Accepted Manuscript Prognostic features in angiosarcoma of the head and neck : A retrospective monocenter study Therese Dettenborn , Kai Wermker , MD , DMD Hans-Joachim Schulze , MD , PhD Martin Klein , MD , DMD , PhD Volker Schwipper , MD , DMD , PhD Christian Hallermann , MD , PhD PII:
S1010-5182(14)00158-9
DOI:
10.1016/j.jcms.2014.05.002
Reference:
YJCMS 1809
To appear in:
Journal of Cranio-Maxillo-Facial Surgery
Received Date: 25 February 2014 Revised Date:
24 March 2014
Accepted Date: 6 May 2014
Please cite this article as: Dettenborn T, Wermker K, Schulze H-J, Klein M, Schwipper V, Hallermann C, Prognostic features in angiosarcoma of the head and neck : A retrospective monocenter study, Journal of Cranio-Maxillofacial Surgery (2014), doi: 10.1016/j.jcms.2014.05.002. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Prognostic features in angiosarcoma of the head and neck : A retrospective monocenter study
Therese Dettenborn
1)
Kai Wermker , MD , DMD
2)
* 3)
Hans-Joachim Schulze , MD , PhD Martin Klein , MD , DMD , PhD
RI PT
Authors :
2)
Volker Schwipper , MD , DMD , PhD 3)
SC
Christian Hallermann , MD , PhD
2)
M AN U
1) Department of Plastic Surgery (Head : Dr. A. Krause-Bergmann), Fachklinik Hornheide, Dorbaumstrasse 300, 48157 Muenster, Germany
2) Department of Oral and Cranio-Maxillofacial Surgery (Head : Prof. Dr. Dr. M. Klein), Head and Neck Cancer Centre, Fachklinik Hornheide, Dorbaumstrasse 300, 48157 Muenster, Germany
TE D
3) Department of Dermatology and Histopathology (Head : Priv.-Doz. Dr. H.-J. Schulze), Skin Cancer Centre,
* Corresponding author :
EP
Fachklinik Hornheide, Dorbaumstrasse 300, 48157 Muenster, Germany
AC C
Kai Wermker , MD , DMD Fachklinik Hornheide
Department of Oral and Cranio-Maxillofacial Surgery , Head and Neck Cancer Centre Dorbaumstrasse 300 , 48157 Muenster , Germany email : [email protected] , [email protected] Fax : 0049 251 3287 424 Tel. : 0049 251 3287 421
Article type : Original Research Article
ACCEPTED MANUSCRIPT
ABSTRACT
Background : Cutaneous angiosarcoma of the head and neck (cAS-HN) is a rare malignancy with poor survival. Most of
RI PT
the histological markers and grading were not proven to be significant for prediction of outcomes in cAS-HN. This study aimed to find prognostic clinical features and histologic markers for cAS-HN.
Material and Methods: We retrospectively analysed primary cAS-HN´s seen in a single institution between 1980 and 2009. Clinical data and specific histologic characteristics were assessed. Outcome parameters were analysed using uni-
SC
and multivariate statistics.
Results: 80 patients (mean age 71.4 (SD 14.4) years, average follow-up time 55.3 (SD 74.4) months) were included. 5-
M AN U
year DSS rate was 62%. Univariate analysis revealed the extent of primary tumour (affecting more than one anatomical region), incomplete resection and initial metastatic disease as significant (p 5 cm), resectability, depth of
M AN U
growth, mitotic activity, pattern of growth, stage at initial diagnosis in AS in general – not differing between AS localization (head and neck, trunk or extremities) or aetiology (after irradiation, as sequelae of lymphoedema or sporadical) (Buehler et al., 2013; Guadagnolo et al., 2011; Ogawa et al., 2012; Perez et al., 2013; Yeang et al., 2013). Results are inconsistent and sometimes even conflicting.
TE D
In addition cAS presents a highly variable clinical presentation that often leads to delayed diagnosis and poor prognosis at the initial diagnosis. Early lesions can simulate many benign lesions, including bruises, haemangioma, infection, or inflammatory disorder. Atypical clinical presentations that have been reported include rosacea, rhinophyma, isolated eyelid ooedema, scarring alopecia, and chronic episodic facial oedema. It may be of importance
EP
to consider the initial clinical presentation of cAS and the delay it takes to achieve the correct diagnosis (Deyrup et al.,
AC C
2008a; Morgan et al., 2004).
More than half of cAS, primary and sporadic, occur in the head and neck (cAS-HN), precisely in the face and scalp, first described by Jones in 1964 (Jones, 1964). The scalp is the most common affected site in up to 48% (Albores-Saavedra et al., 2011; Rouhani et al., 2008). Until now prognostic factors in this disease are very limited. Older studies showed that histological tumour grading is of no or minimal prognostic significance (Holden et al., 1987; Morgan et al., 2004; Stewart and Treves, 1948). Distinct reliable prognostic variables for cAS, and especially cAS of the head and neck region (cAS-HN) are rare.
ACCEPTED MANUSCRIPT
Recently Deyrup and colleagues presented a retrospective study and proposed a new stratification system with prognostic significance in their collection of sporadical AS (Deyrup et al., 2008a). They used epidemiologic data and histologic features as pattern of growth (vasoformative versus solid), nuclear grade (high versus low), necrosis (present or absent), cell type (epithelioid or spindled), extent of inflammatory infiltrate (minimal versus marked), and mitotic rate. Tumours were stratified into low or high risk groups based on necrosis and/or epitheloid features. They
RI PT
showed that most tumours invading the dermis and subcutis, higher age of patients, anatomic site of occurrence, necrosis and epitheloid features correlated with increased mortality. High risk group and age above 70 years were associated with increased mortality. The depth of tumour invasion (dermis and/or subcutis) correlated with the risk of
SC
local recurrence.
The aim our study was to clarify on the statements of Deyrup (Deyrup et al., 2008a) and to evaluate the clinical and
Materials and Methods
Patients
TE D
M AN U
few histological features that could predict behaviour and outcome of this fatal disease.
EP
Patients with angiosarcoma in the head and neck region (cAS-HN), treated in our skin and head and neck cancer
database.
