ABSTRACT
PURPOSE: To determine whether focal peripheral zone enhancement on routine venous-phase CT is predictive of higher-grade (Gleason 4 + 3 and higher) prostate cancer. MATERIALS AND METHODS: IRB approval was obtained and informed consent waived for this HIPAA-compliant retrospective study. Forty-three patients with higher-grade prostate cancer (≥Gleason 4 + 3) and 96 with histology-confirmed lower-grade (≤Gleason 3 + 4 [n = 47]) or absent (n = 49) prostate cancer imaged with venous-phase CT comprised the study population. CT images were reviewed by ten blinded radiologists (5 attendings, 5 residents) who scored peripheral zone enhancement on a scale of 1 (benign) to 5 (malignant). Mass-like peripheral zone enhancement was considered malignant. Likelihood ratios (LR) and specificities were calculated. Multivariate conditional logistic regression analyses were conducted. RESULTS: Scores of "5" were strongly predictive of higher-grade prostate cancer (pooled LR+ 9.6 [95% CI 5.8-15.8]) with rare false positives (pooled specificity: 0.98 [942/960, 95% CI 0.98-0.99]; all 10 readers had specificity ≥95%). Attending scores of "5" were more predictive than resident scores of "5" (LR+: 14.7 [95% CI 5.8-37.2] vs. 7.6 [95% CI 4.2-13.7]) with similar specificity (0.99 [475/480, 95% CI 0.98-1.00] vs. 0.97 [467/480, 95% CI 0.96-0.99]). Significant predictors of an assigned score of "5" included presence of a peripheral zone mass (p < 0.0001), larger size (p < 0.0001), and less reader experience (p = 0.0008). Significant predictors of higher-grade prostate cancer included presence of a peripheral zone mass (p = 0.0002) and larger size (p < 0.0001). CONCLUSION: Focal mass-like peripheral zone enhancement on routine venous-phase CT is specific and predictive of higher-grade (Gleason 4 + 3 and higher) prostate cancer.
Subject(s)
Prostatic Neoplasms/diagnostic imaging , Prostatic Neoplasms/pathology , Radiographic Image Enhancement , Tomography, X-Ray Computed , Aged , Aged, 80 and over , Clinical Competence , Humans , Image Interpretation, Computer-Assisted , Male , Middle Aged , Multivariate AnalysisABSTRACT
An abdominal computed tomographic scan was modified by inserting 10 simulated colonic polyps with use of methods that closely mimic the attenuation, noise, and polyp-colon wall interface of naturally occurring polyps. A shape-based polyp detector successfully located six of the 10 polyps. When settings that enhanced the edge profile of polyps were chosen, eight of 10 polyps were detected. There were no false-positive detections. Shape analysis is technically feasible and is a promising approach to automated polyp detection.
