance has not been found in these cases. Rather, similar to the hypercalcemia in sarcoidosis and other granulomatous diseases, the pathophysiology of hypercalcemia in Hodgkin's disease may be related to increased conversion to 1,25-dihydroxy vitamin D₃ , with increased gastrointestinal absorption of calcium. ^{[97] [98]}

2627

Neurologic complications of Hodgkin's disease include side effects of therapy, infections, direct effects of tumor, and paraneoplastic syndromes. Most common is the peripheral neuropathy associated with vincristine treatment. This is usually initially manifested as loss of Achilles reflexes and distal numbness and tingling, and can progress to foot drop and difficulty with fine-motor control of the fingers. An autonomic neuropathy with obstipation is not infrequent. Vincristine neuropathy is rarely severe before greater than 10 mg cumulative dose, and almost always resolves over several months after the completion of chemotherapy. Radiation can cause brachial plexopathy or transverse myelopathy with paraplegia. These irreversible complications are dose- and fractionation/schedule-dependent and are seldom seen with modern treatment planning techniques. However, a reversible demyelinating process affecting the cord may be seen after standard mantle irradiation or other techniques taking the cord to doses of 40 to 45 Gy. Shock-like sensations along the spine and tingling and pain in the hands, referred to as Lherrmitte's sign, associated with neck flexion or arm extension are experienced transiently starting 2 to 4 months after radiation therapy. ^[103] The most common direct complication of Hodgkin's disease is spinal cord compression by epidural masses originating in vertebrae or in retroperitoneal or mediastinal lymphatics, by direct extension. ^{[104] [105]} Hodgkin's disease involves brain parenchyma only rarely, usually as an end-stage manifestation.

Several debilitating and even fatal paraneoplastic neurologic syndromes have been reported. ^[106] Although the precise pathophysiology of these cases has not been determined, the cases seem to share an inflammatory component leading to the supposition that immune phenomena are involved, perhaps cross-reactivity of tumor antigens with myelin epitopes or normal nerve cell or axonal antigens. Although upper extremity paresthesias, sensory loss, or weakness in a patient with Hodgkin's disease should prompt a search for tumor involvement of spinal nerve roots or brachial plexus and a consideration of radiation-induced brachial plexopathy, a case of inflammatory paraneoplastic brachial plexopathy associated with Hodgkin's disease was recently reported. ^[107] An acute demyelinating neuropathy resulting in Guillain-Barre syndrome has been reported in several cases. ^{[108] [109] [110]} A fatal case of a rapidly progressing Guillain-Barre-like syndrome in a patient with recently diagnosed Hodgkin's disease was recently reported. ^[111] The patient abruptly developed burning dysesthesias rapidly progressing to involve total-body pain and sensory loss, autonomic dysfunction with initial preservation of motor function. The patient died 5 days later. Complete loss of spinal sensory ganglion cells at multiple levels was demonstrated at autopsy. Several well-documented cases of subacute or abrupt paraneoplastic cerebellar degeneration have been reported. ^{[112] [113] [114]} Most patients present with ataxia, diplopia, nystagmus, and vertigo. The neurologic symptoms can be severe and some patients have become bedbound. Both CT and magnetic resonance imaging (MRI) scans are reported to be normal. Anti-Purkinje cell antibodies have been found in the sera of some patients. Effective treatment for Hodgkin's disease has resulted in complete remission of the cerebellar syndrome. A case of stiff-man syndrome in a patient with Hodgkin's disease was recently reported. ^[115] The patient presented with neuromuscular hyperactivity, painful cramps and spasms, and limb and truncal stiffness responsive to diazepam. CT scan, MRI, myelogram, and electroencephalogram (EEG) were normal. Antiglutamic acid decarboxylase (GAD) cerebellar antibodies were present in the patient's serum. Effective chemotherapy of the Hodgkin's disease resulted in disappearance of the symptoms. A psychiatric syndrome characterized by recent memory loss, depression, personality change, and hallucinations associated with Hodgkin's disease and termed the "Ophelia syndrome" has been reported. ^{[116] [117]} In one patient, CT scan, EEG, and cerebrospinal fluid (CSF) studies were normal. The memory and emotional disturbances improve with successful treatment of the Hodgkin's disease.

Hodgkin's Disease and HIV

The incidence of Hodgkin's disease is increased among HIV-infected patients. ^[12] Analogous to non-Hodgkin's lymphoma in HIV-infected patients, Hodgkin's disease has a propensity for presentation with extranodal advanced disease and very aggressive behavior and is associated with short survival. ^{[9] [118] [119] [120] [121]} Two thirds of patients present with extranodal disease, with bone marrow involvement in half, but lung, liver, bone, paraspinal masses, rectum, pleura, tongue, ^[122] and skin ^[123] presentations have also been reported. The frequencies of mixed cellularity and lymphocyte-depleted histologic subtypes are increased and nodular sclerosis frequency is substantially decreased in comparison with Hodgkin's disease in non-HIV-infected patients. Only one case of lymphocyte predominance subtype has been reported. ^[121] Patients are susceptible to increased severity of chemotherapy-induced myelotoxicity, and median survival is less than 12 to 15 months. Nevertheless, treatment for Hodgkin's disease was successful in patients, particularly those with a CD4+ cell count higher than 250/mm³ and in patients without defined AIDS, alowing complete remission and very prolonged survival in some patients.

STAGING

The Ann Arbor staging system for Hodgkin's disease has been universally accepted for nearly three decades as the foundation for reporting and comparing treatment results and for making treatment decisions. ^{[69] [124]}

2628

The anatomic lymph node regions (Fig. 90-1) , originally defined by Kaplan and adopted at the Rye Conference, ^{[127] [128]} used to distinguish stages in the Ann Arbor

Figure 90-1 Anatomic lymph node regions. (From Kaplan HS, Rosenberg SA: The treatment of Hodgkin's disease. Med Clin North Am 50:1591, 1996, with permission.)

In the late 1970s, the poor prognosis associated with large masses, particularly in the mediastinum, was recognized; this led to the designation by a subscript "X" to denote large tumor masses in the revised staging system. Bulky masses in the mediastinum are defined as masses having a transverse diameter equal to or greater than one third the transverse diameter of the thorax at T5-T6 on a posteroanterior (PA) chest x-ray. Outside of the mediastinum, masses greater than 10 cm in diameter are defined as bulky. The Cotswolds committee attempted to remedy ambiguity in the Ann Arbor definition of "E" extranodal disease by emphasizing the necessity that IIE or IIIE be used only to designate a single extranodal site that is either (1) near and proximal to or (2) in contiguous extension from a site of known nodal involvement. This distinction is intended to solidify the original intention that the "E" designation be "for extralymphatic disease so limited in extent and/or location that it can still be subjected to definitive treatment by radiotherapy" (Kaplan, ^[3] p. 346), which defines a group with the same prognosis, when treated with radiation therapy, as a group with a similar distribution of nodal-only disease. Patients with solitary

2629

EVALUATION

Pretreatment Evaluation

Most patients with Hodgkin's disease present with peripheral adenopathy accessible to surgical excision. If possible, an entire lymph node should be excised for pathologic examination. With a confirmed diagnosis of Hodgkin's disease, the subsequent evaluation (Table 90-3) includes a medical history, with careful documentation of fevers, drenching night sweats, weight loss, pruritus, pain elicited by alcohol consumption, and other changes in the patient's physical state; a complete physical examination with special attention directed to all lymphatic areas including Waldeyer's ring (involvement suggests a diagnosis other than Hodgkin's disease). Perpendicular diameters of nodal masses in each region must be measured and the size of liver and spleen noted. Laboratory tests should include complete blood count (CBC), ESR, and complete chemistry panel including biochemical tests of liver and renal function.

A number of small studies have suggested other serum markers of prognosis, including soluble CD25, soluble CD30, soluble CD4, soluble CD8, and certain cytokines such as IL-10. However, none of the studies has adequately evaluated the degree to which any of these markers are independent prognostic factors. At the moment, none of these tests is used routinely in practice. Radiographic evaluation must include PA chest x-ray, CT of the chest (unless plain chest radiograph is normal), and CT of abdomen and pelvis. Other noninvasive tests that may be indicated include technetium bone scan, ultrasonography, and, when pericardial involvement is suspected, echocardiography. MRI studies usually add little information to the CT scans. Gallium-67 ( ⁶⁷ Ga) scanning can occasionally add useful information for initial staging, detecting unsuspected disease or confirming disease, ^[129] but is most useful in evaluating response to treatment, especially in the mediastinum. ^{[130] [131] [132]} A double-dose (10 mCi) ⁶⁷ Ga scan with 72-hour delay in scanning is recommended to re-stage the mediastinum after treatment. The development of single-photon emission computed tomography (SPECT) ⁶⁷ Ga scanning ^[133] has greatly increased the sensitivity of this procedure. Nevertheless, gallium scanning is not very useful in initial staging.

Bipedal lymphangiography (LAG) remains a useful staging procedure, but is used less frequently now than it was in the 1970s and 1980s, even in referral centers. LAG is the most accurate and sensitive method of detecting and following retroperitoneal lymph node involvement including the iliac and para-aortic chains to

TABLE 90-3 -- RECOMMENDATIONS FOR DIAGNOSTIC EVALUATION OF PATIENTS WITH HODGKIN'S DISEASE

Mandatory procedures

Biopsy, with interpretation by a qualified pathologist

History, with recording of (1) age; (2) gender; (3) presence or absence of unexplained fever and its duration; (4) unexplained sweating, especially at night, and its severity; (5) unexplained loss of weight as a percentage of usual body weight and rapidity of loss; (6) unexplained pruritis

Physical examination with special attention to evaluation of lymphadenopathy, size of liver and spleen, and evaluation of bone tenderness

Laboratory tests

Complete blood count

Erythrocyte sedimentation rate

Liver and kidney function

Alkaline phosphatase

Radiographic examinations

Chest radiography, posteroanterior and lateral

Abdominal and pelvis computed tomography

Bilateral bone marrow aspirates and biopsies

Contingent procedures

Thoracic computed tomography if mediastinal, hilar, and/or mediastinal involvement is seen or suspected on chest radiography

Bipedal lymphangiogram if available

Laparotomy, only if decisions regarding management will be influenced

Liver biopsy (percutaneous or computed tomography-guided) if there is a clinical or radiographic indication of hepatic involvement, or if there is evidence of splenic involvement

Optional ancillary procedures

Double-dose gallium-67 scan

Technetium bone scan

Magnetic resonance imaging

Ultrasonography

Echocardiography

the level of the renal veins, but has little if any utility in evaluating upper abdominal or mesenteric adenopathy. The characteristic "foaminess" and/or filling defects can be identified in involved nodes that are too small to be considered abnormal by CT. The superiority of LAG in detecting para-aortic node involvement was demonstrated by Castellino et al. in correlations of lymphographic and CT accuracy with histologic diagnosis after laparotomy. ^[134] ^{[135] [136]} They found an overall accuracy of 95 percent for lymphangiography, with a 3 percent false-positive rate and an 11 percent false-negative rate in predicting para-aortic lymph node involvement, while the overall accuracy of CT scans for predicting histologically positive para-aortic lymph nodes was 87 percent, with a 35 percent false-negative rate. The LAG has the additional advantage of allowing a cheap and readily available monthly follow-up of response to therapy and post-therapy follow-up by an abdominal flat plate x-ray (kidney, ureter, and bladder [KUB]). On the other hand, LAG is tedious and time consuming to perform, requires 24- and 48-hour follow-up radiographs, and interpretation demands considerable expertise of the radiologist. Moreover, if chemotherapy is to be employed, it is no longer of critical importance to detect small sites of disease not detected by CT. LAG is also associated with a transient decrease in lung diffusing

2630

capacity; thus, patients with pre-existing interstitial lung disease should not undergo LAG.

