Chapter 8: Aggressive B-Cell Lymphomas

It is clinically useful to divide non-Hodgkin lymphomas (NHL) into indolent and aggressive based on their clinical behavior (¹). Patients with indolent NHL typically have a survival of many years, even if untreated, but paradoxically are usually incurable. Patients with aggressive NHL have a survival time measured in weeks to months if untreated yet are usually chemosensitive and frequently curable. In this chapter, we focus on the clinical characteristics, pathology, and treatment of aggressive NHL.

The incidence of NHL has increased over the last five decades, as reported by US and international registries (^2,3,4). During the years 1993 to 1995, the age-adjusted incidence increased 3% per year according to data from the Surveillance, Epidemiology, and End Results (SEER) program of the National Cancer Institute (²). Some of this increased incidence can be attributed to the acquired immunodeficiency syndrome (AIDS), but this epidemic does not explain the increase of NHL before 1980. In the elderly population, there has also been a marked increase of NHL, largely the indolent NHLs, which are discussed in Chapter 7.

An estimated 71,850 new cases of NHL will be diagnosed in the United States in 2015, and 19,790 NHL-related deaths will occur. In 2014, NHL was the ninth largest cause of death among men and the eighth largest cause in women (3% of all cancer-related deaths). Diffuse large B-cell lymphoma (DLBCL), the most common NHL subtype, has an aggressive behavior and is more common in whites than African Americans in the United States; however, 5-year survival outcomes are worse in African Americans (⁵).

Most cases of aggressive NHL do not have a well-defined cause. Recent work has shown that the lifetime risk for many cancers correlates with the total number of divisions of normal self-renewing cells (⁶). The implications of these findings suggest that many cancers may not be due to hereditable genetic aberrancies, environmental exposures, infectious etiologies, or other known causes, but instead are attributable to a combination of factors that may include chance. For the NHLs that appear to have currently identifiable drivers, there are four groups of drivers: immune suppression (both acquired and primary), infectious agents, toxic exposure, and familial (Table 8-1).

The strongest association is with immune suppression, both primary and acquired (⁷). Examples of primary immunodeficiency include inherited immune disorders, such as Wiskott-Aldrich syndrome, severe combined immune deficiency, common variable immune deficiency, and ataxia-telangiectasia (⁸). These and other inherited disorders are associated with an increased lifetime risk of developing NHL, with aggressive B-cell NHL being most common.

Patients who are immunosuppressed for therapeutic reasons—for example, after transplantation—are also at increased risk of NHL, especially if treated with cyclosporine, azathioprine, prednisone, or monoclonal antibodies for the removal of T cells (⁹). A loose association can be drawn between the level of immune suppression and lymphoma risk. Transplant patients treated with the highest doses of immunosuppressive agents, such as heart transplant recipients, are at greater risk of developing lymphomas. These lesions are also more likely to be aggressive, extranodal forms. Individuals treated with pharmacologic immune suppression for autoimmune disorders—such as systemic lupus erythematosus, Sjögren syndrome, or rheumatoid arthritis—are also at increased risk for NHL, including the tumor necrosis factor antagonists (¹⁰). A subset of these NHLs is histologically aggressive and associated with Epstein-Barr virus (EBV). These lesions may regress following withdrawal of the immunosuppressive agent, speaking to the complex interaction between the immune system and an immune clonal population (¹¹).

Infectious agents associated with development of aggressive NHL include human immunodeficiency virus (HIV), EBV, human herpesvirus type 8, and human T-cell lymphotropic virus type 1 (¹²). The greatest factor involved in the worldwide increase in NHL, although lessened with the advent of highly active antiretroviral therapy (HAART), is HIV infection (¹³). The risk of NHL is increased by up to 300% in untreated HIV-infected patients, rising in proportion to the duration of the HIV infection. Although the risk of NHL in HIV-infected patients appears to be decreased by HAART, the relative risk of NHL remains much higher than that for those not infected with HIV. Aggressive NHL can occur in HIV-infected patients at any stage of infection, but the risk increases as CD4 counts drop to <100 × 10³/μL, and NHL is considered an AIDS-defining illness.

EBV also plays a role in lymphomagenesis, due in part to chronic antigenic stimulation (¹⁴). EBV is virtually always associated with certain types of NHL, such as endemic (African) Burkitt lymphoma (BL) and extranodal T-cell/natural killer (NK)-cell lymphoma of nasal type, with many other NHL subtypes occasionally involved. Many patients with HIV-related lymphomas are co-infected by EBV, including HIV-associated primary central nervous system (CNS) lymphoma, which is infected in essentially 100% of cases (¹⁵). Human herpesvirus type 8 is associated with primary effusion lymphoma (PEL), which tends to occur in HIV-infected patients. Human T-cell lymphotropic virus type 1 is associated with adult T-cell lymphoma/leukemia.

Environmental and occupational exposures to toxins are associated with an increased risk of NHL (¹⁶). Herbicides, especially phenoxyacetic acid derivatives, are associated with NHL, especially in the farming belts of the United States. Occupations held by individuals reported to be at increased risk for NHL include farming, metalworking, forestry, aircraft maintenance, woodworking, and dry cleaning. One of the common exposures in these industries is the use of organic solvents.

Family history of NHL is also a potential risk factor for some lymphomas. Individuals who have relatives with NHL may have a slightly higher risk of developing NHL, but data are inconclusive and mechanisms are unclear (¹⁷).

The clinical presentation of patients with aggressive NHL varies substantially based on histologic type and anatomic site of disease. The likelihood of B symptoms, including fever greater than 38°C, night sweats, or weight loss greater than 10% of body weight in the preceding 6 months, increases with NHL aggressiveness. Approximately 50% of patients present with these B symptoms, often with fatigue, malaise, and pruritus, although these are less common initially.

Most patients present with painless lymphadenopathy, which is often first treated with antibiotics for presumed infection and eventually biopsied when lymph nodes fail to regress. The most common scenario involves a diagnosis based on examination of peripheral lymph nodes, which may be detectable prior to internal lymph nodes becoming enlarged and causing symptoms. Peripheral lymph nodes are not usually painful, unless they are rapidly enlarging or are massive. Symptoms vary with the anatomic site of internal lymphadenopathy. Patients with mediastinal lymphadenopathy frequently experience cough, chest pain, and less commonly superior vena cava syndrome. Patients with large nodal masses in the abdomen or retroperitoneum frequently experience pain, abdominal fullness, or early satiety. Retroperitoneal lymphadenopathy can cause back pain and discomfort.

Extranodal disease is common in patients with aggressive NHL. The most common extranodal sites are the intestines, tonsils, and skin, although the frequency of involvement of these sites varies across reports. Disease in the gastrointestinal (GI) tract can present with nonspecific symptoms, including obstruction, blood loss with subsequent anemia, or diarrhea. Other extranodal sites include liver, lung, testis, bones, CNS, and spleen; however, aggressive extranodal NHL may involve nearly any organ system.

