Mobile nucleic acid amplification testing (mobiNAAT) for Chlamydia trachomatis screening in hospital emergency department settings

Abstract

Management of curable sexually-transmitted infections (STI) such as Chlamydia can be revolutionized by highly sensitive nucleic acid testing that is deployable at the point-of-care (POC). Here we report the development of a mobile nucleic acid amplification testing (mobiNAAT) platform utilizing a mobile phone and droplet magnetofluidics to deliver NAAT in a portable and accessible format. By using magnetic particles as a mobile substrate for nucleic acid capture and transport, fluid handling is reduced to particle translocation on a simple magnetofluidic cartridge assembled with reagents for nucleic acid purification and amplification. A mobile phone user interface operating in tandem with a portable Bluetooth-enabled cartridge-processing unit facilitates process integration. We tested 30 potentially Chlamydia trachomatis (CT)-infected patients in a hospital emergency department and confirmed that mobiNAAT showed 100% concordance with laboratory-based NAAT. Concurrent evaluation by a nontechnical study coordinator who received brief training via an embedded mobile app module demonstrated ease of use and reproducibility of the platform. This work demonstrates the potential of mobile nucleic acid testing in bridging the diagnostic gap between centralized laboratories and hospital emergency departments.

Figure 1

Technology overview. (a) Photograph of droplet magnetofluidic cartridge next to US quarter for comparison. (b) Time-lapse sequence of particle transport on cartridge. DNA is captured on the magnetic particle surface in (i), followed by successive rinsing steps in wash buffers (ii) and (iii), and subsequent elution and amplification in LAMP reagent (iv). (c) CAD layout of components inside the cartridge-processing unit. The unit integrates several subsystems to facilitate mechanical manipulation for magnetic particle handling, thermal control for assay initiation and optical signal acquisition with the assistance of a mobile phone. (d) Functional overview of instrumentation. The mobile phone interacts with the instrument in two ways: first, the embedded CMOS camera sensor is used to collect signal via an optical signal relay chain, in which the LED illuminates the reaction chamber to generate fluorescent signal that is relayed by a mirror and filtered prior to being magnified by a combination of an external imaging lens and an embedded camera lens on the phone before reaching the sensor. Secondly, the mobile phone facilitates wireless communication with the instrument’s on-board microcontroller to coordinate magnetic particle transport and thermal incubation.

Figure 2

Assay design and sensitivity characterization. (a) Schematic of single-stream LAMP assay as implemented on the mobiNAAT platform. Magnetic particles capture nucleic acid targets from sample lysate via electrostatic interaction, where the acidic pH maintained by the binding buffer (i) generates a positive charge on the polyhistidine-coated bead surface. Affinity between particles and nucleic acids is maintained at acidic pH (<pH 5) of the wash buffers (ii), which is reversed upon entry in the LAMP amplification mixture (iii) with a basic pH of 8.5. Subsequent amplification (iv) generates a green fluorescent signal based on the complexometric indicator calcein, contrasting with unamplified reagent (v). (b) Comparison among amplification of dilutions of synthetic DNA spiked into reaction mixture. A fraction of the replicate reactions were amplified for 10² copies of molecular target input (red marker, 1/3), indicating analytical sensitivity of 10²–10³ copies of molecular target. Shaded areas represent upper and lower bounds of time threshold (n = 3). (c) Evaluation of magnetic particle-based DNA capture, elution and subsequent amplification as a function of DNA quantity in lysis/binding buffer. Time thresholds obtained indicate sensitivity and process efficiency that are comparable to the standard LAMP process without bead coupling. Shaded areas represent upper and lower bounds of time threshold (n = 3).

Figure 3

Signal acquisition and platform validation study. (a) Image processing algorithm on the mobile phone. Initially, image is acquired at the start of the incubation (time = 0). Signal acquisition is performed using an optical pre-processing signal relay chain as described in Fig. 1c. Subsequent images are captured and segmented to extract information from a single row of pixels across the incubation chamber. Only the green value of the RGB data is used. Data acquired at each time point are normalized by subtracting segmented data from time = 0. Afterwards, the area under the resultant vector is integrated in order to yield a single value, which is used either for end-point analysis or plotted sequentially over time for real-time monitoring. (b) Real-time monitoring of select deblinded clinical samples using time-lapse image acquisition. Positive samples show development of signal and are clearly differentiated from signal developed by negative samples beginning at approximately 20 minutes of incubation. Shaded areas represent upper and lower bounds of AUC fluorescence (n = 3). Bottom panel shows image of fluorescence developing over time in a positive control reaction in droplet cartridge. (c) Reproducibility test showing signal generated from chlamydia DNA target-positive and negative cartridges (n = 10 each). Welch’s two sample t-test yields p < 0.000001. (d) Validation of the mobiNAAT platform using blinded vaginal swab samples obtained from an equal number of chlamydia-positive and -negative individuals (n = 20). AUC fluorescence levels of swab samples that are positive (blue) or negative (orange) are shown. Welch’s two sample t-test yields p < 0.00001. Classification threshold (red dotted line) is established using signals obtained from a set of control samples (n = 10).

Figure 4

Platform evaluation at ED. (a) Comparison of standard and mobiNAAT workflows in the emergency room. In a standard workflow, the vaginal swab acquired during a pelvic exam by the provider is collected by a study coordinator, who transports the sample to the microbiology laboratory. The sample is then barcoded and batch processed to be tested using the standard-of-care NAAT assay. Afterwards, the test result is made available on the electronic medical record (EMR) accessible by the provider. In the mobiNAAT workflow, the test is performed directly in the emergency room via workflow as described in Fig. 2a, and the result is made available directly on the mobile device. Standard approach via central laboratory yielded >72 hours due to downtime inherent to batch processing, whereas mobiNAAT-enabled testing in the emergency room yields diagnostic results in approximately 1 hour. (b) Comparison of results obtained using the mobiNAAT platform using clinical samples collected at the emergency room (n = 30). Samples were tested in parallel using the Gen-Probe Aptima Combo 2 assay for verification (positives marked by red bars; negatives marked by green bars; instrument data in Fig. S8). Among all patient samples tested, two patients (ID 7 and 9) reported positive in the verification assay, in agreement with the result obtained using the mobiNAAT platform. Classification threshold (red dotted line) is established using signals obtained from a set of control negative samples (n = 15) (Fig. S9). (c) Comparison of results obtained using mobiNAAT platform by the developers and a naïve operator trained using the mobile app (n = 13). Patients 5–17 were tested in parallel by a clinical staff member at the emergency room (green markers) and showed full correspondence with the results obtained by the developers (11 negative, 2 positive).