An Enhanced Computer Vision Platform for Clinical Diagnosis of Malaria
- *Corresponding Author:
- Seth Salpeter
Sight Diagnostics Ltd.
Jerusalem Technology Park
Jerusalem 96951, Israel
E-mail: [email protected]
Received Date: Jan 19, 2016; Accepted Date: Mar 2, 2016; Published Date: Mar 9, 2016
Citation: Houri-Yafin A, Eshel Y, Lezmy N, Larbi B, Wypkema E, et al. (2016) An Enhanced Computer Vision Platform for Clinical Diagnosis of Malaria. Malaria Contr Elimination 5:138. doi:10.4172/2470-6965.1000138
Copyright: © 2016 Houri-Yafin A, et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Accurate malaria diagnosis is necessary to prevent unnecessary deaths and curb malaria drug resistance related to unnecessary treatment. While numerous diagnostic assays exist, the need for a low-cost, rapid and highly accurate malaria test remains. Here we evaluate the diagnostic performance of a computer vision platform, the Sight Diagnostic P2 device for malaria diagnosis, speciation and parasite quantification. The trial was conducted at two centers on Plasmodium falciparum and Plasmodium vivax samples, using different testing protocols: 374 samples were collected at City Hospital Mangalore India and 167 samples were collected at Lancet Laboratories Johannesburg South Africa. At City Hospital, the device diagnoses were compared to RT-PCR results while at Lancet Laboratories the device diagnoses were compared to a panel of tests provided by the clinic. For identification of malaria, the device demonstrated a sensitivity of 97% and a specificity of 99.5% at City Hospital India, and a sensitivity of 97.8% and a specificity of 97.5% at Lancet Laboratories Johannesburg. For speciation, the device correctly identified 87.5% for Plasmodium Vivax and 93.5% for Plasmodium Falciparum at City Hospital India. Lastly, comparing the device parasite count with that of trained microscopes, produced an average pearsons correlation of 0.87.