Diagnosis and Treatment of Mediastinal Lymphomas
Sonali Smith
Koen van Besien
Lymphomas are heterogeneous disorders with variable clinical presentation and course; correct histopathologic classification is essential for their appropriate management. Recent classification systems rely heavily on disease histology but also incorporate data from immunophenotyping, cytogenetic analysis, and molecular assays. The most commonly used system is the REAL classification (for Revised European American Lymphoma classification)36 or the more recent World Health Organization (WHO) classification.44 A modified classification for clinical use, as provided by Hiddemann,39 is shown in Table 191-1, with an emphasis on subtypes with a mediastinal tropism. Most subtypes of lymphoma can occasionally involve the mediastinum, but several entities have a unique affinity for the mediastinum and their clinical features are often related to the presence of a mediastinal mass. These include Hodgkin lymphoma, primary mediastinal B-cell lymphoma, and lymphoblastic lymphoma. Much of this chapter focuses on Hodgkin lymphoma, which has been reliably diagnosed and studied for close to five decades. Primary mediastinal B-cell lymphoma and lymphoblastic lymphoma have been recognized as unique disease entities with a tropism for the mediastinum in the last two decades and are discussed in some detail. The diagnosis and management of long-term sequelae of treatment for Hodgkin lymphoma represents a unique challenge in which thoracic surgeons are often involved and which is discussed in some detail as well.
Hodgkin Lymphoma
Since Sir Thomas Hodgkin’s first description of Hodgkin’s disease in 1832, Hodgkin’s disease has emerged as a unique malignancy, histologically characterized by the presence of Reed–Sternberg cells, as recently reviewed by Diehl19 and previously by Bonadonna.6 The disease is currently considered a subtype of lymphoma and the term Hodgkin lymphoma (HL) was proposed in the WHO classification by Jaffe et al.44 Radiation was the first treatment modality that demonstrated curability of HL. Staging laparatomy was essential for design of the radiation ports and for successful radiation treatment. More recently, the use of effective combination chemotherapy regimens along with improved radiographic techniques has obviated the need for staging laparotomy in most cases. HL is a very chemosensitive and radiosensitive disease. High cure rates can be obtained in newly diagnosed patients and in a considerable proportion of patients with recurrences. However, the successes have been tempered by serious late side effects resulting from radiation and chemotherapy. In pediatric patients, growth retardation and infertility are of concern. Thus, as the number of long-term HL survivors increases, the major thrust of current clinical research in patients with early stages of HL is to minimize long-term toxicity while maintaining the high cure rate.
Epidemiology and Etiology
HL is a relatively uncommon malignancy with an estimated 8,100 new cases and 1,070 deaths in 2007, as predicted by the Surveillance Epidemiology and End Results (SEER) database.47 The absolute incidence has remained stable over the past several decades, in contrast to non-Hodgkin lymphoma (NHL), which is increasing in incidence. Medeiros and Greiner59 show that there are generally two peaks in incidence: among teenagers/young adults (ages 15–34 years) and among adults >55 years of age. Overall, the incidence is similar in men and women, although mortality appears higher in men.
The cell of origin in HL has long been a mystery, as the pathognomonic Hodgkin’s and Reed–Sternberg (HRS) cell accounts for only 1% of all cells within an affected node and lacks classic hematopoietic surface markers. Several observations in the mid-1990s proved that HRS cells were not only clonal but were derived from “crippled” B cells. The dramatic discovery process of the B-cell origin is beautifully summarized in a review by Kuppers and Rajewsky.54 Despite being a B-cell–derived malignancy, HL and NHL maintain their individual classification by the WHO, as the response to treatment and natural history are quite divergent for the two disorders.
HL has several histologic variants, although the WHO has divided all cases into one of two categories: classic HL (cHL) and nodular lymphocyte predominant Hodgkin lymphoma (NLPHL). cHL includes the following four histologies: nodular sclerosis, mixed cellularity, lymphocyte-rich, and lymphocyte-depleted Hodgkin lymphomas. As reviewed by Nogova,67 NLPHL shares many features with non-Hodgkin lymphomas, including strong CD20 expression, and often has a waxing and waning course similar to that of indolent lymphomas. It
rarely has a mediastinal presentation, and thus the remainder of this discussion pertains mainly to cHL.
rarely has a mediastinal presentation, and thus the remainder of this discussion pertains mainly to cHL.
