Urothelial phase CT for assessment of pathologic complete response after neoadjuvant chemotherapy in muscle-invasive bladder cancer

Purpose: To investigate the diagnostic performance of urothelial phase (UP) CT and identify the appropriate imaging criteria for assessment of pathologic complete response (pCR) after neoadjuvant chemotherapy (NAC) in patients with bladder cancer. Method: Seventy-three patients who underwent NAC and subsequent radical or partial cystectomy between January 2017 and July 2019 were retrospectively analyzed. UP CT findings after NAC were divided into five grades as follows: grade 1, no wall thickening or inner-layer enhancement; grade 2, thin inner-layer enhancement without wall thickening; grade 3, inner-layer enhancement with low-attenuation wall thickening; grade 4, wall thickening with enhancement; and grade 5, nodular enhancement or enhanced soft tissue. Two radiologists independently evaluated these grades on the post-chemotherapy CT. The area under the ROC curve (AUC) was used to evaluate diagnostic performance for pCR. Inter-reader agreement was assessed using the kappa coefficient. Results: Twenty-seven patients (37 %) were confirmed with pCR. The AUCs for the assessment of pCR on UP CT were 0.85-0.86 for the two readers. Using absent or thin inner-layer enhancement as features to predict pCR, the sensitivity, specificity, positive predictive value, and negative predictive value were 74.1-81.5 %, 80.4-84.8 %, 71.0-74.1 %, and 84.8-88.1 % for both readers. The inter-reader agreement for grades <= 2 was almost perfect (kappa = 0.83). Conclusions: Absent or thin inner-layer enhancement on UP CT demonstrated high diagnostic performance and high inter-reader agreement for assessment of pCR after neoadjuvant chemotherapy in bladder cancer, and evaluation of this feature could improve the predictive ability of preoperative imaging for assessing pCR.