Dataset Information

Development and performance of CUHAS-ROBUST application for pulmonary rifampicin-resistance tuberculosis screening in Indonesia.

ABSTRACT:

Background and objectives

Diagnosis of Pulmonary Rifampicin Resistant Tuberculosis (RR-TB) with the Drug-Susceptibility Test (DST) is costly and time-consuming. Furthermore, GeneXpert for rapid diagnosis is not widely available in Indonesia. This study aims to develop and evaluate the CUHAS-ROBUST model performance, an artificial-intelligence-based RR-TB screening tool.

Methods

A cross-sectional study involved suspected all type of RR-TB patients with complete sputum Lowenstein Jensen DST (reference) and 19 clinical, laboratory, and radiology parameter results, retrieved from medical records in hospitals under the Faculty of Medicine, Hasanuddin University Indonesia, from January 2015-December 2019. The Artificial Neural Network (ANN) models were built along with other classifiers. The model was tested on participants recruited from January 2020-October 2020 and deployed into CUHAS-ROBUST (index test) application. Sensitivity, specificity, and accuracy were obtained for assessment.

Results

A total of 487 participants (32 Multidrug-Resistant/MDR 57 RR-TB, 398 drug-sensitive) were recruited for model building and 157 participants (23 MDR and 21 RR) in prospective testing. The ANN full model yields the highest values of accuracy (88% (95% CI 85-91)), and sensitivity (84% (95% CI 76-89)) compare to other models that show sensitivity below 80% (Logistic Regression 32%, Decision Tree 44%, Random Forest 25%, Extreme Gradient Boost 25%). However, this ANN has lower specificity among other models (90% (95% CI 86-93)) where Logistic Regression demonstrates the highest (99% (95% CI 97-99)). This ANN model was selected for the CUHAS-ROBUST application, although still lower than the sensitivity of global GeneXpert results (87.5%).

Conclusion

The ANN-CUHAS ROBUST outperforms other AI classifiers model in detecting all type of RR-TB, and by deploying into the application, the health staff can utilize the tool for screening purposes particularly at the primary care level where the GeneXpert examination is not available.

Trial registration

NCT04208789.

SUBMITTER: Herman B

PROVIDER: S-EPMC7993842 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:BACKGROUND: Rapid new diagnostic methods (including Xpert MTB/RIF assay) use rifampicin resistance as a surrogate marker for multidrug resistant tuberculosis. Patients infected with rifampicin susceptible strains are prescribed first line anti-tuberculosis therapy. The roll out of such methods raises a concern that strains with resistance to other first line anti-tuberculosis drugs including isoniazid will be missed and inappropriate treatment given. To evaluate implications of using such methods review of resistance data from high burden settings such as ours is essential. OBJECTIVE: To determine resistance to first line anti-tuberculosis drugs amongst rifampicin susceptible pulmonary Mycobacterium tuberculosis (MTB) isolates from Pakistan. MATERIALS AND METHODS: Data of pulmonary Mycobacterium tuberculosis strains isolated in Aga Khan University Hospital (AKUH) laboratory (2009-2011) was retrospectively analyzed. Antimicrobial susceptibility profile of rifampicin susceptible isolates was evaluated for resistance to isoniazid, pyrazinamide, ethambutol, and streptomycin. RESULTS: Pulmonary specimens submitted to AKUH from 2009 to 2011 yielded 7738 strains of Mycobacterium tuberculosis. These included 54% (n 4183) rifampicin susceptible and 46% (n: 3555) rifampicin resistant strains. Analysis of rifampicin susceptible strains showed resistance to at least one of the first line drugs in 27% (n:1133) of isolates. Overall isoniazid resistance was 15.5% (n: 649), with an isoniazid mono-resistance rate of 4% (n: 174). Combined resistance to isoniazid, pyrazinamide, and ethambutol was noted in 1% (n: 40), while resistance to isoniazid, pyrazinamide, ethambutol, and streptomycin was observed in 1.7% (n: 70) of strains. CONCLUSIONS: Our data suggests that techniques (including Xpert MTB/RIF assay) relying on rifampicin susceptibility as an indicator for initiating first line therapy will not detect patients infected with MTB strains resistant to other first line drugs (including isoniazid). The roll out of these techniques must therefore be accompanied by strict monitoring ensuring early resistance detection to increase chances of improved patient outcomes.

