Redefining early vs late progressive disease (PD) in chronic lymphocytic leukemia: A pooled analysis of fixed-duration therapies

Abstract

Abstract Background: Progression-free survival (PFS) is strongly correlated with overall survival (OS) after time-limited first-line treatment of chronic lymphocytic leukemia (CLL). Current CLL guidelines recommend a switching of drug classes after “early” PD, while patients (pts) with “late” PD may be re-treated with prior therapy. However, the definition of early vs. late PD varies widely (12-36 months after end of treatment [EOT]) and remains under debate. Methods: Seven randomized phase III trials of the German CLL Study Group (GCLLSG) comprising 3,829 pts who received at least one dose of study medication were analyzed. Included studies were CLL4, CLL5, CLL8, CLL10, CLL11, CLL13 and CLL14. 2,921 pts (76.3%) were treated with CIT (bendamustine + rituximab (R), fludarabine +/- cyclophosphamide +/- R; chlorambucil, chlorambucil + R or + obinutuzumab (O)) and 908 with TT (23.7%) (venetoclax + O, venetoclax + R, venetoclax + O + ibrutinib). CIT was administered for a maximum of 6 cycles. TT duration was 12-36 cycles, with 23 pts (2.5%) treated beyond 15 cycles of therapy. Pts were grouped by first-line treatment received (chemo(immuno)therapy [CIT] vs. targeted therapy [TT]). Cox proportional hazards models were estimated for OS from EOT for various thresholds regarding timing of PD, with both PD as categorical covariate (i.e. early, late, and no PD) and time to PD from EOT as time-dependent continuous covariate. The optimal threshold for defining early PD after EOT was determined by the maximum standardized difference in hazard between early and late PD (assessed via z-value) and the maximum model fit. Hazard ratios (HR) were calculated for early and late vs. no PD. Multivariable multinomial regression analyses were performed to determine associations of early and late vs. no PD with known risk factors. Results: After a median observation time of 65 months from EOT, 2,131 (55.7%) pts experienced PD, and 1,143 (29.9%) pts died. Most PD events (1,435 [67.3%|) occurred >12 months after EOT. PD occurred in 1,837 pts (62.9%) treated with CIT and in 294 pts (32.4%) treated with TT. A 12-month cutoff after EOT was identified as the most effective threshold to distinguish early from late PD with respect to OS. This timepoint showed the largest difference in mortality risk between early and late PD (z-value: 8.72) and yielded the best model fit. The same 12-month threshold was confirmed in subgroup analyses of pts treated with CIT (z = 7.42) and TT (z = 3.84), supporting its consistency across treatment modalities. Overall, 696 pts (18.2%) experienced early PD (≤12 months after EOT). These pts had significantly worse OS compared to those without PD (HR 5.43, 95% CI 4.66–6.33, p<0.001). Late PD (beyond 12 months) was also associated with shorter OS, but to a lesser extent (HR 2.52, 2.09–3.04, p<0.001). For CIT, early PD was associated with an HR of4.57 (3.89–5.37, p<0.001) and late PD with an HR of 2.35 (1.94–2.86, p<0.001). In the TT group, early PD was associated with a higher risk of death (HR 9.23, 5.16 – 16.5, p<0.001), whereas late PD was not significantly different from no PD (HR 1.73, 0.81 – 3.71, p=0.157). When analyzing time to PD from study entry (instead of EOT), early PD occurring before 18 months (CIT) or 24 months (TT) was associated with the highest mortality risk, respectively, supporting the 12-month threshold after EOT as a meaningful definition for early PD. Multivariable multinomial regression analyses identified factors associated with early PD. For CIT, these included unmutated IGHV, TP53 aberrations, del(11q), abscence of B-symptoms, CIRS score >6 and creatinine clearance <70ml/min. For TT, unmutated IGHV and TP53 aberrationswere associated with early PD. However, in multivariable Cox proportional hazards analyses, neither IGHV (HR 1.52, 0.91 – 2.5, p=0.108) nor TP53 status (HR 1.65, 0.71 – 3.87, p=0.245) were prognostic for OS after EOT. Conclusion: This analysis challenges the current definition of early vs. late PD commonly used in CLL treatment guidelines in the context of targeted fixed-duration therapy. Pts who experience PD within 12 months after stopping treatment have a significantly higher risk of death, independent of TP53 aberrations and IGHV mutational status, highlighting the high-risk nature of early PD. This group may represent a distinct high-risk population warranting further clinical investigation and tailored treatment strategies.