Angiosarcoma: clinical outcomes and prognostic factors, a single-center analysis

Purpose This study sought to investigate oncological outcomes and prognostic factors for patients with angiosarcomas (AS). Methods This single-center, retrospective cohort study, analyzed histopathologically confirmed AS cases. Primarily diagnosed, locally recurrent and metastatic AS were included. Overall survival (OS), local control (LC) and local progression-free survival (LPFS) were assessed by Kaplan-Meier estimator. Multivariable Cox regression analysis was performed to detect factors associated with OS and LPFS. Results In total, 118 patients with a median follow-up of 6.6 months were included. The majority presented with localized disease (62.7%), followed by metastatic (31.4%) and locally recurrent (5.9%) disease. Seventy-four patients (62.7%) received surgery, of which 29 (39.2%) were treated with surgery only, 38 (51.4%) with surgery and perioperative radiotherapy or chemotherapy, and 7 (9.4%) with surgery, perioperative radiotherapy and chemotherapy. Multivariable Cox regression of OS showed a significant association with age per year (hazard ratio (HR): 1.03, p = 0.044) and metastatic disease at presentation (hazard ratio: 3.24, p = 0.015). For LPFS, age per year (HR: 1.04, p = 0.008), locally recurrent disease at presentation (HR: 5.32, p = 0.013), and metastatic disease at presentation (HR: 4.06, p = 0.009) had significant associations. Tumor size, epithelioid components, margin status, and perioperative RT and/or CTX were not significantly associated with OS or LPFS. Conclusion Older age and metastatic disease at initial presentation status were negatively associated with OS and LPFS. Innovative and collaborative effort is warranted to overcome the epidemiologic challenges of AS by collecting multi-institutional datasets, characterizing AS molecularly and identifying new perioperative therapies to improve patient outcomes.


Introduction
Angiosarcomas (AS) are rare malignant vascular tumors of endothelial cell origin [1,2].They account for 2-4% of soft tissue sarcoma (STS), have a male predominance, and most frequently affect patients aged 60-70 years [3,4].For the majority of AS, the etiology is unknown [1,2].However, established risk factors are chronic lymphedema of any origin (Stewart-Treves syndrome), exposure to ionizing radiation (typically years after radiotherapy (RT)), chemicals (vinyl chloride, arsenic, radium), vascular malformations (arteriovenous fistulas, pre-existing hemangioma), and rare genetic syndromes (neurofibromatosis, Mafucci syndrome) [1,2,5,6].Moreover, prior trauma or surgery as well as long-term exposure to implanted foreign material are associated with AS development [1,2].More than half of all AS arise from the skin followed by soft tissues of the lower extremities, retroperitoneum, trunk wall and head and neck region [1].
Within the group of STS, AS have one of the highest risks of distant metastatic spread (most commonly to the lung) and, for this reason, among others, one of the worst prognoses [7][8][9].Another remarkable feature of AS among all STS subtypes is the higher tendency to lymph node and central nervous system metastatic spread [9,10].Adverse prognostic factors in AS are older age, retroperitoneal location, large tumor size, and positive surgical margins [1,8,9,11].For localized cutaneous AS, wide surgical resection has repeatedly demonstrated its essential role in improving oncological outcomes [12,13].A number of retrospective studies suggest that the addition of postoperative RT improves overall survival (OS) and local control (LC) in localized cutaneous AS [12,14,15].For metastatic AS, taxane-based chemotherapy (CTX) regimens are recommended [10,16,17].This single-center retrospective study sought to investigate clinical outcomes, prognostic factors and the role of perioperative combination therapies for the management of AS.

Methods
In this retrospective, single-center cohort study adult patients with histopathologically confirmed AS who received treatment at our institution between 2009 and 2023 were included.Patients presenting with primarily diagnosed, locally recurrent or metastatic AS were included.Patients below 18 years of age were excluded.Data on the patient characteristics, imaging, pathology, surgical, oncological, RT treatment characteristics, and oncological outcome data were reviewed.Oncological endpoints of interest included OS, LC and local progression-free survival (LPFS).OS was defined as the time from primary diagnosis to death by any cause.LC was defined as an unchanged or decreased AS volume after surgical excision or last RT treatment (if not resected) or last CTX cycle (if not resected or irradiated) assessed by a board-certified radiologist on follow-up imaging with magnetic resonance imaging (MRI) or computed tomography (CT).LPFS was defined as the time from surgical excision or last RT treatment (if not resected), or last CTX cycle (if not resected or irradiated) to histopathological or radiographic evidence of AS volume increase or local recurrence on clinical or radiographic follow-up examinations or death by any cause.Clinical follow-up was calculated from the date of initial therapy until the last clinical visit.Radiographic follow-up was calculated from the day of initial therapy until the last available MRI or CT.Patients were censored at the last available follow-up if no local recurrence or death were observed.Data on survival status was obtained from the tumor registry of the Charité Comprehensive Cancer Center.In the event of a registered death, the information was integrated into the OS and LPFS analysis.
For descriptive statistics, ranges, medians, interquartile ranges, and means for continuous variables were used.OS and LPFS were assessed using the Kaplan-Meier estimator.Multivariable Cox regression was performed to analyze factors associated with OS and LPFS.A p-value of ≤ 0.05 was considered statistically significant.The proportional hazards assumption was tested and fulfilled for all analyzed endpoints using Schoenfeld residuals.Statistical analysis and figure design was performed with STATA MP 16.0 (Stata-Corp, College Station, TX, USA).The study was approved by the institutional review board (EA1/072/23).

