DeepLensNet：深度学习自动诊断和白内障类型及严重程度的定量分类。
DeepLensNet: Deep Learning Automated Diagnosis and Quantitative Classification of Cataract Type and Severity.
机构信息
Division of Epidemiology and Clinical Applications, National Eye Institute, National Institutes of Health, Bethesda, Maryland.
National Center for Biotechnology Information, National Library of Medicine, National Institutes of Health, Bethesda, Maryland.
出版信息
Ophthalmology. 2022 May;129(5):571-584. doi: 10.1016/j.ophtha.2021.12.017. Epub 2022 Jan 3.
PURPOSE
To develop deep learning models to perform automated diagnosis and quantitative classification of age-related cataract from anterior segment photographs.
DESIGN
DeepLensNet was trained by applying deep learning models to the Age-Related Eye Disease Study (AREDS) dataset.
PARTICIPANTS
A total of 18 999 photographs (6333 triplets) from longitudinal follow-up of 1137 eyes (576 AREDS participants).
METHODS
Deep learning models were trained to detect and quantify nuclear sclerosis (NS; scale 0.9-7.1) from 45-degree slit-lamp photographs and cortical lens opacity (CLO; scale 0%-100%) and posterior subcapsular cataract (PSC; scale 0%-100%) from retroillumination photographs. DeepLensNet performance was compared with that of 14 ophthalmologists and 24 medical students.
MAIN OUTCOME MEASURES
Mean squared error (MSE).
RESULTS
On the full test set, mean MSE for DeepLensNet was 0.23 (standard deviation [SD], 0.01) for NS, 13.1 (SD, 1.6) for CLO, and 16.6 (SD, 2.4) for PSC. On a subset of the test set (substantially enriched for positive cases of CLO and PSC), for NS, mean MSE for DeepLensNet was 0.23 (SD, 0.02), compared with 0.98 (SD, 0.24; P = 0.000001) for the ophthalmologists and 1.24 (SD, 0.34; P = 0.000005) for the medical students. For CLO, mean MSE was 53.5 (SD, 14.8), compared with 134.9 (SD, 89.9; P = 0.003) for the ophthalmologists and 433.6 (SD, 962.1; P = 0.0007) for the medical students. For PSC, mean MSE was 171.9 (SD, 38.9), compared with 176.8 (SD, 98.0; P = 0.67) for the ophthalmologists and 398.2 (SD, 645.4; P = 0.18) for the medical students. In external validation on the Singapore Malay Eye Study (sampled to reflect the cataract severity distribution in AREDS), the MSE for DeepSeeNet was 1.27 for NS and 25.5 for PSC.
CONCLUSIONS
DeepLensNet performed automated and quantitative classification of cataract severity for all 3 types of age-related cataract. For the 2 most common types (NS and CLO), the accuracy was significantly superior to that of ophthalmologists; for the least common type (PSC), it was similar. DeepLensNet may have wide potential applications in both clinical and research domains. In the future, such approaches may increase the accessibility of cataract assessment globally. The code and models are available at https://github.com/ncbi/deeplensnet.
目的
开发深度学习模型，以便从眼前段照片中自动诊断和定量分类年龄相关性白内障。
设计
通过将深度学习模型应用于年龄相关性眼病研究 (AREDS) 数据集来训练 DeepLensNet。
参与者
共有来自 1137 只眼（576 名 AREDS 参与者）的纵向随访的 18999 张（6333 对）45 度裂隙灯照片和 retroillumination 照片。
方法
训练深度学习模型以从 45 度裂隙灯照片中检测和量化核硬化（NS；0.9-7.1 级），并从 retroillumination 照片中量化皮质晶状体混浊（CLO；0%-100%）和后囊下白内障（PSC；0%-100%）。DeepLensNet 的性能与 14 名眼科医生和 24 名医学生进行了比较。
主要观察指标
均方误差 (MSE)。
结果
在整个测试集中，DeepLensNet 的平均 MSE 为 NS 为 0.23（标准差 [SD]，0.01），CLO 为 13.1（SD，1.6），PSC 为 16.6（SD，2.4）。在测试集的一个子集（明显富含 CLO 和 PSC 的阳性病例）中，对于 NS，DeepLensNet 的平均 MSE 为 0.23（SD，0.02），而眼科医生的平均 MSE 为 0.98（SD，0.24；P = 0.000001），医学生的平均 MSE 为 1.24（SD，0.34；P = 0.000005）。对于 CLO，平均 MSE 为 53.5（SD，14.8），而眼科医生的平均 MSE 为 134.9（SD，89.9；P = 0.003），医学生的平均 MSE 为 433.6（SD，962.1；P = 0.0007）。对于 PSC，平均 MSE 为 171.9（SD，38.9），而眼科医生的平均 MSE 为 176.8（SD，98.0；P = 0.67），医学生的平均 MSE 为 398.2（SD，645.4；P = 0.18）。在对新加坡马来眼研究（抽样以反映 AREDS 中的白内障严重程度分布）的外部验证中，DeepSeeNet 用于 NS 的 MSE 为 1.27，用于 PSC 的 MSE 为 25.5。
结论
DeepLensNet 可对所有 3 种类型的年龄相关性白内障进行自动和定量分类。对于最常见的两种类型（NS 和 CLO），准确性明显优于眼科医生；对于最不常见的一种类型（PSC），准确性相似。DeepLensNet 在临床和研究领域都具有广泛的潜在应用。未来，此类方法可能会增加全球白内障评估的可及性。代码和模型可在 https://github.com/ncbi/deeplensnet 上获得。
Zotero 插件