247 Background: MET signaling is implicated in the tumorigenesis of many solid tumors, including CRC. MET-targeted therapies are approved in non-small cell lung cancer with novel agents in clinical trials for tumors with MET exon 14 splice site mutations, MET amplifications, and recently MET expression. There is emerging therapeutic interest in targeting MET in CRC. Methods: Tumor samples from 50,500 cases of clinically advanced CRC were analyzed by hybrid capture-based comprehensive genomic profiling (CGP) that evaluated all classes of genomic alterations (GA). MSI-high status, tumor mutational burden (TMB), genomic ancestry, trinucleotide mutational signatures, and homologous recombination deficiency signature (HRDsig) were assessed for patients with activating MET GA. PD-L1 expression was determined by IHC (Dako 22C3 with TPS scoring system). Results were compared using the Fisher exact test with the Benjamini-Hochberg adjustment. Results: A total of 502 (1.0%) cases of activating GA of MET (METmut) were identified, including 418 cases (0.83%) with MET amplification (METamp), 10 (<0.1%) rearrangements, and 78 (0.1%) short variant (SV) mutations with 19 (<0.1%) cases of MET exon 14 splice site mutations. When compared with MET wild-type (METwt) CRC, the METmut CRC were of similar age (median 62 years), and numerically more likely to be male (61.4% vs 56.0%; P=.05). There were no differences in genomic ancestry (72% European ancestry for both groups). The METmut CRC featured more cases with a POLE signature (3.6% vs 0.4%; P=<0.0001) and more GA per tumor (7 vs 5; P<0.0001). MSI-high status was similar in both groups, with 4.2% in METmut and 6.0% in METwt (P=0.28). TMB levels were numerically higher in the METmut cases (TMB ≥ 10 mutations/Mb at 12.2% vs 9.0%; P=0.058). Low level PD-L1 expression (1-49% TPS) was similar in both groups (13.9% in METmut vs 13.3% in METwt), and while not statistically significant, high level PD-L1 expression (>50% TPS) was higher in the METmut group (4.3% vs 1.6%; P=0.056). The frequencies of HRDsig+ CRC were rare and similar in both groups with 1.5% in METmut and 1.7% in METwt. METmut CRC had lower frequencies of GA in KRAS (34.7% vs 48.8%; P<0.0001), NRAS (1.8% vs 4.2%; P=0.011), APC (72.5% vs 78.2%; P=0.009) and PIK3CA (12.2% vs 18.9%; P=0.0004), and higher frequencies of GA in CDK6 (10.2% vs 0.6%; P<0.0001), EGFR (5.2% vs 0.2%; P=0.0002), ERBB2 (9.6% vs 5.2%; P=0.0004), HGF (8.0% vs 0.8%; P<0.0001), POLE (4.2% vs 0.5%; P=0.005) and TP53 (84.1% vs 75.9%; P<0.0001). In METmut CRC, ERBB2 alterations were 4.1% amplifications, 2.9% sequence mutations, 0.9% with more than one mutation. BRAF V600E was seen in 5.9% in METmut CRC and 8.4% of METwt. Conclusions: CGP reveals significant differences in the genomic landscapes of METmut CRC and METwt CRC, which may guide selection of potential rational drug combinations with novel MET-targeted therapies.
Read full abstract