Outcome and late complications of radiotherapy in patients with unicentric Castleman disease

Abstract

Castleman disease is a rare lymphoproliferative disorder. Surgery is considered standard therapy for the unicentric type. However, case reports have documented favorable responses to radiotherapy. The aim of this study was to analyse the clinical outcomes of five patients with unicentric Castleman disease treated with radiotherapy between 1991 and 2005. Mediastinal lymph nodes were the most common site of disease (four patients). Three patients were treated with radiotherapy alone, two patients with surgery and radiotherapy. Patients were treated with radiotherapy doses ranging from 40 Gy to 50 Gy. The median follow-up was 12 months (range, 3–175 months). During follow-up only one patient had progressive disease and died of Castleman disease. At the time of last follow-up two patients were in complete remission, one patient in partial remission, and one patient had stable disease. One patient showed serious acute and late toxicities. At the end of radiotherapy a paraneoplastic pemphigus vulgaris occurred, and eight to 11 months after radiotherapy a stenosis of the esophagus, of the left bronchus, and of the trachea due to scars. The study shows that unicentric Castleman disease is successfully treated with radiotherapy. However, for detection of possible complications as pemphigus vulgaris or stenosis of the esophagus or trachea an accurate follow-up is necessary.

Castleman disease, also known as angiofollicular lymph node hyperplasia, giant lymph node hyperplasia, and lymphoid hamartoma, was first described in 1954 and later defined by Castleman et al. in 1956 [1]. Castleman disease is an extremely rare and poorly understood lymphoproliferative disorder.

Three histopathologic subtypes have been defined [2]: hyaline vascular (HV), plasma cell (PC), and mixed variant (MV). The HV subtype is the most common histological variant of Castleman disease, accounting for 90% of cases. The HV subtype is characterized by small, hyalinized follicles surrounded by circumferentially arranged layers (“onion skin”) of small lymphocytes interconnected by a prominent vascular stroma. The PC subtype appears as germinal centers with dense plasma cell infiltration in the less vascular interfollicular stroma. The MV subtype is pathologically a mixture of the two other subtypes.

Two clinical entities also have been described: a unicentric presentation with disease confined to a single lymph node region and a multicentric presentation characterized by generalized lymphadenopathy, and systemic symptoms [3], [4]. Symptoms of Castleman disease are fever, night sweats, weight loss, splenomegaly, anemia, and hypoalbuminemia. Lymphadenopathy is found in two third of the cases in the thorax, mostly in the mediastinum. One third of the cases are located in other lymph node regions of the body or in extranodal regions (e.g. subcutaneous, intramuscular).

The etiology of Castleman disease is unclear. Chronic low grade inflammation, hamartomatous process, an immunodeficiency state, and autoimmunity have all been proposed as likely pathogenetic mechanisms [4], [5]. It also has been suggested that a chronic antigenic stimulus triggered by some microorganisms, such as Epstein-Barr virus (EBV), may cause chronic low-grade inflammation and reactive hyperplasia. Recently HHV-8 (Kaposi sarcoma-associated human herpes virus) has been implicated in the etiology of multicentric Castleman disease [6]. In the context of human immunodeficiency virus (HIV) infection, multicentric Castleman disease is associated with HHV-8 infection.

Surgery is considered standard therapy for the unicentric type, with several case reports and retrospective series reporting excellent rates of cure [3], [5], [7], [8]. However, other case reports [9–15] and retrospective studies [16], [17] have shown advantageous responses to radiotherapy in both unicentric and multicentric Castleman disease. Due to the rareness of Castleman disease the largest series of patients with unicentric Castleman disease treated with radiotherapy alone contained four patients [17]. Randomized, prospective studies for treatment of Castleman disease do not exist.

The aim of this study was to analyse the responses to therapy and clinical outcomes of patients with unicentric Castleman disease treated with radiotherapy (alone or after surgery) at a single institution.

Methods

Eligibility criteria for patients were as follows: histologically confirmed unicentric Castleman disease and treatment with radiotherapy. Data were obtained retrospectively using the irradiation protocols and patient documents of the Department of Radiation Oncology, University of Heidelberg and of the Department of Thoracic Surgery, Thorax Clinic at Heidelberg University Medical School.

