Automated Detection of Malarial Retinopathy in Digital Fundus Images for Improved Diagnosis in Malawian Children with Clinically Defined Cerebral Malaria

Vinayak Joshi, Carla Agurto, Simon Barriga, Sheila Nemeth, Peter Soliz, Ian J. MacCormick, Susan Lewallen, Terrie E. Taylor, Simon P. Harding
2017 Scientific Reports  
Cerebral malaria (CM), a complication of malaria infection, is the cause of the majority of malariaassociated deaths in African children. The standard clinical case definition for CM misclassifies ~25% of patients, but when malarial retinopathy (MR) is added to the clinical case definition, the specificity improves from 61% to 95%. Ocular fundoscopy requires expensive equipment and technical expertise not often available in malaria endemic settings, so we developed an automated software system
more » ... o analyze retinal color images for MR lesions: retinal whitening, vessel discoloration, and white-centered hemorrhages. The individual lesion detection algorithms were combined using a partial least square classifier to determine the presence or absence of MR. We used a retrospective retinal image dataset of 86 pediatric patients with clinically defined CM (70 with MR and 16 without) to evaluate the algorithm performance. Our goal was to reduce the false positive rate of CM diagnosis, and so the algorithms were tuned at high specificity. This yielded sensitivity/specificity of 95%/100% for the detection of MR overall, and 65%/94% for retinal whitening, 62%/100% for vessel discoloration, and 73%/96% for hemorrhages. This automated system for detecting MR using retinal color images has the potential to improve the accuracy of CM diagnosis. Cerebral malaria (CM) is an often-fatal clinical complication associated with Plasmodium falciparum (Pf) malaria infection. The 2015 World Malaria Report estimates that malaria affected over 214 million people, and claimed the lives of 438,000 people worldwide, about 67% of whom were African children under 5 years of age; the majority of these pediatric deaths was attributed to CM 1 . Annually, CM results in the loss of 35.4 million disability adjusted life years due to mortality and morbidity of clinically diagnosed CM patients in Africa 1,2 . The standard WHO clinical case definition of CM (physical symptoms such as seizures, convulsions; Blantyre coma score <2, detection of Pf parasites in the peripheral blood, and exclusion of other causes of coma (e.g. meningitis, hypoglycemia) refs 3,4); may misclassify up to 25% of cases 5 . Because asymptomatic Pf infection is common parasitemia in malaria-endemic areas, CM is often over-diagnosed. The WHO clinical case definition overlooks other non-malarial illnesses with similar clinical symptoms "masquerading" as CM. The result is a significant false positive rate (FPR) for CM which leads to inaccurate epidemiological estimates of disease incidence, incorrect classifications in research and clinical settings, and potentially inadequate treatment of the underlying illness 6-8 . If non-malarial illnesses with incidental parasitaemia go undiagnosed and thus improperly treated, potentially avoidable death or neurologic disabilities may result. In summary, though the WHO criteria are highly sensitive
doi:10.1038/srep42703 pmid:28198460 pmcid:PMC5309896 fatcat:gsok7g3z7vhgvfajmzihz34a2u