Strikingly high false positivity of surveillance FDG-PET/CT scanning among patients with diffuse large cell lymphoma in the rituximab era

Transcription

1 Research Article Strikingly high false positivity of surveillance FDG-PET/CT scanning among patients with diffuse large cell lymphoma in the rituximab era Irit Avivi, 1,2 * Ariel Zilberlicht, 1 Eldad J. Dann, 1,2 Ronit Leiba, 3 Tal Faibish, 1 Jacob M. Rowe, 2,4 and Rachel Bar-Shalom 2,5 Predictive value (PV) of surveillance fluorodeoxyglucose positron emission tomography (FDG-PET) in patients with diffuse large B-cell lymphoma (DLBCL) treated with chemotherapy-rituximab (R) versus chemotherapy only, remains unclear. The aim of the current study was to compare the performance of surveillance PET in DLBCL patients receiving CHOP (cyclophosphamide, hydroxydaunorubicin hydrochloride, vincristine, and prednisone) alone versus CHOP-R. Institutional database was retrospectively searched for adults with newly diagnosed DLBCL, receiving CHOP or CHOP-R, who achieved complete remission and underwent surveillance PETs. Follow-up (FU) PET was considered positive for recurrence in case of an uptake unrelated to physiological or known benign process. Results were confirmed by biopsy, imaging and clinical FU. One hundred nineteen patients, 35 receiving CHOP and 84 CHOP-R, who underwent 422 FU-PETs, were analyzed. At a median PET-FU of 3.4 years, 31 patients relapsed (17 vs. 14, respectively; P ). PET detected all relapses, with no false-negative studies. Specificity and positive PV (PPV) were significantly lower for patients receiving CHOP-R vs. CHOP (84% vs. 87%, P ; 23% vs. 74%, P < ), reflecting a higher false-positive (FP) rate in subjects receiving CHOP-R (77% vs. 26%, P < 0.001). In the latter group, FP-rate remained persistently high up to 3 years post-therapy. Multivariate analysis confirmed rituximab to be the most significant predictor for FP-PET. In conclusion, routine surveillance FDG-PET is not recommended in DLBCL treated with rituximab; strict criteria identifying patients in whom FU-PET is beneficial are required. Am. J. Hematol. 88: , VC 2013 Wiley Periodicals, Inc. Introduction Despite significant improvement in the outcome of patients with diffuse large B cell lymphoma (DLBCL) over the last decade, attributable to the introduction of rituximab into first-line therapy [1,2], 40% of patients experience disease relapse, mainly during the first 2 years after completing therapy [3]. Response to second-line therapy, achieved in about 63% of patients [4], appears to depend on several factors, including tumor mass at relapse [5,6]. Hence, employment of follow-up (FU) strategies allowing detection of relapse at the early stage, may improve survival. The current guidelines recommend following DLBCL patients during remission, employing clinical assessment only and reserving imaging studies for patients whose history, physical examination or laboratory tests raise the suspicion of relapse/recurrence[7 10]. However, present clinical practice varies greatly in regard to this matter, and many lymphoma centers do employ routine surveillance PET scans, aiming to ensure early capturing of relapse, given the aggressiveness of this lymphoma type. Adoption of a surveillance scan policy necessitates the proposed technique to be highly sensitive, allowing an early detection of disease recurrence. Imaging must also be relatively specific, possess a low false-positive rate, thereby hampering undesirable overtreatment, or at least, unjustified exposure to irradiation, and mental stress. The surveillance imaging program, such as scheduled computed tomography (CT) [7] or positron emission tomography (PET) [11] scans, both known to be highly sensitive in detecting DLBCL and defining its stage [10,12], has never been confirmed to contribute to survival prolongation [13,14]. Moreover, a recent retrospective study investigating this crucial issue, failed to confirm a survival advantage for patients whose relapse was captured "pre-clinically" by a routine FU-PET scan [15]. The main limitation of FU-PET in lymphoma patients is known to be its relatively low specificity. However, most VC 2013 Wiley Periodicals, Inc. American Journal of Hematology 400 previous studies assessed PET performance in NHL patients receiving different therapeutic protocols, which may also affect the reliability of their results. Recent, though limited, data suggest that rituximab administration in addition