How Reliable Is CT Scan in Staging Right Colon Cancer?

BACKGROUND: The observation of inferior oncologic outcomes after surgery for proximal colon cancers has led to the investigation of alternative treatment strategies, including surgical procedures and neoadjuvant systemic chemotherapy in selected patients. OBJECTIVE: The purpose of this study was to determine the accuracy of CT staging in proximal colon cancer in detecting unfavorable pathologic features that may aid in the selection of ideal candidates alternative treatment strategies, including extended lymph node dissection and/or neoadjuvant chemotherapy. DESIGN: This was a retrospective consecutive series. SETTINGS: Trained abdominal radiologists from 2 centers performed a blinded review of CT scans obtained to locally stage proximal colon cancer according to previously defined prognostic groups, including T1/2, T3/4, N+, and extramural venous invasion. CT findings were compared with histopathologic results as a reference standard. Unfavorable pathologic findings included pT3/4, pN+, or extramural venous invasion. PATIENTS: Consecutive patients undergoing right colectomy in 2 institutions between 2011 and 2016 were retrospectively reviewed from a prospectively collected database. MAIN OUTCOME MEASURES: T status, nodal status, and extramural venous invasion status comparing CT with final histologic findings were measured. RESULTS: Of 150 CT scans reviewed, CT failed to identify primary cancer in 18%. Overall accuracy of CT to identify unfavorable pathologic features was 63% with sensitivity, specificity, positive predictive value, and negative predictive value of 63% (95% CI, 54%-71%), 63% (95% CI, 46%-81%), 87% (95% CI, 80%-94%) and 30% (95% CI, 18%-41%). Only cT3/4 (55% vs 45%; p = 0.001) and cN+ (42% vs 58%; p = 0.02) were significantly associated with correct identification of unfavorable features at final pathology. CT scans overstaged and understaged cT in 23.7% and 48.3% and cN in 28.7% and 53.0% of cases. LIMITATIONS: The study was limited by its retrospective design, relatively small sample size, and heterogeneity of CT images performed in different institutions with variable equipment and technical details. CONCLUSIONS: Accuracy of CT scan for identification of pT3/4, pN+, or extramural venous invasion was insufficient to allow for proper identification of patients at high risk for local recurrence and/or in whom to consider alternative treatment strategies. Locoregional overstaging and understaging resulted in inappropriate treatment strategies in <48%.