AC C
center between 1980 and 2009, were identified consecutively in a retrospective manner from our institutional
This study was approved by the local Ethics committee (Ethical committee of the Westfalian Wilhems-University Muenster, Approval-No. 2006-088-f-S), was conducted in accordance with the Guidelines for Good Clinical Practice (GCP) and in compliance with the Declaration of Helsinki. All participating patients or their relatives (in deceased cases) gave their written informed consent to participate in the study. Inclusion criteria for this study were sporadical cutaneous angiosarcoma of the head and neck (cAS-HN) and complete database with histopathological secured diagnosis of cAS-HN. Exclusion criteria were incomplete datasets concerning the assessed variables, patients suffering from multiple cancers of the head and neck, patients after other malignancies in the head and neck or after already performed cancer or lymph node surgery (e.g. neck dissection).
ACCEPTED MANUSCRIPT
Patients with cAS-HN arising in lymphoedema or after radiation were also excluded from the study. In summary, we assessed only patients with primary sporadical cAS-HN and without history or treatment of other head and neck cancers. All patients received complete appropriate staging, including ultrasound diagnostics of relevant lymph node levels and abdomen and also radiological imaging by either CT or MRI of head, neck, thorax and abdomen. All patients were initially treated surgically (resection of the primary tumour) with curative intention. In total 80 patients were
RI PT
eligible for inclusion in this study.
Methods
With a specially developed data entry form the characteristics of all cases were analysed.
SC
In all cases the following relevant epidemiological and clinical parameters were assessed: age at the time of first diagnosis (divided into two groups : 70 years), sex, primary tumour site in general (face or scalp)
M AN U
and more detailed (scalp, lower third of the face, midface including the nose, upper third of the face, ear and periauricular, tumour including more than one of the above mentioned regions of the face or scalp, tumour of more than one region in the face and the scalp), extent and dimension of the primary tumour (one region versus two or more regions ; 5 cm in diameter), initial locoregional or distant metastasis, resection status (R0=
TE D
histopathological tumour free (i.e.negative) resection margins, R1= microscopically positive margins, R2= macroscopically positive margins), minimal safety margin (>= 1cm versus < 1 cm), adjuvant radiation (yes / no), adjuvant chemotherapy (yes / no), comprehensive treatment protocol (surgery , surgery + radiotherapy, surgery + adjuvant chemotherapy, surgery + combined radio-chemo-therapy). Outcome and course of disease were analysed
EP
concerning tumour recurrence (local relapse, occurrence of distant metastasis), concerning overall survival status and
AC C
time and concerning disease specific survival status and time.
Histological examinations and re-evaluations were performed by two experts in dermatopathology (HJS, CH) independently and blinded from each other. Diagnosis was proven in formalin fixed and paraffin embedded representative biopsies and excisions using conventional H&E staining as well as a panel of immunohistological stainings with antibodies against different vascular markers (CD31, CD34, Podoplanin). In cases of adequately archived tumour material, the following histological parameters were analysed retrospectively: According to the FNCLCC system tumour differentiation (Score 1: sarcoma closely resembling normal adult mesenchymal tissue ,Score 2: sarcoma for which histologic typing is certain , Score 3: Embryonal and undifferentiated sarcoma, sarcomas of uncertain type), mitosis count (Score 1: 0-9/10 HPF (high power fields) , Score 2: 10-19/10 HPF ,
ACCEPTED MANUSCRIPT
Score 3: >= 20/10 HPF) and microscopic tumour necrosis (Score 0: no necrosis, Score 1: 50% tumour necrosis). These 3 Sub-Scores were summed up to build a total score value, from which the histological grade according to the FNCLCC system is deduced (Grade 1= total score 2 or 3, Grade 2 = total score 4 or 5 , Grade 3= total score 6 to 8). The infiltration depth (dermis, subcutis, fascia or deeper), percentage of solid areas (= 80%, Figure 7) and cell type (epitheloid versus spindled cells) were evaluated. A risk stratification according
RI PT
the proposals of Deyrup (Deyrup et al., 2008a) was derived from the parameters tumour necrosis and cell type: the low risk histologic group (LRHG) is defined by absence of necrosis and epitheloid cells, the high risk histologic group (HRHG) displays epitheloid morphology and / or necrosis.
SC
Statistical analysis
To assess significance of differences concerning disease specific survival (between patients having died due to their
M AN U
cAS-HN and patients without tumour dependent death at the end of their follow-up period) Chi-square test and Fisher´s exact test were applied for categorical variables, for metric parameters Kolmogorow-Smirnov (KS) test was used as non-parametric test in not normally distributed data and independent t-test was used in normally distributed variables. Overall survival was calculated according to the Kaplan-Meier method (patients being alive at the end of the
TE D
follow-up period were censored) and group differences were tested on significance using log-rank test (univariate analysis).
Binary logistic regression analysis (BLR, multivariate analysis) was used to model the predictors of disease specific survival including patient, tumour, and treatment characteristics. Potential predictors (p < 0.05 in univariate analysis)
EP
from the Chi-square-, KS- and independent t- test analyses were entered in a stepwise forward fashion using p0.1 for removal. The following parameters were considered to be eligible for BLR: extent of primary tumour, tumour size, resection status, locoregional metastasis, initial distant metastasis, secondary distant metastasis and tumour differentiation.
To analyze potential predictors for disease specific survival time in a multivariate analysis, Cox proportional hazards model (Cox regression analysis, CRA) was used. The following variables and parameters with significant p-value < 0.05 in univariate analysis (log-rank test) were eligible for CRA: age group, extent of primary tumour, resection status, locoregional metastasis, initial distant metastasis, secondary distant metastasis, adjuvant radiation, adjuvant chemotherapy, treatment protocol and solid areas. Again a stepwise forward procedure using p0.1 for removal was chosen.
ACCEPTED MANUSCRIPT
All statistical analyses were performed by a statistician using the Statistical Package for Social Sciences, version 16.0
RI PT
(SPSS Inc., Chicago, IL, USA).
Results
In total 80 patients (30 females, 50 males, mean age at first diagnosis 71.4 +/- 14.4 years) were included in this study.