Diffuse Large B-Cell Lymphoma, Not Otherwise Specified

Diffuse large B-cell lymphoma is the most common type of NHL (^1,18). It occurs mainly in adults, with a median age in the sixth decade. Men are affected slightly more often than women. B-type symptoms or bulky disease occurs in one-third of patients. Nodal presentation is most common, but extranodal sites are involved in approximately 40% of patients (Figs. 8-1 and 8-2), and more than one-third of patients have more than one extranodal site of disease. Slightly more than half of patients have stage III or IV disease (¹⁹). Bone marrow involvement occurs in approximately 10% to 20% of patients. Diffuse large B-cell lymphoma uncommonly involves privileged sites, such as the testis and CNS, of which the latter portends a poor prognosis.

If untreated, DLBCL is invariably fatal, but DLBCL is generally sensitive to chemotherapy, at least initially. With current immune-chemotherapy approaches, a slight majority of patients will be cured of their disease, with 2-year event-free survival of approximately 70% (²⁰). The International Prognostic Index (IPI) score remains a widely used prognostic model, despite not accounting for any tumor-specific biologic features. With rituximab plus chemotherapy–based disease management, the 4-year progression-free and overall survival rates for DLBCL patients with a Revised IPI (R-IPI) of 0, 1 to 2, and 3 to 5 are as shown in Table 8-2 (²¹).

These data are generally useful but unfortunately are not predictive of outcomes for an individual patient. Nearly half of patients will have a poor-risk R-IPI, and of these patients, approximately half will be cured and half will die of disease. Few patients with relapsed disease are cured.

The diagnosis of DLBCL is based on its diffuse growth pattern and large neoplastic cells, with frequent mitotic figures (¹). Cytologically, the neoplastic cells can be subdivided as large centrocytes, centroblasts, or immunoblasts. Large centrocytes range from 13 to 30 mm in size and have irregular or cleaved nuclear contours, relatively small, indistinct nucleoli, and a thin rim of eosinophilic cytoplasm. Centroblasts are 20 to 30 mm in size and have round or oval vesicular nuclei with two or three nucleoli and more abundant amphophilic cytoplasm (Figs. 8-3 and 8-4). Immunoblasts are larger than centroblasts, with an eccentrically located vesicular round or oval nucleus containing a prominent target-like central nucleolus and relatively abundant amphophilic, often plasmacytoid cytoplasm (Fig. 8-5). Elevated proliferation (Ki-67) rates and immunoblastic features appear to correlate with MYC rearrangements (²²).

Immunophenotypic studies have shown that DLBCLs are of mature B-cell lineage. Approximately two-thirds of cases express monotypic immunoglobulin (Ig). These tumors express pan-B-cell antigens, 60% to 70% express BCL2, and a subset is positive for CD10 and BCL6.

Diffuse large B-cell lymphomas are heterogeneous at the molecular level. A subset of cases carries the t(14;18) involving BCL2, as shown by conventional cytogenetic or molecular studies (^23,24). Another subset of DLBCLs has translocations or other abnormalities involving BCL6 at chromosome 3q27. BCL6 is rearranged in approximately 20% to 40% of DLBCLs, more often in tumors arising at extranodal sites (²⁵). Recently, DLBCLs with translocations involving both MYC and BCL2 or BCL6, representing approximately 5% to 10% of DLBCL, have been shown to have an extremely aggressive behavior and are commonly referred to as “double-hit” DLBCL (²⁶). Diffuse large B-cell lymphomas with overexpression of MYC and BCL2 or BCL6 protein levels are more common and are designated double-positive DLBCLs. These tumors are still adverse compared to DLBCLs without protein overexpression but less so than translocation-defined double-hit DLBCLs (^27,28,29).

Gene expression profiling (GEP) can molecularly divide DLBCL into three groups: germinal center B-cell like (GCB), activated B-cell like (ABC), and a third noncharacteristic group. Patients with the GCB type of DLBCL have a better prognosis independent of the IPI (^30,31). Despite the initial GEP studies of DLBCL occurring approximately 15 years ago, this technology is not routinely used in clinical practice due to logistical issues and need for significant bioinformatic analysis. Fortunately, new technologies may allow these issues to finally be overcome by limiting the number of genes analyzed, such as in the Lymph-2Cx Nanostring assay (Nanostring, Seattle, WA) (³²). In lieu of GEP, a limited number of immunohistochemical markers, including CD10, MUM-1, BCL6, GCET1, and FOXP1, can subclassify DLBCL into GCB and non-GCB with relatively good concordance with GEP (^33,34).

Diffuse Large B-Cell Lymphoma Clinicopathologic Subtypes

T-Cell/Histiocyte-Rich B-Cell Lymphoma

T-cell/histiocyte-rich (TCHR) DLBCL is a diffuse neoplasm in which most of the cells are reactive T cells and histiocytes and the large, neoplastic B cells represent <10% of all cells in the infiltrate (Fig. 8-6). Patients with TCHR-DLBCL commonly have a history of nodular lymphocyte-predominant Hodgkin lymphoma, and TCHR-DLBCL may represent a transformation event in some patients.

Primary Diffuse Large B-Cell Lymphoma of the Central Nervous System

This entity includes all primary intracerebral or intraocular lymphomas. Although these lymphomas are remarkable for their unique clinical presentation, histologically and at the immunophenotypic level, these neoplasms closely resemble other cases of systemic DLBCL. In HIV-positive patients, the neoplastic cells commonly have immunoblastic features and are positive for EBV. Approximately, one-third of the cases demonstrate translocations involving the BCL6 gene.

Primary Cutaneous Diffuse Large B-Cell Lymphoma, Leg Type

As the name suggests, DLBCL, leg type, presents as cutaneous lesions originally described in the lower extremities, although all skin sites may be affected. Lesions often are initially confused with an insect bite, but eventually become diffuse and ulcerated, and the clinical course is aggressive. Histologic sections show diffuse sheets of monotonous neoplastic cells with centroblastic or immunoblastic morphology with frequent mitotic figures (³⁶). The neoplastic cells have a non–germinal center B-cell immunophenotype, and fluorescence in situ hybridization studies often detect rearrangements of BCL6, IGH, and/or MYC (³⁷).

Epstein-Barr Virus–Positive Diffuse Large B-Cell Lymphoma of the Elderly

This DLBCL is most common in the elderly, defined arbitrarily as >50 years of age but can occur in younger patients and often presents at an advanced stage. Epstein-Barr virus is a defining feature. Histologically, these neoplasms have a diffuse pattern and show a spectrum of features, and two morphologic variants are recognized: polymorphous, with a broad range of B-cell maturation in the reactive background, and monomorphous, which contains mostly large cells (¹). Geographic necrosis is common, and the neoplastic cells usually have a non–germinal center B-cell immunophenotype.

Intravascular Large B-Cell Lymphomas

These are rare neoplasms in which the lymphoma cells are confined to intravascular spaces (¹). The histologic diagnosis can be subtle, and therefore, diagnosis can be delayed. The neoplastic cells are large, express B-cell antigens, and usually have a non–germinal center B-cell immunophenotype. The molecular basis of intravascular large B-cell lymphoma is unknown (Fig. 8-7).

