Background Accurate preoperative staging is essential in rectal cancer, particularly in patients managed without neoadjuvant therapy, where initial magnetic resonance imaging findings directly inform surgical decision-making. Evidence describing magnetic resonance imaging–pathology concordance in this specific clinical context remains limited. Methods This single-centre retrospective service evaluation included consecutive patients with rectal adenocarcinoma who underwent primary surgery without neoadjuvant therapy between June 2022 and August 2025. Preoperative pelvic MRI staging (T stage, nodal status, and extramural venous invasion) was compared with postoperative histopathology. Concordance, rates of over and understaging, and diagnostic performance metrics were calculated. Given the small sample size, results are descriptive. Results Twenty-nine patients were included. Magnetic resonance imaging correctly staged T category in 62% and nodal status in 55% of cases. For distinguishing ≥T3 disease (n=27), sensitivity was 64.3% (95% CI 38.8–83.7), specificity 76.9% (95% CI 49.7–91.8), and overall accuracy 70.4% (95% CI 51.5–84.1). Nodal staging showed poor sensitivity at 18.2% (95% CI 5.1–47.7) with accuracy of 55.2% (95% CI 37.5–71.6). EMVI concordance was limited, with both false-positive and false-negative Magnetic resonance imaging assessments observed. Conclusions In this small single-centre retrospective cohort, magnetic resonance imaging demonstrated moderate performance for T staging but limited reliability for nodal and EMVI assessment in rectal cancer patients undergoing primary surgery without neoadjuvant therapy. These findings support cautious interpretation of magnetic resonance imaging staging in early disease and highlight the need for improved standardisation and larger prospective studies.
Filipe L.F. Carvalho (Sat,) studied this question.