Innovation Hub India

INDIA

Leveraging AI to Combat Cancer

“We have processed over 2 million images in order to train our deep learning algorithms for ensuring clinical accuracy. We are aiming at processing 500 million images by the end of 2020, to further improve the accuracy of our products,” says Sutirth Vaidya, the firm’s co-founder & Chief Technology Officer.

A company focussed on deep-learning is leveraging AI to create clinical applications for the early detection of certain types of cancer.

Predible Health, a Bengaluru-based startup has been working hard at finding effective solutions to fight cancer, starting with diagnosis. The company was started in 2016 by Suthirth Vaidya, Abhijith Chunduru and Deepak Mohan, all of whom were students at the Indian Institute of Technology (IIT), Madras. Not too long after completing their research – which won the Longitudinal Multiple Sclerosis Lesion Segmentation Challenge held at the IEEE-International Symposium on Biomedical Imaging 2015 in New York, USA – their research drew significant media attention, and led them to seriously consider how they could differentiate their value proposition. They soon narrowed down on solving problems in the field of healthcare using AI and began to develop techniques to extend applications of 3D medical imaging to cancer diagnosis, culminating in the team founding the company.

The problem Predible Health sets out to address pertains to the detection of cancerous nodules at an early stage, finds out Subhalakshmi Roy. The nodules are small and often get missed in screenings, with a 50%-70% chance of escaping detection altogether. This would have an adverse effect on the treatment of the patient. Even when detected early, not all nodules are malignant, bringing the quality of diagnosis into play. The firm’s proprietary AI solution helps radiologists expose cancerous nodules faster and with heightened accuracy, also reducing the subjectivity involved in this kind of detection.

Currently, the firm’s flagship products include Predible Liver and Predible Lung. Predible Liver, a cloud-based solution, offers instant analysis along with 3D visualisation of liver lesions and hepatic structures. This lets surgeons plan for resectioning tumours and liver transplants with enhanced precision. Its presence on the cloud lets any hospital with an internet connection to access the relevant information with ease, contributing to lower morbidity through better-planned diagnosis and surgeries.

The Predible Health team combines Deep Learning and Cloud Computing to enable the future of radiology reporting

The company’s other solution, Predible Lung, supports medical professionals to detect lung nodules in CT scan images and speeds up the process of detection and assessment of malignancy. Early and accurate diagnosis through effective screening can, in turn, improve care for the patients being screened.

According to Suthirth Vaidya, Chief Executive Officer (CEO), Predible Health, “The clinical validation studies we undertook are yielding positive results, and now, we are moving towards putting our products into practical use.”

The company’s offerings have an impact on the healthcare industry by helping enhance access to high quality cancer care which is otherwise a highly specialised field. There aren’t too many centres that cater to cancer care and those that do, often deal with very large volumes and high demand. In order to ensure high quality care, the right kind of software is needed to ensure that very few errors take place at the stage of diagnosis and that the right kind of treatment can be initiated at the right time.

Predible’s products help address these problems, at scale, also helping hospitals run screening programmes in a wider geographical area. The company is piloting one of India’s first lung-cancer screening programmes, targeting people who have been smoking for a long time and asymptomatic patients as well. This is not only more economically viable, but can also have better outcomes for the patients being screened.

Having secured funding from Unitus Seed Fund, the firm now intends to scale its AI platform for screening cancer and also invest in regulatory approval for the products it has developed so far.

To increase its products’ reach, the firm is partnering with leading medical institutions such as Tata Memorial Hospital (Mumbai), Mahajan Imaging (Delhi) and Narayana Health (Bengaluru), for engaging in research on deep learning technologies that can help solve high-impact issues in healthcare, especially in developing countries.

Going forward, with its aim to cover end-to-end cancer care through its solutions including diagnosis, treatment and follow-ups for multiple kinds of cancer, the company is looking at expanding to other markets too, such as the US.