Vault predicting after implantable collamer lens implantation using random forest network based on different features in ultrasound biomicroscopy images.

AIM: To analyze ultrasound biomicroscopy (UBM) images using random forest network to find new features to make predictions about vault after implantable collamer lens (ICL) implantation. METHODS: A total of 450 UBM images were collected from the Lixiang Eye Hospital to provide the patient's preoperative parameters as well as the vault of the ICL after implantation. The vault was set as the prediction target, and the input elements were mainly ciliary sulcus shape parameters, which included 6 angular parameters, 2 area parameters, and 2 parameters, distance between ciliary sulci, and anterior chamber height. A random forest regression model was applied to predict the vault, with the number of base estimators (n_estimators) of 2000, the maximum tree depth (max_depth) of 17, the number of tree features (max_features) of Auto, and the random state (random_state) of 40.0. RESULTS: Among the parameters selected in this study, the distance between ciliary sulci had a greater importance proportion, reaching 52% before parameter optimization is performed, and other features had less influence, with an importance proportion of about 5%. The importance of the distance between the ciliary sulci increased to 53% after parameter optimization, and the importance of angle 3 and area 1 increased to 5% and 8% respectively, while the importance of the other parameters remained unchanged, and the distance between the ciliary sulci was considered the most important feature. Other features, although they accounted for a relatively small proportion, also had an impact on the vault prediction. After parameter optimization, the best prediction results were obtained, with a predicted mean value of 763.688 µm and an actual mean value of 776.9304 µm. The R2 was 0.4456 and the root mean square error was 201.5166. CONCLUSION: A study based on UBM images using random forest network can be performed for prediction of the vault after ICL implantation and can provide some reference for ICL size selection.

Validation of the vault prediction model based on the sulcus-to-sulcus diameter and lens thickness: a 925-eye prospective study.

BACKGROUND: To verify the accuracy and stability of the prediction formula based on the ciliary sulcus diameter and lens thickness and to analyse factors influencing the prediction results. METHODS: In total, 925 eyes from 506 subjects were enrolled in this prospective study between July 1, 2020, and June 30, 2021. Subjects were divided into four seasons, each spanning three months. The target vault was set to be between 300 µm and 700 µm according the prediction formula. The actual vault was measured one month postoperatively. The Bland-Altman test, 95% confidence intervals (95% CI) and 95% limits of agreement (95% LoA) were used to evaluate the agreement between the predicted vault and the actual vault. Eyes with absolute prediction errors greater than 300 µm were further analysed. RESULTS: The mean predicted vaults for the four seasons were 503 ± 99, 494 ± 96, 481 ± 92 and 502 ± 93 µm, while the mean actual vaults were 531 ± 189, 491 ± 179, 464 ± 179 and 529 ± 162 µm, respectively. The predicted and actual vaults of the overall subjects were 493 ± 95 and 500 ± 180 µm, respectively. Of the 925 eyes, 861 eyes (93.08%), 42 eyes (4.54%), and 22 eyes (2.38%) showed a normal vault, high vault, and low vault, respectively. Bland-Altman plots showed that the mean difference between the actual vault and predicted vault overall (± 95% LoA) was 6.43 ± 176.2 µm (-339 to 352 µm). Three UBM features may lead to large prediction errors (more than 300 µm): wide iris-ciliary angle (ICA), iris concavity and anteriorly positioned ciliary body. CONCLUSIONS: This study demonstrated the accuracy and stability of the prediction formula through the validation of a large sample size and a long time span. Wide ICA, iris concavity and anteriorly positioned ciliary body may have an effect on vault.

A prediction model for sulcus-to-sulcus diameter in myopic eyes: a 1466-sample retrospective study.

BACKGROUND: To establish and verify the accuracy and reliability of a sulcus-to-sulcus diameter (STS) prediction model. METHODS: In this retrospective study, the prediction formula was established with the data from 1466 eyes from 733 subjects from July 2020 to April 2021 and verified with the data from 278 eyes from 139 subjects between May 2021 and June 2021. Each subject was measured with a Pentacam, IOLMaster 700, OPD-Scan III, and ultrasound biomicroscope. The prediction formulas were established with multiple linear regression, and intergroup correlation coefficients (ICCs) and Bland-Altman tests were used to assess the agreement between the predicted and actual STS (actual STS was measured by UBM). RESULTS: The explanatory variables relevant to the horizontal STS (STSH) were the Pentacam white-to-white diameter (WTWP; standardized partial regression coefficient [ß] = 0.330; p < 0.001), the flat K value (ß = -0.211; p < 0.001), and the anterior corneal diameter (ACD) (ß = 0.178; p < 0.001). The corresponding multiple regression equation was : STSH (mm) = 8.061 + 0.510 × WTWP - 0.090 × Flat K value + 0.430 × ACD. The explanatory variables relevant to the vertical STS (STSV) were the WTWP (ß = 0.435; p < 0.001), the steep K value (ß = -0.271; p < 0.001), and the ACD (ß = 0.187; p < 0.001). The corresponding multiple regression equation was : STSV (mm) = 8.540 + 0.492 × WTWP - 0.075 × Steep K value + 0.329 × ACD. The bias of the predicted to the actual STSH was - 0.021, with 95% limits of agreement (95% LoA) from - 0.499 to 0.457. The bias of the predicted to the actual STSV was 0.057, with 95% LoA from - 0.462 to 0.575. The ICC was 0.883 between the predicted and actual STSH and 0.859 between the predicted and actual STSV. CONCLUSIONS: The Pentacam-measured WTW, the K value and the ACD are important for predicting the STS diameter. The prediction model has good accuracy and reliability. TRIAL REGISTRATION: Not applicable.

