To evaluate the severity scales of diabetic macular ischemia (DMI) by analyzing the quantity and distribution of capillary nonperfusion using OCT angiography (OCTA) images.
Design
A single-center, prospective case series.
Participants
Three hundred one eyes from 301 patients with diabetic retinopathy.
Methods
We acquired 3 × 3-mm swept-source OCTA images and created en face images within a central 2.5-mm circle. The circle was divided into 15 × 15-pixel squares and nonperfusion squares (NPSs) were defined as those without retinal vessels. Eyes with high-dimensional spatial data were arranged on a 2-dimensional space using the uniform manifold approximation and projection (UMAP) algorithm and classified by clustering into 5 groups: Initial, Mild, Superficial, Moderate, and Severe.
Main Outcome Measures
Development of a severity scale for DMI.
Results
Eyes arranged on a 2-dimensional UMAP space were divided into 5 clusters, based on the similarity of nonperfusion area distribution. Nonperfusion square counts in the deep layer increased in eyes of the Initial, Mild, Moderate, and Severe groups in a stepwise manner. In contrast, there were no significant changes in superficial NPS counts between eyes of the Initial and Mild groups. In the intermediate stage, eyes of the Superficial group exhibited higher NPS counts in the central sector of the superficial layer compared with those of the Moderate group. The foveal avascular zone extended into the temporal subfield of the deep layer in eyes of the Moderate group. Eyes of the Severe group had significantly poorer visual acuity that was more frequently accompanied with proliferative diabetic retinopathy.
Conclusions
The application of dimensionality reduction and clustering has facilitated the development of a novel severity scale for DMI based on the distribution of capillary nonperfusion in OCTA images.
Financial Disclosure(s)
The authors have no proprietary or commercial interest in any materials discussed in this article.
{"title":"Severity Scale of Diabetic Macular Ischemia Based on the Distribution of Capillary Nonperfusion in OCT Angiography","authors":"Miyo Yoshida MD, Tomoaki Murakami MD, PhD, Keiichi Nishikawa MD, Kenji Ishihara MD, PhD, Yuki Mori MD, PhD, Akitaka Tsujikawa MD, PhD","doi":"10.1016/j.xops.2024.100603","DOIUrl":"10.1016/j.xops.2024.100603","url":null,"abstract":"<div><h3>Purpose</h3><div>To evaluate the severity scales of diabetic macular ischemia (DMI) by analyzing the quantity and distribution of capillary nonperfusion using OCT angiography (OCTA) images.</div></div><div><h3>Design</h3><div>A single-center, prospective case series.</div></div><div><h3>Participants</h3><div>Three hundred one eyes from 301 patients with diabetic retinopathy.</div></div><div><h3>Methods</h3><div>We acquired 3 × 3-mm swept-source OCTA images and created en face images within a central 2.5-mm circle. The circle was divided into 15 × 15-pixel squares and nonperfusion squares (NPSs) were defined as those without retinal vessels. Eyes with high-dimensional spatial data were arranged on a 2-dimensional space using the uniform manifold approximation and projection (UMAP) algorithm and classified by clustering into 5 groups: <em>Initial</em>, <em>Mild</em>, <em>Superficial</em>, <em>Moderate</em>, and <em>Severe</em>.</div></div><div><h3>Main Outcome Measures</h3><div>Development of a severity scale for DMI.</div></div><div><h3>Results</h3><div>Eyes arranged on a 2-dimensional UMAP space were divided into 5 clusters, based on the similarity of nonperfusion area distribution. Nonperfusion square counts in the deep layer increased in eyes of the <em>Initial</em>, <em>Mild</em>, <em>Moderate</em>, and <em>Severe</em> groups in a stepwise manner. In contrast, there were no significant changes in superficial NPS counts between eyes of the <em>Initial</em> and <em>Mild</em> groups. In the intermediate stage, eyes of the <em>Superficial</em> group exhibited higher NPS counts in the central sector of the superficial layer compared with those of the <em>Moderate</em> group. The foveal avascular zone extended into the temporal subfield of the deep layer in eyes of the <em>Moderate</em> group. Eyes of the <em>Severe</em> group had significantly poorer visual acuity that was more frequently accompanied with proliferative diabetic retinopathy.</div></div><div><h3>Conclusions</h3><div>The application of dimensionality reduction and clustering has facilitated the development of a novel severity scale for DMI based on the distribution of capillary nonperfusion in OCTA images.</div></div><div><h3>Financial Disclosure(s)</h3><div>The authors have no proprietary or commercial interest in any materials discussed in this article.</div></div>","PeriodicalId":74363,"journal":{"name":"Ophthalmology science","volume":"5 1","pages":"Article 100603"},"PeriodicalIF":3.2,"publicationDate":"2024-09-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.sciencedirect.com/science/article/pii/S2666914524001398/pdfft?md5=9fd86bd124fe988794885bc8b18f64b7&pid=1-s2.0-S2666914524001398-main.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142314355","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-09-05DOI: 10.1016/j.xops.2024.100616
Verônica Vilasboas-Campos MD , Alessandro A. Jammal MD, PhD , Carolina P.B. Gracitelli MD, PhD , Gustavo R. Gameiro MD , Vital P. Costa MD, PhD , Felipe A. Medeiros MD, PhD
Purpose
To evaluate the effect of aging on estimated retinal ganglion cell (RGC) counts over time in healthy eyes, obtained from a combination of structural and functional information.
Design
Longitudinal observational cohort study.
Participants
One hundred healthy eyes of 50 subjects.
Methods
Estimated RGC counts were obtained by a previously described method using standard automated perimetry sensitivity thresholds and OCT retinal nerve fiber layer thickness measurements. Linear mixed-effects models were applied to investigate the effect of aging, as well as other covariates, on rates of change in estimated RGC counts over time.
Main Outcome Measures
Rates of change in estimated RGC counts in healthy eyes.
Results
Subjects had a mean age of 49.6 ± 15.7 years at baseline (range 22.8–89.9 years) and were followed up for 3.5 ± 2.5 years. Thirty-three (66%) patients were female and 11 (22%) self-identified as Black. At baseline, the eyes had an average estimated RGC count of 1 144 010 ± 222 084 cells. After adjusting for confounding factors, the mean rate of change in estimated RGC counts was –6769 RGC/year (95% confidence interval: –10 994 to –2544 RGC/year; P = 0.002), or 0.6%/year. Older age and longer axial length were significantly associated with lower RGC counts at baseline.