Primary Effusion Lymphoma

Primary effusion lymphoma, also known as body cavity–based lymphoma, is a very rare neoplasm of large B cells that involves a body cavity and is almost always associated with AIDS (³⁸). The prognosis for patients with PEL is poor. Human herpesvirus type 8 (also known as Kaposi sarcoma–associated herpesvirus) is present in virtually all cases of PEL, and its presence selects for a distinct cellular gene expression profile (¹). EBV is also present in most cases of PEL. An extracavitary (solid) variant of PEL also can occur, involving lymph nodes or extranodal sites (³⁹).

Primary Mediastinal (Thymic) B-Cell Lymphoma

These neoplasms are thought to arise in the thymus, are usually localized to the thoracic cavity, and occur more frequently in young women, with a female-to-male ratio of 2:1 (¹). Patients with primary mediastinal B-cell lymphoma (PMBL) present frequently with cough and dyspnea mimicking a respiratory infection or, rarely, superior vena cava syndrome. Histologically, PMBL has a diffuse pattern and is composed of large lymphoid cells that exhibit a spectrum of cytologic appearances: centroblastic, immunoblastic, or a mixture. Sclerosis is common, mitotic figures are usually numerous, and the tumor cells can have clear or pale cytoplasm. Immunophenotypic studies have shown that PMBLs are frequently Ig negative and commonly lack major histocompatibility complex class II antigens. Gene expression profiling analysis of PMBL has shown significant overlap with classical Hodgkin lymphoma (⁴⁰). Rearrangements of CIITA occur in about 40% of PMBLs, and amplification of the chromosome 9p24.1 locus (site of programmed death ligands) and mutations of PTPN1 have been reported in about 25% of cases (⁴¹). These molecular abnormalities highlight the interaction between the immune system and PMBL in pathogenesis and suggest potential therapeutic targets.

Mantle Cell Lymphoma

Mantle cell lymphoma (MCL) represents approximately 6% of all NHLs (¹⁹). Patients with MCL have a median age in the seventh decade, and MCL has a male-to-female ratio of approximately 3:1 (⁴²). Most patients present with advanced-stage disease, with bone marrow involvement in approximately 60% of patients, and most patients have low-level involvement of the peripheral blood. Overt leukemia may be associated with a poorer prognosis. Although GI symptoms are uncommon, 85% to 90% of MCL patients have GI involvement (⁴³). Blastoid MCLs have a more aggressive clinical course and often have occult CNS involvement. In the peripheral blood, MCL can sometimes resemble B-cell prolymphocytic leukemia.

Histologically, in classical MCL, the lymph node architecture is replaced by a diffuse or vaguely nodular neoplasm (Fig. 8-8) (¹). In a subset of cases, a mantle zone pattern results when the neoplasm selectively involves the follicular mantle zones surrounding normal or reactive germinal centers (Fig. 8-9). Cytologically, MCL is composed of a monotonous population of small lymphoid cells with slightly or clearly irregular nuclear contours (Fig. 8-10). In about 15% of cases, MCL may show blastoid (lymphoblastic-like) or pleomorphic (large cell–like) features. These neoplasms are associated with a poorer prognosis and have a high frequency of TP53 or TP16 mutations.

Mantle cell lymphomas express monotypic Ig light chain IgM, IgD, pan-B-cell antigens, BCL2, SOX11, alkaline phosphatase, and CD5 (²¹). A high proliferation rate, most often assessed by Ki-67 immunohistochemical analysis, predicts a poorer prognosis. t(11;14)(q13;q32) is present in virtually all cases of MCL (⁴⁴). In this translocation, CCND1 on 11q13 is juxtaposed with IGH on 14q32, resulting in overexpression of cyclin D1 (Fig. 8-11). Cyclin D1 facilitates cell cycle transition from G₁ to S phase (⁴⁵). Although t(11;14) is central to the pathogenesis of MCL, t(11;14) is insufficient to cause lymphomagenesis. Conventional cytogenetic studies have shown a number of additional abnormalities (^46,47), and a number of gene mutations have been reported.

Burkitt Lymphoma

There are three variants of BL: endemic (African), sporadic (nonendemic), and AIDS-associated (^21,48). Endemic BL was first described in equatorial Africa and is associated with EBV infection in 95% of patients (⁴⁸). The median age of patients is approximately 7 years, with a 3:1 male-to-female ratio (⁴⁹). Abdominal and retroperitoneal lymph nodes, the GI tract, and the gonads can be involved.

Sporadic BL occurs in industrialized nations and is associated with EBV infection in approximately 25% of patients. Patients are usually in the second or third decade of life, with a male-to-female ratio of 3:1. Involvement of the jaw is uncommon, and patients may present with large masses in the abdomen, peripheral lymph nodes, pleura, or pharynx (¹⁹). In endemic or sporadic BL, bone marrow and CNS involvement are uncommon at presentation, but they are frequent sites of subsequent dissemination.

Burkitt lymphoma can also occur in the clinical setting of immunosuppression, including HIV infection, posttransplant, or congenital immune deficiency setting (¹⁹). Epstein-Barr virus infection occurs in 30% to 40% of cases.

Morphologically, all variants are similar. The relatively clear histiocytes in a background of blue lymphoma cells imparts a “starry sky” appearance (Fig. 8-12A). This pattern results from rapid cell turnover with individual cell necrosis and scavenging of debris by macrophages. The neoplastic cells are round to ovoid, uniform in shape, and approximately the size of benign histiocyte nuclei. The chromatin is coarse, with two to five prominent basophilic nucleoli. Mitotic figures are numerous.

Burkitt lymphomas of endemic, sporadic, and AIDS-associated types are of mature B-cell lineage and express Ig, pan-B-cell antigens, CD10, and BCL6 (²³). Burkitt lymphomas have a very high proliferation rate (>99%) using an antibody specific for Ki-67 (Fig. 8-12B and C) and are negative for BCL2. MYC translocations are characteristic of BL. Approximately, 80% of cases carry the t(8;14)(q24;q32), with the remaining cases having one of two variant translocations, t(2;8)(p11;q24) or t(8;22) (q24;q11) (^50,51,52). Common to each of these translocations is involvement of chromosome region 8q24, the site of the MYC, which is deregulated. Via these translocations, MYC is juxtaposed with the IgH or the Ig light chain genes.

B-Cell Lymphoma With Features Intermediate Between Diffuse Large B-Cell Lymphoma and Burkitt Lymphoma

This entity in the 2008 edition of the World Health Organization (WHO) classification of hematopoietic neoplasms is designed for cases that do not fit either DLBCL or BL (¹). In the past, these cases were designated as atypical Burkitt/Burkitt-like lymphoma or as high-grade B-cell lymphoma or B-cell lymphoma with high-grade features. This group often exhibits morphologic and immunophenotypic deviation from typical cases of BL and DLBCL. These neoplasms have a diffuse pattern, often with a starry sky appearance, and a high proliferation rate (Fig. 8-13) (^53,54). A significant number of cases demonstrate MYC rearrangement, although in some of the cases, MYC rearrangement involves one of the non-Ig partners (⁵⁵). Many cases of so-called double-hit lymphoma fit within this category. These cases commonly demonstrate strong BCL2 expression and a complex karyotype (⁵⁶). The algorithm for diagnosis of high-grade B-cell lymphoma is illustrated in Fig. 8-14.