to CHOP may increase the incidence of FP PETs, performed either during induction [16], or at the end of therapy, due to immune-mediated inflammation [16,17]. Significance of rituximab-induced inflammatory response for the reliability of surveillance PETs, performed "long time" after exposure to rituximab, has not been fully explored. The aim of the current study was to compare the performance and validity of surveillance PETs in DLBCL subjects treated with CHOP-R versus CHOP alone, assuming that rituximab, but not chemotherapy, adversely affects the 1 Department of Hematology and Bone Marrow Transplantation, Rambam Health Care Campus, Haifa, Israel; 2 Bruce Rappaport Faculty of Medicine, Technion, Israel Institute of Technology, Haifa, Israel; 3 Biostatistics Unit, Rambam Health Care Campus, Haifa, Israel; 4 Department of Hematology, Shaare Zedek Medical Center, Jerusalem; 5 Department of Nuclear Medicine, Rambam Health Care Campus, Haifa, Israel The authors report no potential conflicts of interest. I.A. was the principal investigator, wrote the article and takes primary responsibility for the article. A.Z. performed research, analyzed data. E.J.D. performed research. R.L. analyzed data. T.F. performed research. J.M.R. wrote the paper. R.B.S. designed and performed research, analyzed data, wrote the article. I.A. and A.Z. contributed equally to this work. *Correspondence to: Irit Avivi, Department of Hematology and Bone Marrow Transplantation, Rambam Health Care Campus and Technion, Israel Institute of Technology, P.O. Box 9602, Haifa 31096, Israel. irit_avivi@hotmail.com Received for publication 4 November 2012; Revised 13 January 2013; Accepted 14 February 2013 Am. J. Hematol. 88: , Published online 20 February 2013 in Wiley Online Library (wileyonlinelibrary.com). DOI: /ajh

2 interpretation of PET studies performed long after completion of therapy. Methods The study was approved by the IRB of the Rambam Medical Center, Haifa, Israel (approval number RMB).The institutional database was retrospectively searched for newly diagnosed adult patients with DLBCL, treated with CHOP or CHOP-R between and respectively, who achieved complete remission (CR) [10], had at least one surveillance FDG PET/CT during CR between 2002 and 2010 and were followed until relapse/death or 12 months after the last FU-PET scan. Surveillance PETs were performed according to the departmental protocol during the study period, including a 3- to 6-month interval in the first year of CR, a 6-month interval in the second year, and one scan per year over the next 5 years of CR, or when relapse was suspected due to clinical or imaging findings. Relapse, defined as disease recurrence within 3 months after therapy completion, was confirmed histologically or by further clinical and imaging follow-up (FU), based on the evidence of remarkable progression in a repeated PET scan [10] in the presence of suggestive clinical findings: either the development of significant palpable lymphadenopathy or/and the appearance of definite B symptoms. According to our departmental policy, patients with suspicious findings on FDG PET were referred for biopsy. However, when biopsy was not technically feasible, repeated imaging and clinical assessment, including a repeated PET study, were performed within 6 months. Data collection Patient information was retrospectively retrieved from clinical files and comprised demographic data, disease-related characteristics at diagnosis, including disease stage, LDH level, presence of B symptoms, International Prognostic Index (IPI) score, treatment protocol, disease status at the end of therapy (confirming CR achievement) and after completion of therapy; censored at relapse or at the last FU, performed on June 1, Data of FDG PET/CT studies performed during FU were recorded, focusing on remission duration pre-fu-pet and anatomic location of suspicious sites seen on each FU-PET. Detailed information regarding the cause of patient referral to each of the scans was retrieved from the referral notes themselves, as well as from patients charts containing detailed account on each time point when PET scan was performed (medical history and physical examination). FDG PET/CT technique and interpretation All FDG PET studies were performed using a PET/CT system combining a dedicated PET scanner and a multi-slice CT (GE Discovery LS, GE, Milwaukee, was used till 02/2010, GE Discovery 690 was used after 02/2010), 60 to 90 minutes