Langerhans Cell Histiocytosis Involving the Thymus: A Case Report and Review of the Literature

Langerhans cell histiocytosis (LCH) encompasses a disparate group of diseases and distinct clinical syndromes, including solitary eosinophilic granuloma and the clinical variants of LCH, Letterer-Siwe disease, and Hand-Schuller-Christian disease.1 In most cases, LCH presents in older children and adults as unifocal disease that typically affects bone (eg, skull, femur, pelvic bones) and less commonly lymph node, skin, or lung.2 Multifocal disease commonly affects young children and can be either unisystem, as in Hand-Schuller-Christian disease, or multisystem, as in Letterer-Siwe disease. Although the thymus is typically unaffected by LCH, it has been described rarely as a localized lesion3,4 or more commonly as part of systemic LCH in children.5–7 Most adult patients with isolated thymic involvement by LCH have had coexistent myasthenia gravis3; however, there are exceptional reports of small thymic LCH lesions in patients without myasthenia gravis.4 

Langerhans cell histiocytosis has also been described in association with a number of coexistent malignancies, including both non-Hodgkin and Hodgkin lymphoma8 and other nonhematologic neoplasms, such as adenocarcinoma of the lung and breast.9,10 The occurrence of sarcomas with LCH is extremely rare,9,10 without a report, to our knowledge, of an associated soft tissue sarcoma. We describe the rare case of isolated thymic involvement by LCH in an adult patient with a history of low-grade leiomyosarcoma but without myasthenia gravis and review the literature regarding the association between LCH and other malignancies.

The patient is a 49-year-old woman with a medical history significant for resection of a low-grade leiomyosarcoma of the left gluteal and perineal region 3 years before her current presentation. She was treated with radical surgical excision and adjuvant radiotherapy. The sarcoma measured 14 cm in greatest dimension, and all surgical margins were negative for malignancy. The patient did well for 3 years, with no evidence of local or distant disease. One month before her current presentation, she was admitted to an outside hospital with pleuritic substernal chest pain and a presumed diagnosis of pericarditis. No other symptoms, including muscle weakness or other features of myasthenia gravis, were reported. There was no history of prior chest surgery, and the patient had never undergone diagnostic pneumomediastinoscopy. Physical examination revealed no pleural or pericardial rubs and was remarkable only for volume loss in the gluteal region related to her previous surgery. Her symptoms gradually improved; however, routine surveillance computed tomographic scans revealed a new 4.9 × 2.4-cm anterior mediastinal mass. The patient underwent resection of the mass for definitive diagnosis.

The resected specimen consisted of a 103-g, tan-pink mass that measured 9.0 cm from medial to lateral, 6.5 cm from superior to inferior, and 2.0 cm from anterior to posterior. The external surface was smooth and covered by a thickened membranous capsule (Figure 1, A). Cut section of the specimen revealed a tan-gray, firm, vaguely nodular mass interposed with bands of dense white-gray fibrous tissue (Figure 1, B).

Histologic sections showed fibrous tissue and interposed islands of thymic tissue with a sheetlike infiltrate of cells that possessed convoluted, elongated, and frequently grooved nuclei, with fine chromatin, thin nuclear membranes, and moderately abundant, slightly eosinophilic cytoplasm consistent with Langerhans cells (Figure 2, A and B). Admixed were numerous eosinophils, occurring singly and in large sheets, that focally demonstrated tissue destruction consistent with microabscess formation. Mitoses numbered up to 5 per 10 high-power fields within the population of atypical cells. Only small amounts of residual thymic tissue that demonstrated follicular hyperplasia with prominent germinal center formation were present, a feature commonly associated with myasthenia gravis. Moreover, strands of thymic epithelium were noted throughout the surrounding fibrotic stroma (Figure 2, C). Hassall corpuscles and dystrophic calcification were also seen. Granulation tissue and fibrosis bordered most of the lesion.

Immunohistochemical studies performed on paraffin sections revealed that the atypical Langerhans cells were strongly reactive for CD1a, S100 protein, and Fascin and negative for CD68, which alternatively highlighted only scattered histiocytes (Figure 3, A through D). The B-cell marker CD20 was positive within the reactive secondary follicles of the scant residual thymic tissue. CD3 and O13 (CD99) highlighted scattered thymocytes, with only occasional cells demonstrating reactivity for TdT and CD1a. A stain for keratin (AE1/AE3) highlighted a characteristic dispersed network of thymic epithelial cells (Figure 3, E). No evidence of involvement by the patient's low-grade leiomyosarcoma was seen.