The Ann Arbor Staging system, developed in 1971 for Hodgkin lymphoma, is used to stage NHL (Table 8-3) (¹⁹). NHLs are often disseminated at diagnosis, unlike Hodgkin lymphoma. Of interest, there does not appear to be any meaningful difference in the outcomes of patients with stage III or IV disease, and thus the purpose of staging is to identify the patients with localized NHL who may benefit from additional local therapy.

A careful history and physical examination, including the presence of systemic symptoms, is essential. Performance status and comorbid conditions should be assessed. Physical examination should include a complete survey of all external lymph node groups including cervical, supraclavicular, axillary, epitrochlear, inguinal, and popliteal areas. Examination of abdomen for organomegaly is necessary, and in men, the testes should be examined. A complete neurologic examination must also be performed. Laboratory evaluation includes a complete blood count with differential, lactate dehydrogenase (LDH), β₂-microglobulin, kidney and liver function tests, albumin, calcium, and uric acid. Testing for hepatitis B is indicated prior to rituximab therapy, as the virus may reactivate during or after treatment. Testing for HIV should always be performed, as should bone marrow aspiration and biopsy (bilateral biopsies for certain NHL types; also see comments on fluorodeoxyglucose [FDG] positron emission tomography [PET] and bone marrow findings below). Examination of the cerebrospinal fluid (CSF) should be strongly considered for patients with highly aggressive NHL; DLBCL associated with spinal or paraspinal masses; renal or adrenal, ovarian, breast, or skull lesions; bone marrow involvement; testicular lymphoma; or nasal or sinus lymphomas; or any patient with clinical symptoms leading to suspicion of CNS involvement (⁵⁷).

Additional clinical evaluation is guided by the histologic type of NHL, symptoms, and anatomic sites involved by NHL. Lymphomas of the GI tract, especially in the stomach, require endoscopy for diagnosis unless other disease sites can be found to biopsy. It is especially important that multiple biopsies of different areas of the stomach be obtained because sampling error is frequent. There is no utility to gastrectomy or other surgical management for extranodal disease. Other types of aggressive NHL can involve the GI tract, especially MCL (⁴³). Evaluation of primary CNS lymphoma requires biopsy of the lesion, but a vigorous search for additional disease sites should be undertaken concurrently, because therapy for CNS and systemic disease will need to account directly for both.

Imaging Studies for Initial Staging

The use of imaging studies for evaluating lymphoma patients has evolved greatly. In 2014, FDG-PET–computed tomography (CT) scans were named the preferred imaging technique for FDG-avid lymphomas (essentially all aggressive NHLs) (⁵⁸). Based on these new guidelines, referred to as the Lugano criteria, a routine chest x-ray is no longer required. These guidelines also recommend that bone marrow biopsy could be omitted if the bone or marrow is FDG-avid on initial PET/CT scans. There still is significant controversy regarding the opposite situation: If the PET/CT scan does not show bone or marrow involvement, what is the utility of a bone marrow biopsy? Scant infiltration of DLBCL in the bone marrow (10%) may be missed with lack of FDG avidity, and thus, routine bone marrow biopsy may still be required for staging DLBCL patients (Figs. 8-15 and 8-16).

Limitations of PET/CT scan for staging of aggressive lymphoma include the fact that uptake of FDG is not specific to tumors and that infection and inflammatory processes are common false-positive findings on PET. As a result, an unexpected FDG-avid lesion that will result in a significant change in management should be confirmed by biopsy. The presence of high normal background activity in an organ, for example, in the kidneys or testes, may also make it difficult to identify abnormal FDG sites in that region. Although there is usually high normal metabolic activity within the brain, CNS lymphomas are often positive on FDG-PET scans, showing greater metabolic activity than the adjacent brain. However, additional imaging with magnetic resonance imaging (MRI) may be indicated for confirmation.

Bone Marrow Evaluation

Bone marrow aspiration and biopsy should be performed as part of the initial staging evaluation because involvement suggests widespread disease (stage IV) that affects treatment and prognosis, with the caveat mentioned above (see previous section). Bilateral iliac crest assessment is preferred because sensitivity of detection is higher than unilateral biopsy (⁵⁹). Although several studies found high accuracy of FDG-PET for predicting bone marrow involvement, a recent analysis concluded that the positive predictive value is high, but a negative FDG-PET is not completely concordant with the results of bone marrow biopsy (⁶⁰). Of note, the pattern of uptake within the bone marrow spaces on FDG-PET is important, because a diffuse pattern is commonly seen with activation (eg, with underlying anemia or infection, or after chemotherapy or growth factor treatment), and caution should be taken in interpreting this as diffuse bone marrow involvement by tumor. In contrast, focal or nodular uptake within osseous structures is suspicious (Fig. 8-17).

Pretreatment

Prognostic factors in patients with aggressive NHL can be broadly grouped into pretreatment (tumor-related) and treatment-related characteristics. Tumor-related genetic characteristics of importance, as noted earlier, include germinal (GCB) or non-GCB origin genetic profile and presence of MYC and BCL2 translocations (“double-hit”). Other tumor-related characteristics reported to be of prognostic value include a complex karyotype shown by conventional cytogenetics, high proliferation rate (high Ki-67 expression), and BCL2 and/or MYC expression shown by immunohistochemical staining (¹).

High serum LDH level is a measure of anaerobic metabolism and/or cell turnover and tumor bulk and is associated with a lower probability of complete remission and poorer long-term survival in patients with aggressive NHL. Other pretreatment prognostic factors include serum β₂-microglobulin level, stage, number of disease sites, bulky disease, presence of bone marrow involvement, poor performance status, and age (⁶¹). Of these pretreatment factors, age appears to be the most important, with patients over the age of 60 having lower response rates and a higher rate of relapse (⁶²).

The most commonly used system to provide pretreatment prognostic information in patients with aggressive NHL is the IPI, first developed in 1993 (Table 8-4) (⁶³). These initial data resulted from a cohort of 2,031 patients treated with doxorubicin-containing regimens analyzed for the presence of factors that independently predicted survival. The most commonly used doxorubicin-based regimen at that time was CHOP (cyclophosphamide, doxorubicin, vincristine, and prednisone) (Table 8-5). Significant prognostic factors were serum LDH (abnormal vs normal), age (<60 vs >60), number of extranodal sites (<2 vs >2), performance status (Eastern Cooperative Oncology Group [ECOG] 0-1 vs 2-4), and stage (I and II vs III and IV). Each of the five factors had an equal impact on survival. Risk groups identified were low (zero to one factor), low/intermediate (two factors), high/intermediate (three factors), and high (four to five factors), with 5-year survival rates of 73%, 51%, 43%, and 26%, respectively. Stage, serum LDH level, and performance status were independent predictive prognostic factors in a simplified subanalysis of 1,274 subjects ≤60 years of age in the same study. In this subgroup, the 5-year survival rate was 83% for zero risk factors, 69% for one risk factor, 46% for two risk factors, and 32% for three risk factors. In patients over 60 years of age, the 5-year survival rates were 56%, 44%, 37%, and 21%, respectively. These data highlight the prognostic significance of age on the survival of patients with aggressive NHL.