Project description:BackgroundWhile rifampicin (RFP) and isoniazid (INH) are the most commonly used first-line antituberculosis drugs, multidrug resistance in Mycobacterium tuberculosis poses a threat to the success of tuberculosis (TB) control programs. Clinical practice guidelines and expert consensuses recommend drug susceptibility testing (DST) before the initiation of antituberculosis treatment. However, traditional DST is time-consuming and has high requirements for laboratory conditions. The recently developed molecular diagnostic techniques, such as DNA microarray, offer new options. We thus investigated the diagnostic value of DNA microarray in detecting RFP + INH-resistant TB, with an attempt to identify simple, efficient, and accurate drug-resistant TB testing methods.MethodsThe clinical features and DST results of patients diagnosed with pulmonary tuberculosis by Bactec MGIT 960 liquid culture system (Becton, Dickinson and Company, Franklin Lakes, NJ, USA) who received DNA microarray analysis in our center from July 2019 to July 2020 were retrospectively analyzed. Level of agreement between liquid culture and DNA microarray technology was assessed by using the Cohen kappa coefficient. With the results of liquid culture as the gold standard, the sensitivity and specificity of the DNA microarray were calculated, and the receiver operating characteristic (ROC) curves were used to assess the diagnostic values of the DNA microarray in detecting RFP + INH-resistant TB.ResultsA total of 825 patients were enrolled. The sensitivity and specificity of DNA microarray were 0.84 and 0.94, respectively, in the detection of RFP resistance, with an area under the curve (AUC) of 0.89 [95% confidence interval (CI): 0.87-0.91)] and a Cohen kappa coefficient of 0.78 (95% CI: 0.72-0.83). For INH resistance, the sensitivity and specificity of the DNA microarray were 0.73 and 0.97, respectively, with an AUC of 0.85 (95% CI: 0.82-0.87) and a Cohen kappa coefficient of 0.75 (95% CI: 0.70-0.80).ConclusionsThe DNA microarray had high specificity and sensitivity in detecting RFP + INH-resistant TB. As a rapid, accurate, and practical technique, it can be routinely performed in clinical laboratories.

Project description:BackgroundActive tuberculosis (TB) case finding is important as it helps detect pulmonary TB cases missed by the other active screening methods. It requires periodic mass screening in risk population groups such as prisoners and refugees. Unfortunately, in these risk population groups periodic mass screening can be challenging due to lengthy turnaround time (TAT), cost and implementation constraints. The aim of this study was to evaluate a diagnostic algorithm that can reduce the TAT and cost for TB and Rifampicin resistance (RR) detection. The algorithm involves testing with TB-LAMP followed by Xpert MTB/RIF for positive TB-LAMP cases to diagnose TB during mass campaigns in prisons and refugee camps.MethodsThe National Tuberculosis Control Program (NTCP) organized routine TB mass-screening campaigns in 34 prisons and 3 villages with refugees camps in Cameroon in 2019. TB LAMP was used for initial TB diagnosis and all TB-LAMP positive cases tested with the Xpert MTB/RIF assay to determine RR. TAT and cost benefits analysis of the combined use of TB-LAMP and Xpert MTB/RIF assays was determined and compared to the Xpert MTB/RIF assay when used only.ResultsA total of 4075 sputum samples were collected from TB presumptive, 3672 cases in 34 prisons and 403 samples in 3 villages. Of the 4,075 samples screened with TB-LAMP, 135 were TB positive (3.31%) and run on the Xpert MTB/RIF. Of the 135 positives cases, Xpert MTB/RIF revealed 3 were RR (2.22%). The use of TB-LAMP followed by testing with Xpert MTB/RIF for TB and RR detection reduced the TAT by 73.23% in prisons and 74.92% in villages. In addition to a reduced TAT, the two molecular tests used in synergy is cost benefit from year 2 onwards.ConclusionThis study demonstrates the advantages of a diagnostic algorithm based on an initial testing with TB-LAMP followed by testing with Xpert MTB/RIF for TB diagnosis. This approach improved early and rapid TB detection with an added advantage of providing RR status. The proposed algorithm is effective and less costly from the second year of implementation and should be used by TB control programs.