Fraunhofer Institute for Production Technology IPTResearch project "KIVI"
Challenge
In many cancer surgeries – for example on the liver, pancreas, or bowel – surgeons need to completely remove the tumor while preserving as much healthy tissue as possible. The boundary between healthy and diseased tissue is often hard to differentiate with the naked eye or with imaging modalities like CT and MRI during surgery.
The current gold standard is intraoperative frozen section analysis: during the operation, small tissue samples (biopsies) are taken repeatedly, sliced very thinly, stained, and examined under a microscope in pathology. By the time the result is available, typically 20 to 40 minutes pass per biopsy. In Germany this leads to tens of thousands of hours of waiting time in the operating room each year and costs in the millions. Additionally, anesthesia and operation times are extended for patients – increasing risk of complications. Pathology and operating room capacity are increasingly strained. Thus, there is a strong need for a method that requires fewer biopsies, yields results much faster, and meaningfully complements intraoperative frozen section diagnostics.
Solution approach
In the KIVI project, we are developing a portable imaging system based on optical coherence tomography (OCT), combined with artificial intelligence (AI) methods. OCT can be described as “ultrasound with light,” revealing the tissue structure layer by layer with high resolution, without the need to cut the tissue.
With a hand-held OCT probe that is to be developed within the KIVI project, the surgical team can examine tissue directly in the operative field or on freshly obtained samples in a contactless manner. The resulting three-dimensional images are automatically analyzed by AI algorithms. The system should display within a few seconds whether the tissue is healthy or tumor-bearing with high probability.
The target is a sensitivity and specificity of over 95 percent, initially for liver tissue. The results are presented in a simple, understandable user interface so that surgeons can quickly determine where additional biopsies are needed. In the long term, KIVI is intended to serve as a platform that can also be transferred to other organs and applications – for example in endoscopic diagnostics or in the assessment of donor organs – thereby reducing the number of necessary biopsies and frozen sections.
Project partner
- Fraunhofer Institute for Production Technology IPT (Project leadership)
- Fraunhofer Institute for Digital Medicine MEVIS
- University Hospital Aachen
- HS Analysis GmbH
Associated partners
- Carl Zeiss AG
- Aigora GmbH
Funding
The project "KIVI" is funded by the Federal Ministry of Research, Technology and Space under the funding directive AI-supported precision surgery in oncology.
Funding code: 16SV9537
Project sponsor: VDI/VDE-IT
Project duration: 11/2025 - 10/2028