Introduction

Hemangiopericytoma (HPC) and Solitary Fibrous Tumors (SFTs) are rare vascular malignancies that derive from Zimmerman’s pericytes, modified smooth muscle cells that line capillaries, postcapillary venules and sinusoidal spaces [1]. Primary HPC-SFT tumors occur at various sites across the body; the most common sites for HPC-SFT are the thigh (25.5%), pelvic retroperitoneum (24.5%) and head and neck (16.0%) [2]. HPC-SFT predominantly affect young adults with an average age at diagnosis of 41–48 years [3,4,5].

Intracranial HPC-SFT tumors are rare (0.4% of all primary CNS tumors), slow growing, extra-axial tumors that radiologically and macroscopically resemble meningioma. HPC-SFT account for around 2.4% of suspected meningiomas [3] and often the diagnosis is made during resection where HPC bleed profusely. Despite their similar appearance and location to meningioma on imaging, intracranial HPC are more aggressive, more likely to recur locally and to metastasise [3].

The World Health Organization (WHO) classification of Central Nervous System tumors (1993) defines HPC as a distinct class from meningioma [6] and in the 2016 update from previous classifications, HPC have been reclassified combining previously named solitary fibrous tumors with traditional HPC. The tumors have been grouped together as there is a growing body of evidence that both HPC and SFT share similar histological [7] and immunohistochemical appearances [7, 8] and cannot be reliably differentiated. Conventionally low (WHO I) grade tumors with low cellularity and a ‘patternless architecture’ were classified as SFT, WHO II tumors are described as HPC and higher grade (WHO III) tumors with increased cellularity and mitotic number were classified as anaplastic HPC.

Intra-cranial HPC-SFTs are usually managed with surgical resection. Extent of resection is correlated with survival [4, 5] but does not reduce the probability of local recurrence or extra-cranial metastases [4, 5]. Although it has not been investigated robustly, there is evidence to suggest that grade of tumor is related to likelihood of tumor recurrence [9, 10] but has no confirmed association to extra-cranial metastasis [3, 11, 12]. Post-operative periodic follow-up cranial imaging is routine but screening for extra-cranial disease and the most appropriate imaging modality for this remains a question faced by neuro-oncology multi-disciplinary teams.

In this systematic review we discuss the frequency and location of extra-cranial metastases in patients with confirmed intra-cranial HPC or SFT. We investigate the factors that may predispose patients to extra-cranial metastasis and consider evidence for routine monitoring for extra-cranial disease and the appropriate imaging modality for this purpose.

Materials and methods

Criteria for considering studies for this review

All studies published from 1/1/1980 to 01/02/2018 that describe primary intracranial HPC and/or SFT with extra-cranial metastases were included in this study. Studies published before 1980 or in languages other than English were excluded as were papers in which it was unclear whether the primary tumor was intra or extracranial. Orbital primary tumors were also excluded from the analysis.

Literature search methods

Based on the PRISMA guidelines, the authors searched Pubmed (http://www.ncbi.nlm.nih.gov/pubmed). Studies were identified up to 1st February 2018. Search terms included intracranial, hemangiopericytoma, HPC, solitary fibrous tumor/tumor, SFT, metastases, extra-cranial metastases, recurrence, recurrent, imaging, MRI, CT, PET, monitor and follow up. Reference lists of identified articles were reviewed to detect other relevant citations.

Data extraction and management

Data were extracted using a standard data collection form created for this review. Extracted data consisted of study characteristics (number of participants, mean age, gender, follow-up length), details of the primary disease (age of onset, location, tumor grade, treatment, local recurrence) and details of extra-cranial metastases (location, time to diagnosis, methods of detection ± imaging modality for diagnosis, suggested systemic monitoring protocols where applicable).

Statistical analysis

Univariate logistic regression was performed in SPSS for the following variables: age, gender, primary tumor grade, local recurrence, surgical treatment (gross total resection vs. sub-total resection), adjuvant treatment to initial primary disease (external beam radiotherapy, stereotactic radiotherapy, proton beam therapy, chemotherapy) on the likelihood that patients would develop extra-cranial metastasis. The logistic regression model was considered statistically significant, χ2(4) = 27.402, p < 0.0005.

