Now, cancer diagnosis just click away with artificial intelligence – The Tribune India
Tribune News Service
Naina Mishra
Chandigarh, June 11
The Post Graduate Institute of Medical Education and Research (PGIMER), Chandigarh, has started a multi-institutional research project involving cancer patients using artificial intelligence (AI) tool, with Tata Memor…….
Tribune News Service
Naina Mishra
Chandigarh, June 11
The Post Graduate Institute of Medical Education and Research (PGIMER), Chandigarh, has started a multi-institutional research project involving cancer patients using artificial intelligence (AI) tool, with Tata Memorial Hospital (TMH), Mumbai, AIIMS, New Delhi, and Rajiv Gandhi Cancer Institute, New Delhi, as the collaborating institutes.
All you will need to do is to upload a CT scan or pathology images in the software and the computer will automatically generate the report/diagnosis. It will act as a screening tool by differentiating normal from abnormal and, in turn, help in triaging cases that need expert opinion, saving a lot of time and money. — Dr Mandeep Garg, Professor, Department of Radiodiagnosis, PGI
Through this project, the PGI will collect a dataset of cancer-related radiology and pathology images from Indian patients and store it digitally. The AI tools developed from this database will work with high precision on Indian patients. It will assist a general radiologist or a pathologist in providing accurate diagnosis and expert assistance, just with a click of the button.
“All you will need to do is to load CT scan or pathology images in the software and the computer will automatically generate the report/diagnosis. It will act as a screening tool by differentiating normal from abnormal and in turn, help in triaging cases that will actually need the opinion of an expert saving lot of time and money,” explained Dr Mandeep Garg, Professor, Department of Radiodiagnosis, who is the principal investigator of the project from the PGI, Chandigarh.
“This is the first-of-its-kind project in India. Though there are many such open access databases available, which have tremendously helped cancer research in the West, but unfortunately, the experience of our scientists using images from these repositories is far from satisfactory as this data is primarily derived from western patients and don’t directly replicate on Indian population. Hence it prompted us to create our own Indian database,” said Dr Garg.
“The database that we will create will have many advantages over the existing western databases as it will be much more informative and comprehensive. It will include both radiology DICOM images and pathology whole-slide images. We will be linking it with clinical, genetic, follow-up and treatment outcome information. It will also have expert-verified annotations in place, which is missing in other databases,” Dr Garg added.
“While early detection and treatment of cancer remains crucial for a favorable outcome, there is a severe shortage of qualified onco-radiologists and onco-pathologists in the country. This increases patients’ waiting time to arrive at a correct diagnosis leading to significant delay in the start of their treatment,” said Dr Garg.
“We have started collecting data of patients with lung, head and neck cancers. In the next phase will include other cancers like breast and prostate,” Dr Garg added.
“We hope this project will serve as a template for others to follow and pave the way for similar initiatives not only in healthcare, but in several other areas as well,” said Dr MS Sandhu, head, Department of Radiodiagnosis.
Earlier, Dr Garg has did an AI-based research project with IIT- Kharagpur that was funded by the Department of Information Technology (DIT), Government of India. The current project is funded by the Department of Biotechnology (DBT), Government of India. This project was initiated by Dr Swapnil Rane from the TMH, Mumbai. Prof Amanjit Bal from the Department of Pathology and Dr Chirag Ahuja from the Department of Radiology are the other co-investigators of the project from the PGI.
