In patients with hematologic diseases, candidemia is the most common form of invasive candidiasis and is associated with high mortality rates up to 40% [12]. Neutropenia (neutrophil count <500 cells/mL) is a common risk factor and is usually related to cytotoxic chemotherapy or HSCT, use of broad-spectrum antibiotics gastrointestinal mucosal dysfunction, and indwelling vascular catheters [13]. Acute disseminated candidiasis or hepatosplenic candidiasis might also be seen in patients with prolonged neutropenia. Evidence of defining neutropenia as neutrophil count of less than 500 cells/mL followed previous study result [14].

At least 15 distinct Candida sp. can cause human diseases, but majority of invasive infections are caused by 5 species (Candida albicans, Candida tropicalis, Candida glabrata, Candida krusei and Candida parapsilosis). There is geographical, center-to-center variability in the prevalence of Candida species. Patient’s predisposing factors and previous antifungal use also have a considerable influence on the distribution of Candida species. Candidiasis control should therefore include monitoring the epidemiology of Candida species and resistance to various antifungal agents [15, 16].

Traditionally, C. albicans has been the most common species but through the decades, non-albicans species have been increasing in proportion. We compared the epidemiology of candidemia from previous studies (Table 1). C. glabrata was the most common species among non-albicans species in Korea, similar in the United States [16-19]. C. tropicalis disputed this place in Asia-pacific region [20]. C. tropicalis followed C. glabrata in Korea.

In a single tertiary-care hospital in Korea where all data were collected from adult patients with hematologic diseases, the trend of decreasing C. albicans and increasing non-albicans species was the same. It was notable that C. tropicalis was considerably high in frequency and was already reported in several studies to be the most common non-albicans species causing candidemia in hematology patients [21-24]. C. glabrata recently increased from 9.5% to 24.6%, following the global trend. A main cause of the shift toward non-albicans species could be an increased azole resistance (ex. C. glabrata, C. tropicalis) due to the increased use of prophylactic antifungal agents. Furthermore, there is an increased use of echinocandin in patients with hematologic diseases for prophylaxis and/or empirical use [21] which could lead to epidemiological change.

Uncommon Candida species include C. lusitaniae, C. guilliermondii, and C. auris. Patients with hematologic diseases were at particular risk of C. lusitaniae and C. guillermondii [25]. These species are known to have intrinsic resistance to certain antifungal agent; C. guilliermondii resistant to echinocandins and C. lusitaniae resistant to amphotericin B. Therefore, selecting antifungal agent is challenging and accurate identification of these species is important [26-28]. C. auris, a nosocomial fungus, has become a serious threat for health care facilities around the globe. It can spread readily and has caused numerous healthcare-associated outbreaks. Moreover, most strains are resistant to at least one antifungal agent [29].

Classic diagnostic methods for candidiasis are microscopy, histopathology and culture [15]. Tissue samples and body fluids from sterile sites must be collected aseptically and transported to the laboratory promptly. If it is candidemia, the preferred diagnostic would be blood culture [30]. Yeast isolation from normally sterile tissues or fluids is usually indicative of deep-seated infection; however, negative results do not exclude candidiasis. Identification of the isolate to species level is mandatory. Microscopy should make use of special stains such as Grocott-Gomori methenamine silver (GMS) and Periodic Acid Schiff (PAS). β-D-glucan (BDG) is a cell wall component of most fungi and is not specific to detect Candida species with high false positive rate. In contrast, BDG has high negative predictive value, suggesting that invasive candidiasis is unlikely when BDG is negative [15]. However, careful interpretation is needed as it may not be reliable in the early course of IFI. Immuno-histochemistry, in situ hybridization, and polymerase chain reaction (PCR) based procedures are evaluated but utility and accuracy has not been warranted yet [31].

For patients with hematologic diseases, the possibility of obtaining samples of deep tissues is not warranted due to the patient’s condition [32]. Therefore, many clinicians also rely on empirical evidence with prior exposure to antibacterial agents, existence of a central venous catheter (CVC), and recent abdominal surgery to establish diagnosis [15]. As the number of fungal infections has increased and as hosts have become more vulnerable, newer diagnostic tests are needed to identify the pathogens more quickly [33].

Breakthrough invasive candidiasis is also a problem in patients with hematologic diseases. Breakthrough infection is defined as an infection occurring in a patient receiving antifungal agents. Azole-resistant Candida sp., especially C. krusei, were major cause of breakthrough infection. Up to 45% had resistance to antifungal agent which the patient was receiving when the breakthrough infection occurred [34-36]. Therefore, changing antifungal agent to different class and testing susceptibility of cultured species are needed.

