[Clinical evaluation of endoscopic ultrasonography and CT in the prediction of the resectability of esophageal carcinoma].

OBJECTIVE

To evaluate the role of endoscopic ultrasonography (EUS) and CT in the prediction of the resectability of esophageal carcinoma.

METHODS

A retrospective study was carried out in 746 patients with esophageal carcinoma. These patients were divided into CT group (480 cases), EUS group (151 cases) and EUS+CT group (115 cases). Images of EUS and CT were double-blindly reviewed by radiologists. Relationship of EUS and CT images with surgical and pathological findings was examined.

RESULTS

Resection rates in the EUS group, CT group and EUS+CT group were 93.4%, 91.0% and 93.9%, respectively (chi(2)=1.551, P=0.484). Accuracy, sensitivity, specificity, positive predictive value and negative predictive value in the CT group were 81.7%, 87.4%, 23.3%, 92.0% and 15.4%, respectively; 94.7%, 98.6%, 40.0%, 95.9% and 66.7% in the EUS group; and 96.5%, 99.1%, 57.1%, 97.3% and 80.0% in the EUS+CT group, respectively. When assessing aortic invasion, accuracy, sensitivity, specificity, positive predictive value and negative predictive value were 91.3%, 33.3%, 93.1%, 13.5% and 97.7%, in the CT group, respectively; 98.7%, 87.5%, 99.3%, 87.5% and 99.3% in the EUS group, respectively,and 98.3%, 85.7%, 99.1%, 85.7% and 99.1% in the EUS+CT group, respectively. In assessing tracheobronchial invasion, accuracy, sensitivity, specificity, positive predictive value and negative predictive value were 91.3%, 20.8%, 95.0%, 17.9% and 95.8% in the CT group, respectively; 96.0%, 20.0%, 98.6%, 33.3% and 97.3% in the EUS group, respectively; and 98.3%, 66.7%, 99.1%, 66.7% and 99.1% in the EUS+CT group. Differences in assessing resectability were significant between CT group and EUS group (chi(2)=15.131, P=0.000), between CT group and EUS+CT group (chi(2)=15.662, P=0.000), and between EUS group and EUS+CT group (chi(2)=0.502, P=0.346). Differences in assessing aortic invasion were significant between CT group and EUS group (chi(2)=9.764, P=0.000), and between CT group and EUS+CT group (chi(2)=6.659, P=0.004), but were not significant between EUS group and EUS+CT group (chi(2)=0.076, P=0.581). Differences in assessing tracheobronchial invasion were significant between CT group and EUS+CT group (chi(2)=6.659, P=0.004), but were not significant between CT group and EUS group (chi(2)=3.729, P=0.034) and between EUS group and EUS+CT group (chi(2)=1.117, P=0.248).

CONCLUSIONS

EUS is a better procedure than CT in the prediction of the resectability and aortic invasion in esophageal carcinoma. There is limited value for EUS and CT in assessing tracheobronchial invasion. Combination of CT and EUS does not improve the prediction of resectability significantly.