The SSS revolution in fungal diagnostics: speed, simplicity and sensitivity

Abstract

Introduction

Fungal disease has historically presented a diagnostic challenge due to its often non-specific clinical presentations, relative infrequency and reliance on insensitive and time-intensive fungal culture.

Sources of data

We present the recent developments in fungal diagnostics in the fields of serological and molecular diagnosis for the most clinically relevant pathogens; developments that have the potential to revolutionize fungal diagnosis through improvements in speed, simplicity and sensitivity. We have drawn on a body of evidence including recent studies and reviews demonstrating the effectiveness of antigen and antibody detection and polymerase chain reaction (PCR) in patients with and without concurrent human immunodeficiency virus infection.

Areas of agreement

This includes recently developed fungal lateral flow assays, which have a low cost and operator skill requirement that give them great applicability to low-resource settings. Antigen detection for Cryptococcus, Histoplasma and Aspergillus spp. are much more sensitive than culture. PCR for Candida spp., Aspergillus spp., Mucorales and Pneumocystis jirovecii is more sensitive than culture and usually faster.

Areas of controversy

Effort must be made to utilize recent developments in fungal diagnostics in clinical settings outside of specialist centres and integrate their use into standard medical practice. Given the clinical similarities of the conditions and frequent co-infection, further study is required into the use of serological and molecular fungal tests, particularly in patients being treated for tuberculosis.

Growing points

Further study is needed to clarify the utility of these tests in low-resource settings confounded by a high prevalence of tuberculosis.

Areas timely for developing research

The diagnostic utility of these tests may require revision of laboratory work flows, care pathways and clinical and lab coordination, especially for any facility caring for the immunosuppressed, critically ill or those with chronic chest conditions, in whom fungal disease is common and underappreciated.

Introduction

Fungal infection has an extensive global disease burden, probably causing over 1.5 million deaths annually and affecting up to a billion people each year.¹ Despite this huge burden, fungal disease is underappreciated by the international medical community²; a large and gradually expanding number of fungal infections form part of the WHO’s list of neglected tropical diseases (NTDs).³ Not only are immunocompromised patients at much higher risk of invasive fungal disease, but adults and children in intensive care (including people with severe influenza or coronavirus disease 2019 [COVID-19]), with diabetes, on renal replacement therapy and with chronic lung disease are at substantially higher risk. Endemic and implantation fungal disease primarily affects people in poorer nations and the poorer, more rural populations within those nations.⁴ Fungal disease is often difficult or impossible to accurately diagnose clinically, for example pulmonary fungal infections can mimic malignancy⁵ and tuberculosis.^6,7 This makes clinical diagnosis of fungal disease difficult, particularly considering that initial clinical assessment of fungal disease in developing countries is performed primarily by frontline healthcare workers with only basic training.³ The burden of unidentified fungal disease has been shown to be high via autopsy discovery, even in developed nations.⁸

A missed or mistaken diagnosis in fungal disease can have significant consequences. The correct diagnosis, and therefore the correct treatment, can make a huge difference to morbidity and mortality in many fungal diseases. Cryptococcal meningitis’ mortality falls from as high as 100% to as low as 20% with correct treatment⁹ and disseminated histoplasmosis mortality falls from as high as 100% to as low as 25%.¹⁰ Fungal disease that lacks the ability to kill people can still cause major morbidity without treatment, such as eye loss following fungal keratitis.¹¹ Therefore, fungal diseases are of potentially high consequence when missed, and laboratory testing must rise to meet this diagnostic challenge.

The traditional ‘gold’ standard of testing for fungal disease, fungal culture, is slow, labour intensive and has a poor sensitivity, as summarized in Table 1.

Table 1.

