In 2009, the estimated number of new cases of NHL in the USA was 65,980, and NHL was the sixth most common cancer in men and the fifth most common in women [3]. NHL resulted in about 19,500 deaths from the disease in 2009—ranking as the ninth and sixth most common cause of cancer deaths for men and women, respectively. The incidence of NHL has been rising since the 1970s, most markedly in the elderly population. The overall increase in incidence is out of proportion to that expected due to the aging population alone and the reasons for this are mostly unclear. Despite this, the death rate from NHL has been decreasing since 1997 due to improvements in therapy [3]. Most cases of NHL occur in individuals greater than 60 years of age.

Alterations of the immune system, certain viruses, and environmental exposures have been implicated in the etiology of NHL, with varying degrees of evidence. For example, both immune suppression (e.g., posttransplant) and immune stimulation (e.g., Sjogren’s syndrome, rheumatoid arthritis) have been associated with increased risk of NHL [53–55]. Viral-lymphoma associations have also been suggested. Environmental exposures linked to NHL include organochlorine-based pesticides, organic solvents, and wood products.

NHL consists of a heterogeneous group of diseases that arise from malignant transformation of lymphoid tissues. Initial classification schemes were based primarily on morphology. Advancements in immunological, molecular, and genetic laboratory techniques have led to several updates and revisions of classification systems for NHL [56–58]. The ability to identify cell surface markers, cytogenetic and molecular markers led to the recognition that subtypes of NHL depend on the stage of the maturation when malignant transformation of the lymphocyte occurs as well as the location of the “normal lymphocyte counterpart” within the lymph node.

The currently used World Health Organization classification uses morphologic, immunophenotypic, genetic, and clinical features to define distinct NHL subtypes and was revised in 2008 [57, 58]. The revision further defined and refined definitions of early/in situ lesions, the relevant distinctions in subtypes based on age, and borderline diagnoses (hard to classify as one subtype versus another due to atypical or overlapping features). This is nicely reviewed by Jaffe [57]. The NHL subtypes are grouped by lineage (B cell or T cell/natural killer (NK) cell) and the maturity of the cell from which the malignancy is derived (Table 7.2). Most (80–85%) are B-cell neoplasms [9]. Generally, NHL may behave in a manner that is indolent, aggressive, or very aggressive. The aggressiveness of a tumor within a histologic subtype can be variable between patients, so it is not included in the official classification scheme.

NHL most commonly presents as painless adenopathy, ­similar to HL, but the rapidity of tumor growth and the exact presentation varies with the histologic subtype and aggressiveness of the tumor. NHL spreads in a manner less ­predictable than HL (more hematogenously than by ­contiguous nodal spread) and is more likely to involve extranodal sites than HL. Systemic “B” symptoms are found in 40% of patients, more common in the aggressive than indolent NHL subtypes [59]. Lymphomatous infiltration of the bone marrow may result in cytopenias, and the presentation of infiltration of other organs may be site specific.

A core or excisional biopsy is usually needed for diagnosis, in order to provide sufficient tissue. Fine needle aspiration alone and cytology often do not provide enough tissue for morphologic, immunologic, and genetic assessment. After diagnosis, the initial workup is similar to that of HL. History and physical exam, laboratory tests, and imaging are performed for staging and assessing organ function prior to treatment. Staging is best performed with PET for NHL subtypes that are typically [¹⁸F]FDG avid (i.e., diffuse large B-cell NHL). Although follicular lymphomas are typically [¹⁸F]FDG avid, indolent lymphomas may sometimes have low metabolic activity and in these instances PET for staging may be less helpful. Variability of [¹⁸F]FDG uptake and staging with PET in NHL is discussed further below.

Bone marrow biopsy is performed in nearly all patients at diagnosis of NHL, and lymphomatous involvement is found in 30–50% of them, more commonly with the indolent NHL subtypes [60, 61]. Laboratory testing is primarily used for prognostic scoring and to test baseline organ function prior to treatment.

