Challenges and Pitfalls of Morphologic Identification of Fungal Infections in Histologic and Cytologic Specimens

Transcription

1 Anatomic Pathology / Morphologic Identification of Fungi Challenges and Pitfalls of Morphologic Identification of Fungal Infections in Histologic and Cytologic Specimens A Ten-Year Retrospective Review at a Single Institution Ankur R. Sangoi, MD, 1 William M. Rogers, MD, 1 Teri A. Longacre, MD, 1 Jose G. Montoya, MD, 2 Ellen Jo Baron, PhD, 1 and Niaz Banaei, MD 1 Key Words: Culture; Cytology; Fungal; Fungi; Histology; Hyphae; Identification; Mold; Morphology; Yeast CME/SAM DOI: /AJCP99OOOZSNISCZ Upon completion of this activity you will be able to: discuss histopathologic challenges of identification of fungi in tissue sections and cytologic preparations. describe fungi with similar morphologic features on histopathology sections. describe different growth rates of various fungi. describe measures that improve diagnostic accuracy of morphologic identification of fungal infections in histology and cytology specimens. The ASCP is accredited by the Accreditation Council for Continuing Medical Education to provide continuing medical education for physicians. The ASCP designates this educational activity for a maximum of 1 AMA PRA Category 1 Credit per article. This activity qualifies as an American Board of Pathology Maintenance of Certification Part II Self-Assessment Module. The authors of this article and the planning committee members and staff have no relevant financial relationships with commercial interests to disclose. Questions appear on p 442. Exam is located at Abstract Despite the advantages of providing an early presumptive diagnosis, fungal classification by histopathology can be difficult and may lead to diagnostic error. To assess the accuracy of histologic diagnosis of fungal infections vs culture ( gold standard ), we performed a 10-year retrospective review at our institution. Of the 47 of 338 positive mold and yeast cultures with concurrent surgical pathology evaluation without known history of a fungal infection, 37 (79%) were correctly identified based on morphologic features in histologic and/or cytologic specimens. The 10 discrepant diagnoses (21%) included misidentification of septate and nonseptate hyphal organisms and yeast forms. Errors resulted from morphologic mimics, use of inappropriate terminology, and incomplete knowledge in mycology. The accuracy did not correlate with preceding antifungal therapy (P =.14) or use of special stains (P =.34) and was not operator-dependent. Among 8 discrepancies with clinical follow-up available, 2 potential adverse clinical consequences resulted. While histopathologic identification of fungi in tissue sections and cytologic preparations is prone to error, implementation of a standardized reporting format should improve diagnostic accuracy and prevent adverse outcomes. Histopathologic examination remains one of the major diagnostic tools in mycology because it permits rapid, presumptive identification of fungal infections. Histopathologic and/or cytopathologic examination can also provide insight into the diagnostic significance of some culture isolates. Demonstration of tissue invasion or an inflammatory reaction can help to determine whether an organism represents contamination, colonization, or true infection. 1 Furthermore, histopathologic examination remains the only reliable means to identify certain pathogens, including Pneumocystis jiroveci (formerly Pneumocystis carinii), Loboa loboi, and Rhinosporidium seeberi. 2,3 Several studies have demonstrated improved fungal detection by histopathologic examination in some circumstances, 4-11 but few have investigated the diagnostic specificity of histopathologic examination vs microbiological culture, 3,12-14 while none, to our knowledge, have included a comprehensive formal evaluation of the diagnostic accuracy of surgical pathology for all fungal infections. Although histopathology and microbiology are thought to be complementary, in recent years, we have encountered a number of cases with discrepant histologic and culture results at the time of frozen section or at final diagnosis. Because some of these discrepancies could lead to unnecessary pharmacologic exposure and/or delayed treatment, 15 we undertook a retrospective 10-year review of all positive mold and yeast cultures that were associated with concurrent surgical pathology specimens at Stanford University Medical Center (SUMC), Stanford, CA. The rate of misclassification was determined, and a root cause analysis was performed on discordant cases to develop an improved fungal identification process for general surgical pathologists. 364 Am J Clin Pathol 2009;131: DOI: /AJCP99OOOZSNISCZ

2 Anatomic Pathology / Original Article Materials and Methods A retrospective 10-year review of our laboratory information system database of all positive mold and yeast cultures reported by the Microbiology Division of Stanford Clinical Laboratory, SUMC, was performed for the years September 1997 to September 2007, inclusive. All specimens obtained under the categories of tissue, aspirate, abscess, or biopsy were included in the initial search. Specimens obtained from autopsy cases were excluded. Of the remaining cases, a search for concurrent specimens accessioned into the SUMC Department of Surgical Pathology and verified by date of accession was performed. Of the concomitant cases, those that had any known documented prior fungal diagnosis (by culture or by histologic examination) and those with a reported clinical history of a specific fungal infection were excluded to prevent any potential bias in the pathologist s diagnosis. However, cases that had prior presumptive fungal therapy were