after injecting MBq18F- FDG. Patients were instructed to fast for at least 4 hr and their blood glucose was measured to ensure a serum level lower than 11 mmol/l before injection. Each FU-PET/CT study, including PET and PET/CT images, was reviewed visually as part of the study by a team of at least two nuclear medicine physicians experienced in PET/CT reading, who were fully cognizant of updated clinical data and prior imaging results of each patient. In all cases, consensus regarding the final report was reached among the readers. The FU-PET/CT was considered negative for the presence of lymphoma on visual review when no foci of increased FDG uptake were revealed, other than those related to physiologic bio-distribution of the tracer or to a known or presumed benign process. Any other foci of increased FDG uptake compared with the background, unrelated to physiological bio-distribution or a known benign process, were considered abnormal and suspicious for relapse, except for a mild FDG uptake involving calcified lung hilar nodes or small peripheral lymph nodes (head and neck, axilla, and inguinal region) characterized with morphologically benign imaging features. Results of PET scans interpreted as positive were confirmed by tissue biopsy or repeated imaging and clinical assessment performed within 6 months. A positive FU-PET scan was defined as true-positive (TP), and a negative scan was considered false-negative (FN) if relapse was confirmed within the first 6 months after imaging. A negative FU-PET study was defined as true-negative (TN) and a positive scan was defined as false-positive (FP) if no evidence of relapse was revealed for at least 6 months after evaluation. research article Data analysis Clinical characteristics of relapse were analyzed in the whole study group and compared between patients treated with CHOP or CHOP-R. Assessed parameters included: relapse frequency, remission duration, and incidence of asymptomatic relapse. Correlation of relapse frequency to demographic and pre-therapy risk factors, including gender, age, disease stage, and IPI were assessed and compared between the two patient groups. Statistical performance of FU-PET, defining sensitivity, specificity, accuracy, and predictive values for relapse detection was assessed for the whole cohort and separately for patients treated with CHOP versus CHOP-R. The relationship of FU-PET performance to various clinical parameters, including gender, age, stage of disease, and duration of CR till FU-PET were assessed. Analysis of involved sites included characterization of suspicious sites seen on FU-PET in relation to their anatomic location in the study cohort as a whole and in two study groups. Statistical analysis Data were analyzed using SPSS version 18. Descriptive statistics in terms of frequencies, standard normal distribution of quantitative variables was tested by the Kolmogorov Smirnov test. Mann Whitney U test was used to analyze differences in "abnormally distributed" variables. Fisher s exact test was used to determine relation between categorical variables. Logistic regression was employed to determine the variables predicting an FP result. P < 0.05 was considered as significant. The number of studies required to detect a single truly positive scan was determined by dividing the entire number of surveillance PETs by the number of relapse events. A repeated analysis, assessing the number of scans required to capture a single event of asymptomatic relapse was calculated by dividing the entire number of surveillance PETs performed in the absence of symptoms, by the number of relapse events, captured at the same time. Results Data of 137 DLBCL patients, treated during the study period ( ), who achieved CR, were reviewed. Eighteen patients were excluded from further analysis, as no FU-PETs at remission were available for them. One hundred and nineteen patients treated with CHOP-R (84; 71%) versus CHOP only (35; 29%) underwent FU PET scans between 2002 and Patient characteristics of both cohorts are presented in Table I. The median age of the whole series was 59 years (24 88), 59% presented with an advanced disease stage (3), 26% exhibited B symptoms, and 45% had increased LDH levels. There were no statistically significant differences between the two treatment cohorts except a longer FU period, a higher number of FU-PETs per patient and increased frequency of adjuvant radiotherapy administration in patients receiving CHOP