The Modified Ann Arbor Staging System is also used for staging NHL [19, 20]. The Cotswold modification made some changes to improve the relevance of the Ann Arbor Staging System for NHL (versus HL), but a continued limitation was lack of inclusion of total tumor bulk instead of just the number of disease sites. Additional factors are important for determining prognosis and treatment for patients with NHL.

The histopathologic subtype primarily determines the prognosis for patients with NHL. Within each subtype of NHL, disease-specific genetic alterations and molecular markers may provide additional information for predicting outcomes and guiding management. The International Prognostic Index (IPI) was developed in patients with aggressive NHL (Table 7.1) [62]. Four risk groups were identified (low  =  IPI 0–1, low intermediate  =  IPI 2, high intermediate  =  IPI 3, and high  =  IPI 4–5) with decreasing 5-year overall survival with increasing IPI score (73%, 51%, 43%, and 26%, respectively) [62].

The original IPI was developed and validated prior to the routine use of rituximab in regimens to treat NHL. Sehn et al. [63] retrospectively evaluated the use of the IPI in 365 patients with diffuse large B-cell NHL treated with rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisolone (R-CHOP). Using the original IPI groupings, only 2 (not 4) risk groups were identified. The revised IPI redistributed the risk groups (very good  =  IPI 0, good  =  IPI 1–2, poor IPI 3–5), and this model was felt to better predict outcome. In the rituximab era, the risk group with 4-year survival <50% was no longer identified [63].

Additional variations of the IPI have been developed, including age- and stage-adjusted IPIs to account for the variability of prognoses within these groups [62, 64]. The usefulness of the IPI also has been shown to provide prognostic information at the time of relapse [65–68].

There are two versions of the IPI for patients with indolent follicular NHL (FLIPI 1 and 2), since the original IPI was designed and evaluated in patients with aggressive NHL. Using the first FLIPI version, overall survival decreases with increasing FLIPI score based on three risk groups (low, intermediate, and high) and provides prognostic information for this population, even in the setting of rituximab [69, 70]. Validation of FLIPI 2 is ongoing.

As with the natural history and prognosis of NHL, treatment varies and depends on the specific NHL subtype. Detailed management strategies, including recommended chemotherapy regimens, are provided by a panel of experts in the NCCN guidelines [71]. Excellent review articles are available for specific questions and agents for the treatment of NHL, including new biologic agents and treatment strategies [72–77]. We will review overall management strategies for the most common NHLs—diffuse large B-cell NHL and ­follicular NHL. These are also probably the most commonly encountered by the nuclear medicine physician or radiologist. Discussion of treatment for the other subtypes is beyond the scope of this chapter and can be found elsewhere [71, 78].

Indolent NHL is generally considered incurable at present. Patients respond to therapy but may have multiple relapses and long disease-free or progression-free periods. A major question for those with indolent NHL is when to initiate therapy. Before rituximab, studies did not show an overall survival advantage for early initiation of therapy in asymptomatic patients despite improvements in remission rates and disease-free survival, even in patients with advanced disease [79, 80].

General indications for treatment of indolent NHL are symptoms, organ compromise, rapid or steady progression of tumor, bulky disease, cytopenias due to lymphoma, and patient preference or eligibility for a clinical trial [71]. The treatment plan and aggressiveness of the chemotherapy regimen used should be tailored for the patient’s disease and clinical status. Common regimens include purine analogs, alkylators, anthracyclines, and rituximab alone or in combination with chemotherapy. Zevalin consolidation has also been FDA approved as a first line therapy of NHL.

Sousou and Friedberg nicely reviewed the pivotal studies that investigated the role of rituximab in patients with indolent lymphoma [81]. Treatment with rituximab alone [82–84] and with chemotherapy [85–88] has improved overall response rates and PFS for patients with indolent NHL in both the frontline and maintenance settings [89]. There is a suggestion toward improved overall survival, but longer term follow-up is needed and perhaps more prominently in certain subsets of patients [90, 91]. Areas of continued study for rituximab are used with immune stimulators to improve efficacy, optimal schedule for maintenance dosing, and resistance mechanisms. Other monoclonal antibodies other than CD20 are also being developed and investigated [92].