SC
Mean follow-up time was 55.3 +/- 74.4 months. 39 patients died during their observation period (overall survival), 28 of them due to their cAS-HN (disease specific survival, DSS). Overall survival rates for 3 years, 5 years, 10 years and 20
M AN U
years were 71 %, 54 %, 38 % and 20 %, respectively. Mean overall survival time was 121.6 (95% confidence interval (CI) 82.5 – 160.8) months, median overall survival time was 64.0 (95% CI 48.4 – 75.6) months. Concerning DSS, survival rates for 3 years, 5 years, 10 years and 20 years were 74 %, 62 %, 55 % and 41 % and mean and median DSS time were 174.3 (95% CI 120.9 – 227.7) months and 174.0 (95% CI 30.6 – 317.4) months, respectively. Histological re-evaluation
TE D
of the parameters described was possible in 37 patients in total. Not all clinical and histological variables were available in all cases. The main reason why only 37 histological re-evaluations were possible was that an adequate tumour specimen was no longer available, in older cases the specimen was used up beforehand in previous diagnostic procedures and analyses. In a few cases, the quality of the specimen was of inadequate quality to allow histological
EP
evaluation of some parameters. Some missing information in other variables were caused by single cases with poor or
AC C
insufficient data quality concerning an assessed parameter - mainly due to the retrospective nature of this study. An overview over the study population and considered variables with regard to DSS status are shown in Table 1. There were significant differences (p < 0.05) between patients who died due to cAS-HN and patients who had not died from their tumour, related to the extent of the primary tumour, tumour size, resection status, locoregional metastasis, initial distant metastasis, secondary distant metastasis and tumour differentiation. Concerning DSS rates and time, data of clinical and histological parameters and their impact on DSS time (univariate analysis by log-rank test) is presented in Table 2. Possible predictors (p < 0.05) of DSS time were age group, extent of primary tumour, resection status, locoregional metastasis, initial distant metastasis, secondary distant metastasis, adjuvant radiation, adjuvant chemotherapy, treatment protocol and solid areas (Figure 1). Risk stratification according to Dreyrup et al. into LRHG and HRHG revealed no statistically significant difference concerning DSS time (p = 0.440).
ACCEPTED MANUSCRIPT For calculation of a multivariate prediction model concerning DSS status (dead versus alive), the significant variables of the univariate analysis (Table 1) were entered into BLR. Included variables and details of the model are shown in Table 3. Correct prediction of DSS status by this mathematical model was reached in 75.3 % of all cases and explanation of variance reached 42.6 % (Nagelkerke´s r-square = 0.426). Extension of the primary cAS-HN to more than one region,
RI PT
initial distant metastasis and incomplete resection (R1) increased the risk of tumour dependent death.
To model prediction of DSS time in a multivariate analysis, the significant variables of univariate DSS time analysis (Table 2) were included into CRA. Table 4 displays extracted possible predictors and their statistics. Whereas age
SC
above 70 years (hazard ratio (HR) 4.8, 95% CI 1.3 – 18.0) and incomplete tumour resection (HR 6.0, 95% CI 1.8 – 20.9) increase the risk for earlier tumour dependent death, absence of initial distant metastasis reduces this risk (HR 0.2,
TE D
Discussion
M AN U
95% CI 0.1 – 0.6).
Grading according to the AJCC and staging systems have been widely used to predict behaviour and prognosis of most
EP
soft tissue sarcomas, but AS are different. AS is considered a high grade lesion because the degree of differentiation
AC C
has not been shown to correlate with prognosis and clinical outcome. Epitheloid hemangioendothelioma is considered a low grade lesion, has a better prognosis and might have affected the results of recent studies in cAS (Deyrup et al., 2008b).
Average age of patients at first diagnosis in our study was 71.4 (SD 14.4) years. In comparison to patients in other studies this was slightly younger (Brenn and Fletcher, 2005), although we excluded AS in the trunk (breast) and extremities (Stewart and Treves, 1948). It may account to the currently improving screening of skin cancer, the awareness of rare cutaneous lesions and highly specialized medical settings for further therapy. 5-year disease specific survival (DSS) was 62%. 13 of 80 patients who had distant metastatic disease initially, mainly affecting the lung as first site of metastasis (Mark et al., 1996). DSS 5-year survival might be higher than in other studies due to consequent
ACCEPTED MANUSCRIPT
radiation therapy after surgical resection in 44 patients of 80. It might have improved outcome in locally advanced tumours and incompletely resected cAS compared to patients with smaller tumour load. Nevertheless surgery with adjuvant radiation is the treatment of choice in cAS in general today and might explain the higher survival rates among our patient collective.
RI PT
In our patient collective we observed that men were more likely to develop AS, but not four to five times more likely than women as described in previous studies (Wollina et al., 2013). We attribute this difference to the highly selective subgroup of cAS-HN we investigated. In our study patients with cAS in extremities and trunk were excluded because they are in contrast to cAS-HN, who are mostly patients with postradiation tumours or at least secondary tumours
SC
and the course of the disease and prognosis is proven to be different from that of sporadic AS (Holden et al., 1987). Moreover, one quarter of the patients in the study of Deyrup had surgery alone (Deyrup et al., 2008a). Our patients
M AN U
with locally advanced tumour growth underwent combination therapy by.
Previously tumour diameter and depth of growth was proven to be the strongest predictor for prognosis (AlboresSaavedra et al., 2011; Ogawa et al., 2012). Holden et al already described tumour size as the only clinical factor
TE D
predicting the course of the disease, survival was proven to be significantly better in patients with a tumour diameter less than 5cm (Holden et al., 1987). Patients with lesions greater than 5cm in diameter did not seem to show any histological markers that served for prognostication. In the presented study tumour size (also categorized into above or less than 5cm in diameter) did not show statistically significant differences in survival parameters, whereas
EP
involvement of more than one anatomical region was proven significant (p = 0.048). Definition of margins in this type
AC C
of tumour is difficult to assess due to characteristic skip lesions. Nevertheless AS crossing area boundaries (one region versus two or more) was superior to simple diameter concerning prognostic information. This could be applicable for future work comparing data of tumour extent.
Cases with complete excision by the first interventional step, R0, not depending on the extent of tumour, demonstrated a highly significantly better survival (p < 0.001). This result is in accordance with all previous studies in the literature. Although most patients presented in advanced stages in which no standardized regime was applicable, complete excision remains the therapeutic goal (Rouhani et al., 2008).
ACCEPTED MANUSCRIPT
Weaker anatomical boundaries and a longer sun exposure without proper protection and later diagnosis due to the hair coverage of the tumour suggested a higher occurrence of larger tumour size in the scalp area compared to the face, but a larger sample size did not show a significant difference in size as a prognostic feature. Age, site of occurrence and size, mitotic activity and pattern of growth were already described as significant histological markers for prediction of the clinical course (Perez et al., 2013; Yeang et al., 2013).
