849. Histopathology of Cutaneous Invasive Fungal Infections in a Tertiary Cancer Center: Causes, Discordance with Culture, and Histopathologic Determinants of Outcome

Abstract

Abstract Background Cutaneous invasive fungal infections CIFIs (primary or secondary to hematogenous seeding) are frequent and often fatal in immunocompromised cancer patients (pts). There is a paucity of studies on the prognostic significance and concordance of histopathologic features with cultures. Methods We reviewed all pts with histologically diagnosed CIFIs at MD Anderson Cancer Center (06/2016-06/2020). Demographic, clinical, histopathologic [organism, distribution (dermis/subcutis, blood vessels/nerves/epidermis), density of fungi and host response (inflammation, fibrosis)], culture, and outcome (all-cause mortality) data were recorded. Results We identified 61 pts (median age: 60 years, range: 8-81); 37 (61%) were male. Most had hematologic malignancy (n=58, 95%), especially acute leukemia (n = 40, 66%). CIFI was primary in 53 pts (87%), with acute onset (≤ 1 week) in 66% of pts; 37 pts (61%) had multiple skin lesions. Fungal organisms were seen on H&E-stained sections in 47 cases (77%), whereas ancillary studies (GMS/PAS) were required in 14 cases (23%). Of the 59 concurrent microbiology cultures, only 43 (73%) were positive. In 16 cases, fungal order/genus was identified by both histopathology and culture; 13/16 (81%) were concordant (Fleiss’ kappa 0.67, Fig 1A). The causative fungal order/genus was determined in 55 pts (90%), most commonly Fusarium (n = 22, 36%) or Mucorales (n = 12, 20%, Fig. 1B). Angiotropism was most frequently associated with Fusarium (19/22, 88%), and neurotropism with Mucorales (8/12, 67%, Table 1). Eighty-four-day all-cause mortality rate was 62% (66% and 50% in CIFIs caused by molds and yeasts, respectively). Fungal angiotropism (p = 0.001, Fig 2A) and neurotropism (p < 0.001, Fig 2B) were associated with significantly increased mortality, while lymphocytic inflammation, seen only in 20%, was associated with reduced mortality (p = 0.024, Fig 2C). Figure 1 (A) Concordance of histopathologically determined and cultured fungal order/genus. Fleiss’ kappa 0.67 (“substantial agreement”), p < 0.001. (B) Distribution of causative fungal pathogens. Figure 2 Histopathological features significantly associated with 84-day all-cause mortality in CIFI patients. Error bands denote 95% confidence interval. Mantel-Cox log-rank test. Table 1 Association between type of organism and histopathological characteristics. Six patients with no identified organism were excluded. Fisher’s exact test. Abbreviation: PEH = pseudoepitheliomatous hyperplasia. Conclusion CIFIs have poor prognosis, especially when caused by molds and if fungal angio-/neurotropism is identified. Inflammation may be associated with better prognosis. Since cultures might be false negative (27%) or discordant (19%), more efforts are needed for culture-independent molecular detection of fungi. Incorporation of histopathologic features might inform prognostic risk stratification. Disclosures Dimitrios P. Kontoyiannis, MD, MS, ScD, PhD, AbbVie: Board Member|Astellas: Grant/Research Support|Cidara: Board Member|Gilead: Grant/Research Support|Merck: Advisor/Consultant|Scynexis/MSGERC: Board Member