Conclusions
A significant age-related decline in estimated RGC counts was found in healthy subjects with a combined metric integrating imaging and functional testing. The estimated mean age-related decline was remarkably similar to estimates from previous histologic studies in cadaver eyes, reinforcing the validity of the proposed combined metric and highlighting the importance of considering age when evaluating RGC count changes over time for monitoring glaucoma progression.
Financial Disclosure(s)
Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.
{"title":"Estimated Rates of Retinal Ganglion Cell Loss with Aging","authors":"Verônica Vilasboas-Campos MD , Alessandro A. Jammal MD, PhD , Carolina P.B. Gracitelli MD, PhD , Gustavo R. Gameiro MD , Vital P. Costa MD, PhD , Felipe A. Medeiros MD, PhD","doi":"10.1016/j.xops.2024.100616","DOIUrl":"10.1016/j.xops.2024.100616","url":null,"abstract":"<div><h3>Purpose</h3><div>To evaluate the effect of aging on estimated retinal ganglion cell (RGC) counts over time in healthy eyes, obtained from a combination of structural and functional information.</div></div><div><h3>Design</h3><div>Longitudinal observational cohort study.</div></div><div><h3>Participants</h3><div>One hundred healthy eyes of 50 subjects.</div></div><div><h3>Methods</h3><div>Estimated RGC counts were obtained by a previously described method using standard automated perimetry sensitivity thresholds and OCT retinal nerve fiber layer thickness measurements. Linear mixed-effects models were applied to investigate the effect of aging, as well as other covariates, on rates of change in estimated RGC counts over time.</div></div><div><h3>Main Outcome Measures</h3><div>Rates of change in estimated RGC counts in healthy eyes.</div></div><div><h3>Results</h3><div>Subjects had a mean age of 49.6 ± 15.7 years at baseline (range 22.8–89.9 years) and were followed up for 3.5 ± 2.5 years. Thirty-three (66%) patients were female and 11 (22%) self-identified as Black. At baseline, the eyes had an average estimated RGC count of 1 144 010 ± 222 084 cells. After adjusting for confounding factors, the mean rate of change in estimated RGC counts was –6769 RGC/year (95% confidence interval: –10 994 to –2544 RGC/year; <em>P</em> = 0.002), or 0.6%/year. Older age and longer axial length were significantly associated with lower RGC counts at baseline.</div></div><div><h3>Conclusions</h3><div>A significant age-related decline in estimated RGC counts was found in healthy subjects with a combined metric integrating imaging and functional testing. The estimated mean age-related decline was remarkably similar to estimates from previous histologic studies in cadaver eyes, reinforcing the validity of the proposed combined metric and highlighting the importance of considering age when evaluating RGC count changes over time for monitoring glaucoma progression.</div></div><div><h3>Financial Disclosure(s)</h3><div>Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.</div></div>","PeriodicalId":74363,"journal":{"name":"Ophthalmology science","volume":"5 1","pages":"Article 100616"},"PeriodicalIF":3.2,"publicationDate":"2024-09-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142655076","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-09-05DOI: 10.1016/j.xops.2024.100615
Nicolas J. Heckenlaible BS , Christopher B. Toomey MD, PhD , James T. Handa MD
Purpose
Automated retinal cell layer segmentation empowers OCT as a precise tool for characterizing morphologic features of retinal health throughout age-related macular degeneration (AMD) progression, particularly in advance of more visible biomarkers such as drusen and macular pigmentary changes. Few studies have examined OCT changes in eyes progressing from early to intermediate disease, or combined examinations of cell layer thickness, reflectivity, and heterogeneity. Therefore, this study analyzed OCTs from eyes progressing from early to intermediate AMD to identify changes in retinal morphology and reflectivity that may serve as biomarkers of early progression.
Design
Retrospective cohort study.
Participants
Patients ≥50 years with a diagnosis of AMD and with high-quality ipsilateral OCTs in both early and intermediate stage disease.
Methods
Fifty OCTs from 25 patients were automatically segmented using a previously validated artificial intelligence-driven algorithm. Changes in the mean and standard deviation of cell layer thickness and reflectivity with progression through stages were calculated for 90 retinal volumes with the help of a novel Python-based analysis tool.
Main Outcome Measures
The primary outcomes were significant changes to cell layer thickness, reflectivity, and heterogeneity with progression of AMD.
Results
With progression from early to intermediate disease, photoreceptor outer segments diffusely thinned. Within the ellipsoid zone, the fovea and parafovea were thinned with a simultaneous increase in thickness variability and a decrease in parafoveal reflectivity. The retinal pigment epithelium-Bruch’s membrane complex underwent diffuse thickening and increased thickness variability alongside a decrease in foveal and parafoveal reflectivity.
Conclusions
These findings correlate with the known histopathology of early AMD and identify measurable OCT trends through the earliest stages of disease.
Financial Disclosures
Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.
