Purchase this article with an account.
J. Peter Campbell, Sang Kim, Ryan Swan, Karyn Elizabeth Jonas, Susan Ostmo, Stratis Ioannidis, Deniz Erdogmus, Jayashree Kalpathy-Cramer, Robison Vernon Paul Chan, Michael F Chiang; Is there clinical utility for a continuous severity score for plus disease in ROP?. Invest. Ophthalmol. Vis. Sci. 2017;58(8):4737.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
The International Classification for ROP (ICROP) has defined ROP in terms of zone, stage, and plus disease. However, each of these have been shown to demonstrate high levels of inter-observer variability. We recently published data showing that inter-expert discrepancies in plus disease diagnosis (plus vs. pre-plus vs. normal) are due to differences in cut-points for disease severity along a continuous spectrum of vascular abnormality. The purpose of this study is to propose a more continuous 9-point ROP severity scale, and to evaluate its clinical utility as a screening and diagnostic tool.
281 infants were identified as part of a multi-center, prospective, ROP cohort study. A 9 level ROP plus disease severity score was determined for each image by averaging the image-based disease classifications (plus vs. pre-plus vs. normal) by 3 graders masked to the clinical ophthalmoscopic diagnosis. Levels 1-3 (all with unanimously “normal” vessels) were subdivided based on the stage classification. We analyzed the relationship between the ROP severity score and ICROP classifications, including referral-warranted ROP (RWROP, defined as zone I or stage 3 or plus disease) and treatment-requiring ROP (defined as type 1 ROP). Area under the receiver operating characteristic curve (AUC) scores were calculated for the ROP severity score as a screening tool for type 1 ROP.
1553 study eye examinations from 281 infants were included in the study. 249 examinations were classified as RWROP, and 79 as type 1 ROP. The Table displays the percentage in each severity scale with zone I, stage 3, plus disease, or RWROP. Figure 1 displays the percentage with type 1 ROP by severity score. The AUC for detection of type 1 disease was 0.95.
This 9 point ROP severity score is associated with clinically significant ROP and demonstrates promise as a screening and diagnostic tool for ROP. In this analysis, the severity score was determined by averaging multiple plus disease classifications. We have shown that similar relative scores may be obtained using pairwise comparisons (assigning a numerical score based on relative comparison to standard images) and using computer-based image analysis (CBIA). In the future, these quantitative methods have potential to improve the objectivity of plus disease diagnosis, improve monitoring of disease progression, and provided added value to telemedicine screening programs.
This is an abstract that was submitted for the 2017 ARVO Annual Meeting, held in Baltimore, MD, May 7-11, 2017.
This PDF is available to Subscribers Only