only (Table I). Clinical characteristics of relapse Within a median FU of 3.4 years (range ), 31(26%) patients relapsed: 17 in the CHOP and 14 in the CHOP-R group (accounting for 49% of the CHOP-only cohort vs. 17% of the rituximab cohort, P ). Relapse was confirmed histologically (n 5 17; 11 in CHOP and 6 in CHOP-R groups) or clinically (n 5 14), demonstrating progressive disease in consequent PET/other imaging, or physical examination. Notably, 78% of relapsed patients initially presented with advanced disease, with no differences in these predispositions between patients treated with R versus those receiving chemotherapy only. Characteristics of relapsed patients are presented in Table II. Twenty-two patients with diagnosed relapse, presented with B symptoms and/or peripheral lymphadenopathy. The remaining nine relapsed patients (accounting for 29% of relapses) were asymptomatic at the time of relapse, including 3 of 17 (18%) in the CHOP group and 6 of 14 (43%) in the CHOP-R cohort (P 5 NS). Relapse was confirmed histologically in five of these nine asymptomatic patients, and was diagnosed "clinically" in four, based on consecutive American Journal of Hematology 401

3 research article TABLE I. Characteristics of the Entire Patient Cohort and Groups Subdivided According to the Treatment Protocol Entire cohort (N 5 119) CHOP-R (N 5 84) CHOP (N 5 35) P value Median age, years (range) 59 (24 88) 59 (28 88) 58 (24 86) NS Sex (male) 62 (52%) 45 (54%) 17 (49%) NS Performance status (94%) 79 (94%) 33 (94%) NS (6%) 5 (6%) 2 (6%) B symptoms 31 (26%) 20 (24%) 11 (31%) NS Disease stage I 1 II 48 (40%) 35 (41%) 13 (37%) NS III 1 IV 71 (60%) 49 (59%) 22 (63%) LDH 54 (45%) 42 (50%) 12 (34%) NS BM 11 (9%) 8 (9%) 3 (9%) NS IPI (55%) 45 (54%) 21 (60%) NS (45%) 39 (46%) 14 (40%) Radiotherapy 19 (16%) 7 (8%) 12 (34%) No. FU-PETs per patient (median) 4 (1 12) 4 (1 12) 5 (1 11) Median duration (months) of PET-FU (range) 3.4 ( ) 2.9 ( ) 6.4 ( ) < PET scan results, which demonstrated a significant progression, compatible with disease recurrence. It is noteworthy that by the time the consecutive PET scan was performed (10 13 weeks after the initial positive scan) all patients were symptomatic (one had unexplained fever, two had palpable small lymph nodes involving the groin and sub-mandibular area, and one reported on a moderate weight loss of 5% within 2 months). Notably, the outcome of symptomatic patients was not statistically different from that reported for asymptomatic subjects. Sixteen of the 22 relapsed patients diagnosed in the presence of clinically suspicious signs or symptoms, remained alive (76%), compared with 8 (89%) of those diagnosed while being asymptomatic (P 5 NS). Moreover, all the seven patients who died (six in the symptomatic cohort and one in the asymptomatic group) succumbed to disease progression. There were also no statistically significant differences in the outcome of patients presenting with symptomatic versus asymptomatic relapses, irrespective of therapy applied. TABLE II. Characteristics of Relapsed Patients: Entire Cohort and Groups Subdivided According to the Treatment Protocol Entire relapsed cohort (N 5 31) CHOP-R (N 5 14) CHOP (N 5 17) P value Median age (yr) (range) 61 (24 86) 65 (31 86) 58 (24 86) NS Sex (males) 19 (61%) 8 (57%) 11 (65%) NS Performance status (ECOG score) 0 17 (55%) 6(43%) 11 (65%) 1 11 (36%) 7 (50%) 4 (24%) NS 2 3 (10%) 1 (7%) 2 (12%) 3 B symptoms 12 (39%) 6 (43%) 6 (35%) NS Stage I 3 (10%) 1 (7%) 2 (12%) II 4 (13%) 1 (7%) 3 (18%) NS III 8 (26%) 4 (29%) 4 (24%) IV 16 (52%) 8 (57%) 8 (47%) Abnormal LDH level 23(62%) 9(58%) 14(67%) NS BM involvement 6 (30%) 3 (38%) 3 (25%) NS IPI 0 4 (13%) 4 (24%) 1 7 (23%) 3 (21%) 4 (24%) 2 10 (32%) 6 (43%) 4 (24%) NS 3 7 (23%) 4 (29%) 3 (18%) 4 3 (10%) 1 (7%) 2 (12%) 5 IPI 5 international prognostic index; R 5 rituximab. Statistical performance of FU-PET/CT for relapse detection One hundred and nineteen patients underwent 423 FU- PET scans: 309 following CHOP-R (73%) and 114 (27%) following CHOP only. All the 423 PET scans were reviewed as part of the study. Three hundred forty PETs (80%) were interpreted as negative for relapse, with no evidence for recurrence during further follow-up of at least 6 months, indicating a FN rate of 0% (Table III). Eighty-three PETs (20%) were interpreted as positive for relapse, including 31 (37%) eventually determined as TP and 52 (63%) as FP. Overall, FU-PET had sensitivity, specificity, accuracy, PPV