Fungal culture sensitivity for different fungal genera and TAT in invasive disease

Fungus	Sample type and disease form	Culture sensitivity	TAT	References
Aspergillus	Sputum in chronic pulmonary disease	30%	48–92 hours	12
Candida	Blood in possible sepsis	30–50%	36–128 hours	13–15
Pneumocystis	Not culturable	0%	NA	16
Cryptococcus	CSF in meningitis	80–95%	Up to 7 days	17 , 18
Histoplasma	Sputum in chronic pulmonary disease	<50%	3–12 weeks	7 ,  19
Mucorales	Specimens from multiple organs	50%	3–5 days	20 ,  21

Fungal culture is at best a copper standard. Fungal cultures often become positive too late in the disease course to be clinically useful, if positive at all.²² There is utility for identification of fungi through microscopy which can be rapid and sensitive,³ but this also has limited sensitivity, and requires the availability of skilled operators and can be imprecise given the similar microscopic appearance of several fungi.²³ This has led to the development of serological and molecular tests that have the potential to revolutionize fungal diagnostics in terms of speed, accuracy and cost effectiveness.²³ Most practical for low-resource settings are the lateral flow assays (LFAs) of various types, as summarized in Table 2. However, the variety of monoclonal antibodies and antigens used in LFAs produced by different manufacturers mean that there are variations in their sensitivity and specificity. All data below regarding LFAs must be considered specific to the brand in question. Further study is required to compare available LFAs.²⁴

Table 2.

Examples of commercially available fungal LFAs in 2023

Product	Target	Manufacturer	Examples of published data
Aspergillus GM LFA	Galactomannan	IMMY Diagnostics25	26 ,  27
Aspergillus Galactomannan Detection K-Set	Galactomannan	Era Biology28
AspLFD	Aspergillus J5 antigen	OLM Diagnostics29	26 ,  30,31
Aspergillus ICT IgG–IgM	Aspergillus IgG + IgM	LDBio Diagnostics32	33
Aspergillus IgG Antibody Detection K-Set	Aspergillus IgG	Era Biology28
Cryptococcal antigen LFA	Cryptococcal Capsular Polysaccharide	IMMY Diagnostics34	35 , 36
Cryptococcal antigen LFA	Cryptococcal Capsular Polysaccharide	BioSynex37	38
Cryptococcal antigen LFA	Cryptococcal Capsular Polysaccharide	Dynamiker39	40
CrAg FungiXpert LFA	Cryptococcal Capsular Polysaccharide	Era Biology28	41
Candida Mannan Detection K-Set	Candida mannan antigen	Era Biology28	42
Candida Mannan IgM Antibody Detection K-Set	Candida mannan IgM	Era Biology28	42
Coccidioides Antibody LFA	Coccidioides IgM	IMMY Diagnostics43	44
Histoplasma Urine Antigen LFA	Histoplasma galactomannan	Optimum Imaging Diagnostics45	46
Histoplasma Urine Antigen LFA	Histoplasma galactomannan	MiraVista47	48 , 49

Aspergillosis

Serology

The most appropriate serological testing for disease caused by Aspergillus depends upon the type of infection. Aspergillus primarily affects the lungs and the way in which it does so is determined by a complex interplay between the fungus and the host immune system.⁵⁰ Disease varies from a primarily immune-driven hypersensitivity disease in the form of allergic bronchopulmonary aspergillosis (ABPA) to chronic or invasive chest disease with a high fungal load in the form of chronic pulmonary aspergillosis (CPA) or invasive aspergillosis (IA), usually involving the lungs as invasive pulmonary aspergillosis (IPA).⁵⁰ Antigen testing for Aspergillus primarily involves two tests: galactomannan (GM) and b-D-glucan (BDG). GM is a polysaccharide component of the Aspergillus cell wall, which is released during growth of the fungus,⁵¹ as is BDG.⁵² Within the spectrum of Aspergillus disease, antigen detection (usually by ELISA) is best used for acute disease forms with higher fungal loads, principally IA including IPA (Table 4).⁵³

Table 4.

Summary of utility of Aspergillus investigations for the main forms of aspergillosis

Investigation	Diagnostic utility
	IA including IPA	CPA	ABPA
Microscopy	++	+	−
Culture	+	++	+
GM	++ (serum) +++ (BAL)	+ (serum) ++ (BAL)	−
BDG	++	+	−
IgG/IgM	+	+++	++
IgE	−	+	+++
Asp. PCR serum	++	−	−
Asp. PCR respiratory	++	+++	+

