Sorafenib in patients with progressive epithelioid hemangioendothelioma

A phase 2 study by the French Sarcoma Group (GSF/GETO)


  • Presented in part as Abstract 10020 at the 2012 Annual Meeting of the American Society for Clinical Oncology; June 1-5, 2012; Chicago, Illinois.

  • We thank all the staff of the Oscar Lambret Center's Clinical Research Unit (Yvette Vendel, Laurence Delannoy, and Emilie Decoupigny) and Severine Marchant for editing the article. We also thank the Institut National du Cancer and Bayer HealthCare for their grants and support.



There is no standard treatment for progressive epithelioid hemangioendothelioma (EHE). To investigate the significant vascularization of EHE, the activity/toxicity of sorafenib in patients with progressive EHE was explored.


In this multicenter, 1-stage, phase 2 trial of sorafenib (800 mg daily), the primary endpoint, which was chosen by default, was the 9-month progression-free rate. All patients had documented progressive disease at the time of study entry.


Fifteen patients were enrolled between June 2009 and February 2011. The median age was 57 years (range, 31-76 years), and the ratio of men to women was 9:6. The performance status was zero in 10 patients and 1 in 5 patients. Twelve patients had metastases, mainly in the lung (12 patients), liver (5 patients), and bone (3 patients). Five patients had received prior chemotherapy (doxorubicin in 5 patients and taxane in 3 patients). The median sorafenib treatment duration was 124 days (range, from 27 to >271 days). Seven patients required dose reductions or transient treatment discontinuation. The 9-month progression-free rate was 30.7% (4 of 13 patients). The 2-month, 4-month, and 6-month progression-free rate was 84.6% (11 of 13 patients), 46.4% (6 of 13 patients), and 38.4% (5 of 13 patients), respectively. Two partial responses were observed that lasted 2 months and 9 months.


Further clinical trials exploring sorafenib as treatment of progressive EHE are needed. Cancer 2013;119:2639–2644. © 2013 American Cancer Society.


Soft tissue sarcomas are a heterogeneous group of rare tumors.[1, 2] Locally advanced or metastatic sarcomas are generally treated with doxorubicin-based chemotherapy as first-line treatment.[3] In contrast, there is no standard second-line therapeutic option after failure on or intolerance to doxorubicin (because the available literature is a list of successive phase 2 trials, which identify some drugs with promising activity but with no confirmation by formal, comparative phase 3 trials[4]).

However, recent data on histology-driven or biologic target-oriented strategies seem promising.[3] These studies indicate the necessity of conducting phase 2 trials in some exceptional subtypes.

Among the rarest histologic subtypes of sarcoma, “vascular sarcomas” or “highly vascularized sarcomas” (comprising several separate entities, such as angiosarcoma, epithelioid hemangioendothelioma [EHE], and malignant solitary fibrous tumors [SFT]) are heterogeneous entities.[5-8] No evidence for their optimal treatment is available,[6, 7] because clinical trials focusing on this disease are rare,[6, 7] and some trials include the different entities in the same category.[8] Among these tumors, angiosarcoma is characterized by its aggressiveness (with a median overall survival of approximately 8 months) and the short duration of the response to chemotherapy (with a median time to progression of approximately 2-4 months).[5] In contrast, EHEs are indolent diseases in most patients.[9, 10]

Metastatic/advanced EHE can be stable for years before requiring systemic treatment and can spontaneously regress in some exceptional cases. Until now, there has been no standard treatment for progressive and nonresectable EHE.[9, 10]

The inhibition of angiogenesis is a promising approach for treating sarcoma, especially in the case of highly vascularized sarcoma. Indeed, EHEs express vascular endothelial growth factor (VEGF).[11, 12] For these reasons, sorafenib, a small-molecule B-Raf and VEGF receptor inhibitor, is a logical candidate for the treatment of EHE.