Table 191-1 Clinical Classification of Lymphoid Malignanciesa | ||
|---|---|---|
|
As with most malignancies, the etiology of HL is not definitively known. However, several risk factors have been proposed, including infection with lymphotrophic viruses, higher socio- economic status, and aberrancies of the immune system. The contribution of Epstein–Barr virus (EBV) in the pathogenesis of HL has been investigated for nearly four decades. Reports by Herbst,38 Jarrett,45 Glaser,28 and many others suggest that 25%–50% of HL cases are associated with infection with EBV. Clonal EBV DNA has been detected in Reed–Sternberg cells (the malignant cell of origin for HL) by Anagnostopoulos,2 Gledhill,29 and Pallesen68 and their colleagues and by many others. Skinnider and Mak83 summarized the data supporting the influence of EBV on cytokine production, which may underlie the clinical and pathologic manifestations of HL. As reported by Jarrett, the International Agency for Research on Cancer (IARC) considers the relationship of EBV and HL to be causal, although this may not be true of all cases.46
HL also occurs in the setting of human immunodeficiency virus (HIV); as such, it was one of the first malignancies designated as an AIDS-defining illness. Spina and colleagues86 report a more aggressive course for HIV-associated HL, whereas a review of cases by Tsimberidou and colleagues90 showed a clinical outcome similar to that of patients without HIV.
Clinical Presentation
The majority of patients with HL present with nontender adenopathy, usually in the neck or supraclavicular area. Mediastinal involvement is commonly present and may become very large. It is sometimes the only site of disease. Mauch and colleagues57 reviewed 719 cases of HL and noted that 60% presented with either cervical or mediastinal adenopathy. Regional disease presenting initially below the diaphragm (so-called subdiaphragmatic Hodgkin’s disease) is distinctly rare.
Mediastinal involvement in HL, especially when bulky, is a poor prognostic factor. In the revised Ann Arbor staging system, bulky mediastinal adenopathy is defined as a mediastinal-to-thoracic width ratio of greater than one-third. As reported by Hasenclever and Diehl,37 as many as two-thirds of patients present with mediastinal involvement. Those with very large masses have an inferior rate of freedom from progression and of overall survival in a univariate (but not multivariate) analysis. The adverse prognostic factors in this analysis include low serum albumin, anemia (hemoglobin <10.5 g/dL), male gender, stage IV disease, age ≥45 years, leukocytosis (>15,000/mm3), and lymphopenia (<600/mm3).
Constitutional symptoms occur in approximately one-third of patients and may consist of fatigue, diffuse pruritus, weight loss, fevers, and/or night sweats. The last three symptoms constitute the classic “B” symptoms of malignant lymphomas. When present, they typically confer a more aggressive course of disease. As elegantly summarized by Kaplan,51 other notable but less common symptoms attributable to HL include alcohol-induced pain and Pel–Epstein fevers. The pain associated with alcohol intake typically occurs quickly after consuming the beverage and may last from minutes to hours. Pel–Epstein fevers, first described in 1887, are cyclic temperature elevations typically with a 1- to 2-week periodicity. The fevers usually occur in the late afternoon and evening, only to normalize by morning. The
prognostic significance of constitutional symptoms other than the classic B symptoms described above is not known. However, recurrence of symptoms in a previously treated patient may signal disease relapse.
prognostic significance of constitutional symptoms other than the classic B symptoms described above is not known. However, recurrence of symptoms in a previously treated patient may signal disease relapse.
Diagnostic Evaluation
The diagnostic workup of HL includes obtaining an adequate tissue sample and a staging evaluation. The differential diagnosis of patients presenting with a mediastinal mass includes HL, primary mediastinal B-cell lymphoma, gray-zone lymphoma, lymphoblastic lymphoma, Ki-1–positive anaplastic large-cell lymphoma, primary mediastinal germ cell tumors, and, rarely, thymomas and carcinomas. With rare exceptions, an accurate diagnosis of lymphoma can be established only by obtaining tissue for immunophenotyping, cytogenetics, and molecular assays in addition to classic histologic examination. Many of these tests require viable tissue; it is therefore of the utmost importance that, in addition to fixed specimens, fresh tissue is provided to the pathologist whenever a diagnosis of lymphoma is suspected. Mediastinal lymphomas, be they HL or NHL, are often fibrotic, and small biopsies may be difficult to interpret. A diagnosis of benign versus malignant process may be difficult and subclassification may be impossible. In HL in particular, the malignant cell of origin, the Reed–Sternberg cell, is relatively rare in comparison to the intense background lymphoid reaction. Its recognition is essential for diagnosis, and such cells may not be present if the biopsy is small. Tissue obtained by fine-needle aspiration (FNA) or core biopsy is generally not sufficient for diagnosis. Samples obtained by mediastinoscopy, an excellent technique for staging carcinomas, are not always adequate for diagnosis of mediastinal lymphomas. If no easily accessible peripheral lymph node is present, thoracoscopy, anterior mediastinotomy, or a limited thoracotomy may be required for establishing the diagnosis.
