AIM: To evaluate the accuracy of three commonly used biometric formulae across different axial lengths(ALs) at one United States Veterans Affairs teaching hospital.METHODS: A retrospective chart review was conducted f...AIM: To evaluate the accuracy of three commonly used biometric formulae across different axial lengths(ALs) at one United States Veterans Affairs teaching hospital.METHODS: A retrospective chart review was conducted from November 2013 to May 2018. One eye of each patient who underwent cataract surgery with a monofocal intraocular lens(IOL) was included. The range of postoperative follow-up period was from 3 wk to 4 mo. The Holladay 2, Barrett Universal II, and Hill-Radial Basis Function(Hill-RBF) formulae were used to predict the postoperative refraction for all cataract surgeries. For each formula, we calculated the prediction errors [including mean absolute prediction error(MAE)] and the percentage of eyes within ±0.25 diopter(D) and ±0.5 D of predicted refraction. We performed subgroup analyses for short(AL<22.0 mm), medium(AL 22.0-25.0 mm), and long eyes(AL>25.0 mm).RESULTS: A total of 1131 patients were screened, and 909 met the inclusion criteria. Resident ophthalmologists were the primary surgeons in 710(78.1%) cases. We found no statistically significant difference in predictive accuracy among the three formulae over the entire AL range or in the short, medium, and long eye subgroups. Across the entire AL range, the Hill-RBF formula resulted in the lowest MAE(0.384 D) and the highest percentage of eyes with postoperative refraction within ±0.25 D(42.7%) and ±0.5 D(75.5%) of predicted. All three formulae had the highest MAEs(>0.5 D) and lowest percentage within ±0.5 D of predicted refraction(<55%) in short eyes.CONCLUSION: In cataract surgery patients at our teaching hospital, three commonly used biometric formulae demonstrate similar refractive accuracy across all ALs. Short eyes pose the greatest challenge to predicting postoperative refractive error.展开更多
AIM: To compare the Barrett True-K formula with other formulas integrated in Lenstar 900 to predict intraocular lens(IOL) power after small-incision lenticule extraction(SMILE).METHODS: A theoretical prospective study...AIM: To compare the Barrett True-K formula with other formulas integrated in Lenstar 900 to predict intraocular lens(IOL) power after small-incision lenticule extraction(SMILE).METHODS: A theoretical prospective study was performed to predict the ratio of equivalent IOL power before and after SMILE using the SRK/T(Sanders, Retzlaff, Kraff/theoretical), Holladay 1, Haigis, and Barrett True-K formulas and compare the stability of their predictions. The study included 54 eyes(54 cases) with a manifest refraction spherical equivalent(MRSE) of-4.99±1.45 D. They were divided into two groups: 27 eyes with axial length of 24-26 mm in Group A, and 27 eyes with axial length >26 mm in Group B. All subjects enrolled in this study were examined with the Lenstar 900 before and 6 mo after SMILE including measurements of axial length, corneal curvature, and anterior chamber depth(ACD). RESULTS: The prediction of equivalent IOL power of the two groups was more stable for the Barrett True-K formula, especially in long axial length eyes(Group B). The respective percentages for the SRK/T, Holladay 1, Haigis, and Barrett True-K formulas were 7.4%, 7.4%, 85.19%, and 88.89% for a margin of error within 0.5 D;25.92%, 51.84%, 100%, and 100% for a margin of error within 1.0 D in Group A;33.33%, 40.74%, 44.44%, and 81.48% for a margin of error within 0.5 D;and 44.44%, 59.26%, 66.66%, and 92.59% for a margin of error within 1.0 D in Group B. The respective percentages for Barrett True-K formulas were 100% for a margin of error within 2.0 D in Group B.CONCLUSION: Theoretically, the Barrett True-K formula provides more stable predictions than other formulas for cataract eyes after SMILE.展开更多