included in the analysis. For each case, patient sex and age, history of antifungal therapy, specimen type, diagnosis, the use of special histologic stains, any reported suggestion to correlate with or defer to microbiology, and the individual surgical pathologist were recorded. Any discrepancies between surgical pathology interpretation (frozen section or diagnosis lines, including the comment section) and microbiology culture results (established as the gold standard for this study) based on standard morphologic and biochemical criteria, 16 respectively, were noted. Slides from all discrepant cases were reviewed by 3 of us (A.R.S., T.A.L., and N.B.), and likely causes of misidentification were identified. For statistical analysis, we used the Fisher 2-tailed exact test, with significance set at a P value of less than.05. Results We evaluated 338 positive mold and/or yeast culture reports, of which 68 (20.1%) had a concurrent surgical pathology specimen Figure 1. Of the 68 cases, 17 (25%) were excluded because of a known history of fungal infection by previous culture isolation or histologic diagnosis. An additional 4 cases (6%) were excluded because of a reported history of a particular fungal infection. Of the 47 remaining cases (none of which included mixed fungal infections), 37 (79%) were correctly identified by histopathologic examination; these included 5 fungal categories: Aspergillus spp (n = 23), Coccidioides immitis (n = 8), Scedosporium apiospermum (n = 2), Trichophyton spp (n = 2), and Zygomycetes (n = 2). Of these 37 cases, genus and/or species identification was attempted in 35 (95%). The remaining 2 cases were reported as fungal elements identified with a differential diagnosis that included the organism identified by culture. The distribution of correctly vs incorrectly identified fungi 338 positive molds/yeasts with culture diagnosis (gold standard) 68 with concurrent histology 37 cases with concordant diagnosis 2 negative clinical consequences 270 without concurrent histology (excluded) 21 excluded due to known history 8 major errors in classification (6 with available follow-up) 10 cases with discordant diagnosis No negative clinical consequences 2 minor errors in classification (2 with available follow-up) Figure 1 Schematic overview of how fungal cases were selected for the present study based on a 10-year retrospective review of morphologic classification of all molds and yeasts in tissue or fine-needle aspiration biopsy specimens with concomitant culture diagnosis (taken as the gold standard ) at Stanford University Medical Center, Stanford, CA. by patient age, sex, use of special stains, previous fungal therapy, and other parameters is provided in Table 1. The 10 discrepant diagnoses (21%) included misidentification of Rhizopus (n = 3), Aspergillus spp (n = 2), S apiospermum (n = 2), Fusarium (n = 1), C immitis (n = 1), and Histoplasma capsulatum (n = 1) Table 2. Classification was attempted in all 10 misdiagnosed cases. Of the 10 discrepancies, 8 involved errors in division identification ( major errors ) and uniformly involved discordance between culture diagnosis and the diagnosis line or Table 1 Comparative Parameters in 47 Cases of Fungal Infections for Which Microbiological and Histologic Results Were Available * Correctly Incorrectly Identified Identified P Cases (n = 47) 37 (79) 10 (21) Mean (range) age (y) 44 ( ) 46 (4-73) Sex (No. of M/F) 23/14 9/1 Use of special stains 32/37 (86) 7/10 (70).34 Defer/correlate 21/37 (57) 6/10 (60) 1.00 Previous antifungal therapy 9/37 (24) 5/10 (50).14 No. of operators 13 9 * Data are given as number (percentage) unless otherwise indicated. Final pathology report included recommendation to defer to or correlate with culture. Number of different surgical pathologists signing out each case in each group. Am J Clin Pathol 2009;131: DOI: /AJCP99OOOZSNISCZ 365

3 Sangoi et al / Morphologic Identification of Fungi comment section of the final report. Of the 10 discrepant cases, 2 involved errors in genus identification ( minor errors ); both involved misidentification of Rhizopus on intraoperative frozen section diagnosis. Special stains (Gomori-methenamine silver or periodic acid Schiff with diastase) were used in 86% (32/37) of the correctly identified subgroup and in 70% (7/10) of the incorrectly identified subgroup (P =.34), but neither subgroup incorporated immunohistochemical studies or molecularbased techniques. 1,17-20 For accurate fungal identification by subgroup, 13 surgical pathologists were involved in the 37 cases in the correctly identified subgroup, and 9 surgical pathologists were involved in the 10 cases in the incorrectly identified subgroup, indicating that errors were evenly distributed among pathologists (Table 1). In addition, no apparent trend was identified when analyzing overall operator accuracy in fungal identification. Reports for both subgroups included a clause to defer and/or correlate with culture in nearly equal percentages of cases (Table 1). Follow-up data were available for 8 of 10 incorrectly identified cases (Figure 1). Two of these cases demonstrated the potential for adverse clinical consequences as a result of misdiagnoses by histopathologic examination, with the remaining 6 cases showing no change in patient treatment or management (Table 2). In case 4, amphotericin was added to the patient s regimen of voriconazole after surgery based on the rendered histologic diagnosis of suspicious for mucormycosis, but this was immediately discontinued after cultures revealed Aspergillus niger. While no apparent harm occurred to the patient during this brief period as a result of fungal misclassification, it is designated as a negative clinical consequence because of unnecessary exposure to a drug with known severe potential side effects. 