and negative PV (NPV) of 100%, 87%, 88%, 37%, and 100%, respectively, for detecting relapse during CR. Performance indices of FU-PET for the two groups are presented in Table III. The incidence of FP scans was significantly higher in patients receiving CHOP-R (46/60, 77%) vs. CHOP alone (6/23, 26%, P < 0.001). Both specificity and PPV were significantly lower in the group receiving CHOP-R compared with CHOP alone (84% vs. 94% and 23% vs. 74%, respectively, P and P < ). Sensitivity, accuracy and NPV were similarly high for patients treated with CHOP-R (100%, 85%, and 100% respectively) and CHOP alone (100%, 95%, 100%, respectively). Four hundred twenty-three FU-PETs detected 31 events of relapse, suggesting that 14 FU-PET studies were required to detect a single event of relapse. The number of PETs required to capture a single event of relapse was significantly higher in patients receiving immunochemotherapy versus CHOP-only (22 vs. 7, respectively). Moreover, a repeat analysis, excluding PET scans performed in the presence of symptoms, showed that 47 scans were needed to capture a single event of asymptomatic progression. Relationship between location of suspicions sites and frequency of FP FU-PET results Three hundred thirty-nine FDG-avid sites suspicious for disease recurrence were detected in 83 positive FU-PET/ CT studies, including 211(62%) FP and 128 TP (38%) sites. Of 339 positive sites, 191(56%) were nodal and 148(48%) extra-nodal. An FDG uptake in a nodal site was more likely to be FP in patients receiving CHOP-R (66/101, 65% vs. 38/90, 42%, P ) compared with subjects treated with CHOP (38/90, 42%, P ). This event was particularly prominent in case of an uptake in a cervical lymph node, with 88% FP studies in CHOP-R (32/36) vs. 42% (8/19) in those receiving chemotherapy only (P ). 402 American Journal of Hematology

4 research article TABLE III. Performance of Surveillance PET Scans for the Entire Cohort and Groups Subdivided According to the Treatment Protocol Entire cohort (n 5 423) CHOP-R (n 5 309) CHOP (n 5 114) P value a PET positive (number/%) 83 (19.6%) 60 (19.4%) 23 (20.1%) NS TP (number/%) 31 (37.3%) 14 (23%) 17 (74%) <0.001 FP (number/%) 52 (62.6%) 46 (77%) 6 (26%) <0.001 PET negative (number/%) 340 (100%) 249 (80.5%) 91 (80%) NS TN (number/%) 340 (100%) 249 (100%) 91 (100%) NS FN (number/%) 0 (0%) 0 (0%) 0 (0%) NS Sensitivity (number/%) 31/31 (100%) 14/14 (100%) 17/17 (100%) NS Specificity (number/%) 340/392 (87%) 249/295 (84%) 91/97 (94%) Accuracy 371/423 (88%) 263/309 (85%) 108/114 (95%) NS NPV 340/340 (100%) 249/249 (100%) 91/91 (100%) NS PPV 31/83 (37%) 14/60 (23%) 17/23 (74%) < a CHOP-R versus CHOP. Relationship between CR duration and frequency of false-positive FU-PETs The median time to a positive PET was similar for subjects treated with and without R (1.2 vs months, P 5 NS). However, the frequency of FP PET in patients receiving CHOP only was higher in scans performed within the first 12 months of CR compared with those performed beyond that period. Conversely, the risk for a FP FU-PET remained constantly high up to 3 years after CR achievement in patients treated with CHOP-R (Fig. 1). Relation of FU-PET results to patient-, disease-, and treatment-dependent parameters Relationship between the number of FP FU-PET and patient, disease- and treatment-related parameters was assessed. Evaluation of the whole series by univariate analysis found rituxmab and low IPI to be the only factors associated with increased risk for FP FU-PETs. Age, sex, and disease stage had no statistically significant impact on the risk for FP test (Table IV). Multivariate analysis of the entire cohort confirmed these findings, demonstrating rituximab to be the most powerful factor for a FP PET, followed by a low IPI (Table V). Univariate analysis, restricted to patients receiving R, found age <60 years, early disease stage (I II) and IPI 3, to be associated with increased risk of FP FU-PET (88% vs. 61.5% P ; 92% vs. 67%, P ; 92% vs. 50% P < , respectively). Low IPI remained a statistically significant factor for a FP FU-PET in multivariate analysis (RR , CI95% , P ). Univariate analysis, restricted to patients from the CHOP-only group, found Figure 1. Relationship between remission duration and frequency of false positive surveillance PETs. none of these parameters to be related to increased risk of FP PET during CR. Discussion PET-FDG, a noninvasive imaging technique visualizing physiological and