Once a diagnosis of HL is confirmed, a thorough staging evaluation will allow a determination of clinical stage, which then guides further management. Blood tests for lymphoma staging should include routine blood counts, serum LDH (an important prognostic marker in many lymphomas), liver function tests, and kidney function tests. Radiographic staging for HL has recently been revised to include not only total body computed tomography (CT) scans and a bone marrow biopsy core and aspirate but also functional imaging with 18F-fluorodeoxyglucose (18F-FDG) positron emission tomography (PET), as discussed in greater detail below.
Staging laparotomy with splenectomy and lymph node sampling has been widely used in the past. The importance of staging laparotomy in HL is based on an early observation by Rosenberg and Kaplan77 that HL spreads in an orderly and predictable fashion. This allowed precise pathologic staging, thus avoiding the risk of under- or overtreatment. However, the use of systemic chemotherapy even in early-stage disease, the ability to effectively salvage patients who relapse, and improvements in radiologic staging have all decreased the need for precise surgical staging. Staging laparotomy is therefore no longer considered part of the standard approach to HL. Exploratory laparotomy is still occasionally recommended for patients for whom chemotherapy is contraindicated, usually because of concern for infertility or risk for secondary malignancies. Lower extremity lymphangiography was also part of standard staging in the past because its sensitivity exceeds that of abdominal CT scans. As combined-modality therapy is more commonly used and the expertise with lymphangiography decreases, this test is now rarely used.
The use of 18F-FDG PET has considerably challenged the approach to HL and reshaped the design of all new therapies for this disease. HL is a strongly PET-avid disease, and this examination should be positive in nearly all cases except situations of minimal residual or microscopic disease. 18F-FDG PET, or combined 18F-FDG PET/CT, has essentially replaced the use of another nuclear imaging modality, gallium-67 (67Ga). Although both gallium scans and 18F-FDG PET reflect metabolic activity within a mass and improve the sensitivity of detecting tumor over CT scans, 18F-FDG PET provides superior sensitivity and specificity compared with gallium in a greater variety of lymphoma types and in subdiaphragmatic disease, as summarized in a position paper by Seam, Juweid, and Cheson.81 Potential applications of 18F-FDG PET include routine staging, evaluation of response to treatment, evaluation of residual masses following treatment, and prediction of relapse risk in such patients.
A major advantage of 18F-FDG PET in HL, in addition to the strong uptake, is related to the frequent finding of a residual mass following treatment, especially in cases of bulky presentation. Isasi and colleagues performed a meta-analysis of 20 studies examining this topic and found that 18F-FDG PET can reliably distinguish between tumor and fibrosis with a pooled sensitivity and specificity of 87%. On behalf of the International Harmonization Project in Lymphoma, Juweid et al.49 and Cheson et al.14 presented updated response criteria incorporating 18F-FDG PET, which has now replaced the prior method of structural detail alone provided by CT scans. Perhaps the more interesting and still controversial application of 18F-FDG PET is in the prediction of response to specific chemotherapeutic regimens based on midtreatment 18F-FDG PET findings. Spaepen,85 Kostakoglu,53 and Querellou76 and their colleagues, among others, have found that 18F-FDG PET following one or two cycles of chemotherapy predicts outcome. Several studies by large cooperative groups within the United States and Germany are prospectively evaluating the utility of early or midtreatment 18F-FDG PET as a tool for individualizing treatment in patients with HL.