Between 1991 and 2005 five patients with unicentric Castleman disease were treated with radiotherapy at the Department of Radiation Oncology, University of Heidelberg. In all patients, the diagnosis of Castleman disease was made by lymph node biopsy. The extent of disease was assessed by computed tomography (CT) of the chest, abdomen, and pelvis; ultrasound of the cervical, axillary, and inguinal lymphnodes; routine blood chemistry analyses, and bone marrow biopsy.

Clinical, pathologic, and radiographic data were used to evaluate treatment response. A complete remission was defined as radiographic resolution of disease. A partial remission was defined as a greater than 50% reduction in disease bulk, stable disease as a lack of radiographic evidence of an increase in tumor size. Actuarial rates of overall survival were calculated from the time of last irradiation to the date of death or date of last follow-up. Actuarial progression free survival (PFS) was calculated from the time of last irradiation to the time of documented disease progression or recurrence. Acute and late radiation toxicity was scored according to the RTOG/EORTC morbidity scoring criteria.

Results

The clinical and tumor characteristics of all patients are summarized in Table I. The median age of the patients was 28 years (range, 22–71 years). Two patients were male, three patients were female. Mediastinal lymph nodes were the most common site of disease (four patients). One patient had disease location in cervical lymph nodes. Median tumor size determined by CT was 10 cm (range, 3.9–13 cm). Four patients had the HV subtype of Castleman disease, one patient had the MV subtype. Three patients were treated with radiotherapy alone, two patients with surgery and radiotherapy. One patient had a macroscopic partial resection; the other patient had a microscopic partial resection.

Table I.  Clinical and tumor characteristics of the patients.

Characteristic	No.	%
Age [years]
Median	28
Range	22–71
Sex
Male	2	40
Female	3	60
Histopathology
Hyaline vascular	4	80
Plasma cell	0	0
Mixed variant	1	20
Location
Cervical	1	20
Mediastinal	4	80
Tumor size [cm]
Median	10
Range	3.9–13
Primary treatment
Radiotherapy	3	60
Surgery + radiotherapy	2	40

Tumor and treatment characteristics for each single patient are given in Table II. The median follow-up was 12 months (range, 3–175 months). Patients were treated with radiotherapy doses ranging from 40 Gy to 50 Gy (single dose 1.8–2 Gy). All patients received 3D-radiotherapy. During follow-up only one patient had progressive disease and died of Castleman disease. At the time of last follow-up two patients were in complete remission, one patient in partial remission, and one patient had stable disease.

Table II.  Tumor and treatment characteristics and survival for each single patient.

Patient	Age [years]	Sex	Histo-pathology	Location	Tumor size [cm]	Therapy	Radiotherapy dose [Gy]	Acute toxicity	Late toxicity	Response	Overall survival [mos]	End of follow-up
1	22	m	HV	Mediastinal	12	OP + RT	46	–	–	PR	3	alive
2	24	f	MV	Mediastinal	13	RT	45	Dermatitis, pemphigus vulgaris	Stenosis of esophagus, trachea, and bronchus	SD	12	alive
3	71	m	HV	Mediastinal	10	RT	45	–	–	PD	5	death
4	28	f	HV	Mediastinal	7.5	OP + RT	50	–	–	CR	23	alive
5	38	f	HV	Cervical	3.9	RT	40	–	–	CR	175	alive

M = male; f = female; HV = hyaline vascular; MV = mixed variant; OP = operation; RT = radiotherapy; CR = complete remission; PR = partial remission; SD = stable disease; PD = progressive disease

The two patients treated with surgery (partial resection) and radiotherapy were alive at the time of last follow-up. The patient with microscopic partial resection was in complete remission. The patient with macroscopic partial resection had a tumor size reduction by surgery from 12 cm to 3.7 cm. At the time of last follow-up (three months after radiotherapy) no change of tumor size was observed. After surgery and radiotherapy the patient was in partial remission. Both patients tolerated surgery and radiotherapy without complications. Postoperative radiotherapy of the mediastinum was administered using conformal techniques and 23 MV photons to a total dose of 46 Gy respectively 50 Gy (single dose 2 Gy). CT-based treatment planning was used to design the target volume in both patients.

The patient with cervical Castleman disease (tumor size 3.9 cm) was treated with 40 Gy (single dose 2 Gy) using 18 MV electrons (one irradiation field, 7×8 cm). Treatment planning based on ultrasound and CT. 4 months after radiotherapy a complete remission of the tumor was achieved. 175 months (approximately 14.5 years) after radiotherapy the patient was alive and had no evidence of disease.

