Efficacy and accuracy of artificial intelligence to overlay multimodal images from different optical instruments in patients with retinitis pigmentosa.
Efficacy and accuracy of artificial intelligence to overlay multimodal images from different optical instruments in patients with retinitis pigmentosa.
复制标题
DOI：
10.1111/ceo.14234
复制
发表时间：
2023-07
影响因子：
4
通讯作者：
中科院分区：
文献类型：
--
作者：
视网膜色素变性（RP）是一组进行性、遗传异质性致盲性疾病。最近，需要视网膜功能和结构的测量之间的关系，以帮助确定临床试验的结果测量或生物标志物。在不同平台上拍摄的视网膜多模态图像的对齐能力将允许更好地理解这种关系。我们研究了人工智能（AI）在RP患者中覆盖不同多模态视网膜图像的有效性。我们使用手动对齐和AI将RP患者的微视野红外图像叠加在扫描激光检眼镜和光谱域光学相干断层扫描的近红外图像上。人工智能采用了两步框架，并在单独的数据集上进行训练。使用内部软件进行手动对准，该软件允许标记位于血管分叉处的六个关键点。如果重叠图像上相同关键点之间的距离≤1/2°，则认为手动重叠成功。32例患者的57只眼纳入分析。如线性混合效应模型所证实的，与手动对齐相比，AI在对齐图像方面显著更准确和成功（p < 0.001）。用于计算AI（0.991）和手动（0.835）Dice系数曲线下面积与其各自“真值”的关系的受试者操作特征分析发现，AI在叠加中显著更准确（p < 0.001）。AI在RP患者的重叠多模态视网膜成像中比手动对齐显着更准确，并显示出将AI算法用于未来多模态临床和研究应用的潜力。
Retinitis pigmentosa (RP) represents a group of progressive, genetically heterogenous blinding diseases. Recently, relationships between measures of retinal function and structure are needed to help identify outcome measures or biomarkers for clinical trials. The ability to align retinal multimodal images, taken on different platforms, will allow better understanding of this relationship. We investigate the efficacy of artificial intelligence (AI) in overlaying different multimodal retinal images in RP patients. We overlayed infrared images from microperimetry on near-infrared images from scanning laser ophthalmoscope and spectral domain optical coherence tomography in RP patients using manual alignment and AI. The AI adopted a two-step framework and was trained on a separate dataset. Manual alignment was performed using in-house software that allowed labelling of six key points located at vessel bifurcations. Manual overlay was considered successful if the distance between same key points on the overlayed images was ≤1/2°. Fifty-seven eyes of 32 patients were included in the analysis. AI was significantly more accurate and successful in aligning images compared to manual alignment as confirmed by linear mixed-effects modelling (p < 0.001). A receiver operating characteristic analysis, used to compute the area under the curve of the AI (0.991) and manual (0.835) Dice coefficients in relation to their respective ‘truth’ values, found AI significantly more accurate in the overlay (p < 0.001). AI was significantly more accurate than manual alignment in overlaying multimodal retinal imaging in RP patients and showed the potential to use AI algorithms for future multimodal clinical and research applications.
登录
查看更多内容
影响因子：
3
作者：
Birch DG;Cheng P;Duncan JL;Ayala AR;Maguire MG;Audo I;Cheetham JK;Durham TA;Fahim AT;Ferris FL 3rd;Heon E;Huckfeldt RM;Iannaccone A;Khan NW;Lad EM;Michaelides M;Pennesi ME;Stingl K;Vincent A;Weng CY;Foundation Fighting Blindness Consortium Investigator Group
通讯作者：
Foundation Fighting Blindness Consortium Investigator Group
影响因子：
3
作者：
Cavichini M;An C;Bartsch DG;Jhingan M;Amador-Patarroyo MJ;Long CP;Zhang J;Wang Y;Chan AX;Madala S;Nguyen T;Freeman WR
通讯作者：
Freeman WR
影响因子：
4.2
作者：
Lad, Eleonora M.;Duncan, Jacque L.;Liang, Wendi;Maguire, Maureen G.;Ayala, Allison R.;Audo, Isabelle;Birch, David G.;Carroll, Joseph;Cheetham, Janet K.;Durham, Todd A.;Fahim, Abigail T.;Loo, Jessica;Deng, Zengtian;Mukherjee, Dibyendu;Heon, Elise;Hufnagel, Robert B.;Guan, Bin;Iannaccone, Alessandro;Jaffe, Glenn J.;Kay, Christine N.;Michaelides, Michel;Pennesi, Mark E.;Vincent, Ajoy;Weng, Christina Y.;Farsiu, Sina
通讯作者：
Farsiu, Sina
影响因子：
4.4
作者：
Birch DG;Samarakoon L;Melia M;Duncan JL;Ayala AR;Audo I;Cheetham JK;Durham TA;Iannaccone A;Pennesi ME;Stingl K;Foundation Fighting Blindness Consortium Investigator Group
通讯作者：
Foundation Fighting Blindness Consortium Investigator Group
影响因子：
3
作者：
Duncan JL;Pierce EA;Laster AM;Daiger SP;Birch DG;Ash JD;Iannaccone A;Flannery JG;Sahel JA;Zack DJ;Zarbin MA;and the Foundation Fighting Blindness Scientific Advisory Board
通讯作者：
and the Foundation Fighting Blindness Scientific Advisory Board