RI PT
Deyrup described in 2008 selected histologic features for predicting outcome in cAS in an entity that seem unpredictable in behaviour. In their study epitheloid morphology and necrosis appeared to be significant factors for segregation of the tumour into a high or low risk group (LRHG and HRHG according to FNCLCC). The high risk group
tumour invasion only had a higher risk for recurrence.
SC
had a significantly increased disease specific mortality in patients above 70 years (Deyrup et al., 2008a). The depth of
In our study mitotic activity, necrosis and the cell type did not serve for prognostication, but pattern of growth (solid
M AN U
areas greater than 80%) was correlated with better survival (p = 0.011). Nevertheless tumour differentiation and risk stratification according to Deyrup (Deyrup et al., 2008a) did not show significant predictive values in cAS-HN (p = 0.440) .
TE D
Some limitations of this study have to be mentioned. Retrospectiveness of this study has to be kept in mind with the well-known weaknesses of retrospective study designs concerning data quality and risk for selection bias. Due to the limited number of appropriate available histological samples, the trend in differentiation and risk stratification on
AC C
larger sample.
EP
prognostication of DSS could not be proven with statistical significance, but this could probably be confirmed in a
Conversely, the specific strength of this monocenter study lies in the restriction of cAS to the head and neck region with exclusion of secondary angiosarcomas (e.g. after radiotherapy or chronic lymphoedema) and considerable clinicopathologic data. To date no previous study analysed this well-defined, homogenous subgroup of cAS with an adequate number of included patients in this way.
Conclusion
ACCEPTED MANUSCRIPT In conclusion age below 70 years, low number of affected regions in the head and neck, initial absence of metastatic disease, complete initial resection and pattern of growth with more than 80% solid areas showed significantly lower mortality rates in cAS-HN. Our results give important information for clinical risk assessment of a tumour group in a
RI PT
specific setting and necessary screening considerations for the future in skin cancer.
Sponsorship
SC
There was no financial or material support for this study by any company
M AN U
Role of funding source None
Conflict of interest
TE D
There were no financial and personal relationships with other people or organizations that could inappropriately influence (bias) our work. The authors disclose any commercial associations, current and over the past 5 years, that
Acknowledgements
EP
might pose a conflict of interest. All authors declare that there is no conflict of interest.
AC C
We thank Ms. Margret Leygraf and Coworkers for their skilled technical assistance.
References
Albores-Saavedra J, Schwartz AM, Henson DE, Kostun L, Hart A, Angeles-Albores D, Chablé-Montero F : Cutaneous angiosarcoma. Analysis of 434 cases from the Surveillance, Epidemiology, and End Results Program, 19732007. Ann. Diagn. Pathol. 15, 93–97, 2011.
ACCEPTED MANUSCRIPT
Brenn T, Fletcher CDM : Radiation-associated cutaneous atypical vascular lesions and angiosarcoma: clinicopathologic analysis of 42 cases. Am. J. Surg. Pathol. 29, 983–996. 2005. Buehler D, Rice SR, Moody JS, Rush P, Hafez GR, Attia S, Longley BJ, Kozak KR : Angiosarcoma Outcomes and Prognostic Factors: A 25-Year Single Institution Experience. Am. J. Clin. Oncol. , 2013 Deyrup AT, McKenney JK, Tighiouart M, Folpe AL, Weiss SW : Sporadic cutaneous angiosarcomas: a proposal for risk
RI PT
stratification based on 69 cases. Am. J. Surg. Pathol. 32, 72–77, 2008a. Deyrup AT, Tighiouart M, Montag AG, Weiss SW : Epithelioid hemangioendothelioma of soft tissue: a proposal for risk stratification based on 49 cases. Am. J. Surg. Pathol. 32, 924–927, 2008b.
Guadagnolo BA, Zagars GK, Araujo D, Ravi V, Shellenberger TD, Sturgis EM : Outcomes after definitive treatment for
SC
cutaneous angiosarcoma of the face and scalp. Head Neck 33, 661–667, 2011.
Holden CA, Spittle MF, Jones EW : Angiosarcoma of the face and scalp, prognosis and treatment. Cancer 59, 1046–
M AN U
1057, 1987.
Jones EW : Malignant angioendothelioma of the skin. Br. J. Dermatol. 76, 21–39, 1964. Käcker C, Marx A, Mössinger K, Svehla F, Schneider U, Hogendoorn PCW, Nielsen OS, Küffer S, Sauer C, Fisher C, Hallermann C, Hartmann JT, Blay JY, Mechtersheimer G, Hohenberger P, Ströbel P : High frequency of MYC
TE D
gene amplification is a common feature of radiation-induced sarcomas. Further results from EORTC STBSG TL 01/01. Genes. Chromosomes Cancer 52, 93–98, 2013. Mark RJ, Poen JC, Tran LM, Fu YS, Juillard GF : Angiosarcoma. A report of 67 patients and a review of the literature. Cancer 77, 2400–2406, 1996.
EP
Mitelman F, Johansson B, Mertens : Mitelman Database of Chromosome Aberrations and Gene Fusions in Cancer
AC C
[WWW Document]. URL http://cgap.nci.nih.gov/Chromosomes/Mitelman (accessed 2.18.14), 2014. Morgan MB, Swann M, Somach S, Eng W, Smoller B : Cutaneous angiosarcoma: a case series with prognostic correlation. J. Am. Acad. Dermatol. 50, 867–874, 2004. Ogawa K, Takahashi K, Asato Y, Yamamoto Y, Taira K, Matori S, Iraha S, Yagi N, Yogi A, Haranaga S, Fujita J, Uezato H, Murayama S : Treatment and prognosis of angiosarcoma of the scalp and face: a retrospective analysis of 48 patients. Br. J. Radiol. 85, e1127–1133, 2012. Perez MC, Padhya TA, Messina JL, Jackson RS, Gonzalez RJ, Bui MM, Letson GD, Cruse CW, Lavey RS, Cheong D, Forster MR, Fulp WJ, Sondak VK, Zager JS : Cutaneous angiosarcoma: a single-institution experience. Ann. Surg. Oncol. 20, 3391–3397, 2013.
