# Preoperative Magnetic Resonance Imaging Results Are Concordant with Pathology Staging in Rectal Cancer

## TL;DR

This study compares MRI results with actual pathology findings in rectal cancer patients after treatment, finding MRI moderately reliable for staging.

## Contribution

The study evaluates MRI reliability in rectal cancer staging after neoadjuvant therapy, identifying subgroup-specific performance.

## Key findings

- MRI showed moderate agreement with pathology for lymph node staging (N stage) but not for tumor depth (T stage) or CRM.
- Subgroup analysis revealed better MRI performance for lower rectal tumors and specific N stage predictive value.
- MRI demonstrated 88.6% specificity for N stage prediction with an AUC of 0.776.

## Abstract

Background and Objectives: Magnetic resonance imaging (MRI) is the gold standard for rectal cancer staging; however, its reliability after neoadjuvant therapy (NAT) remains controversial due to treatment-induced tissue changes. This study aimed to compare preoperative MRI findings with postoperative pathologic results in rectal cancer patients following NAT and to assess MRI reliability across clinical subgroups. Materials and Methods: This single-center retrospective study included 47 adult patients with locally advanced rectal adenocarcinoma who received NAT followed by elective rectal resection, with preoperative pelvic MRI and postoperative pathology results. Clinical features, MRI (T stage, N stage, and circumferential resection margin [CRM]), and pathologic staging were recorded. The endpoints were defined as concordance (via kappa coefficients) and predictive performance (via ROC analysis). Results: Among 47 patients (mean age 63.5 ± 9.4 years; 80.9% male), MRI demonstrated slight concordance with pathology for the T stage (kappa = 0.178, p = 0.028) and moderate concordance with the N stage (kappa = 0.489, p < 0.001), but not for CRM (p = 0.154). Subgroup analyses revealed significant concordance for N stage across most subgroups, with lower rectal tumors showing significant agreement for all three parameters. ROC analysis demonstrated significant predictive value for the N stage (AUC = 0.776, p = 0.002) with 88.6% specificity, while the T stage and CRM showed non-significant discriminatory performance. Conclusions: Post-NAT MRI demonstrates moderate reliability relative to pathology, particularly for N staging, and may have even better utility in specific subgroups stratified for sex, NAT type, and tumor site.