TABLE 2

Nonmolecular tests used for diagnosis of the most common invasive fungal infections a

MicroorganismDiagnostic testOptimal specimen typeSensitivity (%)Specificity (%)Reasons for false-positive resultsReasons for false-negative resultsComments
Candida spp.CulturesBlood50–6095NoneNoneGold standard. May take up to 3 days for a positive result.
Beta-glucan assaySerum77.6–81.387.1–92.4 (for patients not infected with fungal pathogens)Other fungal infection, dialysis filters made from cellulose, bacteremia?Hemolyzed samples, higher cutoff valuesA cutoff of 80 pg/ml is associated with higher accuracy; used as screening test for various fungal infections.
CAGTA assaySerum77–8991–100UnknownUnknownCAGTA assay is not affected by Candida colonization or intake of antifungal.
Aspergillus spp.HistopathologyVarious, depending on the infection site100100 Fusarium and Scedosporidium spp. have similar microscopic appearancesFormation of pseudoseptations by the organismMost accurate test is tissue biopsy. Used as a last resort in undiagnosed cases.
CultureVarious, depending on the infection site30–6872–100 Aspergillus sp. colonizationSlow-sporulating organisms, hematopoietic stem transplant recipientsGold standard, but with low sensitivity. PPV largely depends on the population tested.
Galactomannan assaySerum, BAL fluid, or CSF71 for serum, 90 for BAL fluid89 for serum, 94 for BAL Histoplasma sp. and Fusarium sp. infections, fungal colonizationSteroid treatmentOptimal diagnostic cutoff is not yet established. GM levels could be used to monitor the response to treatment.
Beta-glucan assaySerum55–9577–96Other fungal infection, Gram-negative bacteremia, dialysisHemolyzed samples, higher cutoff valuesA cutoff of 80 pg/ml is associated with higher accuracy; screening test for various fungal infections.
Lateral-flow device antigen detectionSerum and BAL fluid48–100100UnknownUnknownInterpretation is subjective; perhaps has a better performance than that of GM assay.
Pneumocystis spp.HistopathologyExpectorated sputum, induced sputum, or BAL fluid33–100, depending on stain and specimen used100UncommonVaries by specimen type and stainMethenamine silver stain on BAL fluid is the current gold standard. Toluidine blue stain on induced sputum may be the most cost-effective method. Pneumocystis does not grow easily in culture.
Beta-glucan assaySerum94.886.3Other fungal infection, bacteremia, dialysisUncommonExcellent screening test for high-risk patients, not useful for monitoring response to treatment.
Cryptococcus spp.CulturesCSF>95100UncommonUncommonGold standard, but takes 3–7 days for a positive result.
HistopathologyMostly CSF75100UncommonLow levels of microorganismIndia ink stain often used as a screening test.
Cryptococcal antigen test (LA, EIA, or LFD)CSF or serum97 for CSF, 87 for serum93–100 Trichosporon sp., Capnocytophaga sp., or Stomatococcus sp. invasive infectionsUncommonMost accurate test when performed on CSF. The three methods are comparable, although LA gives more false-positive results. LFD is best for rapid point-of-care diagnosis.
Histoplasma capsulatum CultureTissue, BAL fluid, or other bodily fluids85 for disseminated and acute pulmonary infections100UncommonLow fungal levels on specimenGold standard, but takes 2–4 weeks to grow.
HistopathologyTissue or BAL fluid76 for disseminated infection100UncommonLow fungal levelsUnacceptably low sensitivity, which is even lower for pulmonary infection.
Antibody tests (CF or ID)Serum75 for disseminated infection, 67 for acute pulmonary infection100UncommonLow fungal levelsBest performance with combination of the two methods. One study showed unacceptably low sensitivity for solid organ transplant patients.
Antigen testUrine and serum88–92100 for patients without fungal infectionHigh cross-reactivity in cases of Blastomyces sp. or Coccidioides sp. infectionUncommonMost accurate test overall, but shows cross-reactivity with other dimorphic fungi.
Blastomyces dermatitidis CultureSputum, BAL fluid, tissue86 for sputum, 92 for BAL fluid100UncommonLow levelsGold standard, grows better on fungal isolator cultures, takes a long time to grow.
HistopathologyVaries based on infection site46 for sputum, 90 for tissue100UncommonIncorrect specimenBroad-based budding.
Antibody test (CF and ID), new EIASerum57 for CF, 65–80 for ID, 88 for novel EIA37 for CF, 100 for ID, 100 for EIACross-reactivity with other dimorphic fungiLow levels of circulating antibodiesID method is clearly preferable due to higher performance. EIA has high specificity even in cases of histoplasmosis.
Antigen testMostly urine9399 for patients free of fungal infectionHigh cross-reactivity with other dimorphic fungi (e.g., 96% with histoplasmosis)UncommonMost accurate test, but with cross-reactivity issues
Coccidioides spp.CultureSputum or tissue90100UncommonLow fungal levelsGrows better than all endemic fungi. Culture is used mainly for hospitalized patients. Can grow within a week, but identification can take longer.
HistopathologySputum or tissue31–42100UncommonLow fungal levels on specimenSpherule detection.
Antibody assays (CF, TP, and novel EIA)Serum9599False-positive results have been reported for asymptomatic individualsUncommonMost commonly used test. EIA seems to have the best accuracy but is still not widely tested. A single positive IgM result must be interpreted based on pretest probability.
  • a BAL, bronchoalveolar lavage; CAGTA, Candida albicans germ tube antibody; CF, complement fixation; CSF, cerebrospinal fluid; EIA, enzyme immunoassay; GM, galactomannan; ID, immunodiffusion; LA, latex agglutination; LFD, lateral-flow device; PPV, positive predictive value; TP, tube precipitin.