Bilateral posterior iliac crest trephine bone marrow biopsies should be performed in patients with clinical stage III or IV disease and those with "B" symptoms. Bone marrow involvement is found in less than 1 percent of patients stage IA or IIA disease. Peripheral lymph node biopsy of suspicious nodes on the opposite side of the diaphragm from sites of known disease should be performed if the result will change the stage. Extralymphatic sites suspected on the basis of physical examination, radiographic studies, or laboratory abnormalities of harboring disease, including bone, lung, or liver, should be biopsied if the results will change the stage or would alter the treatment plan.

Laparotomy staging to detect subdiaphragmatic disease in patients with clinical stage (CS) I or II supradiaphragmatic disease is being done less and less frequently. Staging laparotomy with splenectomy is an expensive procedure and is associated with potential immediate and delayed morbidity. Both acute and late bowel and other abdominal complications are seen. Splenectomy results in a lifelong risk for overwhelming sepsis, notably caused by encapsulated gram-positive bacteria, most frequently Streptococcus pneumoniae (pneumococcus). Vaccination 10 to 14 days before splenectomy with pneumococcal, Haemophilus influenzae, and meningococcal vaccines is quite effective in preventing these infections. In addition, laparotomy results in 4 to 7 weeks' delay in initiating definitive therapy. ^{[137] [138]} Thus, laparotomy should be performed only if radiation therapy alone will be used for pathologically documented early-stage disease. Since chemotherapy is being used with increasing frequency in early-stage patients, it should be emphasized that a decision, for whatever clinical reason, to use chemotherapy, either as sole treatment or in combination with radiation therapy, renders staging laparotomy unnecessary. Even when radiation therapy is planned as sole treatment modality for clinical early-stage patients, the utility of staging laparotomy has been questioned by some. There is now good evidence that treatment success among clinically staged patients treated with radiation therapy is not different from that for surgically staged patients, ^{[129] [139] [140]} but it must be recognized that the large radiation fields used in these patients include the intact spleen and portions of the left ventricle and lung, with potential for increased late toxicity to heart and lung. There are subsets of patients with supradiaphragmatic CS IA disease with a very low probability, in some cases approaching zero, of subdiaphragmatic disease. These patients can clearly be safely treated with radiation therapy alone without laparotomy staging, in some cases with mantle-field irradiation only. With the above exceptions, many North American radiation oncologists continue to recommend surgical staging for CS IA and IIA patients. These issues will be discussed in greater detail below. Staging laparotomy must be carefully planned among the surgeon, the pathologist, and the oncologist, and meticulously executed.

2631

Post-treatment Evaluation

The most important initial result of therapy for Hodgkin's disease is a documented complete remission (CR), the sine qua non of cure. ^[141] One month after completion of the final cycle of chemotherapy or radiation therapy, all known sites of disease should be re-evaluated radiographically and, if they are accessible, pathologically. Previously involved bone marrow, liver, and residual peripheral adenopathy should be biopsied. In general, restaging thoracotomy or laparotomy is not recommended. Unfortunately, needle aspirates of residual masses are useful only if they are positive. Persistent but stable radiographic abnormalities in areas of previous disease that are not accessible to easy rebiopsy, such as mediastinal widening or persistent distortion of lymph node architecture on a lymphangiographic study or residual stable masses on CT scanning, may be classified as CR[U], indicating an unconfirmed or uncertain remission if the patient is asymptomatic and otherwise without evidence of active disease. ^[125] With modern scanning techniques, a positive post-treatment ⁶⁷ Ga scan in an area of residual radiographic abnormality, particularly a mediastinal mass, is very sensitive in detecting residual disease and predicting relapse. ^{[142] [143] [144]} However, a negative ⁶⁷ Ga scan after therapy has a significantly lower predictive value. ^[143] The sensitivity of MRI scans in distinguishing active residual disease from fibrosis is not defined.

Patients in CR should be seen at a minimum of 3-month intervals during the first 2 years following therapy, at 4- to 6-month intervals during the next 3 years, and yearly after the fifth year. Each visit should include careful physical examination; laboratory tests including CBC with differential, ESR, alkaline phosphatase, and liver transaminases; a chest x-ray if mediastinal disease was present; and KUB if the retroperitoneal nodes remain radiopaque. Periodic CT scans may sometimes be indicated. On the other hand, a recent cost-benefit analysis suggests that routine follow-up laboratory and radiographic tests, other than chest x-ray, rarely led to early detection of relapse, and method of relapse detection did not have a significant impact on success of salvage therapy. ^[145] Thus, less frequent follow-up is permissible. On the other hand, the late effects of treatment make it highly desirable to keep in contact with patients and, as noted below, efforts at early diagnosis of treatment complications should be made.

THERAPY

Within 5 years of Rontgen's discovery of x-rays, Hodgkin's disease was first treated with radiation therapy, with dramatic, although transient, reductions in the size of involved lymph nodes. Radiation therapy during the first half of the 20th century was delivered with kilovoltage machines adapted from diagnostic radiology equipment; nevertheless, improvements were made as a result of increasing doses of radiation, within the limits allowable by skin toxicity, and extension of radiation fields beyond the limits of obviously involved lymph nodes. These advances resulted from seminal observations of the clinical behavior and patterns of relapse of Hodgkin's disease, initially made by Gilbert in the 1930s, further developed by Peters in the 1940s and 1950s, to the point that Hodgkin's disease treated in its early stages with aggressive orthovoltgage radiotherapy could be considered occasionally curable. The development of the linear accelerator in the 1950s provided megavoltage therapeutic radiation, thus allowing treatment to large volumes of disease with decreased skin toxicity. Through a series of randomized trials beginning in 1962, Kaplan and his colleagues at Stanford University demonstrated that megavoltage radiation therapy, when delivered to extended fields and to tumor-sterilizing doses, can cure a majority of patients with stage I and II disease. ^{[146] [147]} The careful pathologic and clinical correlations conducted by the Stanford group expanded and clarified Gilbert's and Peters' early suggestions that Hodgkin's disease is of unifocal origin and tends to spread in an orderly fashion to contiguous lymph node groups.

In 1964, the first patient with advanced Hodgkin's disease was treated at the NCI with MOPP (mechlorethamine, vincristine, procarbazine, and prednisone) chemotherapy. It became obvious following the first abstract report of this therapy in 1967 and the subsequent initial full report in 1970 that even patients with far-advanced Hodgkin's disease could be cured. Over the past three decades, incremental advances have been made in both radiation therapy and chemotherapy for Hodgkin's disease, and the two modalities have been successfully integrated for initial treatment of some presentations. Because, historically, radiation therapy was the first effective therapy, attempts to decrease mortality from Hodgkin's disease have frequently involved the use of more radical radiation therapy including extranodal organs in some cases. The importance of late toxicity from radiation therapy, especially the risk of second solid tumors, heart disease, and late

2632

In 1999, most oncologists still agree that the preferred treatment of most cases of early-stage Hodgkin's disease is radiation therapy, that the preferred treatment of advanced-stage disease is combination chemotherapy, and that optimal treatment of bulky disease, regardless of stage, is combined chemotherapy and radiation therapy. Disagreements persist regarding the location of the boundary between early and advanced stage; furthermore, there are proponents of combined-modality therapy for both nonbulky early-stage and some cases of advanced stage disease, ^[148] as well as proponents of the routine use of chemotherapy alone for early-stage disease. ^[149]

Radiation Therapy

Randomized trials from Stanford and elsewhere comparing involved field (IF) with successively larger treatment fields, including subtotal nodal (STNI) and total nodal irradiation (TNI), plus radiation to extranodal organs at risk, demonstrated improved freedom from relapse (but not improved overall survival, as discussed below), with more extensive fields. ^[147] These studies led to the development of the following principles of curative radiation for Hodgkin's disease:

1. Patients must be carefully and accurately staged, in most cases including laparotomy and splenectomy, if radiation is to be used as sole treatment.
2. Treatment must include, in addition to the areas of known involvement, "prophylactic" treatment to extended fields including all nodal areas likely to be subclinically involved. ^[124]
3. Techniques to ensure dose homogeneity and appropriate organ shielding must be employed, including (a) individualized, or custom-tailored, radiation therapy portals, (b) large treatment portals, and (c) equally weighted opposed AP portals.

Based on retrospective evaluation of kilovoltage treatment data from Peters, ^[150] Kaplan concluded that cure requires a tumor dose of 40 to 44 Gy to clinically and subclinically involved nodal areas, ^[151] doses that could not be practically delivered until the development of the linear accelerator. Based on analysis of dose-response data from the megavoltage era, most radiation oncologists now believe these doses to be unnecessarily high, especially to nonbulky and subclinical areas. ^[152] Furthermore, as will be discussed in more detail below, recent treatment programs are de-emphasizing surgical staging as well as extended field radiotherapy for treatment of early-stage disease.

The techniques of radiation therapy for pathologic stage (PS) I and IIA Hodgkin's disease have been reviewed in considerable detail, ^{[3] [147] [153] [154]} and will only be discussed in general terms here. "Standard" treatment fields for pathologically staged supradiaphragmatic disease is mantle and para-aortic fields plus splenic pedicle, referred to as STNI (Fig. 90-2) . A unilateral preauricular field is added for patients with preauricular involvement or with high cervical involvement. In addition, Waldeyer's ring is treated for patients with bulky high cervical involvement. There are some "favorable" subsets of patients with PS IA and IIA disease, discussed further below, who can be safely treated with mantle field irradiation alone. Treatment of CS II disease and most cases of CS I disease with radiation therapy alone requires irradiation of the entire spleen volume. Patients thus treated are rendered functionally asplenic, with the same long-term risks encountered by splenectomized patients; in addition, the apex of the left ventricle and the base of the left lung receive considerable scatter radiation dose.

Most patients with subdiaphragmatic stage I and II disease are treated to para-aortic/splenic pedicle fields plus pelvis fields, with the iliac and inguinal regions treated with an "inverted Y," as demonstrated in Figure 90-2 . When the patient is clinically staged, the spleen must be irradiated. Patients presenting with CS I inguinal and/or femoral adenopathy can be treated with inverted

Figure 90-2 Total nodal irradiation of the mantle, spade, and inverted Y fields. (From Kaplan HS: Hodgkin's Disease, 2nd ed. Cambridge, Harvard University Press, 1980, with permission.)

2633

Chemotherapy

There are several excellent chemotherapy regimens now available, each with apparently similar anticancer efficacy if delivered in optimal doses and schedules, but with considerable differences in patterns of both early and late toxicity. Both the oncologist's familiarity with a regimen and a consideration of the potential toxicities should enter into the decision of which regimen to choose. Despite advances in salvage therapy for relapsed Hodgkin's disease, we believe that the best opportunity a patient has of being cured is afforded by the initial treatment regimen, and any modification of that regimen for any reason other than as necessitated by severe toxicity already manifested may deprive the patient of a chance for cure, and is thus indefensible. We have thus formulated a set of principles that we believe are broadly applicable to the use of any chemotherapeutic regimen that has legitimate claim to being potentially curative for Hodgkin's disease ^[157] :

CHOICE OF CHEMOTHERAPY FOR TREATMENT OF HODGKIN'S DISEASE Several regimens afford similar therapeutic efficacy. Choice of chemotherapty regimens may be based largely on patterns of expected toxicities. If fertility is the overriding issue, ABVD is the treatment of choice. However, is the risk of fatal pulmonary toxicity from ABVD is comparable to the risk of leukemia from MOPP-like regimens. If fertility is not a crucial concern, MOPP/ ABVD or MOPP-ABV hybrid therapy is the treatment of choice. The efficacy is the same as four-drug regimens, and the lower total doses of alkylators bleomycin, and doxorubicin may lower the risk of chronic toxicity.

