2024 World Conference on Lung Cancer (WCLC) - Posters-P2.05B.03 A Novel Lung Nodule Localize Method: Topographic Map Dyeing Model for Target Vessel Watershed Boundary

P2.05B.03 A Novel Lung Nodule Localize Method: Topographic Map Dyeing Model for Target Vessel Watershed Boundary

Pdf Summary

The document introduces a novel method for localizing lung nodules using a topographic map dyeing model, which targets the vessel watershed boundary. Developed by a team from Guangdong Provincial People’s Hospital and Guangdong Optomedic Technique Inc., the method addresses common complications and blind spots in the surgical localization of pulmonary nodules.<br /><br />The process leverages deep learning to enhance pulmonary CT scans, particularly focusing on vessel segmentation. This method oversamples the pulmonary blood vessels up to level 24, allowing for rapid arterial and venous segmentation within three minutes. These segmentations help in accurately defining the boundaries for staining in a 3D model using hemodynamics.<br /><br />A critical discovery made during the study was the irregularity of segment boundaries during surgical procedures like expansion and collapse or ICG (Indocyanine Green) dyeing after blocking target vessels. The model developed predicts these boundaries more accurately, potentially offering a better localization method prior to surgery.<br /><br />The method's effectiveness was validated by comparing simulated boundary textures with actual surgical outcomes. In a study involving 231 patients, 220 (95.24%) showed consistent boundary comparisons between the topographical map and real-life staining results, indicating the model's high accuracy and potential for clinical use.<br /><br />Overall, the topographic map dyeing model provides a reliable, non-invasive preoperative alternative for localizing lung nodules, reducing the risk of complications traditionally associated with surgery.

Asset Subtitle

Zihao Chen

Meta Tag

Speaker Zihao Chen

Topic Pulmonology & Staging

Keywords

lung nodules

topographic map dyeing

vessel watershed boundary

deep learning

pulmonary CT scans

vessel segmentation

hemodynamics

ICG dyeing

preoperative localization

surgical complications