21,22 A similar but somewhat less severe negative repercussion occurred in case 9. In this case, because the patient s lung lesion was a small solitary nodule, active surveillance without treatment was initially intended based on a clinical and radiologic diagnosis of Coccidioides. Because the histologic diagnosis of compatible with Cryptococcus was not congruent with the initial microbiological diagnosis of C immitis, the patient had to return to the clinic for further evaluation. Fortunately, the patient was not treated during this time. More important, if the specimen submitted to the microbiology laboratory had been inadequate or not representative of the lesion in question, the patient may have Table 2 Clinicopathologic Features of Misclassified Molds or Yeast Cases * Case No./ Immune Frozen Section Surgical Pathology Surgical Pathology Special Culture Sex/Age (y) Competent Specimen Diagnosis Diagnosis Comment Stains Diagnosis Defer Treatment 1/M/40 Yes Biopsy (nasal) Consistent Consistent with NA ND Rhizopus Yes No with Mucor Zygomycetes 2/M/47 No (leukemia/ Biopsy Aspergillus vs Consistent with NA ND Rhizopus Yes Yes bone marrow (maxillary Mucor Rhizopus transplant) sinus) 3/M/27 No (heart Biopsy (lung) NA Fungal organisms Aspergillus GMS Rhizopus Yes Yes transplant) identified 4 /M/73 No (relapsed Biopsy (ear Invasive Fungal elements Suggestive of GMS Aspergillus Yes Yes leukemia) canal) hyphal forms present mucormycosis niger 5/M/55 Yes Biopsy (lung) Infectious Consistent with NA GMS Aspergillus No No etiology Coccidioides # fumigatus 6/M/75 Yes Biopsy (nasal NA Aspergillus NA PAS-D Fusarium No Yes turbinate) 7/M/50 Yes Biopsy NA Fungal hyphae Compatible with ND Scedosporium No No (paranasal Aspergillus apiospermum sinus) 8/F/21 Yes FNA (retina) NA Consistent with NA GMS S apiospermum No Yes Aspergillus 9 /M/71 Yes FNA (lung) NA Compatible with NA GMS Coccidioides Yes No Cryptococcus immitis 10/M/4 Yes Biopsy (lung) Granulomatous Yeast forms Resemble GMS Histoplasma Yes No inflammation identified Candida spp capsulatum FNA, fine-needle aspiration; GMS, Gomori-methenamine silver; NA, not available; ND, not done; PAS-D, periodic acid Schiff with diastase. * Underlined items indicate a minor error; boldfaced items indicate a major error. Surgical pathology permanent section (bottom-line) diagnosis. Included recommendation to defer to or correlate with cultures. Previous antifungal therapy. Diagnosis made after microbiology culture results were reported. Demonstrated potential for adverse clinical consequences. # Case amended after culture results obtained. 366 Am J Clin Pathol 2009;131: DOI: /AJCP99OOOZSNISCZ

4 Anatomic Pathology / Original Article undergone a treatment plan based on an erroneous histologic diagnosis with a potential for inappropriate pharmacological coverage and unnecessary drug exposure. Although adverse clinical consequences were demonstrated in only 2 of the 8 cases for which there were followup data, it is important to emphasize that adverse outcomes may have occurred in all of the 8 cases given the wide range of antifungal choices currently available for clinical use Table 3. Analysis of the 10 misclassified cases identified a variety of procedural, terminological, and interpretive factors that may have led to errors in identification. For example, a false sense of the ability to categorize fungal organisms by genus on the basis of morphologic features alone in combination with the use of inappropriate terminology contributed to several minor errors involved in classification (cases 1 and 2, Table 2, labeling a Zygomycetes infection specifically as Mucor). A lack of knowledge of other potential molds within a particular category was an additional factor in many of the major errors in misclassification (cases 6-8, Table 2, failure to include hyalinohyphomyces (see the Discussion section) in the differential diagnosis of a fungus with septate hyphae). Finally, misinterpretation of morphologic mimics in diagnosing yeast and hyphal forms also contributed to diagnostic error. In addition to these factors, there was a fairly pervasive institutional misperception of the role of morphologic classification of fungal elements in directing patient care. Misclassification did not seem to be related to patient characteristics, specimen preparation (histologic section vs cytologic preparation; frozen section vs permanent section), work environment (daily schedule vs on call; stat vs routine), or individual surgical pathologist (Table 1). Discussion In this retrospective study, we found that microscopic identification of fungi in histologic or cytologic specimens has an overall accuracy of 79%. Of the 45 cases for which fungal classification was attempted, the accuracy was 78% (35/45). Because the morphologic features of several fungi may overlap in histologic and cytologic specimens, difficulties arise in recognizing characteristic features that are routinely seen in microbiological culture media. This problematic area in fungal identification has been discussed in detail elsewhere 2,3,12,14,23,24 and is only briefly mentioned here as a segue to a discussion of the misclassified fungal cases in this study. In our study, the group of misclassifications with the most significant potential for adverse consequences involved the misinterpretation of septate vs nonseptate hyphae (Table 2, cases 3 and 4). Although Zygomycetes are recognized on the basis of broad, nonseptate, or occasionally pauciseptate hyphae with nondichotomous branching, significant overlap can occur with septate hyphae with dichotomous branching seen in Aspergillus spp Image 1. Moreover, when fungal hyphae are scant, folded, kinked, or otherwise