Langerhans Cell Histiocytosis: A Population-based Study of Anatomical Distribution and Treatment Patterns

Graphical abstract

• LCH in bone marrow and lymph node are more likely to have multi-system involvement compared to LHC in other sites. • Craniofacial osseous LCH is more likely to be treated with surgery, vertebral LCH is less likely to be treated with surgery. • A racial disparity in surgery utilization is identified in pediatric patients with bone LCH.

Keywords:
Langerhans cell histiocytosis Treatment Surgery Chemotherapy Racial disparity Surveillance Epidemiology and end results

A B S T R A C T
Background: Langerhans cell histiocytosis (LCH) is a rare monoclonal histiocytic neoplasm. Little is known about clinical factors associated with LCH single-vs multi-system involvement at the time of diagnosis. Methods: Data on 1549 LCH patients diagnosed between years 2010 and 2018 were extracted from the Surveillance, Epidemiology and End Results Program. Patterns of single-vs multisystem involvement were examined using multivariable logistic regression analysis. Odd ratio (OR) and 95% confidence interval (CI) were reported. Results: 968 children and adolescents (0-19 years; median: 4 years) and 581 adults (≥20 years; median: 49 years) were included in the analysis. Multi-system LCH was reported for 30.9 % patients. Bone marrow (BM) (OR = 3.776; 95 %CI = 1.939-7.351; P < 0.001) and lymph node (LN) (OR = 3.274; 95 %CI = 1.443-7.427; P = 0.005) involvement were most commonly associated with multi-system LCH at the time of diagnosis; similar pattern was also observed in adult patients (OR = 17.780; 95 %CI = 6.469-48.867; P < 0.001 for BM LCH; and OR = 5.156;

Introduction
LCH is a rare neoplastic disease characterized by clonal expansion of CD1a + /CD207 + histiocytes, also known as Langerhans cells, in the background of cellular inflammatory infiltrate [1,2]. The yearly incidence of LCH varies by age, and is approximately 4.46 per million children and 1.06 per million adults [3]. While the etiology of LCH remains to be elucidated, there are some recognized predisposition risk factors such as family history of LCH, cancer, or thyroid disease; child born to a parent with history of occupational exposure to certain organic solvents; metal, granite or wood dust; tobacco smoking; and infections in the neonatal period [4]. Under normal physiological circumstances, Langerhans cells orchestrate immune cells cross-talk. However, upon acquired somatic BRAF V600E mutation, or aberrant activation of the MAPK/ERK signaling pathway, Langerhans cells abnormally proliferate and activate surrounding reactive lymphocytes, eosinophils and neutrophils, resulting in localized or diffuse tissue distruction [5][6][7].
Solitary LCH lesions, single or multiple, may form in any organ; several body systems may be affected at the same time, too. Bone, skin, bone marrow (BM) and lymph node (LN) are the most common organ systems involved in pediatric patients, while lung, bone, skin and BM are the commonest involved organ systems in adult patients [2]. The clinical manifestations are highly heterogeneous depending on anatomical sites. About 70 % of LCH cases are confined to the single-organ system at the time of diagnosis. However, some patients develop multi-system LCH, often with protracted and debilitating clinical course [8]. Medical management of patients with LCH is quite heterogeneous, and includes close observation, surgery, radiotherapy and pharmacological interventions: systemic and topical chemotherapy, glucocorticoids, nonsteroidal anti-inflammatory drugs such as indomethacin, and signaling pathway inhibitors [9]. There recently have been a number of comprehensive reports summarizing results of multicenter clinical trials addressing discrete research questions [10][11][12][13]. While the results of these ground-breaking clinical trials have changed the treatment paradigms, the patient cohort composition in these trials may not necessarily directly translatable into the real world practice patterns. The Surveillance, Epidemiology and End Results (SEER) is the United States of America-based nationwide program, which prospectively collects data on diagnosis and treatment of various types of cancer [14]. In this study, using the SEER-based data, we aimed to explore the associations of anatomical site distribution with multisystem involvement risk and treatment pattern in pediatric and adult patients with LCH.

Database and patient
In this population-based cohort study, we retrieved data from the  , unknown primary site (n = 25), unknown system involvement information (n = 51), unknown surgery status of primary site (n = 21) and unknown cause of death (n = 3) were excluded from subsequent analysis. Ultimately, data on 1549 patients were analyzed; 968 were pediatric patients (0-19 years) and 581 were adult patients (≥20 years). The patient inclusion flowchart is displayed in Fig. 1.

Clinical variables
Data on the patients' age at diagnosis; sex; race (white vs non-white as defined in the SEER database); primary site; diagnostic confirmation approach; surgery status for primary site; type of therapy; survival time, and cause of death were extracted. Primary sites of the bone were further divided into craniofacial (C41.0-1, including mandible), limb (C40.0-3), vertebral (C41.2), pelvic (C41.4) and chest wall (C41.3). Body system involvement at the diagnosis was extracted using the field "combined summary stage", where "localized" was considered as the single-system involvement, while "distant" as the multi-system involvement. Disease specific survival (DSS) and overall survival (OS) were also determined.

Statistical analysis
Age was expressed by median with quartiles and dichotomized based on the median. Categorical variables, such as sex, race and etc., were expressed by number with percentage, and examined by the Chi-square test. Multivariable logistic regression model was used to determine the associations of the anatomical body site with the outcomes. Odd ratio (OR) and 95 % confidence interval (95 %CI) were calculated. All statistical analyses were conducted using SPSS 26.0 software. A two-sided P < 0.05 was considered as statistically significant.