pathological processes, has been recently established to be the most sensitive and specific method evaluating DLBCL response [10,18], distinguishing between viable lymphoma infiltrates and necrosis or fibrosis[19 25]. The addition of rituximab to the CHOP regimen [1 3] has been suggested to decrease the PPV of PET performed during or at the end of induction [16,17,23,26 33]. However, none of these studies directly compared statistical performance of surveillance PET in patients receiving rituximab versus chemotherapy only, which was done in the current study. Our results confirmed FU-FDG PET/CT to be highly sensitive for detecting relapse in both patient groups. However, less than third of recurrences were detected at an asymptomatic stage. Even the employment of a scheduled intensive surveillance imaging protocol did not ensure early identification of most recurrences, which were actually captured in the presence of already existing symptoms [17,34 37]. Moreover, the outcome of asymptomatic patients appeared to be similar to that reported for symptomatic subjects, though, this statement should be taken with caution, given the small number of preemptively detected relapses. Although surveillance PET has proved to be sensitive, its specificity appears to be insufficient, resulting in a high FP rate. PPV of FU-PET in patients treated with R, was not only low (23%) but appeared to be significantly reduced as compared with patients treated with CHOP only (74%). This observation is in line with the study by Han et al., who observed a low PPV of 19% in surveillance PETs performed for R-treated individuals, retrospectively reviewing 38 DLBCL patients [17], and with more recent findings by El-Galaly et al., who reported on 52 patients with aggressive NHL (including 43 DLBCL patients), in whom PPV of FU-PET approached 21% [17,35]. Interestingly, a large prospective study evaluating statistical performance of surveillance PET in 183 subjects with aggressive NHLs, reported an extremely low FP rate of 1% only [11]. This low incidence is likely to be attributed to different criteria employed to define a FP test in this particular study, where a scan demonstrating an unexplained, nonphysiological uptake, was defined at the time of performance as an inconclusive test and was not included in calculations of FP rate. The definition of PET positivity during CR is still not standardized. In the present study, despite excluding a American Journal of Hematology 403

5 research article TABLE IV. Univariate Analysis for Factors Predicting a FP Surveillance PET for the Entire Cohort of Patients Parameter % of FP tests P value Sex Male 31/50 (62) NS Female 21/33 (64) Age (yr) <60 32/47 (68) NS >60 20/36 (55.6) Stage I II 23/30 (77.6) 0.06 III IV 29/53 (54.7) IPI /49 (77.6) /34 (41) Rituximab CHOP-R 46/60 (76.7) < CHOP only 6/23 (26) substantial number of positive PETs considered as representing a non-specific non-malignant tracer uptake, the false-positive rate remained high, indicating the need for further refinement of interpretation criteria for FU-PET studies. Rituximab, an anti-cd20 monoclonal antibody binding specifically to B-cells, causes lymphoma cell death through complement-mediated cytolysis and antibody-dependent cell cytotoxicity, which lead to inflammation and necrosis, resulting in FDG uptakes. Rituximab administration may therefore induce relatively long inflammatory changes in involved tissues, which would translate into an increased incidence of FP FU-PETs in patients receiving CHOP-R compared with CHOP-only. Additional potential explanation of the low PPV observed in patients receiving CHOP-R, could be the reduced incidence of relapses reported in these patients compared with those treated with CHOP only, which increases the likelihood that a positive PET may not necessarily represent a disease recurrence. The lower relapse rate observed in patients receiving CHOP-R seems to be true rather than related to a shorter FU in this cohort of patients, given that the 2.9-year median duration of PET FU in the former cohort allowed capturing the majority of relapses [1 3]. Furthermore, FP rate seems to be unaffected by both these factors (lower incidence of relapses and relatively short FU of rituximab-treated patients), considering that the incidence of FP FU-PETs in patients treated with CHOP-R remained relatively constant throughout the FU period, accounting for 9% to 15% even at 3 years after completing therapy, while