GM ELISA assay of bodily fluids, particularly of fluid from bronchoalveolar lavage (BAL),⁵⁴ is more sensitive than culture for diagnosis of IA.⁵⁵ IA tends to affect immunocompromised or critically ill patients. The sensitivity of GM detection varies depending on whether the patient is neutropenic or not.⁵⁶ Study of neutropenic patients with haematological malignancies demonstrated sensitivities of GM of up to 97.4% using an optical density cut-off of 0.5,⁵⁶ although prior antifungal therapy reduces this.⁵⁷ Multiple meta-analyses have demonstrated the effectiveness of GM ELISA, with pooled sensitivities and specificities of 71–92 and 89–98% respectively.^58–61 Non-neutropenic intensive care patients have more variable optical densities with higher levels (i.e. > 3.0) having a very high probability of IA, and ODs less than 0.5 making IA unlikely.⁶² GM is effective in patients with solid organ transplants, particularly when testing BAL.⁶³ GM testing of endotracheal aspirates (which are much easier to obtain than BAL) is useful in the diagnosis of IA in intubated patients, as has been studied during the COVID -19 pandemic.⁶⁴ GM has a good specificity in serum (~90%)^14,65 and particularly in BAL (94% in IPA). It is important to note that even prophylactic antifungal therapy reduces the sensitivity of GM testing.⁶⁶ Important causes of false positive GM results include other fungal infections including histoplasmosis, fusariosis and talaromycosis.⁵⁵ Additionally, some medical therapies can induce false positives including Plasmalyte™ infusion⁶⁷ and some antibiotic preparations, though the specific issue with piperacillin-tazobactam induced false positivity is reported to be resolved.⁶⁸

While useful for IA, serum GM is only positive in about 11% of CPA cases,⁶⁹ where Aspergillus antibody detection (IgG), often referred to as precipitins, is preferred.⁵⁵ However, BAL GM has about a 75% sensitivity, if a bronchoscopy is done.⁷⁰ BAL GM testing has been shown to be a useful adjunct to BAL Aspergillus PCR testing.⁷¹ Unsurprisingly given the hypersensitive rather than the infective nature of the disease, there is minimal evidence that GM is diagnostic in ABPA.⁷² It is unclear if GM in sputum has any role in the diagnosis of CPA or ABPA.⁷³

Standard GM testing using ELISA requires sample heating prior to testing, meaning the test turnaround time (TAT) is typically 24–48 hours.¹² Several lateral flow antigen testing kits have recently been commercialized enabling rapid results. These have also been driven in part by the lack of laboratories in resource-constrained nations to run conventional GM ELISA testing,⁷⁴ and the common trend to send away tests to outside laboratories with long TATs. Galactomannan lateral flow assays (GM-LFA) are self-contained sandwich immunochromatographic test with serum or BAL compatibility. TAT can be dropped to less than an hour, and the results can be read by eye or with an automated reader.⁷⁵ Studies show varying results in sensitivity and sensitivity of GM-LFA for IA depending on the patient group and whether an automated reader is used or not (see Table 3). Generally, there is good evidence to support use of GM-LFA in patient with haematological malignancy with suspected IA, but further study is required for other groups.⁷⁵ There is a current lack of studies as to the variation of the performance of GM-LFA in neutropenic versus non-neutropenic patients, as is seen in GM ELISA. An alternative antigen (J5) is also useful as an LFA,^26,76 with direct comparisons with GM assays limited by GM being a component of the diagnostic standard to define cases.

Table 3.

Study findings for GM-LFA testing in different patient populations

Patient group	Specimen tested	Sensitivity	Specificity	References
Haematological malignancy Automated reader Read by eye	Serum BAL Serum BAL	49%* 80% 91% 83–89%	95% 89% 91% 87–88%	77  78  79  26,80
Solid organ transplant Automated reader Read by eye	BAL BAL	100% 50%	41% 48%	78  81
Non-selected patient groups (mixed ICU and non-ICU) Automated reader Read by eye	BAL BAL BAL	80% 61% 88%	75% 80% 84%	78  81  27

^*In this study, the low sensitivity was replicated in conventional GM ELISA testing.⁷⁷

LFAs read by an automated reader vary in sensitivity depending on the ODI selected. Values shown in Table 3 use an ODI of 1.0. See Jenks et al for variation of values at different ODIs.⁷⁸

