Mohamed E. Salem, Hsih-Te Yang, James Thomas Symanowski, Alberto Puccini, Carol J. Farhangfar, Sherif El-Refai, David King, and Edward S. Kim
Background: Molecular changes and associated acquired resistance of rectal tumors following chemoradiotherapy (CRT) have not been well studied. We aimed to examine CRT-induced molecular changes and prognostic associations in rectal cancer patients (pts) undergoing preoperative CRT followed by surgery.
Methods: A paired analysis using pre-CRT biopsies and the corresponding post-CRT resected tissues of rectal cancer patients undergoing preoperative CRT followed by surgery was performed. Pre- and post-CRT tumor samples underwent next-generation sequencing (NGS) by Tempus xT assay, which detects a panel of 596 gene mutations, including single nucleotide variants, insertions/deletions, copy number amplifications, and gene rearrangements. The cancer driver gene was detected based on positional clustering of gained mutations using OncodriveCLUST. The paired t-test was used to examine differences in tumor mutational burden (TMB) and microsatellite instability (MSI) between pre- and post-CRT samples.
Results: In total, 61pts of median age 61yrs—75% white, 18% black, and 57% male—with localized rectal cancer were studied. NGS testing was performed in 57 pt tumor samples; 43 pts had pre-CRT samples, 48 pts had post-CRT samples, and 34 pts had paired samples. The most frequent genetic alterations seen in the 43 pre-CRT tumor samples were APC (37.2%), ARID1B (30.2%), KMT2C (30.2%), ZFXH3 (25.6%), FLT4 (20.9%), MLLT3 (20.9%), and TP53 (20.9%), whereas in the 48 post-CRT tumor samples, the most frequent mutations were APC (54.2%), TP53 (35.4%), KRAS (27.1%), MKI67 (25.0%), KMT2C (18.8%), APOB (14.6%), and CEBPA (12.5%).
Comparing the pre- and post-CRT samples, no significant differences in TMB (median: 5.0 mut/MB vs. 3.3 mut/MB, p=0.922) or MSI status by NGS (p=0.069) were observed. Among the 34 pts with paired samples, 26.5% (9/34) relapsed, and 17.6% (6/34) died. When examining tumor mutation changes between pre- and post-CRT samples (table), the most common gained mutations were seen in APC (29.4%), MKI67 (26.5% ), KTM2C (17.6%), and TP53 (17.67%); and most common losses were in ARID1B (26.5%), ZFHX3 (26.5%), FLT4 (21.0%) and GATA6 (21.0%). Of the gained mutations, OncodriveCLUST analysis showed that MKI67 potentially carries a driver mutation (pG866V) at exon 12 (p = 0.045), which exclusively existed in the two relapsed pts.
Conclusions: Our data suggest that CRT did not alter MSI status nor the level of TMB. However, CRT did result in gained molecular alterations that could be responsible for treatment resistance and predict relapse. Gained MKI67 mutation may be a prognostic biomarker for relapse after CRT. Further prospective studies are needed to validate these findings.
VIEW THE PUBLICATION