For years, the debate around AI has swung between excitement and anxiety, but researchers are now putting this technology to work on humanity’s toughest challenge the mysteries of cancer.
At the centre of one such effort is Debarka Sengupta, associate dean of Innovation, Research and Development, Indraprastha Institute of Information Technology Delhi (IIIT-Delhi), who is using AI and genomics to detect cancer earlier, understand how tumours behave and help doctors choose treatments tailored to individual patients.
Rather than viewing cancer as a single disease or a mutation in one gene, Sengupta’s laboratory studies it as a complex biological system by combining molecular biology, genomics, single-cell analysis, microfluidics and AI.
“The aim is to detect weak cancer signals that are often hidden in blood, tissue or massive biological datasets and turn them into information that doctors can actually use,” Sengupta told PTI.
Traditionally, cancer research has focused on studying one gene or one biomarker at a time. AI, he explained, allows researchers to analyse thousands of genes, different cell types and clinical records simultaneously, uncovering patterns that would be almost impossible to identify manually.
“The real value of AI is not just speed. It helps researchers discover patterns that would be extremely difficult to see manually, especially when the signal is distributed across thousands of genes, many cell types and multiple experimental systems,” he said.
Among the team’s major achievements is the development of an 11-gene blood test based on platelet RNA that could eventually become an affordable screening tool for multiple cancers.
Unlike expensive genome sequencing technologies, the test is designed to work on RT-PCR machines, the same technology widely deployed across India during the COVID-19 pandemic.
“Such a test could potentially be run in the same kind of qPCR-equipped molecular labs that scaled during COVID testing, making broad deployment much more practical in India and similar settings,” Sengupta said.
He added that they have also been working on circulating tumour cell detection in triple-negative breast cancer, where the challenge is to find extremely rare cancer cells in the blood.
“That work was exciting because it combined molecular biology, microfluidics and AI,” he added.
But detection is only one part of the puzzle.
The researcher also focuses on AI models that can predict how individual cancers are likely to respond to different medicines, potentially helping doctors move away from the current trial-and-error approach to treatment.
Through a startup called GeneSilico, the team is building what it calls an “Agentic Digital Twin”, an AI-powered virtual model that combines a patient”s molecular profile, clinical history, tumour biology, treatment guidelines and scientific literature to help oncologists evaluate possible treatment options.
“The goal is not to replace doctors,” Sengupta said.
“It is to provide them with a deeper evidence layer so they can better understand which therapies appear biologically plausible and which treatment strategies have stronger scientific support.”
Despite rapid advances, Sengupta cautioned that AI is still far from becoming an independent decision-maker in hospitals.
“We are closer than many people think, but these technologies need rigorous clinical validation, regulatory oversight and careful integration into medical practice,” he said.
Sengupta believes India has a unique opportunity because of the molecular testing infrastructure built during the COVID-19 pandemic, which could eventually support affordable cancer diagnostics if such technologies are successfully translated into clinical use.
Asked whether AI is sometimes overhyped, Sengupta said the technology delivers its greatest value when it helps researchers manage biological complexity rather than attempting to replace clinicians.
“In oncology, AI should function as a scientific reasoning tool that supports doctors, not as a black box making decisions on its own,” he said.
Looking ahead, his laboratory plans to further validate blood-based cancer detection methods and improve AI systems that can predict drug responses using genomic and clinical data.
If these efforts succeed, cancer patients a decade from now may no longer rely only on a one-time biopsy report.
Instead, doctors could continuously update a patient’s disease profile using blood tests, imaging and genetic information, allowing treatments to evolve alongside the cancer itself.
“The vision is to make cancer care more personalised, more evidence-based and ultimately more accessible,” Sengupta said.
This story has been sourced from a third party syndicated feed, agencies. Mid-day accepts no responsibility or liability for its dependability, trustworthiness, reliability and data of the text. Mid-day management/mid-day.com reserves the sole right to alter, delete or remove (without notice)