In the relapsed setting, indolent NHL may be treated in a similar manner to initially untreated disease. Overall response rates to rituximab in the relapse setting range from 40% to 60% with complete responses in 10–20% and median PFS of 8–24 months [84, 93, 94]. RIT is more active and produces better results than unlabeled antibody alone, and its use in relapsed (as well as the frontline) indolent NHL is further discussed below [95, 96]. Autologous and allogeneic transplants have been performed safely and with antitumor activity. If there is a role for transplant in indolent lymphoma, the optimal role is still being defined.

For aggressive B-cell NHL, systemic chemotherapy is the mainstay of treatment. Diffuse large B-cell NHL is the most common aggressive NHL and is discussed as an example. Several commonly used regimens are R-CHOP and R -CVP, with increasing use of R-bendamustine in the United States. [97–99]. For nonbulky early stage disease, IFRT may or may not be added depending on the number of chemotherapy cycles given. Higher doses of IFRT (30–40 Gy) are added to 6–8 cycles of chemotherapy for early stage, bulky disease. The preference at some major centers is to treat patients with advanced disease on a clinical trial; otherwise, a full course of chemotherapy is recommended [71, 97]. Studies continue to investigate various strategies and schedules for administering R-CHOP.

The role of high-dose therapy and transplant for initial therapy of aggressive NHL is controversial and the subject of clinical trials. [¹⁸F]FDG PET is being used as a prognostic indicator in several of these risk-adapted therapy trials, and these studies are discussed in the section on risk-adapted therapy. In the relapse setting, there are several salvage chemotherapy regimens that may be used with the ultimate goal of transplant in those patients who respond.

T-cell NHL accounts for ∼15% of all lymphomas. Most T-cell lymphomas are clinically aggressive and with a few exceptions (anaplastic large cell, early stage T/NK cell, nasal) most are incurable. CHOP and CHOP-based therapy may be used for initial treatment in these settings [100]. Other approaches have been investigated for T-cell NHL subtypes including investigations of biologic agents (bevacizumab [101] and histone deacetylase inhibitors [102]) and agents against other lymphocyte surface antigens e.g. CD52 (alemtuzumab [103]).

Follow-up for patients with NHL also depends on the subtype of NHL, treatment intent, and the patient’s clinical symptoms. For those treated with curative intent, history and physical exam, laboratory testing, and imaging occur with decreasing frequency over the first 5 years, and then annually. Patients must be monitored for late complications of therapy, which are similar to those discussed for HL.

Most lymphomas accumulate [¹⁸F]FDG [2]. Weiler-Sagie et al. [119] recently readdressed this issue in a large retrospective study of 766 patients with newly diagnosed lymphoma referred for initial staging with [¹⁸F]FDG PET/CT. [¹⁸F]FDG-avid disease was found in all patients with HL, 94% with aggressive NHL and 83% with indolent NHL. The histologic subtypes of lymphoma less likely to be [¹⁸F]FDG-avid were small lymphocytic lymphoma (SLL) (%[¹⁸F]FDG avid—83%), enteropathy-type T-cell lymphoma (67%), extranodal marginal zone lymphoma (67%), MALT marginal zone lymphoma (54%), lymphomatoid papulosis (50%), and primary cutaneous anaplastic large T-cell lymphoma (40%) [119].

To some extent, the location of lymphomatous involvement may also hinder its detection regardless of histopathologic subtype. The cutaneous location accounted for one-third of the 48 non-[¹⁸F]FDG-avid cases in Weiler-Sagie et al.’s study [119], including two with diffuse large B-cell lymphoma. The other most common locations of non-[¹⁸F]FDG-avid disease were bowel, bone marrow, effusions, and mucosal surfaces [119]. Clearly tumor burden has an effect and low tumor burden can be falsely negative.