Langerhans cell histiocytosis is regarded as a clonal, neoplastic proliferation of Langerhans cells, a normal constituent of the skin, lung, lymph node, and thymus.1 Thymic involvement by LCH occurs typically in children and as part of a systemic process. Much less commonly, LCH is seen as an isolated lesion of the thymus. In 1997, Gilcrease and colleagues3 reported 7 cases of localized LCH in the thymus, including their own index case and several reports published previously in the literature. All of these patients were either infants or young children (4 of 7 patients; age range, 0.2–12 years) or young adults (3 of 7 patients; age range, 21–36 years). Interestingly, the young adults had all clinically presented with myasthenia gravis, and LCH was diagnosed at the time of thymectomy. In addition, an isolated LCH lesion of the thymus was recently reported in association with a multilocular thymic cyst in a middle-aged man (age, 58 years) at the time of an unrelated cardiothoracic surgical procedure.4 In contrast to the large mass present in our patient (9.0 cm in greatest dimension), these previously reported cases described small LCH tumors (between 0.3 and 0.6 cm). Histologic examination of the mass in our case revealed extensive involvement by LCH with little residual thymic tissue remaining; features of thymoma were not seen. Of note, microscopic changes commonly associated with most cases of myasthenia gravis were seen in the residual thymic tissue (ie, follicular hyperplasia), although no clinical features of myasthenia gravis existed in this patient.

One interesting aspect of our index case is the prior history of a leiomyosarcoma. Langerhans cell histiocytosis has been found to be associated with a number of other neoplasms, including hematologic malignancies. Neumann and Frizzera8 documented cases of LCH found in association with both Hodgkin and non-Hodgkin lymphoma. In their series, they reported several cases of unifocal LCH coexistent with malignant lymphoma in the same nodal site. Moreover, other cases of LCH in their series were found to occur either before or after the presentation of lymphoma or within anatomic sites distinct from that of the malignant lymphoma. These sites included both the skin and the lung.

Although a number of nonhematologic malignancies have been reported in association with LCH, the finding of a coexistent sarcoma has only rarely been seen. In a recent study by Howarth et al,9 which followed up a large cohort of 314 patients with LCH, 14 patients were reported to have an additional malignancy diagnosed at the same time or after their diagnosis of LCH. The range of neoplasms included lung adenocarcinoma (4 cases), prostate adenocarcinoma (1 case), breast adenocarcinoma (1 case), small cell carcinoma of the lung (1 case), acute myeloid leukemia (1 case), chronic myeloid leukemia (1 case), non-Hodgkin lymphoma (1 case), parathyroid adenoma (1 case), pancreatic cystadenoma (1 case), and a pontine mass (1 case). Pathologic details of the pontine mass were unknown because a biopsy was not performed on the mass. In addition, one patient, a 46-year-old woman with documented LCH of the lung, was noted to have developed osteogenic sarcoma 8 years after her diagnosis of LCH.

Finally, Egeler et al10 documented from the literature 91 cases of LCH associated with malignant neoplasms. Most of these cases (67%) were associated with lymphoma or leukemia; however, 12% were associated with solid tumors other than lung carcinomas, including only 2 sarcomas, both arising in the bone. Interestingly, the LCH was most frequently found to have preceded the malignant neoplasm (78%). The authors believed that these tumors likely arose secondary to the radiation therapy used to treat the LCH. However, as in our case, 11% of the reported cases of LCH appeared unrelated to treatment effects and occurred after solid tumor malignancy.

Although it is clear that LCH can be present in the context of other malignancies, whether these occurrences represent as-yet-unrecognized biological relationships or are merely coincidental to heightened radiologic screening is still to be determined. In addition, it has been proposed that not all dendritic cell proliferations with the immunophenotypic features of Langerhans cells necessarily fall into the clinicopathologic spectrum of LCH and that these lesions might best be considered secondary dendritic cell processes.1 This may particularly be true in dendritic cell proliferations that are concurrently associated with other malignancies. However, making the distinction between a neoplastic and reactive process in this setting is problematic, since it is clear that solitary lesions exist that meet the accepted diagnostic criteria for LCH yet are clinically banal. Clonality studies would help to further characterize the nature of such an infiltrate; however, the status of suspected reactive Langerhans cell infiltrates with regard to clonality has not as yet been reported.1 Additionally, even most clonal Langerhans cell infiltrates have been shown to behave in a benign, “reactive” fashion. Perhaps the best measure of whether a dendritic cell process is reactive is the involution of the lesion once an associated primary disease process is controlled. With regard to the current case, the coexistence of thymic follicular hyperplasia could perhaps support a reactive pathogenesis. In addition, a possible relationship to the patient's recent episode of substernal pleuritic chest pain cannot be excluded, since Langerhans cells can proliferate in response to viral or other infections. Nonetheless, the large size of this lesion (9.0 cm) and the extent of the Langerhans cell infiltrate, together with the lack of other infection-associated clinical symptoms, may suggest that an exaggerated normal response is less likely. However, in the absence of clonality studies, this question cannot be definitively addressed.

Although extremely uncommon, this case illustrates that LCH can present as a localized lesion in the thymus in the adult population and adds to the limited number of reports documented in the literature. In contrast to our patient, nearly all of these previously reported cases have been associated with myasthenia gravis. Interestingly, despite the absence of clinical features of myasthenia gravis exhibited by our patient, follicular hyperplasia within the thymic tissue was seen, a finding present in nearly two thirds of patients with myasthenia gravis. Finally, the additional history of leiomyosarcoma in this case expands the spectrum of associated neoplasms encountered with LCH, although whether there is truly a biological relationship between these 2 entities is currently unknown.