ACCEPTED MANUSCRIPT
Rouhani P, Fletcher CDM, Devesa SS, Toro JR : Cutaneous soft tissue sarcoma incidence patterns in the U.S. : an analysis of 12,114 cases. Cancer 113, 616–627, 2008. Stewart FW, Treves N : Lymphangiosarcoma in postmastectomy lymphoedema; a report of six cases in elephantiasis chirurgica. Cancer 1, 64–81, 1948. Wollina U, Koch A, Hansel G, Schönlebe J, Kittner T, Pabst F, Haroske G, Nowak A : A 10-year analysis of cutaneous
RI PT
mesenchymal tumours (sarcomas and related entities) in a skin cancer center. Int. J. Dermatol. 52, 1189– 1197, 2013.
Yeang MS, Tay K, Ong WS, Thiagarajan A, Tan DSW, Ha TC, Teo PTH, Soo KC, Tan HK, Iyer NG : Outcomes and
case-control study. Ann. Surg. Oncol. 20, 3066–3075, 2013.
SC
prognostic factors of post-irradiation and de novo sarcomas of the head and neck: a histologically matched
Young RJ, Brown NJ, Reed MW, Hughes D, Woll PJ : Angiosarcoma. Lancet Oncol. 11, 983–991, 2010.
AC C
EP
TE D
M AN U
Zahm SH, Fraumeni JF Jr : The epidemiology of soft tissue sarcoma. Semin. Oncol. 24, 504–514, 1997.
ACCEPTED MANUSCRIPT
Tables
Table 1 : Study population and considered variables with regard to DSS.
Variable
Disease specific survival (DSS) Alive
Dead
RI PT
n
Sign. (p-value)
Clinical Parameters
(0.072)
SC
Gender 50
29
21
Female
30
23
7
80
Age
M AN U
Male
70.0
73.7
SD [years]
16.4
9.5
TE D
Mean [years]
Age groups 70 years
(0.815)
AC C
Localisation of Primary Tumour Face
45
30
15
Scalp
35
22
13
(0.034)*
Extent of primary tumour One region
45
34
11
Two or more regions
35
18
17
Anatomical tumour site
(0.210)
(0.327)
Lower third of face
2
1
1
Midface incl. nose
27
18
9
Upper third of face
10
6
4
Ear / periauricular
7
7
0
> 1 region : face or scalp
9
7
2
> 1 region : face and scalp
3
2
1
RI PT
ACCEPTED MANUSCRIPT 22 11 11
Scalp
(0.018)*
Tumour size 36
18
18
> 5 cm
43
33
10
M AN U
Resection status ( pR)
SC
= 1 cm
Yes No
AC C
Locoregional metastasis
50
32
18
29
19
10
EP
Safety margins
TE D
R0
(0.545)
(0.001)**
27
11
16
48
38
10
( 70 years
(0.189)
(0.013)*
23
90
84
248.0
57
66
51
93.3
EP
1 region : face or scalp
9
80
48
59.3
> 1 region : face and scalp
3
60
38
36.0
36
64
61
> 5 cm
43
82
61
Resection status ( pR)
159.2
35
97
97
265.2
R1
43
53
28
69.9
AC C
>= 1 cm
79
66
173.4
29
64
56
98.4
(
S1010-5182(14)00158-9
DOI:
10.1016/j.jcms.2014.05.002
Reference:
YJCMS 1809
To appear in:
Journal of Cranio-Maxillo-Facial Surgery
Received Date: 25 February 2014 Revised Date:
24 March 2014
Accepted Date: 6 May 2014
Please cite this article as: Dettenborn T, Wermker K, Schulze H-J, Klein M, Schwipper V, Hallermann C, Prognostic features in angiosarcoma of the head and neck : A retrospective monocenter study, Journal of Cranio-Maxillofacial Surgery (2014), doi: 10.1016/j.jcms.2014.05.002. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Prognostic features in angiosarcoma of the head and neck : A retrospective monocenter study
Therese Dettenborn
1)
Kai Wermker , MD , DMD
2)
* 3)
Hans-Joachim Schulze , MD , PhD Martin Klein , MD , DMD , PhD
RI PT
Authors :
2)
Volker Schwipper , MD , DMD , PhD 3)
SC
Christian Hallermann , MD , PhD
2)
M AN U
1) Department of Plastic Surgery (Head : Dr. A. Krause-Bergmann), Fachklinik Hornheide, Dorbaumstrasse 300, 48157 Muenster, Germany
2) Department of Oral and Cranio-Maxillofacial Surgery (Head : Prof. Dr. Dr. M. Klein), Head and Neck Cancer Centre, Fachklinik Hornheide, Dorbaumstrasse 300, 48157 Muenster, Germany
TE D
3) Department of Dermatology and Histopathology (Head : Priv.-Doz. Dr. H.-J. Schulze), Skin Cancer Centre,
* Corresponding author :
EP
Fachklinik Hornheide, Dorbaumstrasse 300, 48157 Muenster, Germany
AC C
Kai Wermker , MD , DMD Fachklinik Hornheide
Department of Oral and Cranio-Maxillofacial Surgery , Head and Neck Cancer Centre Dorbaumstrasse 300 , 48157 Muenster , Germany email : [email protected] , [email protected] Fax : 0049 251 3287 424 Tel. : 0049 251 3287 421
Article type : Original Research Article
ACCEPTED MANUSCRIPT
ABSTRACT
Background : Cutaneous angiosarcoma of the head and neck (cAS-HN) is a rare malignancy with poor survival. Most of
RI PT
the histological markers and grading were not proven to be significant for prediction of outcomes in cAS-HN. This study aimed to find prognostic clinical features and histologic markers for cAS-HN.
Material and Methods: We retrospectively analysed primary cAS-HN´s seen in a single institution between 1980 and 2009. Clinical data and specific histologic characteristics were assessed. Outcome parameters were analysed using uni-
SC
and multivariate statistics.
Results: 80 patients (mean age 71.4 (SD 14.4) years, average follow-up time 55.3 (SD 74.4) months) were included. 5-
M AN U
year DSS rate was 62%. Univariate analysis revealed the extent of primary tumour (affecting more than one anatomical region), incomplete resection and initial metastatic disease as significant (p 5 cm), resectability, depth of
M AN U
growth, mitotic activity, pattern of growth, stage at initial diagnosis in AS in general – not differing between AS localization (head and neck, trunk or extremities) or aetiology (after irradiation, as sequelae of lymphoedema or sporadical) (Buehler et al., 2013; Guadagnolo et al., 2011; Ogawa et al., 2012; Perez et al., 2013; Yeang et al., 2013). Results are inconsistent and sometimes even conflicting.
