AUTHOR=Shen Zhan , Hu Bohao , Tao Liyuan , Ma Jiao , Peng Rongmei , Zhao Yinghan , Hong Jing 

TITLE=Development and validation of a diagnostic model for the identification of chronic ocular graft-versus-host disease (oGVHD)

JOURNAL=Frontiers in Medicine

VOLUME=Volume 10 - 2023

YEAR=2023

URL=https://www.frontiersin.org/journals/medicine/articles/10.3389/fmed.2023.1277194

DOI=10.3389/fmed.2023.1277194

ISSN=2296-858X

ABSTRACT=Purpose: By comparing the clinical manifestations and clinical parameters of patients with chronic ocular graft-versus-host disease (oGVHD) and those without chronic oGVHD, we aim to To verify the International Chronic Ocular Graft-Versus-Host Disease (ICCGVHD) Group diagnostic criteria and establish an easy-to-use and reliable diagnosis model for quick identification of chronic oGVHD. Methods: This study included 180 patients (355 eyes) who underwent allogeneic hematopoietic stem cell transplantation (allo-HSCT) and visited the Peking University Third Hospital Cornea and Ocular Surface Disease Specialist Clinic from July 2020 to February 2021. The proportion of chronic oGVHD was 76.06% (279/355). Results: Five complaints, including eye dryness, photophobia, foreign body sensation, eye redness, and burning sensation; six ophthalmic examinations, including Ocular Surface Disease Index (OSDI) score, corneal fluorescein staining (CFS), tear break-up time (TBUT), Schirmer's test score without anesthesia, conjunctival score, tear meniscus height, and non-ocular GVHDinvolved organs were significantly different between patients with chronic oGVHD and control group(P<0.05). Binary logistic regression (backward LR algorithm) selection demonstrated that three variables retained diagnostic significance for chronic oGVHD: CFS (OR=2.71(1.92-3.81), P<0.001), Schirmer's test score without anesthesia (OR=0.83(0.76-0.91), P<0.001), and 3 conjunctival score (OR=1.96(1.13-3.42), P=0.031). A nomogram for the identification of chronic oGVHD was developed, and its performance was examined using an internal validation cohort (118 eyes). The areas under the curve (AUCs) for the three-variable-based nomogram were 0.976 (95% CI (0.959-0.992), P<0.01) and 0.945 (95% CI (0.904-0.986), P<0.01) in the development and internal validation cohorts, respectively. Conclusion: This concise three-variable-based nomogram based on ICCGVHD criteria could serve as an easy-to-use and reliable tool for rapid screening of chronic oGVHD.