Accuracy of integrated Artificial Intelligence (AI) grading at the point of care (POC) using handheld retinal imaging in a community-based diabetic retinopathy (DR) screening program (DRSP)

Recivall Salongcay; Lizzie Anne Aquino; Claude Michael Salva; Glenn Paulo Alog; Kaye Locaylocay; Aileen Viguilla Saunar; Cris Martin P. Jacoba; Jennifer K. Sun; Lloyd P. Aiello; Tunde Peto; Paolo S. Silva

Abstract

Purpose
To prospectively evaluate handheld retinal images assessed by AI at the time of imaging as compared to standard retinal image graders at a centralized reading center (RC).

Methods
Prospective comparative study of AI assessment of referable DR [(refDR) moderate nonproliferative DR (NPDR) or worse, or any level of diabetic macular edema (DME)] and vision threatening DR [(vtDR) severe NPDR or worse, or any level of center involving DME (ciDME)]. AI assessment of disc and macular images performed at the time of imaging was compared with RC evaluation of validated 5-field handheld retinal images [(5F) disc, macula, temporal, superior and inferior]. RC evaluation of the 5F images followed the international DR/DME classification. Sensitivity and specificity (SN/SP) for ungradable images, refDR and vtDR were calculated.

Results
1,733 eyes from 869 diabetic (DM) patients were enrolled in the study. Cohort demographic: age 59.5±10.0, 64.7% female, 98.4% type 2, DM duration 6.6±7.2 years. RC distribution of DR severity: no DR 70.5%, mild NPDR 9.3%, moderate NPDR 7.6%, severe NPDR 3.7%, PDR 2.9%, ungradable 6.0%. DME severity: no DME 82.2%, DME 6.2%, ciDME 4.2%, ungradable 7.3%. RefDR was present in 13.8% and vtDR in 7.8% of eyes. Images were ungradable for DR or DME in 7.7% by RC and 20.7% by AI. Table 1 summarizes the SN/SP and operating characteristics of POC AI grading. SN/SP of AI grading compared to RC evaluation was 0.83/0.95 for refDR and 0.96/0.91 for vtDR. 4 eyes with vtDR (3 severe NPDR and 1 proliferative DR) were missed by AI. Comparisons of performance of the POC AI with existing FDA approved algorithms are presented in table 2.

Conclusions
This study demonstrates that POC AI following a defined retinal imaging protocol at the time of imaging has SN/SP for refDR that meets the current acceptable thresholds of 0.80 and 0.95. Integrating AI at the POC could substantially reduce centralized reading center burden and speeds information delivery to the patient, allowing more prompt eye care referral.

Original language	English
Pages (from-to)	1159
Number of pages	1
Journal	Investigative Opthalmology and Visual Science
Volume	63
Issue number	7
Publication status	Published - 01 Jun 2022
Event	Association for Research in Vision and Ophthalmology Annual Meeting 2022 - New Orleans, USA, Denver, United States Duration: 01 May 2022 → 04 May 2022 https://www.arvo.org/annual-meeting/

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

https://iovs.arvojournals.org/article.aspx?articleid=2782793Licence: CC BY-NC-ND

Cite this

Salongcay, R., Aquino, L. A., Salva, C. M., Alog, G. P., Locaylocay, K., Viguilla Saunar, A., Jacoba, C. M. P., Sun, J. K., Aiello, L. P., Peto, T., & Silva, P. S. (2022). Accuracy of integrated Artificial Intelligence (AI) grading at the point of care (POC) using handheld retinal imaging in a community-based diabetic retinopathy (DR) screening program (DRSP). Investigative Opthalmology and Visual Science, 63(7), 1159. https://iovs.arvojournals.org/article.aspx?articleid=2782793

@article{29338104cb86477d9afd3669ba191cca,

title = "Accuracy of integrated Artificial Intelligence (AI) grading at the point of care (POC) using handheld retinal imaging in a community-based diabetic retinopathy (DR) screening program (DRSP)",

