Localized Langerhans-cell histiocytosis of bone (eosinophilic granuloma of bone) is a rare, benign tumorlike condition that is characterized by a clonal proliferation of Langerhans-type histiocytes18-20,29,33. Although the etiology and pathogenesis remain unknown, the clinical course is usually benign but can be highly variable, with partial or complete healing of the lesion, recurrence after treatment, or progression or spontaneous remission without treatment5,7,9,12,22. The variable clinical course and the lack of a standardized approach to treatment have resulted in a paucity of guidelines for the management of patients who have this condition.
The treatment of localized Langerhans-cell histiocytosis has varied widely and depends in part on the site of the lesion and the clinical presentation. Patients who have a lesion involving the axial or appendicular skeleton usually have localized pain with or without swelling. Radiographically, the differential diagnosis includes osteomyelitis and Ewing sarcoma in children and metastatic carcinoma and other benign and malignant primary tumors of bone in adults. The diagnosis must be confirmed with operative (open) or percutaneous biopsy.
Local therapy is warranted when there is persistent pain, limited range of motion of the adjacent joint, involvement of an adjacent physis, or a risk of pathological fracture. The therapeutic approach remains controversial and reflects in part institutional preferences regarding strategies for diagnosis and treatment. Osseous healing has been observed after various therapeutic interventions, including curettage or low-dose radiation therapy, or both9,12,13,17,21-23,25,26,32; intralesional injection of corticosteroids4,6,8,10,14-16,25,27,28,31,34; and oral chemotherapy35.
Given the generally benign course of the disease, a simple, minimally invasive treatment with a low potential for complications is desirable. Intralesional injection of corticosteroids is a non-operative approach that has been demonstrated to produce rapid relief of pain with predictable healing and without the need for additional intervention4,6,8,10,14-16,25,27,28,31,34. A standardized protocol for the diagnosis and treatment of localized Langerhans-cell histiocytosis has been used at our institution since 1979. The utility of percutaneous biopsy for the diagnosis of this lesion and the early experience with intralesional injection of methylprednisolone have been reported previously6,15. The efficacy of this technique has been manifested clinically by immediate relief of pain and radiographically by predictable healing of the osseous lesion at the time of short-term follow-up.
We present the findings of our sixteen-year experience with use of percutaneous techniques for the diagnosis and treatment of localized Langerhans-cell histiocytosis involving the axial or appendicular skeleton. In addition, we report the clinical, radiographic, and functional outcomes for thirty-nine patients who were so managed.
*No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. No funds were received in support of this study.
†Section of Orthopaedic Oncology, Department of Surgical Oncology (A. W. Y. and J. A. M.), and Department of Pathology (C. V. F. and A. G. A.), University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Boulevard, Houston, Texas 77030. Please address requests for reprints to Dr. Yasko, Box 106.
‡Deceased.
Through a comprehensive review of the database maintained by our Department of Patient Studies, we identified thirty-nine patients who had been diagnosed with and managed for localized Langerhans-cell histiocytosis of bone involving the axial or appendicular skeleton at our institution between 1979 and 1994. An additional thirty-four patients who had Langerhans-cell histiocytosis were excluded because they had a craniofacial lesion, a lesion of the rib, a non-osseous lesion, or multisystem involvement. No patient was managed at our institution because of recurrent lesions during the period of the study. Moreover, no patient had had a biopsy or other treatment at another facility because of a lesion involving the axial or appendicular skeleton before being seen at our institution.
We reviewed the clinical presentation and radiographic findings from the medical records and radiographs of the thirty-nine patients. Procedural notes on the diagnostic technique and pathological reports on the histological findings were reviewed for each patient. The duration between the initiation of treatment and the cessation of pain was noted. Serial radiographs that were made during the period of active treatment and follow-up were examined. Data on the long-term outcome were obtained from a review of the clinical notes; the radiographic studies; and the functional analysis, including an assessment of the range of motion of the affected extremity, the patient's level of activity, and the presence or absence of symptoms.
