Artificial Neural Network for Enhanced Risk Assessment in Thyroid Cancer Patients

Abstract

Thyroid cancer is among the most rapidly increasing endocrine tumors, presenting considerable difficulties for prompt diagnosis and precise risk assessment. Conventional evaluation frameworks, such the American Thyroid Association recommendations and TNM staging, provide valuable classification although are limited in their capacity to include various risk variables for accurate, personalized forecasts. This research utilizes an Artificial Neural Network (ANN) to improve risk assessment in thyroid cancer patients, therefore addressing these constraints. The model was created and assessed using the publicly accessible Thyroid Cancer Risk Dataset from Kaggle, including a comprehensive array of demographic, lifestyle, clinical, and hormonal characteristics. The ANN showed enhanced efficacy in identifying patients at varying risk levels by using its capacity to record complex, non-linear connections. The quantitative assessment indicated that the proposed ANN had an overall accuracy of 99.98%, exceeding traditional approaches such logistic regression and decision trees. These findings validate the capability of ANN-based methodologies to provide more dependable and individualized forecasts, thereby assisting physicians in refining treatment regimens, reducing superfluous treatments, and progressing precision oncology in the management of thyroid cancer.

Ähnliche Arbeiten