目的 自动视网膜细胞层分割使 OCT 成为一种精确的工具，用于描述老年性黄斑变性（AMD）进展过程中视网膜健康的形态特征，特别是在更明显的生物标志物（如色素沉着和黄斑色素变化）出现之前。很少有研究对从早期病变发展到中期病变的眼睛的 OCT 变化进行检查，或对细胞层厚度、反射率和异质性进行综合检查。因此，本研究分析了从早期AMD进展到中期AMD的眼睛的OCT，以确定视网膜形态和反射率的变化，这些变化可作为早期进展的生物标志物。方法使用先前验证过的人工智能驱动算法自动分割25名患者的50幅OCT。在基于 Python 的新型分析工具的帮助下，计算了 90 个视网膜体积的细胞层厚度和反射率的平均值和标准偏差随病程进展的变化。在椭圆区内，眼窝和眼窝旁变薄，同时厚度变异性增加，眼窝旁反射率下降。视网膜色素上皮-布氏膜复合体发生弥漫性增厚，厚度变异性增加，同时眼窝和眼底旁反射率下降。结论这些发现与已知的早期 AMD 组织病理学相关，并确定了疾病最早阶段可测量的 OCT 趋势。
{"title":"OCT Changes Observed during the Progression of Early Age-Related Macular Degeneration","authors":"Nicolas J. Heckenlaible BS , Christopher B. Toomey MD, PhD , James T. Handa MD","doi":"10.1016/j.xops.2024.100615","DOIUrl":"10.1016/j.xops.2024.100615","url":null,"abstract":"<div><h3>Purpose</h3><div>Automated retinal cell layer segmentation empowers OCT as a precise tool for characterizing morphologic features of retinal health throughout age-related macular degeneration (AMD) progression, particularly in advance of more visible biomarkers such as drusen and macular pigmentary changes. Few studies have examined OCT changes in eyes progressing from early to intermediate disease, or combined examinations of cell layer thickness, reflectivity, and heterogeneity. Therefore, this study analyzed OCTs from eyes progressing from early to intermediate AMD to identify changes in retinal morphology and reflectivity that may serve as biomarkers of early progression.</div></div><div><h3>Design</h3><div>Retrospective cohort study.</div></div><div><h3>Participants</h3><div>Patients ≥50 years with a diagnosis of AMD and with high-quality ipsilateral OCTs in both early and intermediate stage disease.</div></div><div><h3>Methods</h3><div>Fifty OCTs from 25 patients were automatically segmented using a previously validated artificial intelligence-driven algorithm. Changes in the mean and standard deviation of cell layer thickness and reflectivity with progression through stages were calculated for 90 retinal volumes with the help of a novel Python-based analysis tool.</div></div><div><h3>Main Outcome Measures</h3><div>The primary outcomes were significant changes to cell layer thickness, reflectivity, and heterogeneity with progression of AMD.</div></div><div><h3>Results</h3><div>With progression from early to intermediate disease, photoreceptor outer segments diffusely thinned. Within the ellipsoid zone, the fovea and parafovea were thinned with a simultaneous increase in thickness variability and a decrease in parafoveal reflectivity. The retinal pigment epithelium-Bruch’s membrane complex underwent diffuse thickening and increased thickness variability alongside a decrease in foveal and parafoveal reflectivity.</div></div><div><h3>Conclusions</h3><div>These findings correlate with the known histopathology of early AMD and identify measurable OCT trends through the earliest stages of disease.</div></div><div><h3>Financial Disclosures</h3><div>Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.</div></div>","PeriodicalId":74363,"journal":{"name":"Ophthalmology science","volume":"5 1","pages":"Article 100615"},"PeriodicalIF":3.2,"publicationDate":"2024-09-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142655075","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-09-04DOI: 10.1016/j.xops.2024.100614
Maria de los Angeles Ramos Cadena MD , Ashley Sohn MD , Heather Livengood PhD , Ting-Fang Lee PhD , Batsheva Rubin BA , Jiyuan Hu PhD , Bernhard A. Sabel PhD , Rachel Matayev BA , Joseph Panarelli MD , Gadi Wollstein MD , Joel S. Schuman MD
Purpose
To determine the efficacy and safety of repetitive transorbital alternating current stimulation (rtACS) treatment by assessing vision-related quality of life and visual function outcome in subjects treated with rtACS versus sham-control.
Sixteen subjects with moderate-to-advanced glaucoma (visual field [VF] mean deviation [MD] ≤−6.00 decibels) randomized into sham (9 subjects) or rtACS intervention (7 subjects) groups.
Methods
Subjects underwent 10 rtACS sessions over 2 weeks. All subjects had comprehensive ocular examination at baseline, 1-week, and 4-weeks posttreatment.
Main Outcome Measures
Visual acuity (VA), contrast sensitivity (CS), VF MD, number of threshold sensitivity points that changed or were unchanged, and vision-related quality of life (VR-QoL) questionnaire scores.
Results
The rtACS group showed a significantly greater improvement from baseline to 4 weeks posttreatment compared with sham in VR-QoL domains including near activities (P < 0.01), dependency (P = 0.03), social functioning (P = 0.03), mental health (P < 0.01) and in the overall composite score (P = 0.04). No significant changes were detected with VA, CS, and VF analyses for either group. No serious adverse events were noted in either study group.
Conclusions
Repetitive transorbital alternating current stimulation therapy showed a significant beneficial effect on several domains of VR-QoL. Further studies will determine its utility in glaucoma.
Financial Disclosure(s)
Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.
{"title":"Transorbital Alternating Current Stimulation in a Double-Masked Randomized Clinical Trial: Visual Functional Effect and Quality of Life","authors":"Maria de los Angeles Ramos Cadena MD , Ashley Sohn MD , Heather Livengood PhD , Ting-Fang Lee PhD , Batsheva Rubin BA , Jiyuan Hu PhD , Bernhard A. Sabel PhD , Rachel Matayev BA , Joseph Panarelli MD , Gadi Wollstein MD , Joel S. Schuman MD","doi":"10.1016/j.xops.2024.100614","DOIUrl":"10.1016/j.xops.2024.100614","url":null,"abstract":"<div><h3>Purpose</h3><div>To determine the efficacy and safety of repetitive transorbital alternating current stimulation (rtACS) treatment by assessing vision-related quality of life and visual function outcome in subjects treated with rtACS versus sham-control.</div></div><div><h3>Study design</h3><div>Double masked, randomized, sham-controlled clinical trial (NCT03188042).</div></div><div><h3>Subjects</h3><div>Sixteen subjects with moderate-to-advanced glaucoma (visual field [VF] mean deviation [MD] ≤−6.00 decibels) randomized into sham (9 subjects) or rtACS intervention (7 subjects) groups.</div></div><div><h3>Methods</h3><div>Subjects underwent 10 rtACS sessions over 2 weeks. All subjects had comprehensive ocular examination at baseline, 1-week, and 4-weeks posttreatment.</div></div><div><h3>Main Outcome Measures</h3><div>Visual acuity (VA), contrast sensitivity (CS), VF MD, number of threshold sensitivity points that changed or were unchanged, and vision-related quality of life (VR-QoL) questionnaire scores.</div></div><div><h3>Results</h3><div>The rtACS group showed a significantly greater improvement from baseline to 4 weeks posttreatment compared with sham in VR-QoL domains including near activities (<em>P</em> < 0.01), dependency (<em>P</em> = 0.03), social functioning (<em>P</em> = 0.03), mental health (<em>P</em> < 0.01) and in the overall composite score (<em>P</em> = 0.04). No significant changes were detected with VA, CS, and VF analyses for either group. No serious adverse events were noted in either study group.</div></div><div><h3>Conclusions</h3><div>Repetitive transorbital alternating current stimulation therapy showed a significant beneficial effect on several domains of VR-QoL. Further studies will determine its utility in glaucoma.</div></div><div><h3>Financial Disclosure(s)</h3><div>Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.</div></div>","PeriodicalId":74363,"journal":{"name":"Ophthalmology science","volume":"5 1","pages":"Article 100614"},"PeriodicalIF":3.2,"publicationDate":"2024-09-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142655034","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-30DOI: 10.1016/j.xops.2024.100613
Shiva Sabazade MD , Marco A. Lumia Michalski MD , Jakub Bartoszek MD , Maria Fili MD, PhD , Mats Holmström MS , Gustav Stålhammar MD, PhD
Purpose
To develop and validate a deep learning algorithm capable of differentiating small choroidal melanomas from nevi.
