Page 102 - Haematologica August 2018

C. Melani et al.
monly used to treat PMBCL but retrospective studies indi-
cate that this therapy alone is inadequate for many patients,9 resulting in the frequent use of consolidative mediastinal radiotherapy, as part of combined modality treatment.6,8,10 It is well documented, however, that medi- astinal radiotherapy is associated with significant late tox- icity including premature death due to cardiovascular complications and second malignancies,11-14 which has led to efforts to minimize its use in mediastinal lymphomas.15- 18 In an effort to reduce mediastinal radiotherapy in PMBCL, we conducted a prospective study of DA- EPOCH-R based on hypothesis-generating evidence that dose-intensive regimens may be more effective and showed DA-EPOCH-R obviated the need for radiothera- py in most patients.4
An important, albeit preliminary observation from this study, was that most patients with a positive end-of-treat- ment (EOT) 18F-fluorodeoxyglucose-positron-emission tomography (FDG-PET) scan achieved durable remissions without further therapy, calling into question the positive predictive value (PPV) of EOT FDG-PET following DA- EPOCH-R.4 This is in line with several other retrospective studies as well as the prospective IELSG-26 study that have eluded to the low PPV and high false-positive rate of EOT FDG-PET imaging in PMBCL; however, a variety of induction chemoimmunotherapy regimens were utilized with most EOT FDG-PET positive patients going on to receive salvage radiotherapy or high-dose chemotherapy with autologous stem-cell transplantation, making the results inapplicable to DA-EPOCH-R or chemotherapy alone.19-22 While it is routine clinical practice to consider a positive (Deauville 4-5) EOT FDG-PET scan indicative of persistent disease and the need for radiotherapy,23,24 our findings raise a potential paradigm shift whereby singular EOT FDG-PET scans are inadequate following DA- EPOCH-R. Indeed, even the significance of a negative EOT FDG-PET following front-line chemoimmunothera- py remains an open question and the subject of a random- ized phase III study of post-treatment radiotherapy versus observation (clinicaltrials.gov identifier 01599559).
To fully characterize the role of EOT and serial FDG- PET imaging on clinical decision making and to provide further data on the clinical outcome of DA-EPOCH-R in PMBCL, we significantly extended enrollment on our phase II trial and independent clinical series. Herein, we provide an in-depth analysis of single EOT and serial FDG-PET scans and long-term patient outcome following DA-EPOCH-R for previously untreated PMBCL.
Methods
Patients/Treatment
Ninety-three PMBCL patients received DA-EPOCH-R on the prospective NCI (N=59) and retrospective Stanford (N=34) study from November 1999 through July 2016. This includes 67 patients from the previously published study4 plus an additional 26 patients; 8 NCI and 18 Stanford. All patients received 6-8 cycles of DA-EPOCH-R (dose-adjusted etoposide, cyclophos- phamide, and doxorubicin with prednisone, vincristine and rit- uximab) with G-CSF support as previously described, without consolidation radiotherapy.4,25 The study was approved by the NCI IRB and all patients provided written informed consent in accordance with the Declaration of Helsinki. clinicaltrials.gov identifier 00001337.
Response Assessment
EOT response assessment was performed using CT in all patients and FDG-PET beginning in September 2002. Published guidelines recommend EOT FDG-PET a minimum of 3 weeks, preferably 6-8 weeks, following completion of chemotherapy.24 All patients with an EOT FDG-PET following the last dose of chemotherapy up to 8 weeks post-therapy (11 weeks post day 1 of the final cycle) were included for analysis. EOT FDG-PET was performed a median 3 weeks (range, 1-10) from day 1 of the final cycle of therapy. Scans were retrospectively scored per the 5-point Deauville scale26 with scores 1-3 negative and 4-5 positive.23,24 Thirty-five of 55 (64%) and 22 of 25 (88%) patients with negative and positive EOT scans, respectively, underwent serial FDG-PET imaging. Tumor biopsy and salvage therapy was implemented per investigator discretion. Surveillance CT scans were performed for up to 5 years post-therapy. One nuclear medicine physician from each institution reviewed and scored all FDG-PET scans from their respective institution without knowledge of clinical outcome. Calculation of metabolic tumor volume (MTV) and total lesion glycolysis (TLG = MTV x SUVmean) was performed on all NCI FDG-PET scans using Osirix version 8.50 (Pixmeo SARL, Bernex, Switzerland).
Statistical Analysis
Overall survival (OS) and event-free survival (EFS) was calculat- ed from the on-study date until date of death or last follow up or date of death, relapse, progression, second lymphoma treatment, or last follow up, respectively. Treatment failure was defined as relapse, progression, or residual disease following therapy. Probabilities of OS/EFS were calculated using the Kaplan-Meier (KM) method,27 with the significance of the difference between a pair of KM curves determined via an exact log-rank test. Characteristics were compared between patients with and with- out evaluable EOT FDG-PET scans and between patients by insti- tution. Dichotomous characteristics, ordered characteristics, and continuous parameters were compared using Fisher’s exact test, an exact Cochran-Armitage test, and an exact Wilcoxon rank sum test, respectively. Linear regression was used in patients with seri- al FDG-PET scans to determine the slope of the change in SUVmax over time. Tests of the slopes being 0 within each group, tests of slopes among the 3 groups, and pairwise comparisons between 2 groups at a time were done using a Wilcoxon signed rank test, an exact Kruskal-Wallis test and Wilcoxon rank sum test, respective- ly. All P-values are two-tailed and not adjusted for multiple com- parisons. Median potential follow up was calculated from the date of enrollment through April 2018, the date of the most recent data update.
Results
Patient Characteristics
Baseline characteristics of the 93 patients from NCI and Stanford were similar aside from a higher proportion of patients with an ECOG of 2-3 (29% vs. 3%, P=0.00058) in the Stanford cohort [Table 1]. Thirteen patients did not have evaluable EOT FDG-PET scans. Reasons included; treatment prior to routine FDG-PET use (N=9), FDG-PET performed prior to the last dose of chemotherapy (N=2), FDG-PET performed later than 8 weeks post completion of chemotherapy (N=1), and extensive brown fat uptake (N=1); exclusion of the 3 patients with FDG-PET scans had no significant impact on the study results or conclu- sions. The 80 remaining patients with evaluable EOT FDG-PET scans had similar baseline characteristics to the
1338
haematologica | 2018; 103(8)