1. Do not modify planned doses or schedules of chemotherapy in anticipation of toxicity that has not yet happened, nor for short-term, non-life-threatening toxicity, such as emesis. Although the ability to deliver full doses can be compromised by the patient's intolerance of the drugs' toxicities, which in turn may be influenced by age and other variables associated with poor prognosis, every effort should be made to deliver maximum tolerated doses.
2. Consider dose escalation of the myelotoxic agents in subsequent cycle for patients who experience minimal neutrophil (i.e., no granulocyte count below 1,500/mm³ ) and platelet (i.e., no platelet count below 100,000/mm³ ) nadirs in the first cycle of chemotherapy.
3. In order to identify slow responders or nonresponders, reassess tumor status with each cycle of chemotherapy with a careful physical exam, a chest radiograph for patients with thoracic disease, and a KUB for those with abnormal lymphangiograms. Tumor progression at any time during treatment should prompt switching to a non-cross-resistant regimen, with plans to proceed to high-dose salvage therapy if possible.
4. Although a minimum of six cycles of all currently used chemotherapy regimens is believed to be necessary, slow responders may require more than six. We recommend treating for two additional cycles following a complete remission or after identification of stable, nonresponding residual masses, usually to a maximum of eight cycles. In our experience, six cycles of therapy has been sufficient in 85 percent of patients and 15 percent have required eight cycles. Thus, complete noninvasive reassessment of all sites of disease after four cycles of chemotherapy is required for comparison with the studies obtained after the sixth cycle. There seems little justification for treating with the same regimen for greater than eight cycles. ^[157] Additional cycles beyond eight for patients with residual tumor masses that seem to be shrinking when reassessed after the eighth cycle are either futile or unnecessary (though it may not be possible at the time to discern which) and will only increase morbidity without

   ---

   2634

   added benefit, and may delay switching to a more effective therapy for those patients with active residual disease.
5. Residual masses following treatment, whether stable for two cycles or still seemingly responding, require careful assessment. This may consist of biopsy, gallium scanning, ^{[142] [143]} or watchful waiting with radiographic reassessment at frequent intervals. Evidence of disease activity will prompt alternative therapy, either radiation therapy to residual masses, conventional dose salvage chemotherapy, or high-dose therapy with stem-cell rescue.

MOPP and Derivatives

MOPP (Table 90-4) was the first successful combination drug regimen for the treatment of Hodgkin's disease. The regimen was designed on the principles of combining drugs with independent antitumor activity and nonoverlapping organ toxicity, delivery of all drugs in full therapeutic dose with a fixed schedule of administration and a sliding scale of dose adjustment based on marrow toxicity, and cyclic delivery to allow normal tissue recovery. The results of long-time follow-up of 188 patients treated at the NCI with MOPP between 1964 and 1976 was updated in 1986 ^{[159] [160] [161]} : overall CR rate was 84 percent (stage II, 100 percent CR; stage III, 82 percent CR; stage IV, 77 percent CR; no B symptoms, 100 percent CR; with B symptoms, 78 percent CR). The overall survival was 48 percent at 19 years, virtually plateaued since the 10th year of follow-up, and the 15-year disease-free survival of the overall cohort was 54 percent. Patients free of Hodgkin's disease comprised 30 percent of the deaths (30 of 98). Thus, the majority of these patients with far-advanced disease and B symptoms were cured of Hodgkin's disease.

Similar success with the MOPP regimen was subsequently reported by several other groups, ^{[162] [163] [164]} and MOPP rapidly became the accepted treatment for advanced-stage Hodgkin's disease. The impact of MOPP and subsequent multidrug regimens on the outlook for Hodgkin's disease has been dramatic. In the first 20

TABLE 90-4 -- CHEMOTHERAPY REGIMENS FOR INITIAL TREATMENT OF HODGKIN'S DISEASE: REGIMENS FOR ADVANCED DISEASE

DRUGS	DOSE	SCHEDULE	CYCLE LENGTH	COMMENTS
MOPP
Nitrogen mustard   Vincristine   Procarbazine   Prednisone	6 mg/m2 1.4 mg/m2 (no cap) 100 mg/m2 40 mg/m2	IV days 1 & 8 IV days 1 & 8 PO days 1-14 PO days 1-14	4 wk	6 cycles min 8 cycles max
MVPP			6 wk	6 cycles
Nitrogen mustard   Vinblastine   Procarbazine   Prednisolone	6 mg/m2 6 mg/m2 100 mg/m2 40 mg	IV days 1 & 8 IV days 1 & 8 PO days 1-15 PO days 1-15
ChlVPP
Chlorambucil   Vinblastine   Procarbazine   Prednisolone	6 mg/m2 (10 mg max) 6 mg/m2 (10 mg max) 100 mg/m2 40 mg	PO days 1-14 IV days 1 & 8 PO days 1-14 PO days 1-14	4 wk	6 cycles min
ABVD
Doxorubicin   Bleomycin   Vinblastine	25 mg/m2 10 U/m2 6 mg/m2 375 mg/m2	IV days 1 & 15 IV days 1 & 15 IV days 1 & 15 IV days 1 & 15	4 wk	6 cycles min
Alternating MOPP/ABVD				4-wk MOPP cycle alternating with 4-wk ABVD cycle
MOPP/ABV hybrid			4 wk	8 cycles total
Nitrogen mustard   Vincristine   Procarbazine   Doxorubicin   Vinblastine   Vincristine   Bleomycin	6 mg/m2 1.4 mg/m2 (max 2.0 mg) 100 mg/m2 40 mg/m2 35 mg/m2 6 mg/m2 10 U/m2	IV day 1 IV day 1 PO days 1-7 PO days 1-14 IV day 8 IV day 8 IV day 8		Reassess after 6 cycles   If CR, 2 more cycles   If PR, give involved field radiation to residual disease, then 2 more cycles chemo
MA/MA hybrid			4 wk	6-8 cycles, then involved field XRT to residual disease
Nitrogen mustard   Vincristine   Procarbazine   Prednisone   Doxorubicin   Vinblastine   Bleomycin   Decarbazine	6 mg/m2 1.4 mg/m2 (max 2.0 mg) 100 mg/m2 40 mg/m2 35 mg/m2 6 mg/m2 10 U/m2 375 mg/m2	IV day 1 IV day 1 PO days 1-7 PO days 1-7 IV day 15 IV day 15 IV day 15 IV day 15

2635

Unfortunately, toxicity of MOPP proved to be significant. Long-term toxicities of MOPP include sterility and secondary myeloid malignancies, including acute myeloid leukemia and myelodysplastic syndrome. Virtually all men are rendered permanently azoospermic following six cycles of MOPP and over half of women over 26 years old become amenorrheic. Thirteen cases of acute leukemia and two cases of aplastic anemia were observed in the NCI series. No cases of leukemia occurred later than 9 years after the completion of therapy and all cases but one occurred in patients who received radiation therapy in addition to MOPP. Many reports have confirmed increased risk for myeloid malignancies occurring 2 to 10 years after treatment with MOPP or other alkylator-based regimens, with a risk of 2 to 10 percent at 10 years. The risk returns to that of the normal population by 10 years after therapy. Radiation therapy adds to the risk of myeloid leukemia in relation to the size of the fields. Total or subtotal nodal radiation approximately doubles the leukemia risk when added to chemotherapy; small field radiation therapy appears to add little to the risk. The risk for both sterility and myeloid malignancies is related to the cumulative dose of alkylator and procarbazine. The lifetime risk of acute leukemia and myelodysplasia from MOPP alone is 2 to 3 percent.

The acute toxicities include reversible bone marrow suppression in nearly all patients, nausea and vomiting, and largely preventable but painful phlebitis at the site of intravenous injection of nitrogen mustard. Peripheral or autonomic neuropathies from vincristine, manifested by paresthesias, weakness, foot drop, and obstipation, are dose-related, but are usually not debilitating. Vincristine neuropathy is usually completely reversible. In the 1970s, the late toxicities of MOPP (i.e., sterility and myelogenous leukemia/myelodysplastic syndromes) were only little appreciated, and the primary focus of many clinical researchers was on decreasing the rate of acute toxicities to MOPP, which were perceived to be the main drawback to routine use of the regimen. It should be emphasized that the limiting acute toxicities of the 1970s may be minor annoyances or less in the era of indwelling catheters and ports, hematopoietic growth factors, and serotonin 5-HT3 receptor antagonist antiemetics.

Many groups introduced modifications of the MOPP program by addition, deletion, or substitution of one or more drugs, yielding alkylator-based regimens that were felt to be easier to tolerate and/or to deliver. Of these MOPP derivatives, MVPP and ChlVPP (chlorambucil, vinblastine, procarbazine, and prednisone) have become the most widely used. MVPP (nitrogen mustard 6 mg/m² , with vinblastine 6 mg/m² substituted for vincristine, and prednisolone used in all six cycles, with cycles lengthened from 4 to 6 weeks) ^{[166] [167] [168] [169]} affords CR rates and relapse-free rates similar to those of MOPP and is associated with only insignificant neurotoxicity; emesis is comparable to and marrow toxicity worse than with MOPP. The ChlVPP regimen was developed as a largely oral regimen with fewer short-term side effects than are seen with either MOPP or MVPP. ^[170] Oral chlorambucil was substituted for nitrogen mustard in the MVPP regimen in order to decrease vomiting and the risk of nitrogen mustard-associated phlebitis, and a shorter cycle length of 28 days was used. Extensive experience, both in England ^{[171] [172] [173] [174]} and in the United States, ^[175] has shown that ChlVPP possesses efficacy comparable to that of MOPP. The results of ChlVPP in 960 patients with early and advanced disease has recently been summarized. ^[176] Although acute toxicities are indeed substantially lower than with MOPP, the frequency of both sterility and secondary myeloid malignancies following ChlVPP is similar to that following MOPP.

In summary, 80 percent of patients with advanced-stage Hodgkin's disease with constitutional symptoms treated with MOPP or one of its alkylator-based variants can be expected to enter complete remission; of these, one third will relapse. Since about one third of the patients who relapse can be salvaged, approximately two thirds of advanced-stage patients with symptoms will be cured. The long-term results are influenced by the tumor burden and accompanying constitutional symptoms, age, and performance status. All of these regimens are associated with a high rate of gonadal damage, and the risks of second myeloid malignancies are approximately the same as with MOPP.

Several retrospective studies have suggested adverse prognostic consequences of decreased dose intensity of one or more of the agents in the MOPP regimen. ^{[161] [177] [178] [179] [180] [181] [182] [183] [184] [185]} It is probable that similar dose-response relationships hold for other regimens as well, and we remain emphatic in recommending that dose modifications of protocol doses be made only for toxicities already experienced, using sliding scale modifications as recommended by the designers.

