EVALUATION OF THE RISK OF SECONDARY LEUKEMIAS IN CANCER SURVIVORS

Therapy-related myeloid neoplasms (t-MNs) are severe late complications of chemotherapy and/or radiotherapy in cancer survivors, characterized by dismal prognosis and resistance to treatment. Chemotherapy may also play a role in increasing clonal hematopoiesis (CH), a clinical entity associated with a higher risk of developing hematological malignancies. Moreover, the presence of germline mutations in cancer predisposing genes can increase the risk of t-MNs. Our hypothesis is that chemotherapy can select clones with somatic mutations, leading to secondary neoplasms. We developed and validated a robust experimental and analytical workflow for the identification of germline variants and subclonal CH-related mutations, enabling longitudinal monitoring of CH dynamics in two clinical settings: 73 lymphomas patients and 58 PARP-inhibitor (PARPi)-eligible ovarian cancer (EOC) patients after platinum-based chemotherapy. We used high-sensitivity error-corrected sequencing (limit of sensitivity VAF>0.1%) for CH analysis, in order to detect mutations well below the cutoff of the clinical threshold of CHIP (variant allele frequency (VAF) >2%). In both cohorts, pathogenic germline mutations were more frequent than previously reported, occurred mostly in DNA damage repair (DDR) genes, and showed no correlation with CH mutational burden or gene targets of CH mutations. In lymphoma patients, the prevalence of CH, the number of mutations, and their VAF remained stable across age groups and timepoints up to 5 years after chemotherapy. Clonal evolution analysis revealed transient fluctuations of low-frequency clones without evidence of selective pressure, suggesting that CH expansion is unlikely to increase the risk of t-MNs in this clinical setting. Conversely, in EOC patients PARPi therapy appeared to promote clonal expansion of clones with mutations in DDR genes, particularly in PPM1D and in clones with VAF ≥1%. Longitudinal follow-up revealed complex clonal dynamics that included loss, acquisition and persistence of CH mutations, also for potential drivers. However, the majority of persistent mutations occurred in PPM1D and showed the strongest increase in VAF up to 30 months of PARPi exposure. Although not statistically significant, multiple CH mutations in PPM1D and TP53 were associated with a higher probability of clonal cytopenia of undetermined significance (CCUS). Notably, cases of CHIP/CCUS nearly doubled in the EOC cohort, rising from 16 to 30 patients within 2 years of PARPi treatment. Overall, our findings suggest that pre-treatment screening for CH mutations, especially in DDR genes, should be considered in EOC patients eligible to PARPi therapy. Moreover, close longitudinal monitoring of clonal dynamics may help early identification of patients at the risk of developing t-MNs, warranting eventual early intervention.