Closing the Gap in Surveillance and Audit of Invasive Mold Diseases for Antifungal Stewardship Using Machine Learning

Utilizing natural language processing and large language models in the diagnosis and prediction of infectious diseases: A systematic review

BACKGROUND

Natural Language Processing (NLP) and Large Language Models (LLMs) hold largely untapped potential in infectious disease management. This review explores their current use and uncovers areas needing more attention.

METHODS

This analysis followed systematic review procedures, registered with the Prospective Register of Systematic Reviews. We conducted a search across major databases including PubMed, Embase, Web of Science, and Scopus, up to December 2023, using keywords related to NLP, LLM, and infectious diseases. We also employed the Quality Assessment of Diagnostic Accuracy Studies-2 tool for evaluating the quality and robustness of the included studies.

RESULTS

Our review identified 15 studies with diverse applications of NLP in infectious disease management. Notable examples include GPT-4's application in detecting urinary tract infections and BERTweet's use in Lyme Disease surveillance through social media analysis. These models demonstrated effective disease monitoring and public health tracking capabilities. However, the effectiveness varied across studies. For instance, while some NLP tools showed high accuracy in pneumonia detection and high sensitivity in identifying invasive mold diseases from medical reports, others fell short in areas like bloodstream infection management.

CONCLUSIONS

This review highlights the yet-to-be-fully-realized promise of NLP and LLMs in infectious disease management. It calls for more exploration to fully harness AI's capabilities, particularly in the areas of diagnosis, surveillance, predicting disease courses, and tracking epidemiological trends.

Navigating fungal infections and antifungal stewardship: drug resistance, susceptibility testing, therapeutic drug monitoring and future directions

Antifungal stewardship refers to the rational use of antifungal agents. Historically, in some instances, the misuse or overuse of antifungal agents has predisposed patients to an elevated risk of systemic side-effects and treatment resistance, as well as increased healthcare costs. Superficial mycoses, such as onychomycosis, are sometimes treated without any diagnostic testing and is associated with a high likelihood of self-diagnosis and self-treatment, potentially leading to the emergence of resistance against commonly used antifungals like terbinafine. Practitioners need to ensure that a proper clinical diagnosis is backed up by appropriate testing. This may include the traditional light microscopy and culture; additionally, molecular techniques (such as polymerase chain reaction, terbinafine gene mutational analysis) and antifungal susceptibility testing are considerations as appropriate. The choice of antifungal agent should be guided by what is the standard of care in the location where the clinician practices as well as more broadly state and national prescription patterns. Recently, reports of treatment resistance concerning both superficial and deep fungal infections have added another layer of difficulty to clinical practice. This review aims to explore the phenomenon of antifungal drug resistance, and highlights the importance of adopting antifungal stewardship programs. We provide an overview of treatment resistance and mechanisms of resistance reported thus far in dermatophytes. Challenges of performing antifungal susceptibility testing and therapeutic drug monitoring are discussed, as well as principles, recommendations and future directions of antifungal stewardship programs.

Artificial Intelligence: Exploring utility in detection and typing of fungus with futuristic application in fungal cytology

Artificial Intelligence (AI) is an emerging, transforming and revolutionary technology that has captured attention worldwide. It is translating research into precision oncology treatments. AI can analyse large or big data sets requiring high-speed specialized computing solutions. The data are big in terms of volume and multimodal with the amalgamation of images, text and structure. Machine learning has identified antifungal drug targets, and taxonomic and phylogenetic classification of fungi based on sequence analysis is now available. Real-time identification tools and user-friendly mobile applications for identifying fungi have been discovered. Akin to histopathology, AI can be applied to fungal cytology. AI has been fruitful in cytopathology of the thyroid gland, breast, urine and uterine cervical lesions. AI has a huge scope in fungal cytology and would certainly bear fruit with its accuracy, reproducibility and capacity for handling big data. The purpose of this systematic review was to highlight the AI's utility in detecting fungus and its typing with a special focus on future application in fungal cytology. We also touch upon the basics of AI in brief.

Antifungal Stewardship in Invasive Fungal Infections, a Systematic Review

INTRODUCTION

Invasive fungal infections (IFI) are a group of life-threatening diseases associated with significant morbidity, mortality and high healthcare costs. Some modern management programs known as AFS (antifungal stewardship programs) have now been developed. The purpose of this systematic review is to evaluate the different declinations of antifungal stewardship programs (AFPs).

