IIT-Madras, CMC build AI toolkit to detect early kidney diseases

CHENNAI: Researchers from IIT Madras and Christian Medical College (CMC), Vellore, have developed a set of artificial intelligence (AI)-based tools designed to assist in the early detection and assessment of kidney diseases, enabling timely intervention before significant damage occurs, officials said.

The technologies comprise a machine learning model that uses clinical and laboratory data to predict the risk of chronic kidney disease (CKD); a deep learning system that automatically analyses CT scans and classifies them into four categories: normal kidney, kidney cyst, kidney stone and kidney tumour; and a 3D imaging platform that recreates kidneys from CT scans to precisely assess tumour volume and the percentage of kidney involvement.

The CT image classifier has been trained with more than 12,000 images and can distinguish healthy kidneys from cysts, stones and tumours.

The 3D imaging framework developed using open-source software offers a low-cost and repeatable method of assessing tumour burden, providing information that could assist treatment planning.

According to GL Samuel, professor at the Department of Mechanical Engineering, IIT Madras, kidney diseases are often asymptomatic in their early stages and are not diagnosed until substantial damage has occurred.

The AI tools aim to assist physicians by delivering consistent results quickly, in a fast-paced healthcare environment.

"These tools can reduce the need for expensive interventions like dialysis. We used machine learning along with clinical knowledge to develop tools that would assist in the earlier detection of kidney diseases and give more detailed information specific to the patient," Samuel said.

Jennifer Delighta, research scholar at IIT Madras, emphasised that early detection is of paramount importance when dealing with kidney diseases.

"These AI tools can help detect at-risk patients early and plan their treatment more effectively. The CKD model has been developed as a user-friendly prototype, with further work aimed at improving its accuracy and interpretability for clinicians," Delighta said.

Supported by IIT-Madras and the SPARC project, the study is also a step toward developing a kidney digital twin.

The researchers plan to validate the models using larger patient datasets and explore their integration with wearable sensing systems for personalised kidney health monitoring.