Chronic disseminated (hepatosplenic) candidiasis (CDC) is a unique clinical manifestation and almost seen in patients with hematologic diseases. CDC usually occur during recovery from neutropenia. Fever, right upper quadrant discomfort, nausea, elevate liver enzymes following recovery from neutropenia implicate CDC [37]. CT, MRI, or ultrasound imaging is helpful and small, peripheral, target-like abscesses are seen in the liver or spleen [31, 38].

The mortality rate is closely correlated with delayed initiation of appropriate antifungal treatment. Therefore, prompt treatment is needed to improve prognostic outcomes in patients [39]. In candidemia, echinocandin antifungals are recommended as initial therapy. Transition from an echinocandin to fluconazole, usually within 5–7 days, is suggested in non-neutropenic patients who are clinically stable, have isolates that are susceptible to fluconazole, and have negative repeat blood cultures following treatment [40]. Fluconazole (intravenous or oral) is an acceptable alternative first line therapy if the patient is not critically ill and is unlikely to have a fluconazole-resistant Candida species [31]. However, antifungal resistance is an emerging problem worldwide and hematology patients might already be exposed to antifungal drugs for prophylaxis, complicating the selection of appropriate antifungal therapy [15]. For example, C. glabrata and C. krusei are more likely to have azole resistance and C. parapsilosis is known to have intrinsic resistance to echinocandin but not clinically significant [41].

Follow-up blood cultures should be performed every day or every other day to check when candidemia has been cleared. Recommended duration of therapy for candidemia without obvious metastatic complications is at least 2 weeks after the documented clearance of Candida species from the bloodstream, provided that neutropenia and other symptoms have resolved. For other complications such as endocarditis, surgical intervention and longer treatment are needed. Therefore, work up with transthoracic (TTE) or transesophageal (TEE) echocardiography should be done in patients with candidemia [32]. To check other disseminated focus, all non-neutropenic patients with candidemia should have a dilated ophthalmological examination within the first week after diagnosis to check for endophthalmitis. In neutropenic patients, ophthalmological findings are usually minimal until recovery from neutropenia; therefore, examination could be performed within the first week after recovery from neutropenia [31]. CVCs should be removed as early as possible when the source of candidemia is presumed to be the CVC. In neutropenic patients, sources of candidemia other than a CVC (ex. gastrointestinal tract) predominate. Removal of CVC should be considered on an individual basis [31, 42].

With CDC, therapy should be continued until lesions resolve on repeated imaging. The usual duration is several months and premature discontinuation of antifungal therapy can lead to relapse. If chemotherapy or HSCT is required, it may not be delayed because of the presence of chronic disseminated candidiasis [43].

For oropharyngeal candidiasis in patients with neutropenia, fluconazole 7–14 days is recommended. Chronic suppressive therapy is usually unnecessary. Esophageal candidiasis always requires systemic antifungal therapy. Diagnostic trial of antifungal therapy is available before endoscopic examination and intravenous therapy is recommended if the patient cannot tolerate oral therapy [31].

Echinocandin, a treatment of choice, has so-called sanctuary sites in the body such as eye, kidney, and meninges; therefore, it might be unsuitable for infection in those sites [44]. In addition, there is an increase in the resistance of C. glabrata to echinocandin [45]. In contrast, C. parapsilosis demonstrates innately higher minimum inhibitory concentrations (MICs) to echinocandins. If echinocandin or fluconazole is inadequate, liposomal amphotericin B could be an alternative.

In patients with hematologic diseases, Aspergillus-related infection is the most common IFI, followed by Candida infections [46]. Incidence rates vary according to local epidemiology and many other factors, including environmental control. Aspergillus fumigatus complex is the most common species, followed by A. flavus, A. niger, and A. terrus [47]. Invasive aspergillosis (IA) is a severe IFI, which manifests mostly as invasive pulmonary aspergillosis (IPA). A. fumigatus is responsible for most cases of IA but more commonly colonizes the respiratory tract. In contrast, A. flavus and A. niger often colonize burn wounds, and A. terreus causes only invasive disease, and usually has a poor prognosis [1, 48-50]. Some species are known to have variable susceptibility to antifungal agents. A. terreus is infrequently susceptible to amphotericin B, whereas A. calidoustus and A. lentulus are resistant to multiple antifungal agents, including amphotericin B and voriconazole [50]. Azole resistance in A. fumigatus is also increasing [51], thus raising a problem.

