After colorectal surgery, one in five patients who develop a leak at the reconnected bowel will die. Surgeons currently lack a reliable way to see whether the tissue they are stitching together has enough oxygen to heal. A team from EnAcuity, Imperial College London, and UCL is building a software-only solution: EnSurge reads standard video feeds from cameras already in the operating theatre and converts them into hyperspectral images that reveal blood oxygen levels in real time. No new hardware is needed. The team will train their deep-learning model on gallbladder surgery videos—seven times more common and simpler than colorectal procedures—then fine-tune it on colorectal footage they already hold. If successful, EnSurge will receive UKCA/CE marking for colorectal surgery and later expand to gallbladder procedures, potentially doubling projected sales. The ultimate impact is straightforward: fewer anastomotic leaks, fewer readmissions, fewer deaths, and lower surgical costs. This is applied medical device software with a clear path to clinical use.
View original technical description
In colorectal surgery, the surgeon removes the tumour-affected bowel section and reconnects the two healthy bowel ends to form an 'anastomosis'. An insufficient oxygenated blood flow to the reconnected bowel can lead to anastomotic leaks (AL). Patients with AL have the risks: * 19.7% being readmitted for further treatment chance * 19.3% getting an infection chance * 15.9% death chance. Hyperspectral Imaging (HSI) allows surgeons to visualise blood oxygen levels. This helps them to avoid performing an anastomosis on poorly-oxygenated bowel and decrease AL risk. Current HSI devices are large, expensive, hard to integrate into surgery and cannot collect both detailed colour and spatial information. Our software innovation is a fast and cost-effective solution to medical HSI. Our software reads the existing hospital camera images and calculates the corresponding HSI. In this way, the surgeon can see the oxygenation without new HSI hardware. In this project, EnAcuity, Imperial, and UCL will develop and evaluate the first software-only HSI solution for colorectal surgery, which will be named "EnSurge". The EnSurge development requires HSI surgical video collection for deep-learning-based algorithm training/testing. The target application is colorectal surgery, but gallbladder surgery is 7x more common and simpler. Both surgeries include common visuals, such as blood vessels and tissue layers. To speed-up development and reduce patient risks, we will collect gallbladder surgery videos. We will use these gallbladder videos to pre-train our deep-learning model for HSI reconstruction. Then, we will fine-tune the model using our target colorectal videos. We already hold colorectal videos from privately-funded work. The project will result in UKCA/CE marking of EnSurge for colorectal surgery. Beyond this project, the gallbladder library will be used to create an EnSurge gallbladder surgery version. This will expand our projected sales two-fold. The ultimate impact is helping surgeons perform safer, more effective, and less expensive surgeries.
Plain English summaries and category classifications on this site are generated by AI and may not perfectly reflect the original research.
Is something wrong? Let us know