Consensus molecular subtypes (CMS) as a marker for treatment and disease biology in metastatic colorectal cancer (CRC).

4089 Background: Consensus molecular subtypes (CMS) categorize colorectal cancer (CRC) into groups based on gene transcription and are prognostic in early-stage and first-line metastatic settings. Their impact on treatment history is unknown. We hypothesized that the best prognosis CMS2 would have higher utilization of liver-directed therapies and maintenance chemotherapy over the worst prognosis CMS4. Methods: Primary surgical resection specimens were annotated for CMS on a translational protocol in a 5FU-refractory metastatic CRC population. Specimens that had neoadjuvant chemotherapy or radiation were excluded. CMS1, CMS3 and indeterminate CMS were also excluded. Liver-directed therapies were defined as any surgery, direct injection of cytotoxics or microspheres, radiation or radiofrequency ablation. Multiple occurrences of liver-directed therapies or maintenance chemotherapy in the same patient were recorded once for statistical tests of association. Results: CMS1 (7.4%), CMS3 (8.2%) and indeterminate calls (4.1%) accounted for 20% of all samples tested. There were 43 (44%) CMS2 and 55 (56%) CMS4 patients eligible. Age, stage at diagnosis, mismatch repair and RAS mutational status were similar in both groups. Left-sided tumors were more common in CMS2 (79%) than CMS4 (42%), p = .001. The median overall survival (OS) from stage IV diagnosis was 40 [34 - 51] versus (vs) 28 [21 – 33] months (p < .0001) for CMS2 vs CMS4 respectively. Liver-directed therapy was greater in CMS2 (53%) vs CMS4 (31%), p = .024. The number further increased when multiple treatments in a single patient were taken into account. Microsphere injection, radiation and liver surgery combined with RFA were the most skewed towards CMS2 over CMS4. No difference in median OS was seen from first liver-directed therapy in CMS2 vs CMS4 (29 vs 27 months, p = .31). There was a trend towards greater maintenance chemotherapy in CMS2 (47%) vs CMS4 (29%), p = .076. Conclusions: Better prognosis with CMS2 is consistent with other studies. Significantly increased liver-directed therapy was observed in CMS2. Selection criteria for liver-directed therapy such as slowly progressing, oligometastatic and hepatic-confined disease may be enriching for CMS2 and gives insight into the natural history of this subtype.