Characterization of tumor mutation load (TML) in solid tumors

Authors:

Mohamed E. Salem, Joanne Xiu, HeinzJosef Lenz, Michael B. Atkins, Philip A. Philip, Jimmy J. Hwang, Zoran Gatalica, 2Nianqing Xiao, Geoffrey T Gibney, Wafik S. ElDeiry, Antoinette R. Tan, Edward S. Kim, Anthony Frank Shields, Derek Raghavan, and John Marshall

Background:

Rapid advances in immunotherapy have created a need for biomarkers to improve patient treatment selection. TML is proposed as a potential predictive biomarker due to its association with tumor immunogenicity.

Methods:

TML was assessed in 7,748 solid tumors from 14 different origins using somatic nonsynonymous missense mutations sequenced with a 592-gene panel. High TML was set at ≥ 17 mutations per megabase (mt/MB) based on an established concordance ( > 99%) with MSI-High status in colorectal cancer (CRC).

Results:

Mean TML was highest in melanoma (Mel, 21 mt/MB), non-small cell lung cancer (NSCLC, 11 mt/MB), and bladder cancer (BLC; 11 mt/MB), whereas prostate cancer (PC), pancreas adenocarcinoma (PA), and renal cell carcinoma (RCC) had the lowest levels (all 6 mt/MB). High TML was seen most frequently in Mel (36%), NSCLC (15%), BLC (15%), and anal cancer (SCCA; 9%), whereas it was seen least frequently in PA (1.6%) and RCC (0.5%). Primary NSCLC carried lower TML than its brain metastases (11 vs. 16 mt/MB, p < 0.001). Older age was associated with higher TML in Mel (p = 0.001), CRC (p = 0.009), breast cancer (BC; p = 0.01), and NSCLC (p = 0.02). Higher TML was seen in males than in females for Mel (p = 0.002) and NSCLC (p < 0.001). Presence of mutations in oncogenic driver genes such as EGFR, ALK, ROS1 and RET fusions, as well as cMET exon 14 skipping mutations correlated with lower TML in NSCLC (6.9 vs. 12 mt/MB, p < 0.001), as did BRAF and NRAS mutations in Mel (17 vs. 26, p = 0.003). Conversely, mutations in tumor suppressor genes such as ARID1A (CRC, NSCLC, and BLC) and NF1 (BC, CRC, Mel, BLC, and NSCLC) were associated with higher TML (p < 0.05). MSI-high was correlated with high TML in CRC and gastric cancers (p < 0.05).

Conclusions:

TML varied significantly among different cancers. High TML was associated with older age, presence of tumor suppressor gene mutations, and absence of other oncogenic mutations. Future studies will assess the impact of TML on clinical outcome and establish the role of TML in selecting patients for immunotherapy

Download Publication