Long Noncoding RNA and Predictive Model To Improve Diagnosis of Clinically Diagnosed Pulmonary Tuberculosis

Abstract

Clinically diagnosed pulmonary tuberculosis (PTB) patients lack microbiological evidence of Mycobacterium tuberculosis, and misdiagnosis or delayed diagnosis often occurs as a consequence. We investigated the potential of long noncoding RNAs (lncRNAs) and corresponding predictive models to diagnose these patients. We enrolled 1,764 subjects, including clinically diagnosed PTB patients, microbiologically confirmed PTB cases, non-TB disease controls, and healthy controls, in three cohorts (screening, selection, and validation). Candidate lncRNAs differentially expressed in blood samples of the PTB and healthy control groups were identified by microarray and reverse transcription-quantitative PCR (qRT-PCR) in the screening cohort. Logistic regression models were developed using lncRNAs and/or electronic health records (EHRs) from clinically diagnosed PTB patients and non-TB disease controls in the selection cohort. These models were evaluated by area under the concentration-time curve (AUC) and decision curve analyses, and the optimal model was presented as a Web-based nomogram, which was evaluated in the validation cohort. Three differentially expressed lncRNAs (ENST00000497872, n333737, and n335265) were identified. The optimal model (i.e., nomogram) incorporated these three lncRNAs and six EHRs (age, hemoglobin, weight loss, low-grade fever, calcification detected by computed tomography [CT calcification], and interferon gamma release assay for tuberculosis [TB-IGRA]). The nomogram showed an AUC of 0.89, a sensitivity of 0.86, and a specificity of 0.82 in differentiating clinically diagnosed PTB cases from non-TB disease controls of the validation cohort, which demonstrated better discrimination and clinical net benefit than the EHR model. The nomogram also had a discriminative power (AUC, 0.90; sensitivity, 0.85; specificity, 0.81) in identifying microbiologically confirmed PTB patients. lncRNAs and the user-friendly nomogram could facilitate the early identification of PTB cases among suspected patients with negative M. tuberculosis microbiological evidence.

Keywords: clinically diagnosed pulmonary tuberculosis; electronic health record; lncRNA; nomogram.

FIG 1

Overview of the strategy for investigating lncRNAs and prediction models for clinically diagnosed PTB patients. Abbreviations: PTB, pulmonary tuberculosis; PBMC, peripheral blood mononuclear cell; non-TB DC, nontuberculosis disease control; DE, differentially expressed; EHR, electronic health record; DCA, decision curve analysis.

FIG 2

Receiver operating characteristic (ROC) curves of different models in predicting clinically diagnosed PTB from suspected patients. (A) ROC curves of the selection cohort between clinically diagnosed PTB cases and non-TB disease controls. The 10-fold cross-validation ROC curve of the EHR+lncRNA model is provided in Fig. S4 in the supplemental material. P values for model AUC comparisons in the selection cohort were 0.00012 (EHR+lncRNA versus EHR only), 1.402 × 10⁻⁷ (EHR+lncRNA versus lncRNA only), and 0.103 (EHR only versus lncRNA only). P values of <0.016 (0.05/3, i.e., alpha divided by the comparison number) were considered statistically significant. (B) ROC curves of the validation cohort between clinically diagnosed PTB cases and non-TB disease controls. P values for model AUC comparisons in the validation cohort were 0.004 (EHR+lncRNA versus EHR only), 0.0003 (EHR+lncRNA versus lncRNA only), and 0.361 (EHR only versus lncRNA only).

FIG 3

Nomogram for the prediction of clinically diagnosed PTB patients. (A) Nomogram to predict the risk of clinically diagnosed PTB patients, in which points were assigned based on the feature rank order of the effect estimates. A vertical line is drawn between the “Points” axis and the corresponding point for each feature to generate a total point score and PTB probability. (B) Calibration plot in the selection cohort (left) and validation cohort (right), with lines indicating the ideal, apparent, and bias-corrected predictions of the nomogram. (C) Decision curve analysis for the nomogram and EHR-only model, with lines indicating the nomogram, the EHR-only model, and assumptions that no patients or all patients have PTB.

FIG 4

Alteration of lncRNAs before and after 2 months of intensive therapy. Shown are lncRNA expression levels before (blue) and after (red) a 2-month intensive anti-TB treatment regimen. Altered lncRNA expression levels were calculated using log₂ lncRNA (posttreatment expression/pretreatment expression) values, and the Wilcoxon matched-paired rank test was used for comparisons among 22 paired samples. The median (interquartile range) log₂ lncRNA values are as follows: −1.91 (−2.74, −1.11) before and −1.55 (−2.61, −0.79) after treatment for ENST00000497872, −3.88 (−4.81, −3.33) before and −2.30 (−2.99, −0.50) after treatment for n333737, and 2.12 (1.05, 2.34) before and 1.29 (0.85, 1.69) after treatment for n335265.