# Diagnostic Accuracy of Procalcitonin in the Diagnosis of Sepsis in Cancer Patients Hospitalized for Infection

## TL;DR

This study evaluates how well procalcitonin (PCT) can diagnose sepsis in cancer patients compared to non-cancer patients, finding that PCT is less accurate in cancer patients.

## Contribution

The study provides new insights into the diagnostic accuracy of PCT for sepsis specifically in cancer patients hospitalized for infection.

## Key findings

- PCT levels were higher in cancer patients compared to non-cancer patients.
- PCT had lower specificity in cancer patients when using a 0.5 ng/mL cut-off.
- A 1 ng/mL PCT cut-off showed better diagnostic accuracy in cancer patients.

## Abstract

Sepsis is defined as a life‐threatening, dysfunctional body‐response to infection. Procalcitonin (PCT) is considered a marker of sepsis due to bacterial infections and it has been extensively used as a guide to antimicrobial management in the general population. The clinical role of PCT in cancer patients admitted to the Emergency Department (ED) for infection is still little researched.

A prospective observational study enrolling all adult patients hospitalized for infection referred to the ED of IRCCS Azienda Ospedaliero‐Universitaria di Bologna between February 1st, 2023 and July 31st, 2023 was conducted. The primary endpoint was to evaluate the accuracy of PCT in the diagnosis of sepsis (defined according to the latest guidelines) in patients with cancer in comparison to non‐cancer patients.

1041 out of 1125 eligible patients were enrolled (559 males and 482 females), out of whom 289 (27.8%) had active cancer. PCT levels differed between cancer and non‐cancer patients (1 ng/mL with IQR 5.85 vs. 0.6 ng/mL with IQR 2.7; p < 0.001). The AUROC of PCT for the diagnosis of sepsis in the entire enrolled population was 0.717 (95% CI 0.683–0.745), whereas it was 0.655 (95% CI 0.592–0.718) in cancer patients and 0.743 (95% CI 0.708–0.778) in non‐cancer patients (p = 0.016). A PCT cut‐off of 0.5 ng/mL (PCT ≥ 0.5 ng/mL) confirmed its accuracy for predicting sepsis in non‐cancer patients (sensitivity 71.5%, specificity 64.1%) but the specificity fell to 44.7% in cancer patients, although sensitivity remained good (sensitivity 78.9%). Conversely, a higher PCT cut‐off of 1 ng/mL, as the most accurate threshold identified in the present study in the cancer population, showed a sensitivity of 66.9% and specificity of 61.2% in predicting sepsis in cancer patients.

Our study confirms the clinical role of PCT as a part of the diagnostic algorithm for sepsis but its diagnostic role is sub optimal in cancer patients.