Prognostic factors
In the multivariable Cox regression analysis, age per year (hazard ratio: 1.03, p = 0.044), and metastatic disease at presentation (hazard ratio: 3.24, p = 0.015) were significantly associated with OS (Table 3).Tumor size, epithelioid components, perioperative RT and/or CTX, and locally recurrent disease at presentation were not significantly associated with OS.For LPFS, age per year (hazard ratio: 1.04, p = 0.008), locally recurrent disease at presentation (hazard ratio: 5.32, p = 0.013), and metastatic disease at presentation (hazard ratio: 4.06, p = 0.009) had significant associations (Table 4).Tumor size, epithelioid components, margin status, and perioperative RT and/or CTX were not significantly associated with LPFS.

Subgroup analyses
The OS and LPFS Kaplan-Meier estimates stratified by localized vs. metastatic disease at initial presentation are displayed in Fig. 2. In patients with localized disease, the median OS of 31.1 months was strikingly longer than the median OS of 6.2 months in patients presenting with metastatic disease (Fig. 2A).Similarly, the median LPFS of 15.1 months in localized disease at presentation was longer than the 4.4 months in patients with metastatic disease (Fig. 2B).

Discussion
In this single-center retrospective cohort study, we report data on 118 AS patients.Despite intensive treatment, the overall prognosis of patients is limited.Age and metastatic disease at presentation were significant adverse prognostic factors for OS, and local disease progression after therapy.
The median age of 67 years at initial presentation is within the range described in large analyses, ranging from 62 to 73 years of age [11,18,19].Age, as an important The combination of surgery and RT was associated with improved OS, LC and disease-specific survival in a retrospective single-center cohort study on 70 localized AS cases from the MD Anderson Cancer Center by Guadagnolo et al. [12].In our study, combination therapies (surgery with RT or surgery with CTX vs. surgery alone) were not significantly associated with OS or LPFS.These findings can be explained by the inclusion of metastatic cases in our study and the smaller number of analyzable cases who were treated with combination therapies.For the LC endpoint, the low number of local recurrences in our cohort impeded multivariable Cox regression analyses.Current European guidelines recommend the combination of wide surgical excision and preoperative CTX or RT for localized AS [10].Thus, larger cohort studies are needed to further investigate the role of perioperative treatments in addition to surgery for localized AS.In our cohort, locally recurrent disease at presentation was significantly associated with LPFS and not significantly associated with OS.Lahat et al. did not detect a significant associations between this factor and local disease recurrence or disease-specific survival in a single-center study on 222 AS patients (11).Although presenting with locally recurrent disease does suggest an aggressive tumor biology and could indicate a higher risk of further local disease recurrences as seen in our cohort, the small subgroup of 7 patients in our cohort may be biased and results should be interpreted in light of this limitation.
Tumor size of > 5 cm is repeatedly described as another adverse prognostic factor for OS and disease-specific survival in AS patients [12,19].We did not identify a significant association with OS or LPFS.This may be on the one hand due to missing values and on the other hand due to difficulties assessing tumor sizes in AS.In comparison to other sarcomas, AS are known to diffusely infiltrate the skin or cause multifocal disease which often impede correct and precise measurements of tumor size [2].Epithelioid features have been reported to correlate with increased rates of local disease recurrence and mortality [8,11,23].In our study cohort, we were not able to confirm this observation and even found a tendency towards reduced risk of a LPFS clinical factor associated with OS and LPFS in our cohort, was repeatedly identified in other studies as well [8,9,18].This association is not very surprising, as, depending on the tumor location, the curative treatment requires intensive multimodal therapies.Moreover, AS are prone to local and distant recurrences requiring repeated demanding therapies, which may not be amenable to patients of a certain age without risking considerable therapy-limiting side effects [9].As expected, distant metastasis at initial presentation was also associated with poor OS and LPFS in our cohort.This finding confirms previous studies and may account for the short OS in our cohort, as a comparably large proportion (31.4%) of patients initially presented with metastatic disease [16,[19][20][21][22]. Additionally, the median OS of 19.3 months and 1 year OS of 57.5% reported herein are comparable to previous cohort studies on localized and metastatic AS [9,11].The authors found striking differences in tumor mutational burden, and the pattern of molecular alterations among the different locations and subtypes of AS.For instance, AS of the head and neck region not only had a significantly higher tumor mutational burden compared to AS of other regions but also a mutational signature commonly found in damage from ultraviolet light indicating solar radiation to be a reproducible etiology for AS in sun-exposed areas [25].A similar, ultraviolet light-induced mutational pattern was also described in melanoma and correlates with response to immune checkpoint inhibitors [26][27][28].Intriguingly, about 21% of all AS samples in the Angiosarcoma Project carried PIK3CA mutations which are targetable with FDA-approved drugs according to the OncoKB precision oncology database [25,29].Although the sample size is yet too small to draw firm conclusions for clinical practice, the project still has taken an important, innovative and patientinclusive step towards overcoming AS as rare cancers and unravelling the group of AS to identify targetable mutations for clinical trials.Future results on this project are awaited with great interest.event with epithelioid features.However, the informative value is limited due to the small number of cases.
Secondary AS due to radiation exposure (stochastic effect of RT) are not the most common radiation-induced sarcomas.However, they do represent the most common subtype of secondary sarcoma after irradiation of breast tissue [24].In line with this, 11 out of 13 radiation-induced AS in our cohort were breast AS.Moreover, the median time of 7.7 years from previous radiation exposure to the development of radiation-induced AS in our study confirms previous studies reporting median times of 6.5-8 years [11,18].
Painter et al. undertook an innovative and inclusive approach to overcome the rarity of AS and gain valuable insights into its mutational landscape in the Angiosarcoma Project [25].In this project, the authors generated an open and easy to use online platform for AS patients and relatives in the US and Canada to consent for uploading medical records and send in saliva and blood samples, which will be verified and analyzed with whole exome sequencing.Within 18 months, 227 patients consented and a total of 47 samples from 36 patients were used for genomic analysis.