ABVD and Hybrid Regimens

The ABVD (doxorubicin, bleomycin, vinblastine, and dacarbazine) regimen (Table 90-4) was designed as a non-cross-resistant regimen for the treatment of patients relapsing after treatment with MOPP, and was initially shown to be effective in this setting. ^{[186] [187] [188]} ABVD was then combined in alternating cycles with MOPP for previously untreated patients with advanced disease in trials beginning in 1974; although ABVD alone was not prospectively evaluated in untreated patients for several years thereafter, it has now become clear that ABVD is an effective regimen for initial treatment of advanced Hodgkin's disease, and in many settings it may be the preferred regimen. By contrast, it is not terribly effective when used as salvage therapy for MOPP-resistant patients. It has a different spectrum of toxicities from the alkylator-based regimens, with virtually no gonadal toxicity or risk for myeloid malignancies. Both acute and long-term pulmonary toxicity due to bleomycin are not infrequent and may be severe;

2636

Several mature randomized comparisons of MOPP or related regimens with ABVD or combinations of MOPP and ABVD (either with cycles of MOPP alternating with cycles of ABVD or hybrid regimens in which all or most of the drugs of both regimens are delivered in each 4-week treatment cycle) have now been reported. Comparison between 12 cycles of MOPP versus six cycles of MOPP alternating with six cycles of ABVD in stage IV Hodgkin's disease was begun in 1974 in Milan, and updated 10-year results have been reported. ^{[189] [190] [191]} Complete remission rates were 74 percent and 89 percent; the relapse-free survivals were 46 percent and 68 percent; and the 10-year overall survival rates were 58 percent and 69 percent for MOPP and the MOPP-ABVD alternating regimens, respectively. The differences between the relapse-free survivals reached statistical significance favoring MOPP-ABVD, but the differences in overall survival were not significant. An Eastern Cooperative Oncology Group (ECOG) study confirmed the effectiveness of alternating cycles of MOPP and ABVD in a comparison with BCVPP (carmustine, cyclophosphamide, vinblastine, procarbazine, and prednisone). ^[192] A European Organization for Research in Cancer Therapy (EORTC) trial in stage IIIB and IV compared eight cycles of MOPP with two cycles MOPP alternating with two cycles of ABVD for a total of four 2-month cycles of both regimens. ^[193] Radiation therapy was given to sites of residual disease for patients with partial responses. Surprisingly low CR rates were observed in both arms (57 percent vs. 59 percent) and no significant differences were seen in relapse-free survival or overall survival.

A Cancer and Acute Leukemia Group B (CALGB) study reported in 1992 allows more detailed analysis of these regimens ^[194] : 361 patients with stage III and IV Hodgkin's disease or relapsing after previous radiation therapy were randomized among 12 cycles of alternating MOPP-ABVD, MOPP alone, and ABVD alone (the four-drug regimens were given for six to eight cycles). The CR rates for both ABVD and MOPP-ABVD were significantly higher than the CR rate for MOPP alone: 67 percent for MOPP, 82 percent for ABVD, and 83 percent for MOPP-ABVD ( P = 0.006 for MOPP vs. the other regimens). Significant differences were also observed between the failure-free survivals (FFS) measured at 5 years: 50 percent for MOPP, 61 percent for ABVD, and 65 percent for MOPP-ABVD ( P = 0.02). On the other hand, the differences between the overall survivals (66 percent, 73 percent, and 75 percent for MOPP, ABVD, and MOPP-ABVD, respectively) were not significant. Subset analysis suggested that the significant difference in FFS between MOPP and the two ABVD-containing regimens could be accounted for entirely by a subset of older patients with two or more extranodal sites of disease (43 percent vs. 62 percent FFS for MOPP and ABVD, respectively). The FFS rates in a low-risk group (<50 years with stage III or stage IV with only one extranodal site) were similar with the three regimens. MOPP was reported to cause substantially greater acute myelotoxicity than did ABVD, particularly in older patients, which resulted in progressive dose reductions and treatment delays in patients treated with MOPP. Only 24 percent of all patients received at least 85 percent of the planned dose of nitrogen mustard by the final cycle, whereas doxorubicin was delivered in near full doses to most patients receiving ABVD. Of serious concern was the 3 percent fatality rate from pulmonary complications on the ABVD arm. The risk of death from ABVD is thus similar to the risk of death from MOPP.

Two regimens were designed to overcome ab initio tumor cell resistance, utilizing the assumptions of the Goldie-Coldman mathematical model of tumor cell resistance, which predicted that a hybrid regimen using all the drugs in two effective and non-cross-resistant regimens in each and every cycle (i.e., use of all of the effective drugs in a single cycle) would minimize emergence of resistant clones. ^{[195] [196]} The Milan group's hybrid regimen, termed "MA/MA," consisted of a half-cycle of MOPP on Day 1 and a half-cycle of ABVD on Day 15 of each 28-day cycle. ^{[190] [197] [198]} The Vancouver MOPP-ABV hybrid regimen differs from MA/MA in that (1) the highly emetogenic agent dacarbazine was eliminated, (2) the dose of doxorubicin was increased from 25 to 35 mg/m² per cycle, (3) the ABV portion is given on Day 8 instead of Day 15, and (4) eight cycles are given instead of six, with involved field radiation therapy administered to sites of residual disease after six cycles of chemotherapy. ^{[158] [199] [200] [201]} An uncontrolled trial of this regimen in 170 previously untreated patients with advanced Hodgkin's disease suggested that this regimen is similar to both ABVD and alternating MOPP-ABVD in efficacy. ^[201] The complete remission rate was 84 percent, the projected 7-year overall survival was 80 percent, and the projected 7-year FFS was 65 percent. A comparison of both sequences of MOPP and ABVD MA/MA hybrid versus alternating MOPP/ABVD was conducted by NCI-Milan. ^[198] Patients with stages IB, bulky IIA, IIB, IIIA and B, and IV were randomized. Radiation therapy was given following chemotherapy to sites of pretreatment bulky disease. No significant differences in CR rate (91 percent vs. 89 percent), 10-year FFP (67 percent vs. 69 percent), or 10-year overall survival (74 percent vs. 72 percent) between alternating MOPP/ABVD and MA/MA hybrid were detected. Similarly, a large trial conducted by NCI-Canada comparing MOPP/ABV hybrid with alternating MOPP/ABVD in 301 patients with stage IIIB or IV Hodgkin's disease and patients relapsed after previous wide-field irradiation detected no differences in CR rate, 5-year FFS, or 5-year overall survival between the two regimens. ^[202] An Intergroup trial comparing MOPP-ABV hybrid with ABVD alone demonstrated insignificant differences in CR rates, FFS, and overall survival. ^[203] These three trials suggest that ABVD, MA/MA hybrid, MOPP/ABV hybrid, and alternating MOPP-ABVD are essentially equivalent in treatment efficacy. It would further appear from these trials that Goldie-Coldman predictions have not led to the development of a more effective therapy in Hodgkin's

2637

As expected, long-term toxicities following MOPP or ABVD are substantially different. The gonadal toxicity and risk for myeloid leukemia subsequent to MOPP therapy are significant and discussed in greater detail below. In contrast, ABVD alone results in only transient germ-cell toxicity in most males ^{[204] [205] [206] [207] [208]} and the alternating regimen MOPP/ABVD appears also to result in about half the incidence of sterility seen with MOPP. ^[207] The incidence of second myeloid malignancy after ABVD is virtually nil, as previously noted; in fact, only four cases have been reported in patients receiving ABVD alone. Although AML/myelodysplasia would be expected to be lower in patients treated with MOPP/ABV than with MOPP alone because of the lower exposure to nitrogen mustard and procarbazine, six cases among patients treated with the hybrid regimen were reported in the first Intergroup trial of MOPP/ABV versus ABVD. ^[203] In fact, this trial was terminated early by the Data Safety and Monitoring Committee because of excessive toxicity in the MOPP/ABV arm, with significant increases in life-threatening or lethal neutropenia, thrombocytopenia, infection, pulmonary toxicity, and second malignancies.

In all clinical trials of ABVD, significant and sometimes life-threatening acute pulmonary toxicity is reported. In the CALGB study, ^{[194] [209]} ABVD was associated with acute fatal pulmonary toxicity in four patients, two late lung carcinomas, and fatal cardiac failure in one patient in complete remission, accounting for 7 of the 12 deaths in patients without evidence of Hodgkin's disease. It would be anticipated that alternating MOPP/ABVD or hybrid MOPP/ABV would be associated with less pulmonary toxicity than ABVD, since cumulative doses of bleomycin are lower. However, as noted above, in the Intergroup trial of MOPP/ABV versus ABVD, pulmonary toxicity was greater in the MOPP/ABV arm. The problem of synergistic heart and lung toxicity in patients treated with both ABVD and radiation therapy is likely to be much more serious, but has yet to be fully defined.

Summary

MOPP and most of the MOPP-derived alkylator-containing regimens, such as ChlVPP or MVPP, are effective therapy for advanced Hodgkin's disease, and cure the majority of patients when delivered in optimal therapeutic doses (i.e., doses at the maximum tolerable levels, as defined by the observed toxicity in each individual patient without ad hoc dose modifications). Gonadal toxicity and induction of secondary myeloid malignancy are expected to be similar for all alkylator-containing regimens. Controlled trials in advanced Hodgkin's disease have demonstrated that ABVD, alternating MOPP/ABVD, MA/MA hybrid, and MOPP/ABV hybrid regimens are virtually equal in therapeutic efficacy. All four ABV(D)-containing regimens have fairly consistently been shown to give modestly improved CR rates and freedom-from-relapse rates in comparison with MOPP alone, with modest or statistically insignificant improvements in overall survival. The advantage of the ABV(D)-containing regimens over MOPP appears to be limited to patients older than 50 years of age. It must be noted that in each of the randomized trials, MOPP dosing has been systematically attenuated from the originally reported NCI regimen, in a manner expected to reduce efficacy.

Both early and late toxicities of ABVD are different from the toxicities of MOPP. The myeloid toxicity of ABVD appears to be more tolerable, especially in older patients, than that of MOPP, allowing less dose modification. Neither gonadal toxicity nor the risk for second leukemias is significant following ABVD. On the other hand, morbidity and mortality from cardiac and pulmonary toxicities of doxorubicin and bleomycin, respectively, particularly when mediastinal radiation therapy is used, may be substantial. As has been noted, 15- to 20-year follow-up data are necessary to appreciate fully the late toxicities of any regimen, data that have unfortunately not been reported for patients treated with ABVD alone. Since late toxicities are related to cumulative doses of the offending drugs, the seven- or eight-drug regimens appear to mitigate each of the principal toxicities associated with both MOPP and ABVD, while exposing the patient to some potential risk for all of them. It is unclear at this time where the balance lies, but it appears that, based largely on the toxicity profile, ABVD may have an overall edge as an initial treatment of choice for most patients with advanced Hodgkin's disease. Clearly, if combined-modality therapy including mantle field irradiation is planned, alternating MOPP-ABVD or MOPP-ABV hybrid is preferable to ABVD, since cumulative doses of bleomycin and doxorubicin are lower and interactive toxicities between the drugs and radiation therapy would likely be less frequent and/or less severe.