Multivariate logistical regression was not performed as a complete data set was only present in 27 patients.

Results

Seventy-one studies [3,4,5, 9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76] were identified documenting 904 cases of HPC-SFT and 251 cases of extra-cranial metastasis. Study characteristics and details of the primary tumor and management, where given, are shown in Table 1 [3,4,5, 9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76]. Follow up duration was not always documented but the mean recorded follow up ranged from 4 [66, 72] to 312 months [74]. Mean age at diagnosis of primary tumor for patients in all studies was 42.0 years.

Table 1 Details of the included studies: number of patients with extra-cranial metastases, average time to metastasis and details on the grade and treatment of the primary tumors for those patients
Full size table

Duration of follow up

The rate of extra-cranial metastases following a diagnosis of primary intra-cranial HPC or SFT ranged from 3.7 [46] to 69.0% [77] in all case series to 6.3 [13] to 69.0% [77] in case series with follow up of over 60 months (5 years). In those case series with a long follow up of almost a decade (> 9 years) there was a 11.6 [5] to 69.0% [77] rate of extra-cranial metastases. Range of time to extracranial metastases was from 3 [76] to 372 months [16] (Table 1).

Location of extra-cranial metastasis

Intra-cranial HPC-SFT can metastasize to multiple extra-cranial sites (Table 2). The most common sites for metastases were Bone NOS (location not otherwise specified) (19.6%) followed by lung and pleura (18.4%), liver (17.6%), and vertebrae (14.1%). Further sites of metastases listed include kidney, pelvic bones, femur, pancreas, retroperitoneum, peritoneum, the soft tissues, skin, muscle, ocular, breast, adrenal gland and metastases not otherwise specified.

Table 2 Location and number (no.) of extra-cranial metastases (metastasis) from primary intracranial HPC-SFT where site of metastasis was specified
Full size table

Predictors of extracranial metastasis

Tumor grade

Details of tumor grade were documented in 301 cases. High grade primary tumors (WHO Grade III n = 108) were 1.88 times as likely to metastasize extra-cranially as low grade tumors (WHO Grade I and WHO Grade II n = 193) (p = 0.016). WHO Grade III HPC (n = 108) were 2.53 times as likely to metastasize as WHO Grade II (n = 175) (p = 0.001).

When comparing the rates of extra cranial metastasis from WHO Grade II (n = 175) to WHO Grade I (n = 18) the Odds Ratio (OR) was 0.89 (p = 0.000017), indicating a higher rate in Grade I patients, which is a surprising result. As the majority of papers (11/13) [15, 17, 21, 22, 27, 54, 70,71,72,73,74,75] that discussed WHO grade I tumors were case reports, this analysis likely over emphasises the frequency of metastatic disease in this tumor grade. Furthermore in six cases of extra-cranial metastasis in WHO Grade 1 tumors [15, 17, 21, 27, 54], the primary tumor was originally classified as a low grade meningioma and later revised to HPC, which might raise the possibility of misdiagnosis.

Age

Age was given in 184 cases. When divided into ‘younger’ [0–40 (n = 101)] and ‘older’ [41–90 (n = 83)] older age at first diagnosis were less likely (OR 0.585) to develop extra-cranial metastasis but this was not statistically significant (p = 0.74).

Sex

Gender was recorded in 184 cases. Men (n = 93) were 1.197 times as likely to develop an extra-cranial metastasis as women (n = 91) but this was not statistically significant (p = 0.544).

Tumor location

Tumor location was accounted for in 155 cases. A posterior fossa location for the tumor had a higher chance of metastasizing when compared to a supratentorial lesion (OR 2.250) but this was not statistically significant (p = 0.76).

Local recurrence

Data was recorded for all patients with extra-cranial metastasis who had local recurrence prior to a diagnosis of metastasis and all patients without extra-cranial metastasis who had local recurrence (n = 352). Local recurrence was not predictive for developing extra-cranial metastasis (OR 0.794, p = 0.343).

Surgery

All patients underwent surgical intervention although the extent of surgery was not always specified. In those patients of whom extent of surgery was reported (n = 333) there was no difference in rates of extra-cranial metastasis between those with gross total resection (GTR) (n = 259) and those with subtotal resection (STR) (n = 74) (OR 0.644, p = 0.151).

