[SS24] Recent Trends in Modeling & Simulation of Cancer

[SS24-02] Efficient Deep Learning Models for Early Cancer Detection and Prognosis

*Bishal Chhetri¹, BV Rathish Kumar¹ (1. Indian Institute of Technology Kanpur, India (India))

Keywords:

Early Cancer Detection,Neural Networks,Explainable AI,Time Series Forecasting,SHAP Values

Cancer remains a major public health challenge, with its incidence continuing to rise every year. Among various types of cancer, breast cancer is the most common cancer with the highest mortality rate among women. Early detection is the key to the successful treatment of almost all cancers. If breast cancer is detected early, the chances of survival exceed 90$\%$. Today, artificial intelligence models have become powerful tools and are widely used for the early detection and diagnosis of various types of cancer. Accurate prediction of these models can help oncologists make informed decisions about treatment plans, save time and resources, and ultimately improve patient outcomes. In this talk, I will discuss about the new deep neural network model that we have proposed for early breast cancer detection. The proposed model was trained for 200 epochs using 30 neurons, the Adam optimizer with a learning rate of 0.001, and the binary cross-entropy loss function for loss computation. The model's performance was evaluated using accuracy, precision, recall, and F1 score, which were found to be 0.992, 1.000, 0.977, and 0.988, respectively. Explainable machine learning techniques such as SHAP and LIME are used to understand the mechanisms of model predictions and assess the contributions of the data features to the decision-making process. Further, I will discuss the power of Deep Neural Network models like LSTM and RNN in time series forecasting of tumor growth based on in silico cancer model time series data.