The application of artificial intelligence (AI) in medicine has increased rapidly with respect to tasks including disease detection/diagnosis, risk stratification, and prognosis prediction. With recent advances in computing power and algorithms, AI has shown promise in taking advantage of vast electronic health data and imaging studies to supplement clinicians. Machine learning and deep learning are the most widely used AI methodologies for medical research and have been applied in pancreatic diseases for which diagnosis and treatment selection are often complicated and require joint consideration of data from multiple sources.
Our group recently trained and tested a convolutional neural network (CNN) to distinguish pancreatic ductal adenocarcinoma (PDAC) and normal pancreas in CT images of 370 PDAC patients and 320 normal controls from Taiwan. In the local (Taiwanese) test sets, the CNN-based analysis achieved 98.6-98.9% accuracy (AUC 0.997– 0.999), with a higher sensitivity compared with radiologist interpretation (98.3% vs 92.9%, difference 5.4% [95% CI 1.1%–9.8%]; p=0.014). Notably, CNN-based analysis achieved 92.1% sensitivity for PDACs smaller than 2 cm and correctly classified 92% of PDACs missed by radiologists. We have also investigated the potential usefulness of radiometer analysis with machine learning in detecting PDAC on CT and identified a panel of distinguishing radiomic features of PDAC. Radiomic analysis with a machine learning model trained with predominantly Taiwanese images could differentiate between patients with PDAC and controls in Taiwanese (accuracy 95.0%) and U.S. (accuracy 86.5%) test images, with 96.9% and 90.9% sensitivity, respectively, for PDACs smaller than 2 cm. Our studies support that AI may supplement radiologist interpretation to facilitate early detection of PDAC. |