TE D
In addition cAS presents a highly variable clinical presentation that often leads to delayed diagnosis and poor prognosis at the initial diagnosis. Early lesions can simulate many benign lesions, including bruises, haemangioma, infection, or inflammatory disorder. Atypical clinical presentations that have been reported include rosacea, rhinophyma, isolated eyelid ooedema, scarring alopecia, and chronic episodic facial oedema. It may be of importance
EP
to consider the initial clinical presentation of cAS and the delay it takes to achieve the correct diagnosis (Deyrup et al.,
AC C
2008a; Morgan et al., 2004).
More than half of cAS, primary and sporadic, occur in the head and neck (cAS-HN), precisely in the face and scalp, first described by Jones in 1964 (Jones, 1964). The scalp is the most common affected site in up to 48% (Albores-Saavedra et al., 2011; Rouhani et al., 2008). Until now prognostic factors in this disease are very limited. Older studies showed that histological tumour grading is of no or minimal prognostic significance (Holden et al., 1987; Morgan et al., 2004; Stewart and Treves, 1948). Distinct reliable prognostic variables for cAS, and especially cAS of the head and neck region (cAS-HN) are rare.
ACCEPTED MANUSCRIPT
Recently Deyrup and colleagues presented a retrospective study and proposed a new stratification system with prognostic significance in their collection of sporadical AS (Deyrup et al., 2008a). They used epidemiologic data and histologic features as pattern of growth (vasoformative versus solid), nuclear grade (high versus low), necrosis (present or absent), cell type (epithelioid or spindled), extent of inflammatory infiltrate (minimal versus marked), and mitotic rate. Tumours were stratified into low or high risk groups based on necrosis and/or epitheloid features. They
RI PT
showed that most tumours invading the dermis and subcutis, higher age of patients, anatomic site of occurrence, necrosis and epitheloid features correlated with increased mortality. High risk group and age above 70 years were associated with increased mortality. The depth of tumour invasion (dermis and/or subcutis) correlated with the risk of
SC
local recurrence.
The aim our study was to clarify on the statements of Deyrup (Deyrup et al., 2008a) and to evaluate the clinical and
Materials and Methods
Patients
TE D
M AN U
few histological features that could predict behaviour and outcome of this fatal disease.
EP
Patients with angiosarcoma in the head and neck region (cAS-HN), treated in our skin and head and neck cancer
database.
AC C
center between 1980 and 2009, were identified consecutively in a retrospective manner from our institutional
This study was approved by the local Ethics committee (Ethical committee of the Westfalian Wilhems-University Muenster, Approval-No. 2006-088-f-S), was conducted in accordance with the Guidelines for Good Clinical Practice (GCP) and in compliance with the Declaration of Helsinki. All participating patients or their relatives (in deceased cases) gave their written informed consent to participate in the study. Inclusion criteria for this study were sporadical cutaneous angiosarcoma of the head and neck (cAS-HN) and complete database with histopathological secured diagnosis of cAS-HN. Exclusion criteria were incomplete datasets concerning the assessed variables, patients suffering from multiple cancers of the head and neck, patients after other malignancies in the head and neck or after already performed cancer or lymph node surgery (e.g. neck dissection).
ACCEPTED MANUSCRIPT
Patients with cAS-HN arising in lymphoedema or after radiation were also excluded from the study. In summary, we assessed only patients with primary sporadical cAS-HN and without history or treatment of other head and neck cancers. All patients received complete appropriate staging, including ultrasound diagnostics of relevant lymph node levels and abdomen and also radiological imaging by either CT or MRI of head, neck, thorax and abdomen. All patients were initially treated surgically (resection of the primary tumour) with curative intention. In total 80 patients were
RI PT
eligible for inclusion in this study.
Methods
With a specially developed data entry form the characteristics of all cases were analysed.
SC
In all cases the following relevant epidemiological and clinical parameters were assessed: age at the time of first diagnosis (divided into two groups : 70 years), sex, primary tumour site in general (face or scalp)
M AN U
and more detailed (scalp, lower third of the face, midface including the nose, upper third of the face, ear and periauricular, tumour including more than one of the above mentioned regions of the face or scalp, tumour of more than one region in the face and the scalp), extent and dimension of the primary tumour (one region versus two or more regions ; 5 cm in diameter), initial locoregional or distant metastasis, resection status (R0=
TE D
histopathological tumour free (i.e.negative) resection margins, R1= microscopically positive margins, R2= macroscopically positive margins), minimal safety margin (>= 1cm versus < 1 cm), adjuvant radiation (yes / no), adjuvant chemotherapy (yes / no), comprehensive treatment protocol (surgery , surgery + radiotherapy, surgery + adjuvant chemotherapy, surgery + combined radio-chemo-therapy). Outcome and course of disease were analysed
EP
concerning tumour recurrence (local relapse, occurrence of distant metastasis), concerning overall survival status and
AC C
time and concerning disease specific survival status and time.
Histological examinations and re-evaluations were performed by two experts in dermatopathology (HJS, CH) independently and blinded from each other. Diagnosis was proven in formalin fixed and paraffin embedded representative biopsies and excisions using conventional H&E staining as well as a panel of immunohistological stainings with antibodies against different vascular markers (CD31, CD34, Podoplanin). In cases of adequately archived tumour material, the following histological parameters were analysed retrospectively: According to the FNCLCC system tumour differentiation (Score 1: sarcoma closely resembling normal adult mesenchymal tissue ,Score 2: sarcoma for which histologic typing is certain , Score 3: Embryonal and undifferentiated sarcoma, sarcomas of uncertain type), mitosis count (Score 1: 0-9/10 HPF (high power fields) , Score 2: 10-19/10 HPF ,
ACCEPTED MANUSCRIPT
Score 3: >= 20/10 HPF) and microscopic tumour necrosis (Score 0: no necrosis, Score 1: 50% tumour necrosis). These 3 Sub-Scores were summed up to build a total score value, from which the histological grade according to the FNCLCC system is deduced (Grade 1= total score 2 or 3, Grade 2 = total score 4 or 5 , Grade 3= total score 6 to 8). The infiltration depth (dermis, subcutis, fascia or deeper), percentage of solid areas (= 80%, Figure 7) and cell type (epitheloid versus spindled cells) were evaluated. A risk stratification according
RI PT
the proposals of Deyrup (Deyrup et al., 2008a) was derived from the parameters tumour necrosis and cell type: the low risk histologic group (LRHG) is defined by absence of necrosis and epitheloid cells, the high risk histologic group (HRHG) displays epitheloid morphology and / or necrosis.
