Abstract 5033: Clonal origin of therapy-related myeloid neoplasms

Abstract

Abstract Background: Therapy-related myeloid neoplasms (t-MNs) are secondary malignancies that develop in patients (pts) treated with chemo-radiation therapy (CRT). Patient-related risk factors for t-MN susceptibility is unknown. Recent studies reported that canonical hematological driver mutations (DMs) could be detected in peripheral blood (PB) from healthy individuals or pts with solid tumors, a phenomenon referred to as clonal hematopoiesis of indeterminate potential (CHIP). We hypothesized that there are canonical DMs likely important in t-MNs that pre-exist as detectable CHIP in pts at time of diagnosis of their original cancer. Methods: We sequenced 300 leukemia-relevant genes on diagnostic bone marrow (BM) samples from 13 t-MN pts to detect DMs. We then assessed the presence of the same DMs in the pts’ matched PB samples, which were obtained after primary cancer diagnosis but before CRT. Results: Four pts (31%) had t-AML and 9 (69%) had t-MDS. The median age at primary cancer diagnosis and t-MN diagnosis was 62 years (y) (range: 25-74 y) and 66 y (range: 28-77 y), respectively. The median latency time to t-MN development was 3 y (range: 1-8 y). Targeted gene sequencing revealed 17 canonical DMs in 8 genes. At least one clonal DM was detected in each pt. The most frequently detected mutation was TP53 mutations in 4 patients (31%), followed by DNMT3A (N = 3, 23%), IDH2 (N = 2, 15%), TET2 (N = 2, 15%), and RUNX1 mutations in 2 (15%). Analysis of deep sequencing PB sample revealed that DMs identical to those detected in t-MN BM were detected in 7 of 13 pts (54%). For example, Case UID6982 had small cell lung cancer and received concurrent CRT with carboplatin and etoposide. He developed t-AML 3 y later and was found to have an IDH2 p. R140Q mutation in the diagnostic BM with a VAF of 40%. His PB obtained before CRT showed the same IDH2 p.R140Q mutation with a VAF of 14%. The most common DM detected in paired PB samples was TP53 mutation in 3 cases, followed by DNMT3A, TET2, IDH2, and RUNX1 mutation in 1 case each. Of note, DMs that were subclonal in t-MN BM were not detected in the PB at the time of primary cancer diagnosis. The median VAF of the detected variants in the paired PB samples was 6% (range: 0.6-27%). We found no significant differences between pts with and pts without evidence of CHIP in terms of median age at primary cancer diagnosis or median latency time to t-MN development. Conclusion: In this study we have demonstrated evidence of detectable mutations in multiple canonical leukemia driver genes at time of diagnosis and before therapy of primary cancer in pts whose subsequent t-MN also harbored identical DMs. This work suggests that a substantial fraction of t-MN pts likely exhibit CHIP with detectable canonical mutations in a number of leukemia driver genes at time of diagnosis for their primary cancer. Further, these data suggest the potential to develop a risk stratification model based on presence of CHIP with canonical DMs at the time of primary cancer diagnosis. Citation Format: Koichi Takahashi, Feng Wang, Hagop Kantarjian, Doss Denaha, Erika Thompson, Curtis Gumbs, Keyur Patel, Sattva Neelapu, Yan Wang, Zachary Bohannan, Courtney DiNardo, Farhad Ravandi, Zeev Estrov, Simona Colla, Jianhua Zhang, Xifeng Wu, Guillermo Garcia-Manero, Andrew Futreal. Clonal origin of therapy-related myeloid neoplasms. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 5033.