abstract = "PurposeTo prospectively evaluate handheld retinal images assessed by AI at the time of imaging as compared to standard retinal image graders at a centralized reading center (RC).MethodsProspective comparative study of AI assessment of referable DR [(refDR) moderate nonproliferative DR (NPDR) or worse, or any level of diabetic macular edema (DME)] and vision threatening DR [(vtDR) severe NPDR or worse, or any level of center involving DME (ciDME)]. AI assessment of disc and macular images performed at the time of imaging was compared with RC evaluation of validated 5-field handheld retinal images [(5F) disc, macula, temporal, superior and inferior]. RC evaluation of the 5F images followed the international DR/DME classification. Sensitivity and specificity (SN/SP) for ungradable images, refDR and vtDR were calculated.Results1,733 eyes from 869 diabetic (DM) patients were enrolled in the study. Cohort demographic: age 59.5±10.0, 64.7% female, 98.4% type 2, DM duration 6.6±7.2 years. RC distribution of DR severity: no DR 70.5%, mild NPDR 9.3%, moderate NPDR 7.6%, severe NPDR 3.7%, PDR 2.9%, ungradable 6.0%. DME severity: no DME 82.2%, DME 6.2%, ciDME 4.2%, ungradable 7.3%. RefDR was present in 13.8% and vtDR in 7.8% of eyes. Images were ungradable for DR or DME in 7.7% by RC and 20.7% by AI. Table 1 summarizes the SN/SP and operating characteristics of POC AI grading. SN/SP of AI grading compared to RC evaluation was 0.83/0.95 for refDR and 0.96/0.91 for vtDR. 4 eyes with vtDR (3 severe NPDR and 1 proliferative DR) were missed by AI. Comparisons of performance of the POC AI with existing FDA approved algorithms are presented in table 2.ConclusionsThis study demonstrates that POC AI following a defined retinal imaging protocol at the time of imaging has SN/SP for refDR that meets the current acceptable thresholds of 0.80 and 0.95. Integrating AI at the POC could substantially reduce centralized reading center burden and speeds information delivery to the patient, allowing more prompt eye care referral.",

author = "Recivall Salongcay and Aquino, {Lizzie Anne} and Salva, {Claude Michael} and Alog, {Glenn Paulo} and Kaye Locaylocay and {Viguilla Saunar}, Aileen and Jacoba, {Cris Martin P.} and Sun, {Jennifer K.} and Aiello, {Lloyd P.} and Tunde Peto and Silva, {Paolo S.}",

year = "2022",

month = jun,

day = "1",

language = "English",

volume = "63",

pages = "1159",

journal = "Investigative Opthalmology and Visual Science",

issn = "0146-0404",

publisher = "Association for Research in Vision and Ophthalmology Inc.",

number = "7",

note = "Association for Research in Vision and Ophthalmology Annual Meeting 2022, ARVO 2022 ; Conference date: 01-05-2022 Through 04-05-2022",

url = "https://www.arvo.org/annual-meeting/",

}

Salongcay, R, Aquino, LA, Salva, CM, Alog, GP, Locaylocay, K, Viguilla Saunar, A, Jacoba, CMP, Sun, JK, Aiello, LP, Peto, T & Silva, PS 2022, 'Accuracy of integrated Artificial Intelligence (AI) grading at the point of care (POC) using handheld retinal imaging in a community-based diabetic retinopathy (DR) screening program (DRSP)', Investigative Opthalmology and Visual Science, vol. 63, no. 7, pp. 1159. <https://iovs.arvojournals.org/article.aspx?articleid=2782793>

Accuracy of integrated Artificial Intelligence (AI) grading at the point of care (POC) using handheld retinal imaging in a community-based diabetic retinopathy (DR) screening program (DRSP). / Salongcay, Recivall; Aquino, Lizzie Anne; Salva, Claude Michael et al.
In: Investigative Opthalmology and Visual Science, Vol. 63, No. 7, 01.06.2022, p. 1159.

Research output: Contribution to journal › Meeting abstract › peer-review

TY - JOUR

T1 - Accuracy of integrated Artificial Intelligence (AI) grading at the point of care (POC) using handheld retinal imaging in a community-based diabetic retinopathy (DR) screening program (DRSP)

AU - Salongcay, Recivall

AU - Aquino, Lizzie Anne

AU - Salva, Claude Michael

AU - Alog, Glenn Paulo

AU - Locaylocay, Kaye

AU - Viguilla Saunar, Aileen

AU - Jacoba, Cris Martin P.

AU - Sun, Jennifer K.

AU - Aiello, Lloyd P.

AU - Peto, Tunde

AU - Silva, Paolo S.

