PATIENT-SPECIFIC 3-D LUNG MODELING: THE FUTURE OF PRECISION LUNG CANCER SURGERY?

Abstract

SESSION TITLE: Fellows Lung Cancer Posters SESSION TYPE: Fellow Case Report Posters PRESENTED ON: October 18-21, 2020 INTRODUCTION: Reconstruction software based on mathematical algorithms and machine learning are beginning to allow the processing of chest CT scans to generate highly specific 3D lung models. We describe our initial experience using this technology for simulation and planning of lung cancer resection. CASE PRESENTATION: DICOM data from patient CT scans was processed (IQQA-Lung, EDDA Technology, Princeton, NJ, USA) to generate 3D maps of patient's airways and vasculature as well as obtain information on lung volumes (cm3), tumor volume and volumes of individual lung lobes. Case 1: The model [Figure 1(A-G), Gif 1] generated demonstrated a 2.2 cm x 2.2 cm x 1.7 cm lesion in the left upper lobe (LUL). The lesion had a total volume of 3.39 cm3. Total lung volume (TLV), right lung volume (RLV) and left lung volume (LLV) were 4029.54 cm3, 2369.08 (58.79%) cm3 and 1660.46 (41.21%) cm3 respectively. LUL and left lower lobe (LLL) had volumes of 951.02 cm3 (57.27%) and 709.45 cm3 (42.73%) respectively. While the lesion appeared to be located in the lingular segment, our model showed that it crossed into the adjacent segments and a segmentectomy would have led to a close or positive margin. This was confirmed in the operating room. Case 2: The model [Figure 1(H-N), Gif 2] generated demonstrated a 2.6 cm x 2.3 cm x 1.3 cm lesion in the right lower lobe (RLL). The lesion had a total volume of 4.72 cm3. TLV, RLV and LLV were 2948.46 cm3, 1579.84 (53.58%) and 1368.62 (46.42%) cm3 respectively. Right upper lobe (RUL), right middle lobe (RML) and RLL had volumes of 738.09 cm3 (46.72%), 178.57 cm3 (11.3%) and 663.18 cm3 (41.98%) respectively. While the lesion appeared to be located in the superior segment, interplanar analysis showed that a segmentectomy would have led to a close or positive margin. This was confirmed intraoperatively. In addition, a small proximal pulmonary artery was identified preoperatively that informed on the need for extra vigilance during dissection of the PA. DISCUSSION: Highly accurate patient-specific lung models provide extensive detail on pulmonary vasculature and intersegmental planes, allowing for preoperative recognition of anatomic variations and guiding the extent of an ideal oncologic resection. The ability to get high-resolution images without the need for IV contrast is an added benefit. These models also provide an accurate measurement of postoperative residual lung volumes hence providing a surrogate marker for postoperative residual lung function and quality of life. These models also carry teaching value for residents and medical students. CONCLUSIONS: Highly accurate 3D lung models give detailed information on patient-specific anatomy as well as preoperative and postoperative lung volumes. These platforms should be utilized to obtain a better understanding of patient anatomy in order to make lung operations safer, more efficient and allow appropriate lung conservation surgery. Reference #1: Nakada, T., Akiba, T., Inagaki, T. and Morikawa, T., 2014. Thoracoscopic anatomical subsegmentectomy of the right S2b + S3 using a 3D printing model with rapid prototyping. Interact Cardiovasc Thorac Surg, 19(4), pp.696-698. Reference #2: Akiba T, Nakada T, Inagaki T. Three-Dimensional Pulmonary Model Using Rapid-Prototyping in Patient with Lung Cancer Requiring Segmentectomy. Annals of Thoracic and Cardiovascular Surgery. 2014;20(Supplement):490-492. DISCLOSURES: No relevant relationships by Mirza Zain Baig, source=Web Response No relevant relationships by Faiz Bhora, source=Web Response No relevant relationships by Zaid Muslim, source=Web Response