The value of routine surveillance imaging in aggressive NHL in first remission is controversial. Despite the lack of evidence, surveillance imaging is often performed with regular intervals during the first years in remission. The rationale behind this may be a presumed better outcome in patients with early and asymptomatic relapse as compared to those patients presenting with overt lymphoma symptoms. Early detection of relapse could in theory be advantageous for the following reasons: (1) lower tumor burden at relapse (and maybe reduced tumor heterogeneity and prevention of treatment resistant clones) and (2) less disease-related impairment of function and organs with better tolerability of treatment. Low disease stage and ECOG performance status 0–1 at relapse are associated with a better outcome in DLBCL [73]. With less than a quarter of the relapses being imaging detected, routine imaging did not contribute significantly to relapse detection in studies based on imaging modalities less sensitive than PET/CT [74–78]. Furthermore, the number of scans per relapse exceeded 100 per preclinical relapse for both DLBCL and PTCLs in first complete remission, and out of all patients in complete remission entering intensive follow-up protocols, only a very small minority (2–3 %) will experience imaging-detected relapse [78]. Few studies have investigated the value of PET/CT in the surveillance of aggressive NHL, mainly reporting on DLBCL. Generally, the rate of imaging-detected relapse seems to be higher when using regular PET/CT surveillance in first remission, and depending on the surveillance protocol, 31–100 % of relapses were detected by imaging [79–81]. However, with the number of PET/CT scans per relapse ranging between 35 and 120 [81, 82] and an estimated cost of up to US$85,550 per preclinical PET/CT-detected relapse, an indiscriminate use of PET/CT surveillance is not likely to be cost-effective. Restricting the use of PET/CT to high-risk patients (IPI > 2) decreases the number of PET/CT scans per relapse to 22 and may tip the balance toward better cost-effectiveness [81]. In patients with transformed indolent lymphomas undergoing routine imaging with PET/CT at least one time during follow-up, the PET/CT-detected relapses were all indolent histologies and less important from a clinical perspective. On the contrary, cases of relapsed DLBCL were symptomatic [83]. Since FDG uptake is unspecific and not restricted to lymphoma recurrence, the positive predictive value (PPV) of PET/CT surveillance in aggressive NHL has been reported low to moderate with values between 21 and 60 % [79, 81, 84]. The increasing use of PET/CT for lymphoma has probably led to safer recognition of uptake patterns characteristic for lymphoma and hence less false-positive reporting. Still, a confirmatory biopsy of lymphoma suspicious findings on PET/CT is mandatory whenever possible. As opposed to the problematic PPV of PET/CT, the negative predictive value has been reported uniformly high and often 100 % [79, 81, 84]. This means that a negative PET/CT study virtually rules out recurrent disease and the need for further investigations in patients with symptoms suggesting lymphoma relapse. Finally, the use of routine surveillance imaging in aggressive NHL can only be justified if early relapse detection is associated with improved survival. Some studies have suggested that imaging-detected relapse of aggressive non-Hodgkin lymphoma could be associated with improved outcome, whereas others found no difference in outcome [76, 78, 81]. However, the retrospective design of these studies does not allow firm conclusions about the causal role of imaging due the presence of important bias (lead time, length time, and guarantee time). Interestingly, a recent randomized trial of surveillance strategies in Hodgkin lymphoma showed that the use of chest X-ray and ultrasound was just as effective as PET/CT in early relapse detection, but at a much lower cost and with higher PPV [85].

While use of routine imaging in the follow-up of aggressive NHL is still widespread, the use of routine imaging in the follow-up of indolent NHL is not recommended. An observational approach without initial treatment is often chosen for patients with an asymptomatic relapse of an indolent lymphoma. An exception could be specific cases where an asymptomatic relapse could lead to unrecognized compression of vital organs such as vascular structures. In any case, PET/CT has no role in this setting.

Like other cancers, lymphoma is characterized by deregulated cell cycle progression and most anticancer drugs are designed to inhibit cell proliferation. So a tracer enabling imaging of cell proliferation could be useful for both initial characterization and treatment monitoring of the disease. FDG uptake is somewhat correlated with cell proliferation, but this correlation is weakened by a number of factors, including FDG uptake in non-malignant lesions [95, 96]. The nucleoside [¹¹C]thymidine was the first PET tracer to specifically address cell proliferation. Early studies showed that [¹¹C]thymidine could determine both disease extent and early response to chemotherapy in aggressive NHL patients [97, 98]. However, the short 20-min half-life of ¹¹C along with rapid in vivo metabolism has limited the clinical application of [¹¹C]thymidine. The thymidine analogue 3′-deoxy-3′-[¹⁸F]fluorothymidine (FLT) offers a more suitable half-life of 110 min and is stable in vivo [99]. More recent studies have shown that FLT-PET can sensitively identify lymphoma sites [100]. FLT uptake is highly correlated with proliferation rate and may thus be able to distinguish between high- and low-grade lymphomas [101, 102]. And furthermore, recent studies have showed a potential of FLT for imaging early response to treatment in NHL [103–106]. Amino acid metabolism of cancer cells is influenced by catabolic processes favoring tumor growth [107]. It has been shown that increased uptake of amino acids reflects the increased transport and protein synthesis of malignant tissue [108, 109]. This is the background for PET imaging of amino acid metabolism with the labeled amino acids L-[methyl-¹¹C]methionine (MET) and O-2-[¹⁸F]fluoroethyl)-L-tyrosine (FET) [110]. Nuutinen et al. [111] studied 32 lymphoma patients and found MET-PET highly sensitive for the detection of disease sites although there was no correlation between MET uptake and patient outcome. Also, MET-PET has an advantage over FDG-PET in imaging of CNS lymphoma [112]. While these results are encouraging, it should be noted that no studies have shown the usefulness or cost-effectiveness of amino acid or nucleoside tracers in large patient cohorts. Furthermore, high physiological tracer uptake in the abdomen limits the usefulness of these tracers for imaging of abdominal and pelvic lymphomas.

With CD20 expressed as a cell surface antigen in most cases of B-NHL, and with CD20 as an important therapeutic target, it is logical to also target CD20 for the purpose of lymphoma imaging. However, most of the available so-called immunoPET tracers are based on full-size, intact antibodies, and as a result of this, it takes a long time for the background activity levels to drop sufficiently to provide acceptable target-to-background ratios. Therefore, Olafsen and colleagues generated a recombinant anti-CD20 fragment which, labeled with a PET tracer, produced high-contrast images in a small-animal model [113]. This and similar targeted PET methods may in the future provide very lymphoma-specific images and prove useful in the staging and treatment monitoring of NHL.

MRI has long been the main structural imaging methods in pediatric lymphomas, due to the reduction in radiation exposure. MRI is used in adult patients to evaluate central nervous system disease and for disease sites in the head-and-neck region (e.g., NK/T-cell lymphomas of nasal type). Recent studies show that diffusion-weighted MRI early during treatment may identify responders before a reduction in tumor size occurs [114].