FP rate for those receiving CHOP only dramatically decreased at one year after the end of therapy. These data are in conformity with the study by El-Galaly et al. that reported a constantly high FP rate (8 17%) in PETs performed within 2 years post-therapy [35]. Interestingly, these "durable" FP uptakes have mainly involved nodal sites, particularly the head and neck lymph nodes. This phenomenon may be explained by lymphocyte regeneration in peripheral lymph nodes, lasting for months after completing R treatment and further enhanced even by minor infections (most commonly involving throat and gums), TABLE V. Multivariate Analysis for Factors Predicting a FP Surveillance PET for the Entire Cohort of Patients Parameter RR (95% CI) P value Rituximab administration 10.1 (3 33.9) < Low IPI (0 I vs. II) 5.42 ( ) RR relative risk which resulted in increased FDG uptake in patients previously exposed to the B-cell depleting agent - rituximab. Multivariate analysis of multiple clinical parameters which may affect PET performance in the entire series, confirmed rituximab administration to be the most important factor accounting for a FP surveillance scan (causing a 10-fold increase in this risk), followed by a low IPI. In summary, the introduction of rituximab significantly increased the FP rate and reduced the PPV of surveillance PET in patients with DLBCL, even as late as 3 years after therapy completion. The extremely low specificity suggests this approach to be unjustified for the routine use in all DLBCL patients in the rituximab era. Further studies, exploring the role of surveillance PET in selected high-risk patients are warranted. References 1. Coiffier B, Lepage E, Briere J, et al. CHOP chemotherapy plus rituximab compared with CHOP alone in elderly patients with diffuse large-b-cell lymphoma. N Engl J Med 2002;346: Pfreundschuh M, Kuhnt E, Trumper L, et al. CHOP-like chemotherapy with or without rituximab in young patients with good-prognosis diffuse large-b-cell lymphoma: 6-Year results of an open-label randomised study of the MabThera International Trial (MInT) Group. Lancet Oncol 2011;12: Sehn LH, Donaldson J, Chhanabhai M, et al. Introduction of combined CHOP plus rituximab therapy dramatically improved outcome of diffuse large B-cell lymphoma in British Columbia. J Clin Oncol 2005;23: Gisselbrecht C, Glass B, Mounier N, et al. Salvage regimens with autologous transplantation for relapsed large B-cell lymphoma in the rituximab era. J Clin Oncol 2010;28: Hamlin PA, Zelenetz AD, Kewalramani T, et al. Age-adjusted International Prognostic Index predicts autologous stem cell transplantation outcome for patients with relapsed or primary refractory diffuse large B-cell lymphoma. Blood 2003;102: Lerner RE, Thomas W, Defor TE, et al. The International Prognostic Index assessed at relapse predicts outcomes of autologous transplantation for diffuse large-cell non-hodgkin s lymphoma in second complete or partial remission. Biol Blood Marrow Transplant 2007;13: Elis A, Blickstein D, Klein O, et al. Detection of relapse in non-hodgkin s lymphoma: role of routine follow-up studies. Am J Hematol 2002;69: Tilly H, Dreyling M. Diffuse large B-cell non-hodgkin s lymphoma: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Ann Oncol 2010;21(Suppl 5):v172 v NCCN Practice Guidelines in Oncology, Hodgkin s disease/lymphoma and Non-Hodgkin s Lymphoma. National Comprehensive Cancer Network (NCCN). Available at: asp. Accessed May Cheson BD, Pfistner B, Juweid ME, et al. Revised response criteria for malignant lymphoma. J Clin Oncol 2007;25: Zinzani PL, Stefoni V, Tani M, et al. Role of [18F]fluorodeoxyglucose positron emission tomography scan in the follow-up of lymphoma. J Clin Oncol 2009;27: Isasi CR, Lu P, Blaufox MD. A metaanalysis of 18F-2-deoxy-2-fluoro-D-glucose positron emission tomography in the staging and restaging of patients with lymphoma. Cancer 2005;104: Cheson B. The case against heavy PETing. J Clin Oncol 2009;27: Liedtke M, Hamlin PA, Moskowitz CH, et al. Surveillance imaging during remission identifies a group of patients with more favorable aggressive NHL at time of relapse: A retrospective analysis of a uniformly-treated patient population. Ann Oncol 2006;17: Goldschmidt N, Or O, Klein M, et al. The role of routine imaging procedures in the detection of relapse of patients with Hodgkin lymphoma and aggressive non-hodgkin lymphoma. Ann Hematol 2011;90: Haioun C, Itti E, Rahmouni A, et