Management
The management of HL has evolved considerably over time. Initial reports showing the efficacy of radiation therapy led to significant optimism that HL could be cured. In the 1970s through the 1990s, knowing the precise stage of HL was critical in determining the treatment plan. The staging of HL based on the Ann Arbor classification and Cotswolds revision, as reported by Lister,56 is summarized in Table 191-2. Staging laparotomies with splenectomy and lymph node sampling determined the radiation ports. This staging method was called “pathologic staging” as opposed to “clinical staging,” which relied on physical examination and radiologic tests. Currently, management of HL, with the exception of very early stage favorable disease, utilizes chemotherapy as the mainstay of treatment, with radiation used mainly as a consolidative modality. Improved radiographic technology and the use of systemic chemotherapy have obviated the need for pathologic staging. Outside of establishing the diagnosis, surgery has a minimal role in the therapy of HL. An excellent summary of the history of HL is provided by Bonadonna.6
Table 191-2 Staging: Ann Arbor Classification (Cotswold Revision) | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| ||||||||||
Early-Stage Disease
The traditional management of patients with early-stage HL (stages IA to IIA) has been entirely radiation-based. Using extended-field radiation and pathologic staging, cure rates of 80% to 90% are achieved in such patients. The high incidence of secondary malignancies and other treatment sequelae (see below) in this otherwise favorable risk group has brought the radiation ports and doses into scrutiny. Modern treatment protocols often involve a moderate amount of chemotherapy and limited (“involved field”) radiation. As reported by Press and colleagues75 on behalf of a large intergroup randomized phase III trial, such combined-modality therapy was superior to subtotal lymphoid radiation in terms of failure-free survival (94% versus 81%, p <0.001). Other studies by Horning and colleagues41 and Brusamolino10 also support the use of combined-modality treatment for early-stage disease. A report from Groupe d’Etude des Lymphomes de l’Adulte (GELA) also convincingly demonstrates superior survival with the addition of systemic chemotherapy.24 Patients with early-stage disease who received three cycles of chemotherapy followed by involved field radiation had a superb 98% event-free survival after 5 years. Others, particularly the Canadian cooperative group, have evaluated the use of chemotherapy alone (without any added radiation) in patients with early-stage Hodgkin’s disease.61 The follow-up for these studies has not yet reached maturity, and although the preliminary results of a chemotherapy-only strategy are encouraging, this cannot be recommended as the standard of care at present. The treatment of early-stage HL continues to evolve, this time with an emphasis on reducing late complications while preserving the excellent cure rate, and whenever possible, patients should be encouraged to participate in prospective studies.
Advanced Disease
The foundation of treatment of patients with advanced-stage HL (i.e., those with bulky stage II, stage III, and stage IV disease as well as those with constitutional symptoms) is systemic chemotherapy. The original MOPP regimen (mechloretha- mine, vincristine, procarbazine, prednisone), has largely been supplanted with the equally effective and less toxic ABVD regimen proposed by Bonadonna7 (Adriamycin [doxorubicin], bleomycin, vinblastine, dacarbazine). Canellos and colleagues12 reported a randomized phase III study comparing MOPP and ABVD and a MOPP-ABVD combination. ABVD produced an 82% complete remission rate and 93% overall response rate compared to 77% complete remission rate and 93% overall response rate for MOPP. The combination of MOPP–ABVD was as efficient as ABVD but more toxic. Canellos and Niedzwiecki13 updated these data with a median follow-up of more than 14 years; patients treated with ABVD maintained a superior failure-free survival rate of approximately 60% at 10 years compared with approximately 40% for those who had MOPP chemotherapy. Other advantages of ABVD compared with MOPP include a decreased incidence of infertility and of secondary malignancies. On the other hand, bleomycin-induced pulmonary toxicity and doxorubicin-related cardiac toxicity are of concern.
Among patients with advanced disease, disease recurrence is still relatively common, and newer, and more intensive regimens, including combined-modality approaches, are actively being investigated. For example, the Stanford group40 has advocated an abbreviated course of intensive chemotherapy, called the Stanford V regimen, followed by adjuvant radiation therapy to sites of bulky disease, whereas the German Hodgkin lymphoma Study Group proposed the BEACOPP regimen (bleomycin, etoposide, doxorubicin, cyclophosphamide, vincristine, procarbazine, and prednisone). Diehl18 reported a randomized study showing that an escalated BEACOPP regimen resulted in improved survival compared to ABVD in patients with advanced Hodgkin’s disease, particularly those with high risk scores by the International Prognostic Scoring system. Despite these data, the escalated BEACOPP regimen has not (yet) found widespread acceptance in the United States, mainly because of concerns over early toxicity and the potential for secondary leukemia. A randomized multi-institutional U.S intergroup study comparing ABVD with Stanford V in advanced HL has completed accrual. An Italian randomized study does not indicate an advantage to the Stanford V regimen, but the study has been criticized for the delayed delivery of scheduled radiation in many patients.30
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
