Ophthalmology. Retina · 2022
Characterizing Early Residual Fluid in Neovascular Age-Related Macular Degeneration using Machine Learning in Routine Clinical Practice.
СтатьяНазвание и аннотация: Перевод пока не подготовлен
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Wu Anna K, Perkins Scott W, Singh Rishi P
Аннотация
PURPOSE: Neovascular age-related macular degeneration (nAMD) often requires intensive therapy with anti-VEGF injections. In prior post hoc studies, early residual fluid (ERF) after the loading phase was associated with poorer treatment outcomes. This retrospective study examined the impact of ERF on vision using machine learning (ML) methods in routine clinical practice.
DESIGN: Retrospective cohort.
PARTICIPANTS: This study included treatment-naïve patients with nAMD who were initiated on anti-VEGF between 2012 and 2018, with at least 1 year of follow-up.
METHODS: Overall, 286 patients with nAMD were included. An ML algorithm quantified intraretinal fluid (IRF), subretinal fluid (SRF), and total retinal fluid from OCTs. The ERF group included those with fluid at week 12 and was further stratified by fluid subtype. Paired t tests and analysis of variance compared best visual acuity (BVA) and fluid among subgroups, and a quartile analysis correlated fluid volumes to week 52 BVA. The risk of ERF was predicted from baseline factors using 3 ML methods: Ridge logistic regression, k nearest neighbors classification, and support vector classification.
MAIN OUTCOME MEASURES: Mean change in BVA from baseline to week 52 according to week 12 fluid status.
RESULTS: At week 12, 58.4% of patients had ERF. The breakdown of those in the ERF group included SRF-only (45.5%), IRF-only (21.6%), and IRF and SRF (32.9%). The ERF and ERF-free groups had similar BVA gains from baseline to week 52 (+5.7 ± 15.4 vs. +4.9 ± 18; P = 0.69). Examining specific ERF subgroups revealed no significant differences among the IRF-only (+4.6 ± 16.4), SRF-only (+5.6 ± 12.5), and IRF and SRF (+6.6 ± 18.5, P = 0.93) groups. Quartile analysis of week 12 fluid revealed no predictive pattern for BVA gains. Three ML methods were developed to predict those at risk for ERF achieved equivalent performance, with F1 score of 0.73 to 0.76.
CONCLUSIONS: These results diverge from prior post hoc studies, in that there was no significant difference in long-term BVA gains between ERF and ERF-free cohorts, as well as between the week 12 fluid subgroups.
DESIGN: Retrospective cohort.
PARTICIPANTS: This study included treatment-naïve patients with nAMD who were initiated on anti-VEGF between 2012 and 2018, with at least 1 year of follow-up.
METHODS: Overall, 286 patients with nAMD were included. An ML algorithm quantified intraretinal fluid (IRF), subretinal fluid (SRF), and total retinal fluid from OCTs. The ERF group included those with fluid at week 12 and was further stratified by fluid subtype. Paired t tests and analysis of variance compared best visual acuity (BVA) and fluid among subgroups, and a quartile analysis correlated fluid volumes to week 52 BVA. The risk of ERF was predicted from baseline factors using 3 ML methods: Ridge logistic regression, k nearest neighbors classification, and support vector classification.
MAIN OUTCOME MEASURES: Mean change in BVA from baseline to week 52 according to week 12 fluid status.
RESULTS: At week 12, 58.4% of patients had ERF. The breakdown of those in the ERF group included SRF-only (45.5%), IRF-only (21.6%), and IRF and SRF (32.9%). The ERF and ERF-free groups had similar BVA gains from baseline to week 52 (+5.7 ± 15.4 vs. +4.9 ± 18; P = 0.69). Examining specific ERF subgroups revealed no significant differences among the IRF-only (+4.6 ± 16.4), SRF-only (+5.6 ± 12.5), and IRF and SRF (+6.6 ± 18.5, P = 0.93) groups. Quartile analysis of week 12 fluid revealed no predictive pattern for BVA gains. Three ML methods were developed to predict those at risk for ERF achieved equivalent performance, with F1 score of 0.73 to 0.76.
CONCLUSIONS: These results diverge from prior post hoc studies, in that there was no significant difference in long-term BVA gains between ERF and ERF-free cohorts, as well as between the week 12 fluid subgroups.