Background:To examine the effectiveness of the use of machine learning for adapting an intraocular lens(IOL)power calculation for a patient group.Methods:In this retrospective study,the clinical records of 1,611 eyes ...Background:To examine the effectiveness of the use of machine learning for adapting an intraocular lens(IOL)power calculation for a patient group.Methods:In this retrospective study,the clinical records of 1,611 eyes of 1,169 Japanese patients who received a single model of monofocal IOL(SN60WF,Alcon)at Miyata Eye Hospital were reviewed and analyzed.Using biometric metrics and postoperative refractions of 1211 eyes of 769 patients,constants of the SRK/T and Haigis formulas were optimized.The SRK/T formula was adapted using a support vector regressor.Prediction errors in the use of adapted formulas as well as the SRK/T,Haigis,Hill-RBF and Barrett Universal II formulas were evaluated with data from 395 eyes of 395 distinct patients.Mean prediction errors,median absolute errors,and percentages of eyes within±0.25 D,±0.50 D,and±1.00 D,and over+0.50 D of errors were compared among formulas.Results:The mean prediction errors in the use of the SRT/K and adapted formulas were smaller than the use of other formulas(P<0.001).In the absolute errors,the Hill-RBF and adapted methods were better than others.The performance of the Barrett Universal II was not better than the others for the patient group.There were the least eyes with hyperopic refractive errors(16.5%)in the use of the adapted formula.Conclusions:Adapting IOL power calculations using machine learning technology with data from a particular patient group was effective and promising.展开更多
Background:To examine the effectiveness of the use of machine learning for adapting an intraocular lens(IOL)power calculation for a patient group.Methods:In this retrospective study,the clinical records of 1,611 eyes ...Background:To examine the effectiveness of the use of machine learning for adapting an intraocular lens(IOL)power calculation for a patient group.Methods:In this retrospective study,the clinical records of 1,611 eyes of 1,169 Japanese patients who received a single model of monofocal IOL(SN60WF,Alcon)at Miyata Eye Hospital were reviewed and analyzed.Using biometric metrics and postoperative refractions of 1211 eyes of 769 patients,constants of the SRK/T and Haigis formulas were optimized.The SRK/T formula was adapted using a support vector regressor.Prediction errors in the use of adapted formulas as well as the SRK/T,Haigis,Hill-RBF and Barrett Universal II formulas were evaluated with data from 395 eyes of 395 distinct patients.Mean prediction errors,median absolute errors,and percentages of eyes within±0.25 D,±0.50 D,and±1.00 D,and over+0.50 D of errors were compared among formulas.Results:The mean prediction errors in the use of the SRT/K and adapted formulas were smaller than the use of other formulas(P<0.001).In the absolute errors,the Hill-RBF and adapted methods were better than others.The performance of the Barrett Universal II was not better than the others for the patient group.There were the least eyes with hyperopic refractive errors(16.5%)in the use of the adapted formula.Conclusions:Adapting IOL power calculations using machine learning technology with data from a particular patient group was effective and promising.展开更多
文摘AIM: To evaluate the accuracy of three commonly used biometric formulae across different axial lengths(ALs) at one United States Veterans Affairs teaching hospital.METHODS: A retrospective chart review was conducted from November 2013 to May 2018. One eye of each patient who underwent cataract surgery with a monofocal intraocular lens(IOL) was included. The range of postoperative follow-up period was from 3 wk to 4 mo. The Holladay 2, Barrett Universal II, and Hill-Radial Basis Function(Hill-RBF) formulae were used to predict the postoperative refraction for all cataract surgeries. For each formula, we calculated the prediction errors [including mean absolute prediction error(MAE)] and the percentage of eyes within ±0.25 diopter(D) and ±0.5 D of predicted refraction. We performed subgroup analyses for short(AL<22.0 mm), medium(AL 22.0-25.0 mm), and long eyes(AL>25.0 mm).RESULTS: A total of 1131 patients were screened, and 909 met the inclusion criteria. Resident ophthalmologists were the primary surgeons in 710(78.1%) cases. We found no statistically significant difference in predictive accuracy among the three formulae over the entire AL range or in the short, medium, and long eye subgroups. Across the entire AL range, the Hill-RBF formula resulted in the lowest MAE(0.384 D) and the highest percentage of eyes with postoperative refraction within ±0.25 D(42.7%) and ±0.5 D(75.5%) of predicted. All three formulae had the highest MAEs(>0.5 D) and lowest percentage within ±0.5 D of predicted refraction(<55%) in short eyes.CONCLUSION: In cataract surgery patients at our teaching hospital, three commonly used biometric formulae demonstrate similar refractive accuracy across all ALs. Short eyes pose the greatest challenge to predicting postoperative refractive error.
