Utility of computed tomography (CT) –based staging of mismatch repair deficient (dMMR) colon cancer.

45 Background: dMMR tumors are characterized by rapid primary tumor growth with reduced likelihood for nodal and distant metastases compared to MMR proficient (pMMR) tumors. Neoadjuvant immune checkpoint inhibitors (ICIs) have demonstrated promising efficacy for localized colon cancer with high pathologic complete response rates and low relapse rates. This raises the question if surgical treatment could be avoided with accurate radiographic staging while avoiding overtreatment of early stage tumors. This study assesses the utility of radiographic staging by CT for dMMR tumors compared to pathologic staging. Methods: Patients with stage I-III colon cancer treated with upfront surgical resection were retrospectively reviewed from a single institution from 2012 to 2023 in two cohorts: 1) dMMR colon cancer and 2) pMMR colon cancer limited to similar sample size as matched controls. Clinical stage was determined based on CT scans prior to surgical resection by 2 independent radiologists blinded to pathologic stage, and the results were correlated to pathologic stage. Clinical characteristics were extracted from the electronic medical record. Statistical analysis was performed using SPSS. Results: We identified 80 patients with dMMR colon cancer and 66 with pMMR tumors. CT clinical tumor staging was discordant from pathologic staging for 68% of dMMR tumors and 61% for pMMR tumors. For T4 dMMR tumors, CT had a 74% sensitivity and 44% specificity. T4 tumors were understaged by radiology to less than T4 for 56% of cases. For dMMR nodal staging, CT had 89% sensitivity and 50% specificity (Table 1). For pathologic node positive tumors, 14% were understaged by radiologist CT read to node negative. For pathologic node negative tumors, 50% of tumors were overstaged to node positive by radiologist CT read. There was poor inter-rater reliability for both T stage and N stage for dMMR and pMMR tumors. The weakest agreement between radiologists was for T stage (dMMR kappa = 0.246, mMMR kappa = 0.145). Conclusions: Radiologic clinical staging of dMMR and pMMR colon cancer does not correlate well with pathologic staging. There were high rates of understaging by CT evaluation for patients who may be candidates for neoadjuvant treatment (T4, N+). Furthermore, there was poor inter-rater reliability between radiologists, indicating that CT staging alone may not be sufficient. Additional diagnostic modalities for nodal detection may be necessary to accurately clinically stage patients before neoadjuvant ICIs in patients with locally advanced disease. Comparing utility of pMMR versus dMMR tumors for T4 and node positive disease. pMMR (n = 66) dMMR (n = 80) T4 versus T3 or less Sensitivity (%) 92 74 Specificity (%) 51 47 PPV (%) 80 77 NPV (%) 82 63 Node positive versus node negative Sensitivity (%) 60 50 Specificity (%) 75 89 PPV (%) 81 89 NPV (%) 53 50 Abbreviations: PPV = positive predictive value; NPV = negative predictive value.