Data on the Patients
Twenty-eight patients were male and eleven were female; the ages ranged from two to forty-nine years (median, ten years). All patients had had pain for two weeks to five months (median, one month) before they were seen by us. Ten of the patients who had a lesion affecting the lower extremity had an associated limp. Plain radiographs were made for all patients when they were first seen at our institution.
Skeletal Distribution
The osseous lesions involved ten different bones: the femur (ten lesions), the scapula (seven lesions), the clavicle (six lesions), the ilium (six lesions), the humerus (four lesions), the tibia (two lesions), the pubis (two lesions), the ischium (two lesions), the ulna (one lesion), and the thoracic spine (one lesion) (Fig. 1).
Each of the thirty-nine patients had a solitary symptomatic lesion that was revealed by a skeletal survey, which did or did not include a technetium bone scan. A second symptomatic lesion developed in two patients, two and four months after the first lesion was diagnosed. Radiographically, the greatest dimension of the lesion ranged from one to 8.5 centimeters, and the appearance ranged from a well defined lytic lesion to an ill defined permeative lytic lesion with a florid periosteal reaction. Four pathological fractures had occurred: one was in the clavicle, one was in the neck of the scapula, one was in the distal end of the humerus, and one was in the femoral neck. On the basis of the initial radiographs, the differential diagnosis included Langerhans-cell histiocytosis, osteomyelitis, Ewing sarcoma, metastatic carcinoma, bone cyst, enchondroma, chondrosarcoma, lymphoma, and plasmacytoma.
Diagnostic Biopsy
Since 1975, percutaneous biopsy has been the standard technique used at our institution to obtain tissue for the diagnosis of all osseous lesions2,3. The procedure routinely involves fine-needle aspiration alone or with core-needle biopsy. Each biopsy is performed with use of fluoroscopic guidance in the interventional radiology suite. At the time of the fine-needle aspiration, the radiologist and the cytologist decide whether additional tissue should be obtained with a core-needle biopsy. There is no specific protocol governing this decision, which is based on the adequacy of the specimen obtained by fine-needle aspiration and the need for special studies.
In the current series, the biopsy was performed with use of general anesthesia in twenty-five patients, twenty-four of whom were less than sixteen years old. Fourteen patients, thirteen of whom were more than fourteen years old, had local anesthesia; in addition, some were sedated but conscious. The fine-needle aspiration was performed with use of an 18 to 22-gauge needle. Cytological smears were made from the aspirated material and were stained with use of Diff-Quik (Harleco, Gibbstown, New Jersey) and rapid Papanicolaou techniques. Cell blocks usually were prepared, and additional material was processed for electron microscopy and immunohistochemical staining for S-100 protein and CD1 antigen when indicated or desired. Sterile samples were taken routinely for bacterial, fungal, and mycobacterial culture.
Core-needle biopsy was performed with use of a Tru-cut needle (Baxter Health Care, Deerfield, Illinois). The tissue that was obtained was fixed and was submitted for routine histological sectioning and staining with hematoxylin and eosin as well as for any special studies requested.
An operative (open) biopsy was done only when adequate but non-diagnostic tissue specimens had been obtained with a percutaneous biopsy.
Treatment
Before 1979, localized Langerhans-cell histiocytosis was treated with curettage, with or without radiation therapy, or with radiation therapy alone at our institution25. In 1979, treatment was standardized; since then, it has uniformly included intralesional injection of methylprednisolone sodium succinate. Radiation therapy was not used to treat any case of localized Langerhans-cell histiocytosis of the axial or appendicular skeleton that was diagnosed at our institution after 1979. Curettage was used only when an open biopsy was necessary to make the diagnosis.