Design
Retrospective multicenter cohort study.
Participants
A total of 802 images from 688 patients diagnosed with choroidal nevi or melanoma.
Methods
Wide field and standard field fundus photographs were collected from patients diagnosed with choroidal nevi or melanoma by ocular oncologists during clinical examinations. A lesion was classified as a nevus if it was followed for at least 5 years without being rediagnosed as melanoma. A neural network optimized for image classification was trained and validated on cohorts of 495 and 168 images and subsequently tested on independent sets of 86 and 53 images.
Main Outcome Measures
Area under the curve (AUC) in receiver operating characteristic analysis for differentiating small choroidal melanomas from nevi.
Results
The algorithm achieved an AUC of 0.88 in both test cohorts, outperforming ophthalmologists using the Mushroom shape, Orange pigment, Large size, Enlargement, and Subretinal fluid (AUC 0.77) and To Find Small Ocular Melanoma Using Helpful Hints Daily (AUC 0.67) risk factors (DeLong’s test, P < 0.001). The algorithm performed equally well for wide field and standard field photos (AUC 0.89 for both when analyzed separately). Using an optimal operating point of 0.63 (on a scale from 0.00 to 1.00) determined from the training and validation datasets, the algorithm achieved 100% sensitivity and 74% specificity in the first test cohort (F-score 0.72), and 80% sensitivity and 81% specificity in the second (F-score 0.71), which consisted of images from external clinics nationwide. It outperformed 12 ophthalmologists in sensitivity (Mann–Whitney U, P = 0.006) but not specificity (P = 0.54). The algorithm showed higher sensitivity than both resident and consultant ophthalmologists (Dunn's test, P = 0.04 and P = 0.006, respectively) but not ocular oncologists (P > 0.99, all P values Bonferroni corrected).
Conclusions
This study develops and validates a deep learning algorithm for differentiating small choroidal melanomas from nevi, matching or surpassing the discriminatory performance of experienced human ophthalmologists. Further research will aim to validate its utility in clinical settings.
Financial Disclosure(s)
Financial DisclosuresProprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.
{"title":"Development and Validation of a Deep Learning Algorithm for Differentiation of Choroidal Nevi from Small Melanoma in Fundus Photographs","authors":"Shiva Sabazade MD , Marco A. Lumia Michalski MD , Jakub Bartoszek MD , Maria Fili MD, PhD , Mats Holmström MS , Gustav Stålhammar MD, PhD","doi":"10.1016/j.xops.2024.100613","DOIUrl":"10.1016/j.xops.2024.100613","url":null,"abstract":"<div><h3>Purpose</h3><div>To develop and validate a deep learning algorithm capable of differentiating small choroidal melanomas from nevi.</div></div><div><h3>Design</h3><div>Retrospective multicenter cohort study.</div></div><div><h3>Participants</h3><div>A total of 802 images from 688 patients diagnosed with choroidal nevi or melanoma.</div></div><div><h3>Methods</h3><div>Wide field and standard field fundus photographs were collected from patients diagnosed with choroidal nevi or melanoma by ocular oncologists during clinical examinations. A lesion was classified as a nevus if it was followed for at least 5 years without being rediagnosed as melanoma. A neural network optimized for image classification was trained and validated on cohorts of 495 and 168 images and subsequently tested on independent sets of 86 and 53 images.</div></div><div><h3>Main Outcome Measures</h3><div>Area under the curve (AUC) in receiver operating characteristic analysis for differentiating small choroidal melanomas from nevi.</div></div><div><h3>Results</h3><div>The algorithm achieved an AUC of 0.88 in both test cohorts, outperforming ophthalmologists using the Mushroom shape, Orange pigment, Large size, Enlargement, and Subretinal fluid (AUC 0.77) and To Find Small Ocular Melanoma Using Helpful Hints Daily (AUC 0.67) risk factors (DeLong’s test, <em>P</em> < 0.001). The algorithm performed equally well for wide field and standard field photos (AUC 0.89 for both when analyzed separately). Using an optimal operating point of 0.63 (on a scale from 0.00 to 1.00) determined from the training and validation datasets, the algorithm achieved 100% sensitivity and 74% specificity in the first test cohort (F-score 0.72), and 80% sensitivity and 81% specificity in the second (F-score 0.71), which consisted of images from external clinics nationwide. It outperformed 12 ophthalmologists in sensitivity (Mann–Whitney <em>U</em>, <em>P</em> = 0.006) but not specificity (<em>P</em> = 0.54). The algorithm showed higher sensitivity than both resident and consultant ophthalmologists (Dunn's test, <em>P</em> = 0.04 and <em>P</em> = 0.006, respectively) but not ocular oncologists (<em>P</em> > 0.99, all <em>P</em> values Bonferroni corrected).</div></div><div><h3>Conclusions</h3><div>This study develops and validates a deep learning algorithm for differentiating small choroidal melanomas from nevi, matching or surpassing the discriminatory performance of experienced human ophthalmologists. Further research will aim to validate its utility in clinical settings.</div></div><div><h3>Financial Disclosure(s)</h3><div>Financial DisclosuresProprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.</div></div>","PeriodicalId":74363,"journal":{"name":"Ophthalmology science","volume":"5 1","pages":"Article 100613"},"PeriodicalIF":3.2,"publicationDate":"2024-08-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142423211","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
To evaluate the OCT images of eyes with fovea-involved complete retinal pigment epithelial and outer retinal atrophy (cRORA) as well as best-corrected visual acuity (BCVA) to explore the pathogenesis of visual impairment and atrophy.
Design
Retrospective observational study.
Subjects
Data of eyes with cRORA associated with age-related macular degeneration with foveal center involvement were collected from 10 hospitals in Japan.
Methods
Ophthalmic examination data, BCVA, and extents of retinal pigment epithelial and outer retinal atrophy (RORA), represented by choroidal hyper-transmission, and outer plexiform layer (OPL) deterioration, central retinal thickness (CRT), and central choroidal thickness (CCT) measured using built-in software on the sectional OCT images were evaluated.
Main Outcome Measures
Relationship between BCVA and extents of RORA and OPL deterioration.