Therapy for Specific Presentations

Supradiaphragmatic Stages I and II

Radiation therapy

Radiation therapy is generally considered the preferred therapy for most patients with stage IA and IIA supradiaphragmatic Hodgkin's disease. ^{[146] [210] [211] [212] [213] [214] [215]} The radiation therapy techniques for treatment of PS I to II disease developed at Stanford, and widely adapted in North America and elsewhere, are summarized in Table 90-5 . ^[147] Patients with supradiaphragmatic PS IA and IIA are treated with subtotal lymphoid irradiation (STLI); whereas TLI has been used for patients with "B" symptoms. When patients are treated in this manner, complete remission rates approaching 100 percent are expected,

2638

Although the staging laparotomy is being used with decreasing frequency, the lessons about the clinical course of Hodgkin's disease taught by the practice of meticulous surgical staging must not be forgotten. Abdominal CT scanning has only about 50 percent accuracy in predicting splenic involvement. ^{[135] [136]} One third of patients with enlarged or abnormal-appearing spleens are found at splenectomy to have no splenic involvement and, conversely, about one third of patients without CT abnormalities in the spleen or elsewhere below the diaphragm will be upstaged by laparotomy. There are prognostic factors in a subset of CS IA patients that predict an acceptably low likelihood (i.e., <10 percent) of finding subdiaphragmatic disease. ^{[217] [218]} These groups include (1) all women with CS IA, (2) men with nonbulky mediastinal disease, (3) men with lymphocyte-predominant histology, and (4) men with cervical disease above the hyoid bone. Limited radiation therapy without staging laparotomy and without chemotherapy can be recommended for these patients. The remaining men with CS IA disease and all patients with CS IIA (80 percent of all CS I to II patients) have approximately 30 to 40 percent risk of

2639

Certain prognostic factors have been reported to predict an increased probability of relapse after treatment with radiation therapy alone. Most significant of these have been the presence of bulky masses and constitutional symptoms. An analysis of 180 patients with PS IB and IIB disease treated at Stanford and at the Joint Center for Radiation Therapy revealed that in these symptomatic patients, the most significant adverse characteristic at initial presentation was the presence of both fevers and weight loss. ^[220] Patients with both fevers and weight loss had a 52 percent 7-year relapse rate and only 57 percent 7-year actuarial survival, which was worse than for patients with massive mediastinal disease treated with radiation therapy. The relapse rate for patients with fever only was 31 percent, and was 19 percent for patients with weight loss only. The presence of night sweats was not an adverse prognostic factor. Patients with PS IB or IIB disease on the basis of night sweats only had the same freedom-from-relapse rates and survivals as patients with PS IA or IIA disease, respectively, and the presence of night sweats in addition to either fever or weight loss or both did not worsen the prognoses compared with similar patients lacking night sweats. Prognostic factors reported in other studies, ^{[221] [222]} but not confirmed by the analysis of the combined Stanford-Joint Center cohorts, included gender (male is worse), histologic subtype, number of sites of involvement, and the presence of extranodal disease. Clearly, tumor burden is the most important adverse factor, ^[223] and most prognostic indicators serve as surrogate markers for tumor bulk. Age has only modest influence on relapse rate but has a significant influence on survival, due to decreased effectiveness of salvage therapy in the older age group. ^{[212] [221]}

Combined-modality therapy

About 40 percent of the remainder of clinical early-stage patients (i.e., all patients with CS II disease and men with CS I lacking the above favorable characteristics) will have occult disease below the diaphragm. Whether or not pathologic staging is necessary in these patients is the key question driving many recent trials in early stage disease. Recent studies have focused on the treatment of clinically staged patients with radiation therapy alone, chemotherapy alone, or combined-modality therapy. More than 20 randomized trials comparing radiation therapy with combined chemotherapy plus radiation therapy have been conducted. Although improvements in relapse rates in early-stage patients treated with combined-modality therapy are easily demonstrable, no trial has yet shown superior survival for combined-modality therapy in patients with PS I and II disease, even among subsets with adverse prognostic features at presentation, as compared with radiation therapy alone. ^{[138] [213] [224] [225] [226]} A meta-analysis of 23 reports with 2,999 early-stage patients did not demonstrate survival differences between patients treated with extended field radiation and those treated with combined-modality therapy. ^[210] Nevertheless, many clinicians still believe that avoidance of relapse should be paramount and continue to recommend combined-modality therapy, in part to avoid the psychological trauma for patients who relapse and must then be treated with salvage therapy. There would be little controversy about this practice were it not for the additive, or even synergistic, toxicity of radiation and most chemotherapy regimens currently known to be effective, including the possibly increased risk of secondary myeloid malignancy with alkylator-based regimens and the cardiac and pulmonary toxicities associated with doxorubicin and bleomycin.

Several strategies have been explored in efforts to circumvent the late toxicities observed following extended field radiation combined with chemotherapy, including programs utilizing smaller radiation treatment fields, lower radiation doses, and less toxic chemotherapy regimens. Included in the latter are programs utilizing decreased numbers of cycles of standard chemotherapy regimens and less aggressive

2640

Another approach to reducing the toxicity of combined-modality therapy is to deliver only modest doses of involved field radiation therapy (15 to 30 Gy) following full courses of combination chemotherapy. This approach has proved effective in children, in whom avoidance of both splenectomy and high-dose radiation therapy to growing bones is mandatory because of the risks of postsplenectomy sepsis and cessation of bone growth, respectively. ^{[231] [232]} Furthermore, with their longer overall survival, children treated with radiation therapy have a greater cumulative risk of second solid tumors. However, it has not been shown that radiation therapy is necessary in either children or adults treated with effective chemotherapy. Recently the Pediatric Oncology Group (POG) has shown conclusively that low-dose radiation does not improve the results in children with stages IIB through IV disease following alternating MOPP-ABVD. ^[233]

In summary, combined-modality therapy has not been shown to be superior to radiation therapy alone in early-stage patients, nor has it been demonstrated to be superior to chemotherapy alone in advanced-stage patients, except those with bulky tumor masses. ^[234] Advocates of combined-modality therapy in adults cite data largely from uncontrolled trials ^{[148] [235] [236]} or from flawed randomized studies ^[237] in advanced Hodgkin's disease to suggest superior results. We have cautioned against the routine use of this practice outside the clinical trial setting in advanced-stage patients because of the risk of second cancers and fatal heart attacks associated with radiation therapy. ^[234]

Chemotherapy alone in early stage HD

A chemotherapy-only approach to the treatment of early-stage Hodgkin's disease avoids both staging laparotomy and the toxicities of combined-modality therapy. Since chemotherapy is more effective in treating low-bulk than high-bulk disease in advanced-stage patients, it would seem unlikely a priori that it would be less effective in early-stage than in advanced Hodgkin's disease, or that optimal results in early-stage disease require even more intensive therapy (i.e., combined-modality therapy) than is required to cure advancedstage disease. Chemotherapy is effective in children with early-stage disease ^{[174] [238] [239]} and has been shown in several randomized trials to be equivalent to combined-modality therapy in adults with stage I to IIIA disease. ^{[167] [240] [241]}

Two prospective randomized trials have compared chemotherapy alone with radiation therapy alone for early-stage adult patients. ^{[242] [243] [244]} Both studies randomly assigned surgically staged patients to receive either six cycles of MOPP chemotherapy or standard radiation therapy to mantle and para-aortic fields. However, the trials differed in their inclusion criteria for randomization, and have come to different conclusions. At the NCI, 110 patients with central PS IA and IB (thoracic and abdominal only), IIA, IIB, and PS III₁ A were randomized between MOPP and STLI. ^{[149] [242]} Patients with peripheral IA (disease above the clavicles or below the inguinal ligament) were treated with radiation therapy, either to a minimantle or an inverted Y, without randomization. After the trial was begun, it became clear that radiation therapy alone was inappropriate therapy for patients with massive mediastinal disease and patients with PS III₁ A disease; these patients were not entered after the first 5 years of patient accrual. The CR rates were 96 percent for both approaches. When all randomized patients were analyzed, statistically significant improvements in both 17-year disease-free survival (85 percent vs. 64 percent) and overall survival at 17 years (93 percent vs. 76 percent) were seen for MOPP as compared with radiation. When these results were reanalyzed after exclusion of seven patients with massive mediastinal disease and 13 patients with PS III₁ A who had been randomized in the early years of the trial and who fared very poorly with radiation alone (leaving 86 randomized patients, of whom 79 were stage II), there was no difference in overall survival and a small but statistically insignificant advantage in disease-free survival (DFS) for patients treated with MOPP. From this analysis, we conclude that MOPP and radiation therapy afford equivalent DFS and overall survivals, at least for stage IIA and IIB patients. Since only seven patients with PS I were randomized, it is difficult to draw firm conclusions for these patients, although we believe it unlikely that they would fare less well with chemotherapy than do stage II patients.

In the Italian trial, all patients with PS IA and IIA were randomized between MOPP and extended field radiation therapy; patients with B symptoms were excluded. ^[244] The authors concluded that radiation therapy was superior. Both the CR rates and relapse rates were similar in the two arms; however, significantly better overall survival (93 percent vs. 56 percent at 8 years' median follow-up) was found in the patients randomized to radiation therapy. The poor survival in the

2641

Subdiaphragmatic Stage I or II

Subdiaphragmatic stage I and II Hodgkin's disease accounts for only 3 to 4 percent of all new cases of Hodgkin's disease. Eighty percent of these patients present with groin adenopathy (inguinal, femoral, or superficial iliac) and the remainder will present with disease confined to the abdomen. ^{[71] [246] [247] [248] [249] [250] [251] [252] [253] [254] [255] [256]} Patients presenting with abdominal Hodgkin's disease tend to be older, more likely to be male, and more likely to present with constitutional symptoms than those with either peripheral subdiaphragmatic or supradiaphragmatic presentations; mixed cellularity histology is more common among central abdominal presentations than are the other histologies. ^{[248] [249]} In a series of 473 patients of all stages from Stanford, four patients presented with fever of unknown origin and were found to have an abdominal mass that proved to be Hodgkin's disease. ^[70] Patients may also present with lower extremity edema due to lymphatic or venous obstruction. Hodgkin's disease has been diagnosed at splenectomy after presentations with splenomegaly, splenic rupture, or apparent splenic abscess. Extranodal small bowel, stomach, or large bowel presentations are quite rare.

Initial diagnosis of intra-abdominal disease will frequently be made at exploratory laparotomy; however, when the diagnosis is secure and abdominal disease is known to be present, laparotomy to define extent of abdominal disease would appear to contribute little additional information. For patients with femoral, inguinal, or superficial iliac presentations, abdominal CT scans and bipedal lymphangiography is adequate staging; staging laparotomy for peripheral subdiaphragmatic presentations would appear to have limited value, since treatment will usually not be altered by the results. A review of 76 patients with CS IA to IIB Hodgkin's disease with inguinal presentations who had staging laparotomies showed that laparotomy resulted in a change in stage in one third, but a clinically important upstaging to PS III or IV occurred in only 5 percent. ^[249] Only one of 21 patients with CS IA groin presentations with negative lymphangiograms was found at laparotomy to have disease outside the true pelvis (splenic involvement) and another was found to have a positive iliac node. Of 55 patients with CS IB, IIA, or IIB, six were down-staged from CS IIA to PS IA, three were upstaged to IVA on the basis of liver disease, and one was upstaged to IIIA (after a blind scalene biopsy). Splenic involvement was found in 28 percent.

Evaluation of treatment in these patients is difficult because of the small numbers and the variability in treatment strategies. Most patients reported in the literature have been treated with extended field subdiaphragmatic radiation therapy fields or total nodal radiation therapy with or without adjuvant chemotherapy. It appears from subset analysis of the small number of patients with subdiaphragmatic presentations who have been randomly allocated either to radiation therapy or combined-modality treatment programs that the prognoses of these patients are similar to those of supradiaphragmatic early-stage patients ^{[71] [248] [250] [257]} ; on the other hand, the prognosis for patients with intra-abdominal presentations appears to be worse than that for patients with either supradiaphragmatic presentations or peripheral subdiaphragmatic disease. ^[248] This could be associated with age of patients, the presence of bulky masses, or other factors. Krikorian et al. recommended clinical staging only, and treatment with inverted Y plus splenic radiation, for patients with CS IA disease presenting in the groin. ^[249] For patients with CS IB, IIA, and IIB inguinal presentations, this group recommended staging laparotomy and treatment with TNI including the mediastinum, or combined-modality therapy, for those patients with PS IB or IIA disease. We, on the other hand, believe that clinical staging should be adequate for all peripheral subdiaphragmatic presentations. The only finding at laparotomy resulting in upstaging is occult liver disease, which would be expected in fewer than 5 percent of patients with CS IIA. Inverted Y plus splenic irradiation is adequate therapy for patients with CS IA. Patients with CS IIA can be treated with TNI therapy, and combination chemotherapy alone for patients with CS IIA is also acceptable treatment. Patients with B symptoms have a high incidence of extensive splenic involvement and hepatic involvement and should be treated with combination chemotherapy. ^[250] Therapy for patients with large abdominal masses should be individualized, but should probably include both chemotherapy and involved-field radiation therapy of the mass.