Adjuvant therapy

Patients whose primary tumor was treated with adjuvant radiotherapy [either external beam radiotherapy (EBRT) or Stereotactic radiosurgery (SRS) n = 209/413] were more likely to develop extra cranial metastasis but this was not statistically significant (OR 1.24, p = 0.335). This may reflect the fact that radiotherapy is more likely to be used in the higher grade tumors.

When the types of adjuvant treatment to the primary tumor were subdivided [EBRT (n = 180), SRS (n = 24), Proton beam therapy (PBT) (n = 1), EBRT + chemotherapy (n = 3), EBRT + SRS (n = 1)] there was no statistically different outcomes compared with those who had no adjuvant treatment.

Diagnostic techniques for extra-cranial metastasis

Of the 71 papers identified, documentation of methods used to monitor patients with intra-cranial HPC for extracranial metastases was poor (Table 3). One group followed up patients with metastases with annual clinical review and Computed Tomography (CT) of abdomen, pelvis and chest [25]. One study documents discovery of a non-symptomatic pancreatic metastasis on a routine follow up CT scan 24 years after the initial tumor diagnosis. In one study where a patient developed pulmonary metastases, the authors recommend that all patients with intra-cranial HPC are monitored with chest X-rays at 6–12 month intervals to screen for metastases [39]. Overall the interval length of scans, length of follow up, imaging modality of choice and regions imaged were not well defined [18].

Table 3 Methods of detecting extra-cranial metastasis
Full size table

Discussion

In this review we have demonstrated that extra-cranial metastases of intra-cranial HPC-SFT are common, occurring in 28% of cases reviewed (n = 251/904). Removing case reports from this statistic, which are inherently biased to discuss the rarer cases with metastatic disease, the prevalence of extra-cranial metastasis becomes 23% (n = 202/868). When this is adjusted for studies with follow up of greater than one decade, it is shown that 1 in 1.4–8.6 people with primary intra-cranial HPC-SFT will develop extra-cranial metastases. The broad width of this estimate reflects the lack of prospective observational studies looking into the natural progression of these patients; all studies assessed were case series with variable durations of follow up. Furthermore our review demonstrates that metastases from HPC/SFT can develop at variable times after the primary tumor diagnosis, both around the time of the primary HPC or as long as 31 years after the initial diagnosis which may mean the above value underestimates extra-cranial metastases. HPC can affect patients at all stages of life but is most commonly observed in the young. Given the length of time at which metastasis can occur, life-long close clinical monitoring for intra-cranial recurrence and extra-cranial metastases is therefore recommended and supported by the literature.

Our review has demonstrated that the common sites for metastases of primary intra-cranial HPC are bone, lung, liver, other abdominal structures as well as multiple other sites across the body. This is in contrast to locations of primary HPC-SFT tumors which commonly occur at the thigh, pelvic retroperitoneum and head and neck [2]. Due to the multiple sites of metastases from intra-cranial HPC-SFT, monitoring for metastases would require whole body imaging.

Radiological surveillance of patients with HPC for extra-cranial metastases would enable early recognition of disease occurrence and enable early intervention. For a diagnostic surveillance test to be useful, it should be inexpensive, easy to administer, cause minimal discomfort or harm, and demonstrate a high sensitivity and specificity. An imaging surveillance protocol should consider (1) the time period for maximal risk of recurrence and interval between examinations (2) the most likely sites for metastases (3) treatment options if lesions were identified (4) the risks associated with the imaging modality [78]. So far there are no biological markers for HPC and imaging modalities (predominantly CT, Positron Emission Tomography (PET)-CT and MRI) are the main methods used to detect HPC. Some studies have recommended the use of whole body CT/PET in the regular monitoring of HPC patients [61]. PET provides metabolic information and the tracer is actively taken up by these highly vascular tumors. However whole body PET/CT scanning exposes patients to a substantial radiation dose and thus increased lifetime cancer risk [79]. For this reason we would argue PET-CT is not a suitable method for periodic, long-term monitoring of a relatively young patient population (although it has a potential role in diagnosis). Annual whole body CT scans in a 45 year old up till the age of 75 (30 scans) would contribute an additional lifetime risk of cancer mortality of 1.9% [80].