In our early experience with the use of percutaneous techniques in patients who had localized Langerhans-cell histiocytosis, both the cytological and the histological findings were reviewed before the lesion was treated. Additional experience with cytological evaluation of aspirated material prompted the initiation, in 1983, of a one-stage procedure whereby the biopsy and injection are performed during one anesthesia session. Fine-needle aspiration is performed, the adequacy of the aspirated material is assessed immediately, and a diagnosis is made. When the diagnosis is Langerhans-cell histiocytosis, the surgeon proceeds with the intralesional injection of the corticosteroid. When a diagnosis cannot be established, a core-needle biopsy is done to obtain additional tissue. Thus, treatment is delayed only when the diagnosis cannot be determined with fine-needle aspiration.
During the period of the current study, core-needle biopsy and fine-needle aspiration were often done concurrently. When the lesion was treated on the basis of a diagnostic fine-needle aspiration, the additional tissue obtained with the core-needle biopsy was processed for routine histological evaluation and special studies, if desired. When the fine-needle aspiration was not diagnostic, the additional tissue was examined to establish a diagnosis.
After the intralesional injection, the affected extremity was protected from unrestricted use until the pain resolved and radiographic healing occurred.
Diagnosis
Fine-needle aspiration was performed for all forty-one lesions and was diagnostic for thirty-six (88 per cent). The cytological smears consistently demonstrated Langerhans histiocytes and a mixture of inflammatory cells, including eosinophils, lymphocytes, and plasma cells. The histiocytes had round-to-oval nuclei with prominent nuclear grooves or infolding and moderately abundant cytoplasm. These cells were identified easily with both the Diff-Quik and the Papanicolaou techniques (Fig. 2-A). Four aspirations yielded adequate tissue for analysis but were non-diagnostic, typically demonstrating inflammatory cells without characteristic Langerhans cells. The findings of one aspiration were suggestive of osteomyelitis.
The core-needle biopsy and the fine-needle aspiration were done concurrently in twenty-four lesions. Twenty (83 per cent) of these twenty-four core-needle biopsies were diagnostic for Langerhans-cell histiocytosis; the fine-needle aspiration was also diagnostic for Langerhans-cell histiocytosis in nineteen of the twenty lesions and was suggestive of osteomyelitis in one. Histological examination consistently demonstrated characteristic Langerhans histiocytes and a variable distribution of inflammatory cells (Fig. 2-B). For four lesions, both the core-needle biopsy and the fine-needle aspiration were non-diagnostic. The diagnosis for these four lesions was determined with open biopsy.
All cultures of the tissue samples were negative.
Treatment
Of the thirty-seven lesions for which percutaneous biopsy provided the correct diagnosis, thirty-five were treated with an intralesional injection of methylprednisolone at a dose of 100 to 150 milligrams (median, 125 milligrams). The injection was administered in twenty-nine lesions on the basis of the cytological findings of the fine-needle aspiration. One of these lesions was associated with a pathological fracture of the clavicle. The duration of the procedures that included both biopsy and injection ranged from twenty-two to 105 minutes (median, seventy minutes) for patients who had general anesthesia. Six patients had a two-stage procedure: the corticosteroid was injected a median of seven days after the core-needle biopsy, after the diagnosis had been confirmed on histological evaluation. The fine-needle aspiration had been diagnostic of localized Langerhans-cell histiocytosis in five of these six patients, but the prevailing strategy at the time was to perform the diagnostic procedure and to delay the therapeutic intervention until all tissue specimens had been analyzed. In the one patient for whom the fine-needle aspiration was suggestive of osteomyelitis, treatment was delayed until the cell block prepared from aspirated material and the histological specimen obtained with the core-needle biopsy had been evaluated. The diagnosis of Langerhans-cell histiocytosis was established after examination of both types of specimens. The cultures were negative.
Two lesions that were diagnosed with the percutaneous biopsy were not treated with an injection of methylprednisolone. One lesion, in the femoral neck, was associated with a non-displaced pathological fracture. Treatment included immobilization in a spica cast and external-beam radiation therapy after the diagnosis had been confirmed with both fine-needle aspiration and core-needle biopsy. This was the last time that radiation was used as the primary local therapy for localized Langerhans-cell histiocytosis of an extremity at our institution. Treatment of the other lesion was delayed because the patient chose to wait until all specimens had been analyzed. Despite the establishment of the correct diagnosis with both fine-needle aspiration and core-needle biopsy, curettage was subsequently performed at another institution.
