Author Block: Gabriela Lahaie Luna, Jean-Marie Parel, Alex Gonzalez, Cornelis Rowaan, Wilma Hopman, Sarit Khimdas, Martin tenHove

Author Disclosure Block: G. Lahaie Luna: None. J. Parel: None. A. Gonzalez: None. C. Rowaan: None. W. Hopman: None. S. Khimdas: None. M. tenHove: None.

Abstract Body:

Purpose: To validate the use of a mechanized remotely operated stereoscopic drone slit lamp (DSL) in assessing anterior segment pathology in ophthalmology patients, comparing it to a conventional slit lamp (CSL).

Study Design: Cross Sectional Study Validating the use of a medical device.

Methods: 43 patients recruited from clinics at Hotel Dieu Hospital (Kingston, ON). 42 patients met the inclusion criteria and were assessed by two examiners with a level of training ranging from PGY2 residents to attending staff. Patients were assessed using the DSL and then a CSL. Examiners used both the DSL and CSL to assess the anterior chamber (AC) depth using Van Herick technique and a simplified grading scheme (open, closed or needing gonioscopy). The presence or absence of cells and/or flare, using a modified SUN criteria grading scheme was also assessed. Qualitative data was collected on the ability to assess corneal integrity, foreign bodies, epithelial defects, stromal infiltrates and conjuctival injection using the DSL.

Results: 48 eyes of 42 participants were examined using both the DSL and CSL. When comparing the data for each examiner, no statistically significant differences were identified between the novice and experienced examiners. There was substantial agreement between the DSL and CSL when assessing AC cell and flare (kappa 72.6 and 60.4 respectively) and moderate agreement when assessing AC depth (kappa 42.5). The DSL compared to CSL had a sensitivity and specificity of 98.3% (95% CI= 94-100) and 100% (95% CI= (98.7-100) respectively, for detecting AC cell. The DSL had sensitivity and specificity of 100% (95% CI= 97.5-100) and 88.2% (95% CI= 80.2-96.1) respectively for detecting AC flare.

Conclusions: There was substantial agreement between the drone and conventional slit lamp when assessing subtle anterior segment findings including AC depth, cell and flare. The sensitivity and specificity of the drone slit lamp for assessing these findings ranged from 88.2% to 100%. This drone slit lamp provides excellent capability for examination of anterior segment pathology in live patients when compared to a conventional slit lamp.