Study Population and Eligibility Criteria

The patients who were considered for this study were required to be aged ≥18 years. All had histologically proven EHE, as reviewed by an independent pathologist (Y.M.R.), with metastasis or EHE of a locally advanced stage not amenable to radiotherapy or curative-intent surgery after multidisciplinary decision making. Prior systemic treatment for EHE was allowed. Disease that was measurable or assessable by computed tomography scanning was required according to Response Evaluation Criteria in Solid Tumor (RECIST) version 1.1 guidelines.[13] Additional key eligibility criteria were as follows: at least 1 lesion measurable according to RECIST 1.1; evidence of progression at the time of study entry; no brain or meningeal metastasis; no more than 2 prior episodes of chemotherapeutic treatment (irrespective of the indication); World Health Organization performance status ≤2; white blood cell count ≥3000/mm; platelet count ≥100,000/mm3; hemoglobin ≥9 g/dL; and international normalized ratio, activated partial thromboplastin time, liver transaminase levels, total bilirubin, serum creatinine, and amylase and lipase levels all ≤1.5 times the upper normal limit. Disease progression within the past 6 months was defined as progression according to RECIST 1.1 guidelines.[13]


Patients received a starting dose of oral sorafenib 400 mg twice daily until they reached intolerable toxicity, had tumor progression, or withdrew informed consent. Dose reduction to 400 mg daily and then to 200 mg daily was permitted for patients who experienced severe toxicities (grade 3 or recurrent grade 2 toxicities). In the event of treatment discontinuation for toxicity, if the toxicity grade was 1 or 0, then the drug was reintroduced within 3 weeks; in contrast, treatment was definitively discontinued for toxicity grade ≥2.

Study Endpoints

Given the indolent course of the disease, we chose 9-month progression-free survival as the primary endpoint. The secondary endpoints were as follows: 1) safety and toxicity assessed according to the National Cancer Institute Common Toxicity Criteria (version 3.0); 2) response rates at 2 months, 6 months, and 9 months; 3) overall survival; and 4) the median time to progression.

During the study, patients underwent clinical, hematologic, and biologic evaluations at baseline and on days 1, 7, 15, 30, 60, 120, 180, and 270. Response to treatment was assessed by comparing unidimensional tumor measurements (computed tomography scans) in pretreatment and per-treatment imaging studies at 2 months, 4 months, 6 months, and 9 months. We assessed the treatment response according to RECIST 1.1 guidelines. An independent, third-party radiologist reviewed selected imaging studies from all imaging performed during the treatment period with the trial drug to ensure the consistent and unbiased application of RECIST.

Sample Size Calculation and Statistical Analysis

This study was designed as an exploratory, hypothesis-generating, proof-of-concept study. The sample size was chosen (n = 15) based on practical considerations rather than statistical power and type I error rate calculations, with the objective of demonstrating the biologic activity of this targeted therapy on a very homogenous group of patients with EHE.

Ethical and Regulatory Considerations

The sponsor of this study was the Oscar Lambret Cancer Center on behalf of the French Sarcoma Group (GSF/GETO). The study was opened in 9 regional centers: Claudius Regaud Institute, Toulouse; Gustave Roussy Institute, Villejuif; Leon Berard Center, Lyon; Bergonie Institue, Bordeaux; Georges-Francois Leclerc Center, Dijon; Oscar Lambret Center, Lille; Jean Minjoz Hospital, Besancon; Jean Perrin Center, Clermont-Ferrand; and Rene Gauducheau Center, Nantes. Study investigations were conducted after approval by the regional ethics committee (approved June 16, 2009) and after declaration to the French Health Products Safety Agency (approved June 1, 2009). Informed consent was obtained from each patient. The study was registered in the European Union Clinical Trials Register (EudraCT no. 2007-004651-10) and on the website (trial no. NCT00874874). It was conducted in accordance with the Declaration of Helsinki and the International Conference on Harmonization of Good Clinical Practice guidelines. The AngioNexT Study includes 5 different strata: superficial angiosarcoma, visceral angiosarcoma (both already published[7]), EHE (reported here), solitary fibrous tumor or hemangiopericytoma, and chordoma (available at: =633547&version=healthprofessional [accessed September 2012]).


Patient Characteristics

From June 2009 to February 2011, 15 patients (9 men and 6 women) were enrolled in 8 centers. The median age was 57 years (range, 31-76 years). All patients were eligible and had evidence of tumor progression at the time of study entry. Performance status was zero in 10 patients (67%) and 1 in 5 patients (33%). The most common primary sites were the liver in 4 patients, the head and neck in 2 patients, and the superficial trunk in 2 patients. The other primary sites (1 patient per location) were as follows: lower limb, upper limb, buttock, pleura, retroperitoneum, lung, and unknown primary site. One tumor arose in irradiated fields (in a patient with head and neck EHE). The tumor was locally advanced and was not amenable to surgery in 3 patients. Metastases were present at study entry in 12 patients (80%), and the most common metastatic sites were as follows: lung (12 v patients), liver (5 patients), and bone (3 patients). Prior treatments included surgery in 7 patients (46%), radiotherapy in 5 patients (33%), and chemotherapy in 5 patients (33%). The best responses obtained with the prior chemotherapy were disease progression in 3 patients, stable disease (lasting 11 months) in 1 patient, and not assessable in 1 patient. The number of prior line(s) of chemotherapy was 1 line in 2 patients, 2 lines in 2 patients, and 3 lines in 1 patient. Five patients had previously received anthracycline, and 3 had previously received taxanes. The median interval between diagnosis and study entry was 29 months (range, 1-168 months).

