Diagnostic accuracy of Candida albicans germ tube antibody for invasive candidiasis: systematic review and meta-analysis

Early diagnosis of candidemia with explainable machine learning on automatically extracted laboratory and microbiological data: results of the AUTO-CAND project

BACKGROUND

Candidemia is associated with a heavy burden of morbidity and mortality in hospitalized patients. The availability of blood culture results could require up to 48-72 h after blood draw; thus, early treatment decisions are made in the absence of a definite diagnosis.

METHODS

In this retrospective study, we assessed the performance of different supervised machine learning algorithms for the early differential diagnosis of candidemia and bacteremia in adult patients on a large dataset automatically extracted within the AUTO-CAND project.

RESULTS

Overall, 12,483 episodes of candidemia (1275; 10%) or bacteremia (11,208; 90%) were included in the analysis. A random forest classifier achieved the best diagnostic performance for candidemia, with sensitivity 0.98 and specificity 0.65 on the training set (true skill statistic [TSS] = 0.63) and sensitivity 0.74 and specificity 0.57 on the test set (TSS = 0.31). Then, the random classifier was trained in the subgroup of patients with available serum β-D-glucan (BDG) and procalcitonin (PCT) values by exploiting the feature ranking learned in the entire dataset. Although no statistically significant differences were observed from the performance measures obtained by employing BDG and PCT alone, the performance measures of the classifier that included the features selected in the entire dataset, plus BDG and PCT, were the highest in most cases.

CONCLUSIONS

Random forest classifiers trained on large datasets of automatically extracted data have the potential to improve current diagnostic algorithms for candidemia. However, further development through implementation of automatically extracted clinical features may be necessary to achieve crucial improvements.

Comparison of β-1-3-D-Glucan and Candida Mannan Biomarker Assays with Serological Tests for the Diagnosis of Candidemia

Invasive candidiasis, including bloodstream infection (candidemia), encompasses the most severe forms of Candida infection. Several species-specific and non-specific serological assays are commercially available to aid in diagnosis. This study compared the performance of five such biomarker assays. Serum samples from 14 patients with proven or probable invasive candidiasis, and from 10 control patients, were included in the analysis. A total of 50 serum samples were tested using C. albicans germ tube antibody (CAGTA) assay (Vircell), C. albicans IgM, C. albicans IgG and Candida mannan assays (Dynamiker Biotechnology). Among these samples, the β-1-3-D-glucan (BDG) assay (Fungitell), a laboratory standard for the diagnosis of invasive candidiasis, was positive in 20 (40%), intermediate in five (10%) and negative in 25 (50%). In cases of proven or probable candidemia, the sensitivity and specificity of the BDG assay was 86% and 80%, respectively; the Candida mannan assay, 14% and 86%; the CAGTA test, 57% and 60%; the C. albicans IgM assay, 71% and 60%; and C. albicans IgG assay 29% and 90%. In 4/8 (50%) cases with multiple serum samples, C. albicans IgM was positive sooner than BDG. Thus, when used as a rule-out test for invasive candidiasis, our data suggest that the C. albicans IgM assay may assist antifungal stewardship (over serum BDG).

Invasive candidiasis: current clinical challenges and unmet needs in adult populations

Invasive candidiasis (IC) is a serious infection caused by several Candida species, and the most common fungal disease in hospitals in high-income countries. Despite overall improvements in health systems and ICU care in the last few decades, as well as the development of different antifungals and microbiological techniques, mortality rates in IC have not substantially improved. The aim of this review is to summarize the main issues underlying the management of adults affected by IC, focusing on specific forms of the infection: IC developed by ICU patients, IC observed in haematological patients, breakthrough candidaemia, sanctuary site candidiasis, intra-abdominal infections and other challenging infections. Several key challenges need to be tackled to improve the clinical management and outcomes of IC patients. These include the lack of global epidemiological data for IC, the limitations of the diagnostic tests and risk scoring tools currently available, the absence of standardized effectiveness outcomes and long-term data for IC, the timing for the initiation of antifungal therapy and the limited recommendations on the optimal step-down therapy from echinocandins to azoles or the total duration of therapy. The availability of new compounds may overcome some of the challenges identified and increase the existing options for management of chronic Candida infections and ambulant patient treatments. However, early identification of patients that require antifungal therapy and treatment of sanctuary site infections remain a challenge and will require further innovations.

Developments in Fungal Serology

Purpose of Review

The true incidence of fungal disease is hampered by conventionally poor diagnostic tests, limited access to advanced diagnostics, and limited surveillance. The availability of serological testing has been available for over two decades and generally underpins the modern diagnosis of the most common forms of fungal disease. This review will focus on technical developments of serological tests for the diagnosis of fungal disease, describing advances in clinical performance when available.

Recent Findings

Despite their longevity, technical, clinical, and performance limitations remain, and tests specific for fungal pathogens outside the main pathogens are lacking. The availability of LFA and automated systems, capable of running multiple different tests, represents significant developments, but clinical performance data is variable and limited.

Summary

Fungal serology has significantly advanced the diagnosis of the main fungal infections, with LFA availability increasing accessibility to testing. Combination testing has the potential to overcome performance limitations.

