BACKGROUND: Erdheim-Chester disease (ECD) is a rare histiocytosis that may overlap with Langerhans Cell Histiocytosis (LCH). This "mixed" entity is poorly characterized. We here investigated the clinical phenotype, outcome, and prognostic factors of a large cohort of patients with mixed ECD-LCH.

METHODS: This retrospective study was performed at two referral centers in France and Italy (Pitié-Salpêtrière Hospital, Paris; Meyer Children's Hospital, Florence). We included children and adults with ECD diagnosed in 2000-2022 who had biopsy-proven LCH, available data on clinical presentation, treatment and outcome, and a minimum follow-up of one year. Outcomes included differences in clinical presentation and survival between mixed ECD-LCH and isolated ECD; we also investigated response to treatments and predictors of survival in the mixed cohort. Survival was analyzed using the Kaplan-Maier method and differences in survival with the long-rank test. Cox regression models were used to evaluate the potential impact of age and gender on survival and to identify predictors of non-response and survival.

FINDINGS: Out of a cohort of 502 ECD patients, 69 (14%) had mixed ECD-LCH. Compared to isolated ECD, mixed ECD-LCH occurred more frequently in females (51 vs. 26%, p < 0.001) and in patients with multisystem disease (≥4 sites). Mixed ECD-LCH more frequently involved long bones (91 vs. 79%, p = 0.014), central nervous system (51 vs. 34%, p = 0.007), facial/orbit (52 vs. 38%, p = 0.031), lungs (43 vs. 28%, p = 0.009), hypothalamic/pituitary axis (51 vs. 26%, p < 0.001), skin (61 vs. 29%, p < 0.001), and lymph nodes (15 vs. 7%, p = 0.028); the BRAFV600E mutation was also more frequent in mixed ECD-LCH (81 vs. 59%, p < 0.001). Targeted treatments (BRAF and/or MEK inhibitors) induced response more frequently than conventional therapies (interferon-α, chemotherapy), either as first-line (77 vs. 29%, p < 0.001) or as any line (75 vs. 24%, p < 0.001). After a median follow-up of 71 months, 24 patients (35%) died. Survival probability was comparable between ECD alone and mixed ECD-LCH (log-rank p = 0.948). At multivariable analysis, age at diagnosis (HR 1.052, 95% CI 1.008-1.096), associated hematologic conditions (HR 3.030, 95% CI 1.040-8.827), and treatment failure (HR 9.736, 95% CI 2.919-32.481) were associated with an increased risk of death, while lytic bone lesions with a lower risk (HR 0.116, 95% CI 0.031-0.432).

INTERPRETATION: Mixed ECD-LCH is a multisystem disease driven by the BRAFV600E mutation and targeted treatments are effective. Age at diagnosis, bone lesion patterns, associated hematologic conditions, and treatment failure are the main predictors of death in mixed ECD-LCH.

FUNDING: None.