Retrospective Evaluation of BCR::ABL Kinase Domain Mutation Profiles and Treatment Outcomes in Patients with Chronic Myeloid Leukemia to Confirm Clinical Relevance of in Vitro Sensitivity-Based Treatment Switch: Real-World Experience

Introduction

The constitutively active BCR::ABL fusion protein leads to dysregulated tyrosine kinase activity and leukemogenesis in chronic myeloid leukemia (CML). The BCR::ABL kinase domain mutation (KDM) activates signalling pathways, promoting uncontrolled proliferation, evasion of apoptosis, and drug resistance. The emergence of ABL1 KDMs confers resistance to conventional therapeutic agents, including tyrosine kinase inhibitors (TKIs). A strategy to overcome ABL1 KDM involves TKI switch based on the IC50 experiments generated from the BaF3 cell lines harboring ABL1 KDMs. While therapeutic TKI switches based on in vitro IC50 data have been applied in the clinic, there is limited data on real-world outcomes. This study evaluated clinical outcomes of TKI therapy for different ABL1 KDMs in resistant CML patients.

Patients and method

We retrospectively reviewed the ABL1 KDM result of 1,004 patients (pt) registered in our institutional CML registry from 2000 to 2023.The ABL1 KDM test used was Sanger sequencing with a detection limit of 15%. For analysis, KDMs were grouped according to the structural motifs within the KD: 1) T315I, 2) P-loop, and 3) others (including C-helix, drug contact site, SH2 contact, SH3 contact, and A-loop). Primary endpoints were failure-free survival (FFS) and major molecular response (MMR). Outcomes were calculated from the date of KDM diagnosis. FFS was measured until treatment failure or last follow-up. Major molecular response (MMR) was defined as BCR::ABL <0.1% IS.

Results

Out of 1,004 pts, 113 ABL1 KDMs were detected in 104 patients (10.4%), with a median of 40.2 months to detect (range: 3.1 - 270.2). From this cohort, 96 pts developed a single ABL1 KDM, 9 two mutations, and one a triple mutation, and 7 pts developed frameshift, nonsense, or deletion mutations. At the time of KD mutation detection, 11 patients (10%) were in the accelerated phase, 8 (7%) in blast phase, and 80 (71%) in the chronic phase. Regarding the distribution of ABL1 KD mutations, T315I mutations were the most frequent (n=34, 30.1%), followed by P-loop mutations (n=24, 26.4%), and others including: A-loop mutations (n=6, 6.6%), C-helix mutations (n=5, 5.5%), and SH2 contact mutations (n=6, 6.6%). At the time of ABL1 KDM detection, 58 pts (56%) were on imatinib, and 42 pts (41%) were on second or beyond TKI generation (including DAS: 18, NIL: 16, PON: 3, BOS: 2, and ASC: 1), while 4 pts were not on any TKI therapy. Pts on DAS (9 out of 19), NIL (9 out of 17), and PON (2 out of 3) had higher rates of T315I mutations, whereas patients on BOS (2 out of 2) and IMA (37 out of 58) had higher rates of mutations other than T315I or P-loop mutation (p = 0.012). Particularly with IMA, a significant proportion of the patients developed IMA-resistant but 2G-TKI sensitive mutations, including I293V, D276G, M244V or M351T.

MMR was achieved in 46 (40.7%) pts, with a median onset of 280 days (range: 63-5420 days). The MMR rate was lower for patients with T315I mutations (23.5%) compared to P-loop mutations (45.8%) and other mutations (49.0%), (p = 0.049). Median FFS was 507 days (range: 92-3037 days), with FFS at 1 year 54.5% (95% CI: [44.36-63.7]) in overall population.

At the time of KDM detection, 52 pts (58%) switched their TKI therapy. In the patients with T315I mutations: 11 out of 17 switched to PON, 3 to 2G-TKI, and 3 to ASC, with MMR rates of 36%, 33%, and 67%, respectively. Patients with P-loop mutations: 10 out of 12 switched to DAS, 1 to PON, and 1 to NIL, with MMR rates of 50%, 100%, and 100% for DAS, PON, and NIL, respectively. Patients in the other mutation group: 14 out of 21 switched to DAS, 4 to PON, and 3 to NIL, with MMR rates of 64%, 0%, and 67% for DAS, PON, and NIL, respectively. In terms of FFS, there were no significant differences among those with T315I, P-loop, and other mutations. FFS at 1 year was 53.9% [34.7-69.7%] for T315I mutations, 52.2% [30.5-70.0%] for P-loop mutations, and 56.51% [41.4-69.1%] for other mutations (p = 0.7104).

Conclusion

Ultimately, a comprehensive understanding of the implication of BCR::ABL KDM on its selection for TKI switch and clinical outcomes is crucial for optimizing treatment strategies and improving long-term outcomes in patients with ABL1 KDM. Further study with an expanded larger cohort of CML patients using clinical treatment data is strongly warranted to confirm the clinical relevance of the in vitro IC50 data-based TKI switch strategy