Rapidly adaptable automated interpretation of point-of-care COVID-19 diagnostics

BACKGROUND

Point-of-care diagnostic devices, such as lateral-flow assays, are becoming widely used by the public. However, efforts to ensure correct assay operation and result interpretation rely on hardware that cannot be easily scaled or image processing approaches requiring large training datasets, necessitating large numbers of tests and expert labeling with validated specimens for every new test kit format.

METHODS

We developed a software architecture called AutoAdapt POC that integrates automated membrane extraction, self-supervised learning, and few-shot learning to automate the interpretation of POC diagnostic tests using smartphone cameras in a scalable manner. A base model pre-trained on a single LFA kit is adapted to five different COVID-19 tests (three antigen, two antibody) using just 20 labeled images.

RESULTS

Here we show AutoAdapt POC to yield 99% to 100% accuracy over 726 tests (350 positive, 376 negative). In a COVID-19 drive-through study with 74 untrained users self-testing, 98% found image collection easy, and the rapidly adapted models achieved classification accuracies of 100% on both COVID-19 antigen and antibody test kits. Compared with traditional visual interpretation on 105 test kit results, the algorithm correctly identified 100% of images; without a false negative as interpreted by experts. Finally, compared to a traditional convolutional neural network trained on an HIV test kit, the algorithm showed high accuracy while requiring only 1/50th of the training images.

CONCLUSIONS

The study demonstrates how rapid domain adaptation in machine learning can provide quality assurance, linkage to care, and public health tracking for untrained users across diverse POC diagnostic tests.

A case of Burkholderia pseudomallei mycotic aneurysm linked to exposure in the Caribbean via whole genome sequencing

Melioidosis, an infection caused by Burkholderia pseudomallei, has a very high risk of mortality when treated, with an even higher risk of fatality if undiagnosed or not treated appropriately. It is endemic to Asia, Australia, South America, and the Caribbean; however, the number of melioidosis cases reported in the United States has been increasing. Therefore, physicians should be aware of this clinical entity and its possible presentations. Mycotic aneurysms due to B. pseudomallei are extremely rare accounting for approximately 1-2% of cases. Here we describe a rare case of melioidosis presenting as a mycotic aneurysm in the United States; highlight the potential for diagnostic challenges and epidemiologic concerns; as well as provide a review of mycotic aneurysm cases due to B. pseudomallei published to date.

Rapid Antimicrobial Susceptibility Testing on Clinical Urine Samples by Video-Based Object Scattering Intensity Detection

To combat the ongoing public health threat of antibiotic-resistant infections, a technology that can quickly identify infecting bacterial pathogens and concurrently perform antimicrobial susceptibility testing (AST) in point-of-care settings is needed. Here, we develop a technology for point-of-care AST with a low-magnification solution scattering imaging system and a real-time video-based object scattering intensity detection method. The low magnification (1-2×) optics provides sufficient volume for direct imaging of bacteria in urine samples, avoiding the time-consuming process of culture-based bacterial isolation and enrichment. Scattering intensity from moving bacteria and particles in the sample is obtained by subtracting both spatial and temporal background from a short video. The time profile of scattering intensity is correlated with the bacterial growth rate and bacterial response to antibiotic exposure. Compared to the image-based bacterial tracking and counting method we previously developed, this simple image processing algorithm accommodates a wider range of bacterial concentrations, simplifies sample preparation, and greatly reduces the computational cost of signal processing. Furthermore, development of this simplified processing algorithm eases implementation of multiplexed detection and allows real-time signal readout, which are essential for point-of-care AST applications. To establish the method, 130 clinical urine samples were tested, and the results demonstrated an accuracy of ∼92% within 60-90 min for UTI diagnosis. Rapid AST of 55 positive clinical samples revealed 98% categorical agreement with both the clinical culture results and the on-site parallel AST validation results. This technology provides opportunities for prompt infection diagnosis and accurate antibiotic prescriptions in point-of-care settings.

Direct Antimicrobial Susceptibility Testing on Clinical Urine Samples by Optical Tracking of Single Cell Division Events

With the increasing prevalence of antibiotic resistance, the need to develop antimicrobial susceptibility testing (AST) technologies is urgent. The current challenge has been to perform the antibiotic susceptibility testing in short time, directly with clinical samples, and with antibiotics over a broad dynamic range of clinically relevant concentrations. Here, a technology for point-of-care diagnosis of antimicrobial-resistant bacteria in urinary tract infections, by imaging the clinical urine samples directly with an innovative large volume solution scattering imaging (LVSi) system and analyzing the image sequences with a single-cell division tracking method is developed. The high sensitivity of single-cell division tracking associated with large volume imaging enables rapid antibiotic susceptibility testing directly on the clinical urine samples. The results demonstrate direct detection of bacterial infections in 60 clinical urine samples with a 60 min LVSi video, and digital AST of 30 positive clinical samples with 100% categorical agreement with both the clinical culture results and the on-site agar plating validation results. This technology provides opportunities for precise antibiotic prescription and proper treatment of the patient within a single clinic visit.

Rapid Antimicrobial Susceptibility Testing of Patient Urine Samples Using Large Volume Free-Solution Light Scattering Microscopy

The emergence of antibiotic resistance has prompted the development of rapid antimicrobial susceptibility testing (AST) technologies that will enable evidence-based treatment and promote antimicrobial stewardship. To date, many rapid AST methods have been developed, but few are able to be performed on clinical samples directly. Here we developed a large volume light scattering microscopy technique that tracks phenotypic features of single bacterial cells directly in clinical urine samples without sample enrichment or culturing. The technique demonstrated rapid (90 min) detection of Escherichia coli in 24 clinical urine samples with 100% sensitivity and 83% specificity and rapid (90 min) AST in 12 urine samples with 87.5% categorical agreement with two antibiotics, ampicillin and ciprofloxacin.