Nomogram to predict the outcomes of patients with microsatellite instability-high metastatic colorectal cancer receiving immune checkpoint inhibitors

Background The efficacy of immune checkpoint inhibitors (ICIs) in patients with microsatellite instability (MSI)-high metastatic colorectal cancer (mCRC) is unprecedented. A relevant proportion of subjects achieving durable disease control may be considered potentially ‘cured’, as opposed to patients experiencing primary ICI refractoriness or short-term clinical benefit. We developed and externally validated a nomogram to estimate the progression-free survival (PFS) and the time-independent event-free probability (EFP) in patients with MSI-high mCRC receiving ICIs. Methods The PFS and EFP were estimated using a cure model fitted on a developing set of 163 patients and validated on a set of 146 patients with MSI-high mCRC receiving anti-programmed death (ligand)1 (PD-(L)1) ± anticytotoxic T-lymphocyte antigen 4 (CTLA-4) agents. A total of 23 putative prognostic factors were chosen and then selected using a random survival forest (RSF). The model performance in estimating PFS probability was evaluated by assessing calibration (internally—developing set and externally—validating set) and quantifying the discriminative ability (Harrell C index). Results RFS selected five variables: ICI type (anti-PD-(L)1 monotherapy vs anti-CTLA-4 combo), ECOG PS (0 vs >0), neutrophil-to-lymphocyte ratio (≤3 vs >3), platelet count, and prior treatment lines. As both in the developing and validation series most PFS events occurred within 12 months, this was chosen as cut-point for PFS prediction. The combination of the selected variables allowed estimation of the 12-month PFS (focused on patients with low chance of being cured) and the EFP (focused on patients likely to be event-free at a certain point of their follow-up). ICI type was significantly associated with disease control, as patients receiving the anti-CTLA-4-combination experienced the best outcomes. The calibration of PFS predictions was good both in the developing and validating sets. The median value of the EFP (46%) allowed segregation of two prognostic groups in both the developing (PFS HR=3.73, 95% CI 2.25 to 6.18; p<0.0001) and validating (PFS HR=1.86, 95% CI 1.07 to 3.23; p=0.0269) sets. Conclusions A nomogram based on five easily assessable variables including ICI treatment was built to estimate the outcomes of patients with MSI-high mCRC, with the potential to assist clinicians in their clinical practice. The web-based system ‘MSI mCRC Cure’ was released.