2642

Stage IIIA

In this era, stage IIIA Hodgkin's disease should not be considered early-stage disease and should not be treated with radiation therapy alone. Although high CR rates (90 to 95 percent) are expected with TNI for stage IIIA disease, ^{[258] [259]} between one third and two thirds of patients can be expected to relapse within 5 years. ^{[260] [261]} Retrospective studies in laparotomy-staged patients with PS IIIA disease have identified two subsets with different prognoses following TNI, and these studies have been used to justify laparotomy staging of these patients and treatment with TNI alone in the better-prognosis subset. The first subgroup, termed PS III₁ A, has subdiaphragmatic disease limited to lymphatic structures that accompany the celiac axis group of arteries the spleen, splenic nodes, celiac nodes, and/or portal nodes (disease limited to above the renal hila). PS III₂ A involves the lower abdominal nodes para-aortic, iliac, or mesenteric nodes (disease below the renal hila), with or without involvement of the spleen or the upper abdominal nodes. Patients with PS III₁ A had a prognosis following radiation therapy similar to that of PS IIA patients, whereas the second subgroup, PS III₂ A, had a significantly higher relapse rate. ^{[260] [261] [262] [263] [264] [265] [266]} Relapse-free survivals at 5 years following TNI for PS III₁ A and PS III₂ A were 78 percent and 41 percent, respectively, and the overall 5-year survivals were 93 percent and 57 percent, respectively; these differences were significant. ^[261] Although TNI and combined-modality therapy afforded similar CR rates and relapse rates for patients with PS III₁ A,combined-modality therapy was significantly better than TNI for patients with PS III₂ A. ^[266] These trends were confirmed in a review of patients treated at four institutions. ^[265]

On the other hand, retrospective analysis with much longer follow-up duration of patients with PS III₁ A disease treated at the Joint Center for Radiation Therapy with either radiation alone (STLI or TNI) or combined-modality therapy demonstrates a clear benefit, not only in relapse-free survival, but also in overall survival, for patients treated with combined-modality therapy, strongly suggesting that radiation therapy alone is not sufficient therapy for either subclass of PS IIIA disease. ^[268] Furthermore, both uncontrolled ^[159] and controlled trials ^{[240] [269]} suggest that radiation therapy may not be necessary, since chemotherapy alone and combined radiation therapy plus chemotherapy yield similar results for PS IIIA. A randomized trial comparing MVPP with TNI demonstrated improved complete remission rate and remission duration with combination chemotherapy. ^[270] It should be noted that no randomized trial has demonstrated superior outcome for patients treated with combined-modality therapy over those treated with chemotherapy alone. We categorize stage IIIA disease as advanced-stage disease and recommend chemotherapy alone for these patients. As a corollary, laparotomy is never recommended for any CS III patient, since no germane information is obtained.

Massive Mediastinal Disease

The most significant adverse prognostic factor in otherwise early-stage Hodgkin's disease is massive mediastinal disease, and it is now clear that, even when disease is found only above the diaphragm, disease massively involving the mediastinum should be considered "advanced." Most authors have used the straightforward definition of Mauch et al.: a mediastinal mass is "massive" if the transverse diameter of the mediastinal nodal mass is greater than one third of the widest diameter of the thorax on a standing PA chest x-ray. ^[270] The revised staging classification schema adopted at the Cotswolds conference ^[125] recognized the poor prognosis of bulky disease by subclassifying all patients with bulky disease with a subscript "X" following the anatomic stage, and defined bulky mediastinal disease as greater than one third the diameter at T5-T6.

As many as one quarter to one third of all newly diagnosed patients with Hodgkin's disease and an estimated 15 to 20 percent of those with stage I and II will have massive mediastinal disease. Nodular sclerosis and mixed cellularity histologies comprise the bulk of patients with mediastinal disease; LPHD virtually never involves the mediastinum massively. Relapse rates for patients with stage II disease with large mediastinal masses treated with radiation therapy alone have ranged between 50 and 74 percent. ^{[213] [272]} Mauch et al. found that patients of all stages with massive mediastinal disease had higher relapse rates than comparably staged patients without massive mediastinal disease, when treated with radiation therapy; in fact, stage I or II patients with large mediastinal adenopathy had a higher relapse rate than did patients with stage IV intrathoracic disease with small mediastinal involvement. ^{[271] [273]} Half of PS IA and IIA patients with mediastinal masses greater than one third the chest diameter treated with TNI relapsed, while only 2 percent of the similarly treated patients without mediastinal masses and 12 percent of those with small mediastinal masses relapsed; 8-year actuarial survivals were 86 percent and 96 percent, respectively, for large and small mediastinal masses. Patients with stage IIE or stage IV massive mediastinal disease had a 64 percent relapse-free survival and 66 percent overall survival at 5 years, compared with 90 percent relapse-free and 90 percent overall survival among stage IV patients with small mediastinal masses. Most subsequent analyses have confirmed the increased relapse rate in radiation therapy-treated patients with large mediastinal masses, although not all have demonstrated significant survival differences. ^{[213] [272] [274] [275] [276] [277] [278] [279] [280] [281] [282] [283] [284] [285]} An assessment of the results in massive mediastinal disease treated with chemotherapy alone is difficult, since most of the comparatively few patients who have been treated without radiation therapy have had stage IV disease, while most of the patients treated with either radiation therapy alone or combined-modality therapy have had anatomically limited disease. Retrospective analysis of 49 patients with massive mediastinal disease treated at the NCI with MOPP suggest that chemotherapy alone is not adequate

2643

Because both radiation therapy alone and chemotherapy alone have proven to be unsatisfactory when patients have bulky mediastinal disease, most centers now routinely employ combined-modality therapy for this subset of patients, irrespective of anatomic stage, using one of the chemotherapy regimens described above for treatment of advanced disease, plus mantle or involved field radiation therapy. A randomized trial comparing radiation therapy alone and combined-modality therapy (MOPP plus mantle and para-aortic irradiation) demonstrated significantly improved 10-year freedom-from-relapse for patients treated with combined-modality therapy as compared with radiation therapy (81 percent vs. 45 percent), although differences in overall survivals were not significant. ^[213] Other centers have subsequently reported salutary results with a combined chemotherapy-radiation therapy approach to patients with large mediastinal masses. ^{[278] [280] [281] [282] [283] [285] [287]} The M.D. Anderson group has summarized its experience with combined-modality therapy for massive mediastinal Hodgkin's disease with four different chemotherapy regimens. ^[288] Complete remission rates (CR, 85 percent overall), 3-year freedom-from-progression (FFP, 76 percent overall), and 3-year freedom-from-tumor-mortality (FTM, 90 percent overall) were similar among the four regimens. The Stanford group has recently updated results of combined-modality therapy with the Stanford V regimen (doxorubicin, vinblastine, nitrogen mustard, vincristine, bleomycin, etoposide, and prednisone administered over 12 weeks) followed by radiation therapy to a dose of 36 Gy to bulky sites only. ^{[230] [289]} Forty-nine patients with massive mediastinal disease were included among a larger group including patients with stages IIIB and IV, and although the results for massive mediastinal patients were not reported separately, the overall results were excellent: the 16-year actuarial overall survival and FFP were 93 percent and 89 percent, respectively.

We have recently reported results of a prospective trial in 80 patients with massive mediastinal disease using alternating MOPP-ABVD chemotherapy followed by mantle field irradiation. ^[290] The CR rate was 89 percent, and with median follow-up duration of 10 years, the projected 15-year DFS for responders is 78 percent and overall survival at 15 years is 75 percent. Radiation treatment planning was performed before chemotherapy was begun and following chemotherapy, 10 Gy radiation to mantle field including the original full extent of disease was delivered, followed by cone-down treatment to residual disease to a total dose of 35 to 45 Gy. This regimen allows preservation of fertility more frequently and is associated with a lower risk for leukemia than with MOPP, and is associated with a lower incidence of pulmonary and/or cardiac toxicity than ABVD. Acute toxicities in this trial were modest, though 21 percent of patients had some pulmonary reaction within 3 months of completing therapy and 10 percent had cardiac symptoms. One patient, an older patient with a history of heavy cigarette smoking, died of lung toxicity. Three patients developed second malignancies.

Interestingly, no published studies have employed combination chemotherapy alone in the treatment of patients with early-stage disease accompanied by a large mediastinal mass. In our own early-stage Hodgkin's disease trial, eight patients with massive mediastinal disease were randomly assigned to receive MOPP alone before such patients were eliminated from the study. All eight achieved a complete response and remain alive and in their initial complete remission a median of 13 years after treatment. This experience is certainly not large enough to suggest that radiation therapy can be safely omitted in such patients. However, a large cooperative group study of the need for radiation therapy in stage I and II massive mediastinal Hodgkin's disease would be of interest.

Based on the above studies, we recommend treatment of massive mediastinal Hodgkin's disease with combined-modality therapy, irrespective of anatomic stage. Because laparotomy findings will not alter treatment decisions, and because patients with bulky mediastinal masses are at an increased risk of death during attempts to extubate following general anesthesia, laparotomy staging should never be done in patients with massive mediastinal HD. The alternating MOPP-ABVD regimen or the MOPP-ABV hybrid regimen should be preferable to either MOPP or ABVD in this setting since the long-term toxicities of both MOPP and ABVD may be increased by radiation exposure. Chemotherapy should precede radiation therapy, since shrinking the tumor mass before radiation allows significantly lower radiation dose to normal lung, and initial chemotherapy allows earlier treatment to unirradiated sites. Although a small dose of radiation was delivered to the original tumor volume in the NCI trial described above, ^[290] including a large volume of normal lung in the initial treatment field in many patients, it is not clear that this practice is necessary.

Stage III and IV Disease

The treatment for advanced-stage Hodgkin's disease has advanced very little in the last 20 years. The choice of chemotherapy regimen has been discussed above. ABVD alone, ChlVPP, MVPP or MOPP alone, MOPP alternating with ABVD, or one of the hybrid regimens MA/MA or MOPP/ABV are all reasonable choices. ABVD has pulmonary and cardiac toxicities as its only drawbacks

2644

Efforts to improve upon the outcome of chemotherapy alone in advanced-stage patients are based on the success of high-dose therapy and autologous stem cell support in the treatment of relapsed patients. It has been difficult to justify high-dose therapy as a primary treatment approach because no subsets of patients appeared to have a sufficiently poor prognosis with combination chemotherapy that high-dose therapy was warranted. Two approaches have been taken; development of a more effective conventional-dose regimen and development of a prognostic factor model to identify high-risk patients. Stanford V, alluded to above, consists of seven drugs given in alternating combinations weekly for 12 weeks. The BEACOPP (carmustine, etoposide, cytarabine, cyclophosphamide, vincristine, prednisone, and procarbazine) regimen ^[291] was dose-escalated by increasing the doses of etoposide, doxorubicin, and cyclophosphamide and administering together with granulocyte colony-stimulating factor (G-CSF) in 3-week cycles. Preliminary analysis of escalated BEACOPP versus COPP/ABVD appears to favor BEACOPP. The regimen might be even more effective if given to patients with advanced-stage disease with clinical features suggesting a poorer prognosis. Both the Stanford V and BEACOPP regimens were followed by involved field radiation therapy; there remains no controlled data to support that practice.

