Towards Image-Based Intraoperative Quality Control: Automated Extraction Of Femtosecond Laser Centration Metrics And Their Association With Early Refractive Deviation

To investigate the feasibility of automated extraction of femtosecond laser centration parameters from intraoperative images and to explore their potential role as intraoperative indicators of early refractive deviation following femtosecond laser–assisted cataract surgery (FLACS). 

Single specialized ophthalmic surgical center

Fifty homogeneous eyes undergoing uncomplicated FLACS were retrospectively analyzed. Preoperative biometry was obtained using IOLMaster. Intraoperative femtosecond laser centration screenshots were evaluated using an automated image-processing workflow to identify the pupil center and laser treatment center. Centration offset was calculated as the Euclidean distance between centroids. 

Early postoperative refractive deviation at 1 week was defined as deviation from target emmetropia. Associations between centration offset and refractive outcomes were assessed using correlation and logistic regression analyses. 

Automated extraction of centration metrics was successfully performed in all analyzed eyes. Measurable variability in centration offset was observed despite standardized surgical settings. Greater centration offset was associated with increased likelihood of early refractive deviation. 

Automated quantitative analysis of femtosecond laser centration using routine intraoperative imaging is feasible and may provide objective intraoperative quality indicators associated with early refractive outcomes. Further studies are warranted to evaluate its predictive value and potential integration into surgical workflows