Weiler-Sagie et al. did not address the degree of [¹⁸F]FDG uptake in the histologic subtypes of lymphoma, but this has been done previously [119]. Broadly speaking, low-grade lymphomas take up less [¹⁸F]FDG than high-grade lymphomas [120–122]. Standardized uptake values (SUV—a semiquantitative measure of tumor metabolism) of greater than 10 had a high likelihood of being aggressive and excluded low-grade lymphoma with a specificity of 81% in one study of patients with NHL [122]. Despite this, there is substantial overlap of [¹⁸F]FDG uptake between lymphoma grades with some low-grade tumors exhibiting high levels of [¹⁸F]FDG uptake (Fig. 7.1). The level of [¹⁸F]FDG uptake likely lies along a continuum based on the histologic grade of the lymphoma, but in an individual patient, the level of [¹⁸F]FDG uptake is probably not an adequate single surrogate marker for histologic grading of lymphoma.

One exception may be when a disproportionately high accumulation of [¹⁸F]FDG is seen in a single lesion compared to the others (Fig. 7.1). This may be indicative of transformation of a low-grade lymphoma to a high-grade lymphoma in the same patient, and [¹⁸F]FDG PET and PET/CT could help guide biopsies of these lesions [122]. Bruzzi et al. [123] retrospectively evaluated 37 patients with chronic lymphocytic leukemia (CLL), with and without signs and symptoms of transformation, with PET/CT. Using a SUV_max cutoff of 5, the negative predictive value of PET for detecting transformation was high (97%); however, the specificity and positive predictive value were low (80% and 53%, respectively) [123].

Bodet-Milin et al. [124] used PET/CT to prospectively guide biopsy in 38 patients with indolent NHL (n  =  23 follicular, n  =  10 CLL, n  =  2 Waldenstrom macroglobulinemia, n  =  3 marginal zone NHL) and clinical suspicion for transformation. Biopsy was obtained at the most accessible site with the highest SUV_max and 17 patients had histologic transformation on biopsy. The median SUV_max and gradient SUV_max (highest SUV_max  −  lowest SUV_max) were higher for those with histologic transformation. Using receiver operator curve analysis, a SUV_max threshold of 14 was optimal for identifying transformation (sensitivity/PPV  ∼  94%, specificity/NPV  ∼  95%) [124]. As expected, sensitivity or specificity could be improved one way or the other with higher and lower thresholds. Bodet-Milin’s study differed from Bruzzi’s in several regards—all patients had a clinical suspicion of transformation, and higher threshold values for transformation likely decreased the false-positive rate. They suggest that a higher threshold may be reasonable for patients with follicular versus chronic lymphocytic leukemias/small lymphocytic lymphoma (CLL/SLL), as CLL/SLL typically has less [¹⁸F]FDG uptake at baseline [124]. These data together suggest that PET may be useful for detecting transformation, but tissue confirmation is still required for confirmation at the present time.

It is well established that [¹⁸F]FDG PET is superior to ⁶⁷Ga-citrate scintigraphy [114–118] and improves primary staging of lymphoma as compared to CT [125–129]. For nodal status, this is primarily due to the ability of [¹⁸F]FDG PET to detect lymphoma in normal sized lymph nodes that would be considered benign by CT size criteria. [¹⁸F]FDG PET also detects more extranodal lesions, for example, in the liver, spleen, bone, and lung, than CT does (Fig. 7.2), with less radiation exposure to the patient.

Biopsy is the gold standard to detect lymphomatous involvement in the bone marrow and is a routine component of staging procedures. Bone marrow biopsy and [¹⁸F]FDG PET have been reported to be approximately 80% concordant [127, 130, 131]. [¹⁸F]FDG PET is more likely to detect bone marrow disease when greater than 30% of the marrow is involved. False-negative [¹⁸F]FDG PET scans for bone marrow disease (PET negative, biopsy positive) occur if less bone marrow is involved, in lower grade histologies, or if the primary nodal disease is not [¹⁸F]FDG-avid. Misinterpretation of diffuse [¹⁸F]FDG uptake in the marrow due to concurrent systemic illness or recent administration of hematopoietic growth factors can potentially lead to false-positive [¹⁸F]FDG PET scans. This can be avoided by a carefully elicited clinical history. Sampling error on biopsy may also result in a “false-positive” PET study, which is actually true positive.

