Strain Elastography Evaluation of Rectal Tumors: Inter- and Intraobserver Reproducibility

Abstract

Purpose: Elastography is a promising method for the identification and differentiation of malignant tissue in several organ systems. The primary aim was to evaluate the inter- and intraobserver reproducibility of endorectal strain elastography differentiation of adenomas and adenocarcinomas. The secondary aim was to compare the performance of strain elastography to endorectal ultrasonography (ERUS) examinations.

Materials and methods: Consecutive inclusion of 95 ERUS examinations and 110 elastography video loops with ERUS overlay mode. Video loops were randomized and evaluated by eight observers on two separate occasions. Observers were blinded to all clinical information except the circumferential location of the tumor. A continuous visual analog scale (VAS) and a categorical scale (W-score) were used for elastography evaluation. ERUS loops were T-staged according to the TNM classification system. Histopathological evaluation of surgical resection specimen was used as the reference standard.

Results: Strain elastography visual evaluation yielded intraobserver variability from 0.86 to 0.97 and interobserver variability of 0.99. VAS strain elastography differentiation of adenomas (pT0) and adenocarcinomas (pT1 - 4) yielded sensitivity, specificity, accuracy, positive and negative predictive values of 0.94, 0.71, 0.89, 0.92 and 0.78, respectively. The corresponding ERUS values were 0.83, 0.64, 0.79, 0.88 and 0.54, respectively.

Conclusion: Visual evaluation of elastography loops is highly reproducible in an offline setting with blinded observers, and correlates significantly with pT-stages. Strain elastography performs better than ERUS and might consequently improve staging.