Abstract
BACKGROUND: Given the rising incidence, acute onset of Pneumocystis jirovecii pneumonia (PJP) in HIV-negative patients, and the limitations of existing diagnostic methods (poor sensitivity, low accessibility, and difficulty distinguishing infection from colonization), this study aims to develop and validate a diagnostic prediction model for PJP in this population. METHODS: A retrospective observational study was conducted on hospitalized patients with suspected PJP at Zhejiang Provincial People's Hospital from 2017 to 2022. Based on the inclusion and exclusion criteria, 115 HIV-negative patients were enrolled, including 37 with confirmed PJP and 78 with PJP ruled out. Clinical data including age, gender, comorbidities, laboratory indicators, and treatment history were collected. The LASSO method was used to screen predictive variables for constructing a model, which was internally validated via Bootstrap resampling and visualized as a nomogram. RESULTS: A 7-variable PJP diagnostic prediction model was constructed following LASSO variable screening. The model exhibited good performance in distinguishing whether HIV-negative patients were afflicted with PJP, with an area under the ROC curve (AUC) of 0.895 (95% CI: 0.834-0.956). The calibration curve revealed high consistency between predicted and actual probabilities, and decision curve analysis demonstrated its superior clinical net benefit over all-intervention and no-intervention strategies. Bootstrap internal validation yielded a stable AUC of 0.895 (95% CI: 0.831-0.947) with consistent calibration. CONCLUSION: This study established a predictive model for Pneumocystis jirovecii pneumonia (PJP) in HIV-negative patients, and this effort is clinically meaningful. However, due to several limitations of the present study, it should be emphasized that the final nomogram only serves as a supplement to, rather than a replacement for, existing diagnostic approaches. The findings of this study warrant further validation in future multicenter or prospective investigations.