Prognosis and immunotherapy efficacy in dMMR&MSS colorectal cancer patients and an MSI status predicting model

Abstract The inconsistency between mismatch repair (MMR) protein immunohistochemistry (IHC) and microsatellite instability PCR (MSI‐PCR) methods has been widely reported. We aim to investigate the prognosis and the effect of immunotherapy in dMMR by IHC but MSS by MSI‐PCR (dMMR&MSS) colorectal cancer (CRC) patients. A microsatellite instability (MSI) predicting model was established to help find dMMR&MSS patients. MMR and MSI states were detected by the IHC and MSI‐PCR in 1622 CRC patients (ZS6Y‐1 cohort). Logistic regression analysis was used to screen clinical features to construct an MSI‐predicting nomogram. We propose a new nomogram‐based assay to find patients with dMMR&MSS, in which the MSI‐PCR assay only detects dMMR patients with MSS predictive results. We applied the new strategy to a random cohort of 248 CRC patients (ZS6Y‐2 cohort). The consistency of MMR IHC and MSI‐PCR in the ZS6Y‐1 cohort was 95.7% (1553/1622). Both pMMR&MSS and dMMR&MSS groups experienced significantly shorter overall survival (OS) than those in dMMR by IHC and MSI‐H by MSI‐PCR (dMMR&MSI‐H) group (hazard ratio [HR] = 2.429, 95% confidence interval [CI]: 1.89–3.116, p < .01; HR = 21.96, 95% CI: 7.24–66.61, p < .01). The dMMR&MSS group experienced shorter OS than the pMMR&MSS group, but the difference did not reach significance (log rank test, p = .0686). In the immunotherapy group, the progression‐free survival of dMMR&MSS patients was significantly shorter than that of dMMR&MSI‐H patients (HR = 13.83, 95% CI: 1.508–126.8, p < .05). The ZS6Y‐MSI‐Pre nomogram (C‐index = 0.816, 95% CI: 0.792–0.841, already online) found 66% (2/3) dMMR&MSS patients in the ZS6Y‐2 cohort. There are significant differences in OS and immunotherapy effect between dMMR&MSI‐H and dMMR&MSS patients. Our prediction model provides an economical way to screen dMMR&MSS patients.