PredictTB - Using Biomarkers to Predict TB Treatment Duration

Study Director
Clifton Barry, Gerhard Walzl
Start Date
6 / 2017
Trial Phase
Phase II
Trial Status
Current Enrollment
2 - June 28, 2017
Target Enrollment

Predict TB (Using Biomarkers to Predict TB Treatment Duration) will identify the set of criteria needed to reduce TB treatment times. The PredictTB project team will enroll up to 620 patients with drug sensitive pulmonary TB in a clinical trial with five collaborating sites in Cape Town, South Africa and four in Henan Province, China. Over five years, the team will test criteria using data from radiographic lung scans, assays of patient body fluids and analysis of bacterial cultures. They will evaluate the initial burden of disease, the rate of response to therapy and the presence of TB-causing bacteria in drug adherent patients. Assessment of these risk factors will be used to identify patients who are successfully treated and those that need additional therapy. Patients will then be followed for a year after completing treatment to assess relapse rates.

This protocol builds upon the historical trials and several successful small studies that suggest that directly monitoring lung pathology using [18F]- FDG PET/CT correlates better with treatment outcome than culture status. We will prospectively identify patients at low risk based on their baseline radiographic extent of disease, and further refine this risk score by evaluating the rate of resolution of the lung pathology (CT) and inflammation (PET) at one month as well as checking an end-oftreatment GeneXpert test for the sustained presence of bacteria. Patients classified as low risk will be randomized to receive a shortened 4- month or a full 6-month course of therapy. If successful, this trial will both offer a badly needed alternative to culture status as a trial-level surrogate marker for outcome as well as provide critical information for preclinical and early clinical efforts to identify new agents and combinations with the potential to shorten therapy.

Hypothesis: A combination of radiographic characteristics at baseline, the rate of change of these features at one month, and markers of residual bacterial load at the end of treatment will identify patients with tuberculosis who are cured with 4 months (16 weeks) of standard treatment.

Compound Associations
Regimen Associations