PII-028 - CLONAL HEMATOPOIESIS AS A PRECISION MEDICINE BIOMARKER FOR PATIENTS WITH CANCER

N. Gillis¹, Y. Tang², X. Lu², C. Colin-Leitzinger², J. Teer², M. Teng²; ¹Moffitt Cancer Center and Research Institute, Moffitt Cancer Center and Research Institute, USA, ²Moffitt Cancer Center and Research Institute, USA, United States.

Background: Clonal hematopoiesis of indeterminate potential (CHIP), hematologic malignancy-associated somatic mutations in blood of patients (pts) with normal hematologic parameters, is common in pts with solid tumors and is associated with worse outcomes. We aim to understand the impact of CHIP across cancer types and identify differences in its impact on tumor biology to inform precision medicine decisions.

Methods: We included pts with a diagnosis of breast, colorectal, prostate, or lung cancer with DNA whole exome sequencing (WES) from blood and tumor samples. Raw WES data was aligned to a corrected human genome (GRCh38) reference. Mutations in blood and tumor samples were called using Mutect2 and Lofreq. Filtering removed poor quality and germline variants. Exonic mutations in 80 genes were considered CHIP if the variant allele frequency (VAF) was 0.02-0.4 and VAF in blood was >2x VAF in tumor or if they were known CHIP mutations.

Results: Included were 4071 pts with breast (n=1338), colorectal (n=1132), prostate (n=1032), or lung (n=569) cancer. Overall, 29% of pts had CHIP. The most mutated genes were DNMT3A, TET2, KMT2D, PPM1D, ASXL1, and TP53. CHIP was more common in non-Hispanic than Hispanic individuals (30% v 19%, p< 0.001). Pts with CHIP were older than those without CHIP (65 v 60 yr, p< 0.001) and were more likely to have a history of smoking (53% v 44%, p< 0.001). CHIP was most prevalent in pts with lung (36%) and otherwise similar in prostate (29%), breast (28%), and colorectal (27%) cancers (Fig 1A). There were no differences in genes mutated or VAFs across cancers. Overall, 15% (281/1869) of blood-derived CHIP mutations were detected in the tumor; pts with lung cancer had the most CHIP in tumor (26%, p< 0.001; Fig 1A). CHIP in DNMT3A and TET2 were more likely to be detected in tumors compared to other genes (q < 0.001). We did not observe significantly worse overall (OS) or progression free survival (PFS) in pts with CHIP overall. However, pts with KMT2D-CHIP had worse PFS (adj p=0.009) and a trend for worse OS (p=0.08) compared to pts without KMT2D-CHIP (Fig 1B).

Conclusion: Approximately 30% of pts with solid tumors had CHIP; CHIP was even more common in pts with lung cancer. CHIP mutations are commonly detected in tumors but have variable effects on pt outcomes. Identification of CHIP mutations driving poor outcomes offers opportunities for precision therapeutics.