RESUMO
The purpose of the study was to assess the performance of readers in diagnosing thoracic anomalies on standard chest radiographs (CXRs) with and without a deep-learning-based AI tool (Rayvolve) and to evaluate the standalone performance of Rayvolve in detecting thoracic pathologies on CXRs. This retrospective multicentric study was conducted in two phases. In phase 1, nine readers independently reviewed 900 CXRs from imaging group A and identified thoracic abnormalities with and without AI assistance. A consensus from three radiologists served as the ground truth. In phase 2, the standalone performance of Rayvolve was evaluated on 1500 CXRs from imaging group B. The average values of AUC across the readers significantly increased by 15.94%, with AI-assisted reading compared to unaided reading (0.88 ± 0.01 vs. 0.759 ± 0.07, p < 0.001). The time taken to read the CXRs decreased significantly, by 35.81% with AI assistance. The average values of sensitivity and specificity across the readers increased significantly by 11.44% and 2.95% with AI-assisted reading compared to unaided reading (0.857 ± 0.02 vs. 0.769 ± 0.02 and 0.974 ± 0.01 vs. 0.946 ± 0.01, p < 0.001). From the standalone perspective, the AI model achieved an average sensitivity, specificity, PPV, and NPV of 0.964, 0.844, 0.757, and 0.9798. The speed and performance of the readers improved significantly with AI assistance.
RESUMO
OBJECTIVES: To assess the diagnostic performance of a new device for in situ label-free fluorescence spectral analysis of breast masses in freshly removed surgical specimens, in preparation for its clinical development. METHODS: Sixty-four breast masses from consenting patients who had undergone either a lumpectomy or a mastectomy were included. Label-free fluorescence spectral acquisitions were obtained with a 25G fibre-containing needle inserted into the mass. Data from benign and malignant masses were compared to establish the most discriminating thresholds and measurement algorithms. Accuracy was verified using the bootstrap method. RESULTS: The final histological examination revealed 44 invasive carcinomas and 20 benign lesions. The maximum intensity of fluorescence signal was discriminant between benign and malignant masses (p < .0001) whatever their sizes. Statistical analysis indicated that choosing five random measurements per mass was the best compromise to obtain high sensitivity and high negative predictive value with the fewest measurements. Thus, malignant tumours were identified with a mean sensitivity, specificity, negative and positive predictive value of 98.8%, 85.4%, 97.2% and 93.5%, respectively. CONCLUSION: This new in situ tissue autofluorescence evaluation device allows accurate discrimination between benign and malignant breast masses and deserves clinical development. KEY POINTS: ⢠A new device allows in situ label-free fluorescence analysis of ex vivo breast masses ⢠Maximum fluorescence intensity discriminates benign from malignant masses (p < .0001) ⢠Five random measurements allow a high negative predictive value (97.2%).