Recently, an international collaboration set out to define prognostic subsets within patients with advanced-stage disease treated with the most effective combination chemotherapy programs. Twenty-five centers from all over the world pooled their data on patients with advanced-stage disease for analysis. The result is a new International Prognostic Factor Index for advanced Hodgkin's disease. The following seven clinical factors emerged as having prognostic relevance: serum albumin less than 4 g/dl, hemoglobin less than 10.5 g/dl, male gender, stage IV disease, age older than 45 years, white blood cell count greater than 15,000/mm³ , and lymphocyte count less than 600/mm³ . Each factor exerted roughly an equivalent effect on freedom from progression, about 8 percent. The results are summarized in Table 90-7 . For the purposes of clinical studies, it is convenient to divide patients into two groups: patients with 0 to 2 factors comprise about 58 percent of all advanced-stage patients and have a freedom from progression at 5 years of about 74 percent; patients with three or more factors comprise about 42 percent of all advanced-stage patients and have a freedom from

TREATMENT OF ADVANCED-STAGE HODGKIN'S DISEASE Despite some retrospective analyses that support its use, patients with advanced-stage disease should not routinely receive radiation therapy because of the absence of data from randomized trials and the increased risks of second cancers and fatal heart attacks associated with the use of radiation therapy. However, about 10 percent of patients with advanced-stage disease do not completely respond to the chemotherapy; such patients may be converted to complete responders with the addition of radiation therapy.

progression at 5 years of 55 percent. Efforts to intensify the therapy for the latter group may well lead to improved outcome in advanced disease. ^[292]

Relapsed and Refractory Disease

Relapse after radiation

Chemotherapy has been shown to be excellent salvage therapy for the 20 to 25 percent of patients treated initially with radiation therapy who relapse. ^{[160] [292] [293] [294] [295]} It appears that any of the front-line chemotherapy regimens previously discussed is capable of curing patients who relapse after radiation therapy, with expected complete remission rates between 75 and 90 percent and long-term survival after relapse ranging between 45 and 70 percent. An analysis of prognostic factors in 109 patients treated with MOPP or MOPP-like chemotherapy at Stanford University after relapse from radiation therapy ^[297] suggests that extent of disease at the time of relapse, as measured by anatomic stage at relapse (termed relapse stage [RS]) is the most important factor predicting second CR, freedom from second relapse (FF2ndR), and overall survival after first relapse. For the overall group with median follow-up of 8.3 years, the 10-year FF2ndR was 57 percent. The 10-year FF2ndR rate for 26 patients with RS IA was 90 percent; for 34 patients with RS IIA and IIIA, 60 percent; and for 49 patients with IB, IIB, IIIB, and IV the FF2ndR rate was 30 percent. Age greater than 50 years was the only

TABLE 90-7 -- RADIATION FOR PS I-II:STANFORD RESULTS ^*

SUBGROUP	10-YEAR RELAPSE RATE (% )	5-YEAR SURVIVAL (% )	10-YEAR SURVIVAL ( % )
All PS I-II	23	96	84
PS I	0	100	--
PS II	25	95	--
PS IIA	25	97	--
PS IIB	17	92	--

*Data from Hoppe RT, Coleman CN, Cox RS, et al: The management of stage I-II Hodgkin's disease with irradiation alone or combined modality therapy: the Stanford experience. Blood 59:455, 1982.

2645

Relapse after chemotherapy

The likelihood of response to salvage therapy and long-term prognosis of patients relapsing after initial chemotherapy depend on, among other prognostic variables, the response to initial therapy and duration of the first remission, segregating those who are not cured by their initial therapy into three groups: (1) induction failures patients who never enter complete remission; (2) relapse less than 12 months after achieving complete remission; and (3) relapse greater than 12 months after achieving complete remission. Median survival from date of relapse for each group is 1.3, 2.6, and 4.3 years, respectively. ^[298] The last group of patients is recognized to have an excellent response to salvage chemotherapy, whether the same regimen as the first-line regimen or a different regimen is used, whereas the

TREATMENT OF RELAPSED AND REFRACTORY HODGKIN'S DISEASE Any Hodgkin's disease patient in whom cure is not achieved with the initial course of combination chemotherapy should be considered a potential candidate for high-dose therapy with hematopoietic support. Even in patients whose initial complete response was longer than 12 months, a group that is usually very responsive to conventional dose chemotherapy, high-dose therapy may result in better long-term survival than is obtained in such patients treated with conventional-dose chemotherapy alone.

first two groups of patients fare considerably less well. Analysis of the NCI experience with patients relapsing after chemotherapy suggests that even those patients whose initial remission was longer than 12 months and who achieve second complete remissions have a very low probability of being cured by conventional-dose salvage therapy. ^[298] Of a total of 439 patients treated with MOPP or a related regimen between 1964 and 1990, 107 relapsed after achieving complete remission (approximately half of these relapses occurred <1 year after first CR). The overall survival of all relapsed patients is projected to be 17 percent at 20 years. Both the probability of achieving a second complete response and the overall survival were predicted by the duration of the first remission. Response rate and survival were not significantly affected by initial stage or symptoms or initial chemotherapy regimen. Second CR rate for patients whose initial complete remission was longer than 12 months, most of whom were retreated with MOPP, was 79 percent; whereas for patients with initial remission durations less than 12 months (62 percent of whom were treated with regimens other than MOPP), the second remission rate was 49 percent. Second remissions were significantly more durable for patients with long initial remissions than for those with initial remissions less than 1 year: 45 percent actuarial 22-year disease-free survival for patients achieving second complete remission after a long initial remission compared with 14 percent 5-year disease-free survival for patients with first relapse after less than 1 year. Due largely to deaths from acute leukemias and other treatment-related complications, the overall survival was unfortunately only 24 percent for those patients with long first remissions and 11 percent for those whose first remissions were less than 12 months.

The ABVD regimen was designed specifically to treat MOPP-resistant patients. ^[186] An early report found a 59 percent CR rate in 55 patients with MOPP-resistant Hodgkin's disease treated with ABVD salvage therapy. ^[299] No correlation between response rate and duration of MOPP response was noted; in fact, complete remission was obtained in 52 percent of 29 patients who had progressive disease during MOPP chemotherapy. About two thirds of the responders relapsed, leading the authors to conclude that ABVD "can probably cure one third of complete responders and therefore about 20 percent of all MOPP-resistant patients." ^[299] Other groups employing ABVD or similar doxorubicin-containing salvage regimens have not been able to duplicate these results, with CR rates generally ranging between 10 and 45 percent. ^{[300] [301] [302] [303] [304] [305] [306] [307] [308] [309] [310]} Indeed, in the aforementioned CALGB study, patients randomly assigned to MOPP or ABVD were crossed over to the other four-drug regimen upon relapse; interestingly, MOPP was significantly more effective at rescuing ABVD treatment failures than was ABVD at rescuing MOPP treatment failures.

One retrospective analysis of the outcome of salvage therapy for first relapse following MOPP or a MOPP-like regimen suggests that the outcome is highly dependent on the presence of three risk factors symptoms

2646

Clearly, many relapsed patients have complete remissions and prolonged disease-free intervals with conventional salvage therapy. Re-treatment with MOPP or salvage with a doxorubicin-containing regimen is very good reinduction therapy for patients whose initial remission is greater than 1 year, while ABVD or one of its derivatives appears to be a better choice for patients whose first remission is less than 1 year (unless ABVD was part of the initial treatment). An alkylating agent-containing regimen (ChlVPP or MOPP) is effective in patients who relapse after ABVD treatment. Extended field radiation therapy may be considered for low-risk patients relapsing with limited disease at the time of relapse. However, even for the group of patients lacking putative risk factors, there is a low probability of cure with conventional salvage chemotherapy, and most of these patients will, at some point in the course of their subsequent treatment, be candidates for a high-dose salvage regimen. For patients relapsing after less than 1 year, or those with one of the other high-risk factors mentioned above, high-dose therapy with ABMT may be considered following re-induction with conventional-dose salvage therapy, with the caveats discussed below. ^[311]

Patients who do not achieve a complete remission with initial therapy, who have first complete remissions lasting less than 1 year, or who are in second or subsequent relapse are generally considered candidates for high-dose salvage regimens with either autologous bone marrow transplant (ABMT) or autologous peripheral blood stem cell transfusions (APBSCT). ^{[312] [313] [314]} The drugs used in these programs have myelosuppression as the primary dose-limiting toxicity and are used in doses above the lethal dose to bone marrow in the absence of stem cell re-infusion, but below the limit of life-threatening nonhematologic toxicity. The most commonly used regimens are BEAM (BCNU, etoposide, cytarabine, and melphalan) and CBV (cyclophosphamide, BCNU, and etoposide). Initial debulking with conventional-dose regimens before ablative therapy increases the effectiveness of the high-dose therapy, and several intensive salvage programs for initial debulking, generally including hematopoietic growth factor support, have been developed. These include MINE (mitoguazone, ifosfamide, navelbine, and etoposide), ^[315] dexa-BEAM, ^[316] mini-BEAM, ^[317] and EPOCH (etoposide, prednisone, vincristine, cyclophosphamide, and doxorubicin). ^[318]

About 35 to 45 percent of the highly selected patients treated in various reported trials have prolonged disease-free survival following high-dose salvage therapy and, because of perceived superiority over conventional salvage programs, this approach has become standard treatment, unfortunately without the proof of benefit afforded by large randomized clinical trials. Critical assessment of the overall benefit or the optimal timing of high-dose salvage therapy is difficult. The selection process for high-dose therapy favors younger patients with good performance status and little or no end-organ damage (i.e., the same patients who would be expected to do well with conventional salvage therapy). One randomized trial was attempted by the British National Lymphoma Investigation, comparing high-dose BEAM with ABMT versus mini-BEAM. Early results suggested that high-dose BEAM was better than mini-BEAM, but accrual was stopped after only 40 patients were randomized, and before proof of benefit could be obtained, because physicians refused further participation. ^[319]

Toxicity of ablative regimens is, of course, a major concern. High rates of infection occur during the 3 weeks or so required for marrow engraftment and production of adequate levels of neutrophils. The mortality rate in most of the reported trials averages 5 to 10 percent and somewhat higher when the preparative regimen includes radiation. Interstitial pneumonitis is the cause of one half to two thirds of the treatment-related deaths, followed by bacterial or fungal sepsis and cardiac failure. Although no prospective study that would allow a judgment about optimal timing of high-dose salvage therapy has been performed, theoretical considerations suggest that using high-dose therapy earlier in the course of relapsed disease rather than later would be advantageous ^[320] : significant decreases in morbidity and mortality should be realized by employing high-dose salvage before damage to marrow and other organs caused by multiple previous chemotherapy salvage regimens and radiation therapy accumulates; furthermore, tumor burden at the time of treatment is more likely to be low and, consequently, the likelihood of encountering biochemical resistance to the drugs should be lower. Experience in 22 high-risk patients in first remission following alternating MOPP-ABVD, then treated with CBV and ABMT, has been reported. ^[321] Seventeen of the patients (77 percent) were alive and continuously in CR at a median of 86 months following ABMT, with an actuarial overall survival of 80

2647

For patients whose initial remission was longer than 12 months, the evidence is good that conventional-dose salvage chemotherapy is able to induce second remissions in the majority of patients. However, the durability of remissions is usually poor. For that reason, some investigators have also used high-dose therapy with autologous bone marrow or stem cell transplantation in the relapsed patients with better prognosis. Reece and colleagues reported long-term disease-free survival in 79 percent of such patients, a result that is much better than would be expected from conventional-dose salvage therapy. ^[322] Therefore, it is reasonable to consider high-dose therapy for any patient not cured with their initial course of combination chemotherapy.