Safety and Toxicity

Safety was assessable in 15 patients. The median duration of treatment was 124 days (range, from 21 to ≥271 days). Three patients (20%) required dose reductions (usually in the first 4 weeks), and 5 patients (33%) required transient drug discontinuation. The causes of transient treatment discontinuation were as follows: grade 3 hand-foot syndrome in 2 patients, grade 3 diarrhea in 1 patient, and grade 3 mucositis plus anorexia in 1 patient. Eight patients (53%) did not require transient drug continuation or dose adaptation. Table 1 lists the toxicities observed. The safety profile was consistent with previous reports. No toxic deaths were noted.

Table 1. Maximum Treatment-Related Toxicity, n = 15
 No. of Patients (%)a
ToxicityGrade 2Grade 3Grade 4
  1. a

    Grade 2 through 4 adverse events according to National Cancer Institute Toxicity Criteria, version 4, in all patients indicate the incidence of maximal toxicity considered by the investigator as possibly, most likely, or definitely related to sorafenib.

Hand-foot syndrome1 (6.6)3 (20)0 (0)
Other skin reactions3 (13.3)1 (6.6)0 (0)
Mucositis1 (6.6)0 (0)0 (0)
Anemia0 (0)0 (0)1 (6.6)
Thrombocytopenia0 (0)0 (0)0 (0)
Neutropenia0 (0)0 (0)0 (0)
Fatigue3 (20)1 (6.6)0 (0)
Loss of appetite1 (6.6)1 (6.6)0 (0)
Vomiting0 (0)0 (0)0 (0)
Diarrhea1 (6.6)1 (6.6)0 (0)
Abdominal pain1 (6.6)0 (0)0 (0)
Constipation1 (6.6)0 (0)0 (0)
Acute pancreatitis1 (6.6)0 (0)0 (0)
Arthromyalgia1 (6.6))0 (0)0 (0)
Bleeding0 (0)0 (0)0 (0)
Arterial hypertension0 (0)0 (0)1 (6.6)
Left ventricular function drop0 (0)0 (0)0 (0)
Delirium1 (6.6)0 (0)0 (0)
Convulsion0 (0)1 (6.6)0 (0)


The median follow-up was 12 months, the median progression-free survival was 6 months (Fig. 1), and the median overall survival (OS) was not reached (2-year OS rate, 26%; 4 of 15 patients) (Fig. 2). The treatment response was assessable for 13 patients (Table 2). According to the study protocol, patients who were assessable for response were those who had received sorafenib for at least 28 days (2 patients were not considered assessable because of definitive drug discontinuation for unacceptable toxicity within the first 28 days). The nonprogression rate at 9 months was 30.7% (4 of 13 patients). The nonprogression rates at 2 months, 4 months, and 6 months were 84.6% (11 of 13 patients), 46.4% (6 of 13 patients), and 38.4% (5 of 13 patients), respectively. No complete responses were observed (Fig. 3). Two patients experienced partial responses lasting 2 months and ≥14 months.

Figure 1.

Progression-free survival is illustrated.

Figure 2.

Overall survival is illustrated.

Table 2. Activity Endpoints, n = 15
 No. of Patients (%)
EndpointAt Day 60At Day 120At Day 180At Day 290
Disease status    
Not assessable2222
Progressive disease2789
Stable disease9543
Partial response2111
Complete response0000
Overall response rate2 (13.3)1 (7.7)1 (7.7)1 (7.7)
Overall nonprogression rate11 (84.6)6 (46.4)5 (38.4)4 (30.7)
Figure 3.

This waterfall plot indicates the best objective response according to Response Evaluation Criteria in Solid Tumor (RECIST) 1.1 guidelines.


We conducted a phase 2 trial of sorafenib in patients with EHE as a subgroup from a larger phase 2 trial investigating sorafenib in “vascular” sarcoma. All patients had progressive disease at the time of study entry. The tolerance to the drug was as expected. Sorafenib provided long-lasting (>6 months) tumor growth arrest in >33% of patients. Partial responses were rare (2 of 15 patients).