In one study, bone marrow involvement could be confirmed in 7 of 10 patients with negative iliac crest biopsies by [¹⁸F]FDG PET directed rebiopsy (N  =  4) or anatomic imaging (N  =  3) [131]. A meta-analysis investigating the use of [¹⁸F]FDG PET for detecting bone marrow disease revealed sensitivities in the 54–74% range and specificities in the 91–95% range [132]. Only a small number of studies evaluating Hodgkin’s lymphoma were included. When evaluating an [¹⁸F]FDG-avid focus in the bone, it is important to note that the concurrent CT scan may appear normal up to 50% of the time [133] and the context and entire distribution of uptake must be considered. [¹⁸F]FDG PET is complementary to, but cannot replace, bone marrow biopsy (Fig. 7.3).

The addition of [¹⁸F]FDG PET to lymphoma staging procedures correctly alters stage and leads to changes in management in 8–48% of cases [125, 127, 129, 134, 135]. Patients are upstaged more often and consequently receive more aggressive initial therapy; however, downstaging also occurs (Fig. 7.4). The inclusion of [¹⁸F]FDG PET in the primary staging of lymphoma has substantial impact on patient care. Although this has also been shown to be true for low-grade follicular NHL, the role of PET in the staging of other indolent NHLs (i.e., CLL/SLL) may be more limited [136–139].

Despite the above support for the use of [¹⁸F]FDG PET in lymphoma staging, the early studies had several limitations. The studies often included small, heterogeneous patient populations, lacked gold standards at each disease site, and due to the nature of the disease, all [¹⁸F]FDG-positive (or negative) lesions could not be verified histologically. Furthermore, the definition of “PET positive” and interpretative criteria may not have been as well refined before experience was gained in image interpretation. For example, in at least 1.2% of patients referred for an oncology PET scan, a second, incidental primary tumor will be found [140].

A majority of studies investigating the use of metabolic imaging in lymphoma have evaluated [¹⁸F]FDG PET alone. PET/CT improves lesion localization and provides more ­reliable tumor measurements. Tatsumi et al. [141] reported that combined PET/CT imaging may be better than either modality alone. In 53 patients with lymphoma, 48 of 1,537 findings were discrepant between PET and CT; 40 were correct PET findings, 5 were correct CT findings, and 3 remained uncertain. Nine patients were only staged correctly with PET/CT versus CT alone [141]. Allen-Auerbach et al. [142] found that PET/CT was more accurate than PET alone for the detection of lymphoma (94% versus 84%, respectively). PET/CT improves lesion localization, allows detection of lesions that might be missed by either modality alone, and provides more reliable tumor measurements.

Residual masses on anatomic imaging in patients with ­lymphoma are common posttherapy but may not always indicate the presence of active malignancy [143]. A major advantage of PET is its ability to detect functionally viable tumor cells within the mass, distinguishing active lymphoma from fibrosis/necrosis. The information gained from posttherapy PET also provides prognostic information [129, 144–148]. Patients with PET-positive disease posttherapy (Fig. 7.5) have lower 1- and 2-year progression-free and overall survival rates (0–54%) than those who are PET negative (83–95%) (Fig. 7.6), regardless of the presence of a residual mass on CT. PET positivity posttherapy predicts subsequent relapse. Therefore, the results of a posttherapy PET scan may have important therapeutic implications.

Despite its improved prognostic implications, a negative posttherapy PET scan does not exclude the presence of residual microscopic disease that falls below the resolution of the PET scanner [129, 144–148]. The resolution of modern PET scanners falls between 0.5 and 1.0 cm [149] which converts to about 10⁸–10⁹ tumor cells [150]. Patients who are PET negative with a residual mass are more likely to recur than those who are PET negative without residual mass [129, 144–148].

The optimal time to obtain a posttherapy PET is a frequent question of referring oncologists, but the answer is not completely resolved. A minimum of 10 days postchemotherapy has been recommended to avoid false-negative scans due to the early treatment effect of “stunning” [151]. Longer and more variable times post-external beam radiation have been suggested [152], and optimal interval times post-RIT are beginning to be investigated and discussed in the section on RIT [153, 154].