Future perspective
A number of different treatment schedules for AS are investigated in clinical trials that are currently recruiting or have been completed with pending results.Based on preclinical data showing high ß-adrenergic receptor expression in AS and hemangioma and the clinical benefits of beta-blockers for infantile hemangioma, a drug-repurposing trial using propranolol in addition to standard therapy for AS has recently reached completion [30][31][32].The results are pending.The aforementioned similarity in mutational patterns of head and neck AS to melanomas and the proven efficacy of immune checkpoint inhibition in melanomas and in multiple other cancer types with high tumor mutational burden lay a valuable foundation for future studies (25,27,28,33).Several smaller studies reported durable responses with checkpoint inhibitors in the treatment of AS (34)(35)(36).An ongoing multi-center Scandinavian phase 2 study is examining this rationale using pembrolizumab combined with the betablocker propranolol for advanced AS or undifferentiated pleomorphic sarcoma to study the effect on progression-free survival (37).Another prospective, multi-center, single-arm trial examines the overall response rate after 6 months of checkpoint inhibition with cemiplimab in cutaneous and secondary AS (38).A different approach is investigating synergistic effects of paclitaxel with RT for localized cutaneous AS in an ongoing clinical trial treating patients with 6 weekly cycles of paclitaxel followed by paclitaxel with RT (39).The phase 3 TAPPAS trial compared the non-VEGF pathway anti-angiogenic monoclonal antibody carotuximab combined with pazopanib to pazopanib alone in patients with advanced AS (40).The study did not reach its primary endpoint of superior progression-free survival by the combination therapy over pazopanib alone (40).Despite its negative results, the study demonstrated how multi-center efforts can overcome the challenges in rare cancers by achieving a remarkable number of 114 recruited AS patients within two years of recruitment.Future multicenter joint projects are necessary to identify molecular targets and test new combination therapies to improve outcomes for AS patients.

Limitations
The present study carries the limitations inherent to retrospective, single-center studies.The total number of reported LC events was too small to apply multivariable Cox regression analyses for this endpoint.Importantly, our study also included patients presenting with locally recurrent or metastatic disease which may skew the observed survival times.Missing values in certain subdomains (epithelioid features, tumor size, primary tumor location) also limit the informative value.However, it also reflect the clinical reality, as  assessing epithelioid features in a rare STS subtype may not be a standard procedure and measuring tumor size in often diffusely infiltrating, multifocal AS is challenging or impossible.

Conclusions
AS are rare and aggressive subtypes of STS with dismal prognoses.Older age and metastatic disease at initial presentation status were adverse prognostic factors for local recurrence and OS in our study cohort.Innovative and collaborative effort is warranted to overcome the challenges of AS through collecting multi-institutional datasets, further characterizing AS molecularly and identifying new perioperative therapies to improve patient outcomes.

Fig. 1
Fig. 1 Overall survival (A), local progression-free survival (B) and local control (C) in the entire study cohort proved by the local institutional review board (EA1/072/23).Conflict of interestThe authors declare that they have no competing interest related to the presented work.Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made.The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material.If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this licence, visit http://creativecommons. org/licenses/by/4.0/.

Table 1
Patient and treatment characteristics

Table 3
Multivariable Cox proportional hazards model for overall sur-

Table 4
Multivariable Cox proportional hazards model for local pro-