The literature on the treatment of EHE with antiangiogenic agents is very limited and consists of a large retrospective study that was presented as an abstract and is not yet published,[14] some case reports, and 2 prospective series, as summarized in Table 3.[15-21] Cioffi et al reported a retrospective study of 42 other patients with metastatic EHE. Thirty-four of those patients received systemic treatment, including conventional doxorubicin-containing regimens (16 patients), other chemotherapy regimens (6 patients), and angiogenesis-targeting drugs or tyrosine kinase inhibitors (12 patients: sorafenib in 6 patients, thalidomide in 2 patients, metronomic cyclophosphamide in 2 patients, and imatinib mesylate in 2 patients). No objective responses were reported. Fourteen patients experienced stable disease for 6 months. The patients with stable disease included 4 of 16 patients (25%) who received doxorubicin-containing regimens, 4 of 6 patients (25%) who received other chemotherapy regimens, and 6 of 12 patients who received angiogenesis-targeting drugs or tyrosine kinase inhibitors (3 of 6 patients who received sorafenib [50%], 2 of 2 patients who received metronomic chemotherapy, and 1 of 2 patients who received thalidomide). The median progression-free survival was 4.8 months, and the median OS was 41.0 months. Nevertheless, in that retrospective study, the number of patients who experienced progressive disease before starting systemic treatment was not reported.[14] Limited experience with treatments like thalidomide, interferon 2a-α, and bevacizumab have been reported (Table 3). These treatments seem to provide long-lasting stable disease in all patients and produced a partial response in 1 patient.[16] Agulnik et al recently reported the activity of bevacizumab in patients with vascular sarcoma (mainly angiosarcoma[21]). Seven patients with EHE received bevacizumab in that phase 2 trial. The reported activity was a response in 2 of 7 patients, a median progression-free survival of 10 months, and a median OS of 36 months. Nevertheless, the number of patients who had documented disease progression before study entry was not stated in the report by Agulnik et al. Regarding disease heterogeneity, the status of disease heterogeneity before treatment, the heterogeneity of treatment, and direct comparison of the objective response rate, median progression-free survival, and median OS across the reported data were not relevant.

Table 3. Activity of Antiangiogenic Agents in Patients With Epithelioid Hemangioendothelioma
TreatmentNo. of PatientsORTTP, moOS, moReference
  1. Abbreviations: IFN, interferon; OR, objective response; OS, overall survival; PFS, progression-free survival; TTP, time to progression.

Case reports     
IFN2a-α, 3 million U/d, 3 d/wk subcutaneously10≥3Radzikowska 2008[15]
IFN2a-α, 6 million U/d, 3 d/wk subcutaneously11≥24≥24Calabro 2007[16]
Thalidomide, 200 mg/d orally10≥36≥36Mascarenhas 2004[17]
Thalidomide, 400 mg/d orally10≥84Raphael 2010[18]
Thalidomide, 300 mg/d orally10≥108Salech 2011[19]
Bevacizumab, 7.5 mg/m2/3 wk intravenously10≥16Trautmann 2011[20]
TreatmentNo. of PatientsORPFS, moOS, moReference
Prospective series: Part of phase 2 trials     
Bevacizumab, 15 mg/kg/3 wk intravenously7210≥36Agulnik 2013[21]
Current series1526 

This study had several limitations. The major methodological limitation was the choice of primary endpoint (9-month progression-free survival). We chose this long-term outcome because most EHEs have an indolent course, and we believe that clinical benefit has to be measured after several months (at least after 6 months). We cannot justify this choice based on previously reported clinical trials, because there is still no published trial focusing on this exceptional form of sarcoma. Furthermore, because of the lack of a statistical hypothesis, the observed data remain difficult to interpret. Nevertheless, in a previously reported retrospective study,[14] there was no objective response, whereas sorafenib provided 2 partial responses in the current study. However, 1 of those objective responses lasted 2 months, which may be considered an unconfirmed partial response.

We conclude that sorafenib is a potentially active drug for the treatment of patients with EHE. These findings require further clinical investigation and, if possible, a randomized trial, eg a double-blind, placebo-controlled trial or a phase 2 randomized discontinuation trial (maintenance of sorafenib vs drug holiday in nonprogressive patients treated with sorafenib)m[22, 23] However, this tumor is so rare that an international effort will be needed to achieve this goal.


The authors made no disclosures.


This work was supported by Institut National du Cancer (PHRC-2009/grant) and Bayer HealthCare France.