Association of anatomical sites with treatments in pediatric patients
Pediatric patients with LCH of the bone were less likely to undergo chemotherapy compared to their counterparts with LCH of other sites (OR = 0.532; 95 %CI = 0.334-0.848; P = 0.008). The pediatric patients of 5-19 years of age were more likely to undergo surgery (OR = 1.733; 95 %CI = 1.291-2.327; P < 0.001) and less likely to undergo chemotherapy (OR = 0.439; 95 % CI = 0.329-0.585; P < 0.001) compared to those of 0-4 years of age (Fig. 3).

Skeletal LCH in pediatric patients
The demographic and clinical characteristics of pediatric patients with LCH affecting different skeletal anatomical sites are detailed in Table 2. The frequency of successful diagnostic biopsy with positive histological findings were high in craniofacial (88.2 %), limb (90.8 %), pelvic (88.9 %) bone LCH, and somewhat lower for lesions located in vertebral column (75.3 %). LCH of craniofacial bones was commonly treated with surgery (63.6 %), while LCH involving vertebral column bones were treated surgically rather rarely (9.9 %). Radiotherapy was more commonly used for treatment of vertebral (4.9 %) and pelvic (6.7 %) LCH, but less in craniofacial (0.6 %) and limb (1.7 %) bone LCH. Similarly, chemotherapy was a treatment of choice for LCH of bones of chest wall and limbs only in 12.0 % and 18.3 % cases, respectfully.
A racial disparity in surgery utilization was identified in pediatric patients with bone LCH: non-white patients were less likely to be treated with surgery compared to white patients (OR = 0.470; 95 %CI = 0.272-0.812; P = 0.007) (Fig. 4).

Discussion
Skeleton is the most commonly affected organ system in pediatric patients with LCH, which accounted for 65.9 % of all LCH in our data analysis. Craniofacial osseous LCH is the most common location, accounting for 53.6 % of all skeletal LCH. Our study demonstrates that craniofacial osseous LCH is treated predominantly with surgery and chemotherapy compared to skeletal LCH in other anatomical compartments. The lesion curettage is the commonest surgical method, without need of craniotomy unless the dura mater is involved. In our analysis, radiotherapy (0.6 %) was uncommonly used for craniofacial osseous LCH. Of note, recently, Hiroshima Y., et al reported that, for a skull LCH even with dura infiltration, a non-high intensity radiotherapy can rapidly shrink the lesion resulting in a complete response without local recurrence [15]. Therefore, the feasibility and safety of radiotherapy in certain clinical scenarios may need to be explored. Osseous LCH involving spine is insidious. We previously demonstrated that, once the pathological process involves the vertebral arch or canal, patients may experience neck or back pain, limb numbness, muscle weakness and limping [16,17]. In the last decade, 18-fluorodeoxyglucose positron emission tomography-computed tomography ( 18 FDG PET-CT) plays an increasingly important role in diagnosis of LCH [18][19][20]. 18 FDG PET-CT can help detect asymptomatic LCH lesions and differentiate metabolically active lesions from inactive disease [21]. In the multivariable analysis we demonstrate that osseous LCH of the spine was less likely to be treated with surgery compared to osseous LCH of other bones. Immobilization and observation are the preferred managements while surgery is only applicable for patients with evolving neurological complications [17].
In addition, our study identified a racial disparity in surgery utilization in pediatric patients with bone LCH: the non-white patients were less likely to be treated with surgery than the whites. In the United States, the racial disparity, whether or not mediated by patient's socioeconomic status, is an important factor influencing the care and outcome of various childhood neoplasms [22][23][24]. Our work suggests that equitable care should be stressed in pediatric patients with osseous LCH as well. Hispanic mothers are more likely to have children with LCH compared to non-Hispanic whites (OR = 1.51; 95 %CI = 1.02-2.25); this risk increases further when both parents are Hispanic (OR = 1.80, 95 % CI = 1.13-2.87) [25]. Higher age-standardized incidence rate of LCH has also been observed for Hispanics compared to non-Hispanics (RR = 1.63; 95 % CI = 1.15-2.29) [26].
Skin is the second most commonly affected organ (13.4 % in our study) in pediatric patients with LCH. Our study demonstrated that skin LCH had the highest diagnostic rate via positive histology (92.4 %). Skin LCH has a tendency for spontaneous resolution and is less likely to be associated with multisystem LCH [27]. We demonstrated that skin LCH was less likely to be treated with surgery and chemotherapy compared to LCH of other sites. Indomethacin may be a useful option for recurrent skin LCH refractory to other frontline therapies [28].
In summary, LCH is a highly heterogeneous disease with a wide variety of clinical manifestations. In this study we demonstrate that the risks of multi-system involvement and management strategies depend upon anatomic systems involved. This work also highlights the needs to further study racial disparity in treatment option application in pediatric patients with skeletal LCH.

Ethics approval and consent to participate
Not applicable.

Availability of data and materials
All data were available from the US Surveillance, Epidemiology, and End Results Program (https://seer.cancer.gov/).

Funding
This work was supported by the National Natural Science Foundation of China (#82003351; recipient: Qinyuan Zhu) and Shanghai Municipal Health Commission Research Project (# 20194Y0242; recipient: Yangbai Sun).

Author contribution
All authors equally contributed to the analysis and writing of the manuscript.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.