Inhibitors of tropomyosin receptor kinase (TRK) have shown promising activity against neurotrophic TRK (NTRK) fusion-driven cancers. NTRK gene fusions are observed in less than 1% of colorectal cancers (CRCs). CRCs harboring wild-type BRAF, KRAS and MisMatch Repair deficiency (dMMR)/MicroSatellite Instability (MSI) due to MLH1 hypermethylation have been associated with NTRK fusions in small cohorts of non-metastatic tumors. We evaluated the frequency and associated clinical characteristics of NTRK fusions among dMMR/MSI metastatic CRCs (mCRCs) patients (pts).
Tumor samples of dMMR/MSI mCRC pts were obtained from a French multicenter retrospective cohort and from a single-center cohort of pts treated with immune checkpoint inhibitors (ICIs). Clinico-pathological data including KRAS and BRAF V600E status, MMR protein and MLH1 methylation status were available for all pts. All samples were screened for TRK expression by immunohistochemistry (IHC) using a pan-TRK antibody (clone EPR17341, Abcam) and for NTRK1/2/3 gene rearrangements, by FISH. Targeted RNAseq (FusionPlex® Lung SK0133 ARCHER kit) was performed on IHC or FISH-positive and doubtful cases.
A total of 172 pts [paired samples (primary + metastasis): n=38; primary only: n=134] were screened. Tumor samples of 10 pts (5.8%) harbored NTRK fusion genes by FISH (NTRK1=8; NTRK3=2), out of which only 5 were pan-TRK+. Archer assay detected TPM3-NTRK1 (n=2), LMNA-NTRK1 (n=2), ETV6-NTRK3 (n=1) and EML4-NTRK3 (n=1) fusion transcripts in 6 pts, including 4 cases pan-TRK+. Eight cases were sporadic and two had germline MMR mutations (Lynch syndrome) with loss of MSH2/MSH6 and isolated loss of MSH6 expression, respectively. One Lynch-related tumor was KRAS mutated, one sporadic tumor was BRAF V600E mutated. Four NTRK positive pts received ICIs, resulting in 3 objective responses per iRECIST (two complete and one partial responses with 25 to 54 months of follow up) and one primary resistance.
Frequency of NTRK1/2/3 fusions was 5.8% in our dMMR/MSI mCRCs cohort. These fusions were not restricted to sporadic cases. The diagnostic accuracy of pan-TRK IHC was low. Optimal testing algorithms for theranostic purpose remain to be defined in this setting.
Clinical trial identification
Legal entity responsible for the study
R. Cohen: Financial Interests, Personal, Advisory Board: Amgen; Sanofi. T. André: Financial Interests, Personal, Invited Speaker: BMS; HalioDx; Roche; Sanofi. F. Penault-Llorca: Financial Interests, Personal, Research Grant, consulting and travels: Bayer; Financial Interests, Personal, Advisory Board, Research funding and travels: Nanostring; Financial Interests, Personal, Research Grant, Advisory board: Illumina; Financial Interests, Personal, Advisory Role, Travels: Roche. All other authors have declared no conflicts of interest.