Field validation of a magneto-optical detection device (Gazelle) for portable point-of-care Plasmodium vivax diagnosis

Abstract

A major challenge for malaria is the lack of tools for accurate and timely diagnosis in the field which are critical for case management and surveillance. Microscopy along with rapid diagnostic tests are the current mainstay for malaria diagnosis in most endemic regions. However, these methods present several limitations. This study assessed the accuracy of Gazelle, a novel rapid malaria diagnostic device, from samples collected from the Peruvian Amazon between 2019 and 2020. Diagnostic accuracy was compared against microscopy and two rapid diagnostic tests (SD Bioline and BinaxNOW) using 18ssr nested-PCR as reference test. In addition, a real-time PCR assay (PET-PCR) was used for parasite quantification. Out of 217 febrile patients enrolled and tested, 180 specimens (85 P. vivax and 95 negatives) were included in the final analysis. Using nested-PCR as the gold standard, the sensitivity and specificity of Gazelle was 88.2% and 97.9%, respectively. Using a cutoff of 200 parasites/μl, Gazelle's sensitivity for samples with more than 200 p/uL was 98.67% (95%CI: 92.79% to 99.97%) whereas the sensitivity for samples lower than 200 p/uL (n = 10) was 12.5% (95%CI: 0.32% to 52.65%). Gazelle's sensitivity and specificity were statistically similar to microscopy (sensitivity = 91.8, specificity = 100%, p = 0.983) and higher than both SD Bioline (sensitivity = 82.4, specificity = 100%, p = 0.016) and BinaxNOW (sensitivity = 71.8%, specificity = 97.9%, p = 0.002). The diagnostic accuracy of Gazelle for malaria detection in P. vivax infections was comparable to light microscopy and superior to both RDTs even in the presence of low parasitemia infections. The performance of Gazelle makes it a valuable tool for malaria diagnosis and active case detection that can be utilized in different malaria-endemic regions.

Fig 1. Study sites (blue circles) located in the city of Iquitos and surrounding communities.

Map created using open data from openstreetmap.org: OpenStreetMap contributors.

Fig 2. Gazelle device testing procedure.

15 μL of whole blood are put on a cartridge with 80 μL of whole blood Gazelle buffer. The sample is placed on the Gazelle device which lyse the sample by sonication and then pass magnets multiple times through the sample. Malaria detection is assessed by comparing the amount of light that traverse a sample with and without a magnetic field.

Fig 3. Sensitivity (blue dots), specificity (red dots) and 95% confidence intervals for detection of malaria by microscopy, SD Bioline, BinaxNOW and Gazelle in P. vivax using 18S rRNA-nested PCR as reference.

Fig 4. ROC analysis.

The figure shows the ROC curves for BinaxNow, SD BIOLINE, microscopy and Gazelle compared with Nested PCR as reference test. Gazelle presented an AUC of 0.93 which is comparable to microscopy with an AUC of 0.96 and higher than BinaxNow (AUC = 0.85) and SD BIOLINE (AUC = 0.91).

Fig 5. Parasite density distribution.

The boxplot shows the distribution of parasite densities estimated by microscopy and PET-PCR along the median and lower and upper quartiles. The optimal parasitemia limit for a Gazelle positive result is shown as a dashed red line.