Risk factor analysis of relapse risk during FLT3-inhibitor post-transplant maintenance therapy following allogeneic hematopoietic stem cell transplantation in Acute Myeloid Leukemia with FLT3-ITD

Abstract Background: FLT3-ITD AML is associated with high relapse rates despite allogeneic hematopoietic stem cell transplantation (HCT). Post-transplant maintenance (PTM) with FLT3 inhibitors (FLT3i) can reduce relapse and improve outcomes. However, relapses still occur, prompting this study to assess the incidence and risk factors for relapse during FLT3i PTM. Methods: We conducted a retrospective multicenter study of 233 patients (pts) who received FLT3i PTM out of 577 FLT3-ITD AML pts who underwent HCT from 2007–2024. The primary endpoint was cumulative incidence of relapse (CIR); secondary endpoints included relapse-free survival (RFS), overall survival (OS), non-relapse mortality (NRM), and chronic GVHD (cGVHD). CIR and NRM were analyzed using cumulative incidence methods considering competing risks. Kaplan-Meier was used for RFS/OS. Mantel-Byar test evaluated the effect of cGVHD while avoiding immortal time bias. Multivariate analysis was performed using Cox or Fine-Gray models. A principal component analysis was used to identify the independence of the risk factors for relapse. Results: FLT3i PTM was given to 233 pts (40.4%), of which 120 (51.5%) were females. The median age at HCT was 46 years (14–73). FLT3-ITD allelic ratio (AR) ≥0.5 at diagnosis, was present in 80 pts (34.3%), while 153 pts (65.7%) had AR <0.5. At HCT, 195 pts (83.7%) were in CR1, 15 (6.4%) in CR2, and 23 (9.9%) beyond CR2 (including active disease up to 10% blasts). The most used FLT3i PTM was sorafenib (n=194, 83.3%), followed by gilteritinib (n=50, 21.5%); 1 pt received midostaurin. Seven pts switched from sorafenib to gilteritinib and 4 vice versa, mostly due to toxicity. PTM was started at a median of 90 days (range: 17–661) post-HCT. The starting doses for sorafenib were mainly 400 mg daily (n=90, 46.4%) or 200 mg daily (n=81, 41.8%); gilteritinib 80 mg daily (n=19, 38%), 120 mg daily (n=16, 32%), or 40 mg daily (n=8, 16%). Dose modifications occurred in 30–40%. At a median of 19-month follow-up, 153 (65.6%) had discontinued PTM due to toxicity (n=63, 41.0%), planned completion (n=48, 31.4%; at a median of 2.1 years), or relapse (n=31, 20.2% at a median of 8.6 (0.55-58.9) months). At 3 years from PTM initiation, the OS was 79.1%, RFS was 74.1%, GRFS was 55.8%, CIR was 20.4%, NRM was 7.0%, with cGVHD of 41%. A total of 40 pts (17.2%) relapsed at a median of 17 months post-PTM and mortality was noted in 45 pts (19.3%). Regarding risk factors for relapse, a higher CIR was associated with remission status at HCT (16.3% in CR1/2 vs 56.9% beyond CR2; p<0.001) and disease risk index (DRI) (CIR: 19.8% for DRI-1, 17.2% for DRI-2, and 62.6% for DRI-3; p<0.001). The presence of cGVHD was noted to be protective with a CIR of 7.4% with cGVHD, vs 30.1% without (p=0.002). Of the 87 pts (37.3%) who developed cGVHD, only 8 pts (9.2%) relapsed. With cGVHD as a time-dependent covariate, the CIR was reduced by 66% in the patients who developed cGVHD (HR 0.34; 95% CI 0.17–0.93; p=0.039). The CR status at HCT as well as cGVHD were confirmed to be independent risk factors for post PTM relapse in a MVA. No significant association with relapse was found for FLT3-ITD AR at diagnosis (≥0.5 vs <0.5), cytogenetics risk (adverse vs non-adverse), concurrent mutations (with DNMT3A, NPM1, WT1, TET2, IDH2, RUNX1 or ASXL1), the use of FLT3i pre-transplant (with induction, re-induction or consolidation), the conditioning intensity (myeloablative vs reduced intensity regimen), or the development of acute GVHD. When remission status at transplant, DRI, and cGVHD were incorporated into a composite risk score, the CIR increased with the number of adverse factors. Patients with no risk factors had a CIR of 12.0%, while those with one and two risk factors had CIRs of 29.4% and 47.4%, respectively. The comparison across groups was statistically significant (p = 0.0000014). CIR could not be reliably estimated for the highest-risk group (score = 3) due to small sample size. Conclusion: The current study strongly suggests a significantly increased risk of relapse in the group of FLT3-ITD AML patients, who do not achieve at least CR2 pre-transplant, have a high DRI or do not develop cGVHD, despite receiving FLT3i PTM. This suggests that those pts should continue FLT3i PTM indefinitely and other methods should be incorporated such as prophylactic DLI within the first 6 months post-transplant to reduce the risk of relapse