Frontiers in public health · 2025
Artificial intelligence: a useful tool in active tuberculosis screening among vulnerable groups in Romania - advantages and limitations.
СтатьяНазвание и аннотация: Перевод в очереди
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Mahler Beatrice, Stoichiță Alexandru, Băiceanu Dragoș, Panciu Traian-Constantin, Dendrino Dragoș, Mihai Mihaela, Bobocea Raluca, Ibraim Elmira, Bălteanu Mara, Popescu Oana, Burecu Mădălina Oana, Moșteanu Ioana Mădălina, Veronese Vanessa, Matache Radu, Munteanu Ioana, Popa Cristina, Dragomir Justina Antonela
Аннотация
INTRODUCTION: Despite advances in diagnostic technologies for tuberculosis (TB), global control of this disease requires improved technologies for active case finding in selected vulnerable populations. The integration of artificial intelligence (AI) into imaging modalities has been anticipated to assume a pivotal position in conjunction with traditional bacteriological diagnostic approaches, especially in the active diagnosis of vulnerable groups.
METHODS: The study was conducted as a prospective investigation spanning from November 2019 to October 2023, in Romania's national TB screening project. From total of 92,368 tested participants, 404 patients were included in the study, with 202 individuals diagnosed with TB and 202 individuals serving as controls. The initial interpretation of radiological images was performed by AI X-Vision software and patients with suspicious findings were confirmed to have TB using GeneXpert testing. The objective of this study is to discover a threshold at which the AI score can accurately assess the risk of TB, regardless of the patient's medical background.
RESULTS: The study involved a number of 404 people, among whom 202 were diagnosed with TB out of a total of 92,368 participants, and the remaining 202 patients represent the control group. The current study highlighted, at an AI threshold value of 0.4, that 89% of the screened people benefited from a correct assessment using the AI associated with the radiological examination. ROC curve analysis indicates an AUC of 0.923 (95% CI:0.893-0.947; significance level p < 0.0001) which shows that the test has a high capacity to accurately detect individuals with TB and also to rule out those who do not have the disease, with sensitivity 87.1, specificity 91.6 and criterion >0.3585.
CONCLUSION: Our study brings to the fore the significance of integrating AI software X-vision with bacteriological diagnosis in detecting TB among vulnerable groups in Romania. This underscores the imperative at the global level to develop solutions in the prompt diagnosis of TB, particularly within vulnerable groups.
METHODS: The study was conducted as a prospective investigation spanning from November 2019 to October 2023, in Romania's national TB screening project. From total of 92,368 tested participants, 404 patients were included in the study, with 202 individuals diagnosed with TB and 202 individuals serving as controls. The initial interpretation of radiological images was performed by AI X-Vision software and patients with suspicious findings were confirmed to have TB using GeneXpert testing. The objective of this study is to discover a threshold at which the AI score can accurately assess the risk of TB, regardless of the patient's medical background.
RESULTS: The study involved a number of 404 people, among whom 202 were diagnosed with TB out of a total of 92,368 participants, and the remaining 202 patients represent the control group. The current study highlighted, at an AI threshold value of 0.4, that 89% of the screened people benefited from a correct assessment using the AI associated with the radiological examination. ROC curve analysis indicates an AUC of 0.923 (95% CI:0.893-0.947; significance level p < 0.0001) which shows that the test has a high capacity to accurately detect individuals with TB and also to rule out those who do not have the disease, with sensitivity 87.1, specificity 91.6 and criterion >0.3585.
CONCLUSION: Our study brings to the fore the significance of integrating AI software X-vision with bacteriological diagnosis in detecting TB among vulnerable groups in Romania. This underscores the imperative at the global level to develop solutions in the prompt diagnosis of TB, particularly within vulnerable groups.