Use of Machine Learning on Contact Lens Sensor–Derived Parameters for the Diagnosis of Primary Open-angle Glaucoma

Research Consortium

Research output: Contribution to journalArticle

4 Citations (Scopus)

Abstract

Purpose: To test the hypothesis that contact lens sensor (CLS)-based 24-hour profiles of ocular volume changes contain information complementary to intraocular pressure (IOP) to discriminate between primary open-angle glaucoma (POAG) and healthy (H) eyes. Design: Development and evaluation of a diagnostic test with machine learning. Methods: SUBJECTS: From 435 subjects (193 healthy and 242 POAG), 136 POAG and 136 age-matched healthy subjects were selected. Subjects with contraindications for CLS wear were excluded. PROCEDURE: This is a pooled analysis of data from 24 prospective clinical studies and a registry. All subjects underwent 24-hour CLS recording on 1 eye. Statistical and physiological CLS parameters were derived from the signal recorded. CLS parameters frequently associated with the presence of POAG were identified using a random forest modeling approach. MAIN OUTCOME MEASURES: Area under the receiver operating characteristic curve (ROC AUC) for feature sets including CLS parameters and Start IOP, as well as a feature set with CLS parameters and Start IOP combined. Results: The CLS parameters feature set discriminated POAG from H eyes with mean ROC AUCs of 0.611, confidence interval (CI) 0.493–0.722. Larger values of a given CLS parameter were in general associated with a diagnosis of POAG. The Start IOP feature set discriminated between POAG and H eyes with a mean ROC AUC of 0.681, CI 0.603–0.765. The combined feature set was the best indicator of POAG with an ROC AUC of 0.759, CI 0.654–0.855. This ROC AUC was statistically higher than for CLS parameters or Start IOP feature sets alone (both P <.0001). Conclusions: CLS recordings contain information complementary to IOP that enable discrimination between H and POAG. The feature set combining CLS parameters and Start IOP provide a better indication of the presence of POAG than each of the feature sets separately. As such, the CLS may be a new biomarker for POAG.

Original languageEnglish
Pages (from-to)46-53
Number of pages8
JournalAmerican Journal of Ophthalmology
Volume194
DOIs
Publication statusPublished - 1 Oct 2018

Fingerprint

Contact Lenses
Intraocular Pressure
Area Under Curve
Confidence Intervals
ROC Curve
Primary Open Angle Glaucoma
Machine Learning
Healthy Volunteers
Routine Diagnostic Tests
Registries
Biomarkers

ASJC Scopus subject areas

  • Ophthalmology

Cite this

Use of Machine Learning on Contact Lens Sensor–Derived Parameters for the Diagnosis of Primary Open-angle Glaucoma. / Research Consortium.

In: American Journal of Ophthalmology, Vol. 194, 01.10.2018, p. 46-53.

