Virtual reality and artificial intelligence for 3-dimensional planning of lung segmentectomies

Amir H Sadeghi; Alexander P W M Maat; Yannick J H J Taverne; Robin Cornelissen; Anne-Marie C Dingemans; Ad J J C Bogers; Edris A F Mahtab

doi:10.1016/j.xjtc.2021.03.016

Virtual reality and artificial intelligence for 3-dimensional planning of lung segmentectomies

JTCVS Tech. 2021 Mar 16:7:309-321. doi: 10.1016/j.xjtc.2021.03.016. eCollection 2021 Jun.

Authors

Amir H Sadeghi¹, Alexander P W M Maat¹, Yannick J H J Taverne¹, Robin Cornelissen², Anne-Marie C Dingemans², Ad J J C Bogers¹, Edris A F Mahtab¹

Affiliations

¹ Department of Cardiothoracic Surgery, Thoraxcenter, Erasmus Medical Center Rotterdam, Rotterdam, The Netherlands.
² Department of Pulmonary Medicine, Erasmus Medical Center Cancer Institute, Rotterdam, The Netherlands.

Abstract

Background: There has been an increasing trend toward pulmonary segmentectomies to treat early-stage lung cancer, small intrapulmonary metastases, and localized benign pathology. A complete preoperative understanding of pulmonary anatomy is essential for accurate surgical planning and case selection. Identifying intersegmental divisions is extremely difficult when performed on computed tomography. For the preoperative planning of segmentectomies, virtual reality (VR) and artificial intelligence could allow 3-dimensional visualization of the complex anatomy of pulmonary segmental divisions, vascular arborization, and bronchial anatomy. This technology can be applied by surgeons preoperatively to gain better insight into a patient's anatomy for planning segmentectomy.

Methods: In this prospective observational pilot study, we aim to assess and demonstrate the technical feasibility and clinical applicability of the first dedicated artificial intelligence-based and immersive 3-dimensional-VR platform (PulmoVR; jointly developed and manufactured by Department of Cardiothoracic Surgery [Erasmus Medical Center, Rotterdam, The Netherlands], MedicalVR [Amsterdam, The Netherlands], EVOCS Medical Image Communication [Fysicon BV, Oss, The Netherlands], and Thirona [Nijmegen, The Netherlands]) for preoperative planning of video-assisted thoracoscopic segmentectomies.

Results: A total of 10 eligible patients for segmentectomy were included in this study after referral through the institutional thoracic oncology multidisciplinary team. PulmoVR was successfully applied as a supplementary imaging tool to perform video-assisted thoracoscopic segmentectomies. In 40% of the cases, the surgical strategy was adjusted due to the 3-dimensional-VR-based evaluation of anatomy. This underlines the potential benefit of additional VR-guided planning of segmentectomy for both surgeon and patient.

Conclusions: Our study demonstrates the successful development and clinical application of the first dedicated artificial intelligence and VR platform for the planning of pulmonary segmentectomy. This is the first study that shows an immersive virtual reality-based application for preoperative planning of segmentectomy to the best of our knowledge.

Keywords: 2D, 2 dimensional; 3D, 3 dimensional; AI, artificial intelligence; CT, computed tomography; DICOM, digital imaging and communication in medicine; NSCLC, non–small cell lung cancer; S, segment; VATS, video assisted thoracoscopic surgery; VR, virtual reality; lung cancer; preoperative planning; segmentectomy; video-assisted thoracoscopic surgery; virtual reality.