BDG plays a role in the diagnosis of IA but, but given that it is found in a much broader range of fungi, it is of limited use alone (Table 4) and should be combined with GM or molecular testing.⁵⁵ Potential causes of false positivity include β-lactam antibiotic use, human blood component transfusion and haemodialysis.⁵⁰ BDG has been suggested as a good test of exclusion for IA, but neither its sensitivity not specificity exceed 80% for IA, and in patients in critical care the utility of BDG is in doubt due to this suboptimal specificity.⁸² BDG testing of BAL samples have found a poor specificity⁸³ so this is not recommended. Like GM, BDG is inferior to fungal IgG testing for diagnosis of CPA⁸⁴ and plays essentially no role in the diagnosis of ABPA.⁸⁵

Serum testing for Aspergillus IgG (previously known as the fungal precipitins test) is fundamental for the diagnosis of CPA (Table 4).⁸⁶ It is also often positive in ABPA, and Aspergillus-associated bronchitis usually seen in patients with bronchiectasis.¹² Studies in CPA show a sensitivity for Aspergillus IgG of 70–90%, depending on the assay and cut-off used.⁸⁶ The use of Aspergillus IgM alone for diagnosis in CPA has not been fully evaluated but has been shown to have inferior sensitivity to IgG.⁸⁷  Aspergillus IgG antibody detection also has utility in non-neutropenic patients with subacute and acute IA, with about a 50–65% sensitivity in a mixed population of IA, including community-acquired IA.^88,89

There are a variety of Aspergillus IgG assays available which have a degree of variation in ease of use, cross reactivity with other fungi, and sensitivity.⁵³ The available ELISA platform sensitivities and specificities in CPA range from 75–96 to 97–100%, respectively.⁹⁰ A large meta-analysis of ELISA platforms found sensitivities and specificities in CPA ranging from 50–100 to 83–100%, respectively.⁹¹ Another meta-analysis concurred with this variability in sensitivity findings.⁹² ELISA platforms have a higher sensitivity and far faster TAT than classical precipitin detection or immunodiffusion in agar.⁹³ Western blot testing is also available with good sensitivities but with less supporting clinical data.⁵³ Subtle immune defects in the patient can lead to false-negative results of IgG testing.⁹⁴ While Aspergillus IgG plays an adjunct role in diagnosis of ABPA, raised serum Aspergillus IgE is the principal serological test for its diagnosis.⁵⁰

Recently, immune chromatography technology (ICT) (lateral flow) tests have been commercialized to detect Aspergillus IgG and IgM antibodies.³³ A study of 154 CPA patients in the UK demonstrated a sensitivity and specificity of 91.6% and 98.0%, respectively, with a TAT of less than 30 minutes,³³ and a French study of over 300 CPA cases demonstrated sensitivity and specificity of 88.9% and 96.3%.⁹⁵ A further study of 90 post tuberculosis CPA patients in Indonesia showed 80% sensitivity and 70% specificity.⁹⁶ A meta-analysis of four studies showed pooled sensitivity of 90%, specificity of 91%, an ROC area under the curve of 0.94 and diagnostic odds ratio of 57.2, for CPA diagnosis by ICT tests.⁹⁷ However, another recent study from India has demonstrated a lower sensitivity and specificity of 67.6% and 81%, respectively, in a group of 74 CPA patients with the same ICT test, though the sensitivity of serum Aspergillus IgG was also lower than usual in this cohort at 82.4%.⁹⁸ A recent study from India indicates that finger-prick blood samples can be tested successfully with this LFA.⁹⁹ Another Aspergillus IgG LFA has recently been launched, but no published data are yet available.

PCR tests

Polymerase chain reaction (PCR) for Aspergillus DNA is used to test both BAL samples and blood, with there being no significant difference between serum and whole blood (EDTA) testing.^55,100 Blood testing has a high sensitivity in neutropenic patients and has been highlighted as useful in the early diagnosis of IA as Aspergillus DNA is detectable in the blood at an earlier stage of infection than GM.⁸⁵  Aspergillus PCR testing sensitivity in blood is much higher in neutropenic than non-neutropenic patients; a study of a non-neutropenic ICU patient group found blood PCR sensitivities as low as 11%.⁸⁵  Aspergillus PCR testing is less impacted by antifungal prophylaxis than GM detection. Consensus is that Aspergillus PCR should be used in conjunction with other testing, particularly computerized tomography (CT) scanning and GM testing, to lead to early and rapid diagnosis of IPA, particularly in patients with haematological malignancy.^{49,53,55,84,101} However, there is currently no consensus as to how to incorporate serum or blood PCR into current diagnostic algorithms or which exact patient groups do or do not benefit from PCR testing. Years ago, standardization was problematic, but now, commercial PCR systems have comprehensively addressed this problem.⁵³

