Clinical characteristics of epithelioid hemangioendothelioma: a single-center retrospective study

We retrospectively assessed patients with EHE in a Han Chinese Population who were diagnosed and treated at Peking Union Medical College Hospital between January 2000 and June 2018. All the medical records were studied systematically by two independent doctors. Patients were selected based on the following criteria: pathological diagnosis of EHE via surgery or biopsy and complete medical records. Patients with hemangioma, hemangiosarcoma, and other subtypes of hemangioendothelioma (HE) (Dabska tumor, retiform HE, kaposiform HE, pseudomyogenic HE, and composite HE) were excluded. Discordance in the reviews of the medical records was resolved through a discussion.

Data, including demographic characteristics, clinical symptoms, test results, treatment patterns, pathological results, and prognosis, were collected from both outpatient and inpatient medical records, and a retrospective database was constructed for the analysis. This study was approved by the Peking Union Medical College Hospital Institutional Review Board, and the need for informed consent was waived due to the retrospective nature of the study.

The feasibility of surgery was assessed via preoperative imaging tests. Patients with extensive intrahepatic lesions or distant metastases were considered to have unresectable disease, and intractable cases were managed through multidisciplinary consultation. The patients with resectable EHE underwent surgery. Frozen section was used to ensure negative margin, and diagnoses were confirmed via postoperative pathology. Meanwhile, patients with unresectable primary lesion or distant metastases were diagnosed via biopsy. Complications were defined as any adverse event occurring within 30 days after surgery or biopsy. Outpatient interviews, telephone calls, and e-mails were used for follow-up. Computed tomography (CT) or ultrasound was performed every 6 months during the first 2 years and annually thereafter.

The study endpoints were death or patient status on last follow-up (October 2018).

Statistical analysis was performed by an independent statistician. Linear variables were described using mean ± standard deviation, while categorical variables were presented as absolute number or frequency. Differences between groups were analyzed using Fisher’s exact test or Student’s t test as appropriate. Survival probability was estimated using the Kaplan–Meier method with log-rank test. All statistical analyses were conducted using Statistical Package for Social Sciences software (SPSS, version 19.0, Chicago, IL, USA), and a p value of < 0.05 was considered statistically significant.

Of the 40 patients with HE treated at our institution, 33 were included in the analysis (Fig. 1). The demographic data and preoperative signs and symptoms are presented in Table 1. There were 12 male patients (36.4%), and the male-to-female incidence ratio was 1:1.75. Eighteen (54.5%) patients were asymptomatic and were diagnosed with a space-occupying lesion either incidentally or during routine physical examination. Except for one patient with rectal adenocarcinoma, the remaining patients had no coexisting tumor.

All patients underwent laboratory tests including complete blood count, coagulation function, and tumor marker screening. The results of the complete blood count and coagulation function test were unremarkable. However, CA 125 levels were elevated in five patients (38.3 U/mL, 49.7 U/mL, 52.5 U/mL, 131.7 U/mL, and 1436 U/mL; reference level, 0–35.0 U/mL), and CA 19-9 levels were elevated in two patients (44.0 U/mL and 356.5 U/mL; reference level, 0–37.0 U/mL). Ultrasonography was performed in 24 patients and revealed mixed echo or hypoechoic lesion with clear or unclear boundary. CT was performed in 19 patients and showed low- or mixed-density tumors that might have delayed enhancement. Magnetic resonance imaging (MRI) was performed in six patients and revealed a space-occupying lesion with low–middle signal intensity in T1-weighted imaging and high signal intensity in T2-weighted imaging. Overall, the imaging characteristics were relatively nonspecific.

Twenty-one patients underwent surgery. Within them, two underwent local radiotherapy after scalp mass resection, and one patient who had comorbid rectal adenocarcinoma underwent chemotherapy for the rectal cancer. All the other patients received no adjuvant therapy. There were 12 patients who underwent biopsy only due to unresectable primary lesion or distant metastases. Of them, six were administered chemotherapy, while two underwent radiotherapy. Postoperative complications occurred in six patients and included fever (n = 2), wound infection (n = 1), wound inflammation (n = 1), urinary retention (n = 1), and intra-abdominal hemorrhage (n = 1). Fever, wound inflammation, and urinary retention were managed using antipyretics, wound dressing, and catheterization, respectively. Wound infection was treated via reoperation. The patient who developed intra-abdominal hemorrhage had severe underlying disease and died after splenectomy. In total, four, one, and one patients were classified into grade I, grade IIIb, and grade V complication following the Clavien–Dindo classification of surgical complications [7]. The postoperative morbidity rate was 28.6% (6/21).

All patients were pathologically diagnosed with EHE (Fig. 2). The anatomical sites of the primary lesions are shown in Table 2. The most common sites of the primary tumor were the liver, scalp, and spine. The mean tumor diameter was 5.8 ± 2.9 cm (range, 0.6–10.7 cm). Immunohistochemical staining was performed for all patients. The most common positive markers were CD34 (32/33, 97.0%) and CD31 (32/33, 97.0%), followed by vimentin (14/16, 87.5%) and F8-R (16/19, 84.2%).

Twenty-six patients (78.8%) were followed up at a mean period of 73.1 ± 60.0 months (range, 1–201 months). Six patients died during the follow-up period. One patient died from intra-abdominal hemorrhage after splenectomy, while five died from tumor-related diseases. The cumulative survival rates for the followed-up patients are shown in a Kaplan–Meier curve (Fig. 3). The 1-, 3-, and 5-year cumulative survival rates were 96.2, 87.0, and 75.3%, respectively. The 26 patients were further divided into two groups based on the follow-up results: the survivors (n = 20) and non-survivors (n = 6). There were no significant differences in sex distribution and mean age between the two groups. However, the patients who had metastases or underwent biopsy only showed significantly higher mortality than those who underwent surgery (Table 3).