SC
Statistical analysis
To assess significance of differences concerning disease specific survival (between patients having died due to their
M AN U
cAS-HN and patients without tumour dependent death at the end of their follow-up period) Chi-square test and Fisher´s exact test were applied for categorical variables, for metric parameters Kolmogorow-Smirnov (KS) test was used as non-parametric test in not normally distributed data and independent t-test was used in normally distributed variables. Overall survival was calculated according to the Kaplan-Meier method (patients being alive at the end of the
TE D
follow-up period were censored) and group differences were tested on significance using log-rank test (univariate analysis).
Binary logistic regression analysis (BLR, multivariate analysis) was used to model the predictors of disease specific survival including patient, tumour, and treatment characteristics. Potential predictors (p < 0.05 in univariate analysis)
EP
from the Chi-square-, KS- and independent t- test analyses were entered in a stepwise forward fashion using p0.1 for removal. The following parameters were considered to be eligible for BLR: extent of primary tumour, tumour size, resection status, locoregional metastasis, initial distant metastasis, secondary distant metastasis and tumour differentiation.
To analyze potential predictors for disease specific survival time in a multivariate analysis, Cox proportional hazards model (Cox regression analysis, CRA) was used. The following variables and parameters with significant p-value < 0.05 in univariate analysis (log-rank test) were eligible for CRA: age group, extent of primary tumour, resection status, locoregional metastasis, initial distant metastasis, secondary distant metastasis, adjuvant radiation, adjuvant chemotherapy, treatment protocol and solid areas. Again a stepwise forward procedure using p0.1 for removal was chosen.
ACCEPTED MANUSCRIPT
All statistical analyses were performed by a statistician using the Statistical Package for Social Sciences, version 16.0
RI PT
(SPSS Inc., Chicago, IL, USA).
Results
In total 80 patients (30 females, 50 males, mean age at first diagnosis 71.4 +/- 14.4 years) were included in this study.
SC
Mean follow-up time was 55.3 +/- 74.4 months. 39 patients died during their observation period (overall survival), 28 of them due to their cAS-HN (disease specific survival, DSS). Overall survival rates for 3 years, 5 years, 10 years and 20
M AN U
years were 71 %, 54 %, 38 % and 20 %, respectively. Mean overall survival time was 121.6 (95% confidence interval (CI) 82.5 – 160.8) months, median overall survival time was 64.0 (95% CI 48.4 – 75.6) months. Concerning DSS, survival rates for 3 years, 5 years, 10 years and 20 years were 74 %, 62 %, 55 % and 41 % and mean and median DSS time were 174.3 (95% CI 120.9 – 227.7) months and 174.0 (95% CI 30.6 – 317.4) months, respectively. Histological re-evaluation
TE D
of the parameters described was possible in 37 patients in total. Not all clinical and histological variables were available in all cases. The main reason why only 37 histological re-evaluations were possible was that an adequate tumour specimen was no longer available, in older cases the specimen was used up beforehand in previous diagnostic procedures and analyses. In a few cases, the quality of the specimen was of inadequate quality to allow histological
EP
evaluation of some parameters. Some missing information in other variables were caused by single cases with poor or
AC C
insufficient data quality concerning an assessed parameter - mainly due to the retrospective nature of this study. An overview over the study population and considered variables with regard to DSS status are shown in Table 1. There were significant differences (p < 0.05) between patients who died due to cAS-HN and patients who had not died from their tumour, related to the extent of the primary tumour, tumour size, resection status, locoregional metastasis, initial distant metastasis, secondary distant metastasis and tumour differentiation. Concerning DSS rates and time, data of clinical and histological parameters and their impact on DSS time (univariate analysis by log-rank test) is presented in Table 2. Possible predictors (p < 0.05) of DSS time were age group, extent of primary tumour, resection status, locoregional metastasis, initial distant metastasis, secondary distant metastasis, adjuvant radiation, adjuvant chemotherapy, treatment protocol and solid areas (Figure 1). Risk stratification according to Dreyrup et al. into LRHG and HRHG revealed no statistically significant difference concerning DSS time (p = 0.440).
ACCEPTED MANUSCRIPT For calculation of a multivariate prediction model concerning DSS status (dead versus alive), the significant variables of the univariate analysis (Table 1) were entered into BLR. Included variables and details of the model are shown in Table 3. Correct prediction of DSS status by this mathematical model was reached in 75.3 % of all cases and explanation of variance reached 42.6 % (Nagelkerke´s r-square = 0.426). Extension of the primary cAS-HN to more than one region,
RI PT
initial distant metastasis and incomplete resection (R1) increased the risk of tumour dependent death.
To model prediction of DSS time in a multivariate analysis, the significant variables of univariate DSS time analysis (Table 2) were included into CRA. Table 4 displays extracted possible predictors and their statistics. Whereas age
SC
above 70 years (hazard ratio (HR) 4.8, 95% CI 1.3 – 18.0) and incomplete tumour resection (HR 6.0, 95% CI 1.8 – 20.9) increase the risk for earlier tumour dependent death, absence of initial distant metastasis reduces this risk (HR 0.2,
TE D
Discussion
M AN U
95% CI 0.1 – 0.6).
Grading according to the AJCC and staging systems have been widely used to predict behaviour and prognosis of most
EP
soft tissue sarcomas, but AS are different. AS is considered a high grade lesion because the degree of differentiation
AC C
has not been shown to correlate with prognosis and clinical outcome. Epitheloid hemangioendothelioma is considered a low grade lesion, has a better prognosis and might have affected the results of recent studies in cAS (Deyrup et al., 2008b).
