TUMOR CELL CLASSIFICATION: AN APPLICATION OF MULTIVARIATE DATA PROCESSING METHOD
Calwin S. Parthibaraj, Freeda Mary Selvaraj, Anna P. Joseph · JP Journal of Biostatistics · 2023
Oral Squamous Cell Carcinoma (OSCC) and precancerous lesion Oral Leukoplakia (OL) are usually diagnosed manually by microscopic study of biopsy slides. Such manual microscopic evaluation and classification of histopathological images are subjective, time-consuming, and complex among diagnosing experts due to the influence of multiple factors. Early and precise diagnosis of tumor cells is a prerequisite for reducing oral cancer mortality rate. Hence, the combined influence of multiple factors on tumor prognosis must be statistically studied to reveal the correct inferences and knowledge from clinical data to train deep learning models efficiently. In our retrospective study, clinical experiments were performed using Galectin-3 biomarkers, and the factors were graded by experts to obtain a historical clinical dataset. Multivariate Analysis of Variance (MANOVA) statistical data analysis was performed and implemented using Python code on the experimental dataset, and its inferences on specific questions were found helpful to the tumor grading experts. Received: April 15, 2023 Accepted: July 3, 2023