Purpose: Using a fully automated multitask deep learning method, which enabled simultaneous segmentation and quantification of all major anterior segment structures with swept-source optical coherence tomography (SS-OCT), we aimed to investigate the three-dimensional (3D) alterations in iris morphology before and after implantable collamer lens (ICL) surgery.
Methods: All enrolled patients underwent anterior segment SS-OCT (ANTERION) within one week before and after ICL surgery. A multitask network automatically performed iris SS-OCT image segmentation and quantitative measurements of 3D iris morphology (iris thickness and volume of the inner 1-mm annular area and the outer 1-2-mm annular area, iris curvature [I-Curve], and iris smooth index [SI]).
Results: Forty-six eyes of 27 patients (average age, 26.0 ± 5.3 years) were included in the final data analysis. The spherical equivalent ranged from -15.75 to -4.00 diopters (D). The mean ICL vault values were 537.92 ± 161.69 μm, which positively and significantly correlated with the pre- (r = 0.333, P = 0.024) and postoperative (r = 0.351, P = 0.017) anterior chamber depth (ACD). The mean pupil diameter significantly decreased, whereas both ACD and lens thickness significantly increased postoperatively. The iris thickness, iris volume of the inner 1-mm annular area and the outer 1-2-mm annular area, and SI decreased significantly within one week after ICL implantation. All mean values of the I-Curve increased dramatically postoperatively.
Conclusion: Iris morphology, ACD, and lens thickness were significantly altered after ICL implantation.
Keywords: Deep learning; ICL vault; Implantable collamer lens; Iris morphology; Swept-source optical coherence tomography.
© 2025 The Authors. Published by Elsevier Ltd.