Results
Of the 64 eyes of 64 patients (mean age: 76.8 ± 9.5 years old), 38 eyes (59.4%) belonged to men. Mean BCVA was 0.602 ± 0.475 (median: 0.523; range, −0.079 to 1.523) in logarithm of the minimum angle of resolution (logMAR). Mean extent of RORA was 2921 ± 1291 (median: 3172; range: 479–5985) μm. BCVA in logMAR positively correlated with extents of RORA (P = 0.004) and OPL deterioration (P = 0.004) and negatively correlated with CRT (P = 0.022). Best-corrected visual acuity ≥0.5 was associated with extents of RORA ≥3000 μm (odds ratio [OR], 4.227; 95% confidence interval [CI], 1.440–12.408; P = 0.009) and OPL deterioration ≥1700 μm (OR, 2.984; 95% CI, 1.034–8.609; P = 0.043), and presence of complete central outer plexiform layer defect (cCOD) (OR, 12.700; 95% CI, 2.439–66.132; P = 0.003), after adjusting for age and sex. The extent of RORA ≥3000 μm was associated with BCVA ≥0.5 (OR, 4.213; 95% CI, 1.437–12.356; P = 0.009), extent of OPL deterioration ≥1700 μm (OR, 58.682; 95% CI, 6.865–501.592; P < 0.001), and presence of cCOD (OR, 4.107; 95% CI, 1.339–12.604; P = 0.014), after adjusting for age and sex. The extent of RORA positively correlated with that of OPL deterioration (P < 0.001), CRT (P = 0.001), and CCT (P = 0.041).
Conclusions
A longer extent of cRORA in the OCT images with foveal center involvement was associated with a longer extent of OPL deterioration and the presence of cCOD and worse BCVA. Further studies focusing on OPL changes are warranted for understanding the pathogenesis of RORA and vision loss.
Financial Disclosures
Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.
{"title":"Extent of Complete Retinal Pigment Epithelial and Outer Retinal Atrophy with Foveal Center Involvement is Associated with Visual Acuity","authors":"Norihiro Nagai MD , Hisashi Matsubara MD , Hiroto Terasaki MD , Takao Hirano MD , Aki Kato MD , Akiko Miki MD , Hiromasa Hirai MD , Fumiko Murao MD , Hiroko Imaizumi MD , Fumi Gomi MD , Yoshinori Mitamura MD , Nahoko Ogata MD , Sentaro Kusuhara MD , Tsutomu Yasukawa MD , Toshinori Murata MD , Taiji Sakamoto MD , Mineo Kondo MD , Hajime Shinoda MD , Yoko Ozawa MD","doi":"10.1016/j.xops.2024.100612","DOIUrl":"10.1016/j.xops.2024.100612","url":null,"abstract":"<div><h3>Purpose</h3><div>To evaluate the OCT images of eyes with fovea-involved complete retinal pigment epithelial and outer retinal atrophy (cRORA) as well as best-corrected visual acuity (BCVA) to explore the pathogenesis of visual impairment and atrophy.</div></div><div><h3>Design</h3><div>Retrospective observational study.</div></div><div><h3>Subjects</h3><div>Data of eyes with cRORA associated with age-related macular degeneration with foveal center involvement were collected from 10 hospitals in Japan.</div></div><div><h3>Methods</h3><div>Ophthalmic examination data, BCVA, and extents of retinal pigment epithelial and outer retinal atrophy (RORA), represented by choroidal hyper-transmission, and outer plexiform layer (OPL) deterioration, central retinal thickness (CRT), and central choroidal thickness (CCT) measured using built-in software on the sectional OCT images were evaluated.</div></div><div><h3>Main Outcome Measures</h3><div>Relationship between BCVA and extents of RORA and OPL deterioration.</div></div><div><h3>Results</h3><div>Of the 64 eyes of 64 patients (mean age: 76.8 ± 9.5 years old), 38 eyes (59.4%) belonged to men. Mean BCVA was 0.602 ± 0.475 (median: 0.523; range, −0.079 to 1.523) in logarithm of the minimum angle of resolution (logMAR). Mean extent of RORA was 2921 ± 1291 (median: 3172; range: 479–5985) μm. BCVA in logMAR positively correlated with extents of RORA (<em>P</em> = 0.004) and OPL deterioration (<em>P</em> = 0.004) and negatively correlated with CRT (<em>P</em> = 0.022). Best-corrected visual acuity ≥0.5 was associated with extents of RORA ≥3000 μm (odds ratio [OR], 4.227; 95% confidence interval [CI], 1.440–12.408; <em>P</em> = 0.009) and OPL deterioration ≥1700 μm (OR, 2.984; 95% CI, 1.034–8.609; <em>P</em> = 0.043), and presence of complete central outer plexiform layer defect (cCOD) (OR, 12.700; 95% CI, 2.439–66.132; <em>P</em> = 0.003), after adjusting for age and sex. The extent of RORA ≥3000 μm was associated with BCVA ≥0.5 (OR, 4.213; 95% CI, 1.437–12.356; <em>P</em> = 0.009), extent of OPL deterioration ≥1700 μm (OR, 58.682; 95% CI, 6.865–501.592; <em>P</em> < 0.001), and presence of cCOD (OR, 4.107; 95% CI, 1.339–12.604; <em>P</em> = 0.014), after adjusting for age and sex. The extent of RORA positively correlated with that of OPL deterioration (<em>P</em> < 0.001), CRT (<em>P</em> = 0.001), and CCT (<em>P</em> = 0.041).</div></div><div><h3>Conclusions</h3><div>A longer extent of cRORA in the OCT images with foveal center involvement was associated with a longer extent of OPL deterioration and the presence of cCOD and worse BCVA. Further studies focusing on OPL changes are warranted for understanding the pathogenesis of RORA and vision loss.</div></div><div><h3>Financial Disclosures</h3><div>Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.</div></div>","PeriodicalId":74363,"journal":{"name":"Ophthalmology science","volume":"5 1","pages":"Article 100612"},"PeriodicalIF":3.2,"publicationDate":"2024-08-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142423204","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-28DOI: 10.1016/j.xops.2024.100610
Petra P. Larsen MD , Catherine Féart PhD , Jean-Paul Pais de Barros PhD , Laure Gayraud MSc , Marie-Noëlle Delyfer MD, PhD , Jean-François Korobelnik MD , Cédric Schweitzer MD, PhD , Cécile Delcourt PhD
Purpose
Lipopolysaccharide (LPS)-type endotoxins are naturally found in the gut microbiota and there is emerging evidence linking gut microbiota and neuroinflammation leading to retinal neurodegeneration. Thinning of the retinal nerve fiber layer (RNFL) is a biomarker of retinal neurodegeneration, and a hallmark of glaucoma, the second leading cause of blindness worldwide. We assessed the association of a blood biomarker of LPS with peripapillary RNFL thickness (RNFLT) and its longitudinal evolution up to 11 years.
Design
The Alienor study is a single center prospective population-based cohort study.
