Background and study aims: This multicenter, prospective, country-wide quality-assurance study at more than 300 hospitals in Germany was designed to characterize and analyze the diagnostic accuracy of rectal endoscopic ultrasound (EUS) in the routine clinical staging of rectal carcinoma (depth of tumor infiltration).
Patients and methods: Patients were surveyed between 1 January 2000 and 31 December 2008. Those who received neoadjuvant therapy after EUS were excluded. The correspondence between the EUS assessment of tumor depth (uT) and that determined by histology (pT) was calculated, and the influence of hospital volume upon the sensitivity, specificity, and positive and negative predictive values was investigated.
Results: At 384 hospitals providing care at all levels, 29 206 patients were included; of the 27 458 treated by surgical resection, EUS was performed for 12 235 (44.6 %). Of these, 7096 did not receive neoadjuvant radiochemotherapy, allowing a uT-pT comparison. The uT-pT correspondence was 64.7 % (95 % confidence interval [CI] 63.6 % - 65.8 %); the frequency of understaging was 18 % (95 %CI 17.1 % - 18.9 %) and that of overstaging was 17.3 % (95 %CI 16.4 % - 18.2 %). The kappa coefficient was greatest in the category T1 (κ = 0.591). For T3 tumors κ was 0.468. The poorest correspondence was found for T2 and T4 tumors (κ = 0.367 and 0.321, respectively). A breakdown by hospital volume showed that the uT-pT correspondence was 63.2 % (95 %CI 61.5 % - 64.9 %) for hospitals undertaking ≤ 10 EUS/year, 64.6 % (95 %CI 62.9 % - 66.2 %) for doing 11 - 30 EUS/year, and 73.1 % (95 %CI 69.4 % - 76.5 %) for those hospitals performing > 30 EUS/year.
Conclusions: In clinical routine, the diagnostic accuracy of transrectal ultrasound in staging rectal carcinoma does not attain the very good results reported in the literature. Only in the hands of diagnosticians with a large case volume of rectal carcinoma patients can EUS lead to therapy-relevant decisions.
© Georg Thieme Verlag KG Stuttgart · New York.