Source: www.auntminnie.com Author: Erik L. Ridley, AuntMinnie.com staff writer Radiology and pathology artificial intelligence (AI) algorithms can help in diagnosing and assessing the prognosis of oral squamous cell carcinoma (SCC), according to a literature review published August 19 in JAMA Otolaryngology -- Head & Neck Surgery. After reviewing published studies in the literature on the use of AI with pathology and radiology images in patients with oral SCC, researchers from the University of Hong Kong concluded that the technology yielded good classification accuracy. "The successful use of deep learning in these areas has a high clinical translatability in the improvement of patient care," wrote the authors, led by first author Chui Shan Chu and senior author Dr. Peter Thomson, PhD. In radiology applications for oral SCC, a convolutional neural network (CNN) was able to predict disease-free survival with 80% accuracy, sensitivity, and specificity from PET images, the researchers reported. Another CNN showed lower performance -- 66.9% sensitivity, 89.7% specificity, and 84% accuracy -- when used on CT for predicting disease-free survival. A deep-learning algorithm also yielded 90% sensitivity for detecting lymph node metastasis from oral SCC on CT. In addition to providing prognosis predictions, AI could help facilitate personalized treatment from CT images, according to the researchers. One model was 76% accurate for predicting xerostomia, or dry mouth, an adverse effect of radiotherapy caused by toxicity. Another study determined that radiation dose distribution is the most crucial factor for predicting toxicity. The researchers noted that, to the best of their [...]