Dealing with inter-expert variability in retinopathy of prematurity: A machine learning approach

• Inter-expert variability in clinical decision making is an important problem.
• Retinopathy of prematurity is a disease that suffers from inter-expert variability.
• We propose a methodology for understanding the causes of disagreement.
• The methodology provides a framework to identify important features for experts.
• An automatic system was also developed to deal with this problem.

Background and objectiveUnderstanding the causes of disagreement among experts in clinical decision making has been a challenge for decades. In particular, a high amount of variability exists in diagnosis of retinopathy of prematurity (ROP), which is a disease affecting low birth weight infants and a major cause of childhood blindness. A possible cause of variability, that has been mostly neglected in the literature, is related to discrepancies in the sets of important features considered by different experts. In this paper we propose a methodology which makes use of machine learning techniques to understand the underlying causes of inter-expert variability.MethodsThe experiments are carried out on a dataset consisting of 34 retinal images, each with diagnoses provided by 22 independent experts. Feature selection techniques are applied to discover the most important features considered by a given expert. Those features selected by each expert are then compared to the features selected by other experts by applying similarity measures. Finally, an automated diagnosis system is built in order to check if this approach can be helpful in solving the problem of understanding high inter-rater variability.ResultsThe experimental results reveal that some features are mostly selected by the feature selection methods regardless the considered expert. Moreover, for pairs of experts with high percentage agreement among them, the feature selection algorithms also select similar features. By using the relevant selected features, the classification performance of the automatic system was improved or maintained.ConclusionsThe proposed methodology provides a handy framework to identify important features for experts and check whether the selected features reflect the pairwise agreements/disagreements. These findings may lead to improved diagnostic accuracy and standardization among clinicians, and pave the way for the application of this methodology to other problems which present inter-expert variability.