The four lesions that were diagnosed with an open biopsy were treated concurrently with curettage, with or without autogenous bone-grafting. Two of these four lesions were associated with a pathological fracture—one involved the neck of the scapula and the other, the distal end of the humerus.
Follow-up
Thirty-four (97 per cent) of the thirty-five lesions healed after injection of methylprednisolone. Thirty-one of the thirty-four healed after a single injection, and the patients became pain-free within two weeks after the injection. Twenty-five (86 per cent) of the twenty-nine lesions that had been treated with the biopsy and injection concurrently, and all six lesions that had a two-stage procedure, healed after one injection.
Three lesions were treated with more than one injection. One was treated with a second injection, two months after the first one, because the patient had persistent pain. Another lesion was treated with three injections because the patient had persistent moderate pain. That patient had an associated lesion, which had been detected two months earlier and which healed after a single injection. The third lesion was treated with a second injection because, although the first injection had resolved the pain, radiolucency was evident at four months.
One of the thirty-five lesions was not treated successfully with the methylprednisolone; pain persisted for sixteen weeks after the injection. The patient refused to have a second injection and chose to be managed with curettage and bone-grafting at another institution.
All of the lesions that responded to a single injection had healed radiographically within two to four months (Figs. 3-A and 3-B), with subsequent bone-remodeling (Figs. 4-A, Figs. 4-B, Figs. 4-C and 4-D). The two lesions that were treated with a second injection were seen to have healed on radiographs made one and six months after that procedure. The lesion that was treated with three injections healed two months after the last injection. There were no complications associated with the biopsy or the injection.
At a median of ninety months (range, twenty-four to 199 months), the symptoms had not recurred in any patient. Clinically, all patients had a full range of motion of the affected joint, had full strength without residual muscular atrophy, and had resumed unrestricted activity. There was no radiographic evidence of recurrence of any lesion at the time of the latest follow-up (Figs. 5-A, 5-B, 5-C and 5-D).
The literature is replete with studies of the treatment of localized Langerhans-cell histiocytosis. There have been reports of many different therapeutic approaches that may result in healing of the osseous lesion4,6,8-10,12-17,21-23,25-28,31,32,34,35. However, because of the absence of controlled studies, the variability in the clinical course and the treatment, and the possibility of spontaneous resolution, it is difficult to draw firm conclusions regarding the effectiveness of any one therapeutic approach when healing of the osseous lesion is the end point of analysis.
Our strategy for treating localized Langerhans-cell histiocytosis of bone is based on the usefulness and reliability of percutaneous biopsy. Fine-needle aspiration, with or without core-needle biopsy, is effective for the procurement of tissue for diagnosis. The procedure can be performed after infiltration of a local anesthetic, with the patient sedated but conscious, or with use of general anesthesia. The morbidity associated with the procedure has been reported to be negligible, even when relatively inaccessible regions of the spine or pelvis are involved2,3.
The correct diagnosis of osseous lesions admittedly requires expertise in the techniques of obtaining specimens of bone and in the cytological and histological analysis of such specimens. Over the past twenty years at our institution, we have been able to refine the procedures for fine-needle aspiration and for immediate on-site cytological evaluation to assess the adequacy of the biopsy specimen. An analysis of our experience with percutaneous biopsy between 1976 and 1987 demonstrated an overall accuracy of approximately 87 and 83 per cent for the diagnosis of primary malignant and benign bone tumors3. More recently, an unpublished subset analysis of the accuracy of fine-needle aspiration alone between 1987 and 1996 at our institution revealed that the overall concordance with the diagnosis made on the basis of the final resected specimen was 92 per cent for primary malignant bone tumors and 83 per cent for primary benign tumors. Because of the rarity of all osseous lesions, the importance of expert pathologists and cytologists who are familiar with the morphology of bone tumors cannot by overemphasized.
