Tuberculosis (TB) is among the leading causes of mortality worldwide with an estimated 2 billion individuals currently infected. Latent tuberculosis infection (LTBI) is the most common form of TB infection affecting 13 million Americans. While many with LTBI remain asymptomatic, an estimated 10% of immunocompetent patients with LTBI will reactivate to active TB, and will become infectious. LTBI is treatable with a prolonged antibiotic treatment; however, potential side effects motivate the development of new diagnostic approaches that can identify with high specificity patients at the highest risk of reactivation, for who therapy would be most beneficial. The tuberculin skin test (TST) and interferon-? release assays (IGRAs) are commonly used for TB and LTBI screening. Both tests provide good measures of TB exposure; however, neither is effective at diagnosing LTBI (positive predictive values <5%). Moreover, neither provide any prognostic stratification based upon reactivation risk. Both the TST and IGRAs probe immunological memory to TB-related antigen challenges and we hypothesize that a more nuanced and personalized approach to monitoring immune responses to both TB- specific and non-specific antigens might reveal new approaches to LTBI diagnosis and patient stratification. Enabling a new, individualized approach to LTBI diagnostics, we propose to combine high throughput, multiplexed inflammatory biomarker detection strategies and powerful bioinformatics tools that allow for the identification of previously obscured multi-marker diagnostic signatures of LTBI status and reactivation risk. Silicon photonic microring resonators are an enabling technology for biomarker analysis due to their intrinsic scalability and multiplexing capabilities. Applied to the detection of cytokine panels, this technology supports the rapid immune profiling of individual samples under both TB-specific and non-specific antigen stimulation conditions. Machine learning algorithms will be utilized to analyze the resulting dense data streams to facilitate selection of key diagnostic signatures forming the basis for predictive model development and deployment. This powerful analytical combination is supplemented by deep expertise in clinical diagnosis and treatment of TB and LTBI, and an enabling collaboration and connection to subjects from an international location with high TB burden and exposure in a healthcare worker population subjected to regularly-scheduled and repeated LTBI screening. The resulting diagnostic workflow and machine learning feature selection approaches will reveal multiplexed biomarker signatures that have strong positive predictive correlation with LTBI status (+ or -). This approach will also further stratify LTBI+ subjects on the basis of reactivation potential, thus providing a fundamentally new approach to identifying subjects that are most likely to benefit from therapeutic intervention. The end result of this project will be a new precision medicine-based diagnostic strategy that is vastly superior to the current state-of-the-art and offers the potential to transform current clinical practice.

Public Health Relevance

Tuberculosis (TB) affects an estimated one third of the world?s population and an asymptomatic latent state of tuberculosis infection (LTBI) is extremely common. Unfortunately, there are not any good clinical tests that can definitely diagnose LTBI, making it difficult to identify patients that should be treated to prevent reactivation to active TB, which is infectious. We will integrate cutting edge measurement technologies and machine learning bioinformatic approaches to identify and test multiplexed biomarker signatures that will transform clinical TB management by enabling personalized diagnosis of LTBI and the stratification of individuals with the highest potential for reactivation.

Agency
National Institute of Health (NIH)
Institute
National Institute of Allergy and Infectious Diseases (NIAID)
Type
Research Project (R01)
Project #
5R01AI141591-02
Application #
10006790
Study Section
Enabling Bioanalytical and Imaging Technologies Study Section (EBIT)
Program Officer
Lacourciere, Karen A
Project Start
2019-09-05
Project End
2023-08-31
Budget Start
2020-09-01
Budget End
2021-08-31
Support Year
2
Fiscal Year
2020
Total Cost
Indirect Cost
Name
University of Michigan Ann Arbor
Department
Chemistry
Type
Schools of Arts and Sciences
DUNS #
073133571
City
Ann Arbor
State
MI
Country
United States
Zip Code
48109