Yoshida Miyo, Murakami Tomoaki, Ishihara Kenji, Mori Yuki, Tsujikawa Akitaka
Department of Ophthalmology and Visual Sciences, Kyoto University Graduate School of Medicine, Kyoto, Japan.
Ophthalmol Sci. 2025 Apr 24;5(5):100804. doi: 10.1016/j.xops.2025.100804. eCollection 2025 Sep-Oct.
To explore clinically significant diabetic retinal neurodegeneration in OCT images using explainable artificial intelligence (XAI) and subsequent evaluation by retinal specialists.
A single-center, retrospective, consecutive case series.
Three hundred ninety-seven eyes from 397 diabetic retinopathy patients for XAI-based screening and 244 fellow eyes for subjective human evaluation.
We acquired 30° horizontal OCT images centered on the fovea. An artificial intelligence (AI) model was developed to infer visual acuity (VA) reduction using fine-tuned RETFound-OCT. Attention maps highlighting regions contributing to VA inference were generated using layer-wise relevance propagation. Retinal specialists assessed OCT findings based on salient regions indicated by XAI. Two newly described findings, a needle-like appearance of the ganglion cell layer (GCL)/inner plexiform layer (IPL) ("ice-pick sign") and dot-like alterations in the outer nuclear layer (ONL) ("salt-and-pepper sign"), were evaluated alongside 2 established findings: EZ disruption and choroidal hypertransmission.
Identification of clinically significant OCT findings associated with diabetic retinal neurodegeneration.
The AI model effectively discriminated eyes with poor vision (decimal VA ≤0.5) from those with good vision (VA ≥1.0) (area under the receiver operating characteristic curve of 0.947). Explainable artificial intelligence-based attention maps highlighted salient regions in the GCL/IPL (65.2% or 70.0%), ONL (52.2% or 28.3%), EZ (39.1% or 21.7%), and choroid (26.1% or 5.00%) in eyes with poor or good vision, respectively. Subjective evaluation by retinal specialists revealed the frequencies of these 4 findings as follows: ice-pick sign (32.4%), EZ disruption (25.0%), salt-and-pepper sign (16.0%), and choroidal hypertransmission (13.5%). Eyes with decimal VA ≤0.9 had these findings more frequently than those with VA ≥1.0 ( < 0.001 for all comparisons). Salt-and-pepper sign and choroidal hypertransmission exhibited high specificity for identifying eyes with poor vision. Statistical analyses demonstrated more significant associations between EZ disruption, salt-and-pepper sign, and hypertransmission compared with their relationships with the ice-pick sign.
Artificial intelligence-assisted exploration of OCT findings identified 2 established lesions and 2 novel OCT biomarkers indicative of clinically significant diabetic retinal neurodegeneration.
The author(s) have no proprietary or commercial interest in any materials discussed in this article.
利用可解释人工智能（XAI）探索光学相干断层扫描（OCT）图像中具有临床意义的糖尿病视网膜神经变性，并随后由视网膜专家进行评估。
单中心、回顾性、连续病例系列研究。
397例糖尿病视网膜病变患者的397只眼睛用于基于XAI的筛查，244只对侧眼用于主观人工评估。
我们获取了以黄斑为中心的30°水平OCT图像。开发了一种人工智能（AI）模型，使用微调后的RETFound-OCT来推断视力（VA）下降情况。使用逐层相关性传播生成突出有助于VA推断区域的注意力图。视网膜专家根据XAI指示的显著区域评估OCT结果。评估了两个新描述的发现，即神经节细胞层（GCL）/内丛状层（IPL）的针状外观（“冰锥征”）和外核层（ONL）的点状改变（“椒盐征”），以及两个已确定的发现：椭圆体带（EZ）破坏和脉络膜高透过率。
识别与糖尿病视网膜神经变性相关的具有临床意义的OCT结果。
AI模型有效地将视力差（小数视力≤0.5）的眼睛与视力好（VA≥1.0）的眼睛区分开来（受试者操作特征曲线下面积为0.947）。基于可解释人工智能的注意力图分别突出了视力差或视力好的眼睛中GCL/IPL（65.2%或70.0%）、ONL（52.2%或28.3%）、EZ（39.1%或21.7%）和脉络膜（26.1%或5.00%）中的显著区域。视网膜专家的主观评估显示这4种发现的频率如下：冰锥征（32.4%）、EZ破坏（25.0%）、椒盐征（16.0%）和脉络膜高透过率（13.5%）。小数视力≤0.9的眼睛比VA≥1.0的眼睛更频繁出现这些发现（所有比较P<0.001）。椒盐征和脉络膜高透过率在识别视力差的眼睛方面具有较高的特异性。统计分析表明，与它们与冰锥征的关系相比，EZ破坏、椒盐征和高透过率之间的关联更显著。
人工智能辅助的OCT结果探索识别出2种已确定的病变和2种新的OCT生物标志物，表明存在具有临床意义的糖尿病视网膜神经变性。
作者对本文讨论的任何材料均无所有权或商业利益。
