SimBioSys, an Illinois-based startup, has developed an AI-powered imaging technology that creates detailed three-dimensional models of breast tumours, offering surgeons unprecedented visualisation capabilities for cancer operations, according to an NVIDIA corporate blog post released October 31.
The FDA-cleared technology, called TumorSight, transforms standard black-and-white MRI scans into colour-coded 3D visualisations, with veins displayed in red, tumours in blue, and surrounding tissue in grey. The system provides crucial measurements, including tumour volume and distance from vital structures, helping surgeons make informed decisions about preservation versus mastectomy procedures.
"Typically, surgeons will get a radiology report, which tells them, 'Here's the size and location of the tumour,' and they'll get one or two pictures of the patient's tumour," Jyoti Palaniappan, chief commercial officer at SimBioSys, told NVIDIA's blog. "If the surgeon wants to get more information, they'll need to find the radiologist and have a conversation with them — which doesn't always happen."
The technology's development comes as breast cancer continues to affect millions globally. According to the World Health Organization statistics cited in the report, nearly 2.3 million women are diagnosed with breast cancer annually, with the disease claiming over 500,000 lives each year. In the United States alone, approximately 100,000 women undergo some form of mastectomy annually, according to Brigham and Women's Hospital data.
"We're trying to move the surgical process from an art to a science by harnessing the power of AI to improve surgical planning," Dr. Barry Rosen, SimBioSys's chief medical officer, explained to NVIDIA.
The company is also developing additional AI applications, including technology that converts MRI images taken with patients lying face down into visualisations showing how tumours will appear during surgery when patients are face up. The system accounts for gravity's effects on different tissue types and skin elasticity.
Furthermore, SimBioSys has created an AI model that can analyse tumour characteristics and patient data to generate rapid risk assessments for cancer recurrence. The NVIDIA blog reports that, according to Palaniappan, while traditional methods typically take up to six weeks for comprehensive molecular analysis, the new method can deliver results in hours while matching or exceeding traditional scoring approaches.
The startup utilises NVIDIA A100 Tensor Core GPUs in the cloud for model training, along with NVIDIA MONAI for training and validation data, and NVIDIA CUDA-X libraries for running its imaging technology. SimBioSys is a member of the NVIDIA Inception program for startups.