Single-test Diagnosis of Lyme Disease Using Computational Point-of-Care Sensors
Outcome/Accomplishment
UCLA researchers supported by the National Science Foundation (NSF)-funded Precise Advanced Technologies and Health Systems for Under-resourced Populations (NSF PATHS-UP) Engineering Research Center (ERC), headquartered at Texas A&M University, have developed a novel point-of-care (POC)-compatible serologic test for Lyme disease that offers a significant advance over current tests. Lyme disease is the most prevalent vector-borne disease in North America and Europe. With the growing prevalence of emerging infections and vector-borne illnesses, there is a vital need for testing platforms that can combat the emergence and transmission of diseases. Platforms employing tests like that developed at UCLA that can be deployed rapidly and reliably in POC settings or for at-home testing can play a leading role in this effort.
Impact/Benefits
The PATHS-UP UCLA researchers developed a single-stage test approach for Lyme disease diagnosis. Instead of relying on the traditional two-step testing process, this new test can detect multiple Lyme disease markers at once from a stacked set of paper layers to control sample flow and measure multiple patient-specific immune functions reflecting Lyme infection in parallel. State-of-the art AI algorithms process the resulting data to rapidly predict the presence or absence of Lyme disease.
This technology is attracting commercial interest from diagnostic test manufacturers, suggesting a potential path to real-world use. If successfully developed and adopted, this single-test approach could help patients receive answers sooner, support earlier treatment, and reduce the uncertainty that too many people face during the search for a Lyme disease diagnosis.
Explanation/Background
Computational point-of-care sensors enable rapid, low-cost, and accessible diagnostics in emergency, remote, and resource-limited areas that lack access to centralized medical facilities. First reported in a 2024 Nature Communications publication, the UCLA stacked paper test showed highly promising results, including 95.5% sensitivity (few false negatives) and 100% specificity (no false positives) in blinded validation studies, and strong agreement with standard laboratory testing.
In a 2026 ACS Nano paper, the same team demonstrated new processing methods to boost performance. To overcome inaccuracies in the use of neural network models for processing POC signals, the PATHS-UP researchers used a Monte Carlo dropout (MCDO)-based autonomous uncertainty quantification technique. MCDO was able to identify and exclude erroneous predictions based on understanding uncertainty in a neural network model, significantly improving the sensitivity and reliability of their diagnostic test.
They applied this approach to the stacked paper-based and computational readout platform developed for rapid POC diagnosis of Lyme disease. The resulting sensor platform integrates a disposable paper-based assay, a hand-held optical reader, and an MCDO-based inference algorithm, providing rapid and cost-effective Lyme disease diagnostics in under 20 minutes using only 20 μL of patient serum. Blinded testing using new patient samples demonstrated increased performance for MCDO-based uncertainty quantification, reflecting the robustness of neural network-driven computational POC sensing systems.
Location
College Station, Texaswebsite
Start Year
Biotechnology and Healthcare
Biotechnology and Healthcare
Lead Institution
Core Partners
Fact Sheet
Outcome/Accomplishment
UCLA researchers supported by the National Science Foundation (NSF)-funded Precise Advanced Technologies and Health Systems for Under-resourced Populations (NSF PATHS-UP) Engineering Research Center (ERC), headquartered at Texas A&M University, have developed a novel point-of-care (POC)-compatible serologic test for Lyme disease that offers a significant advance over current tests. Lyme disease is the most prevalent vector-borne disease in North America and Europe. With the growing prevalence of emerging infections and vector-borne illnesses, there is a vital need for testing platforms that can combat the emergence and transmission of diseases. Platforms employing tests like that developed at UCLA that can be deployed rapidly and reliably in POC settings or for at-home testing can play a leading role in this effort.
Location
College Station, Texaswebsite
Start Year
Biotechnology and Healthcare
Biotechnology and Healthcare
Lead Institution
Core Partners
Fact Sheet
Impact/benefits
The PATHS-UP UCLA researchers developed a single-stage test approach for Lyme disease diagnosis. Instead of relying on the traditional two-step testing process, this new test can detect multiple Lyme disease markers at once from a stacked set of paper layers to control sample flow and measure multiple patient-specific immune functions reflecting Lyme infection in parallel. State-of-the art AI algorithms process the resulting data to rapidly predict the presence or absence of Lyme disease.
This technology is attracting commercial interest from diagnostic test manufacturers, suggesting a potential path to real-world use. If successfully developed and adopted, this single-test approach could help patients receive answers sooner, support earlier treatment, and reduce the uncertainty that too many people face during the search for a Lyme disease diagnosis.
Explanation/Background
Computational point-of-care sensors enable rapid, low-cost, and accessible diagnostics in emergency, remote, and resource-limited areas that lack access to centralized medical facilities. First reported in a 2024 Nature Communications publication, the UCLA stacked paper test showed highly promising results, including 95.5% sensitivity (few false negatives) and 100% specificity (no false positives) in blinded validation studies, and strong agreement with standard laboratory testing.
In a 2026 ACS Nano paper, the same team demonstrated new processing methods to boost performance. To overcome inaccuracies in the use of neural network models for processing POC signals, the PATHS-UP researchers used a Monte Carlo dropout (MCDO)-based autonomous uncertainty quantification technique. MCDO was able to identify and exclude erroneous predictions based on understanding uncertainty in a neural network model, significantly improving the sensitivity and reliability of their diagnostic test.
They applied this approach to the stacked paper-based and computational readout platform developed for rapid POC diagnosis of Lyme disease. The resulting sensor platform integrates a disposable paper-based assay, a hand-held optical reader, and an MCDO-based inference algorithm, providing rapid and cost-effective Lyme disease diagnostics in under 20 minutes using only 20 μL of patient serum. Blinded testing using new patient samples demonstrated increased performance for MCDO-based uncertainty quantification, reflecting the robustness of neural network-driven computational POC sensing systems.