Late Complications of Therapy

As has been mentioned throughout this chapter, prolonged survival following effective therapy for Hodgkin's disease has been accompanied by several late sequelae, notably second malignancies, gonadal toxicity, hypothyroidism, and heart and lung toxicities. ^{[323] [324]} In a large epidemiologic study from the Netherlands Cancer Institute of 1,984 patients treated for Hodgkin's disease between 1966 and 1986, the cumulative risk of death at 20 years from diagnosis was 33 percent from HD or acute treatment toxicity, 14 percent from second malignancies, and 20 percent from all other causes, the most important of which was cardiovascular disease. ^[325] The risk of dying from second malignancies and other treatment-related causes continued to increase even after 20 years' follow-up. Awareness of the risks of these late toxicities is essential during initial treatment planning and during the lifelong follow-up period following treatment. When therapies of equal curative potential have different toxicities, the risk for potential late complications should guide choice of therapy. If an inherently increased risk of a treatment-related complication can be exacerbated by patient behavior, as is the case for the association of cigarette smoking with lung cancer following radiation therapy to the lungs, the patient at risk must be emphatically urged to eliminate the risky behavior. The physician must periodically reassess for the occurrence of late-appearing side effects and, in some cases, institute routine screening when early detection and treatment may make a difference, as is the case with radiation-induced breast cancer. Finally, we must not forget the truism that a prerequisite for susceptibility to death of a late complication of treatment 15 to 25 years after diagnosis of Hodgkin's disease is successful initial therapy.

Myeloid Malignancies

Cases of secondary myeloid malignancy, including myelodysplastic syndrome (MDS) and acute nonlymphocytic leukemia (ANLL) occur between 2 and 10 years following completion of therapy for Hodgkin's disease that includes alkylators, especially nitrogen mustard, chlorambucil, nitrosoureas, and procarbazine; thereafter, the incidence returns to that of the untreated population. Antileukemic chemotherapy for secondary ANLL seldom yields long-term remissions, and death within 4 to 8 months is usual. Multiple cytogenetic abnormalities, especially involving chromosomes 5 and 7, are characteristic of secondary myeloid leukemias. ^{[326] [327] [328]}

The cumulative risk of secondary leukemia and myelodysplastic syndrome for patients in first remission after six to eight cycles of MOPP or similar alkylator-containing chemotherapy without radiation therapy or maintenance therapy is 1.5 to 3 percent at 10 years and there is a dose-related increase in leukemia risk with increasing alkylator exposure and number of total treatment cycles. Thus, it is especially important to avoid exposures to alkylating agents that have been shown to be ineffective, such as maintenance chemotherapy after remission induction. The risk of myeloid malignancy after ABVD alone is nil. An epidemiologic study from the Netherlands Cancer Institute demonstrates clearly a lower incidence of leukemia/MDS in patients treated in the 1980s as compared with those treated in the 1970s, due to wide use of the MOPP/ABVD combination in the 1980s. ^[325] Evidence from several epidemiologic studies suggests that splenectomy increases the risk for leukemia. ^{[325] [329] [330]} Although several earlier reports suggested synergy between radiation therapy and alkylator therapy in causing myeloid malignancies, ^{[161] [331] [332] [333]} recent epidemiologic studies have suggested that radiation therapy adds little or no additional risk to alkylator-containing chemotherapy. ^{[325] [334] [335] [336]} The increased risk from radiation therapy is probably associated with larger field size and higher doses.

Non-Hodgkin's Lymphomas

Long-term Hodgkin's disease survivors are at increased risk for B-cell non-Hodgkin's lymphomas, usually of intermediate to high histologic grade. ^{[325] [332] [335] [337] [338] [339] [340] [341]} These lymphomas have a propensity for extranodal presentation, particularly in the gastrointestinal tract with a predilection for the ileocecal valve, not infrequently presenting with abdominal pain due to intussusception of the mass. Secondary lymphomas generally respond well to therapy that is effective in sporadic cases of aggressive lymphoma.

The similarity between the presentation of these lymphomas and non-Hodgkin's lymphomas that occur with high frequency in immunodeficiency states such as organ transplants or AIDS has suggested to many observers that non-Hodgkin's lymphoma is a consequence not of direct lymphomagenicity of the treatment itself but rather of a residual immunodeficiency

2648

Solid Tumors

Most of the excess risk of dying in Hodgkin's disease patients greater than 10 years after treatment is from solid tumors occurring within the field of previous radiation. Estimates of cumulative risk for solid tumors range from 8 to 13 percent at 15 years ^{[333] [335] [342]} to as high as 24 percent at 24 years in a cohort of pediatric patients. ^[329] Lung cancer is the most frequent second solid tumor, followed by stomach cancer, melanoma, and soft tissue sarcomas. ^[335] Lung cancer occurs almost exclusively in patients who smoked cigarettes, but at a much higher incidence than in nonirradiated smokers. Other tumors occurring with a greater than expected frequency include breast cancer, thyroid cancer (67-fold relative risk for children with Hodgkin's disease irradiated to the neck), bladder carcinoma, and head and neck cancer. ^{[343] [344]}

An increased risk for breast cancer following mantle field radiation therapy has been firmly established. ^{[325] [345]} Among 885 women treated with radiation therapy at Stanford University followed for a median of 10 years, 25 developed breast cancer, in comparison with about six cases expected in age-matched controls, for a relative risk (RR) of 4.1. The risk of developing breast cancer was highly dependent on the age at irradiation: for girls treated before age 15, the RR was 136; for those 15 to 24, RR was 19; and for those 24 to 29, RR was 7. The risk was not increased for women treated after age 30. The average interval from treatment to the diagnosis of breast cancer was 15 years, and the risk increased significantly with follow-up time greater than 15 years. A further increased risk was detected in women who had been treated with both MOPP and radiation therapy, although it was thought that estrogen-replacement therapy following iatrogenic menopause may have contributed to this risk. The actuarial and disease-free survivals following breast cancer diagnosis were slightly lower than expected for breast cancer in the general population. Both the finding of an elevated risk for breast cancer limited to women who were irradiated before age 30 and the 15-year latency before increased risk could be detected were corroborated in the large epidemiologic study from the Netherlands. ^[325]

A clinical analysis of 37 women with Hodgkin's disease and subsequent breast cancer (from an undefined population at risk) was reported. ^{[346] [347]} All patients had received radiation therapy to the mediastinum, 80 percent to a dose greater than 35 Gy. The median age at diagnosis of breast cancer in these women was 43 years with a median interval since treatment of Hodgkin's disease of 15 years; 68 percent of the patients had been treated for Hodgkin's before age 30. Eight patients (22 percent) developed bilateral breast cancer. In comparison with breast cancer patients from the general population, there was an increased propensity for presentation in the medial half of the breast (the portion with the greatest radiation exposure). The prognosis of these patients was strongly dependent on the axillary nodal status; survival in both node-negative and node-positive categories was similar to the survival in comparable primary breast cancer patients. Findings were positive in 81 percent of 32 mammograms of involved breasts available for review. Although it was not stated what prompted the mammographic studies (i.e., whether the mammograms were obtained for screening or after identification of palpable breast masses or other symptoms), a previous analysis of 29 (90 percent positive) of these mammograms had shown that in 15 of the 29 tumors (52 percent), both physical examination and mammography were positive; in 11 (38 percent), mammography was positive while physical exam was negative; whereas, in three (10 percent), the mammogram was negative and physical exam was positive. ^[347]

Based on the data in these studies, we recommend routine screening for breast cancer for all women who receive radiation to mediastinal or mantle fields for Hodgkin's disease before age 30 beginning 8 to 10 years after completion of radiation therapy, to include yearly breast examination and mammography.

Gonadal Failure

Sterility following chemotherapy, especially MOPP, or direct radiation therapy to the gonads, can be an important sequela of treatment. ^{[348] [349] [350]} The risk of sterility in men or women is nil following treatment with ABVD. ^[206] Virtually 100 percent of males have azoospermia immediately following a course of therapy with MOPP, and fewer than 10 percent will recovery normal sperm counts 6 months to 4 years after completion of chemotherapy. ^[351] Potency is not affected. The risk for sterility following MOPP in women increases with age of treatment. Affected women experience premature menopause due to injury to the rapidly dividing follicular epithelial cells of the ovary. ^{[348] [352] [353]} Some of the menopausal symptoms can be misattributed to recurrence of Hodgkin's disease and can provoke unnecessary testing, even including laparotomy. In the NCI series 89 percent of women who were older than 25 at the time of therapy became amenorrheic, whereas only 20 percent of the women who were younger than 25 at the time of treatment experienced menopause. ^[354] The onset of amenorrhea occurred less than 1 year after the completion of therapy in women older than 31. Women treated with MOPP before age 30 who underwent menopause frequently noted a gradual decrease in frequency of menses over several years, accompanied by perimenopausal hormone levels, culminating invariably in menopause. In the Stanford series ^[355] a sigmoidal increase in probability of ovarian failure with age was noted, with less than 50 percent probability of retaining regular cyclic menses when treatment was given after age 25. Even in women who

2649

If fertility is a concern for either men or women with Hodgkin's disease for whom chemotherapy is indicated, consideration should be given to the use of ABVD instead of MOPP. It appears that three cycles of MOPP, as would be given in the MOPP-ABVD alternating regimen, probably is the limit of exposure that will allow recovery of spermatogenesis in about half of treated men. If MOPP chemotherapy is planned, or if testicular radiation is necessary, sperm banking can be considered for men who desire to father children. In general, men with Hodgkin's disease have oligospermia related to the disease even before the onset of treatment. However, pregnancy following artificial insemination can be successful even for men with oligospermia before therapy, although the success rate of this procedure may be lower than for the normal population.

Late Cardiac and Pulmonary Toxicity

Depending on the radiation therapy techniques used and on the amount of intrathoracic tumor, mediastinal irradiation can result in an increased incidence among long-term survivors of Hodgkin's disease of constrictive pericardial disease, cardiomyopathy, and radiation pneumonitis and fibrosis. ^[356] These sequelae can be exacerbated by doxorubicin and bleomycin. Many of these late side effects have been greatly decreased due to refinement of radiation therapy techniques, ^{[357] [358]} but radiation therapy to the entire cardiac volume still results in significant reductions in left ventricular function. ^[359] Furthermore, a recent reassessment of patients treated with radiation therapy for Hodgkin's disease at Stanford has revealed a substantial 3.2-fold relative risk for death from myocardial infarction over 20 years among patients treated with radiation therapy as compared with an age- and sex-matched normal population. ^[155] The long-term risk for heart and lung toxicity among patients treated with ABVD or a related regimen without radiation therapy has not been carefully assessed. However, bleomycin in other regimens ^[360] has been associated with late toxicity and the report of toxicity in the CALGB comparative study ^[194] suggests that acute and chronic pulmonary toxicity and heart failure may be serious problems following ABVD.

Infections

Patients who have been treated for Hodgkin's disease generally have readily measurable defects in the function of their T cells, and some evidence suggests that these defects may be present even before the onset of Hodgkin's disease. Thus, patients with Hodgkin's disease have some immune compromise that is not necessarily a consequence of the treatment. In general, the immune compromise does not lead to frequent infections; however, infections are the third leading cause of death in patients cured of Hodgkin's disease behind heart disease and second cancers. ^[361] Patients without spleens are at the greatest risk. Prophylactic antibiotics have not been shown to be effective and pneumococcal vaccine is only partially protective after splenectomy. Thus, patients require education regarding the importance of seeking medical attention promptly if they develop symptoms of fever or signs of infection.