Validation of an Automated System for the Extraction of a Wide Dataset for Clinical Studies Aimed at Improving the Early Diagnosis of Candidemia

There is increasing interest in assessing whether machine learning (ML) techniques could further improve the early diagnosis of candidemia among patients with a consistent clinical picture. The objective of the present study is to validate the accuracy of a system for the automated extraction from a hospital laboratory software of a large number of features from candidemia and/or bacteremia episodes as the first phase of the AUTO-CAND project. The manual validation was performed on a representative and randomly extracted subset of episodes of candidemia and/or bacteremia. The manual validation of the random extraction of 381 episodes of candidemia and/or bacteremia, with automated organization in structured features of laboratory and microbiological data resulted in ≥99% correct extractions (with confidence interval < ±1%) for all variables. The final automatically extracted dataset consisted of 1338 episodes of candidemia (8%), 14,112 episodes of bacteremia (90%), and 302 episodes of mixed candidemia/bacteremia (2%). The final dataset will serve to assess the performance of different ML models for the early diagnosis of candidemia in the second phase of the AUTO-CAND project.

Fungal infections diagnosis - Past, present and future

Despite the scientific advances observed in the recent decades and the emergence of new methodologies, the diagnosis of systemic fungal infections persists as a problematic issue. Fungal cultivation, the standard method that allows a proven diagnosis, has numerous disadvantages, as low sensitivity (only 50% of the patients present positive fungal cultures), and long growth time. These are factors that delay the patient's treatment and, consequently, lead to higher hospital costs. To improve the accuracy and quickness of fungal infections diagnosis, several new methodologies attempt to be implemented in clinical microbiology laboratories. Most of these innovative methods are independent of pathogen isolation, which means that the diagnosis goes from being considered proven to probable. In spite of the advantage of being culture-independent, the majority of the methods lack standardization. PCR-based methods are becoming more and more commonly used, which has earned them an important place in hospital laboratories. This can be perceived now, as PCR-based methodologies have proved to be an essential tool fighting against the COVID-19 pandemic. This review aims to go through the main steps of the diagnosis for systemic fungal infection, from diagnostic classifications, through methodologies considered as "gold standard", to the molecular methods currently used, and finally mentioning some of the more futuristic approaches.

EORTC/MSGERC Definitions of Invasive Fungal Diseases: Summary of Activities of the Intensive Care Unit Working Group

The EORTC/MSGERC recently revised and updated the consensus definitions of invasive fungal disease (IFD). These definitions primarily focus on patients with cancer and stem cell or solid-organ transplant patients. They may therefore not be suitable for intensive care unit (ICU) patients. More in detail, while the definition of proven IFD applies to a broad range of hosts, the categories of probable and possible IFD were primarily designed for classical immunocompromised hosts and may therefore not be ideal for other populations. Moreover, the scope of the possible category of IFD has been diminished in the recently revised definitions for classically immunocompromised hosts. Diagnosis of IFD in the ICU presents many challenges, which are different for invasive candidiasis and for invasive aspergillosis. The aim of this article is to review progresses made in recent years and difficulties remaining in the development of definitions applicable in the ICU setting.

Invasive Candida Infections in Neonates after Major Surgery: Current Evidence and New Directions

Infections represent a serious health problem in neonates. Invasive Candida infections (ICIs) are still a leading cause of mortality and morbidity in neonatal intensive care units (NICUs). Infants hospitalized in NICUs are at high risk of ICIs, because of several risk factors: broad spectrum antibiotic treatments, central catheters and other invasive devices, fungal colonization, and impaired immune responses. In this review we summarize 19 published studies which provide the prevalence of previous surgery in neonates with invasive Candida infections. We also provide an overview of risk factors for ICIs after major surgery, fungal colonization, and innate defense mechanisms against fungi, as well as the roles of different Candida spp., the epidemiology and costs of ICIs, diagnosis of ICIs, and antifungal prophylaxis and treatment.

Invasive candidiasis in critical care: challenges and future directions

Invasive candidiasis is the most common critical care-associated fungal infection with a crude mortality of ~ 40-55%. Important factors contributing to risk of invasive candidiasis in ICU include use of broad-spectrum antimicrobials, immunosuppressive drugs, and total parenteral nutrition alongside iatrogenic interventions which breach natural barriers to infection [vascular catheters, renal replacement therapy, extracorporeal membrane oxygenation (ECMO), surgery]. This review discusses three key challenges in this field. The first is the shift in Candida epidemiology across the globe to more resistant non-albicans species, in particular, the emergence of multi-resistant Candida glabrata and Candida auris, which pose significant treatment and infection control challenges in critical care. The second challenge lies in the timely and appropriate initiation and discontinuation of antifungal therapy. Early antifungal strategies (prophylaxis, empirical and pre-emptive) using tools such as the Candida colonisation index, clinical prediction rules and fungal non-culture-based tests have been developed: we review the evidence on implementation of these tools in critical care to aid clinical decision-making around the prescribing and cessation of antifungal therapy. The third challenge is selection of the most appropriate antifungal to use in critical care patients. While guidelines exist to aid choice, this heterogenous and complex patient group require a more tailored approach, particularly in cases of acute kidney injury, liver impairment and for patients supported by extracorporeal membrane oxygenation. We highlight key research priorities to overcome these challenges in the future.

Biomarkers of fungal infection: Expert opinion on the current situation

The introduction of non-culture-based diagnostic techniques is revolutionizing the world of microbiological diagnosis and infection assessment. Fungi are no exception, and the introduction of biomarkers has opened up enormous expectations for better management of these entities. Biomarkers are diverse, their targets are also diverse and their evaluation has been done preferably in an individualized use and with deficient designs. Less is known about the value of the combined use of biomarkers and the impact of the negativity of two or more biomarkers on antifungal treatment decisions has been poorly studied. Given the paucity of prospective, randomized and definitive studies, we have convened experts from different fields, with an interest in invasive fungal infections, to answer some questions about the current relevant use of fungal biomarkers. This document summarizes the answers of these experts to the different questions.