P1187: END OF TREATMENT 18F FDG-PET/CT RESPONSE IS PROGNOSTIC FOR OVERALL AND PROGRESSION-FREE SURVIVAL IN PERIPHERAL T-CELL LYMPHOMA: RESULTS OF THE UK NCRI PHASE 2 RANDOMISED CHEMO-T TRIAL PET/CT SUBSTUDY

Abstract

Background: The role of 18F FDG PET (PET) in peripheral T-cell lymphoma (PTCL) remains under evaluation. The UK NCRI phase 2 randomised CHEMO-T trial compared cyclophosphamide, doxorubicin, vincristine and prednisolone (CHOP) with gemcitabine, cisplatin and methylprednisolone (GEM-P) in previously untreated PTCL patients. The primary endpoint, to demonstrate a superior complete response (CR) rate for GEM-P by contrast-enhanced CT (CECT) at end of treatment (EOT) was not met. Aims: To investigate the role of PET in determining baseline bone marrow (BM) involvement in PTCL and as a prognostic tool at EOT. Methods: Patients required a baseline PET (acquired as per protocol) within 28 days of randomisation and a PET at EOT at least 28 days post chemotherapy. Site accreditation, data collation, and quality control for PET imaging was performed by the UK PET Core Lab. BM biopsy was required within 6 weeks of randomisation. PET scans were centrally reviewed at The Royal Marsden by 2 expert nuclear medicine physicians (YD and NH). Baseline BM involvement by PTCL on PET was evaluated. PET response was assessed according to the revised response criteria for malignant lymphoma and by the Deauville score (DS) 5-point scale. Where discordance existed between reviewers, a third reviewer (SC) adjudicated. EOT PET response and PET status (DS 1-3=PET-negative, DS 4-5=PET-positive) were correlated with PFS and OS. Results: Baseline PET data was available for 82/84 patients, 80/82 (98%) had FDG-avid disease (n=2 with non-avid disease (n=1 AITL, n=1 EATL) were excluded). Clinical characteristics were similar to the main trial cohort: median age=63 yrs, male=73% and high-risk IPI=23%. PTCL subtypes were AITL n=33 (41%), PTCL-NOS n=32 (40%), ALK- ALCL n=14 (18%) and EATL n=1 (1%). For 20/79 (25%) patients with biopsy-proven BM infiltration 9/19 (47%) had BM uptake on baseline PET. Of 20/80 patients (25%) with BM uptake on baseline PET, 9/20 (45%) had biopsy-confirmed infiltration. PET had a sensitivity of 47% and a specificity 81% for determination of BM involvement. EOT PET was available for 69/80, n=4 were excluded as data was acquired following a later line (n=3) or complete imaging was not retrievable (n=1); therefore 65/80 had evaluable EOT PET data. The median number of days from EOT to PET was 41. At EOT, PET response was CR n=41 (63%), partial response (PR) n=15 (23%), stable disease n=0 and progressive disease (PD) n=9 (14%); 41 (63%) were PET-negative and 24 (37%) were PET-positive. At a median follow-up of 27 months, PET-negative vs. PET-positive patients at EOT had a 2-yr OS of 72% vs. 52% [HR 0.58; 95% CI 0.24–1.41), p=0.231] and 2-yr PFS of 55% and 29% [HR 0.45 (95% CI 0.23–0.88), p=0.021] respectively. EOT PET response correlation with 2-yr OS was CR=72%, PR=67%, PD=29%; and with 2-yr PFS was CR=55%, PR=32%, PD=25%. A CR by PET (vs. PD) was associated with improved OS [HR 0.245 (95% CI 0.083–0.730), p=0.012] (Figure 1) and PFS [HR 0.265 (95% CI 0.102–0.687), p=0.006]. Adjusting for stratification parameters at randomisation (IPI risk group, PTCL subtype) PET response remained prognostic for both OS and PFS while EOT PET status (positive vs. negative) did not. CECT response remained prognostic for PFS only. Image:Summary/Conclusion: Our findings indicate that PET has insufficient sensitivity for detection of BM involvement for baseline staging. Furthermore, while both EOT CECT and PET response were prognostic for PFS and OS, only PET response remained prognostic for OS following adjustment for stratification factors suggesting it is a superior prognostic tool in PTCL.