Role of Artificial Intelligence in Oral Cancer Metastasis Detection Examined

Serdar HocamAuthor & Editor

A systematic review evaluating the effectiveness of artificial intelligence methods for detecting occult lymph node metastasis in oral cancer analyzed 15 studies involving 5,758 patients.

◉ 0 views
PRISMA-2020-flow-diagram-illustrating-the-study-selection-process.

A comprehensive systematic review examining the effectiveness of artificial intelligence approaches in detecting occult lymph node metastasis in oral cancer was conducted, analyzing 15 studies encompassing 5,758 patients.

Background and Scope of the Study

Oral cancer, and particularly oral squamous cell carcinoma, is a significant global public health issue, with metastasis being the primary cause of cancer-related deaths. Lymph node metastasis in the cervical region is the most critical prognostic factor.

Occult Lymph Node Metastasis

Occult lymph node metastasis refers to the presence of cancer cells in regional lymph nodes that cannot be detected by conventional clinical or radiological evaluations. Traditional imaging methods have limited sensitivity.

The Role of Artificial Intelligence in Oncology

Artificial intelligence, including machine learning, deep learning, and convolutional neural networks, has shown the potential to transform oncological imaging and diagnostic processes.

Review Methods and Data Collection

The systematic review followed PRISMA 2020 guidelines and was registered in the PROSPERO database. A literature search was conducted via PubMed, Scopus, Embase, Web of Science, and Google Scholar.

Included Studies and Results

A total of 15 studies meeting the eligibility criteria and included in the qualitative synthesis involved 5,758 patients. Artificial intelligence-based approaches were observed to exhibit variable diagnostic performance.

Performance and Evidence Quality Assessment

While AUC values reported in the studies ranged from 0.729 to 0.961, sensitivity was found to range from 0.65 to 1.00, and specificity from 0.576 to 0.98. Due to methodological heterogeneity, a quantitative meta-analysis was not performed.