Colonoscopic high frequency mini‐probe ultrasound is more accurate than conventional computed tomography in the local staging of colonic cancer

Aim  Colonoscopic high frequency mini‐probe ultrasound was compared prospectively with CT in the local staging of colonic cancer. Method  Consecutive patients undergoing surgical resection for colonic cancer were recruited. Preoperative 64‐slice CT staging with multiplanar reconstruction was compared with colonoscopic high frequency mini‐probe ultrasound using 12 MHz and 20 MHz probes. The three methods of staging (CT, 12 MHz ultrasound and 20 MHz ultrasound) were compared with the histological stage of the resected specimen. This was done using weighted kappa coefficients where weights of 0.7–0.8 were given to penalize disagreements of one level in either direction and weights of zero were given to penalize disagreements of more than one level in any direction. Results  In total, 38 patients with colonic cancer were included. They were located in the sigmoid ( n  =   20), descending ( n  =   5), ascending ( n  =   2) and transverse colon ( n  =   1) and in the caecum ( n  =   7) and splenic ( n  =   2) and hepatic ( n  =   1) flexure. Histopathological assessment revealed seven pT1, four pT2, 25 pT3 and two pT4 cancers. In relation to the pathology the weighted kappa coefficients were 0.36 (SE = 0.14), 0.81 (SE = 0.16) and 0.81 (SE = 0.17) for CT, ultrasound 12 MHz and ultrasound 20 MHz. Histopathologically 15 (39.5%) patients were lymph node positive. The sensitivity, specificity and kappa coefficient for detection of nodal disease for CT were 80%, 47.8% and 0.25 (SE = 0.14) compared with 80%, 82.5% and 0.62 for 12 MHz ultrasound (SD = 0.14) and 23%, 90.5% and 0.15 (SD = 0.13) for 20 MHz ultrasound. Conclusion  Colonoscopic ultrasound is significantly more accurate than CT for T staging of colonic cancers. With respect to nodal status, 12 MHz ultrasound offers superior accuracy to CT or 20 MHz ultrasound.