Abstract:
Objective To evaluate the 30-day mortality from fungemia in patients with malignant tumor and establish a predictive nomogram.
Methods A total of 112 patients from Tianjin Medical University Cancer Institute & Hospital were enrolled in the study. Clinical data of the patients with malignant tumor, having fungemia, from January 2010 to December 2021 were retrospectively analyzed. The cases were assigned into survival and non-survival groups based on each patient’s 30-day survival status. Independent risk factors of 30-day mortality from fungemia were identified using multiple Logistic regression and a Stepwise algorithm, and a predictive nomogram was developed.
Results The 30-day mortality from fungemia was 38.4%. The most common strains of fungi were Candida albicans (54 cases, 48.2%), Candida glabrata (26 cases, 23.2%), Candida tropicalis (13 cases, 11.6%), and Candida parapsilosis (12 cases, 10.7%). Kidney dysfunction (OR: 6.818, 95% CI: 2.244-23.310), ICU stay >5 days (OR: 8.737, 95%CI: 2.918-28.543), and metastasis (OR: 6.384, 95%CI: 2.067-21.647) were identified as independent risk factors of 30-day mortality from fungemia. The consistency index (C-index) of the model was 0.898, and the area under the receiver operating characteristic (ROC) curve (AUC) was 0.898 (95% CI: 0.839-0.957). When the cutoff value was set at 28.7, the false positive rate and sensitivity were 21.7% and 88.4%, respectively.
Conclusions The predictive nomogram established in the current study could predict the risk of death in patients with malignant tumor having fungemia, which would benefit the healthcare providers’ decisions regarding individual patient’s treatment and lead to an improved prognosis.