METHODS

Articles were systematically reviewed using the PRISMA checklist 2020. EMBASE and MEDLINE/PubMED were searched using the term "antifungal stewardship" (2012-2022 data) on 2 January 2023. Eligible studies were those that described an AFS and included an intervention, performance evaluation and outcome measures.

RESULTS

A total of 22/796 studies were included. Approximately two-thirds (16) were published between 2018 and 2022. 16 (72.7%) stated a minimal complete AFS team. 12 (54.5%) adopted a non-compulsory AFS approach, 6(27.3%) had an Educational AFS and 4(18.2%) a compulsory AFS. Cost analyses of 12 studies showed a decrease for 7 (31.8%) compared to an increase for 5 (22.7%). In terms of outcomes, 18 studies showed a lower (10;45.5%) or the same (8;36.4%) pre-post intervention mortality rate.

CONCLUSION

AFS programs seem to be related to lower costs and better outcomes and should thus be implemented in tandem with antimicrobial stewardship programs.

Outbreaks of Fungal Infections in Hospitals: Epidemiology, Detection, and Management

Nosocomial clusters of fungal infections, whilst uncommon, cannot be predicted and are associated with significant morbidity and mortality. Here, we review reports of nosocomial outbreaks of invasive fungal disease to glean insight into their epidemiology, risks for infection, methods employed in outbreak detection including genomic testing to confirm the outbreak, and approaches to clinical and infection control management. Both yeasts and filamentous fungi cause outbreaks, with each having general and specific risks. The early detection and confirmation of the outbreak are essential for diagnosis, treatment of affected patients, and termination of the outbreak. Environmental sampling, including the air in mould outbreaks, for the pathogen may be indicated. The genetic analysis of epidemiologically linked isolates is strongly recommended through a sufficiently discriminatory approach such as whole genome sequencing or a method that is acceptably discriminatory for that pathogen. An analysis of both linked isolates and epidemiologically unrelated strains is required to enable genetic similarity comparisons. The management of the outbreak encompasses input from a multi-disciplinary team with epidemiological investigation and infection control measures, including screening for additional cases, patient cohorting, and strict hygiene and cleaning procedures. Automated methods for fungal infection surveillance would greatly aid earlier outbreak detection and should be a focus of research.

The role of diagnostics-driven antifungal stewardship in the management of invasive fungal infections: a systematic literature review

Antifungal stewardship (AFS) programs are key to optimizing antifungal use and improving outcomes in patients with invasive fungal infections. Our systematic literature review evaluated the impact of diagnostics in AFS programs by assessing performance and clinical measures. Most eligible studies were from Europe and the United States (n = 12/17). Diagnostic approaches included serum β-1–3-D-glucan test (n/N studies, 7/17), galactomannan test (4/17), computed tomography scan (3/17), magnetic resonance (2/17), matrix-assisted laser desorption and ionization time-of-flight mass spectrometry (MALDI-TOF MS; 2/17), polymerase chain reaction (1/17), peptide nucleic acid fluorescent in situ hybridization (PNA-FISH) assay (1/17), and other routine methods (9/17). Time to species identification decreased significantly using MALDI-TOF and PNA-FISH (n = 2). Time to targeted therapy and length of empiric therapy also decreased (n = 3). Antifungal consumption decreased by 11.6%–59.0% (7/13). Cost-savings ranged from 13.5% to 50.6% (5/10). Mortality rate (13/16) and length of stay (6/7) also decreased. No negative impact was reported on patient outcomes. Diagnostics-driven interventions can potentially improve AFS measures (antifungal consumption, cost, mortality, and length of stay); therefore, AFS implementation should be encouraged.

Core Recommendations for Antifungal Stewardship: A Statement of the Mycoses Study Group Education and Research Consortium

In recent years, the global public health community has increasingly recognized the importance of antimicrobial stewardship (AMS) in the fight to improve outcomes, decrease costs, and curb increases in antimicrobial resistance around the world. However, the subject of antifungal stewardship (AFS) has received less attention. While the principles of AMS guidelines likely apply to stewarding of antifungal agents, there are additional considerations unique to AFS and the complex field of fungal infections that require specific recommendations. In this article, we review the literature on AMS best practices and discuss AFS through the lens of the global core elements of AMS. We offer recommendations for best practices in AFS based on a synthesis of this evidence by an interdisciplinary expert panel of members of the Mycoses Study Group Education and Research Consortium. We also discuss research directions in this rapidly evolving field. AFS is an emerging and important component of AMS, yet requires special considerations in certain areas such as expertise, education, interventions to optimize utilization, therapeutic drug monitoring, and data analysis and reporting.