Diagnosis of IA remains difficult in many reasons. Both microscopy and culture should be attempted and tissue invasion by hyphae would provide a definitive diagnosis of IA. However, examinations such as tissue sampling may be difficult in hematology patients because of neutropenia and thrombocytopenia and yield of culture is frequently suboptimal [52]. Therefore, non-invasive diagnostic tests including non-culture-based tests have gained interest for improving diagnostic accuracy.

A high level of suspicion with meticulous physical examination is important since an immunocompromised patient may be relatively asymptomatic, interrupting early diagnosis [50]. The most prevalent but unspecific sign of IPA reported in approximately 100% of patients is persistent fever despite treatment with broad-spectrum antibiotics [53]. Imaging is a critical component in the diagnosis, and CT scan is recommended whenever there is clinical suspicion or when the patients present with 72–96 hours of persistent neutropenic fever to find IPA or sinusitis regardless of simple chest radiograph results [54]. Routine contrast imaging is not recommended but could be considered if vessel involvement is suspected. Typical findings of IPA include consolidation with surrounding ground-glass opacity (halo sign) or cavitation [55]. Lung ultrasound (US) imaging has been considered as an alternative modality to diagnose invasive fungal pneumonia, including US-guided transthoracic needle aspiration. Considering it is a rapid, bedside, non-invasive technique, lung US could have a significant role but more study is required for expected accuracy [56-58].

Bronchoscopy with bronchoalveolar lavage (BAL) is recommended in patients in the suspicion of IPA. It seems to be a valuable diagnostic tool with high yield and low complication rate even in neutropenic patients with hematologic diseases. Therefore, regarding the diagnostic delay and impact on mortality, bronchoscopy with BAL should be considered if possible [59-61]. Routine fungal culture, cytology, and non-culture-based methods should be performed. Galactomannan (GM) detection is more sensitive than culture in the diagnosis of IA. Additionally, BAL fluid GM is more sensitive compared to serum GM in terms of diagnostic performance [62]. Serum assays for BDG with comparison to GM suggest higher sensitivity but are not a specific marker for Aspergillus and may produce frequent false positive results. Nucleic acid testing in clinical specimens such as PCR has become available and promising but still controversial due to the lack of conclusive validation [8]. Its negative predictive value is high but positive predictive value is low, therefore, conjunction with other diagnostic tests and clinical context is needed [63, 64]. If imaging findings implicate fungal infection with or without positive biomarkers (ex. GM, PCR), probable/possible IFI could be considered. As pulmonary infection is the most frequent form of IFIs in patients with hematologic diseases, we depicted the diagnostic process of invasive fungal pneumonia as an example (Fig. 1).

For suspected IA, current guideline recommends voriconazole or isavuconazole as the treatment of choice. Isavuconazole is a mold-active triazole antifungal agent and is also susceptible to other molds including mucormycosis. Therefore, isavuconazole could be considered when coinfection of IA with other molds cannot be excluded [65]. If contraindicated or not tolerated, liposomal amphotericin B may be considered as an alternative. Amphotericin B deoxycholate is no longer a primary option but can be used in limited settings when no other antifungal agents are available [66]. Itraconazole should not be used as first line therapy although it is a potential option due to drug-drug interactions. Echinocandin also is not recommended because outcomes were not favorable with monotherapy [67]. Treatment should be continued for a minimum of 6–12 weeks, largely dependent on the degree and duration of immunosuppression, sites of infection, and evidence of disease improvement. Higher GM level is known as a risk factor related to mortality and follow up can be used to monitor disease progression or therapeutic response [68].

When persistent fever exists but no infiltrate seen in chest imaging, possible focus other than the lung should be considered. If there are no suspected focus and biologic markers such as GM, and PCR is negative, no empirical treatment would be required. However, in patients under mold-active prophylaxis, therapeutic drug monitoring (TDM) for prophylactic azoles such as posaconazole and voriconazole should be done and breakthrough IA should be considered [69]. Triazole is not recommended for empiric therapy of breakthrough infection during triazole prophylaxis and a different class of mold-active antifungal agent is required. When fever of unknown origin persists in patients without mold-active prophylaxis, empiric antifungal therapy with triazole might be considered [14, 70].

Reducing the dose or cessation of immunosuppressive agents could be a component of aspergillosis treatment. Colony-stimulating factors in neutropenic patients may also be considered. Granulocyte transfusion can be considered if the patient is unlikely to respond to standard therapy and if the anticipated duration of neutropenia is more than one week [71]. Surgery should be considered in localized disease [ex. invasive fungal sinusitis, endocarditis, osteomyelitis, focal invasion of central nervous system (CNS)]. IA is not an absolute contraindication to additional chemotherapy or stem cell transplantation. Decision must consider the risk of progressive aspergillosis during anti-neoplastic therapy versus the risk of clinical deterioration due to the aggravation of underlying hematologic diseases during delayed treatment [72, 73].

