# Discordance between radiological and pathological response to neoadjuvant immunotherapy in mismatch repair-deficient/microsatellite instability-high colorectal cancer: a meta-analysis

## TL;DR

This study finds that radiological assessments often don't match up with actual tissue results in colorectal cancer patients treated with immunotherapy, especially in colon cancer.

## Contribution

The study quantifies the high discordance rate between radiological and pathological responses to neoadjuvant immunotherapy in dMMR/MSI-H CRC.

## Key findings

- A 59.6% discordance rate was found between radiological and pathological responses to neoadjuvant immunotherapy.
- Colon cancer patients showed a significantly higher discordance rate (64.2%) compared to rectal cancer patients (34.9%).
- Radiological assessments frequently misdiagnosed pathological complete responses as residual disease.

## Abstract

Mismatch repair deficiency (dMMR) and microsatellite instability (MSI-H) cancers exhibit high immunogenicity and are highly responsive to immune checkpoint inhibitors. In patients with locally advanced dMMR/MSI-H colorectal cancer (CRC), neoadjuvant immunotherapy (NIT) has demonstrated unprecedented pathological complete response (pCR) rates, suggesting nonoperative management strategies may be possible. There remains a discrepancy between radiological assessment and pathological responses to NIT in CRC.

We conducted a systematic review and meta-analysis of studies published between February 2015 and February 2025 to determine if radiological and pathological assessments following neoadjuvant immune checkpoint inhibitor therapy (NIT) were consistent in patients with non-metastatic dMMR/MSI-H CRC. Using PubMed, Embase, and Web of Science, the literature was retrieved, with inclusion criteria focusing on studies that reported both imaging data and pathological results. A random-effects model was used to calculate pooled odds ratios (ORs) and 95% confidence intervals (CIs). Subgroup analyses were conducted based on tumor location (colon versus rectum) and type of response (cCR versus pCR).

In 12 studies, 396 patients were included. A total rate of 59.6% discordance was found between radiological and pathological responses. Compared to rectal cancer patients (34.9%), colon cancer patients exhibited a significantly higher rate of discordance (64.2%). A total of 238 patients with confirmed pCR were incorrectly diagnosed as having residual disease on radiological assessment (OR = 61.41; 95% CI: 10.05–375.27;P < 0.00001). A high level of heterogeneity was observed across studies (I2 = 85%), but no publication bias was observed.

In dMMR/MSI-H CRC, radiologic assessment alone cannot reliably assess the efficacy of NIT, particularly in colon cancer. It should be integrated with additional modalities—such as endoscopic evaluation and biomarker analysis—to ensure an accurate appraisal of treatment efficacy.