Machine Learning for Adrenal Gland Segmentation and Classification of Normal and Adrenal Masses at CT.

Background Adrenal masses are common, but radiology reporting and recommendations for management can be variable. Purpose To create a machine learning algorithm to segment adrenal glands on contrast-enhanced CT images and classify glands as normal or mass-containing and to assess algorithm performance. Materials and Methods This retrospective study included two groups of contrast-enhanced abdominal CT examinations (development data set and secondary test set). Adrenal glands in the development data set were manually segmented by radiologists. Images in both the development data set and the secondary test set were manually classified as normal or mass-containing. Deep learning segmentation and classification models were trained on the development data set and evaluated on both data sets. Segmentation performance was evaluated with use of the Dice similarity coefficient (DSC), and classification performance with use of sensitivity and specificity. Results The development data set contained 274 CT examinations (251 patients; median age, 61 years; 133 women), and the secondary test set contained 991 CT examinations (991 patients; median age, 62 years; 578 women). The median model DSC on the development test set was 0.80 (IQR, 0.78-0.89) for normal glands and 0.84 (IQR, 0.79-0.90) for adrenal masses. On the development reader set, the median interreader DSC was 0.89 (IQR, 0.78-0.93) for normal glands and 0.89 (IQR, 0.85-0.97) for adrenal masses. Interreader DSC for radiologist manual segmentation did not differ from automated machine segmentation (P = .35). On the development test set, the model had a classification sensitivity of 83% (95% CI: 55, 95) and specificity of 89% (95% CI: 75, 96). On the secondary test set, the model had a classification sensitivity of 69% (95% CI: 58, 79) and specificity of 91% (95% CI: 90, 92). Conclusion A two-stage machine learning pipeline was able to segment the adrenal glands and differentiate normal adrenal glands from those containing masses. © RSNA, 2022 Online supplemental material is available for this article.

A simplified classification of proximal femoral fractures improves accuracy, confidence, and inter-reader agreement of hip fracture classification by radiology residents.

PURPOSE: To demonstrate the effect of teaching a simplified treatment-based classification of proximal femoral fractures on the accuracy, confidence, and inter-reader agreement of radiology residents. The authors hypothesize that these measures will improve after viewing an educational presentation. MATERIALS AND METHODS: Three radiology residents independently classified 100 operative proximal femoral fractures, both before and after viewing a 45-min educational video describing the simplified classification scheme, with a washout period of at least 12 weeks between sessions. Based on the gold standard established by consensus of two radiologists and an orthopedic trauma surgeon utilizing intraoperative fluoroscopic imaging, operative reports, and pre-procedural imaging, accuracy of classification was calculated for each reader before and after viewing the educational video. Reader confidence was recorded on a 0-10 scale, and inter-reader agreement was calculated with Fleiss's kappa. McNemar's test was used to compare accuracy, a paired t test was used to compare confidence, and the Z-test was used to compare kappa values after bootstrapping to determine the standard error of the mean. RESULTS: The study cohort included 60/100 females, with a mean age of 76.6 years. The pooled classification accuracy was initially 65%, which improved to 80% in the second reading session after viewing the educational video (p < 0.0001). Confidence improved from 6.9 initially to 8.6 (p < 0.0001). Inter-reader agreement improved from a kappa of 0.45 (moderate agreement) to 0.74 (substantial agreement) (p < 0.0001). CONCLUSIONS: A simplified treatment-based classification of proximal femoral fractures is easily taught to radiology residents and resulted in increased accuracy, increased inter-reader agreement, and increased reader confidence.