fragmented, accurate assessment for septation and type of branching may not be possible. In areas of extensive necrosis, hyphae may assume a more swollen and distorted appearance, and the Table 3 Therapeutic Implications for Select Invasive Fungal Infections Initially Visualized in Histologic/Cytologic Specimens as Fungal Organisms by Surgical Pathology and Subsequently Identified to the Species or Genus Level by the Microbiology Laboratory Organism * Antifungals With Efficacy Antifungals With Inadequate or No Efficacy Aspergillus fumigatus Voriconazole, amphotericin B, itraconazole, Fluconazole an echinocandin (for salvage or used by clinicians in combination with another agent), possibly posaconazole Aspergillus terreus Voriconazole Fluconazole, amphotericin B Fusarium solani Voriconazole plus amphotericin B Fluconazole, itraconazole, any single antifungal agent Mucorales genera Amphotericin B, posaconazole Fluconazole, voriconazole, itraconazole, an echinocandin as a sole agent Scedosporium apiospermum Voriconazole Fluconazole, amphotericin B, itraconazole, an echinocandin Coccidioides immitis Amphotericin B, fluconazole, itraconazole, Possibly echinocandin voriconazole, possibly posaconazole Candida spp (if pseudohyphae) Echinocandin, fluconazole, amphotericin B, Fluconazole, itraconazole, or voriconazole for voriconazole, itraconazole, possibly posaconazole certain Candida spp (eg, Candida krusei) Cryptococcus neoformans Amphotericin B + 5-flucytosine (5FC), fluconazole, An echinocandin or, less preferred, itraconazole Histoplasma capsulatum Amphotericin B, itraconazole, possibly posaconazole Fluconazole, an echinocandin * Final identification of fungal structures observed by microbiological methods. Organisms in the first 5 rows are commonly mistaken for each other on histologic examination, as are the organisms in the last 3 rows. Coccidioides can be confused with organisms in either of the other 2 groups. Echinocandins include caspofungin, micafungin, and anidulafungin. Am J Clin Pathol 2009;131: DOI: /AJCP99OOOZSNISCZ 367

5 Sangoi et al / Morphologic Identification of Fungi resultant atypical morphologic features may generate an erroneous diagnosis. In all such cases, there is a potential for inappropriate pharmacotherapy (Table 3) and/or surgical intervention, especially during intraoperative consultation Another well-described yet more frequently overlooked diagnostic problem occurred with the misidentification of the group of nonpigmented, branched, uniform, septate hyphae as Aspergillus spp Image 2. 2,14,20 Although aspergillosis is a common culprit, it is important to remember that not all such dichotomous hyphae represent Aspergillus spp. Hyalinohyphomyces (ie, Fusarium, Scedosporium, and Penicillium spp) and other hyaline septate molds (dermatophytes) may all share a similar morphologic appearance to Aspergillus spp, yet receive treatment strategies that may include different antifungal agents (Table 3). In the absence of fruiting heads, which are usually seen only in cavities or other well-oxygenated areas, aspergillosis should be diagnosed with caution in tissue and cytologic preparations and always with A B C D Image 1 A, H&E-stained sections demonstrate extensive infiltration by thin, septate hyphae showing frequent septation with dichotomous 45 branching ( 10; inset, 60). B (Case 3, Table 2), Although classified as Aspergillus by histopathologic examination based on Gomori-methenamine silver (GMS) stain (arrow), scant irregular hyphal forms and rare right-angle branching (arrowhead) correlate with the culture diagnosis of Rhizopus ( 40). C and D (Case 4, Table 2), In this case, the H&E stain (C, 40) outlines faint, irregular hyphal elements that on GMS stain (D, 60) appear pauciseptate and ribbon-like, prompting a histopathologic diagnosis suggestive of mucormycosis. Given the additional presence of rare narrow-angle branching (center), the morphologic appearance is most compatible with the culture diagnosis of Aspergillus niger. 368 Am J Clin Pathol 2009;131: DOI: /AJCP99OOOZSNISCZ

6 Anatomic Pathology / Original Article a recommendation to correlate with culture results. 2,23 Some authors have suggested that identification of adventitious sporulation in tissue sections (eg, phialides, phialoconidia) warrants a conclusive diagnosis of hyalohyphomycosis, 29,30 but we are reluctant to recommend this strategy, particularly for surgical pathologists without an extensive background in infectious disease histopathologic examination. As discussed by Watts 3 and Watts and Chandler, 14 exhaustive examination of pristine stained sections with a 100 oil-immersion objective lens is typically necessary to detect these unique specialized structures, whereas other experts propose the use of scanning or transmission electron microscopy for definitive identification. 2,12,23 These specialized sporulation structures are rarely encountered and/or recognized in surgical pathology specimens. As such, when they occur, we cautiously recommend a provisional diagnosis of hyalohyphomycosis with a more extended differential diagnosis and recommendation for culture confirmation. Misclassification also occurred in the identification of yeast forms (Table 2). Tangentially sectioned bulbous hyphae with incomplete spherule-like structures may simulate empty Coccidioides spherules Image 3. Alternatively, endospores in proximity and/or empty, overlapping spherules can mimic budding yeast, creating confusion with Blastomyces, Paracoccidioides, and Cryptococcus Image 4. 2,12 Free endospores from a ruptured spherule, particularly when they occur within histiocytes, can also imitate intracellular yeast. 