Full text
NTDscope: A multi-contrast portable microscope for disease diagnosis Díaz de León Derby et al. Figures Fig 1. NTDscope design, illumination, and resolution. A: Diagram showing NTDscope and straight channel capillary external dimensions. B: Exploded view of NTDscope and main components. C: NTDscope illumination modes, showing brightfield, darkfield, and fluorescence illumination. D: NTDscope resolution measurements, showing design of custom slide with USAF resolution targets (D.i.), image of USAF resolution target array acquired on NTDscope used to measure resolution, with the resolution targets used in calculations color coded with pink (center), yellow (inner ring) and green (outer ring) (D.ii.), close up of center resolution target (D.iii.), and resolution measurements at each target location, showing root mean square of horizontal and vertical resolution measurements (D.iv.). Data was processed and visualized using GraphPad Prism 10.
Fig 2. NTDscope for live imaging of microfilariae in blood and plasma samples. A: Time series showing the movement of a Brugia malayi microfilaria in plasma of an infected cat. B: (top) image of a capillary filled with peripheral blood of a human patient infected with Loa loa, collected and imaged in Cameroon, (middle) screenshots of the videos corresponding to the 7 FOVs imaged for the capillary, (bottom) diQerence images generated from the 7 videos. The patient had calibrated thick smear counts of 4001 mf/ml. A video of the middle (highlighted) FOV is shown in Supplementary Video 1. C: DiQerence images generated from videos from four diQerent patients in Cameroon, three of whom were infected with Loa loa mf (thick smear counts 115-1047 mf/ml) and one patient who was negative for Loa loa. D: Sensitivity and specificity of manual identification of microfilariae on videos of patient peripheral blood acquired on the NTDscope in Gabon, compared to the presence of microfilariae of Loa loa and Mansonella perstans in calibrated thick smears. E: Snapshots of videos and corresponding diQerence images of samples containing M. perstans (top) and B. malayi (bottom) mf. The mf identified in videos and diQerence images are circled in green. Data was processed and visualized using GraphPad Prism 10.
Fig 3. NTDscope can image parasitic eggs in urine and stool. A: Images of Schistosoma haematobium eggs from a urine sample of a patient in Côte d’Ivoire, acquired using brightfield (BF) and darkfield (DF) imaging. B: Image of soil-transmitted helminth (STH) eggs from a patient stool sample processed using a flotation-based assay and imaged using the NTDscope in Cameroon. Insets showing Trichuris trichiura (left) and Ascaris lumbricoides (right) eggs captured in diQerent locations of the NTDscope tapered capillary. C: (top) images of A. lumbricoides eggs from a patient stool sample, processed with a flotation solution and imaged using BF and DF illumination on the NTDscope in Bangladesh. (bottom) Schistosoma japonicum and Schistosoma mansoni eggs isolated from hamster livers and imaged on the NTDscope using BF and DF illumination. All patient samples were confirmed as positive via gold standard microscopy.
Fig 4. The NTDscope can image parasites on fixed and stained slides. A: (top left) picture of a thick smear capillary designed to be imaged with the NTDscope, including 30μL of blood from a cat infected with Brugia malayi. (middle left) Thick smear capillary after Giemsa staining. (bottom) stitched images of Giemsa-stained capillary, collected with the NTDscope, with an inset showing six B. malayi microfilariae (top right). The full video is shown in Supplementary Video 3. B: Images of fixed and stained slides with parasites, collected on the NTDscope, showing (left) Entamoeba histolytica, (middle) Trichomonas vaginalis, and (right) Leishmania donovani promastigotes.
Fig 5. The NTDscope can be used to image lateral flow test strips and CRISPR-Cas molecular assays. A: Lateral flow assays imaged on the NTDscope. A.i. Time series (15 min) of the test line on a COVID-19 antigen test, tested on a positive patient. A.ii. Measurement of the dye intensity over time, of the test line shown in A.i., calculated by dividing the mean pixel intensity of a rectangular region of interest (ROI) over the test line by an ROI outside of the test line. Pixel intensity was measured using Fiji (ImageJ). A.iii. End-point images after 15 minutes for three COVID-19 tests (2 positive, 1 negative), showing the test and control lines. A.iv. Scan of the entire COVID-19 test strip, showing the control and test bands for a positive patient after 15 minutes. B: Results of a CRISPR-Cas13 assay run on NTDscope. B.i. Measurement of fluorescence intensity over time for a Cas13 assay with a guide RNA specific to SARS-CoV-2, in the presence (SARS-CoV-2 1e6 copies/μL) and absence (non-target control — NTC) of synthetic SARS-CoV-2 RNA. Fluorescence intensity was measured during the first 60 minutes of assay time; measurements were acquired every 10 minutes. The data from the first 10 minutes was discarded to allow assay components to equilibrate to temperature. Experiment performed in triplicate. B.ii. Slopes of the curves calculated by performing simple linear regression of the data for each replicate of B.i., slopes of the positive samples were compared to the no target RNA controls (NTC) using an unpaired ttest. * is p≤0.05. Data were processed and visualized using GraphPad Prism 10.
Supplementary Figure 2. CRISPR-Cas13a assay results on plate reader. i. Measurement of fluorescence intensity over time for a Cas13 assay with a guide RNA specific to SARS-CoV-2, in the presence (SARS-CoV-2 1e6 copies/μL) and absence (non-target control — NTC) of synthetic SARS-CoV-2 RNA. Fluorescence intensity was measured on a Tecan Spark plate reader during the first 60 minutes of assay time; measurements were acquired every 10 minutes. The data from the first 10 minutes was discarded to allow assay components to equilibrate to temperature. Experiment performed in triplicate. ii. Slopes of the curves calculated by performing simple linear regression of the data for each replicate of i., slopes of the positive samples were compared to the no target RNA controls (NTC) using an unpaired t-test. Data are expressed as mean ± standard error of the mean (SE). * is p<0.05; ** is p<0.01; *** is p<0.001; **** is p<0.0001. Data was processed and visualized using GraphPad Prism 10 Supplementary Figure 1. Comparison of manual counts of microfilarial movement in videos to counts of Brugia malayi microfilariae in calibrated thick smears. Comparison of B. malayi microfilariae counted manually from videos acquired on the NTDscope and microfilariae counts in calibrated thick smears, from a blood sample of a cat infected with B. malayi
