Quantitative Office-Based Laryngoscopy: Bench Validation of a Simple Calibration Method for Submillimeter Accuracy...

Past successful approaches to quantitative laryngoscopy relied on research prototypes and laser-add-ons which were not approved as medical devices, preventing the acquisition of quantitative laryngoscopic data in routine-diagnostics and complicating large-scale studies. We present a simple one-shot reference-grid calibration for a CE-marked stereo-laryngoscope commercially available for clinical use and validate it on an optical bench.

Using a 3D stereo-laryngoscope (XION GmbH, Berlin, Germany), we measured target distances (TD) of 2.5, 10, and 20 mm across working distances (WD) of 30, 45, 60, 75, and 90 mm in three runs, refocusing each time (30 measurements per target distance). A point-grid reference image with known spacing was captured. Relating each frame's stereo disparity to the reference grid yielded its px/mm scale. Precision was assessed as standard deviation (SD) of repeated measurements and accuracy was evaluated by Bland-Altman-analysis, reported as bias and 95% limits of agreement (LoA).

Over all, measurements showed a SD of 1.0%, a small bias of -0.3% with LoA from -2.2% to +1.6%. Precision and accuracy remained stable across TD, indicating even measurement performance across the evaluated range. No relevant edge-of-field distortion was observed.

The proposed calibration method for a commercially available, CE-marked, off-the-shelf stereo endoscope yields precise, submillimeter distance measurements and is adequate for clinical and scientific use in office-based transoral laryngoscopy, without custom hardware.