Validation of a Photographic Retinopathy of Prematurity Severity Score (ROP–SS) and Comparison to the Modified Activity Score (mROP–ActS) in a Retrospective SUNDROP Cohort

Abstract

Purpose : Telemedicine for identification and treatment of retinopathy of prematurity (ROP) is becoming ubiquitous, necessitating a grading system to gauge ROP disease severity. The modified ROP Activity Scale (mROP-ActS) addresses the totality of binocular burden via indirect ophthalmoscopy. There is the unmet need for an alternative score, which would be used to facilitate ROP identification and to gauge the level of disease improvement or deterioration specifically on photographic telemedicine exams. Here, we propose a Photographic ROP scoring system (ROP Severity Score [ROP-SS]) which we have validated and compared against the modified mROP-ActS system.

Methods : The SUNDROP database includes data from 12 NICUs over ≥17 years beginning in 2005.We abstracted data from SUNDROP, anonymized and de-identified it and placed into a secure HIPAA-compliant online storage folder.

We assessed the features of ROP that resulted in a treatment intervention and then created an ROP-SS ranging from 0-100, with disproportionate weighting for Zone I, Plus, and Stage 3 (Table 1). We also introduced severity levels, such as "treatment-warranted" which scores from 55-85.

Each eye per visit was scored for both ROP-SS and mROP-ActS. We conducted descriptive analysi and simple arithmetic was applied to determine how often the ROP-SS and mROP-ActS returned a score for the SUNDROP dataset for the following outcomes: 1) overall, 2) treated eyes.

Results : The ROP-SS was able to return a score more often than mROP-ActS. Specifically, the ROP-SS returned a score 99.1% of the time, while the mROP-ActS returned a score 80.8% of the time.

ROP-SS outperformed mROP-ActS in predicting subsequent need for treatment. For TW-ROP, overall, there were 39 treatment cases. Of these cases, the ROP-SS identified correctly 38 of the 39 cases (97.4% accuracy); the mROP-ActS identified correctly 26 of the 39 cases (66.7% accuracy) (Table 2).

Conclusions : Photographic telemedicine is important for ROP screening. Here, we demonstrate the rationale for a photographic ROP Severity Scale, and we propose a scoring system with weighted elements that predict treatment interventions and adverse outcomes. We compare ROP-SS to the mROP-ActS and overall, the ROP-SS returned data at a higher percentage and was more accurate in predicting subsequent treatment as compared to the mROP-ActS.

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