Study of the Effectiveness of Artificial Intelligence in Diagnosing Pulmonary Tuberculosis from Chest Radiographs at Samrongthap Hospital, Surin Province

Abstract

Pulmonary tuberculosis (TB) remains a significant public health concern, particularly among populations with limited access to diagnostic services. The integration of artificial intelligence (AI) in chest radiograph analysis presents a promising approach to enhance the efficiency of TB screening. This study aimed to evaluate the diagnostic performance of an AI-based system in detecting pulmonary TB from chest radiographs within the context of a community hospital setting. Secondary data were retrieved from the radiology department’s image processing system, including AI-generated TB scores and reference results from sputum acid-fast bacilli (AFB) staining using the Kinyoun carbol fuchsin method. A purposive sampling method was employed to select 1,238 individuals aged 18 years and older who underwent both chest radiography and AFB testing between October 2024 and June 2025 at Samrong Thap Hospital. Data were analyzed using IBM SPSS version 25 to assess the accuracy and applicability of the AI model. The AI system demonstrated a sensitivity of 71.11%, specificity of 81.31%, positive predictive value (PPV) of 12.55%, negative predictive value (NPV) of 98.87%, and overall accuracy of 80.94%. The area under the ROC curve (AUC) was 0.754, indicating good diagnostic performance. These findings suggest that the AI model is capable of reliably identifying TB cases and holds potential as a preliminary screening tool in resource-limited community hospitals. Given its high specificity and NPV, the system is particularly effective in ruling out non-TB cases in high-volume settings. However, it is not recommended as a standalone diagnostic tool and should be used in conjunction with clinical evaluation and other diagnostic methods. This approach supports the development of efficient TB screening strategies that enhance accessibility, coverage, and resource utilization in Thailand’s healthcare system.