The evolving role of ¹⁸F-FDG PET scans in patients with aggressive non-Hodgkin’s lymphoma

Abstract

Functional imaging by ¹⁸F-fluorodeoxyglucose positron emission tomography (¹⁸F-FDG PET) is being increasingly incorporated into the evaluation of patients with aggressive non-Hodgkin lymphoma (NHL). Its use for the initial staging in combination with computed tomography has now become standard. PET has recently been included in consensus criteria for response after therapy for Hodgkin lymphoma and aggressive NHL. At the end of therapy, PET has a high positive and negative predictive value (PPV, NPV) for relapse in the pre-rituximab era. However, in the rituximab era, there appears to be a higher rate of false-positive PET scans, making the PPV lower, while the NPV remains high. Interim PET scans are an attractive concept for early evaluation of response, and possibly to guide intensification or de-escalation of therapy. Similar to the end-of-therapy scans, the PPV for mid-therapy scans appears to be low in the rituximab era. Trials testing the modification of therapy based on an interim PET scan are ongoing. For surveillance of patients in remission from aggressive NHL, there is as yet no convincing data that monitoring with PET is superior to traditional surveillance. The evidence to date suggests that a positive PET scan during or after rituximab-based therapy for aggressive NHL should be confirmed by a biopsy if major treatment decisions will be made using the results.

Introduction

It is estimated that over 66,000 people in the United States and over 300,000 worldwide are diagnosed with Non-Hodgkin lymphoma (NHL) every year^1,2. There are over 50 distinct subtypes of NHL³, ranging in their clinical behavior from low grade indolent neoplasms to aggressive lymphomas. Although major advances have been made over the last decade in the treatment of these disorders in the form of combination chemotherapy and monoclonal antibody therapy⁴, not all patients achieving a response go on to achieve a cure. Relapses are almost universally observed in low-grade lymphomas and in a significant proportion of patients with aggressive NHL’s. Consequently, the challenge remains to identify patients with aggressive NHL who are at risk for relapse and may benefit from intensification of therapy.

The tools for detecting residual disease in lymphoma have improved over the years. Traditionally, this assessment had been based on anatomical imaging and pathological criteria, for example, disappearance of an involved lymph node by computerized tomography (CT) scanning and clearance of disease in the bone marrow⁵. However, response assessment by CT scanning is limited in that a residual mass does not equate with residual malignancy. Scarring or necrosis can occur in a lymph node after therapy and CT imaging cannot reliably differentiate viable tumor and necrosis or fibrosis in residual masses. The low specificity of CT scanning has been conclusively demonstrated by reports in which biopsies done on post-treatment residual masses had viable tumor only in 5-20% of cases^6,7.

Functional imaging by whole-body ¹⁸Flourine-fluorodeoxyglucose positron emission tomography (¹⁸F-FDG PET) is increasingly being incorporated into the initial staging workup as well as the post-therapy evaluation of patients undergoing treatment for NHL. The whole-body ¹⁸F-FDG PET scan is the most commonly used type of PET scan in aggressive NHL and this modality will be referred to simply as PET for the rest of this review. PET exploits the higher glucose requirements of metabolically active malignant cells compared with normal cells. Thus, the nature of a tumor at presentation and a residual mass after therapy can be characterized; FDG-avid lesions are more likely to have viable tumor as opposed to non-FDG-avid lesions. Multiple studies have suggested the superiority of adding PET to conventional scanning techniques. Herein we will review the current state of the art on the use of PET scans for diagnosis, assessment of treatment response and surveillance of patients with aggressive NHL. There is little evidence to support the use of PET in indolent lymphomas and this will not be covered here.

Review search strategy

Medline/PubMed was searched for English-language articles using the terms “positron emission tomography,” “lymphoma,” “non-Hodgkin lymphoma,” “predictive value,” “false-positive,” “interim restaging,” “surveillance,” and “rituximab.” Original as well as review articles, clinical trials, case series, and meta-analyses were identified. The bibliography of each relevant article was screened for additional references. Abstracts presented at the American Society of Hematology annual meeting were also searched using the above terms. Aggressive NHL subtypes were considered to be diffuse large B-cell lymphoma (DLBCL), mantle cell lymphoma and follicular lymphoma (grade 3). Studies including Hodgkin lymphoma (HL) or indolent NHL were considered whenever the majority of the data pertained to aggressive NHL. For the data presented in tables 1 - 3, individual patient data was not retrieved from the authors; rather data were extracted and extrapolated from the original publications. Because of the limited published data on some areas of this topic, studies were not excluded based on sample size.