Voriconazole undergoes extensive hepatic metabolism and share metabolic pathways with many drugs. Therefore, TDM and drug interaction (ex. cyclosporin, tacrolimus, sirolimus and other CYP3A4 substrates) must be monitored for optimized therapeutic efficacy and to avoid potential toxicities. Genetic polymorphisms in CYP2C19 also contribute to serum drug levels. Common side effects include photopsia, photosensitivity, CNS disturbance, and prolonged QT interval. Amphotericin B deoxycholate, which is no longer used as primary IA treatment has many side effects including infusion-related toxicity and renal toxicity. However, some centers in Korea still use amphotericin B deoxycholate as the first line of treatment due to insurance and economic issues [73, 74].

In the mid-20th century, diabetes mellitus used to be a major risk factor for mucormycosis. However, in recent years, underlying malignancy emerged as another important risk factor due to the increasing number of patients undergoing chemotherapy, immunotherapy, or HSCT [81, 82]. The most frequently reported pathogens are Rhizopus spp., Mucor spp., Cunninghamella, Lichtheimia spp. and Rhizomucor spp. Incidence rates vary by region or center, indicating geographical variation.

The diagnosis of mucormycosis depends on the availability of imaging techniques, trained personnel, and mycological and histological investigations. In patients with hematologic diseases and suspected pulmonary mucormycosis, chest CT scan is recommended. Typical finding includes ground glass opacity surrounded by a ring of consolidation (reversed halo sign).

If mucormycosis is a potential diagnosis, biopsy is strongly recommended. It is usually suspected based on the results of direct microscopy. However, diagnosis with histomorphological basis is challenging because misidentification of Mucorales as Aspergillus species is common [83]. Culture of specimen is also strongly recommended for species identification and antifungal susceptibility testing. Frankly, identification of causative Mucorales does not obviously guide the choice of antifungal treatment because treatment should be started before the report. However, clinical course may be different depending on the species, leading to a better epidemiologic understanding. Immunohistochemistry with monoclonal antibodies or PCR could also be done, although still not widely available [75]. BDG is usually negative in patients with mucormycosis, as these fungi do not produce BDG [84]. However, about one-third of patients with pulmonary mucormycosis had concomitant opportunistic fungal infection according to previous study [85]. Cases of mixed fungal infection have also been reported [86]. Therefore, BDG, GM is recommended in suspected mucormycosis to support differential diagnosis or mixed IFI [75].

Suspected and confirmed mucormycosis are medical and surgical emergencies and require rapid management. Breakthrough mucormycosis during mold-active prophylaxis is also possible [87, 88] and must be considered. Early complete surgical debridement with clean margins is strongly recommended for disease control, histopathology, and microbiologic diagnosis [89, 90]. In neutropenic patients, immediate surgery may be required; however, some still prefer surgical resection after the resolution of neutropenia and thrombocytopenia in IA [91].

For first line antifungal therapy, liposomal amphotericin B with dose of 5–10 mg/kg/day is recommended. When CNS involvement is suspected, dose of 10 mg/kg/day is needed [92]. Increased dose tends to have an increased response rate, but doses higher than 10 mg/kg/day did not result in higher blood concentrations and instead increased the creatinine level [93, 94]. Amphotericin B deoxycholate has been the drug of choice for decades but its use is limited due to toxicity. It is only recommended when there is no other option. Isavuconazole has similar efficacy with liposomal amphotericin B and could be used as first line treatment for mucormycosis [95]. Posaconazole is recommended as an alternative therapy if first line medication has side effect or is ineffective [75]. Two oral forms (tablet and syrup) are available. Posaconazole syrup is influenced by food and concomitant use of other drugs such as proton pump inhibitors. Therefore, TDM must be monitored for optimal blood trough levels [96, 97].

The duration of therapy in mucormycosis is unknown. In general, weeks to months of therapy are usually given. If the underlying immune defect (ex. neutropenia or use of immunosuppressants) is resolved, therapy can be continued until there is resolution of signs and symptoms or radiographic improvement. There is moderate support for intravenous treatment until stable disease is achieved. The decision to switch to oral monotherapy with isavuconazole or posaconazole depends on the patient’s response to therapy and severity of the illness [75, 98].