PCR can also be performed on respiratory samples including sputum and BAL. PCR of sputum has a similar sensitivity as high-volume culture for Aspergillus, though this remains < 50% (Table 4).¹⁰² Sputum detection of Aspergillus by PCR has been suggested as a screening tool for IA in patients with chronic lung disease.¹⁰³ PCR of BAL is useful and well-studied for IA, it had a sensitivity and specificity of 89% in a review,¹⁰⁴ and of high utility when paired with GM testing of BAL samples.⁵⁹ However, PCR testing of respiratory fluids is not necessarily quick. While laboratory processes for DNA extraction and PCR take only a few hours, in practice, TAT is 24–48 hours in a category 3 laboratory.¹² Cell-free serum Aspergillus DNA detection is currently being actively commercialized, which may speed up TAT.^105,106

Microscopy and culture

Culture and direct microscopy are important investigations for IA.⁵⁵ Direct microscopy has a sensitivity of less than 50%, but is rapid and can allow differentiation of Aspergillus from Mucorales.¹⁰¹ Culture of Aspergillus in suspected IPA should always be attempted but interpreted in the clinical context and other tests, as no growth does not exclude IA and growth may simply reflect colonization.⁵⁰ Positive culture may represent simple colonization of Aspergillus rather than invasive or chronic infection or airways aspergillosis (i.e. ABPA or Aspergillus bronchitis), especially if A. niger complex is grown.⁸⁶ Culture has an overall positivity rate for all forms of aspergillosis of less than 30% and a TAT of 48–96 hours. High-volume culture and multiple samples can increase yield.^12,102,107

A positive microscopy or culture is helpful confirmation of disease when antigen or BDG is positive; either test being negative is not sufficient to rule out IA. Positive direct microscopy should alert clinicians to the possibility of Aspergillus tracheobronchitis. However, culture of the fungus can be contributory to the diagnosis as it is far more commonly grown from bronchial samples in patients with CPA than with simple colonization.¹⁰⁶ In CPA, the role of microscopy on sputum or bronchoscopy for diagnosis has been poorly studied.

If a biopsy can be done (which is infrequent) histopathological visualization of tissue invasion by hyphae is diagnostic and has a sensitivity of approximately 90%.¹⁰⁸ Sampling location errors and insensitivity of histological staining of tissue are responsible for the less than perfect performance of biopsy and histology for fungal infections generally.

Candidiasis

Aside from mucosal candidiasis, life-threatening forms of Candida infections include candidaemia (bloodstream infection), tissue invasive candidiasis (deep-seated candidiasis) and peritoneal or intra-abdominal candidiasis, which may occur concurrently or independently of candidaemia.¹⁰⁹ Usually these are grouped together as ‘invasive candidiasis’ (IC) as it is rarely clear if tissue invasion has taken place when candidaemia is found. Tissue invasion can be in almost any organ, with commoner presentations being endocarditis, osteomyelitis, retina and vitreous infection, meninges or brain (especially in premature neonates) and the kidney. The overall mortality of all these forms of invasive candidiasis is similar—45–55% when diagnosed and treated. Early treatment (i.e. with 24 hours of a positive blood culture being drawn) of candidaemia brings down mortality by about 25%, especially if the patients do not already have septic shock.

Candidaemia is primarily diagnosed by blood culture, but fluconazole and other antifungal therapies substantially reduce its sensitivity.^110,111 Sensitivities can be as low as 30–50% and have a long TAT.¹³ Taking serial blood cultures can increase sensitivity up to 75%,¹¹¹ probably partly because of the increased volume of blood drawn. Recently, culture techniques have been developed to improve yield with one technique using culture on magnetic beads coated with a recombined human mannan-binding lectin leading to increased sensitivities of 90% compared to standard blood culture with a significantly shorter TAT.¹¹²

Additional serological tests of value in candidaemia and invasive candidiasis (IC) are serum mannan (Mn) and anti-mannan (A-Mn) antibodies and Candida albicans germ tube antibody (CAGTA) test.^13,113 Used together, Mn and A-Mn serum testing provides a sensitivity of 83% and specificity of 86% in IC (compared to blood culture) and the same study showed positivity of at least one of Mn and A-Mn in 73% of patients with candidaemia.¹³ Mn and A-Mn have been shown to become positive up to six days prior to blood cultures becoming positive.¹¹⁴ The CAGTA test alone had a sensitivity and specificity of 60% and 86% respectively, but when combined with BDG this rose to 97 and 47%, respectively, with a 97% negative predictive value (any test positive).

