ACADSTAFF UGM
| Title | : | The potential of an electronic nose (e-Nose) as a screening tool for tuberculosis (TB) in Indonesia |
| Author | : |
dr. Antonia Morita Iswari Saktiawati., Ph.D (1) Prof. Dr. Eng. Kuwat Triyana, M.Si. (2) dr. Riris Andono Ahmad, MPH, Ph.D (3) Prof. dr. Ari Probandari, MPH, Ph.D (4) Prof. dr. Yodi Mahendradhata, M.Sc., PhD, FRSPH (5) |
| Date | : | 27 2021 |
| Keyword | : | tuberculosis,diagnosis,artificial intelligence,sensitivity,specificity,electronic nose tuberculosis,diagnosis,artificial intelligence,sensitivity,specificity,electronic nose |
| Abstract | : | Background and Aims: TB screening is challenging in low-resourced, highly burdened settings. An accurate, easy-to-use, fast, and cheap diagnostic tool is needed. We develop an e-Nose to screen TB by examining the patient’s exhaled breath. In this first (training) phase, we aim to investigate the diagnostic accuracy of e-Nose in differentiating TB patients from healthy controls. Methods: TB patients were recruited consecutively from the Respira Lung Hospital, Yogyakarta. The healthy controls were recruited from TB patients’ neighborhood to represent the same socioeconomic conditions. All participants were characterized by clinical symptoms, Chest X-ray, smear microscopic, Xpert MTB/Rif, culture, HIV test, and eNose-TB breath test. The participants breathed normally through a disposable mouthpiece of pipe connected to the electronic nose. Afterward, the electronic-nose was connected to a laptop, and the breath data was analyzed. A leave-one-out cross-validation was used in the hyper-parameters tuning process of Support Vector Machine and Recursive Feature Elimination for a feature selection method. Logistic regression was used as an estimator. Results: We included 27 TB patients and 24 healthy controls. Median age was 45; range: 20-70 years; and median BMI was 20; range: 14.7-32.9 kg/m2. The e-Nose device had sensitivity of 0.95 (95%CI=0.75-0.99), specificity of 0.82 (95%CI=0.59-0.94), positive predictive value of 0.84 (95%CI: 0.63-0.95), and negative predictive value of 0.95 (CI: 0.71-0.99). Conclusion: In this training phase, the e-Nose had high diagnostic accuracy. It is also an easy-to-use device. It needs to be further developed in a validation study in presumptive TB patients that is currently in progress. |
| Group of Knowledge | : | |
| Level | : | Internasional |
| Status | : |
Published
|
| No | Title | Action |
|---|---|---|
| 1 |
Bukti sertifikat presentasi oral seminar internasional.pdf
Document Type : [PAK] Sertifikat Seminar
|
View |