Flexible endorectal ultrasound for predicting pathologic stage of rectal cancers.

Background: Endorectal ultrasound (ERUS) is an accurate method for preoperative staging of rectal cancers. Most often, a rigid 360-degree rotating probe is used. We studied whether flexible probes could attain equivalent accuracy for bowel wall penetration.

Methods: Forty-five patients were prospectively evaluated with flexible devices. Results were compared with 20 rigid and 10 flexible probe studies. To assess learning curves, we used logistic regression analysis and coefficients of correlation on accuracy data to compare ERUS accuracy with the number of examinations.

Results: Level of invasion was correct in 49%. Nodal examinations were correct in 78%. Learning curves leveled out at 100 examinations with 87% accuracy for the rigid probe (R = 0.46) and 77% for the flexible devices (R = 0.31).

Conclusions: The coefficient of correlation for each method portends a more reliable learning curve for the rigid devices. Flexible devices were less accurate for level of invasion than the literature reported for rigid devices.