Abstract: Based on the very good measurements, we outline a therapeutic procedure which should enable the production of patient-specific NSP implants based on endoscopic data only.
Introduction
Nasal septum defects (NSD) represent a significant clinical challenge in rhinology. These perforations in the cartilaginous or bony septum can lead to a variety of symptoms, including impaired nasal breathing, crusting, and recurrent epistaxis. While surgical closure is the preferred treatment, many patients are managed conservatively with prosthetic "septum buttons."
The efficacy of these implants depends heavily on a precise fit. Currently, clinicians must choose between invasive silicone impressions, which often necessitate general anesthesia, or CT imaging, which involves ionizing radiation. This study, published in the journal HNO (2022), introduces a novel stereo-endoscopic approach designed to provide high-accuracy measurements in a radiation-free, outpatient setting.
Methods: Stereo-Endoscopic 3D Reconstruction
The proposed method utilizes a stereo-endoscopic pipeline to reconstruct the 3D geometry of the defect. The researchers employed specialized medical hardware, including a XION Medical stereo endoscope and a Schölly Fiberoptic 3D laparoscope.
The technical workflow consists of several critical stages:
- Calibration: An offline process using checkerboard patterns to establish internal camera parameters and the spatial relationship between the stereo sensors.
- Feature Matching and Disparity Estimation: The software identifies robust landmarks across the stereo image pairs. By calculating the pixel-shift (disparity) between these views, the system derives depth information.
- Metric Reconstruction: Calibration data allows the conversion of pixel-based depth maps into actual metric dimensions (millimeters).
- Measurement Acquisition: The system calculates the axial extension, coronal extension, and total circumference of the NSD.
Key Results and Validation
The system was validated against high-resolution CT scans (0.5mm–0.625mm slice thickness) in a clinical cohort of six patients with varying defect etiologies, including trauma and autoimmune disease.
The quantitative results demonstrated impressive accuracy, with all measurements falling within a 5% margin of error compared to CT (excepting defects exceeding the endoscope's field of view):
- Axial Axis Error: Mean absolute difference of 0.28 mm.
- Coronal Axis Error: Mean absolute difference of 0.16 mm.
- Circumference Error: Mean absolute difference of 0.904 mm.
A key clinical insight emerged regarding the difference between mucosal and structural defects. While CT scans primarily visualize the bony or cartilaginous defect, the endoscope captures the mucosal defect—the actual "functional" hole. In many cases, the mucosal opening is smaller than the underlying structural gap, making endoscopic measurement superior for prosthetic fitting.
Discussion and Future Directions
The study confirms that image-based 3D reconstruction can match the accuracy of CT scans for NSD measurement, potentially eliminating the need for radiation-heavy imaging in prosthetic planning.
The primary limitation encountered was the field of view (FOV) of the endoscope, which hindered the measurement of very large defects (e.g., >4 cm). Future research will focus on "3D Mosaicking" and Simultaneous Localization and Mapping (SLAM) to stitch multiple views into a comprehensive map. Ultimately, this technology aims to enable a seamless "Scan-to-Print" workflow, where custom 3D-printed implants are generated directly from endoscopic data.
Summary updated