Katz et al. described the cytological, histological, and ultrastructural morphology of localized Langerhans-cell histiocytosis in tissue obtained with percutaneous biopsy and emphasized that the procedure was a rapid and effective method for the procurement of tissue for diagnosis15. The application of this diagnostic technique has been demonstrated for appendicular, axial, and craniofacial lesions of Langerhans-cell histiocytosis at our institution6,25,31 and in series from other centers4,8,28.
In the present series, percutaneous biopsy established the diagnosis of Langerhans-cell histiocytosis for thirty-seven (90 per cent) of forty-one lesions. Fine-needle aspiration yielded a representative specimen, with features characteristic of Langerhans-cell histiocytosis, from thirty-six lesions (88 per cent). We did not routinely perform enzyme histochemistry, immunocytochemistry, or ultrastructural studies to establish the diagnosis of Langerhans-cell histiocytosis, despite suggestions that such analyses are essential37. However, when only scant material was obtained with the fine-needle aspiration or the cytological findings were equivocal, a core-needle biopsy with a Tru-cut needle was used to obtain additional material for histological evaluation and for one of these special studies or more, if needed. Since the differential diagnosis includes osteomyelitis, we recommend that specimens be obtained for culture at the time of biopsy.
Core-needle biopsy resulted in the correct diagnosis of twenty (83 per cent) of twenty-four lesions for which it was used. Core-needle biopsy also provided the correct diagnosis of one lesion that had been misinterpreted as osteomyelitis with use of fine-needle aspiration. No incorrect diagnosis was made when both fine-needle aspiration and core-needle biopsy were done. On review of our institutional bone-tumor registry, we found no case in which fine-needle aspiration or core-needle biopsy, or both, established a diagnosis of localized Langerhans-cell histiocytosis of the axial or appendicular skeleton that was later found to be incorrect. Moreover, no patient who had been diagnosed with another condition was later found to have Langerhans-cell histiocytosis. Percutaneous biopsy did not preclude an open biopsy when the latter was deemed necessary to establish the diagnosis. No inappropriate treatment was initiated.
Several investigators have reviewed their institution's experience with this rare condition and have all reported variability in the therapeutic approach and the rates of healing30,36. Because of the lack of published reports on a standardized approach to local therapy, it is difficult to draw firm conclusions regarding the preferred treatment of a solitary bone lesion of Langerhans-cell histiocytosis. Since 1979, we have used a standardized strategy at our institution to treat both the clinical symptoms and the radiographic abnormalities associated with localized Langerhans-cell histiocytosis.
To our knowledge, Scaglietti et al. were the first to use intralesional injection for Langerhans-cell histiocytosis, and on the basis of their excellent results with injection of forty to 200 milligrams of methylprednisolone sodium acetate in nine patients they recommended the injection of corticosteroids as the treatment of choice28. The benefit of this treatment compared with other methods is that it promotes early relief of pain and predictable osseous healing without operative morbidity, the late effects of radiation therapy, or the toxicity of chemotherapy.
Cohen et al., in 1980, reported their initial experience at our institution with this technique for lesions of an extremity and craniofacial lesions6. The first eight patients in that series had a two-stage procedure for the establishment of the diagnosis and treatment. All eight patients had relief of symptoms as well as radiographic evidence of osseous healing after an intralesional injection of sixty to 150 milligrams of methylprednisolone sodium succinate.
Capanna et al. reviewed their experience with eleven patients who had localized Langerhans-cell histiocytosis4. According to their treatment protocol, a patient who had a lesion with a radiographic appearance characteristic of Langerhans-cell histiocytosis had a percutaneous biopsy and an intralesional injection of forty to 160 milligrams of methylprednisolone sodium acetate before the diagnosis was confirmed. An open biopsy was done only when the lesion was atypical. A patient who had a new or recurrent lesion had an injection without confirmation of the diagnosis with a biopsy. All eleven lesions healed, but more than one injection was necessary in seven. No patient had complications associated with the treatment.