Research output: Contribution to journalArticle

@article{47031e4cd57b42579ee4c513b6e08dad,
title = "Use of Machine Learning on Contact Lens Sensor–Derived Parameters for the Diagnosis of Primary Open-angle Glaucoma",
abstract = "Purpose: To test the hypothesis that contact lens sensor (CLS)-based 24-hour profiles of ocular volume changes contain information complementary to intraocular pressure (IOP) to discriminate between primary open-angle glaucoma (POAG) and healthy (H) eyes. Design: Development and evaluation of a diagnostic test with machine learning. Methods: SUBJECTS: From 435 subjects (193 healthy and 242 POAG), 136 POAG and 136 age-matched healthy subjects were selected. Subjects with contraindications for CLS wear were excluded. PROCEDURE: This is a pooled analysis of data from 24 prospective clinical studies and a registry. All subjects underwent 24-hour CLS recording on 1 eye. Statistical and physiological CLS parameters were derived from the signal recorded. CLS parameters frequently associated with the presence of POAG were identified using a random forest modeling approach. MAIN OUTCOME MEASURES: Area under the receiver operating characteristic curve (ROC AUC) for feature sets including CLS parameters and Start IOP, as well as a feature set with CLS parameters and Start IOP combined. Results: The CLS parameters feature set discriminated POAG from H eyes with mean ROC AUCs of 0.611, confidence interval (CI) 0.493–0.722. Larger values of a given CLS parameter were in general associated with a diagnosis of POAG. The Start IOP feature set discriminated between POAG and H eyes with a mean ROC AUC of 0.681, CI 0.603–0.765. The combined feature set was the best indicator of POAG with an ROC AUC of 0.759, CI 0.654–0.855. This ROC AUC was statistically higher than for CLS parameters or Start IOP feature sets alone (both P <.0001). Conclusions: CLS recordings contain information complementary to IOP that enable discrimination between H and POAG. The feature set combining CLS parameters and Start IOP provide a better indication of the presence of POAG than each of the feature sets separately. As such, the CLS may be a new biomarker for POAG.",
author = "{Research Consortium} and Martin, {Keith R.} and Kaweh Mansouri and Weinreb, {Robert N.} and Robert Wasilewicz and Christophe Gisler and Jean Hennebert and Dominique Genoud and Tarek Shaarawy and Carl Erb and Norbert Pfeiffer and Trope, {Graham E.} and Medeiros, {Felipe A.} and Yaniv Barkana and Liu, {John H.K.} and Robert Ritch and Andr{\'e} Mermoud and Delan Jinapriya and Catherine Birt and Ahmed, {Iqbal I.} and Christoph Kranemann and Peter H{\"o}h and Bernhard Lachenmayr and Yuri Astakhov and Enping Chen and Susana Duch and Giorgio Marchini and Stefano Gandolfi and Marek Rekas and Alexander Kuroyedov and Andrej Cernak and Vicente Polo and Jos{\'e} Belda and Swaantje Grisanti and Christophe Baudouin and Nordmann, {Jean Philippe} and {De Moraes}, {Carlos G.} and Zvi Segal and Moshe Lusky and Haia Morori-Katz and Noa Geffen and Shimon Kurtz and Ji Liu and Budenz, {Donald L.} and Knight, {O'Rese J.} and Mwanza, {Jean Claude} and Anthony Viera and Fernando Castanera and {Che Hamzah}, Jemaima",
year = "2018",
month = "10",
day = "1",
doi = "10.1016/j.ajo.2018.07.005",
language = "English",
volume = "194",
pages = "46--53",
journal = "American Journal of Ophthalmology",
issn = "0002-9394",
publisher = "Elsevier USA",

}

TY - JOUR

T1 - Use of Machine Learning on Contact Lens Sensor–Derived Parameters for the Diagnosis of Primary Open-angle Glaucoma

AU - Research Consortium

AU - Martin, Keith R.

AU - Mansouri, Kaweh

AU - Weinreb, Robert N.

AU - Wasilewicz, Robert

AU - Gisler, Christophe

AU - Hennebert, Jean

AU - Genoud, Dominique

AU - Shaarawy, Tarek

AU - Erb, Carl

AU - Pfeiffer, Norbert

AU - Trope, Graham E.

AU - Medeiros, Felipe A.

AU - Barkana, Yaniv

AU - Liu, John H.K.

AU - Ritch, Robert

AU - Mermoud, André

AU - Jinapriya, Delan

AU - Birt, Catherine

AU - Ahmed, Iqbal I.

AU - Kranemann, Christoph

AU - Höh, Peter

AU - Lachenmayr, Bernhard

AU - Astakhov, Yuri

AU - Chen, Enping

AU - Duch, Susana

AU - Marchini, Giorgio

AU - Gandolfi, Stefano

AU - Rekas, Marek

AU - Kuroyedov, Alexander

AU - Cernak, Andrej

AU - Polo, Vicente

AU - Belda, José

AU - Grisanti, Swaantje

AU - Baudouin, Christophe

AU - Nordmann, Jean Philippe

AU - De Moraes, Carlos G.

AU - Segal, Zvi

AU - Lusky, Moshe

AU - Morori-Katz, Haia

AU - Geffen, Noa

AU - Kurtz, Shimon

AU - Liu, Ji

AU - Budenz, Donald L.

AU - Knight, O'Rese J.

