Jérémy Abadou, Simon Dahan, Juliette Knoeri, Loic Leveziel, Nacim Bouheraoua, Vincent M Borderie
{"title":"人工智能与传统角膜塑形镜验配方法的比较","authors":"Jérémy Abadou, Simon Dahan, Juliette Knoeri, Loic Leveziel, Nacim Bouheraoua, Vincent M Borderie","doi":"10.1016/j.clae.2024.102321","DOIUrl":null,"url":null,"abstract":"<p><strong>Background: </strong>To compare the efficiency of three artificial intelligence (AI) frameworks (Standard Machine Learning (ML), Multi-Layer Perceptron (MLP) and Convolution Neural Networks (CNN)) with a reference method (Mean radius of curvature (K)) to predict the posterior radius of curvature of the best-fitted rigid contact lens (RCBFL) in keratoconus eyes.</p><p><strong>Methods: </strong>This retrospective study included 197 keratoconus eyes of 135 patients fitted with Rose K2® (Menicon®, Nagoya, Japan) rigid contact lenses with one or more topographies available (MS39®, CSO®, Ferrara, Italy) between January 2020 and September 2022. Two types of topographic data (indices and reconstructed maps from raw data) were used for AI analysis. Three distinct approaches were utilized for leveraging AI: Standard ML methods and MLPs based on topographic indices and CNNs based on topographic maps (i.e., corneal thickness, sagittal, and tangential maps). Seventeen AI framework's accuracies were compared with the r<sup>2</sup> determination coefficient of linear regression between predicted and best-fitted radii. Framework accuracies were compared with the Fisher z-transformation of Pearson correlation coefficients.</p><p><strong>Results: </strong>In multiple linear regression, only three topographic indices (i.e., 3- & 5-mm mean K and Kmax) were significantly associated with RCBFL (p ≤ 0.0001). Compared with the reference method (mean-K; r<sup>2</sup> = 0.36), a significantly better RCBFL prediction was achieved with Random Forest using the three topographic indices, MLP using all indices, ResNet18 CNN using anterior topographic maps and CNNs using combined parameters (0.69 ≤ r<sup>2</sup> ≤ 0.80; p < 0.05). The best accuracy was obtained with the EfficientNetB0 CNN trained with three maps (r<sup>2</sup> = 0.80).</p><p><strong>Conclusions: </strong>Artificial intelligence methods, particularly CNNs, with corneal topography data of MS39® topographer, have demonstrated superiority over conventional approaches in predicting the posterior curvature radius of Rose K2® rigid contact lenses in patients with keratoconus.</p>","PeriodicalId":49087,"journal":{"name":"Contact Lens & Anterior Eye","volume":" ","pages":"102321"},"PeriodicalIF":4.1000,"publicationDate":"2024-11-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Artificial intelligence versus conventional methods for RGP lens fitting in keratoconus.\",\"authors\":\"Jérémy Abadou, Simon Dahan, Juliette Knoeri, Loic Leveziel, Nacim Bouheraoua, Vincent M Borderie\",\"doi\":\"10.1016/j.clae.2024.102321\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<p><strong>Background: </strong>To compare the efficiency of three artificial intelligence (AI) frameworks (Standard Machine Learning (ML), Multi-Layer Perceptron (MLP) and Convolution Neural Networks (CNN)) with a reference method (Mean radius of curvature (K)) to predict the posterior radius of curvature of the best-fitted rigid contact lens (RCBFL) in keratoconus eyes.</p><p><strong>Methods: </strong>This retrospective study included 197 keratoconus eyes of 135 patients fitted with Rose K2® (Menicon®, Nagoya, Japan) rigid contact lenses with one or more topographies available (MS39®, CSO®, Ferrara, Italy) between January 2020 and September 2022. Two types of topographic data (indices and reconstructed maps from raw data) were used for AI analysis. Three distinct approaches were utilized for leveraging AI: Standard ML methods and MLPs based on topographic indices and CNNs based on topographic maps (i.e., corneal thickness, sagittal, and tangential maps). Seventeen AI framework's accuracies were compared with the r<sup>2</sup> determination coefficient of linear regression between predicted and best-fitted radii. Framework accuracies were compared with the Fisher z-transformation of Pearson correlation coefficients.