The IPI has been broadly applied as the standard for prognosis in patients with aggressive NHL, although corrections or changes to the IPI have been proposed including the National Comprehensive Cancer Network IPI (⁶⁴), the IPI24 (⁶⁵), and the R-IPI, which accounts for the addition of rituximab to the frontline treatment of DLBCL (²¹). In the R-IPI, there are only three groups—low risk with zero risk factors, intermediate risk with one or two risk factors, and high risk with three or more factors—with 4-year progression-free survival (PFS) rates of 94%, 80%, and 53%, respectively.

Prognostic Factors in Mantle Cell Lymphoma

The same prognostic factors in the IPI for aggressive lymphomas are of utility in patients with MCL. Other adverse prognostic factors include p53 mutations or deletion, elevated Ki-67, and blastoid histology. A prognostic model for MCL patients treated with chemotherapy followed by high-dose chemotherapy followed by autologous stem cell transplantation (Mantle Cell International Prognostic Index) was proposed using age, performance status, LDH, and leukocyte count (⁶⁶). Patients were divided in low-, intermediate-, and high-risk groups, with overall survival (OS) times of not reached, 51 months, and 29 months, respectively. In patients receiving rituximab plus hyperfractionated cyclophosphamide, vincristine, doxorubicin, and dexamethasone (R-hyper-CVAD), alternating with rituximab-methotrexate and cytarabine, this model could not be reproduced (⁶⁷). However, it was reproducible in patients treated with CHOP-rituximab–like regimens consolidated with high-dose chemotherapy with stem cell transplantation (⁶⁸).

Prognostic Factors in Primary Central Nervous System Lymphoma

Age and LDH are important prognostic factors in patients with HIV-negative primary CNS lymphoma; however, the most important factor is performance status at the time of treatment. Elevated LDH, CSF protein, and tumor mass location(s) are also contributors to prognosis (⁶⁹). Many patients can improve their condition by use of corticosteroids and thus be candidates for intensive chemotherapy-based regimens that are potentially curative.

Therapy-Associated Prognostic Factors

An important posttreatment prognostic indicator is tumor response to induction chemotherapy. In patients with aggressive NHL, dramatic response to induction with early complete remission (by the third cycle of therapy) is associated with a superior outcome (⁷⁰). Fluorodeoxyglucose PET has been found to be highly sensitive for the detection of aggressive NHL in posttreatment residual masses, but its ability to detect interim therapy response is controversial. Moskowitz et al found that DLBCL patients with persistent FDG avidity after four cycles of rituximab plus CHOP (R-CHOP) had an 86% false-positive rate (PET/CT positive, biopsy negative for persistent disease) (⁷¹). Similar data have been shown by many others, highlighting the need for biopsy confirmation of a positive PET/CT scan prior to therapeutic decisions.

Patients who fail to achieve at least a good partial response to induction chemotherapy have primary refractory disease and short survival despite all efforts. Another important indicator of prognosis is duration of remission obtained after induction chemotherapy, because patients with relapses occurring at <1 year have a worse outcome (⁷²).

Early-Stage Aggressive Non-Hodgkin Lymphoma

Early-stage (localized) aggressive NHLs (stages I and contiguous II) were historically treated with radiation therapy (RT) alone, and the results were highly variable (⁷³). The 5-year survival with involved-field RT for stage I/II disease was approximately 50%. Patients with bulky disease (>5 cm) suffered a higher relapse rate. Although many studies were undertaken to improve results by adjusting dosages and field coverage, it was the addition of combination chemotherapy to RT regimens that improved outcome most dramatically. Four randomized trials were conducted in patients with early-stage aggressive NHL before rituximab therapy was incorporated into the CHOP regimen. The first study was by the Southwest Oncology Group (SWOG); eight cycles of CHOP were compared to three cycles of CHOP followed by involved-field RT (40-55 Gy) in limited-stage DLBCL (⁷⁴). The combined-modality arm achieved an OS of 82%, versus 72% for the CHOP alone arm. The ECOG randomized patients with bulky stage I or II disease to eight cycles of CHOP with or without involved-field RT. Patients achieving a complete remission were randomized to involved-field RT (30 Gy) or no further therapy. Patients achieving partial remission received involved-field RT at a higher dose (40 Gy). Disease-free survival at 5 years was higher in patients who received radiation (73% vs 58%) after achieving complete remission (⁷⁵). The Groupe d’Etude des Lymphomes de l’Adulte (GELA) conducted a similar study comparing aggressive chemotherapy (dose-intensified doxorubicin, cyclophosphamide, vindesine, bleomycin, and prednisone [ACVBP]) alone versus abbreviated chemotherapy (three cycles of CHOP) followed by involved-field RT for stage I or II mostly low-risk aggressive lymphoma. All patients in this study were younger than age 60 years. Both the 5-year event-free survival (82%) and OS (90%) rates were significantly better in the chemotherapy group than in the combined-modality group (74% and 81%, respectively); however, the chemotherapy group had significant toxicity and ACVBP is not available in the United States. Although the addition of RT reduced relapses at the initial disease sites, this was not enough to overcome the excessive number of relapses in the abbreviated chemotherapy group (⁷⁶). Despite these results indicating the inability of abbreviated chemotherapy plus RT to prevent out-of-field relapses, the GELA group conducted another trial, GELA LNH 93-4, comparing CHOP with CHOP plus involved-field RT to 40 Gy, this time for patients older than 60 years (⁷⁷). The use of ACVBP had been dropped by the time this trial was undertaken because of excessive toxicity. At a median follow-up time of 7 years, no significant differences were evident in 5-year event-free survival rates (61% for chemotherapy alone vs 64% for chemoradiation) or OS rates (72% vs 68%, respectively). The results in the chemotherapy-only group were similar to the results from the group that received eight cycles of CHOP in the SWOG trial discussed earlier. However, because relapses could appear beyond 5 years, another 5 to 10 years should be allowed to elapse before any approach is widely adopted. A SWOG trial evaluated the addition of rituximab to three cycles of CHOP, followed by involved field RT, for localized DLBCL and found favorable comparisons to nonrituximab historical controls (⁷⁸). The MD Anderson Cancer Center (MDACC) approach to the treatment of DLBCL is shown in Fig. 8-18.

Localized MCL in general has been treated in the same way as extensive disease, but the use of RT with or without chemotherapy has been reported by investigators in British Columbia to be effective as well (⁷⁹).

Advanced-Stage Aggressive Non-Hodgkin Lymphoma

Initial cures using chemotherapy for patients with large-cell lymphoma were reported in the 1970s (^80,81). The SWOG initially reported that CHOP induced complete response in 50% of patients, with long-term disease-free survival in 35%, and CHOP has since represented the standard of care in the treatment of patients with aggressive NHL despite intensive research into newer regimens. Subsequent trials of combination chemotherapy over the last 25 years can be thought of as “generations” (⁸²). The initial generation included CHOP, M-BACOD (methotrexate, bleomycin, doxorubicin, cyclophosphamide, etoposide, and leucovorin), BACOD, ProMACE-MOPP (addition of etoposide), ProMACE-CytaBOM (addition of cytarabine), and MACOP-B (methotrexate with leucovorin rescue, doxorubicin, cyclophosphamide, vincristine, prednisone, and bleomycin). These regimens showed an increased rate of complete remission of nearly 80% in early studies with greater than 60% long-term disease-free survival. The SWOG undertook a landmark phase III trial comparing CHOP, MACOP-B, M-BACOD, and ProMACE-CytaBOM, which showed that, despite early reports of improved response, OS at 3 years varied from 50% to 54%, with disease-free survival ranging from 41% to 46%. In this trial, there was no apparent advantage to increased intensity of therapy. Another finding noted in the 1980s was that inclusion of an anthracycline in the chemotherapy regimen was important to long-term disease-free survival. Other approaches have included alternating regimens, higher dose therapy, and dose-dense therapy. The first two have not been shown to have a survival benefit, whereas the third is still under scrutiny.

