Preliminary single-center feasibility study on the application of LNW-LacriMap based on fusion of lacrimal-nasal 3D model from CT-DCG and endoscopic image in EnDCR

To investigate the accuracy, reproducibility, and practicality of the Lateral Nasal Wall Lacrimal Map (LNW-LacriMap), generated by fusing a three-dimensional (3D) lacrimal-nasal model based on computed tomographic dacryocystography (CT-DCG) with endoscopic images, in assisting lacrimal sac localization during endoscopic dacryocystorhinostomy (EnDCR). This study aims to provide technical support for performing minimally invasive endoscopic lacrimal surgery. This preliminary single‑center feasibility study was conducted without a control group. This study was conducted in two phases: a simulation phase using archived EnDCR videos (16 patients, 16 eyes) and a prospective clinical application phase (14 patients, 16 eyes). LNW-LacriMap was generated by aligning CT-DCG and endoscopic images. Under the guidance of an experienced lacrimal specialist, intraoperative lacrimal sac location maps created from EnDCR videos were compared with LNW-LacriMaps. Accuracy metrics, including horizontal absolute error, vertical absolute error, absolute distance error, absolute angle error, and distance ratio, were measured and calculated. In the simulation group, error performance and potential differences in accuracy metrics among different operators were analyzed to evaluate the technology’s accuracy and reproducibility. Subsequently, clinical application in the practice group was conducted to assess its real-world performance. Overall error analysis for both groups was performed to evaluate the accuracy of this technique. Comparative error analysis between groups and across different anatomical characteristics was conducted to assess its practicality under various conditions. IBM SPSS Statistics 25 was used for statistical analysis of the obtained accuracy metrics. All LNW-LacriMaps accurately reflected the intraoperative lacrimal sac position. In the simulation group, no statistically significant differences were found among three independent operators for any accuracy metric ( P > 0.05), indicating high reproducibility. In the overall analysis, the mean absolute errors for LNW-LacriMap in the horizontal and vertical directions were 36.84 px (0.62 mm) and 20.90 px (0.35 mm), respectively. The mean absolute distance error was 21.78 px (0.36 mm), the mean absolute angle error was 5.35° (2.98, 8.98), and the mean distance ratio was 1.03 ± 0.12. Comparison of accuracy metrics between the simulation and practice groups showed no significant statistical differences for most indicators ( P > 0.05), indicating good practicality and reproducibility of the method. Correlation analysis revealed a suggestive negative correlation between lacrimal sac volume and distance ratio ( r = -0.379, P = 0.032), and a suggestive positive correlation between the distance from the mid-portion of the lacrimal sac to the nasal cavity inner surface and the distance ratio ( r = 0.439, P = 0.012). All patients in the practice group successfully underwent EnDCR surgery without significant complications, achieving ideal postoperative outcomes. The LNW-LacriMap, based on the fusion of a CT-DCG-derived 3D lacrimal-nasal model with endoscopic images, demonstrates good accuracy and reproducibility in assisting lacrimal sac localization during EnDCR surgery. It appears to be a potentially economical and practical technique, providing strong support for clinical practice in performing minimally invasive endoscopic lacrimal surgery.