ESCRS - POCAT319 - Quantifying Discrepancies In Subjective Nuclear Cataract Grading: Implications For Predictive Accuracy And Grading Standardization

Quantifying Discrepancies In Subjective Nuclear Cataract Grading: Implications For Predictive Accuracy And Grading Standardization

Published 2026 - 44th Congress of the ESCRS

Reference: POCAT319 | Type: Poster | DOI: 10.82333/nre2-8523

Authors: Dr Jesus Zarallo-Gallardo* 1 , Dr Vicente Miralles-Pechuán 2 , Dr Maria-Begoña Pastor-Nieto 3 , Dr Julio González-Martín-Moro 1 , Dr Rosario Cobo-Soriano 1

1Ophthalmology,Hospital Universitario del Henares,Coslada,Spain;Medicine,Universidad Francisco de Vitoria,Pozuelo de Alarcón,Spain, 2Ophthalmology,Hospital Universitario del Henares,Coslada,Spain, 3Ophthalmology,Hospital Universitario La Paz,Madrid,Spain

Purpose

To evaluate inter- and intra-grader variability in subjective nuclear cataract grading and determine how these discrepancies affect the predictability of effective phacoemulsification time (EPT), with particular emphasis on grading limitations in dense cataracts.

Setting

Department of Ophthalmology, Hospital Universitario del Henares, Coslada, Spain. All surgeries were performed by a single experienced surgeon (18 years’ experience) using a standardized Stellaris® phacoemulsification platform.

Methods

Retrospective study of 774 cataract surgeries (January 2023–December 2024). From 658 Stellaris® cases, 641 with complete effective phacoemulsification time (EPT) records were analyzed. Preoperative nuclear density was graded (LOCS III, Grades 1–6) under pharmacological mydriasis. Data from six high-volume graders (>60 cases each; n=474) were evaluated. Inter-grader variability was assessed by comparing EPT distributions for identical grades. Grading consistency was quantified using SD and IQR of EPT within each grade. An Ordinary Least Squares (OLS) regression model evaluated the predictive value of nuclear grade on EPT.

Results

Visual distribution analysis demonstrated variability in EPT across graders for identical nuclear grades, indicating differences in grading thresholds. Intra-grader consistency was heterogeneous and grade-dependent: some graders showed high precision in low-density cataracts (e.g., Grader M04 at Grade 1: SD=0.028, IQR=0.020), whereas others exhibited marked dispersion in advanced cataracts (e.g., Grader M02 at Grade 5: SD=2.895, IQR=2.725).

Nuclear grade was a significant predictor of EPT (β=1.1708; 95% CI [1.008–1.334]; p<0.001), explaining 29.6% of the variance (R²=0.296; F=198.8; p<0.001). Residuals were non-normally distributed, suggesting the presence of additional determinants beyond subjective grading.

Conclusions

Subjective LOCS III grading demonstrates significant inter-grader heterogeneity and grade-dependent intra-grader inconsistency, particularly in dense cataracts where accurate energy prediction is most relevant. The moderate explanatory power observed suggests an intrinsic limitation of ordinal subjective grading scales when applied to a continuous surgical energy outcome. These findings support the integration of objective densitometric systems. In their absence, more granular classification scales or standardized high-quality reference imaging may enhance grading consistency in advanced nuclear sclerosis.