The Thai Journal of Radiological Technology

Efficacy of tuberculosis screening using the AI chest 4All (DMS–TU) for Thai people innovation at Udonthani Cancer Hospital

Sirinthorn Sridubdim — 2025-10-25

Background: The chest X–ray images could be interpreted utilizing "AI Chest 4All (DMS–TU) For Thai People". This study aimed to evaluate the performance of this artificial intelligence system in tuberculosis screening. Methods: A total of 7,175 chest X–ray images from the check–up group at Udonthani Cancer Hospital were employed as the subjects. Data were collected retrospectively between July 2^nd, 2020, and September 30^th, 2023. The interpretation results of the AI Chest 4All (DMS–TU) For Thai People were compared with those of the radiologist, focusing on sensitivity, specificity, and accuracy. Results: The findings revealed that the values for sensitivity, specificity, and accuracy were 87.50, 98.60, and 98.53 respectively. Conclusion: These results demonstrate the potential of AI Chest 4 All (DMS–TU) For Thai People in efficiently screening for tuberculosis, rapidly identifying patients, enabling immediate treatment, achieving more effective tuberculosis control, and reducing the number of tuberculosis cases.

Development of specialized radiological examination appointment program in the Radiology Department at Phichit Hospital

Warunyupa Nuiyen — 2025-10-26

Background: Specialized radiological examinations are crucial tools for diagnosing and monitoring various diseases. However, the traditional appointment scheduling system faces issues such as delays, illegible handwriting, the duplicate appointments, and inefficiencies, causing inconvenience to patients and reducing the overall service efficiency of hospital. To address these problems, the researchers proposed the development of a specialized radiology appointment scheduling program. Methods: The study began with an analysis of the issues in the traditional appointment system, followed by the design and development of the program. The research evaluated user satisfaction with the new system and compared it with the traditional appointment system. Results: After implementing the program, user satisfaction improved significantly: appointment procedures increased by 70%, data security by 62.50%, convenience by 67.50%, accuracy of data entry by 65%, appointment data retrieval by 72.50%, and reduction of duplicate appointments by 65%. Additionally, the waiting time for appointment slips decreased from 10-15 minutes per case to 2-3 minutes, and no duplicate appointment incidents were reported. Conclusion: The developed specialized radiology appointment program can effectively replace the traditional system, improving service efficiency, reducing appointment steps, decreasing waiting times, enhancing user satisfaction, and supporting better management and delivery of high-quality medical services.

Evaluation of CT calibration curve impact on proton range accuracy in treatment planning

Kanyarat Purivikrai — 2025-12-07

Introduction: In proton therapy, the accuracy of dose calculation and proton range determination is critically dependent on the conversion of CT number to proton stopping power ratio (SPR). Any inaccuracy in the CT calibration curve can lead to proton range uncertainties and potential deviations in dose distribution within the target and surrounding organs at risk. Therefore, verification of the accuracy of the CT calibration curve used in treatment planning is essential to ensure precise dose delivery. Objective: To evaluate the accuracy and reliability of the CT calibration curve currently implemented in the proton therapy treatment planning system at the Department of Radiology, King Chulalongkorn Memorial Hospital, The Thai Red Cross Society. Methods: A calibration curve was generated by correlating CT numbers with the proton stopping powers of Gammex tissue-equivalent materials. The resulting curve was applied in the treatment planning system to calculate proton ranges, which were then compared with measurements obtained using a Giraffe dosimeter on a Varian ProBeam system. Both phantom and real tissue samples were evaluated for comparison. Results: No statistically significant difference was found between calculated and measured proton ranges (P = 0.21). The mean differences of R80 and R90 in the phantom and real tissue were 1.53 ± 3.85 mm and 0.51 ± 3.08 mm, and 1.54 ± 3.82 mm and 0.83 ± 2.43 mm, respectively. Six of the Gammex materials met the AAPM TG-185 criteria, while the remaining materials showed consistent results with previous studies using the same CT and proton systems. Conclusion: The current CT calibration curve used in the proton therapy planning system at King Chulalongkorn Memorial Hospital demonstrates acceptable accuracy and reliability for clinical dose calculation and treatment planning applications.