Beyond traditional methods-artificial intelligence in detection of oral cancer using smartphone-based oral photographs: a systematic review

Oral cancer is a significant global health concern that affects people of various age groups worldwide. According to Globo Can, reports from 2022 show that approximately 377,713 new cases and 177,757 deaths are reported each year worldwide. Smartphones and Artificial intelligence (AI) are increasing in healthcare for diagnosis and treatment planning. This systematic review aims to appraise the existing evidence on the effectiveness of various artificial intelligence algorithms in the detection of oral cancer based on smartphone-based oral photographs from previously published articles. A systematic electronic search was carried out through various databases that emphasize current studies on the detection or diagnosis of oral cancer using different artificial intelligence algorithms. A modified Newcastle Ottawa scale was used to evaluate the quality of the included research and the PROBAST (Prediction model Risk of Bias Assessment tool) was used to assess the risk of bias. Among 13 articles, 8 show good quality and 5 fair qualities, mostly at low bias risk. Machine learning (support vector machines) sensitivity and specificity range from 89% to 92% and 75% to 82%; deep learning (MobileNet v2, ResNet) ranges from 85.12% to 90.23% and 87.64% to 90%. The diagnostic effectiveness of artificial intelligence models differs among machine learning and deep learning techniques. According to these results, machine learning has demonstrated encouraging outcomes in identifying oral cancer. The findings demonstrate the effectiveness of smartphone photographic images in detecting oral cancer.