Table 1.

Predictive value of an end-of-therapy ¹⁸F-FDG PET scan for relapse in NHL

Series	Year	Rituximab (n)	Total no. of cases	No. of HL cases	Relapsed	PET Positive		PET Negative
						Total	Relapsed n (%)	Total	Relapsed n (%)
De Wit et al18	1997	No	34	17	8	17	8 (47)	17	0
Jerusalem et al 19	1999	No	54	19	14	6	6 (100)	48	8 (17)
Zinzani et al 20	1999	No	44	13	14	13	13 (100)	31	1 (3)
Mikhaeel et al 21	2000	No	33	0	11	6	6 (100)	27	5 (19)
Spaepen et al 22	2001	No	93	0	37	26	26 (100)	67	11 (16)
Kostakoglu et al 24	2002	No	23	10	11	6	5 (83)	17	6 (35)
Mikosch et al 25	2003	No	62	0	27	29	22 (76)	33	5 (15)
Zinzani et al 26	2004	No	75	10	14	16	14 (88)	59	0
Kumar et al 23	2004	No	18	0	7	6	6 (100)	12	1 (8)
Reinhardt et al 27	2005	No	101	34	28	24	20 (83)	77	8 (10)
Juweid et al 28	2005	Yes (29)	54	0	20	19	13 (68)	35	7 (20)
Han et al 29	2008	Yes	48	0	7	9	1 (11)	39	6 (15)
Total			639	127	198	177	140 (79)	462	58(13)

Abbreviations: ¹⁸F-FDG PET: ¹⁸Flourine-Flourodeoxyglucose Positron Emission Tomography, NHL: Non-Hodgkin Lymphoma, HL: Hodgkin Lymphoma.

Table 3.

Predictive value of mid-therapy ¹⁸F-FDG PET scanning for outcome in NHL (excluding risk-adapted studies)

Series	Year	n	Rituximab (n)	Median follow-up (months)	No. of cycles before PET	PPV (%)	NPV (%)	Sens (%)	Spec (%)	Survival endpoint	PET positive outcome (%)	PET negative outcome (%)
Jerusalem et al 31	2000	28	No	17.5	2-5	100	67	42	100	2-year PFS	0	62
Mikhaeel et al 21	2000	23	No	30	2-4	88	100	100	92	NR	NR	NR
Spaepen et al 32	2002	70	No	36	3-4	100	100	85	100	2-year PFS	4 ¥	85 ¥
Kostakoglu et al 24	2002	30 †	No	19	1	87	87	87	87	18-month PFS	10 ¥	75 ¥
Mikhaeel et al 33	2005	102	Some NR	24	2-3	71	90	88	75	5-year PFS	16	89
Haioun et al 34	2005	90	Yes (37)	24	2	44	90	76	70	2-year EFS	43	82
Querellou et al 35	2006	48 §	Yes (21)	16	2-4	83	83	63	94	2-year EFS	20	80
Fruchart et al 36	2006	35	Yes (26)	19	2-3	60	95	90	76	2-year EFS	30	85
Ng et al 37	2007	44	Yes (18)	28	2-4	67	89	67	89	5-year-PFS	29	90
Han et al 29	2008	40	Yes	24	2-4	33	68	33	68	2-year PFS	77	83
Cashen et al‡ 38	2008	50	Yes	15	2-3	25	85	60	53	15-month PFS	70 ¥	90 ¥
Pregno et al‡ 39	2008	42	Yes	22	2	NR	NR	NR	NR	2-year PFS	75	71
Gigli et al‡ 40	2008	42	Yes	15	3	43	93	75	76	2-year EFS	55	90
Dupuis et al41	2009	103	Yes (50)	53	4	71	79	52	89	5-year EFS	36	80

^†13 HL

^§24 HL

^¥Estimated from Kaplan-Meier PFS curve

^‡These data were presented in abstract form only

Abbreviations: ¹⁸F-FDG PET: ¹⁸Flourine-Flourodeoxyglucose Positron Emission Tomography, NHL: Non-Hodgkin Lymphoma, PPV: Positive Predictive Value, NPV: Negative predictive value, Sens: Sensitivity, Spec: Specificity, PFS: Progression-Free Survival, RFS: Relapse-Free Survival, EFS: Event-Free Survival, NR: Not Reported, HL: Hodgkin Lymphoma.