2 Because of these common pitfalls, a diagnosis of Coccidioides in tissue or cytologic preparations should be made only in the presence of at least 1 unequivocal, intact spherule that contains endospores. Cryptococcus is recognized by pleomorphic, mucicarminepositive yeast cells surrounded by a distinctive empty space owing to the presence of a thick capsule; the latter organism is also more likely to occur in an immunocompromised host. 2,12 A B C Image 2 (Case 6, Table 2) A and B, Elongated, slender hyphal organisms in H&E-stained sections (A, 20) show septate, parallel sides with narrow (arrow) and wide-angle (arrowheads) branching (B, 60). C, Periodic acid Schiff stain with diastase reiterates these findings, but also outlines infrequent irregular, aseptate forms (at right) ( 60). Darkly stained ovoid forms, representing tangentially sectioned hyphae, are also present. These morphologic findings were interpreted as Aspergillus in tissue sections, although cultures revealed Fusarium spp. Am J Clin Pathol 2009;131: DOI: /AJCP99OOOZSNISCZ 369

7 Sangoi et al / Morphologic Identification of Fungi A B C D Image 3 (Case 5, Table 2) A, Vague, circular structures emulating Coccidioides are seen on H&E stain ( 60). B and C, Gomori-methenamine silver stained sections also show suspicious empty and incomplete spherule-like structures (B, 60; arrow) or endospore-like elements (C, 60), leading to an initial misdiagnosis of Coccidioides. D, Although intraoperative sampling error is a possibility, the additional presence of oblique, darkly rimmed structures suggests all of the forms depicted likely represent tangentially sectioned bulbous hyphal elements of cultured Aspergillus fumigatus (Gomori methenamine silver, 60). A definitive diagnosis of Coccidioides should not be made in tissue or cytologic preparations in the absence of unequivocal endosporecontaining spherules. Histologic differentiation of Histoplasma from other smaller yeast can also be challenging Image 5. Although typically present extracellularly, intracellular Candida spp can mimic Histoplasma spp. 2,12 Similarly, extracellular Histoplasma spp demonstrate significant overlap with Candida spp. The unique small size and lack of pseudohyphae production of Candida glabrata (vs other species of Candida) can further obscure accurate identification. 16 Clues to differentiation of Candida include the variably sized yeast cells, lack of a pseudocapsule, and better staining with H&E and Gram stain. 2,16 Furthermore, Candida spp typically generate a suppurative tissue reaction, 16 whereas Histoplasma spp tend to elicit a more granulomatous reaction. Several studies have demonstrated the usefulness of special stains in the identification and classification of fungal organisms in tissue sections. 2,18,23,31-36 Although the use of special stains was more prevalent in the correctly identified subgroup, the difference was not statistically significant (Table 1). Fluorescent antibody staining is another method reported to be useful in fungal recognition and categorization, 2,23 but this method requires procedural and interpretative experience. Although newer immunohistochemical and/or molecular methods have demonstrated improved identification and classification of yeast and hyphal organisms compared with conventional histologic (H&E, special stains) methods, 1,17-20 many of these technologies are not commercially available or are not in widespread use and may also require technical and interpretative proficiency. Of these newer techniques, in situ hybridization is the most promising 1,17,19,20 and may prove useful in providing a rapid, accurate preliminary diagnosis for slow-growing fungal cultures or for cases in which the morphologically visible organism may actually not grow. Familiarity of likely fungal organisms by body site location and immune status can also be advantageous in establishing a differential diagnosis Table 4. Finally, because in many cases fungal-specific serologic studies may be more precise 370 Am J Clin Pathol 2009;131: DOI: /AJCP99OOOZSNISCZ

8 Anatomic Pathology / Original Article A B C D Image 4 (Case 9, Table 2) Adjacent or overlapping ovoid yeast-like forms (A and B, 60) distributed as pairs (C, 60) and single organisms (D, 60) mimic budding yeast in this Gomori-methenamine silver stained fine-needle aspiration biopsy specimen. Although Coccidioides was identified in the cultured specimens, diagnostic endospore-containing spherules were not present in tissue isolates. Misclassification as other yeast forms (eg, Cryptococcus in this case) can readily occur without the inclusion of a comprehensive differential diagnosis. A B Image 5 (Case 10, Table 2) Extracellular location of small, ovoid yeast with scattered budding forms (arrows) led to misclassification as Candida spp in Gomori-methenamine silver stained tissue sections (A, 60; B, 100). Despite the unusual extracellular location, Histoplasma capsulatum was identified by culture. than morphologic diagnosis, by histologic or culture examination, correlation with these results (if available) is essential. 37 As demonstrated in Table 1, accurate fungal identification by histopathologic examination was not operator-dependent, with a stable distribution of diagnostic errors identified among the various surgical pathologists involved. Although fungal therapy is frequently perceived as affecting morphologic characteristics, we found no reproducible evidence of previous antifungal therapy altering fungal morphologic features. Moreover, there was no statistically significant difference attributed to previous antifungal therapy and accurate fungal identification between study subgroups (Table 1). Perhaps one of the more pervasive motivations underlying the errors in morphologic misclassification of fungal elements is the desire to offer a specific diagnosis in a pathology report. This notion is supported by this study showing that Am J Clin Pathol 2009;131: DOI: /AJCP99OOOZSNISCZ 371