al. [18F]fluoro-2-deoxy-D-glucose positron emission tomography (FDG-PET) in aggressive lymphoma: an early prognostic tool for predicting patient outcome. Blood 2005;106: Han HS, Escalon MP, Hsiao B, et al. High incidence of false-positive PET scans in patients with aggressive non-hodgkin s lymphoma treated with rituximab-containing regimens. Ann Oncol 2009;20: Cheson BD, Horning SJ, Coiffier B, et al. Report of an international workshop to standardize response criteria for non-hodgkin s lymphomas. NCI Sponsored International Working Group. J Clin Oncol 1999;17: Jerusalem G, Warland V, Najjar F, et al. Whole-body 18F-FDG PET for the evaluation of patients with Hodgkin s disease and non-hodgkin s lymphoma. Nucl Med Commun 1999;20: Spaepen K, Stroobants S, Dupont P, et al. Can positron emission tomography with [(18)F]-fluorodeoxyglucose after first-line treatment distinguish Hodgkin s disease patients who need additional therapy from others in whom additional therapy would mean avoidable toxicity? Br J Haematol 2001;115: Zinzani PL, Chierichetti F, Zompatori M, et al. Advantages of positron emission tomography (PET) with respect to computed tomography in the follow-up of 404 American Journal of Hematology

6 lymphoma patients with abdominal presentation. Leuk Lymphoma 2002;43: Lavely WC, Delbeke D, Greer JP, et al. FDG PET in the follow-up management of patients with newly diagnosed Hodgkin and non-hodgkin lymphoma after first-line chemotherapy. Int J Radiat Oncol Biol Phys 2003;57: Zinzani PL, Fanti S, Battista G, et al. Predictive role of positron emission tomography (PET) in the outcome of lymphoma patients. Br J Cancer 2004;91: Zijlstra JM, Lindauer-van der Werf G, Hoekstra OS, et al. 18F-fluoro-deoxyglucose positron emission tomography for post-treatment evaluation of malignant lymphoma: A systematic review. Haematologica 2006;91: Zinzani PL, Poletti V, Zompatori M, et al. Bronchus-associated lymphoid tissue lymphomas: An update of a rare extranodal maltoma. Clin Lymphoma Myeloma 2007;7: Jerusalem G, Beguin Y, Fassotte MF, et al. Whole-body positron emission tomography using 18F-fluorodeoxyglucose for posttreatment evaluation in Hodgkin s disease and non-hodgkin s lymphoma has higher diagnostic and prognostic value than classical computed tomography scan imaging. Blood 1999;94: Spaepen K, Stroobants S, Dupont P, et al. Prognostic value of positron emission tomography (PET) with fluorine-18 fluorodeoxyglucose ([18F]FDG) after first-line chemotherapy in non-hodgkin s lymphoma: Is [18F]FDG-PET a valid alternative to conventional diagnostic methods? J Clin Oncol 2001;19: Mikosch P, Gallowitsch HJ, Zinke-Cerwenka W, et al. Accuracy of whole-body 18F-FDP-PET for restaging malignant lymphoma. Acta Med Austriaca 2003;30: research article 29. Reinhardt MJ, Herkel C, Altehoefer C, et al. Computed tomography and 18F- FDG positron emission tomography for therapy control of Hodgkin s and non- Hodgkin s lymphoma patients: When do we really need FDG-PET? Ann Oncol 2005;16: Jerusalem G, Beguin Y, Fassotte MF, et al. Persistent tumor 18F-FDG uptake after a few cycles of polychemotherapy is predictive of treatment failure in non-hodgkin s lymphoma. Haematologica 2000;85: Mikhaeel NG, Timothy AR, O Doherty MJ, et al. 18-FDG-PET as a prognostic indicator in the treatment of aggressive Non-Hodgkin s Lymphoma-comparison with CT. Leuk Lymphoma 2000;39: Kostakoglu L, Goldsmith SJ. Fluorine-18 fluorodeoxyglucose positron emission tomography in the staging and follow-up of lymphoma: is it time to shift gears? Eur J Nucl Med 2000;27: Spaepen K, Stroobants S, Dupont P, et al. Early restaging positron emission tomography with (18)F-fluorodeoxyglucose predicts outcome in patients with aggressive non-hodgkin s lymphoma. Ann Oncol 2002;13: Guppy AE, Tebbutt NC, Norman A, et al. The role of surveillance CT scans in patients with diffuse large B-cell non-hodgkin s lymphoma. Leuk Lymphoma 2003;44: El-Galaly T, Prakash V, Christiansen I, et al. Efficacy of routine surveillance with positron emission tomography/computed tomography in aggressive non- Hodgkin lymphoma in complete remission: status in a single center. Leuk Lymphoma 2011;52: Hutchings M. Routine follow-up scanning of patients with lymphoma: Who, when, how, and why? Leuk Lymphoma 2011;52: Petrausch U, Samaras P, Haile SR, et al. Risk-adapted FDG-PET/CT-based follow-up in patients with diffuse large B-cell lymphoma after first-line therapy. Ann Oncol 2010;21: American Journal of Hematology 405