</p><p><strong>Results: </strong>In multiple linear regression, only three topographic indices (i.e., 3- & 5-mm mean K and Kmax) were significantly associated with RCBFL (p ≤ 0.0001). Compared with the reference method (mean-K; r<sup>2</sup> = 0.36), a significantly better RCBFL prediction was achieved with Random Forest using the three topographic indices, MLP using all indices, ResNet18 CNN using anterior topographic maps and CNNs using combined parameters (0.69 ≤ r<sup>2</sup> ≤ 0.80; p < 0.05). The best accuracy was obtained with the EfficientNetB0 CNN trained with three maps (r<sup>2</sup> = 0.80).</p><p><strong>Conclusions: </strong>Artificial intelligence methods, particularly CNNs, with corneal topography data of MS39® topographer, have demonstrated superiority over conventional approaches in predicting the posterior curvature radius of Rose K2® rigid contact lenses in patients with keratoconus.</p>\",\"PeriodicalId\":49087,\"journal\":{\"name\":\"Contact Lens & Anterior Eye\",\"volume\":\" \",\"pages\":\"102321\"},\"PeriodicalIF\":4.1000,\"publicationDate\":\"2024-11-04\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Contact Lens & Anterior Eye\",\"FirstCategoryId\":\"3\",\"ListUrlMain\":\"https://doi.org/10.1016/j.clae.2024.102321\",\"RegionNum\":3,\"RegionCategory\":\"医学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"OPHTHALMOLOGY\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Contact Lens & Anterior Eye","FirstCategoryId":"3","ListUrlMain":"https://doi.org/10.1016/j.clae.2024.102321","RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"OPHTHALMOLOGY","Score":null,"Total":0}
Artificial intelligence versus conventional methods for RGP lens fitting in keratoconus.
Background: To compare the efficiency of three artificial intelligence (AI) frameworks (Standard Machine Learning (ML), Multi-Layer Perceptron (MLP) and Convolution Neural Networks (CNN)) with a reference method (Mean radius of curvature (K)) to predict the posterior radius of curvature of the best-fitted rigid contact lens (RCBFL) in keratoconus eyes.
Methods: This retrospective study included 197 keratoconus eyes of 135 patients fitted with Rose K2® (Menicon®, Nagoya, Japan) rigid contact lenses with one or more topographies available (MS39®, CSO®, Ferrara, Italy) between January 2020 and September 2022. Two types of topographic data (indices and reconstructed maps from raw data) were used for AI analysis. Three distinct approaches were utilized for leveraging AI: Standard ML methods and MLPs based on topographic indices and CNNs based on topographic maps (i.e., corneal thickness, sagittal, and tangential maps). Seventeen AI framework's accuracies were compared with the r2 determination coefficient of linear regression between predicted and best-fitted radii. Framework accuracies were compared with the Fisher z-transformation of Pearson correlation coefficients.
Results: In multiple linear regression, only three topographic indices (i.e., 3- & 5-mm mean K and Kmax) were significantly associated with RCBFL (p ≤ 0.0001). Compared with the reference method (mean-K; r2 = 0.36), a significantly better RCBFL prediction was achieved with Random Forest using the three topographic indices, MLP using all indices, ResNet18 CNN using anterior topographic maps and CNNs using combined parameters (0.69 ≤ r2 ≤ 0.80; p < 0.05). The best accuracy was obtained with the EfficientNetB0 CNN trained with three maps (r2 = 0.80).
Conclusions: Artificial intelligence methods, particularly CNNs, with corneal topography data of MS39® topographer, have demonstrated superiority over conventional approaches in predicting the posterior curvature radius of Rose K2® rigid contact lenses in patients with keratoconus.
期刊介绍:
Contact Lens & Anterior Eye is a research-based journal covering all aspects of contact lens theory and practice, including original articles on invention and innovations, as well as the regular features of: Case Reports; Literary Reviews; Editorials; Instrumentation and Techniques and Dates of Professional Meetings.