Deciphering prognostic indicators in non-HIV cryptococcal meningitis: Constructing and validating a predictive...

Area of Science:

• Infectious Diseases
• Mycology
• Neurology

Background:

• Cryptococcal meningitis (CM) poses a significant mortality risk, particularly in individuals with human immunodeficiency virus (HIV).
• The epidemiology, risk factors, and outcomes of CM in non-HIV adults are not well-established.
• There is a need for improved diagnostic and prognostic tools for non-HIV CM patients.

Purpose of the Study:

• To investigate the clinical characteristics and prognostic indicators of CM in adult patients without HIV.
• To develop and validate a predictive model (Nomogram) for guiding clinical decision-making in non-HIV CM.
• To enhance the understanding and management of CM in immunocompetent adults.

Main Methods:

• Retrospective cohort analysis of 64 non-HIV adult CM patients.
• Assessment of demographic, clinical, cerebrospinal fluid, and neuroimaging data.
• Development of a Nomogram prediction model using LASSO and multivariate logistic regression, validated with Bootstrap methods, ROC analysis, and decision curve analysis.

Main Results:

• Significant predictors identified include age, parenchymal, meningeal, and ventricular involvement on brain imaging, and prior immunosuppressive agent use.
• The Nomogram demonstrated satisfactory predictive performance with a C-index of 0.723 and AUC of 0.723.
• Goodness-of-fit testing indicated excellent model calibration (P = 0.995).

Conclusions:

• This study provides a comprehensive summary of clinical and imaging features of non-HIV adult CM.
• A validated Nomogram prediction model is introduced to assist in the management of non-HIV CM patients.
• The findings offer potential clinical implications for personalized care and improved outcomes in this population.