The major criteria for evaluating response to therapy in lymphoma, the International Workshop Criteria (IWC) [155], were revised in 2007 to include a posttherapy assessment by [¹⁸F]FDG PET [156]. Previously, the IWC were based on clinical and laboratory findings and changes in tumor size on anatomic imaging [155]. The revision was based on the numerous studies showing the predictive value of the posttherapy PET scan and a retrospective study in 54 patients with aggressive NHL [157].

In this study, PET scans were obtained 1–16 weeks after the completion of therapy (94% within 2–12 weeks) and were interpreted qualitatively as positive or negative. A positive scan was defined as focal or diffuse uptake in a location incompatible with normal anatomy [157]. A complete response required a negative PET scan, regardless of the size of a residual mass on CT and a negative bone marrow biopsy if it was positive before therapy. The integrated PET and IWC response criteria more accurately classified lymphoma response by identifying a subset of partial responders by IWC who had a complete metabolic response by [¹⁸F]FDG PET and a more favorable prognosis [157]. The criteria were subsequently found to be applicable to indolent NHL and HL [158, 159].

In addition to the response criteria, recommendations were also made about when to use PET in clinical trials in lymphoma. Pretreatment and posttreatment response assessments with PET were routinely recommended for typically [¹⁸F]FDG-avid tumors (DLBCL and HL), although the pretreatment scan was strongly recommended, but not mandatory. Our view is that a pretreatment PET scan is mandatory. Generally, for tumors that would have been considered by many to be incurable (FL and MCL) and tumors that are inconsistently [¹⁸F]FDG-avid, PET was only recommended if the response rate was a trial endpoint (for both) and if the baseline scan was positive since PET is more accurate than CT, a baseline and follow up PET/CT is often obtained in these lymphomas, as well.

A consensus report from the Imaging Subcommittee of the International Harmonization Project in Lymphoma on the acquisition and interpretation of posttherapy PET scans accompanied the revised IWC [160]. The interpretation criteria proposed are summarized in Table 7.3. Several additional notable recommendations were also made: (1) baseline scans are “strongly encouraged” to aid posttherapy scan interpretation, (2) IV contrast CT scans may complement the posttherapy PET scan depending on clinical and baseline information, but the size of lesions on CT should always be reported, (3) new lung lesions on a posttherapy scan must be interpreted with caution to avoid false positives due to inflammation or infection, and (4) specific recommendations on interpretation of bone marrow and residual extranodal disease are discussed. The incremental value of contrast CT appears modest, in some studies, over a non contrast CT. (e.g. Schafer).

The evaluation of posttherapy PET for lymphoma is based on visual qualitative assessment of tumor uptake (Table 7.3), and this seems to differentiate progressers from nonprogressers well. Whether a semiquantitative-based analysis using SUV is needed in the posttherapy setting has not been studied in a large number of patients. An SUV-based analysis may be more useful, however, in the interim assessment setting and is discussed further below.

Multiple consensus reports using PET for response assessment (interim and posttherapy) in lymphoma and other tumors stress the importance of obtaining high-quality, standardized PET and CT scans on dedicated PET systems. Detailed discussions on obtaining such scans are found in reports by the National Cancer Institute [161] and the Netherlands multicenter trial group [162] and reviewed in the revised IWC [160] and PET Response Criteria in Solid Tumors (PERCIST) [163].

The use of [¹⁸F]FDG PET for routinely monitoring the response of low-grade follicular and other indolent lymphomas is debatable. Low-grade follicular and indolent lymphomas are currently considered incurable by many; therefore, primary endpoints for these lymphomas are usually PFS and overall survival (OS). These tumors also tend to have lower levels of [¹⁸F]FDG uptake compared to their aggressive counterparts. In the revised IWC criteria, [¹⁸F]FDG PET scans at baseline and posttherapy for follicular and indolent lymphomas are only recommended in clinical trials when overall or complete response rates are the primary study endpoints and only if the pretreatment scan is “positive” [156]. We believe that since [¹⁸F]FDG PET/CT is more accurate than CT, it is quite reasonable to perform [¹⁸F]FDG PET/CT in place of CT in such patients.