Explainable AI Based Feature Importance Analysis for Ovarian Cancer Classification with Ensemble Methods

Ashwini Kodipalli ^1* V Susheela Devi ¹ Shyamala Guruvare ^2* Taha Ismail ³

• ¹ Indian Institute of Science (IISc), Bangalore, India
• ² Department of Obstetrics & Gynaecology, Kasturba Medical College, Manipal, India
• ³ Department of Radiology, Kanachur Institute of Medical Sciences, Mangaluru, India

Ovarian Cancer (OC) is one of the leading cancers with a significant impact on women. Despite recent advances in the medical field, such as surgeries, chemotherapy, and radiotherapy interventions, there are only marginal improvements in diagnosis using clinical parameters, as the symptoms at the early stages are very nonspecific. Due to advances in computational algorithms, such as ensemble machine learning, it is now possible to identify complex patterns in clinical parameters. However, these complex patterns do not provide deeper insights into prediction and diagnoses. Explainable Artificial Intelligence models, such as LIME and SHAPLEY Kernels, can provide insights into the decisionmaking process of ensemble models, thus increasing their applicability. The main aim of this research work is to design a computer-aided diagnostic system that accurately classifies and detects ovarian cancer. To achieve this objective, a 3-stage ensemble model and a game-theoretic approach based on SHAPLEY values were built to evaluate and visualize the results, thus analyzing the important features responsible for the prediction. The results demonstrate the efficacy of the proposed model with an accuracy of 98.66%. The proposed model's consistency and advantages are compared with single classifiers. The SHAPLEY values of the proposed model are validated by conventional statistical methods such as the p-test and Cohen's d-test to highlight the efficacy of the proposed method. To further validate the ranking of the features, we compared the p-values and Cohen's d-values of the top five and bottom five features. Proposing and validating an AI-based OC detection, diagnosis, and prognosis method on multi-modal real-life data, which mimics the move of a clinician approach with a demonstration of high performance, is the contribution of this study. The proposed strategy can lead to reliable, accurate, and consistent AI solutions for the detection and management of OC with higher patient experience and outcome at low cost, low morbidity, and low mortality; it can help millions of women living in resource-constrained and challenging economies.