Possible Risk Factors and Clinical Outcomes of Black Areas in Small-Incision Lenticule Extraction

Prediction of Early Visual Outcome of Small-Incision Lenticule Extraction (SMILE) Based on Deep Learning

INTRODUCTION

Deep learning (DL) has been widely used to estimate clinical images. The objective of this project was to create DL models to predict the early postoperative visual acuity after small-incision lenticule extraction (SMILE) surgery.

METHODS

We enrolled three independent patient cohorts (a retrospective cohort and two prospective SMILE cohorts) who underwent the SMILE refractive correction procedure at two different refractive surgery centers from July to September 2022. The medical records and surgical videos were collected for further analysis. Based on the uncorrected visual acuity (UCVA) at 24 h postsurgery, the eyes were divided into two groups: those showing good recovery and those showing poor recovery. We then trained a DL model (Resnet50) to predict eyes with early postoperative visual acuity of patients in the retrospective cohort who had undergone SMILE surgery from surgical videos and subsequently validated the model's performance in the two prospective cohorts. Finally, Gradient-weighted Class Activation Mapping (Grad-CAM) was performed for interpretation of the model.

RESULTS

Among the 318 eyes (159 patients) enrolled in the study, 10,176 good quality femtosecond laser scanning images were obtained from the surgical videos. We observed that the developed DL model achieved a high accuracy of 96% for image prediction. The area under the curve (AUC) value of the DL model in the retrospective cohort was 0.962 and 0.998 in the training and validation datasets, respectively. The AUC values in two prospective cohorts were 0.959 and 0.936. At the video level, the trained machine learning (ML) model (XGBoost) also accurately distinguished patients with good or poor recovery. The AUC value of the ML model was 0.998 and 0.889 in the retrospective cohort (training and test datasets, respectively) and 1.000 and 0.984 in the two prospective cohorts. We also trained a DL model which can accurately distinguish suction loss (100%), black spots (85%), and opaque bubble layer (96%). The Grad-CAM heatmap indicated that our models can recognize the area of scanning and precisely identify intraoperative complications.

CONCLUSIONS

Our findings suggest that artificial intelligence (DL and ML model) can accurately predict the early postoperative visual acuity and intraoperative complications after SMILE surgery just using surgical videos or images, which may display a great importance for artificial intelligence in application of refractive surgeries.

Laser-assisted subepithelial keratomileusis substituted for an aborted small-incision lenticule extraction due to large black area formation

A 25-year-old woman with moderate myopia presented for refractive surgery. Bilateral femtosecond laser-assisted small-incision lenticule extraction (SMILE) was scheduled and her right eye was completed smoothly. However, during the lenticule cutting of her left eye, a large inferior black area was noted. The operation was abandoned after an immediate optical coherence tomography examination, which revealed the corneal epithelium defect with no laser scanning line at the corresponding site. The secondary surgery was assigned to laser-assisted subepithelial keratomileusis (LASEK) after 1 week. The uncorrected distance visual acuity of her left eye recovered to 20/25 on the 12th day and to 20/20 at 3-month follow-up, with ideal corneal topography profiles. Corneal epithelium defect induced by accidental alcohol contact during disinfection was suspected to cause the black area. The management of black area had to be determined according to the location and size. LASEK was a rational substitution for the aborted SMILE.

Development of low laser energy levels in small-incision lenticule extraction: clinical results, black area, and ultrastructural evaluation

PURPOSE

To evaluate the outcomes of 4 low laser energy levels after small-incision lenticule extraction (SMILE) surgery.

SETTING

Zhongshan Ophthalmic Center, Guangzhou, China.

DESIGN

Prospective randomized clinical trial.

METHODS

This study evaluated consecutive patients who had SMILE to correct myopia or myopia with astigmatism. Eyes were placed into groups based on the laser energy used during surgery (ie, 105 nJ, 110 nJ, 115 nJ, or 120 nJ). All patients had a thorough ophthalmic examination preoperative and at 4 timepoints over 3 months postoperatively. Black areas and surface regularity of the extracted lenticules were observed and evaluated qualitatively and quantitatively.

RESULTS

The study comprised 124 eyes of 62 patients (40 women, 22 men), with 31 eyes in each laser energy group. The incidence of black areas was 45.16% (14 of 31 eyes), 12.90% (4 of 31 eyes), 16.13% (5 of 31 eyes), and 12.90% (4 of 31 eyes) for 105 nJ, 110 nJ, 115 nJ, and 120 nJ, respectively. The mean time for lenticule creation was the longest in the 105 nJ group (P = .015). The greatest increase in corneal thickness postoperatively occurred with 105 nJ (P < .05). Regression was highest in the 105 nJ group at 3 months (P < .01). However, corneal horizontal coma (C8) was lowest in the 105 nJ group at 1 week (P = .032). The lenticular surface in the 110 nJ group was the smoothest (P = .011). All contrast sensitivity values varied with time and recovered to preoperative levels by 1 week or 1 month. In all eyes, the uncorrected distance visual acuity and corrected distance visual acuity were good, with no statistically significant differences between the 4 energy groups.

CONCLUSIONS

The 105 nJ group, in which the lowest energy was used, had the highest risk for black areas, serious postoperative corneal edema, and a significant healing response.