9 Sangoi et al / Morphologic Identification of Fungi Table 4 Relationship Between Body Site and Significant Disease Associated With Selected Fungal Species Fungal Organism Aspergillus Mucorales Scedosporium Fusarium Candida Coccidioides Histoplasma Cryptococcus Body Site spp Genera spp spp spp immitis capsulatum neoformans Brain parenchyma Meninges Sinuses Pulmonary nodules Pneumonia Joints Bone Gastrointestinal tract Upper Lower Blood fungal classification was attempted in 96% of the cases (45/47). Although it remains true that histopathologic examination can offer rapid, preliminary identification of infectious fungal organisms and is often the only diagnostic platform when material is not submitted for culture, pressure to offer provisional diagnoses (which are often deemed as conclusive) by our surgical and clinical colleagues may extend a surgical pathologist s interpretative skills beyond a point of diagnostic veracity. Surgical literature supporting the usefulness of frozen section analysis in specific fungal identification 38,39 supplements the misperception by our clinical colleagues of the diagnostic specificity of histologic examination in fungal identification. From a pathologist s perspective, the diagnosis line often serves a different purpose from that listed in the comment section. For example, a malignant liver biopsy in a patient with a questionable history of colon cancer will frequently prompt adenocarcinoma, see comment as the diagnosis line and, within the comment section, how the morphologic findings are suggestive of metastatic adenocarcinoma from a colonic primary site will be discussed with the notation that correlation with the patient s clinical history is required. In this scheme, the diagnosis line is intended to be most definitive and the comment section more tentative and cautious, elaborating potential causes beyond the diagnosis line. Case 4 illustrates this idea: fungal elements present was listed in the diagnosis line, and the comment section discussed how the morphologic findings were suspicious for mucormycosis (also demonstrated in cases 3, 7, and 10). Clinicians may not always appreciate such a distinction between the diagnosis line and the comment section and, instead, simply scan the report without differentiating the two. This behavior has the potential for erroneous or inappropriate treatment and management (Table 3), as occurred in case 4. Further complicating the situation is the fact that many pathologists incorporate gradation modifiers in the diagnosis line and comment sections, such that the phrases consistent with, compatible with, and suggestive of are intended to connote different levels of diagnostic probability. A lack of standardized grading using these terms increases the potential misunderstanding that can occur in the clinical interpretation of a pathology report. To provide more standardized reporting for fungal infections in tissue and cytologic preparations and to deemphasize attempts at definitive species identification, we devised a series of templates Table 5, Table 6, and Table 7 for reporting fungal organisms in final surgical pathology reports (Tables 5 and 6) and intraoperative reports (Table 7), with several key points warranting elaboration. In Tables 5 and 6, Table 5 Permanent Section Templates for Hyphal Fungal Organisms Diagnosis Septate hyphal elements identified (see comment) Comment While the morphologic findings are suggestive of [X, choose one], the differential diagnosis also includes other septate hyphae such as [all other X], as well as nonseptate hyphae (Zygomycetes). Because the accuracy of correct fungal identification on tissue sections is estimated at 79% based on a recent comprehensive institutional review (present study), correlation with cultures is recommended. Diagnosis Nonseptate hyphal elements identified (see comment) Comment While the morphologic findings are suggestive of Zygomycetes, the differential diagnosis also includes other septate hyphae such as [X]. Because the accuracy of correct fungal identification on tissue sections is estimated at 79% based on a recent comprehensive institutional review (present study), correlation with cultures is recommended. Additional relevant comments Invasive? Extent (quantity)? Organism viability? Tissue necrosis? X indicates hyaline septate molds (dermatophytes, Aspergillus spp.), hyalinohyphomcyes (including Fusarium, Scedosporium, and Penicillium spp), or dematiaceous molds (if pigmented). 372 Am J Clin Pathol 2009;131: DOI: /AJCP99OOOZSNISCZ

10 Anatomic Pathology / Original Article additional organisms in the differential diagnosis are included to call attention to the nonspecificity of the morphologic features identified; these are selected to represent the most frequent and clinically significant fungal infections at our institution. For example, of the many hyalinohyphomyces, only the most common at our institution are included in our template, and modification based on individual institutional prevalence may be necessary. In some infrequent and specific cases, the diagnosis line may provide a more definitive interpretation (eg, if the unique fruiting heads of Aspergillus spp or at least 1 unequivocal, intact, endospore-filled spherule of Coccidioides are seen), but all reports should include a suggestion to correlate findings with culture results and an assessment for invasion, organism viability, tissue necrosis, and estimate of quantity of organisms present. These templates are useful because, by incorporating a 4-tiered general grouping (septate hyphae/nonseptate hyphae and yeast with pseudohyphae/yeast