Serial whole body MRI scans are more expensive and more time consuming than CT [81]. However MRI does not expose the patient to harmful radiation and it is more sensitive in picking up soft tissue abnormalities; including early bone marrow infiltration in bony metastatic disease. Previous studies have demonstrated the use of whole body MRI in the monitoring of metastases in patients with saracomas [81]. The sensitivity and specificity of whole body MRI for detecting HPC metastases requires further investigation, as there is no literature on this.

Our review has shown no paradigm for imaging surveillance for HPC/SFT metastatic disease. Most extra-cranial disease was recognized on symptomatic presentation. Although survival data in the studies reviewed was lacking, we found no evidence to suggest that early detection of metastases affects outcome in this population and so cannot conclude that routine screening imaging is warranted in surveillance of extra-cranial metastases.

The only risk factor observed for extra-cranial metastasis, identified through the univariate logistic regression of the pooled studies, was the higher tumor grade (WHO Grade III have 1.88 increased risk when compared to WHO Grade I + II p = 0.016). None of the studies reviewed had independently observed this relationship. Mena et al. [9] found anaplastic tumors had a 3.3 higher risk of recurrence than lower grade (p = 0.043) and that extracranial metastasis were more frequent in anaplastic tumors but without statistical significance (OR 2.81 p = 0.41). Mena et al. [9] also found male gender and infiltration of the brain parenchyma to be associated with higher rates of metastasis (OR 7.0 and 4.8 respectively, p = < 0.05), which was not supported by our review. Whilst our analysis indicates that higher grade tumors have higher rates of metastasis, we have also shown extra-cranial metastasis occurring in patients of all tumor grades including SFT, which were previously described as benign and less likely to metastasize [69,70,71,72,73,74,75]. Clinicians therefore should be more cautious with tumors of higher grade but close clinical monitoring should not be restricted only to this group. This is an insight not previously emphasized in the literature on this tumor.

No previous studies found extent of surgical resection of the primary tumor nor the use of adjuvant radiotherapy to be related to rates of extra-cranial metastasis [3,4,5, 38, 39, 77], which is supported by our analysis. Schiariti et al. [16] reported a longer duration to extra-cranial metastasis in patients who had undergone gross total resection (170 vs. 100 months respectively, p = 0.5) and adjuvant external beam radiotherapy (139 vs. 68 months respectively, p = 0.2) but these findings were without statistical significance. Chen et al. [38] found radical resection with post-operative radiotherapy to improve overall survival and recurrence free interval (p = < 0.05) but had no effect on the metastasis free interval (p = 0.245).

Local recurrence was not an independent risk factor for extra-cranial metastasis, which is also supported by the independent findings of the studies assessed.

Conclusion

In this paper we have demonstrated that intracranial HPC can metastasize to extra-cranial sites over a long time course and variety of locations. A number of imaging modalities have been used in diagnosing extra-cranial disease including: X-ray, CT, CT-PET and MRI after patients present with symptoms. Few groups practice routine imaging surveillance to detect extra-cranial disease. There is no evidence based protocol and a wide variation in clinical practice. Higher grade of primary tumor (WHO Grade III) has a 1.88 increase risk of extra-cranial metastasis when compared to low grades (WHO I + II) but extra-cranial metastases have been seen in all tumor grades. Extent of surgical resection, location of primary tumor, use of adjuvant radiotherapy to the primary tumor, gender and patient age at first diagnosis were not influential on extra-cranial metastasis.

Our review is limited by the reliance on retrospective observational papers including 49 case reports, with varying degrees of patient follow up and clinical information. We were unable to perform a multi-variate logistic regression due to incomplete detail in many of the studies assessed (only 27 cases included information on all the variables assessed). Our univariate logistic regressions were also limited by numbers due to poor study detail.

Despite its limitations this review supports the importance of long term follow up and consistently high clinical suspicion for the possibility of extracranial metastases in HPC-SFT; including in the lower grade tumors. The use of diagnostic screening as a routine part of clinical follow up would require a multi-institutional discussion/study on the best options and given the heterogeneous group of patients and disease presentation may prove difficult to implement.