Short-term changes in and preoperative factors affecting vaulting after posterior chamber phakic Implantable Collamer Lens implantation.

BACKGROUND: To describe the very early vault changes in the first month after Implantable Collamer Lens (ICL) implantation and to evaluate the effect of preoperative biometric factors on vault. METHODS: Eighty-three eyes from eighty-three subjects with complete data who met follow-up requirements were recruited in this retrospective study between May 2019 and March 2020. We quantitatively assessed the postoperative vault at 2 h, 1 day, 1 week, and 1 month following implantation. Associations between the postoperative vault and age, ICL size, spherical equivalent (SE), axial length (AL), central corneal thickness (CCT), flat keratometry (K), steep K, mean K, anterior chamber depth (ACD), crystalline lens thickness (LT), white-to-white (WTW) diameter obtained by three devices, horizontal and vertical sulcus-to-sulcus (STS) diameter, bright and dark pupil sizes (BPS and DPS) and DPS-BPS were investigated using Spearman's correlation analysis and stepwise multiple regression analysis. RESULTS: The mean vault values at 2 h, 1 day, 1 week, and 1 month after ICL implantation were 672.05 ± 30.72, 389.15 ± 28.33, 517.23 ± 30.76 and 530.12 ± 30.22 µm, respectively. Significant differences were found in the vault values at 2 h, 1 day and 1 week after the operation. The ICL size (ß = 0.942; p < 0.001), followed by horizontal STS (ß = -0.517; p < 0.001), crystalline LT (ß = -0.376; p < 0.001) and vertical STS (ß = -0.257; p = 0.017), significantly influenced the vault at 1 month after the operation. The multiple regression equation was expressed as follows: central vault (µm) = -1369.05 + 657.121 × ICL size- 287.408 × horizontal STS - 432.497 × crystalline LT - 137.33 × vertical STS (adjusted R2 = 0.643). CONCLUSIONS: After ICL implantation, the vault decreased and then increased, but it did not return to the vault value 2 h after surgery. The ICL size, horizontal and vertical STS and crystalline LT are key factors for predicting postoperative vaulting.

Analysis of macular microvasculature and thickness after ICL implantation in patients with myopia using optical coherence tomography.

AIM: To investigate changes in macular vessels and thickness in myopic eyes after intraocular collamer lens (ICL) implantation using quantitative optical coherence tomography angiography (OCTA). METHODS: This retrospective included 73 myopic eyes of 73 patients (average age, 27.53±6.16y) who underwent ICL implantation (28 eyes were Toric ICL). Axial length (AL), uncorrected visual acuity (UCVA), refractive dioptre (RD), intraocular pressure (IOP) and OCTA were measured and compared with before and 1wk, 1, and 3mo after surgery. OCTA was used to image vessel density (VD) and skeleton density (SD) in both the superficial (SCP) and deep capillary plexus (DCP). Central retinal thickness (CRT) and ganglion cell-inner plexiform layer thickness (GCT) were also measured. Changes between pre- and postoperative measurements were analysed by repeated measures analysis of variance. RESULTS: Compared with preoperative data, postoperative data on UCVA revealed significant improvements in all patients (P<0.05). However, there was no significant difference in IOP. After the operation, CRT and GCT exhibited significant changes (P<0.05). Among these measures, CRT was significantly higher at one and three months postoperative (all P<0.01). GCT was significantly higher at 1wk, 1, and 3mo postoperative (all P<0.01). Changes in VD and SD were nonsignificant in both the SCP and DCP. There was no difference in postoperative changes between the ICL and Toric ICL groups. CONCLUSION: ICL and Toric ICL implantation both have good efficacy and safety for myopic eyes, but macular area changes that occur after surgery need attention.