Subjects
The studied sample of this study includes 1062 eyes of 548 participants receiving ≥1 gradable RNFL measurement.
Methods
Plasma esterified 3-hydroxy fatty acids (3-OH FAs) were measured as a proxy of LPS burden. Retinal nerve fiber layer thickness was acquired using spectral-domain OCT imaging every 2 years from 2009 to 2020 (up to 5 visits).
Main Outcome Measures
Associations of plasma esterified 3-OH FAs with RNFLT were assessed using linear mixed models.
Results
Mean age of the included 548 participants was 82.4 ± 4.3 years and 62.6% were women. Higher plasma esterified 3-OH FAs was significantly associated with thinner RNFLT at baseline (coefficient beta = −1.42 microns for 1 standard deviation-increase in 3-OH FAs, 95% confidence interval [−2.56; −0.28], P = 0.02). This association remained stable after multivariate adjustment for potential confounders. No statistically significant association was found between 3-OH FAs and longitudinal RNFLT change.
Conclusions
Higher plasma esterified 3-OH FAs were associated with thinner RNFLT at baseline, indicating an involvement of LPS in the early processes of optic nerve neurodegeneration and highlighting the potential importance of the human microbiota in preserving retinal health.
Financial Disclosure(s)
Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.
{"title":"Association of Lipopolysaccharide-Type Endotoxins with Retinal Neurodegeneration: The Alienor Study","authors":"Petra P. Larsen MD , Catherine Féart PhD , Jean-Paul Pais de Barros PhD , Laure Gayraud MSc , Marie-Noëlle Delyfer MD, PhD , Jean-François Korobelnik MD , Cédric Schweitzer MD, PhD , Cécile Delcourt PhD","doi":"10.1016/j.xops.2024.100610","DOIUrl":"10.1016/j.xops.2024.100610","url":null,"abstract":"<div><h3>Purpose</h3><div>Lipopolysaccharide (LPS)-type endotoxins are naturally found in the gut microbiota and there is emerging evidence linking gut microbiota and neuroinflammation leading to retinal neurodegeneration. Thinning of the retinal nerve fiber layer (RNFL) is a biomarker of retinal neurodegeneration, and a hallmark of glaucoma, the second leading cause of blindness worldwide. We assessed the association of a blood biomarker of LPS with peripapillary RNFL thickness (RNFLT) and its longitudinal evolution up to 11 years.</div></div><div><h3>Design</h3><div>The Alienor study is a single center prospective population-based cohort study.</div></div><div><h3>Subjects</h3><div>The studied sample of this study includes 1062 eyes of 548 participants receiving ≥1 gradable RNFL measurement.</div></div><div><h3>Methods</h3><div>Plasma esterified 3-hydroxy fatty acids (3-OH FAs) were measured as a proxy of LPS burden. Retinal nerve fiber layer thickness was acquired using spectral-domain OCT imaging every 2 years from 2009 to 2020 (up to 5 visits).</div></div><div><h3>Main Outcome Measures</h3><div>Associations of plasma esterified 3-OH FAs with RNFLT were assessed using linear mixed models.</div></div><div><h3>Results</h3><div>Mean age of the included 548 participants was 82.4 ± 4.3 years and 62.6% were women. Higher plasma esterified 3-OH FAs was significantly associated with thinner RNFLT at baseline (coefficient beta = −1.42 microns for 1 standard deviation-increase in 3-OH FAs, 95% confidence interval [−2.56; −0.28], <em>P</em> = 0.02). This association remained stable after multivariate adjustment for potential confounders. No statistically significant association was found between 3-OH FAs and longitudinal RNFLT change.</div></div><div><h3>Conclusions</h3><div>Higher plasma esterified 3-OH FAs were associated with thinner RNFLT at baseline, indicating an involvement of LPS in the early processes of optic nerve neurodegeneration and highlighting the potential importance of the human microbiota in preserving retinal health.</div></div><div><h3>Financial Disclosure(s)</h3><div>Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.</div></div>","PeriodicalId":74363,"journal":{"name":"Ophthalmology science","volume":"5 1","pages":"Article 100610"},"PeriodicalIF":3.2,"publicationDate":"2024-08-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.sciencedirect.com/science/article/pii/S2666914524001465/pdfft?md5=d87b134146a791f80cb8b2ca7c30456e&pid=1-s2.0-S2666914524001465-main.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142314354","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-23DOI: 10.1016/j.xops.2024.100600
Jimmy S. Chen MD , Akshay J. Reddy BS , Eman Al-Sharif MD , Marissa K. Shoji MD , Fritz Gerald P. Kalaw MD , Medi Eslani MD , Paul Z. Lang MD , Malvika Arya MD , Zachary A. Koretz MD, MPH , Kyle A. Bolo MD , Justin J. Arnett MD , Aliya C. Roginiel MD, MPH , Jiun L. Do MD, PhD , Shira L. Robbins MD , Andrew S. Camp MD , Nathan L. Scott MD , Jolene C. Rudell MD, PhD , Robert N. Weinreb MD , Sally L. Baxter MD, MSc , David B. Granet MD, MHCM
Objective
Large language models such as ChatGPT have demonstrated significant potential in question-answering within ophthalmology, but there is a paucity of literature evaluating its ability to generate clinical assessments and discussions. The objectives of this study were to (1) assess the accuracy of assessment and plans generated by ChatGPT and (2) evaluate ophthalmologists’ abilities to distinguish between responses generated by clinicians versus ChatGPT.
Design
Cross-sectional mixed-methods study.
Subjects
Sixteen ophthalmologists from a single academic center, of which 10 were board-eligible and 6 were board-certified, were recruited to participate in this study.
Methods
Prompt engineering was used to ensure ChatGPT output discussions in the style of the ophthalmologist author of the Medical College of Wisconsin Ophthalmic Case Studies. Cases where ChatGPT accurately identified the primary diagnoses were included and then paired. Masked human-generated and ChatGPT-generated discussions were sent to participating ophthalmologists to identify the author of the discussions. Response confidence was assessed using a 5-point Likert scale score, and subjective feedback was manually reviewed.
Main Outcome Measures
Accuracy of ophthalmologist identification of discussion author, as well as subjective perceptions of human-generated versus ChatGPT-generated discussions.
Results
Overall, ChatGPT correctly identified the primary diagnosis in 15 of 17 (88.2%) cases. Two cases were excluded from the paired comparison due to hallucinations or fabrications of nonuser-provided data. Ophthalmologists correctly identified the author in 77.9% ± 26.6% of the 13 included cases, with a mean Likert scale confidence rating of 3.6 ± 1.0. No significant differences in performance or confidence were found between board-certified and board-eligible ophthalmologists. Subjectively, ophthalmologists found that discussions written by ChatGPT tended to have more generic responses, irrelevant information, hallucinated more frequently, and had distinct syntactic patterns (all P < 0.01).