BDG also has a role in the diagnosis of IC though is not a specific test.¹¹¹ When using lower cut-off values for BDG, very high sensitivities can be generated.^115,116 Therefore, its high NPV can be used in critical care patients to help exclude IC,¹¹⁷ useful for antimicrobial stewardship. For example, in the same study in critical care, where BDG was used in combination with CAGTA test, BDG alone had a sensitivity and specificity of 87% and 53%, with an NPV of 90%.¹¹³ It has been highlighted that BDG has a reduced diagnostic utility in patients who have haematological disease.¹¹⁸

Deep-seated candidiasis, with or without candidaemia, has lower rates of positive blood cultures, and the deep-seated tissue infection can persist after the resolution of candidaemia.¹¹⁹ At least one-third of ICU IC cases have negative blood cultures.¹²⁰ In the absence of detectable candidaemia, direct microscopy, histopathology and culture of infected tissues can be diagnostic but severely limited by the ability to acquire a sample.¹¹¹ Mn and A-Mn have been shown to be useful in the diagnosis of IC in general¹²¹; other studies have highlighted its limitations,¹²² and there is a lack of evidence for their use as a diagnostic tool in IC without candidaemia. Mn and A-Mn in combination can provide a high negative predictive value for IC leading to the suggestion that it could be used to exclude IC and avoid unnecessary empirical antifungal therapy.^111,123 The same arguments pertain to BDG and CAGTA.

Candida PCR provides an alternative, and potentially faster, diagnostic tool to culture, though its potential appears underdeveloped.¹²⁴ A meta-analysis of over 4500 patients showed a sensitivity and specificity for the diagnosis of candidaemia of 95% and 92%, respectively.¹⁵ However, in deep tissue candidiasis, the role of PCR is less clear, partly because autopsy to confirm IC is so often not done. In one study of IC without candidaemia, PCR had a reduced sensitivity for IC only; deep-seated candidiasis (IC) (45%) compared to candidaemia (83%).¹²⁵ There is consideration as to whether testing of whole blood compromises PCR testing in patients with a low-level fungaemia.¹²⁴ There is interest in development of techniques to improve the utility of Candida PCR through concentration and lysis of blood samples to improve detection.^125–127

Opportunistic infections in AIDS

Cryptococcosis

For cryptococcosis and cryptococcal meningitis, diagnosis centres around detection of the fungus via microscopy, culture or antigen.¹⁷ Microscopy of cerebrospinal fluid with India ink is readily available but has a lower sensitivity (~86%) compared with culture or antigen; meaning it is unreliable as a sole investigation. This is a particularly a problem of low fungal burden within the CSF, seen in early disease, non-human immunodeficiency virus (HIV) infected patients, and in those on anti-retroviral therapy (ART).¹⁸ Culture of CSF is often considered the gold standard to diagnose cryptococcal meningitis, but growth can take 7 days and there are false negatives in those with low fungal burden.¹⁷ Culturing larger volumes of CSF can increase sensitivity of culture, one study showing using 100 μL of undiluted CSF versus 10 μL at five 1:10 dilutions improved the sensitivity from 82.4 to 94.2%.¹⁸ Studies with a specialized birdseed-based (Guizotia abyssinica) agar have shown faster detection rates and higher sensitivities.¹²⁸ Quantification of viable yeast cells in CSF is a valuable surrogate endpoint for clinical trials.

