A formal agreement between PGI physicians and IIT Delhi lays the groundwork for building a domestically developed AI-powered tool aimed at detecting breast cancer at an early stage among women in India. Calibration data drawn from patient records within India — rather than the Western-population cohorts on which most global diagnostic frameworks rely — gives the tool a population-specific foundation.
Prof Veenu Singla, of PGI's radiodiagnosis and imaging department, has framed that distinction as consequential in the context of related research: breast cancer among women in India tends to appear at younger ages, frequently in the third or fourth decade of life, and the higher tissue density common in this group can reduce the effectiveness of standard mammography.
At present, a compact hardware unit that sits on a desktop and connects to a computer feeds mammography scans through digital analysis software, with the setup still being evaluated in a pilot programme. Measuring roughly 1.5 times the footprint of a standard puzzle cube, the unit is housed in a form factor that runs approximately five inches on each side.
PGI is drawing on its own patient records — spanning the full range of breast density classifications — to strengthen the tool's precision. Prof Veenu noted that once the system clears development and validation, the team intends to roll it out to healthcare facilities nationwide as a domestically built breast cancer screening solution.
Efforts are also under way within the department to broaden what the technology can handle, extending its analytical reach into three-dimensional breast imaging — known as tomosynthesis — as well as ultrasound video footage. The ultrasound module's software is designed to analyse tissue in real time, distinguishing fluid-filled cysts from tumours while also penetrating dense breast tissue, an area where conventional X-ray mammography has known performance constraints.
During routine check-ups, the tool functions as a secondary analytical layer, reducing the likelihood of misleading alerts and unwarranted tissue sampling while delivering a uniform standard of care across the patient population.
Three institutions — AIIMS, Maulana Azad Medical College, and PGIMER — have pooled their resources to introduce a specialised training programme covering musculoskeletal and breast imaging, with operations under that initiative taking hold at PGI. Crafted by faculty members in the field, the curriculum traces how radiology's tools and techniques have changed over the past ten years.
Among the focal areas are high-resolution imaging, the integration of artificial intelligence into diagnostic workflows, radiomics, functional imaging, and procedures guided by real-time image data.
Published health data puts the likelihood of a woman in India receiving a breast cancer diagnosis at some point in her life at roughly one in 28. Among women in India's urban centres, breast cancer accounts for roughly 23 to 30 percent of all female cancer diagnoses, making it the most frequently recorded type.
Research indicates that breast tissue density plays a measurable role in cancer likelihood, with those whose mammograms show highly dense tissue facing a rate of diagnosis approximately twice that recorded among patients whose tissue registers as fatty or low in density. Breasts classified as dense contain a higher proportion of glandular and fibrous tissue relative to fatty tissue, and standard mammography and other established detection tools lose reliability under those circumstances.
Designed to scan digital mammography images, the tool works by flagging irregularities that could indicate the presence of breast cancer. Rollout of the tool now covers breast tomosynthesis, a scanning method that builds layered, three-dimensional views of breast tissue.
An ultrasound attachment under development would scan for abnormalities as they occur and provide a means of distinguishing fluid-filled cysts from masses with solid tissue. The tool is designed to enable more precise diagnostic readings, reduce unnecessary needle biopsies, and lower the rate of false alarms.
