Novel Risk Factors for Predicting the Development of Invasive Fungal Infections in Acute Myeloid Leukemia Patients Undergoing Induction and Re-Induction Chemotherapy

BACKGROUND: Invasive fungal infections (IFI) occur in 5% to 40% of patients with hematologic malignancies, most commonly in patients with acute myeloid leukemia (AML). Historically, at the University of Virginia (UVA), antifungal prophylaxis was not routinely used in patient undergoing induction chemotherapy for AML. With an antifungal treatment naïve patient cohort, this study had the unique opportunity to examine if specific patient characteristics, comorbidities, or diagnostic criteria were predictive for the development of an IFI.

METHODS: A retrospective chart review of 126 consecutive patients undergoing induction or re-induction chemotherapy for AML at UVA from July 2011 to December 2015 was conducted. The primary endpoint was to determine the rate of probable or biopsy proven IFIs within ninety days of chemotherapy initiation as characterized by the European Organization for Research and Treatment of Cancer/Invasive Fungal Cooperative Group and the National Institute of Allergy and Infectious Disease Mycoses Study Group criteria. Secondary endpoints include patient comorbidity and disease specific indices associated with IFI. Each induction or re-induction regimen was counted as a distinct event. Statistical methods include pairwise Wilcoxon tests, univariate analyses, and Kruskal-Wallis tests. Statistical significance was defined as a p-value <0.05.

RESULTS: Of 126 patients, there were a total of 176 distinct induction or re-induction events. Thirty-eight patient events (22%) received antifungal prophylaxis with fluconazole per attending physician preference. In patients (n=102, 81%) who did not receive antifungal prophylaxis during induction, there were a total of 19 (19%) patients with probable/proven IFI. There was a significant association with probable/proven IFI in regards to decreased bone marrow cellularity (p-value 0.018). There were no correlations between probable/proven IFI and body mass index, age of diagnosis, gender, nicotine abuse, length of neutropenia, cytogenetics, and bone marrow blast percentage. In the antifungal prophylaxis cohort, 3 patients (8%) had probable/proven IFI.

Another aspect reviewed is the potential concept of an immune reserve ratio (IRR). The IRR is the percentage of the bone marrow core cellularity divided by blasts - representing the bone marrow functionality at the time of leukemia diagnosis or relapse. IRR was 1.9 vs 1.4 vs 1.2 (p-value 0.038) for none vs possible vs probable/proven IFI cohorts.

DISCUSSION: This study reveals that decreased bone marrow cellularity and decreased IRR may be used to predict which AML patients undergoing induction chemotherapy are at higher risk of developing an IFI. Traditional risk factors, such as tobacco use, length and degree of neutropenia, were not found to be significant in this review. Other patient co-morbidities, such as hypertension, COPD, and chronic kidney disease, and AML disease specific characteristics, such as cytogenetic and bone marrow blasts, also did not demonstrate clinical significance.

To date, studies have not developed a method to quantify a patient's immune reserve at the time of AML diagnosis and risk of IFI. The IRR attempts to use the ratio of bone marrow cellularity and blasts as a surrogate marker of functional immune activity. This review suggests that the IRR is highly correlative of IFI development. Future studies need to be performed to validate this prospectively. UVA is currently conducting a prospective clinical trial examining the utility of IRR to predict IFI in adults with AML induction chemotherapy.