Utility of ¹⁸F-FDG PET for detection of NHL and initial staging

The most common subtypes of aggressive NHL are routinely FDG-avid on PET scanning. The sensitivity and specificity of PET in these histologies exceeds 80% and 90% respectively for detection of nodal and extranodal involvement^8-10. In the pretreatment staging of lymphoma, PET can add to conventional diagnostic testing with CT, leading to upstaging in 10-20% of cases^9,11. Typical examples of upstaging include FDG-avidity in sub-centimeter lymph nodes, which are not significant by CT criteria, and splenic and hepatic infiltration which may not be obvious on CT.

For detection of bone marrow involvement, PET is not reliable with a reported sensitivity of only 43%¹². The specificity of PET in this setting is also low since a diffuse pattern of bone marrow FDG uptake can be seen with reactive myeloid hyperplasia. Therefore, if lymphomatous involvement of the bone marrow is suspected based on a positive PET, this should be confirmed by biopsy if this information will affect management¹³.

Based on the superior sensitivity and specificity of PET compared with contrast-CT for detecting aggressive NHL, PET is being increasingly incorporated into pretreatment staging guidelines. Although upstaging by PET in the initial workup changes the treatment plan in only 10-15% of cases¹⁴, one advantage of having a pretreatment PET is to facilitate the interpretation of equivocal scans after therapy. Currently, there is sufficient evidence to support the use of a combined PET/CT scan as a single imaging modality for the initial staging of patients with lymphomas that are routinely FDG-avid¹³.

Response criteria in lymphoma and the role of ¹⁸F-FDG PET

Due to the variability in the definition of endpoints in older NHL studies, a multidisciplinary international working group (IWG) was convened in 1998 in an attempt to standardize response criteria for NHL⁵. These criteria defined four categories of response – complete response (CR), complete response unconfirmed (CRu), partial response (PR) and relapse/progression. Patients were classified using physical examination, laboratory parameters (e.g. LDH), bone marrow evaluation and CT to measure the size of lymph nodes. The IWG achieved their goal of standardization and these response criteria were widely adopted by clinical trialists, regulatory agencies and clinicians alike.

In 2007, the IWG reconvened to update the response criteria in order to incorporate emerging data on PET scanning and molecular testing¹⁵. This International Harmonization Project (IHP) also expanded the response criteria to include HL. These criteria recommended using post-therapy PET scanning which effectively removed the CRu category as follows; in lymphomas that are routinely FDG-avid such as DLBCL and HL, a post-therapy residual mass greater than 1.5cm (which would have previously been designated as a CRu or PR), would now be classified as a CR if the lesion was PET-negative. The evidence for incorporating post-therapy PET scans into the response criteria is described in the next section.

A major strength of the 2007 IHP guidelines is the precise definition of a positive PET scan, since this was a source of significant variation in older studies. Visual assessment is considered to be adequate and the standardized uptake value is not necessary. A positive scan is defined as focal or diffuse FDG uptake above background in a location incompatible with normal anatomy or physiology. Exceptions include mild and diffusely increased FDG uptake at the site of moderate- or large-sized masses with an intensity that is lower than or equal to the mediastinal blood pool, hepatic or splenic nodules greater than 1.5 cm with FDG uptake lower than the surrounding liver/spleen uptake, and diffusely increased bone marrow uptake within weeks after treatment¹⁵.

¹⁸F-FDG PET for end-of-therapy assessment

The large trials that have shaped the standard of care for the treatment of NHL over the last decade have generally not incorporated PET scanning into their protocols due to limited availability and high cost of PET scans in the past. More recently, prospective trials conducted by the Eastern Cooperative Oncology Group¹⁶ and the Cancer and Leukemia Group B¹⁷ have started utilizing PET scans and the data is forthcoming. Therefore, up until now, all the data on the role of PET for end-of-therapy assessment are derived from small, heterogeneous, mostly retrospective series conducted in patients treated with chemotherapy before incorporation of rituximab as part of the standard therapeutic approach. These data strongly support the utility of PET to predict outcome at the end of chemotherapy for aggressive NHL.

The value of PET scanning after therapy for NHL to distinguish residual viable tumor versus fibrosis or other non-malignant conditions was first described by De Wit et al in 1997¹⁸. This modest series included 34 patients with lymphoma who had conventional imaging as well as PET scans at the end of therapy. Among the 34 patients, 17 were PET-negative and none of these relapsed, while 17 were PET-positive, and 8 of these patients relapsed. Subsequently Jerusalem et al reported similar findings among 54 patients with HL or high grade NHL¹⁹. In this series, all 6 patients with positive post-therapy PET scans relapsed while only 8 out of 48 patients with a negative post-therapy PET scan relapsed. This positive predictive value (PPV) of 100% for post-therapy PET scanning was reproduced in subsequent reports by Zinzani et al²⁰, Mikhaeel et al²¹, Spaepen et al²² and Kumar et al²³. Other authors have reported PPV’s in this setting ranging from 47% to 100%^24-29.