文摘AIM: To compare the Barrett True-K formula with other formulas integrated in Lenstar 900 to predict intraocular lens(IOL) power after small-incision lenticule extraction(SMILE).METHODS: A theoretical prospective study was performed to predict the ratio of equivalent IOL power before and after SMILE using the SRK/T(Sanders, Retzlaff, Kraff/theoretical), Holladay 1, Haigis, and Barrett True-K formulas and compare the stability of their predictions. The study included 54 eyes(54 cases) with a manifest refraction spherical equivalent(MRSE) of-4.99±1.45 D. They were divided into two groups: 27 eyes with axial length of 24-26 mm in Group A, and 27 eyes with axial length >26 mm in Group B. All subjects enrolled in this study were examined with the Lenstar 900 before and 6 mo after SMILE including measurements of axial length, corneal curvature, and anterior chamber depth(ACD). RESULTS: The prediction of equivalent IOL power of the two groups was more stable for the Barrett True-K formula, especially in long axial length eyes(Group B). The respective percentages for the SRK/T, Holladay 1, Haigis, and Barrett True-K formulas were 7.4%, 7.4%, 85.19%, and 88.89% for a margin of error within 0.5 D;25.92%, 51.84%, 100%, and 100% for a margin of error within 1.0 D in Group A;33.33%, 40.74%, 44.44%, and 81.48% for a margin of error within 0.5 D;and 44.44%, 59.26%, 66.66%, and 92.59% for a margin of error within 1.0 D in Group B. The respective percentages for Barrett True-K formulas were 100% for a margin of error within 2.0 D in Group B.CONCLUSION: Theoretically, the Barrett True-K formula provides more stable predictions than other formulas for cataract eyes after SMILE.
文摘Background:To examine the effectiveness of the use of machine learning for adapting an intraocular lens(IOL)power calculation for a patient group.Methods:In this retrospective study,the clinical records of 1,611 eyes of 1,169 Japanese patients who received a single model of monofocal IOL(SN60WF,Alcon)at Miyata Eye Hospital were reviewed and analyzed.Using biometric metrics and postoperative refractions of 1211 eyes of 769 patients,constants of the SRK/T and Haigis formulas were optimized.The SRK/T formula was adapted using a support vector regressor.Prediction errors in the use of adapted formulas as well as the SRK/T,Haigis,Hill-RBF and Barrett Universal II formulas were evaluated with data from 395 eyes of 395 distinct patients.Mean prediction errors,median absolute errors,and percentages of eyes within±0.25 D,±0.50 D,and±1.00 D,and over+0.50 D of errors were compared among formulas.Results:The mean prediction errors in the use of the SRT/K and adapted formulas were smaller than the use of other formulas(P<0.001).In the absolute errors,the Hill-RBF and adapted methods were better than others.The performance of the Barrett Universal II was not better than the others for the patient group.There were the least eyes with hyperopic refractive errors(16.5%)in the use of the adapted formula.Conclusions:Adapting IOL power calculations using machine learning technology with data from a particular patient group was effective and promising.
文摘Background:To examine the effectiveness of the use of machine learning for adapting an intraocular lens(IOL)power calculation for a patient group.Methods:In this retrospective study,the clinical records of 1,611 eyes of 1,169 Japanese patients who received a single model of monofocal IOL(SN60WF,Alcon)at Miyata Eye Hospital were reviewed and analyzed.Using biometric metrics and postoperative refractions of 1211 eyes of 769 patients,constants of the SRK/T and Haigis formulas were optimized.The SRK/T formula was adapted using a support vector regressor.Prediction errors in the use of adapted formulas as well as the SRK/T,Haigis,Hill-RBF and Barrett Universal II formulas were evaluated with data from 395 eyes of 395 distinct patients.Mean prediction errors,median absolute errors,and percentages of eyes within±0.25 D,±0.50 D,and±1.00 D,and over+0.50 D of errors were compared among formulas.Results:The mean prediction errors in the use of the SRT/K and adapted formulas were smaller than the use of other formulas(P<0.001).In the absolute errors,the Hill-RBF and adapted methods were better than others.The performance of the Barrett Universal II was not better than the others for the patient group.There were the least eyes with hyperopic refractive errors(16.5%)in the use of the adapted formula.Conclusions:Adapting IOL power calculations using machine learning technology with data from a particular patient group was effective and promising.