Cryptococcal antigen (CrAg) in CSF, serum, or plasma is an essential diagnostic tool and should be performed on CSF of any patient with HIV with suspected meningitis¹⁷ and any patient with unexplained lymphocytic meningitis or CSF pleocytosis. Detection of CrAg by latex agglutination is well established,¹²⁸ and studies have shown good sensitivities (90%) and specificities (100%) within the CSF.¹⁸ ELISA tests are also available for CrAg but are impractical in healthcare settings within minimal laboratory facilities.¹²⁹ A revolutionary step in low-resource diagnosis of cryptococcal meningitis is CrAg LFA testing which rapidly detects cryptococcal polysaccharide capsule using gold-conjugated anti-cryptococcal monoclonal antibodies directed against Cryptococcus neoformans.¹⁷ CrAg LFA can be used in low-resource settings and has a slightly higher sensitivity than standard CrAg tests.^18,130 CrAg detection has a 99% sensitivity in the blood when it is also present in the CSF, meaning CrAg blood testing has good utility.¹⁷ Studies in low-resource settings demonstrated the utility of the CrAg LFA using finger-prick blood testing.^131–133 CrAg LFA use in serum has shown effectiveness in screening AIDS patients for early asymptomatic cryptococcal infection.^134,135 This has led to the development of CrAg screening programs in many African countries.¹³⁶ However, sensitivity for C. gattii complex infections varies by LFA, the broadest species sensitivity being that of Immy.¹³⁷ BDG is a nonspecific marker of fungal infection and, while it can be detected in CSF in patients with cryptococcal meningitis and very high CrAg titres, it should not be used to rule in or out the disease.¹⁷

Pulmonary cryptococcal disease tends to go underdiagnosed due to the lack of diagnostic tests available.¹³⁸ Culture of Cryptococcus from BAL samples or pleural fluid with consistent radiology are the fundamentals of diagnosis.¹³⁹ CrAg testing is understudied in its utility for diagnosis of pulmonary cryptococcosis,¹³⁰ and serum analysis is usually negative unless there is also disseminated disease.¹⁴⁰ In a prospective screening approach in one US laboratory, the sensitivity and specificity of CrAg were 71 and 99% of 1454 BAL specimens tested.¹⁴⁰ While most common in the immunocompromised, both pulmonary and meningeal cryptococcal disease can be found in the apparently immunocompetent.^141,142 The performance of serum and CSF CrAg in cryptococcosis in immunocompetent individuals is understudied but is clearly valuable.¹⁴² Such infections are more likely to be caused by other species of Cryptococcus, notably the gattii variety, and some CrAg LFA tests have less sensitivity for these pathogens.¹⁴³

Pneumocystis jirovecii and Pneumocystis pneumonia

Pneumocystis jirovecii has co-evolved with humans and is usually acquired for the first time early in life. Patterns of infection are distinct in HIV-infected and other immunocompromised people, with much higher fungal burdens typically seen in HIV patients with established pneumonia. It is not culturable on microbiological media. P. jirovecii PCR in general has shown excellent utility in both diagnosing and excluding PCP.¹⁴⁴ This has been replicated in HIV-positive and -negative patient groups, with BAL testing having the highest sensitivities (100%).¹⁴⁵ There is an urgent need for the development of rapid, point of care testing for P. jirovecii in low resource healthcare settings.

Sputum, induced sputum and BAL fluids are all good samples for microscopy and PCR. Microscopic visualization of P. jirovecii in BAL or lung tissue samples is considered the gold standard of diagnosis of infection though it has been found to have a sensitivity of 55–78%.¹⁴⁶ Microscopy with the aid of anti-P. jirovecii immunofluorescent antibodies improves its sensitivity.¹⁴⁷ Microscopy is highly dependent on operator skill and can be markedly less sensitive if the patient takes ART or anti-Pneumocystis prophylaxis.¹⁴⁶ Some progress has been made in in vitro culture of P. jirovecii, but the results of successful studies have proven to be poorly replicable and culture currently plays no role in PCP diagnosis.¹⁶

BDG has been noted as a useful diagnostic test in PCP but has lower sensitivities and specificities in patients without HIV.¹⁴⁸ Recent meta-analysis has shown that as an individual test BDG lacks the sensitivity or specificity to diagnose or exclude PCP.¹⁴⁹ It should therefore be used in conjunction with other investigations, as has been studied.¹⁵⁰ Amongst cancer patients, combining BDG and PCR has been shown to be effective at diagnosis and exclusion of PCP.¹⁵¹

Histoplasmosis

The diagnostic utility of tests for Histoplasma infection depends upon the disease form of histoplasmosis. Culture remains the gold standard but has a very long incubation time (3–12 weeks) and low sensitivity.¹⁹ In disseminated disease, most commonly seen in the immunosuppressed, this is further complicated by the need to biopsy deep tissues, such as bone marrow, colon, lymph nodes or adrenal gland, in order to perform culture or histopathological examination.¹⁵² Microscopically, Histoplasma is a small intracellular yeast and can easily be mistaken for Leishmania or Candida glabrata; it is also similar in size to the fission yeast Talaromyces marneffei.