Table 1 summarizes the published NHL series looking at the utility of a post-therapy PET scan for predicting relapse with an aggregate PPV of approximately 79% and a negative predictive value (NPV) of 87%. The pooled estimate of the sensitivity of post-therapy PET scans for detecting relapse is 70% and the specificity is 92%. These data formed the basis for the recommendation for incorporating PET scans into the post-therapy response assessment of aggressive NHL. It is important to note that most of the studies examining post-therapy PET scans in NHL also included patients with HL. In Table 1, 20% of the patients included in these studies had HL. Therefore, it remains possible that the conclusions may be different if the NHL patients are analyzed separately but in most of these reports, insufficient information was available to exclude the contribution of the HL patients to the overall outcome.

Many of the aforementioned reports looking at post-therapy PET response also included data comparing the accuracy of PET versus CT in predicting relapse. The series in which these data are available are summarized in Table 2. When the PET and CT were both positive (PET⁺/CT⁺), the PPV was approximately 79%. When the PET and CT were both negative (PET^-/CT^-), the NPV was 88%. When the PET and CT findings were discordant, the outcome was strongly associated with the PET result; the relapse rate for the PET⁺/CT^- group was 90% as opposed to only 11% for the PET^-/CT⁺ group. These data suggest that PET is a better predictor of relapse than CT overall, and when the findings of PET and CT are discordant, the PET result is a better predictor of relapse.

Table 2.

Comparison of the predictive value of an end-of-therapy ¹⁸F-FDG PET scan versus CT scan for relapse in NHL

Series	No of HL cases	N		PET+ and CT+		PET+ and CT-		PET- and CT+		PET- and CT-
		Total	Relapse	Total	Relapse	Total	Relapse	Total	Relapse	Total	Relapse
De Wit et al 18	17	34	8	16	7	1	1	16	0	1	0
Jerusalem et al 19	19	54	14	5	5	1	1	19	5	29	3
Zinzani et al 20	13	44	14	13	13	0	0	24	1	7	0
Mikhaeel et al 21	0	33	11	5	5	1	1	12	2	15	3
Spaepen et al 22	0	93	37	12	12	14	14	14	1	53	10
Kostakoglu et al 24	10	23	11	1	0	5	5	1	1	16	5
Zinzani et al 26	10	75	14	11	10	5	4	30	0	29	0
Kumar et al 23	0	18	7	6	6	0	0	3	0	9	1
Reinhardt et al 27	34	101	28	24	20	0	0	57	8	20	0
Han et al 29	0	48	7	7	1	2	0	9	2	30	4
Total	127	523	151	100	79	29	26	185	20	209	26
Relapse rate		29%		79%		90%		11%		12%

Abbreviations: ¹⁸F-FDG PET: ¹⁸Flourine-Flourodeoxyglucose Positron Emission Tomography, CT: Computed Tomography, NHL: Non-Hodgkin Lymphoma, HL: Hodgkin Lymphoma.

Most of the available data on post-therapy PET scans in NHL were published before rituximab was routinely used, with only two of the 12 series included in Table 1 containing patients treated with rituximab. In the report by Juweid et al²⁸, 29 out of 54 patients received rituximab and the overall PPV and NPV of PET scanning were 68% and 80% respectively. In our recent report²⁹, all 48 patients who were included in the study received rituximab. Only one out of 9 patients with a positive end-of-therapy PET scan relapsed. The NPV in the study remained high at 85%. However, the PPV for relapse in these two studies of patients treated with rituximab in combination with chemotherapy declined to 11-63% raising concern on the predictive value of post-therapy positive PET scan in NHL patients treated with rituximab. The reduced end of therapy PPV of the PET scan observed in patients treated with rituximab might be attributed to inflammatory changes associated with the recruitment of immune cells to the tumor by rituximab observed in preclinical studies³⁰. Further data is needed to better define the predictive value of PET scans at the end of rituximab-based therapy for aggressive NHL.