When Azoles Cannot Be Used: The Clinical Effectiveness of Intermittent Liposomal Amphotericin Prophylaxis in Hematology Patients

Background

Patients unable to take azoles are a neglected group lacking a standardized approach to antifungal prophylaxis. We evaluated the effectiveness and safety of intermittent liposomal amphotericin B (L-AMB) prophylaxis in a heterogenous group of hematology patients.

Methods

A retrospective cohort of all hematology patients who received a course of intravenous L-AMB, defined as 1 mg/kg thrice weekly from July 1, 2013 to June 30, 2018, were identified from pharmacy records. Outcomes included breakthrough-invasive fungal disease (BIFD), reasons for premature discontinuation, and acute kidney injury.

Results

There were 198 patients who received 273 courses of L-AMB prophylaxis. Using a conservative definition, the BIFD rate was 9.6% (n = 19 of 198) occurring either during L-AMB prophylaxis or up to 7 days from cessation in patients who received a course. Probable/proven BIFD occurred in 13 patients (6.6%, 13 of 198), including molds in 54% (n = 7) and non-albicans Candidemia in 46% (n = 6). Cumulative incidence of BIFD was highest in patients with acute myeloid leukemia (6.8%) followed by acute lymphoblastic leukemia (2.7%) and allogeneic stem cell transplantation (2.5%). The most common indication for L-AMB was chemotherapy, or anticancer drug-azole interactions (75% of courses) dominated by vincristine, or acute myeloid leukemia clinical trials, followed by gut absorption concerns (13%) and liver function abnormalities (8.8%). Acute kidney injury, using a modified international definition, complicated 27% of courses but was not clinically significant, accounting for only 3.3% (9 of 273) of discontinuations.

Conclusions

Our findings demonstrate a high rate of BIFD among patients receiving L-AMB prophylaxis. Pragmatic trials will help researchers find the optimal regimen of L-AMB prophylaxis for the many clinical scenarios in which azoles are unsuitable, especially as targeted anticancer drugs increase in use.

Ushering in Antifungal Stewardship: Perspectives of the Hematology Multidisciplinary Team Navigating Competing Demands, Constraints, and Uncertainty

Background

The social, contextual, and behavioral determinants that influence care in patients at risk for invasive fungal diseases (IFD) are poorly understood. This knowledge gap is a barrier to the implementation of emerging antifungal stewardship (AFS) programs. We aimed to understand the barriers and enablers to AFS, opportunities for improvement, and perspectives of AFS for hematology patients at a major medical center in Australia.

Methods

Semistructured, face-to-face interviews were conducted with 35 clinicians from 6 specialties (hematology, infectious diseases, pharmacy, nursing, radiology, respiratory), followed by thematic analysis mapped to a behavioral change framework.

Results

Access to fungal diagnostics including bronchoscopy was identified as the key barrier to rational prescribing. Collective decision making was the norm, aided by an embedded stewardship model with on-demand access to infectious diseases expertise. Poor self-efficacy/knowledge among prescribers was actually an enabler of AFS, because clinicians willingly deferred to infectious diseases for advice. A growing outpatient population characterized by frequent care transitions was seen as an opportunity for AFS but neglected by an inpatient focused model, as was keeping pace with emerging fungal risks. Ad hoc surveillance, audit, and feedback practices frustrated population-level quality improvement for all actors. Antifungal stewardship was perceived as a specialized area that should be integrated within antimicrobial stewardship but aligned with the cultural expectations of hematologists.

Conclusions

Antifungal stewardship is multifaceted, with fungal diagnostics a critical gap and outpatients a neglected area. Formal surveillance, audit, and feedback mechanisms are essential for population-level quality improvement. Resourcing is the next challenge because complex immunocompromised patients require personalized attention and audit of clinical outcomes including IFD is difficult.