Thirona’s artery-vein and segment segmentation helps in lung cancer surgeries

A new paper has just been published in high-impact journal JTCVS Techniques, showing the added value of AI-based lung segmentation and VR technology in lung cancer surgery planning. In the study from Erasmus Medical Center, Thirona’s lung CT segmentation was visualized in a 3D virtual reality environment from MedicalVR for planning complex lung segmentectomies. The paper shows that using this technology, surgeons made different chirurgical decisions in 40% of the cases, sparing healthy lung tissue.

Segmentectomies are increasingly used as a lung-sparing resection technique to remove small metastases from the lungs. Using this method, only a small segment of the lungs is removed resulting in less impact on the patient and a quicker recovery. However, the success of a segmentectomy depends on the surgeon’s understanding of the lung anatomy. Previously, the segmental anatomy was determined from 2D or 3D CT images, but it is difficult to accurately understand the complete anatomy of the lungs this way. MedicalVR, together with Erasmus MC, Thirona, and Fysicon, developed PulmoVR: a dedicated immersive 3D-VR and AI-based segmentectomy surgical planning tool.

Thirona applies artificial intelligence to identify lobes, airways, arteries and veins in the lung and annotate the segments in the target lobe. The annotations are loaded onto the VR platform PulmoVR where the surgeon can see the anatomy of the patient in 3D.

Cardiothoracic surgeons of Erasmus MC used our lung segmentations to assess 10 patients before surgery. In 40% of the surgeries, they decided upon a different strategy for the surgery after reviewing the lung anatomy in virtual reality. Moreover, all oncological resection procedures were performed successfully and adequately using the PulmoVR solution.

Surgeons reported a better understanding of the patient’s anatomy and could better prepare for surgery when they had assessed the anatomy using the PulmoVR platform. The successful application of our artery-vein and segmental segmentation in a VR model proves that AI can make a big impact in healthcare.

Link to full article: https://www.jtcvstechniques.org/article/S2666-2507(21)00253-4/fulltext