Search

Objective This study aims to establish an informative dynamic prediction model of treatment outcomes using follow-up records of tuberculosis (TB) patients, which can timely detect cases when the current treatment plan may not be effective. Materials and Methods We used 122 267 follow-up records from 17 958 new cases of pulmonary TB in the Republic of Moldova. A dynamic prediction framework integrating landmark modeling and machine learning algorithms was designed to predict patient outcomes during the course of treatment. Sensitivity and positive predictive value (PPV) were calculated to evaluate performance of the model at critical time points. New measures were defined to determine when follow-up laboratory tests should be conducted to obtain most informative results. Results The random-forest algorithm performed better than support vector machine and penalized multinomial logistic regression models for predicting TB treatment outcomes. For all 3 outcome classes (ie, cured, not cured, and died after 24 months following treatment initiation), sensitivity and PPV of prediction models improved as more follow-up information was collected. Specifically, sensitivity and PPV increased from 0.55 to 0.84 and from 0.32 to 0.88, respectively, for the not cured class. Conclusion The dynamic prediction framework utilizes longitudinal laboratory test results to predict patient outcomes at various landmarks. Sputum culture and smear results are among the important variables for prediction; however, the most recent sputum result is not always the most informative one. This framework can potentially facilitate a more effective treatment monitoring program and provide insights for policymakers toward improved guidelines on follow-up tests.

Background As part of its End TB strategy, the WHO has identified the need for non-sputum-based diagnostics that meet target product profiles (TPP) of 90% sensitivity and 70% specificity for diagnosis of Active TB (ATB) and 75% sensitivity and specificity for predicting progression from Latent TB (LTB) to ATB. The successful translation of a 3-gene blood-based signature, identified using diverse datasets, into a prototype point-of-care diagnostic, that meets the WHO TPPs, has demonstrated the power of integrating large amounts of heterogeneous data to identify generalizable disease signatures. We hypothesized that integration of more diverse datasets, comprising patients with ATB or other inflammatory lung diseases, would identify novel, robust signatures, for diagnosing ATB and predicting progression from LTB to ATB, that meet the WHO TPPs. Methods Using multi-cohort analyses, we integrated and analyzed data from 3,615 peripheral blood samples, in 49 publicly available transcriptomic datasets (discovery cohorts), from healthy controls and patients with LTB, ATB, and other diseases (COPD, viral infections, sarcoidosis, etc.). We used data from (1) 3,836 blood samples in 28 retrospective datasets and (2) 360 prospectively collected blood samples, from a household contact study in Moldova, as validation cohorts. Results Using the discovery cohorts we identified a 9-gene signature for diagnosing ATB patients from healthy controls, or individuals with LTB or other diseases. The signature achieved 90% sensitivity and 82% specificity in retrospective validation (Figure 1A) and 90% sensitivity and 69% specificity in the prospective cohort from Moldova (Figure 1B). In a longitudinal cohort of adolescents, the 9-gene signature predicted progression from LTB to ATB up to 1 year prior to sputum conversion with 76% sensitivity and 83% specificity (Figure 1C). Finally, the signature predicted prolonged lung inflammation post-treatment in the Catalysis Treatment Response Cohort (Figure 1D). 9-gene TB signature validates in independent retrospective and prospective cohorts. (A) ROC curves for comparing ATB vs. all other samples (All), or individually comparing ATB vs. Healthy, LTBI or Other Disease (OD) samples in independent retrospective validation and (B) a prospective Moldova cohort. (C) ROC curves comparing progressor and non-progressor samples collected at different time points in the Adolescent Cohort Study (ACS), for the 9-gene signature in solid lines and a previous 3-gene signature in dashed lines, (D) Distribution of the 9-gene score at different time points post-treatment in the Catalysis Treatment Response Cohort Study, for individuals with persistent lung inflammation at 24 weeks compared with those who had clear lungs at 24 weeks. Conclusion Overall, the 9-gene signature meets the WHO TPPs required for the End TB strategy. Disclosures Purvesh Khatri, PhD, Cepheid, Inc.: Advisor/Consultant|Inflammatix, Inc.: Advisor/Consultant|Inflammatix, Inc.: Inflammatix is in negotiations to license the 9-gene signature discussed here.|Inflammatix, Inc.: Stocks/Bonds.