Dose-dense therapy was reported to be feasible in 2003 by the German High-Grade NHL Study Group (DSHNHL) (⁸³). Three variants of CHOP-like therapy were evaluated, including CHOP-14, CHOEP-14 (addition of etoposide 100 mg/m² on days 1-3), and CHOP-21, each with hematopoietic growth factor support. An interim analysis of 959 patients showed that adherence to the dose-dense regimens was excellent, although dose reductions were more frequently required for the addition of etoposide. To evaluate younger patients, 710 patients with good-prognosis aggressive NHL age 18 to 60 years were randomized to receive six cycles of CHOP-21, CHOP-14, CHOEP-21, or CHOEP-14 (⁸⁴). Patients in the 2-week regimens received granulocyte colony-stimulating factor (G-CSF) from day 4. Initial sites of bulky or extranodal disease were treated with 36 Gy of RT. Patients receiving CHOEP achieved a higher complete response rate (87.6% vs 79.4%) and 5-year event-free survival (69.2% vs 57.6%) than patients treated with CHOP. Dose density (the 2-week regimens) improved OS in a multivariate analysis. Patients receiving CHOEP had a higher rate of myelosuppression, but generally, the regimen was well tolerated.

While the German group was exploring dose density, GELA reported their results of trial LNH98-5 in which 399 patients with DLBCL were randomized to receive either R-CHOP every 21 days or standard CHOP alone for a total of eight cycles. Patients with stage II to IV DLBCL who were between 60 and 80 years old were eligible for this trial. No RT or intrathecal chemotherapy was administered. The complete response rate (76% vs 63%) and the 5-year PFS, disease-free survival, and OS rates were better in the rituximab arm (⁸⁵).

Based on the GELA LNH98-5 results and their own data, the DSHNHL designed a four-arm randomized study in patients older than 60 years (RICOVER-60) that compared CHOP-14 with or without rituximab for six cycles versus CHOP-14 for eight cycles with or without rituximab (in the six-cycle R-CHOP arm, the patients received a total of eight doses of rituximab) (⁸⁶). The CHOP alone group was inferior, and six cycles of R-CHOP plus two cycles of rituximab were as effective as eight cycles of R-CHOP. Of note, this trial also found that patients who achieved a partial response after four cycles received no additional benefit from receiving a total of eight cycles, as compared to the standard six cycles. Subsequent to this study, two other studies addressing the question of R-CHOP every 21 days versus every 14 days in DLBCL, one from the United Kingdom and the other from GELA, have not shown any benefit in the dose-dense group compared to standard R-CHOP every 21 days (^87,88).

An ongoing trial is directly comparing rituximab plus etoposide, prednisone, vincristine, cyclophosphamide, doxorubicin (R-EPOCH) versus R-CHOP in patients with DLBCL, with highly anticipated results expected in 2016. In addition, there are numerous phase II and III trials evaluating the effect of novel agents, including lenalidomide and ibrutinib, when added to standard-of-care therapies. At MDACC, our preference is that all patients with DLBCL be evaluated for a clinical trial at each treatment stage.

Special Types and Situations in Diffuse Large B-Cell Lymphoma

Primary Central Nervous System and Ocular Lymphoma

The treatment of patients with primary CNS lymphoma is limited to drugs that can cross the blood-brain barrier. Standard chemotherapies such as R-CHOP do not cross to the brain, and they have limited activity in this condition. The initial evaluation must include slit-lamp evaluation of the eyes, MRI of the brain and spine, lumbar puncture, and the standard studies for any other lymphoma to exclude systemic disease. The most common histology in primary CNS lymphoma is DLBCL, and the most important drug is high-dose methotrexate, in general at doses higher than 3.0 g/m². The combination of chemoimmunotherapy using rituximab, high-dose methotrexate, procarbazine, and vincristine is generally considered the standard of care (⁸⁹). In this approach, patients have historically received consolidation with low-dose RT if in complete remission, although there is controversy regarding the role of RT. Lower doses of radiation have decreased the long-term neurotoxicity seen in prior studies (⁹⁰). In the most widely used approach, patients receive consolidation with high-dose cytarabine after completion of RT. Recent studies have shown promising results without the use of RT, although larger studies are required to evaluate whether omitting RT due to concerns about neurotoxicity is viable (⁹¹). The use of intrathecal chemotherapy is controversial in patients with no evidence of leptomeningeal involvement, but it is used in patients with CSF disease. Radiation fields should include the eyes if those are thought to be involved.

Testicular Lymphomas

Most patients with testicular lymphoma have DLBCL, but other lymphoma types can be seen. Patients with testicular lymphoma often have a worse prognosis compared to other DLBCL patients without testicular involvement. The treatment in this group of patients should include prophylaxis of CNS relapses with intrathecal chemotherapy and RT to the contralateral testicle to decrease the risk for localized relapses (¹⁹).

Intravascular Lymphomas

These lymphomas are traditionally considered to have poor prognosis. They have better outcomes with the addition of rituximab to the standard chemotherapy regimen. They have a high incidence of CNS relapse, but there are no standard CNS prophylaxis recommendations in this group of patients (¹⁹).

Primary Mediastinal B-Cell Lymphoma

Historically, PMBL had been an NHL subtype with a poor prognosis, despite intensive therapy and young fit patients (⁹²). These tumors are usually CD20⁺, and thus rituximab is now incorporated with standard-of-care chemotherapy. Consolidation with RT after chemotherapy has been a standard practice, but investigators from the National Institutes of Health have shown that dose-adjusted R-EPOCH without radiotherapy can achieve excellent results (⁹³). These data have not yet been confirmed in a multicenter trial, and thus controversy exists regarding whether RT can be omitted in all patients or may still play a role in patients with a residual large non–FDG-avid mass. At MDACC, our current practice is treat PMBL patients with dose-adjusted R-EPOCH and omit RT for all patients who achieve a complete response.