PY - 2022/6/1

Y1 - 2022/6/1

N2 - PurposeTo prospectively evaluate handheld retinal images assessed by AI at the time of imaging as compared to standard retinal image graders at a centralized reading center (RC).MethodsProspective comparative study of AI assessment of referable DR [(refDR) moderate nonproliferative DR (NPDR) or worse, or any level of diabetic macular edema (DME)] and vision threatening DR [(vtDR) severe NPDR or worse, or any level of center involving DME (ciDME)]. AI assessment of disc and macular images performed at the time of imaging was compared with RC evaluation of validated 5-field handheld retinal images [(5F) disc, macula, temporal, superior and inferior]. RC evaluation of the 5F images followed the international DR/DME classification. Sensitivity and specificity (SN/SP) for ungradable images, refDR and vtDR were calculated.Results1,733 eyes from 869 diabetic (DM) patients were enrolled in the study. Cohort demographic: age 59.5±10.0, 64.7% female, 98.4% type 2, DM duration 6.6±7.2 years. RC distribution of DR severity: no DR 70.5%, mild NPDR 9.3%, moderate NPDR 7.6%, severe NPDR 3.7%, PDR 2.9%, ungradable 6.0%. DME severity: no DME 82.2%, DME 6.2%, ciDME 4.2%, ungradable 7.3%. RefDR was present in 13.8% and vtDR in 7.8% of eyes. Images were ungradable for DR or DME in 7.7% by RC and 20.7% by AI. Table 1 summarizes the SN/SP and operating characteristics of POC AI grading. SN/SP of AI grading compared to RC evaluation was 0.83/0.95 for refDR and 0.96/0.91 for vtDR. 4 eyes with vtDR (3 severe NPDR and 1 proliferative DR) were missed by AI. Comparisons of performance of the POC AI with existing FDA approved algorithms are presented in table 2.ConclusionsThis study demonstrates that POC AI following a defined retinal imaging protocol at the time of imaging has SN/SP for refDR that meets the current acceptable thresholds of 0.80 and 0.95. Integrating AI at the POC could substantially reduce centralized reading center burden and speeds information delivery to the patient, allowing more prompt eye care referral.

AB - PurposeTo prospectively evaluate handheld retinal images assessed by AI at the time of imaging as compared to standard retinal image graders at a centralized reading center (RC).MethodsProspective comparative study of AI assessment of referable DR [(refDR) moderate nonproliferative DR (NPDR) or worse, or any level of diabetic macular edema (DME)] and vision threatening DR [(vtDR) severe NPDR or worse, or any level of center involving DME (ciDME)]. AI assessment of disc and macular images performed at the time of imaging was compared with RC evaluation of validated 5-field handheld retinal images [(5F) disc, macula, temporal, superior and inferior]. RC evaluation of the 5F images followed the international DR/DME classification. Sensitivity and specificity (SN/SP) for ungradable images, refDR and vtDR were calculated.Results1,733 eyes from 869 diabetic (DM) patients were enrolled in the study. Cohort demographic: age 59.5±10.0, 64.7% female, 98.4% type 2, DM duration 6.6±7.2 years. RC distribution of DR severity: no DR 70.5%, mild NPDR 9.3%, moderate NPDR 7.6%, severe NPDR 3.7%, PDR 2.9%, ungradable 6.0%. DME severity: no DME 82.2%, DME 6.2%, ciDME 4.2%, ungradable 7.3%. RefDR was present in 13.8% and vtDR in 7.8% of eyes. Images were ungradable for DR or DME in 7.7% by RC and 20.7% by AI. Table 1 summarizes the SN/SP and operating characteristics of POC AI grading. SN/SP of AI grading compared to RC evaluation was 0.83/0.95 for refDR and 0.96/0.91 for vtDR. 4 eyes with vtDR (3 severe NPDR and 1 proliferative DR) were missed by AI. Comparisons of performance of the POC AI with existing FDA approved algorithms are presented in table 2.ConclusionsThis study demonstrates that POC AI following a defined retinal imaging protocol at the time of imaging has SN/SP for refDR that meets the current acceptable thresholds of 0.80 and 0.95. Integrating AI at the POC could substantially reduce centralized reading center burden and speeds information delivery to the patient, allowing more prompt eye care referral.

M3 - Meeting abstract

SN - 0146-0404

VL - 63

SP - 1159

JO - Investigative Opthalmology and Visual Science

JF - Investigative Opthalmology and Visual Science

IS - 7

T2 - Association for Research in Vision and Ophthalmology Annual Meeting 2022

Y2 - 1 May 2022 through 4 May 2022

ER -

Accuracy of integrated Artificial Intelligence (AI) grading at the point of care (POC) using handheld retinal imaging in a community-based diabetic retinopathy (DR) screening program (DRSP)

Abstract

UN SDGs

Access to Document

Fingerprint

Cite this