without pseudohyphae) with individually tailored comprehensive differential diagnoses (including a favored morphologic interpretation), they prompt the treating physician to consider other possibilities in the choice of an antifungal regimen rather than fixate on one particular organism. Moreover, the templates encourage surgical pathologists to withhold definitive species identification for the unique aforementioned histologic circumstances or when previous or concurrent culture results are available. Similarly, such a change in the thought process may lead to consultation with an infectious disease specialist who may not only offer more appropriate suggestions in patient management but also, by increased familiarity with atypical or anomalous organisms, be more capable of navigating the list of differential diagnostic entities provided by the template. Finally, the emphasis on correlation with culture results is significant not only to ensure proper identification and tailored pharmacotherapy but also because many organisms (particularly hyphal fungi) demonstrate such a rapid rate of growth that a preliminary (if not final) diagnosis can be offered within a reasonable time frame Table 8. As newer immunohistochemical and/or molecular methods become more readily accessible, they may prove useful in situations in which definitive classification of morphologically identified fungal organisms require longer incubation or simply do not grow, and findings could eventually be incorporated into the standardized report. It should be noted that because reporting formats may be institutionally specific, these templates are only a recommendation for improved standardization; contents could also be incorporated into the written comment section. It should also be mentioned that while for this study we established microbiological culture as the gold standard, we acknowledge that often a true gold standard necessitates clinical appraisal of an organism as a contaminant, colonizer, or true infectious agent. Table 6 Permanent Section Templates for Yeast-like Organisms Diagnosis Yeast with pseudohyphae identified (see comment) Comment While the morphologic findings are suggestive of Candida spp, the differential diagnosis also includes other molds (septate/ nonseptate hyphae). Because the accuracy of correct fungal identification on tissue sections is estimated at 79% based on a recent comprehensive institutional review (present study), correlation with cultures is recommended. Diagnosis Yeast without pseudohyphae identified (see comment) Comment While the morphologic findings are suggestive of [X, choose one], the differential diagnosis also includes [all other X]. Because the accuracy of correct fungal identification on tissue sections is estimated at 79% based on a recent comprehensive institutional review (present study), correlation with cultures is recommended. Additional relevant comments Invasive? Extent (quantity)? Organism viability? Tissue necrosis? X indicates Candida spp, Coccidioides, Cryptococcus, Histoplasma spp, Blastomyces, Paracoccidioides, and Sporothrix. Table 7 Intraoperative Frozen Section Template for Fungal Organisms Diagnosis Hyphal elements identified Diagnosis Yeast-like elements identified Additional relevant comments Invasive? Extent (quantity)? Organism viability? Tissue necrosis? Ensure intraoperative cultures are obtained Table 8 Expected Turnaround Times for Identification of Common Fungi by Culture Methods Time (d) Very rapid (1) Rapid (1-2) Moderately rapid (2-3) Moderately slow (5-12) Slow (7-13) * May take up to 3 d. May take up to 4 d. Organism Zygomycetes Aspergillus Fusarium Penicillium Cryptococcus Dematiaceous * Candida * Scedosporium Sporothrix Blastomyces Coccidioides Dermatophytes Histoplasma Paracoccidioides Am J Clin Pathol 2009;131: DOI: /AJCP99OOOZSNISCZ 373

11 Sangoi et al / Morphologic Identification of Fungi These templates were developed in collaboration with surgical pathologists, microbiologists, and infectious disease specialists at SUMC and formally presented at a surgical pathology quality improvement meeting and at infectious disease grand rounds. Following introduction of the standardized reporting format, there has been an increased number of infectious disease service consultations involving fungal infections from the general medicine service, many of which in the opinion of infectious disease specialists led to the provision of improved health care (data not shown). A more structured, follow-up analysis of reporting accuracy using the standardized reporting format is in progress. Histopathologic examination can offer prompt provisional identification of infectious fungal organisms and remains the only available reliable means to identify certain pathogens. However, significant morphologic overlap in fungal organisms, a desire to provide unequivocal fungal categorization owing to pressure from clinical colleagues, and idiosyncratic language in surgical pathology reporting contribute to errors. Based on our findings, we recently instituted a standardized report using a series of templates to be used in reporting fungal results in the surgical pathology report. We believe that when histopathologic and culture examination are used in a complementary manner, potential clinical oversights in patient treatment and management can be prevented. From the Departments of 1 Pathology and 2 Medicine (Division of Infectious Diseases and Geographic Medicine), Stanford University, Stanford, CA. Supported in part by a grant from Pfizer (Dr Banaei). Presented in part at the College of American Pathology Society Meeting; September 2008; San Diego, CA.. Address reprint requests to Drs Sangoi and Banaei: Dept of Pathology, 300 Pasteur Dr, Room L-235, Stanford, CA Acknowledgments: We thank A. James for excellent technical assistance in photomicrograph preparation. References 1. Hayden RT, Qian X, Roberts GD, et al. In situ hybridization for the identification of yeastlike organisms in tissue section. Diagn Mol Pathol. 