¹⁸F-FDG PET for interim assessment of response

Although modern chemoimmunotherapy can cure many patients with aggressive NHL, there remains a subset of high-risk patients who fail to respond or relapse after standard frontline therapy. If these high-risk patients who are not responding or destined to relapse can be identified early, therapy can potentially be changed with the aim of achieving a better outcome. With this in mind, interim PET scanning during therapy for NHL has been advocated as a tool for evaluating early response to treatment. Many reports have attempted to correlate early PET response with overall outcome^{21,24,29,31-41}, and more recently, investigators have attempted to modify therapy based on the results of interim PET scanning^42,43.

The evidence for interim PET scanning as a predictive tool in NHL is inconsistent and a recent systematic analysis attempted to organize and interpret the existing data⁴⁴. The sensitivity of interim PET for DLBCL in this analysis was 50-100% and the specificity was 73-100%. The wide range for the sensitivity and specificity was due to a lack of uniformity in the timing of the scans, interpretation of the scans and study endpoints. The authors concluded that interim PET scans could not be recommended for routine clinical use due to the heterogeneity of the existing data and recommended that interim scans should only be done in the context of a clinical trial.

Table 3 summarizes the studies included in the aforementioned systematic analysis and also includes some additional data that has been published or presented since then. Although this dataset is also very heterogeneous, some trends can be observed. The most striking is the drop in the PPV since 2005, which coincides with the widespread use of rituximab in standard treatment protocols. Prior to 2005, the PPV and NPV of interim PET scans were impressively high and this led to enthusiasm for implementing interim PET scans into everyday practice. However, since 2005, the PPV in studies using rituximab has fallen to as low as 25% in one series³⁸. Despite the apparent fall in the PPV in the rituximab era, the NPV has remained relatively high, suggesting that a negative interim PET scan may be reassuring but a positive interim PET may not provide reliable predictive information. Of note, no similar trend was observed in studies evaluating prognostic value of PET in Hodgkin lymphoma patients, in which rituximab is not used as part of the therapeutic regimens.

In our retrospective analysis of the predictive value of mid-therapy PET in 40 patients with aggressive NHL treated with rituximab-based chemotherapy, the PPV, NPV, sensitivity and specificity of the mid-therapy PET for the prediction of relapse were 33%, 68%, 33%, and 68%, respectively²⁹. There was no difference in progression free survival and overall survival of mid-therapy PET positive and negative patients thus suggesting that the mid-therapy PET is not predictive of survival in patients with aggressive NHL treated with rituximab-based regimens.

There are now two reports from investigators who have taken the initial promising data on interim PET as the basis for designing prospective risk-adapted trials using a positive interim PET as a decision point for escalating therapy^42,43. Kasamon et al at the Johns Hopkins University conducted a prospective risk-adapted study of 59 patients with aggressive NHL, 56 of who had DLBCL. All the patients with DLBCL received R-CHOP (rituximab, cyclophosphamide, doxorubicin, vincristine and prednisone) for 2-3 cycles and then had an interim PET scan. Patients with a negative PET scan went on to complete standard therapy while those with a positive PET (using the IHP criteria) underwent salvage chemotherapy followed by myeloablative therapy and autologous stem cell transplant (ASCT). Among the 56 DLBCL patients, 32 (57%) had a positive interim PET, 27 of whom went on to ASCT. The two-year event-free survival (EFS) was 88% in the PET negative group versus 66% in the PET positive group. The investigators concluded that these results represented an encouraging improvement in outcome in this historically poor risk group of patients. However, it should be noted that persistence of lymphoma, as suggested by a positive mid-therapy scan, was not confirmed by biopsy in this study. Consequently, it is possible, that at least in some of these patients, therapy escalation was done without persistent lymphoma, thus potentially leading to over-treatment and complicating interpretation of the published results.

The Memorial Sloan Kettering Cancer Center also conducted a prospective risk-adapted study in patients with DLBCL based on an interim PET scan after 4 cycles of R-CHOP⁴³. For those with a positive interim PET, a biopsy of the positive site was performed and if evidence of lymphoma was present, the therapy was intensified to include myeloablative chemotherapy followed by ASCT. For those with a negative interim PET, or a negative biopsy after a positive PET, a short consolidative regimen was used. At the interim restaging, 31 out of 86 patients had a positive PET scan, but only 4 of these 31 had a positive biopsy. The authors concluded that 27 out of 86 patients (31%) would have been over-treated if the positive PET scan alone were used to escalate therapy. This prospective trial argues against basing treatment decisions on interim PET scans alone without biopsy confirmation of residual lymphoma and further data from another prospective multicenter trial is forthcoming¹⁶.