Conclusions
Large language models have the potential to synthesize clinical data and generate ophthalmic discussions. While these findings have exciting implications for artificial intelligence-assisted health care delivery, more rigorous real-world evaluation of these models is necessary before clinical deployment.
Financial Disclosures
The author(s) have no proprietary or commercial interest in any materials discussed in this article.
{"title":"Analysis of ChatGPT Responses to Ophthalmic Cases: Can ChatGPT Think like an Ophthalmologist?","authors":"Jimmy S. Chen MD , Akshay J. Reddy BS , Eman Al-Sharif MD , Marissa K. Shoji MD , Fritz Gerald P. Kalaw MD , Medi Eslani MD , Paul Z. Lang MD , Malvika Arya MD , Zachary A. Koretz MD, MPH , Kyle A. Bolo MD , Justin J. Arnett MD , Aliya C. Roginiel MD, MPH , Jiun L. Do MD, PhD , Shira L. Robbins MD , Andrew S. Camp MD , Nathan L. Scott MD , Jolene C. Rudell MD, PhD , Robert N. Weinreb MD , Sally L. Baxter MD, MSc , David B. Granet MD, MHCM","doi":"10.1016/j.xops.2024.100600","DOIUrl":"10.1016/j.xops.2024.100600","url":null,"abstract":"<div><h3>Objective</h3><p>Large language models such as ChatGPT have demonstrated significant potential in question-answering within ophthalmology, but there is a paucity of literature evaluating its ability to generate clinical assessments and discussions. The objectives of this study were to (1) assess the accuracy of assessment and plans generated by ChatGPT and (2) evaluate ophthalmologists’ abilities to distinguish between responses generated by clinicians versus ChatGPT.</p></div><div><h3>Design</h3><p>Cross-sectional mixed-methods study.</p></div><div><h3>Subjects</h3><p>Sixteen ophthalmologists from a single academic center, of which 10 were board-eligible and 6 were board-certified, were recruited to participate in this study.</p></div><div><h3>Methods</h3><p>Prompt engineering was used to ensure ChatGPT output discussions in the style of the ophthalmologist author of the Medical College of Wisconsin Ophthalmic Case Studies. Cases where ChatGPT accurately identified the primary diagnoses were included and then paired. Masked human-generated and ChatGPT-generated discussions were sent to participating ophthalmologists to identify the author of the discussions. Response confidence was assessed using a 5-point Likert scale score, and subjective feedback was manually reviewed.</p></div><div><h3>Main Outcome Measures</h3><p>Accuracy of ophthalmologist identification of discussion author, as well as subjective perceptions of human-generated versus ChatGPT-generated discussions.</p></div><div><h3>Results</h3><p>Overall, ChatGPT correctly identified the primary diagnosis in 15 of 17 (88.2%) cases. Two cases were excluded from the paired comparison due to hallucinations or fabrications of nonuser-provided data. Ophthalmologists correctly identified the author in 77.9% ± 26.6% of the 13 included cases, with a mean Likert scale confidence rating of 3.6 ± 1.0. No significant differences in performance or confidence were found between board-certified and board-eligible ophthalmologists. Subjectively, ophthalmologists found that discussions written by ChatGPT tended to have more generic responses, irrelevant information, hallucinated more frequently, and had distinct syntactic patterns (all <em>P</em> < 0.01).</p></div><div><h3>Conclusions</h3><p>Large language models have the potential to synthesize clinical data and generate ophthalmic discussions. While these findings have exciting implications for artificial intelligence-assisted health care delivery, more rigorous real-world evaluation of these models is necessary before clinical deployment.</p></div><div><h3>Financial Disclosures</h3><p>The author(s) have no proprietary or commercial interest in any materials discussed in this article.</p></div>","PeriodicalId":74363,"journal":{"name":"Ophthalmology science","volume":"5 1","pages":"Article 100600"},"PeriodicalIF":3.2,"publicationDate":"2024-08-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.sciencedirect.com/science/article/pii/S2666914524001362/pdfft?md5=1fc56cec0e121016c01c38686515b525&pid=1-s2.0-S2666914524001362-main.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142270456","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-08-23DOI: 10.1016/j.xops.2024.100602
Abdulla Shaheen MD , Gabriele Gallo Afflitto MD , Swarup S. Swaminathan MD
Purpose
To evaluate the performance of a large language model (LLM) in classifying electronic health record (EHR) text, and to use this classification to evaluate the type and resolution of hemorrhagic events (HEs) after microinvasive glaucoma surgery (MIGS).
Design
Retrospective cohort study.
Participants
Eyes from the Bascom Palmer Glaucoma Repository.
Methods
Eyes that underwent MIGS between July 1, 2014 and February 1, 2022 were analyzed. Chat Generative Pre-trained Transformer (ChatGPT) was used to classify deidentified EHR anterior chamber examination text into HE categories (no hyphema, microhyphema, clot, and hyphema). Agreement between classifications by ChatGPT and a glaucoma specialist was evaluated using Cohen’s Kappa and precision-recall (PR) curve. Time to resolution of HEs was assessed using Cox proportional-hazards models. Goniotomy HE resolution was evaluated by degree of angle treatment (90°–179°, 180°–269°, 270°–360°). Logistic regression was used to identify HE risk factors.
Main Outcome Measures
Accuracy of ChatGPT HE classification and incidence and resolution of HEs.
Results
The study included 434 goniotomy eyes (368 patients) and 528 Schlemm’s canal stent (SCS) eyes (390 patients). Chat Generative Pre-trained Transformer facilitated excellent HE classification (Cohen’s kappa 0.93, area under PR curve 0.968). Using ChatGPT classifications, at postoperative day 1, HEs occurred in 67.8% of goniotomy and 25.2% of SCS eyes (P < 0.001). The 270° to 360° goniotomy group had the highest HE rate (84.0%, P < 0.001). At postoperative week 1, HEs were observed in 43.4% and 11.3% of goniotomy and SCS eyes, respectively (P < 0.001). By postoperative month 1, HE rates were 13.3% and 1.3% among goniotomy and SCS eyes, respectively (P < 0.001). Time to HE resolution differed between the goniotomy angle groups (log-rank P = 0.034); median time to resolution was 10, 10, and 15 days for the 90° to 179°, 180° to 269°, and 270° to 360° groups, respectively. Risk factor analysis demonstrated greater goniotomy angle was the only significant predictor of HEs (odds ratio for 270°–360°: 4.08, P < 0.001).