Testing for Histoplasma antigen is preferred to diagnose disseminated disease. Antigen testing of the urine is non-invasive and has a sensitivity of 75% within the first few weeks of infection, higher than serum.¹⁵³ Antigen testing for histoplasmosis has shown to be so effective in disseminated disease that it has been suggested to replace fungal culture as the gold standard diagnostic test.¹⁵⁴ Little published data from prospective evaluations is available for the only commercialized Histoplasma antigen LFA, but it is slightly less sensitive than the ELISA, and less specific.^155,156 The antigen detected is galactomannan, and there may be cross-reactivity with Aspergillus antigen and GM derived from food.

There are two commonly used antibody tests for Histoplasma antibody: immunodiffusion and complement fixation. However, immunodiffusion assays have an unacceptably low sensitivity in disseminated disease where patients are immunosuppressed. Complement fixation sensitivity is higher in disseminated disease at 73–94%. Both tests are liable to cross-reactivity with other fungi such as Blastomyces.¹⁵³  Histoplasma antibody tests are a more effective diagnostic tool (sensitivity > 85%) in chronic pulmonary histoplasmosis as patients are usually immunocompetent.⁷

Multiple PCR tests have been developed in research laboratories for Histoplasma detection for use in disseminated disease and demonstrated high sensitivities.¹⁵⁷ However, none are commercialized. One study examining use of Histoplasma PCR in serum and blood from HIV-positive patients showed a sensitivity and specificity of 96–98 and 88–97%, respectively.¹⁵⁸ Other small-scale studies have been promising,¹⁵⁹ but other studies have shown sensitivities as low as 18% showing that further study is needed.¹⁶⁰

Mucoromycosis

Laboratory diagnosis of mucoromycosis remains primarily based around microscopy, culture and histopathology.¹⁶¹ Histopathology is the gold standard method of diagnosis but requires tissue biopsy, fungal staining of tissue and a well-trained histopathologist.¹⁶² Culture is insensitive with 50% of histopathological diagnoses having negative cultures.²⁰ Histopathology cannot speciate the organism, which is possibly important due to the varying antifungal resistance profiles between different species of Mucorales.¹⁶³ While insensitive, culture is relatively rapid, taking 2–5 days, due to the rapid growth of Mucorales.²¹

Emerging molecular tests support the diagnosis of mucoromycosis in two ways; through rapid and accurate species identification of a positive fungal culture (using PCR or sequencing), or through direct PCR testing of tissues or blood.¹⁶¹ The ITS region of Mucorales DNA as a target of PCR is the present first line test for species identification of Mucorales from fungal culture.¹⁶⁴ Regarding PCR testing of other samples, studies have shown high sensitivities and specificities for testing of BAL samples.^160,165 Recent developments have begun to assess PCR testing of non-invasive samples with blood and urine testing both showing good sensitivity and specificities.^166,167 There are two commercialized PCR assays for Mucorales detection.

Conclusions

The development of antigen, antibody and molecular testing for fungal disease has been transformed over the last 10–15 years. The WHO has recognized this with inclusion of Cryptococcal, Aspergillus and Histoplasma antigen, Aspergillus IgG antibody, and Pneumocystis PCR on its Essential Diagnostic list, along with blood culture, fungal culture, direct microscopy and histopathology. This article has not addressed in any detail new developments in species identification nor antifungal susceptibility testing. A further important area of study involves the use of recently developed fungal serological tests, particularly LFAs, in the context of patients with suspected tuberculosis, where there is risk for misdiagnosis and cross-reactivity. Further verifying the practical validity and cost-effectiveness of fungal LFAs in low-resource settings, where HIV and tuberculosis prevalence are also high, could have great utility for healthcare in developing economies. Laboratories may need to revise their workflow for fungal disease, especially if their medical practitioners care for many patients who are immunocompromised, in critical care, or with chronic chest conditions, the groups most at risk.

Conflict of interest statement

The authors have no potential conflicts of interest.

Data availability

No new data were generated or analysed in support of this review.