AU - Mwanza, Jean Claude

AU - Viera, Anthony

AU - Castanera, Fernando

AU - Che Hamzah, Jemaima

PY - 2018/10/1

Y1 - 2018/10/1

N2 - Purpose: To test the hypothesis that contact lens sensor (CLS)-based 24-hour profiles of ocular volume changes contain information complementary to intraocular pressure (IOP) to discriminate between primary open-angle glaucoma (POAG) and healthy (H) eyes. Design: Development and evaluation of a diagnostic test with machine learning. Methods: SUBJECTS: From 435 subjects (193 healthy and 242 POAG), 136 POAG and 136 age-matched healthy subjects were selected. Subjects with contraindications for CLS wear were excluded. PROCEDURE: This is a pooled analysis of data from 24 prospective clinical studies and a registry. All subjects underwent 24-hour CLS recording on 1 eye. Statistical and physiological CLS parameters were derived from the signal recorded. CLS parameters frequently associated with the presence of POAG were identified using a random forest modeling approach. MAIN OUTCOME MEASURES: Area under the receiver operating characteristic curve (ROC AUC) for feature sets including CLS parameters and Start IOP, as well as a feature set with CLS parameters and Start IOP combined. Results: The CLS parameters feature set discriminated POAG from H eyes with mean ROC AUCs of 0.611, confidence interval (CI) 0.493–0.722. Larger values of a given CLS parameter were in general associated with a diagnosis of POAG. The Start IOP feature set discriminated between POAG and H eyes with a mean ROC AUC of 0.681, CI 0.603–0.765. The combined feature set was the best indicator of POAG with an ROC AUC of 0.759, CI 0.654–0.855. This ROC AUC was statistically higher than for CLS parameters or Start IOP feature sets alone (both P <.0001). Conclusions: CLS recordings contain information complementary to IOP that enable discrimination between H and POAG. The feature set combining CLS parameters and Start IOP provide a better indication of the presence of POAG than each of the feature sets separately. As such, the CLS may be a new biomarker for POAG.

AB - Purpose: To test the hypothesis that contact lens sensor (CLS)-based 24-hour profiles of ocular volume changes contain information complementary to intraocular pressure (IOP) to discriminate between primary open-angle glaucoma (POAG) and healthy (H) eyes. Design: Development and evaluation of a diagnostic test with machine learning. Methods: SUBJECTS: From 435 subjects (193 healthy and 242 POAG), 136 POAG and 136 age-matched healthy subjects were selected. Subjects with contraindications for CLS wear were excluded. PROCEDURE: This is a pooled analysis of data from 24 prospective clinical studies and a registry. All subjects underwent 24-hour CLS recording on 1 eye. Statistical and physiological CLS parameters were derived from the signal recorded. CLS parameters frequently associated with the presence of POAG were identified using a random forest modeling approach. MAIN OUTCOME MEASURES: Area under the receiver operating characteristic curve (ROC AUC) for feature sets including CLS parameters and Start IOP, as well as a feature set with CLS parameters and Start IOP combined. Results: The CLS parameters feature set discriminated POAG from H eyes with mean ROC AUCs of 0.611, confidence interval (CI) 0.493–0.722. Larger values of a given CLS parameter were in general associated with a diagnosis of POAG. The Start IOP feature set discriminated between POAG and H eyes with a mean ROC AUC of 0.681, CI 0.603–0.765. The combined feature set was the best indicator of POAG with an ROC AUC of 0.759, CI 0.654–0.855. This ROC AUC was statistically higher than for CLS parameters or Start IOP feature sets alone (both P <.0001). Conclusions: CLS recordings contain information complementary to IOP that enable discrimination between H and POAG. The feature set combining CLS parameters and Start IOP provide a better indication of the presence of POAG than each of the feature sets separately. As such, the CLS may be a new biomarker for POAG.

UR - http://www.scopus.com/inward/record.url?scp=85051248326&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85051248326&partnerID=8YFLogxK

U2 - 10.1016/j.ajo.2018.07.005

DO - 10.1016/j.ajo.2018.07.005

M3 - Article

AN - SCOPUS:85051248326

VL - 194

SP - 46

EP - 53

JO - American Journal of Ophthalmology

JF - American Journal of Ophthalmology

SN - 0002-9394

ER -