June 2023
Volume 64, Issue 8
Open Access
ARVO Annual Meeting Abstract  |   June 2023
Tear Cytokine-based Predictive Models of Dry Eye Disease Diagnosis and Severity: a Tool for Clinical Trials
Author Affiliations & Notes
  • Margarita Calonge
    IOBA, Universidad de Valladolid, Valladolid, Castilla y León, Spain
  • Itziar Fernández
    IOBA, Universidad de Valladolid, Valladolid, Castilla y León, Spain
  • Alberto López-Miguel
    IOBA, Universidad de Valladolid, Valladolid, Castilla y León, Spain
  • María J. González-García
    IOBA, Universidad de Valladolid, Valladolid, Castilla y León, Spain
  • Carmen Garcia-Vazquez
    IOBA, Universidad de Valladolid, Valladolid, Castilla y León, Spain
  • Amalia Enriquez-De-Salamanca
    IOBA, Universidad de Valladolid, Valladolid, Castilla y León, Spain
  • Footnotes
    Commercial Relationships   Margarita Calonge Thea Laboratories, Novartis, Code C (Consultant/Contractor), Santen, Code F (Financial Support); Itziar Fernández None; Alberto López-Miguel None; María J. González-García None; Carmen Garcia-Vazquez None; Amalia Enriquez-De-Salamanca None
  • Footnotes
    Support  PDC2021-121035-I00 funded by MCIN/AEI/ 10.13039/501100011033 and by the “European Union NextGenerationEU/PRTR”.
Investigative Ophthalmology & Visual Science June 2023, Vol.64, 3974. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Margarita Calonge, Itziar Fernández, Alberto López-Miguel, María J. González-García, Carmen Garcia-Vazquez, Amalia Enriquez-De-Salamanca; Tear Cytokine-based Predictive Models of Dry Eye Disease Diagnosis and Severity: a Tool for Clinical Trials. Invest. Ophthalmol. Vis. Sci. 2023;64(8):3974.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

Purpose : To develop predictive models for dry eye (DE) diagnosis and severity based on levels of tear cytokines.

Methods : Clinical and molecular data from 406 subjects (mean age of 53.2±15.9 years) including healthy (controls) and DE patients were used. Patients were clinically classified as predisposed DE (pDE), mild-moderate DE (mDE), and severe DE (sDE) according to published guidelines. Basal tear samples were collected by capillarity and 28 cytokines were analyzed by multiplex immune bead-based assays. Data comparison among groups were analyzed with one way-ANOVA. Post-hoc pairwise analysis was done with student t-test for independent samples or Welch test for two samples. False positive rate was controlled using Benjamini-Hochberg corrected p-values for multiple comparisons. Discriminant diagnostic models were fitted using the support vector machine and selection of the most relevant cytokines for each model was based on the minimum redundancy-maximum relevance algorithm. Data from 271 subjects were used for fitting the models and data from 135 subjects were used for external validation. Receiver operating characteristic curve analysis (AUC) was used to assess the discriminant ability of the fitted models.

Results : We classified 54 subjects as controls, 136 as pDE, 185 as mDE, and 31 as sDE. Compared to controls, DE patients had levels of GRO, IL-1RA, IL-8, IL-12p70, MCP-1, MMP-9 and RANTES significantly elevated. pDE had significantly increased levels of IL-1RA and decreased levels of IL-1b, IL-6, IL-10, and TNF-a compared to controls. sDE had significantly increased levels of GRO, IL-1RA, MMP-9 and VEGF compared to mDE. Best discriminative models included: 1) IL-6 for controls+pDE vsmDE+sDE; 2) IL-4 for controls+pDE vs pDE; and 3) and MMP-9, EGF, and IL-1RA for mDE vs sDE patients. AUC values were >0.6, sensitivity was 60-97%, specificity was 71-92.9%, and accuracy was 68-94.2% All models had an AUC value >0.5 when externally validated. Classification was particularly good (85% accuracy) at identifying mDE patients.

Conclusions : The developed tear cytokine-based predictive models can differentiate DE patients form healthy subjects and can also classify different DE stages. Then, they could be an accurate tool to increase success in DE clinical trials, specifically to improve recruitment uniformity and evaluation endpoints.

This abstract was presented at the 2023 ARVO Annual Meeting, held in New Orleans, LA, April 23-27, 2023.

×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×