Egeler et al. reported the results of intralesional injection of forty to 150 milligrams of methylprednisolone sodium acetate or methylprednisolone sodium succinate in eight patients who had localized Langerhans-cell histiocytosis that had been diagnosed with either open or needle biopsy8. All eight lesions healed, and seven did so after a single injection.
Subsequently, similar clinical and radiographic results have been described in case reports14,16,27,34 and for small series5 of patients who had monostotic and polyostotic lesions involving craniofacial and long bones.
The administered dose of methylprednisolone in these series has ranged from forty to 200 milligrams. Both methylprednisolone sodium succinate and methylprednisolone sodium acetate have been effective. To our knowledge, no comparative analysis of these two preparations has been reported and no dose-response study has been performed to determine the optimum dose of corticosteroid for this uncommon lesion.
The standard dose that is administered at our institution currently is 125 milligrams. We have used methylprednisolone sodium succinate routinely. In our series, thirty-one (89 per cent) of thirty-five lesions healed, with complete resolution of symptoms and radiographic evidence of healing, after a single intralesional injection of methylprednisolone. Overall, thirty-four (97 per cent) of thirty-five lesions for which one intralesional injection or more was the only treatment healed. One patient chose not to have a second injection and had curettage at another institution. In all of the reported series, the symptoms were relieved early after the injection, usually within two weeks4,6,8,25,28,31.
Although the precise mechanism of action of intralesional injection of methylprednisolone for the treatment of localized Langerhans-cell histiocytosis has not been defined, to our knowledge, several in vitro studies have demonstrated the production of interleukins (such as IL-1 [interleukin-1]) and prostaglandins (such as prostaglandins E2 and D2) by suspensions of Langerhans cells1,11. The inhibition of IL-1-induced bone resorption and prostaglandin production by methylprednisolone may account for the dramatic response demonstrated both clinically and radiographically24. Definitive proof that injection of the corticosteroid is responsible for the observed response will be difficult to obtain. The rarity of the condition may preclude prospective, randomized trials, which are necessary to establish the true impact of corticosteroids.
To our knowledge, we are reporting the largest study to date evaluating the strategy of concurrent percutaneous techniques for the diagnosis and treatment of localized Langerhans-cell histiocytosis involving the axial and appendicular skeleton. The reliability of diagnosis with fine-needle aspiration allowed intralesional injection of methylprednisolone to be administered during the same anesthesia session in twenty-nine (83 per cent) of thirty-five lesions in our series. Moreover, a review of our experience since 1985 demonstrated that concurrent diagnosis and treatment was effective for twenty-three (92 per cent) of twenty-five of our most recently treated patients.
None of our patients had a complication of treatment, recurrence of symptoms at the affected site, or radiographic evidence of persistent or recurrent disease at a median follow-up of ninety months. All patients had full and unrestricted use of the affected extremity at the time of the latest evaluation.
The long-term results in our series support the treatment of localized Langerhans-cell histiocytosis with percutaneous biopsy for diagnosis and with intralesional administration of methylprednisolone. The technique is minimally invasive, is associated with negligible morbidity, and can be performed as an outpatient procedure. The current health-care environment challenges physicians to perform diagnostic and therapeutic procedures at the lowest cost with use of the least invasive (if appropriate) method to attain an expedient diagnosis and achieve the objectives of treatment. We use a local anesthetic or sedation that leaves the patient conscious for all percutaneous biopsies of bone so as to eliminate the costs associated with general anesthesia and an operative procedure.
It is extremely difficult to establish guidelines for the treatment of a rare condition such as Langerhans-cell histiocytosis; however, in the absence of a controlled study to clearly delineate a standard therapeutic approach and given the generally benign course of a solitary bone lesion, we maintain that the least invasive procedure appropriate for the level of expertise at a given institution should be used. We believe that our treatment protocol is extremely desirable because it predictably results in the rapid resolution of symptoms and in radiographic evidence of healing of the osseous lesions without recurrence.