Treatment of Advanced Mantle Cell Lymphoma

Mantle cell lymphoma is considered a special case because of its recognized aggressiveness and frequent refractory behavior (Fig. 8-19). In many studies of patients with MCL, the disease has been shown to be the NHL type with the poorest prognosis overall, with complete and partial response rates of 29% and 45%, respectively, when treated with a CHOP-like regimen. Investigators at MDACC have investigated hyper-CVAD, a regimen of fractionated cyclophosphamide and continuous infusion doxorubicin, vincristine, and dexamethasone alternating with methotrexate and cytarabine, which had previously been used for patients with acute lymphoblastic leukemia (Table 8-6) (⁹⁴). In long-term follow-up of R-hyper-CVAD alternating with rituximab plus methotrexate and cytarabine in untreated MCL, Romaguera and colleagues reported a 97% response rate, 87% complete response rate, and median time to treatment failure of 4.6 years (⁹⁴). The median OS had not been reached at a median follow-up of 8 years. Because the most toxic portion of this treatment is the high-dose methotrexate-cytarabine cycle, Kahl et al used a modified R-hyper-CVAD with maintenance rituximab for 2 years, obtaining a 77% overall response rate, a complete response rate of 64%, and a median PFS of 37 months (⁹⁵). Many other groups have incorporated a consolidation phase with autologous stem cell transplantation after induction chemotherapy with various regimens with remarkable results (⁹⁶). Bendamustine with rituximab, in comparison to R-CHOP, achieved a statistically significant prolongation in PFS and thus has become a new standard-of-care option for patients unfit to undergo more intensive approaches (⁹⁷).

Special Considerations

Patients with double-hit lymphomas, defined as the presence of MYC translocation with BCL2 or BCL6 or another oncogene, have a very poor outcome even with aggressive treatments (²⁹). Consolidation with transplant in first response should be considered, even though transplant outcome data have not shown clear benefit due to small sample sizes in the largest retrospective series (^98,99). At MDACC, our approach to double-hit DLBCL currently includes dose-adjusted R-EPOCH and consideration of autologous stem cell transplantation consolidation if the patient is fit, but clinical trials are always preferred. The best treatment approach for patients with DLBCL with double protein expression is unclear (¹⁰⁰), and additional clinical trials are needed.

Central nervous system prophylaxis remains controversial. However, it is recommended in patients with high-grade lymphomas or BL, bone marrow involvement with DLBCL, renal or adrenal involvement, two or more extranodal sites, testicular involvement, and disease involvement in areas close to the CNS (^19,57).

With the widespread use of effective antiretroviral therapy, patients with HIV-associated lymphomas have an improved prognosis and should be treated with curative intent (¹⁰¹). Standard treatments with rituximab-containing regimens, such as R-CHOP, dose-adjusted R-EPOCH, or continuous infusion cyclophosphamide, doxorubicin, and etoposide, reported acceptable results, especially in the era of the HAART. In HIV-associated BL, investigators from the National Cancer Institute (NCI) have shown that a short course of R-EPOCH can be highly effective (¹⁰²). It is important to note, however, that when EPOCH-based treatments are used, it is often recommended that HAART be stopped while on treatment.

Treatment of Recurrent/Refractory Diffuse Large B-Cell Lymphoma

Approximately 10% of patients treated for aggressive NHL fail to achieve a complete remission after induction therapy; their disease is termed primary refractory (Fig. 8-20). A larger portion of patients with aggressive NHL, up to a third of all patients, will relapse after initially responding to chemotherapy. Although these patients may be sensitive to a second chemotherapy regimen, most patients with refractory and relapsed aggressive NHL have a poor prognosis. Conventional salvage therapy includes rituximab combined with standard chemotherapeutics such as ifosfamide, etoposide, taxanes, and platinum compounds. Among the most commonly used are DHAP (dexamethasone, cytarabine, and cisplatin), ICE (carboplatin replacing cisplatin), GDP (gemcitabine, dexamethasone, and cisplatin), TTR (paclitaxel, topotecan, and rituximab) and ESHAP (etoposide, methylprednisolone, cytarabine, and cisplatin) (^103,104,105). The salvage regimens tend to have higher toxicity and require greater support for administration, often including hospitalization.

Because of the poor prognosis in patients with relapsed disease, the purpose of many of the chemotherapy regimens offered in this clinical setting is to attain remission followed by high-dose chemotherapy with stem cell support. The Parma trial examined autologous bone marrow transplantation versus salvage chemotherapy in patients with relapsed, chemotherapy-sensitive NHL (¹⁰⁶). Patients were randomized to four more cycles of DHAP versus high-dose therapy with stem cell support, showing an event-free survival rate of 46% in the high-dose arm but only 12% in the DHAP alone arm. This is considered strong evidence that high-dose therapy with stem cell support is the treatment of choice for patients with chemosensitive relapsed or primary refractory aggressive NHL.

Gisselbrecht et al reported a GELA study known as the CORAL trial for second-line treatment for 400 recurrent/refractory CD20⁺ DLBCL patients randomized to receive rituximab plus DHAP (R-DHAP) versus rituximab plus ICE (R-ICE); responding patients received autologous stem cell transplantation. No difference in the response rates (~50% in prior rituximab-exposed patients), PFS, and OS were noted between the R-ICE and R-DHAP groups (¹⁰³).

Patients who do not respond to a second-line treatment or are unfit for an aggressive approach should be evaluated for a clinical trial. If not eligible for a trial, they should then be considered for a palliative treatment that may provide meaningful transient benefit or palliative care. Patients who are responding to second-line treatment, but who are unable to mobilize stem cell treatment, should be considered for alternative donor transplant. Treatment response failure beyond second relapse indicates an incurable disease. A small portion of patients, however, may be rescued with salvage chemotherapy and a second stem cell transplantation, often of allogeneic donor origin; however, this is unlikely, and investigational agents should be prioritized over further chemotherapy. At MDACC, our priority for patients who are either unfit for aggressive therapy or resistant to second-line chemotherapy is to aggressively pursue a clinical trial.

Patients with a history of follicular lymphoma with subsequent transformation to DLBCL, who had prior doxorubicin-based treatment, should be treated with salvage therapy for recurrent DLBCL.

Treatment of Recurrent Mantle Cell Lymphoma

Decisions regarding salvage therapy for patients with recurrent MCL should be individualized, depending on their candidacy for stem cell transplantation. Aside from aggressive combination chemotherapy, targeted therapy agents are active against MCL, and bortezomib, lenalidomide, and ibrutinib have each been approved by the US Food and Drug Administration (FDA) for relapsed MCL.

Recurrent Primary Central Nervous System and Ocular Lymphoma

The treatment of recurrent primary CNS lymphoma is limited because of the inability of many drugs to penetrate into the CNS. Retreatment with high-dose methotrexate can be attempted if there was a long duration of the first remission. Patients with recent prior whole-brain RT may be at high risk for methotrexate-induced encephalopathy. Reports using temozolomide in combination with rituximab have been encouraging. Responding patients may benefit with consolidation with high-dose chemotherapy (especially with preparative regimens including thiotepa), followed by autologous stem cell transplantation.

Patients with highly aggressive NHL have largely benefited from the successful application of pediatric therapy regimens to the adult population, with long-term remissions approaching 80% to 90% in some series (Fig. 8-21). The most important principle for treating patients with highly aggressive NHL is prompt systemic therapy, as these are medical emergencies. Attempts should be made to maintain dose intensity and density using supportive therapies, such as growth factor support, prophylaxis for tumor lysis syndrome, and CNS prophylaxis.