Conclusions
Large language models can be effectively used to classify longitudinal EHR free-text examination data with high accuracy, highlighting a promising direction for future LLM-assisted research and clinical decision support. Hemorrhagic events are relatively common self-resolving complications that occur more often in goniotomy cases and with larger goniotomy treatments. Time to HE resolution differs significantly between goniotomy groups.
Financial Disclosure(s)
Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.
目的 评估大型语言模型（LLM）在对电子健康记录（EHR）文本进行分类时的性能，并利用该分类评估微创青光眼手术（MIGS）后出血事件（HEs）的类型和解决情况。方法 分析2014年7月1日至2022年2月1日期间接受MIGS手术的眼球。使用 Chat Generative Pre-trained Transformer (ChatGPT) 将去标识化的 EHR 前房检查文本分为 HE 类别（无红斑、微红斑、血块和红斑）。ChatGPT 和青光眼专家分类之间的一致性采用 Cohen's Kappa 和精确度-召回 (PR) 曲线进行评估。采用 Cox 比例危险模型评估 HE 的消退时间。巩膜切开术 HE 解决时间按角膜治疗程度（90°-179°、180°-269°、270°-360°）进行评估。主要结果测量ChatGPT HE分类的准确性以及HE的发生率和缓解率。结果该研究包括434只眼球切开术眼睛（368名患者）和528只Schlemm's管支架（SCS）眼睛（390名患者）。Chat Generative Pre-trained Transformer（聊天生成预训练变换器）为出色的 HE 分类提供了便利（Cohen's kappa 0.93，PR 曲线下面积 0.968）。使用 ChatGPT 分类，在术后第 1 天，67.8% 的眼球切开术眼和 25.2% 的 SCS 眼发生了 HE（P < 0.001）。270° 至 360° 角膜切开术组的 HE 发生率最高（84.0%，P < 0.001）。术后第 1 周，分别有 43.4% 和 11.3% 的眼球切开术和 SCS 眼睛出现 HE（P < 0.001）。到术后第 1 个月，开导晶体植入术眼和 SCS 眼的 HE 发生率分别为 13.3% 和 1.3% （P < 0.001）。不同视神经切开角度组的 HE 缓解时间不同（对数秩 P = 0.034）；90°至 179°、180°至 269°和 270°至 360°组的中位缓解时间分别为 10 天、10 天和 15 天。结论大语言模型可有效地用于对纵向电子病历自由文本检查数据进行分类，且准确率很高，这为未来的 LLM 辅助研究和临床决策支持指明了方向。出血事件是相对常见的自愈性并发症，更常发生在声门切开术病例中，也更常发生在较大的声门切开术治疗中。不同开腹手术组的 HE 解决时间差异很大。
{"title":"ChatGPT-Assisted Classification of Postoperative Bleeding Following Microinvasive Glaucoma Surgery Using Electronic Health Record Data","authors":"Abdulla Shaheen MD , Gabriele Gallo Afflitto MD , Swarup S. Swaminathan MD","doi":"10.1016/j.xops.2024.100602","DOIUrl":"10.1016/j.xops.2024.100602","url":null,"abstract":"<div><h3>Purpose</h3><p>To evaluate the performance of a large language model (LLM) in classifying electronic health record (EHR) text, and to use this classification to evaluate the type and resolution of hemorrhagic events (HEs) after microinvasive glaucoma surgery (MIGS).</p></div><div><h3>Design</h3><p>Retrospective cohort study.</p></div><div><h3>Participants</h3><p>Eyes from the Bascom Palmer Glaucoma Repository.</p></div><div><h3>Methods</h3><p>Eyes that underwent MIGS between July 1, 2014 and February 1, 2022 were analyzed. Chat Generative Pre-trained Transformer (ChatGPT) was used to classify deidentified EHR anterior chamber examination text into HE categories (no hyphema, microhyphema, clot, and hyphema). Agreement between classifications by ChatGPT and a glaucoma specialist was evaluated using Cohen’s Kappa and precision-recall (PR) curve. Time to resolution of HEs was assessed using Cox proportional-hazards models. Goniotomy HE resolution was evaluated by degree of angle treatment (90°–179°, 180°–269°, 270°–360°). Logistic regression was used to identify HE risk factors.</p></div><div><h3>Main Outcome Measures</h3><p>Accuracy of ChatGPT HE classification and incidence and resolution of HEs.</p></div><div><h3>Results</h3><p>The study included 434 goniotomy eyes (368 patients) and 528 Schlemm’s canal stent (SCS) eyes (390 patients). Chat Generative Pre-trained Transformer facilitated excellent HE classification (Cohen’s kappa 0.93, area under PR curve 0.968). Using ChatGPT classifications, at postoperative day 1, HEs occurred in 67.8% of goniotomy and 25.2% of SCS eyes (<em>P</em> < 0.001). The 270° to 360° goniotomy group had the highest HE rate (84.0%, <em>P</em> < 0.001). At postoperative week 1, HEs were observed in 43.4% and 11.3% of goniotomy and SCS eyes, respectively (<em>P</em> < 0.001). By postoperative month 1, HE rates were 13.3% and 1.3% among goniotomy and SCS eyes, respectively (<em>P</em> < 0.001). Time to HE resolution differed between the goniotomy angle groups (log-rank <em>P</em> = 0.034); median time to resolution was 10, 10, and 15 days for the 90° to 179°, 180° to 269°, and 270° to 360° groups, respectively. Risk factor analysis demonstrated greater goniotomy angle was the only significant predictor of HEs (odds ratio for 270°–360°: 4.08, <em>P</em> < 0.001).</p></div><div><h3>Conclusions</h3><p>Large language models can be effectively used to classify longitudinal EHR free-text examination data with high accuracy, highlighting a promising direction for future LLM-assisted research and clinical decision support. Hemorrhagic events are relatively common self-resolving complications that occur more often in goniotomy cases and with larger goniotomy treatments. Time to HE resolution differs significantly between goniotomy groups.</p></div><div><h3>Financial Disclosure(s)</h3><p>Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.</p></div>","PeriodicalId":74363,"journal":{"name":"Ophthalmology science","volume":"5 1","pages":"Article 100602"},"PeriodicalIF":3.2,"publicationDate":"2024-08-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.sciencedirect.com/science/article/pii/S2666914524001386/pdfft?md5=6a7c056392e56a9af8bd6168c9dd77cb&pid=1-s2.0-S2666914524001386-main.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142270369","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号