TUMOR MARKERS A DIAGNOSTIC TOOL FOR ORAL CANCERS CONCERNING ARTIFICIAL INTELLIGENCE

Abstract

Objective: This research aimed to unravel the intricacies of demographic and molecular markers, along with AI prediction scores, in identifying the risk and presence of oral cancer. The goal was to offer a comprehensive analysis of the predictive power these markers hold and their potential integration into clinical practice.

Study design: Retrospective Cohort Study

Place and duration time: The study utilized data from patients who visited a medical center between January 2021 and December 2022. Rigorous analysis and evaluations were conducted over subsequent months.

Materials and methods: The study encapsulated data from 1,200 patients, extracting details on age, gender, ethnicity, smoking habits, personal and familial cancer histories, molecular markers (CK19, TPA, CEA, and p53 Antibodies Levels), and AI prediction scores. Statistical tools such as logistic regression models, Pearson correlations, and chi-square tests were employed to decipher patterns and relationships.

Results: The analysis exhibited weak correlations between most variables and AI Prediction Scores. Age had a faint positive influence on the prediction scores, and history of any cancer showed a slight negative tilt. Notably, a significant correlation was observed between family history of oral cancer and p53 Antibodies Levels. However, logistic regression results indicated high standard errors, suggesting potential issues with the model's specification.

Conclusion: While AI and molecular markers present a promising future for early oral cancer detection, this study underlines the complexities involved and the paramount importance of holistic patient assessment. Technological advancements, though pivotal, should be harmoniously integrated with clinical insights. More robust models and further research are imperative to streamline the utilization of AI and molecular markers in predictive diagnostics.