The patient with the MV subtype of Castleman disease was treated with radiotherapy alone using a conformal technique (CT-based treatment planning) and 23 MV photons (total dose 45 Gy, single dose 1.8 Gy). This patient showed serious acute and late toxicities. At the end of radiotherapy a dermatitis (grade 3) of the irradiated skin and a paraneoplastic pemphigus vulgaris (locations: oropharynx, esophagus, forearms, hands, labia minora) occured. The pemphigus vulgaris was treated with steroids and immunosuppression. Six months after radiotherapy the pemphigus vulgaris was cured. Eight months after radiotherapy a stenosis of the esophagus due to scars required endoscopic dilatation. Ten months after radiotherapy a stenosis of the left bronchus was treated with endoscopic stent implantation. Eleven months after radiotherapy a stenosis of the trachea was treated with endoscopic dilatation. Despite these complications the tumor size was reduced (13 cm to 10.5 cm) until the time of last follow-up and showed stable disease.

The 71 year old patient had a tumor which infiltrated diffusely the whole mediastinum with accompanying pericardial and pleural effusions. At the beginning of radiotherapy he was in bad general condition and suffered of dyspnoea. Nine years before the diagnosis of Castleman disease he was irradiated at the vertebral column because of bone metastases of a prostate cancer. The mediastinum was irradiated using 23 MV photons to a total dose of 45 Gy (single dose 1.8 Gy). A non-conformal technique (two opponent irradiation fields) was used up to a dose of 39.6 Gy, the last fractions were administered using a conformal technique (CT-based treatment planning). At the end of radiotherapy the dyspnoea was reduced. Three months after radiotherapy the patient showed progressive disease. The infiltration of the mediastinum and pericardium continued. Five months after radiotherapy the patient died of progressive Castleman disease.

Discussion

In previous retrospective studies patients with unicentric Castleman disease presented at a significantly younger age than those with multicentric disease [16], [17]. Chronowski et al. analyzed 21 patients with Castleman disease: median age of patients with unicentric disease was 37 years, and of patients with multicentric disease 53 years [17]. The present study confirms this observation. Median age of the five patients with unicentric Castleman disease was 28 years.

The HV subtype is the most common subtype found in unicentric disease [16], [17]. In our study four of five patients with unicentric Castleman disease had the HV subtype. Traditionally, unicentric Castleman disease has been described as a localized, asymptomatic mass most commonly found in the mediastinum [2], [3]. However, other studies observed patients with systemic constitutional symptoms [16], [17]. In the present study mediastinal lymph nodes also were the most common site of disease (four of five patients). The patients had no systemic constitutional symptoms.

After surgical resection of unicentric Castleman disease cure rates approach 100% [3], [5], [7], [8]. Keller et al. retrospectively examined 61 patients with unicentric disease who were treated with surgery alone over a 20-year period [2]. Their study demonstrated that for patients with unicentric Castleman disease a complete resection afforded the best chance of cure. Relief of constitutional symptoms occured even after partial resection [16], [17]. However, continued follow-up is necessary, because recurrences have been documented more than nine years after incomplete surgical resection.

Although surgical resection is the most popular treatment strategy for unicentric Castleman disease, radiotherapy does have therapeutic activity and is a treatment option for patients not deemed good surgical candidates or in patients with incomplete surgical excision. The review of the literature documents that radiotherapy has the ability to achieve complete radiographic and clinical resolution of disease in unicentric Castleman disease.

Chronowski et al. described four patients with unicentric Castleman disease treated exclusively with radiotherapy [17]. Two fractionation schedules were used: 2 Gy per day to a total dose of 40 Gy and 1.8 Gy per day to a total dose of 39.6 Gy. The median follow-up was 30 months. Two patients were alive and clinically free of disease. Two patients died with no evidence of disease at last follow-up. Three patients had complete, radiographic resolution of their lymphadenopathy. Other studies have described eight patients with unicentric disease treated with radiotherapy [9–16]. Three of the eight patients had complete resolution of their lymphadenopathy, two patients were rendered asymptomatic but had persistence of lymphadenopathy, one patient had a decrease in the size of his tumor, one patient had a minimal response, and one patient had disease recurrence after initial resolution. Patients were treated with doses ranging from 27 to 45 Gy.

