BackgroundPneumocystis jiroveci pneumonia (PJP) is the most common opportunistic infection in immunocompromised patients. The accurate prediction of PJP development in patients undergoing immunosuppressive therapy remains challenge.MethodsPatients undergoing immunosuppressive treatment and with confirmed pneumocystis jiroveci infection were enrolled. Another group of matched patients with immunosuppressant treatment but without signs of infectious diseases were enrolled to control group.ResultsA total of 80 (40 PJP, 40 non-PJP) participants were enrolled from Tongji Hospital. None of the patients were HIV positive. The routine laboratory indicators, such as LYM, MON, RBC, TP, and ALB, were significantly lower in PJP patients than in non-PJP patients. Conversely, LDH in PJP patients was significantly higher than in non-PJP controls. For immunological indicators, the numbers of T, B, and NK cells were all remarkably lower in PJP patients than in non-PJP controls, whereas the functional markers such as HLA-DR, CD45RO and CD28 expressed on CD4+ or CD8+ T cells had no statistical difference between these two groups. Cluster analysis showing that decrease of host immunity markers including CD3+, CD4+ and CD8+ T cells, and increase of tissue damage marker LDH were the most typical characteristics of PJP patients. A further established model based on combination of CD8+ T cells and LDH showed prominent value in distinguishing PJP from non-PJP, with AUC of 0.941 (95% CI, 0.892-0.990).ConclusionsA model based on combination of routine laboratory and immunological indicators shows prominent value for predicting the development of PJP in HIV-negative patients undergoing immunosuppressive therapy.

Background: Discriminating active tuberculosis (ATB) from latent tuberculosis infection (LTBI) remains challenging. The present study aims to evaluate the performance of diagnostic models established using machine learning based on routine laboratory indicators in differentiating ATB from LTBI. Methods: Participants were respectively enrolled at Tongji Hospital (discovery cohort) and Sino-French New City Hospital (validation cohort). Diagnostic models were established based on routine laboratory indicators using machine learning. Results: A total of 2619 participants (1025 ATB and 1594 LTBI) were enrolled in discovery cohort and another 942 subjects (388 ATB and 554 LTBI) were recruited in validation cohort. ATB patients had significantly higher levels of tuberculosis-specific antigen/phytohemagglutinin ratio and coefficient variation of red blood cell volume distribution width, and lower levels of albumin and lymphocyte count than those of LTBI individuals. Six models were built and the optimal performance was obtained from GBM model. GBM model derived from training set ( n = 1965) differentiated ATB from LTBI in the test set ( n = 654) with a sensitivity of 84.38% (95% CI, 79.42%-88.31%) and a specificity of 92.71% (95% CI, 89.73%-94.88%). Further validation by an independent cohort confirmed its encouraging value with a sensitivity of 87.63% (95% CI, 83.98%-90.54%) and specificity of 91.34% (95% CI, 88.70%-93.40%), respectively. Conclusions: We successfully developed a model with promising diagnostic value based on machine learning for the first time. Our study proposed that GBM model may be of great benefit served as a tool for the accurate identification of ATB.

BackgroundRapid and effective discrimination between active tuberculosis (ATB) and latent tuberculosis infection (LTBI) remains a challenge. There is an urgent need for developing practical and affordable approaches targeting this issue.MethodsParticipants with ATB and LTBI were recruited at Tongji Hospital (Qiaokou cohort) and Sino-French New City Hospital (Caidian cohort) based on positive T-SPOT results from June 2020 to January 2021. The expression of activation markers including HLA-DR, CD38, CD69, and CD25 was examined on Mycobacterium tuberculosis (MTB)-specific CD4+ T cells defined by IFN-γ, TNF-α, and IL-2 expression upon MTB antigen stimulation.ResultsA total of 90 (40 ATB and 50 LTBI) and another 64 (29 ATB and 35 LTBI) subjects were recruited from the Qiaokou cohort and Caidian cohort, respectively. The expression patterns of Th1 cytokines including IFN-γ, TNF-α, and IL-2 upon MTB antigen stimulation could not differentiate ATB patients from LTBI individuals well. However, both HLA-DR and CD38 on MTB-specific cells showed discriminatory value in distinguishing between ATB patients and LTBI individuals. As for developing a single candidate biomarker, HLA-DR had the advantage over CD38. Moreover, HLA-DR on TNF-α+ or IL-2+ cells had superiority over that on IFN-γ+ cells in differentiating ATB patients from LTBI individuals. Besides, HLA-DR on MTB-specific cells defined by multiple cytokine co-expression had a higher ability to discriminate patients with ATB from LTBI individuals than that of MTB-specific cells defined by one kind of cytokine expression. Specially, HLA-DR on TNF-α+IL-2+ cells produced an AUC of 0.901 (95% CI, 0.833–0.969), with a sensitivity of 93.75% (95% CI, 79.85–98.27%) and specificity of 72.97% (95% CI, 57.02–84.60%) as a threshold of 44% was used. Furthermore, the performance of HLA-DR on TNF-α+IL-2+ cells for differential diagnosis was obtained with validation cohort data: 90.91% (95% CI, 72.19–97.47%) sensitivity and 68.97% (95% CI, 50.77–82.73%) specificity.ConclusionsWe demonstrated that HLA-DR on MTB-specific cells was a potentially useful biomarker for accurate discrimination between ATB and LTBI.