P-1961. Quantitative Pneumocystis jirovecii Testing for Prediction of Pneumocystis Disease

TL;DR

This study evaluates a PCR test's ability to distinguish Pneumocystis jirovecii colonization from disease in immunocompromised patients.

Contribution

The study introduces a quantitative PCR threshold to differentiate colonization from proven/probable Pneumocystis disease.

Findings

A 3-log10 threshold (1000 genome copies/mL) showed 92% specificity but only 48% sensitivity for predicting disease.

Higher Pneumocystis DNA levels were associated with proven/probable disease compared to colonization.

Many disease cases had DNA quantities below the proposed threshold, suggesting the need for refined cutoffs.

Abstract

Pneumocystis jirovecii pneumonia (PCP) is a serious opportunistic fungal infection. Polymerase chain reaction (PCR) testing of respiratory specimens for P. jirovecii is widely recognized as a key diagnostic tool for PCP1. However, many individuals can be colonized by P. jirovecii without progressing to PCP2. This study aims to evaluate the performance characteristics and establish the optimal quantitative threshold of a P. jirovecii real-time PCR assay to differentiate proven/probable PCP from colonization.Figure 1Pneumocystis Threshold Receiver Operating Characteristic CurveTable 1Cohort characteristics

Pneumocystis Threshold Receiver Operating Characteristic Curve

Cohort characteristics

This single-center cohort study included patients with a positive P. jirovecii PCR between September 15th, 2023 and September 6th, 2024. Patients were categorized as having P. jirovecii colonization or proven/probable disease using the EORTC/MSGERC criteria by an infectious disease fellow blinded to the quantitative result. PCR testing was performed on lower respiratory specimens with an assay targeting the multicopy mitochondrial large subunit rRNA (mtLSU) gene. The cutoff of 3-log10 (1000 genome copies/mL) was chosen based on a receiver-operating curve conducted on a separate validation cohort3.

85 patients were included (24 colonization; 61 proven/probable). Most had at least one predisposing condition for PCP (Table 1). Median Pneumocystis DNA was lower in the colonization group vs. the proven/probable group (146.5 copies/mL vs 621.0 copies/mL, p< 0.001). Among PCR-positive individuals, a 3-log10 cutoff had a sensitivity of 48% (95% CI 35-61%), specificity 92% (95% CI 73-99%), positive predictive value 94% (95% CI 0.79-0.99%), and negative predictive value 41% (95% CI 28-55%), for prediction of proven/probable disease over colonization (Figure 1).

Among an immunocompromised population, higher quantitative Pneumocystis burden in lower respiratory samples was associated with proven/probable disease. However, many patients meeting the criteria for disease had Pneumocystis DNA quantity below the 3-log10 threshold. Future studies to establish a separate upper threshold—above which patients meet criteria for disease—and lower threshold—below which patients meet criteria for colonization—may be needed.

Kyle D. Brizendine, MD, Pfizer: Advisor/Consultant Hannah Wang, MD, Cepheid: Grant/Research Support|Hologic: Advisor/Consultant|Hologic: Grant/Research Support|Moderna: Honoraria

Figures