Infection and drug resistance · 2026
Clinical Characteristics and Machine Learning-Based Severity Classification of Chlamydia psittaci Pneumonia: A Retrospective Cohort Study.
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.
Lei Xiao, Zhao Liyuan, Zhong Ziwei, Lin Haiyan, Xie Zhiting, Guo Yaping, Zhang Sufeng, Zhang Chengsheng, Gong Tian
Аннотация
BACKGROUND: Chlamydia psittaci infection can cause severe community-acquired pneumonia with significant mortality. Distinguishing severe from non-severe disease remains challenging. This study used clinical data from patients with C. psittaci pneumonia and multiple machine-learning methods to develop a model for severity classification.
METHODS: We retrospectively analyzed 231 hospitalized patients with C. psittaci pneumonia, including 84 severe cases, between January 2022 and April 2025 in Jiangxi Province, China. Severe pneumonia was defined according to the IDSA/ATS and Chinese adult community-acquired pneumonia criteria. The model outcome was the composite clinical label of severe versus non-severe pneumonia rather than mortality. A comprehensive machine-learning framework incorporating 11 algorithms and six feature-selection strategies was used to develop the severity-classification model.
RESULTS: C. psittaci pneumonia cases were sporadic and widely distributed across Jiangxi Province. Elderly patients, especially those with cardiovascular disease or diabetes, had an increased risk of severe illness. Laboratory tests in severe cases showed higher neutrophils, D-dimer, CRP, PCT, IL-6, IL-8, and IL-10, with lower lymphocytes, NK cells, albumin, and serum calcium. CT commonly showed large patchy opacities in the lower lung lobes. Additional microorganisms were co-detected by mNGS in 74.03% of patients. In the internal hold-out test set (n = 69), the RF+SVM model achieved an area under the receiver operating characteristic curve (AUC) of 0.845 (95% CI, 0.754-0.935) for distinguishing severe from non-severe pneumonia. The final RF+SVM severity-classification model incorporated nine laboratory predictors.
CONCLUSION: This study identified key clinical and laboratory differences between non-severe and severe C. psittaci pneumonia and developed a predictive model using comprehensive machine learning approaches. The model may support admission-based severity stratification and identify patients who may warrant closer monitoring. This is a single-center retrospective study, and the model needs further validation in prospective multicenter cohorts.
METHODS: We retrospectively analyzed 231 hospitalized patients with C. psittaci pneumonia, including 84 severe cases, between January 2022 and April 2025 in Jiangxi Province, China. Severe pneumonia was defined according to the IDSA/ATS and Chinese adult community-acquired pneumonia criteria. The model outcome was the composite clinical label of severe versus non-severe pneumonia rather than mortality. A comprehensive machine-learning framework incorporating 11 algorithms and six feature-selection strategies was used to develop the severity-classification model.
RESULTS: C. psittaci pneumonia cases were sporadic and widely distributed across Jiangxi Province. Elderly patients, especially those with cardiovascular disease or diabetes, had an increased risk of severe illness. Laboratory tests in severe cases showed higher neutrophils, D-dimer, CRP, PCT, IL-6, IL-8, and IL-10, with lower lymphocytes, NK cells, albumin, and serum calcium. CT commonly showed large patchy opacities in the lower lung lobes. Additional microorganisms were co-detected by mNGS in 74.03% of patients. In the internal hold-out test set (n = 69), the RF+SVM model achieved an area under the receiver operating characteristic curve (AUC) of 0.845 (95% CI, 0.754-0.935) for distinguishing severe from non-severe pneumonia. The final RF+SVM severity-classification model incorporated nine laboratory predictors.
CONCLUSION: This study identified key clinical and laboratory differences between non-severe and severe C. psittaci pneumonia and developed a predictive model using comprehensive machine learning approaches. The model may support admission-based severity stratification and identify patients who may warrant closer monitoring. This is a single-center retrospective study, and the model needs further validation in prospective multicenter cohorts.