2001;10: Schwarz J. The diagnosis of deep mycoses by morphologic methods. Hum Pathol. 1982;13: Watts JC. Surgical pathology and the diagnosis of infectious diseases [editorial]. Am J Clin Pathol. 1994;102: Cody DT Jr, Neel HB III, Ferreiro JA, et al. Allergic fungal sinusitis: the Mayo Clinic experience. Laryngoscope. 1994;104: Granville L, Chirala M, Cernoch P, et al. Fungal sinusitis: histologic spectrum and correlation with culture. Hum Pathol. 2004;35: Morpeth JF, Rupp NT, Dolen WK, et al. Fungal sinusitis: an update. Ann Allergy Asthma Immunol. 1996;76: Renshaw AA. The relative sensitivity of special stains and culture in open lung biopsies. Am J Clin Pathol. 1994;102: Steer A, Rothlauf MV, Morse WC, et al. Cryostat tissue section culture for Mycobacterium tuberculosis. Am Rev Respir Dis. 1967;96: Tarrand JJ, Lichterfeld M, Warraich I, et al. Diagnosis of invasive septate mold infections: a correlation of microbiological culture and histologic or cytologic examination. Am J Clin Pathol. 2003;119: Ulbright TM, Katzenstein AL. Solitary necrotizing granulomas of the lung: differentiating features and etiology. Am J Surg Pathol. 1980;4: Weydert JA, Van Natta TL, DeYoung BR. Comparison of fungal culture versus surgical pathology examination in the detection of Histoplasma in surgically excised pulmonary granulomas. Arch Pathol Lab Med. 2007;131: Kenney M. Misleading morphology of parasites and fungi in tissue sections. Pathol Annu. 1981;16(pt 1): Schofield CM, Murray CK, Horvath EE, et al. Correlation of culture with histopathology in fungal burn wound colonization and infection. Burns. 2007;33: Watts JC, Chandler FW. Morphologic identification of mycelial pathogens in tissue sections: a caveat [editorial]. Am J Clin Pathol. 1998;109: Lopez FA, Crowley RS, Wastila L, et al. Scedosporium apiospermum (Pseudallescheria boydii) infection in a heart transplant recipient: a case of mistaken identity. J Heart Lung Transplant. 1998;17: Larone DH. Medically Important Fungi: A Laboratory Guide to Identification. 3rd ed. Washington, DC: ASM Press; Abbott JJ, Hamacher KL, Ahmed I. In situ hybridization in cutaneous deep fungal infections: a valuable diagnostic adjunct to fungal morphology and tissue cultures. J Cutan Pathol. 2006;33: Eyzaguirre E, Haque AK. Application of immunohistochemistry to infections. Arch Pathol Lab Med. 2008;132: Hayden RT, Isotalo PA, Parrett T, et al. In situ hybridization for the differentiation of Aspergillus, Fusarium, and Pseudallescheria species in tissue section. Diagn Mol Pathol. 2003;12: Hayden RT, Qian X, Procop GW, et al. In situ hybridization for the identification of filamentous fungi in tissue section. Diagn Mol Pathol. 2002;11: Cornely OA, Vehreschild JJ, Ullmann AJ. Is there a role for polyenes in treating invasive mycoses? Curr Opin Infect Dis. 2006;19: Girois SB, Chapuis F, Decullier E, et al. Adverse effects of antifungal therapies in invasive fungal infections: review and meta-analysis. Eur J Clin Microbiol Infect Dis. 2006;25: Anthony PP. A guide to the histological identification of fungi in tissues. J Clin Pathol. 1973;26: Kradin RL, Mark EJ. The pathology of pulmonary disorders due to Aspergillus spp. Arch Pathol Lab Med. 2008;132: Mohindra S, Mohindra S, Gupta R, et al. Rhinocerebral mucormycosis: the disease spectrum in 27 patients. Mycoses. 2007;50: Rajshekhar V. Surgical management of intracranial fungal masses. Neurol India. 2007;55: Siddiqui AA, Shah AA, Bashir SH. Craniocerebral aspergillosis of sinonasal origin in immunocompetent patients: clinical spectrum and outcome in 25 cases. Neurosurgery. 2004;55: Am J Clin Pathol 2009;131: DOI: /AJCP99OOOZSNISCZ

12 Anatomic Pathology / Original Article 28. Young RF, Gade G, Grinnell V. Surgical treatment for fungal infections in the central nervous system. J Neurosurg. 1985;63: Ciesla MC, Kammeyer PL, Yeldandi V, et al. Identification of the asexual state of Rhizopus species on histologic tissue sections in a patient with rhinocerebral mucormycosis. Arch Pathol Lab Med. 2000;124: Liu K, Howell DN, Perfect JR, et al. Morphologic criteria for the preliminary identification of Fusarium, Paecilomyces, and Acremonium species by histopathology. Am J Clin Pathol. 1998;109: Fukuzawa M, Inaba H, Hayama M, et al. Improved detection of medically important fungi by immunoperoxidase staining with polyclonal antibodies. Virchows Arch. 1995;427: Jensen HE, Schonheyder HC, Hotchi M, et al. Diagnosis of systemic mycoses by specific immunohistochemical tests. APMIS. 1996;104: Kaufman L. Immunohistologic diagnosis of systemic mycoses: an update. Eur J Epidemiol. 1992;8: Kimura M, McGinnis MR. Fontana-Masson stained tissue from culture-proven mycoses. Arch Pathol Lab Med. 1998;122: Powers CN. Diagnosis of infectious diseases: a cytopathologist s perspective. Clin Microbiol Rev. 1998;11: Reed JA, Hemann BA, Alexander JL, et al. Immunomycology: rapid and specific immunocytochemical identification of fungi in formalin-fixed, paraffin-embedded material. J Histochem Cytochem. 1993;41: Elias Costa MR, Da Silva Lacaz C, Kawasaki M, et al. Conventional versus molecular diagnostic tests. Med Mycol. 2000;38(suppl 1): Ghadiali MT, Deckard NA, Farooq U, et al. Frozen-section biopsy analysis for acute invasive fungal rhinosinusitis. Otolaryngol Head Neck Surg. 2007;136: Hofman V, Castillo L, Betis F, et al. Usefulness of frozen section in rhinocerebral mucormycosis diagnosis and management. Pathology. 2003;35: Am J Clin Pathol 2009;131: DOI: /AJCP99OOOZSNISCZ 375