Patients with BL should not be treated with CHOP or CHOP-like regimens due to poor long-term disease-free survival (¹⁹). Combined-modality therapy appears to add toxicity without any proven benefit. Patients at MDACC have been treated with the hyper-CVAD/methotrexate/cytarabine regimen, with intrathecal methotrexate/cytarabine CNS prophylaxis with some success (¹⁰⁷) (see Table 8-6). The NCI has developed an alternating regimen of cyclophosphamide, doxorubicin, vincristine, methotrexate, leucovorin, and ifosfamide, etoposide, and cytarabine with intrathecal cytarabine and methotrexate called CODOX-M/IVAC (¹⁰⁸). This is administered for four cycles, with CODOX-M as cycles 1 and 3 and IVAC as cycles 2 and 4. A complete remission rate of 92% was reported, with a 3-year event-free survival rate of 85%. At MDACC, rituximab was added to standard hyper-CVAD alternating with methotrexate-cytarabine (see Table 8-6), resulting in impressive long term outcomes. Adverse risk factors for these patients include elevated LDH, age, and leukemic presentation (¹⁰⁹). Finally, the EPOCH chemotherapy regimen developed at the NCI has recently shown great promise in a small phase II trial, in both HIV-positive and HIV-negative BL, with freedom from progression rates in excess of 90% with 5 years of follow-up (¹⁰²).

As has been repeatedly shown, patients with aggressive NHL over the age of 60 years have a worse prognosis (^21,110). Unfortunately, more than 50% of patients with aggressive NHL are over 60 years old, and most of these patients do not experience extended long-term survival. Treatment of older patients is complicated by higher overall toxicity rates and lower tolerance of aggressive therapies. At MDACC, our practice is to screen older patients with DLBCL for clinical trials with novel therapies.

Although it is sometimes necessary to reduce the doses of chemotherapeutic regimens to treat elderly patients with comorbid conditions, CHOP is generally well tolerated by patients without contraindications to doxorubicin. Growth factor support should be used. Unfortunately, patients with a contraindication to doxorubicin also frequently have contraindications to other therapies, such as platinum-containing regimens. The substitution of etoposide for doxorubicin in patients who have contraindications to anthracyclines appears to be highly effective in a population-based retrospective review (¹¹¹). A multicenter trial found that R-mini-CHOP, a significantly reduced dosing of conventional R-CHOP, resulted in an acceptable compromise between efficacy and toxicity in patients age 80 years and older (¹¹²).

The field of new cancer drugs has evolved rapidly and is difficult to adequately capture in a book chapter. Since 2000, there has been an incredible number of new drugs evaluated in clinical trials for patients with aggressive lymphomas. The typical pattern of drug development includes an evaluation of toxicity and dosing in the relapsed/refractory setting. If tolerable and an early signal of efficacy is found, drugs are then evaluated in larger phase II trials, often in a population enriched for likely responders. The authors of this chapter believe that the evaluation of efficacy in heavily pretreated, refractory aggressive NHL may actually underestimate the true response rate of targeted therapies (¹¹³).

The drugs listed here are not yet FDA approved for DLBCL, BL, or MCL, but appear promising thus far in clinical trials (⁶¹). The goal of this section is not to recommend use of these nonapproved drugs off of a clinical trial but to serve as an indication of the large progress on the horizon. The Bruton tyrosine kinase inhibitor ibrutinib has now been approved for several B-cell malignancies, including MCL. Early trials confirmed a pathway-based prediction of increased efficacy in the non-GCB subtype of DLBCL (¹¹⁴). Based on promising efficacy and acceptable toxicity, ibrutinib is now being evaluated in a randomized clinical trial with R-CHOP (vs R-CHOP + placebo) (¹¹⁵).

Other inhibitors of the B-cell receptor pathway, including Syk, are ongoing but to date have demonstrated modest results in heavily pretreated patients. Inhibitors of the PI3K/Akt/mTOR pathway have shown significant promise across lymphoma subtypes, although as single agents, they have proven more effective in indolent lymphomas (¹¹⁶). Inhibitors of the various isoforms of PI3K are in numerous clinical trials in various B-cell malignancies, with idelalisib already being approved for some NHL subtypes. Inhibitors of TORC1 have shown response rates in the 30% response range across NHL subtypes, and temsirolimus has received an orphan drug approval for MCL in Europe. Newer inhibitors that block both TORC1 and TORC2 may prove more effective, but further trials are needed. Lenalidomide is a drug with a diverse impact on numerous targets and cell types but is often described as a potent immunomodulatory drug and recently was approved for relapsed MCL. In the ABC subtype of DLBCL, lenalidomide blocks expression of interferon regulatory factor 4 (IRF4 or MUM1) and leads to a synthetic lethal response (¹¹⁷). Based on these and other findings, two randomized phase III trials are planned to evaluate R-CHOP with lenalidomide or placebo (¹¹⁸). Other agents appear to have great promise but are at a very early stage of clinical trial development, including selinexor, BCL2 inhibitors including venetoclax, and immune checkpoint inhibitor therapies.

Definitions of Response

Response to therapy is assessed according to criteria based on the anatomic and metabolic changes that occur in disease-involved nodal and extranodal sites (¹¹⁹). A recent consensus statement was issued by leading clinical investigators to attempt standardization of response criteria to be used in clinical trials. There remains controversy about what constitutes an abnormal PET, and it is generally recommended that uptake be compared to the mediastinal blood pool and liver as internal controls and that the results be scored based on these results (¹²⁰) (Table 8-7).

Restaging

Fluorodeoxyglucose PET has proven very useful in assessing responses to therapy and is now considered standard of care for initial posttherapy restaging in FDG-avid lymphomas. Despite this recommendation, FDG-PET/CT should not be used for long-term follow-up imaging after initial response is confirmed (¹²¹) (Figure 8-22).

Surveillance

Follow-up of patients with aggressive NHL after complete remission and cessation of therapy is typically done every 3 to 6 months for 2 years, then annually until year 5, although there is significant controversy about optimal use of surveillance imaging (¹⁹). A large retrospective evaluation of the utility of surveillance imaging to detect recurrence, as compared to patient-reported complaints, found that most DLBCL recurrences were identified based on symptoms and not imaging alone and that outcomes were not different between imaging and symptom-identified recurrences (¹²²). The emergence of blood-based minimal residual disease detection techniques may ultimately make the debate about the utility of imaging moot.

Relapse or Recurrence

The presence of a new lesion, either by anatomic criteria or on FDG-PET scan, is considered relapsed or progressive disease, but at MDACC, we view a biopsy to confirm imaging findings to be essential. Fluorodeoxyglucose PET is nonspecific, and uptake may occur in both benign and malignant tumors, in inflammatory or infectious lesions, and with normal physiologic processes. Sarcoidosis and fungal infections may mimic lymphoma, and biopsy is often necessary to exclude recurrence (Fig. 8-23). A single persistent or new focus of activity, with paradoxical response at other sites of disease, requires further evaluation with a biopsy, because findings may represent a premalignant lesion, such as a thyroid or colonic adenoma, or an incidental second malignant tumor (Fig. 8-24).

Over the last 20 years, remarkable advances have been made in the diagnosis, characterization, and treatment of patients with aggressive NHL. The molecular characterization of disease is finally gaining clinical traction, due in large part to efficacy in clinical trials being preferential for a particular disease subtype. The future is bright for basic science, translational, and clinical research for aggressive lymphomas with a multitude of new therapeutic agents. It is our strong recommendation that all patients be considered for clinical trials to move the field forward.