AI-Powered Medical Devices Advance in Cancer and Surgery
Recent developments in medical technology are highlighting the growing role of embedded AI in healthcare. A 3D whole-body melanoma scanner that uses embedded AI and advanced imaging has launched in New Zealand for non-invasive cancer screening. Separately, a new study in *Nature Machine Intelligence* demonstrates the use of synthetic X-rays to track and control miniature medical devices inside the body, enabling more precise robotic interventions.
- The VECTRA WB360, a 3D whole-body scanner, utilizes 92 cameras to capture a patient's entire skin surface in under a second, creating a 3D avatar that maps all moles and lesions. This technology is designed to help dermatologists detect skin cancer at its earliest stages and avoid unnecessary biopsies. - The AI integrated into some 3D scanners is trained on extensive image datasets to automatically identify, measure, and compare lesions over time, flagging suspicious changes for dermatologists. In one study, an AI model demonstrated high accuracy in distinguishing melanoma from other skin lesions based on automated analysis of 3D images. - Synthetic X-ray technology addresses the challenge of obtaining large, well-annotated datasets for training medical AI models, which is often hindered by patient privacy concerns and the risks of radiation exposure. By generating realistic, labeled X-ray images, this technology enables the development of AI for applications like guiding needles in biopsies and optimizing C-arm positioning during orthopedic procedures. - In robotic surgery, AI is being leveraged to enhance precision and reduce complications. A meta-analysis of 25 studies showed that AI-assisted robotic surgeries led to a 25% reduction in operating time and a 30% decrease in intraoperative complications compared to manual methods. - Companies like NVIDIA are providing the foundational hardware and software for these advancements, with platforms like the NVIDIA Holoscan for real-time processing of medical device data at the edge. This enables the development of software-defined medical devices that can perform real-time AI inference, which is critical during surgical procedures. - Intel is also a key player, providing a range of hardware from CPUs to neural processing units that form the infrastructure for healthcare AI. Their technologies are used in medical imaging devices to help radiologists triage patient scans and to enable edge computing, which allows for rapid data processing directly on or near the device without needing to send data to the cloud. - Looking ahead, the integration of AI in robotic surgery is expected to become more intelligent and data-driven, with potential advancements in remote surgery capabilities and smarter decision-support systems in operating rooms. For cancer detection, AI is being explored for its potential to predict an individual's risk of developing cancer by analyzing their medical records. - The use of synthetic data is a growing trend in medical imaging AI development. Generative AI models, such as denoising diffusion probabilistic models (DDPMs), are being used to create synthetic images that can enhance the diversity of training datasets. Some studies have shown that AI trained on synthetic X-ray data can outperform models trained solely on real-world clinical images.