The present study emphasizes radiotherapy as a successful treatment option for patients with unicentric Castleman disease. Two patients treated with surgery (partial resection) and postoperative radiotherapy showed no progression of disease during follow-up: one patient was in complete remission, the other patient in partial remission. Among the three patients treated with radiotherapy alone two patients showed no progression of disease: one patient had stable disease, the other patient was in complete remission. The patient with complete remission had no evidence of disease for about 14.5 years. Only one patient treated with radiotherapy alone showed progressive disease and died of Castleman disease five months after radiotherapy. This patient differed in clinical and tumor characteristics from the other patients: the patient was older (71 years), in bad general condition at the beginning of radiotherapy, and had bone metastases of a prostate cancer which were irradiated nine years before the diagnosis of Castleman disease. The tumor infiltrated diffusely the whole mediastinum with accompanying pericardial and pleural effusions. The other patients had all circumscribed tumors in the affected lymph nodes. It seems that radiotherapy of Castleman disease has limited therapeutic activity in cases with diffusely infiltrating tumors.

In the present study one patient treated with radiotherapy alone showed serious acute and late toxicities. At the end of radiotherapy, a dermatitis (grade 3) of the irradiated skin and a paraneoplastic pemphigus vulgaris occurred. The pemphigus vulgaris was treated with steroids and immunosuppression. Six months after radiotherapy the pemphigus vulgaris was cured. Castleman disease has been associated with a very high incidence of autoimmune phenomena such as autoimmune cytopenias, peripheral neuropathy, systemic lupus erythematosus and paraneoplastic pemphigus. Some speculate that deregulated cytokine production, especially increased production of interleukin-6, may be an important trigger for the development of autoimmunity in Castleman disease [18]. The patient of the present study who developped a paraneoplastic pemphigus vulgaris had no expression of EBV, HHV-8, HIV, or interleukin-6 in blood samples or skin biopsy. Paraneoplastic pemphigus (PNP) is an autoimmune syndrome that was first described by Anhalt et al. in 1990 [19]. The syndrome is mostly associated with lymphoproliferative neoplasms, chronic lymphocytic leukaemia, thymoma, and Castleman disease. The criteria for PNP that are found in all patients include: intractable stomatitis; polymorphous cutaneus lesions that show histologically features of acantholysis, lichenoid or interface dermatitis; presence of autoantibodies against high molecular weight keratinocyte proteins; progressive disease that is refractory to treatment, with fatal outcome in most cases. Often patients develop respiratory injury evolving to bronchiolitis obliterans. Patients with Castleman disease and PNP have a high mortality rate that is due to the almost uniform finding of pulmonary involvement leading to respiratory failure and death. In a retrospective study including 28 patients with Castleman disease and PNP 22 patients (79%) died from respiratory failure [20]. To our knowledge, the present study describes for the first time the development of a paraneoplastic pemphigus vulgaris during radiotherapy of Castleman disease. It is unprovable whether the radiotherapy has induced the pemphigus vulgaris or not. But the induction of pemphigus vulgaris by radiotherapy may be possible. Radiation-induced pemphigus has been recorded only rarely in the literature. After radiotherapy for breast cancer, non-Hodgkin's lymphoma or squamous cell carcinoma of the lip patients had developed pemphigus vulgaris [21–23].

The same patient who developed the pemphigus vulgaris had eight months after radiotherapy a stenosis of the esophagus due to scars, ten months after radiotherapy a stenosis of the left bronchus, and 11 months after radiotherapy a stenosis of the trachea. The patient was treated with a total dose of 45 Gy (single dose 1.8 Gy). The tolerance dose (TD 5/5, 5% late complications in the first five years after radiotherapy) of the esophagus (55 Gy) and of the trachea (65 Gy) were not exceeded. The occurrence of stenosis of the esophagus and the trachea was unexpected. The review of the literature documents no case of stenosis of the esophagus or trachea after radiotherapy of Castleman disease. However, autoimmune disorders such as systemic lupus erythematosus and scleroderma have been reported to be associated with an increased incidence of long term radiation sequelae involving the skin and soft tissues [24], [25]. Late complications reported include severe fibrosis and/or necrosis of the skin, breast, bowel, and pelvis. The present study suggests that in patients with Castleman disease who experience autoimmune paraneoplastic syndromes radiotherapy should be used with caution.

Our data show that unicentric Castleman disease is successfully treated with radiotherapy. However, for detection of possible complications as pemphigus vulgaris or stenosis of the esophagus or trachea due to scars an accurate follow-up is necessary.