EARLY POSITRON EMISSION TOMOGRAPHY RESPONSE‐ADAPTED TREATMENT IN LOCALIZED DIFFUSE LARGE B‐CELL LYMPHOMA (AAIPI=0) : RESULTS OF THE PHASE 3 LYSA LNH 09‐1B TRIAL

Abstract

Purpose: R-CHOP is the standard treatment of diffuse large B-cell lymphoma (DLBCL). In localized disease, the FLYER trial (Poeschel et al., Lancet, 2019) showed that 4 cycles of R-CHOP-21 could be as effective as 6 cycles in the younger population. Several studies have shown that response assessment by early positron emission tomography (ePET) was effective to drive treatment for patients (pts) with DLBCL and aaIPI score ≥1 (Casasnovas et al., Blood, 2017). In the LNH09-1B trial, using ePET response at central review, we assessed whether 4 cycles of R-CHOP were non-inferior to 6 cycles of R-CHOP in a limited stage DLBCL patients with favorable prognosis. Methods: We performed a randomized trial to evaluate treatment adaptation on the basis of ePET after 2 cycles of R-CHOP-21 in previously untreated stage I-II DLBCL pts, aged 18-80 with aaIPI=0. The standard arm consisted of 6 cycles of R-CHOP-21, regardless of ePET results. In the experimental arm, ePET-negative pts received 4 cycles of R-CHOP-21 only, whereas ePET-positive patients received 6 cycles, if second PET performed after 4 cycles showed a complete metabolic response (defined by a Deauville score ≤3). Progressionfree survival (PFS) was the primary endpoint. Intention-to-treat (ITT) analysis was used. Results: From December 2010 to May 2017, we enrolled 650 pts, 331 and 319 in the standard and experimental arms respectively, in 74 centers in France and Belgium. Forty-four percent of pts were older than 59 yrs, 4% had a bulky disease (>10 cm), 53% had extranodal disease (head and neck: 51%, digestive tract: 17%, testis: 7%, others: 25%). A central histological review was performed for 90.5% of the cases. With a median follow-up of 5.1 years (IQR: 3.5-6.5), 87 patients (13.4%) had a PFS event. The 3-yr PFS was 89.2% (95% CI 85.3-92.2) in the standard arm and 92.0% (95% CI 88.3-94.5) in the experimental arm. The non-inferiority of the experimental arm versus the standard arm was demonstrated (hazard ratio 0.724, 90% CI 0.504-1.040, p value from Com-Nougue test <0.0001). Superiority of experimental arm was also tested but was not observed (one-sided stratified log-rank p value=0.0702). Toxic deaths were very rare, 0 and 1 in the experimental and standard arms, respectively. Sixty-nine patients relapsed with median time of 25.9 months (range, 4.8 to 75.7), suggesting that late relapses may occurred. Conclusion: This study demonstrates a non-inferiority of 4 cycles of R-CHOP versus 6 R-CHOP for early good responders, confirming that 4 R-CHOP could be the new standard of care of the large majority of limited stage DLBCL patients. Occurrence of late relapses shows the need for long-term follow-up for all pts, even if outcome is very good in this population.