ИСТИНА |
Войти в систему Регистрация |
|
ИПМех РАН |
||
The calcifications seen at the X-ray images can be the result of either initial TB process or active TB, requiring anti-TB drugs. Immunological tests do not allow to determine the activity of the TB process. To predict it, we used machine learning. We obtained a prognostic model, which could predict reactivation of the disease. Overall, the method showed a sensitivity of 97.5% and a specificity of 96%. The developed prognostic model didn’t require any special equipment and allowed to observe patients with calcifications on time.