¹⁸F-FDG PET for surveillance of patients in remission

The surveillance of patients in complete remission from lymphoma usually consists of regular history, physical examinations and routine laboratory testing. These measures have been shown to detect lymphoma relapse in over 80% of cases, even without the use of screening CT scans^45-48. There is a paucity of evidence regarding the value of screening PET scans for patients who are in remission from lymphoma. In a series by Jerusalem et al of 36 HL patients in remission who were followed prospectively with PET scans every 4-6 months for 2 years⁴⁹, PET identified disease before the onset of symptoms in only 3 of 36 patients with confirmed relapse, but produced false-positive studies in 6 patients, resulting in an additional burden of invasive testing for these patients. Similarly, in our cohort of 40 patients with aggressive NHL, surveillance PET scans were positive in 7 out of 8 patients at the time of lymphoma relapse²⁹. Among the 7 patients with positive PET scans at the time of the relapse, preceding new symptoms or findings on physical examination were present in 4 patients and in 2 relapse was initially diagnosed by follow-up CT scan. PET scan was the initial modality to diagnose relapse in only one asymptomatic patient with non-diagnostic CT scan. There were also 8 false positive follow-up PET scans for detection of lymphoma. In two patients PET findings led to biopsy of residual mediastinal mass and neck lymph node that showed non-specific inflammatory changes, while in an additional 3 patients, watch and wait clinical follow-up demonstrated no evidence of lymphoma recurrence. In these cases the subsequent PET scan studies became negative. Noticeably, in the remaining 3 patients, a positive follow-up PET scans led to diagnosis of unrelated secondary primary cancers.

Contrasting results were recently reported in a large prospective trial by Zinzani et al⁵⁰. This trial included 421 patients - 160 with HL, 183 with aggressive NHL and 78 with indolent FL, all of whom were in complete remission after therapy with a negative end-of-therapy PET scan. These patients underwent serial PET scans at 6, 12, 18 and 24 months and yearly thereafter, and the scans were reported as positive, negative or inconclusive. Patients with inconclusive scans went on to biopsy if there was a high index of suspicion for relapse or if the PET remained positive when repeated after 1-2 months. Almost 1800 PET scans were performed during the course of this study and a number of interesting findings emerged.

Among the 183 NHL patients, the number of relapses as defined by a positive PET scan detected at 6, 12 and 18 months was 10, 5 and 11, respectively. In other words, 14% had a positive PET scan within the first 18 months, and the risk of relapse became insignificant after 18 months. In patients with a PET scan that was interpreted as unequivocally positive, no information was given regarding histological confirmation of relapse. All 36 patients (including 8 with aggressive NHL) with inconclusive PET scans underwent biopsy and 24 (67%) were proven to have relapse. The overall rate of false-positive PET scans in this study was reported as 16/1789 scans (1%) occurring in 12/421 patients (3%). The caveat is that this false positive rate was derived only from the inconclusive-positive group (who underwent biopsies) and all the scans considered as true positives by the investigators were treated as relapse without any information about histological confirmation.

Furthermore, this study did not report whether the identification of early relapse at an asymptomatic stage by PET affected survival. Frequent PET scanning is a costly undertaking and the comparative cost-effectiveness of this approach remains to be determined⁵¹. A recent report also suggests that late pulmonary toxicity after rituximab-based chemotherapy may lead to a false-positive PET scan, and therefore potentially decrease the PPV in this setting⁵². At present there is still insufficient evidence to recommend PET scans for surveillance of patients in remission from aggressive NHL.

Conclusion

PET scans are being widely used in a variety of lymphoma subtypes and at a variety of time points during treatment of patients with aggressive NHL. Because of the heterogeneous application of PET scanning in research trials, considerable effort has been made to standardize the use of PET as well as its interpretation. These efforts as well as ongoing trials will help to elucidate the role of PET scans in particular settings, such as interim restaging and surveillance where there is currently inadequate conclusive data to recommend its use. The evidence reviewed above suggests that PET is useful in the pre-treatment staging and the end-of-therapy evaluation of patients with aggressive NHL treated with chemotherapy alone. While the NPV of negative PET remains high, there is emerging evidence that false positive scans are occurring more frequently in patients treated with rituximab-based regimens. Consequently, a positive PET scan should be confirmed by a biopsy if major treatment changes will be made based on the PET results. Further studies are needed to clarify the predictive value of positive PET both in the mid and end of therapy setting in patients with aggressive NHL treated with rituximab-based regimens.