Average age of patients at first diagnosis in our study was 71.4 (SD 14.4) years. In comparison to patients in other studies this was slightly younger (Brenn and Fletcher, 2005), although we excluded AS in the trunk (breast) and extremities (Stewart and Treves, 1948). It may account to the currently improving screening of skin cancer, the awareness of rare cutaneous lesions and highly specialized medical settings for further therapy. 5-year disease specific survival (DSS) was 62%. 13 of 80 patients who had distant metastatic disease initially, mainly affecting the lung as first site of metastasis (Mark et al., 1996). DSS 5-year survival might be higher than in other studies due to consequent
ACCEPTED MANUSCRIPT
radiation therapy after surgical resection in 44 patients of 80. It might have improved outcome in locally advanced tumours and incompletely resected cAS compared to patients with smaller tumour load. Nevertheless surgery with adjuvant radiation is the treatment of choice in cAS in general today and might explain the higher survival rates among our patient collective.
RI PT
In our patient collective we observed that men were more likely to develop AS, but not four to five times more likely than women as described in previous studies (Wollina et al., 2013). We attribute this difference to the highly selective subgroup of cAS-HN we investigated. In our study patients with cAS in extremities and trunk were excluded because they are in contrast to cAS-HN, who are mostly patients with postradiation tumours or at least secondary tumours
SC
and the course of the disease and prognosis is proven to be different from that of sporadic AS (Holden et al., 1987). Moreover, one quarter of the patients in the study of Deyrup had surgery alone (Deyrup et al., 2008a). Our patients
M AN U
with locally advanced tumour growth underwent combination therapy by.
Previously tumour diameter and depth of growth was proven to be the strongest predictor for prognosis (AlboresSaavedra et al., 2011; Ogawa et al., 2012). Holden et al already described tumour size as the only clinical factor
TE D
predicting the course of the disease, survival was proven to be significantly better in patients with a tumour diameter less than 5cm (Holden et al., 1987). Patients with lesions greater than 5cm in diameter did not seem to show any histological markers that served for prognostication. In the presented study tumour size (also categorized into above or less than 5cm in diameter) did not show statistically significant differences in survival parameters, whereas
EP
involvement of more than one anatomical region was proven significant (p = 0.048). Definition of margins in this type
AC C
of tumour is difficult to assess due to characteristic skip lesions. Nevertheless AS crossing area boundaries (one region versus two or more) was superior to simple diameter concerning prognostic information. This could be applicable for future work comparing data of tumour extent.
Cases with complete excision by the first interventional step, R0, not depending on the extent of tumour, demonstrated a highly significantly better survival (p < 0.001). This result is in accordance with all previous studies in the literature. Although most patients presented in advanced stages in which no standardized regime was applicable, complete excision remains the therapeutic goal (Rouhani et al., 2008).
ACCEPTED MANUSCRIPT
Weaker anatomical boundaries and a longer sun exposure without proper protection and later diagnosis due to the hair coverage of the tumour suggested a higher occurrence of larger tumour size in the scalp area compared to the face, but a larger sample size did not show a significant difference in size as a prognostic feature. Age, site of occurrence and size, mitotic activity and pattern of growth were already described as significant histological markers for prediction of the clinical course (Perez et al., 2013; Yeang et al., 2013).
RI PT
Deyrup described in 2008 selected histologic features for predicting outcome in cAS in an entity that seem unpredictable in behaviour. In their study epitheloid morphology and necrosis appeared to be significant factors for segregation of the tumour into a high or low risk group (LRHG and HRHG according to FNCLCC). The high risk group
tumour invasion only had a higher risk for recurrence.
SC
had a significantly increased disease specific mortality in patients above 70 years (Deyrup et al., 2008a). The depth of
In our study mitotic activity, necrosis and the cell type did not serve for prognostication, but pattern of growth (solid
M AN U
areas greater than 80%) was correlated with better survival (p = 0.011). Nevertheless tumour differentiation and risk stratification according to Deyrup (Deyrup et al., 2008a) did not show significant predictive values in cAS-HN (p = 0.440) .
TE D
Some limitations of this study have to be mentioned. Retrospectiveness of this study has to be kept in mind with the well-known weaknesses of retrospective study designs concerning data quality and risk for selection bias. Due to the limited number of appropriate available histological samples, the trend in differentiation and risk stratification on
AC C
larger sample.
EP
prognostication of DSS could not be proven with statistical significance, but this could probably be confirmed in a
Conversely, the specific strength of this monocenter study lies in the restriction of cAS to the head and neck region with exclusion of secondary angiosarcomas (e.g. after radiotherapy or chronic lymphoedema) and considerable clinicopathologic data. To date no previous study analysed this well-defined, homogenous subgroup of cAS with an adequate number of included patients in this way.
Conclusion
ACCEPTED MANUSCRIPT In conclusion age below 70 years, low number of affected regions in the head and neck, initial absence of metastatic disease, complete initial resection and pattern of growth with more than 80% solid areas showed significantly lower mortality rates in cAS-HN. Our results give important information for clinical risk assessment of a tumour group in a
RI PT
specific setting and necessary screening considerations for the future in skin cancer.
Sponsorship
SC
There was no financial or material support for this study by any company
M AN U
Role of funding source None
Conflict of interest
TE D
There were no financial and personal relationships with other people or organizations that could inappropriately influence (bias) our work. The authors disclose any commercial associations, current and over the past 5 years, that
Acknowledgements
EP
might pose a conflict of interest. All authors declare that there is no conflict of interest.
AC C
We thank Ms. Margret Leygraf and Coworkers for their skilled technical assistance.
References
Albores-Saavedra J, Schwartz AM, Henson DE, Kostun L, Hart A, Angeles-Albores D, Chablé-Montero F : Cutaneous angiosarcoma. Analysis of 434 cases from the Surveillance, Epidemiology, and End Results Program, 19732007. Ann. Diagn. Pathol. 15, 93–97, 2011.
ACCEPTED MANUSCRIPT
Brenn T, Fletcher CDM : Radiation-associated cutaneous atypical vascular lesions and angiosarcoma: clinicopathologic analysis of 42 cases. Am. J. Surg. Pathol. 29, 983–996. 2005. Buehler D, Rice SR, Moody JS, Rush P, Hafez GR, Attia S, Longley BJ, Kozak KR : Angiosarcoma Outcomes and Prognostic Factors: A 25-Year Single Institution Experience. Am. J. Clin. Oncol. , 2013 Deyrup AT, McKenney JK, Tighiouart M, Folpe AL, Weiss SW : Sporadic cutaneous angiosarcomas: a proposal for risk
RI PT
stratification based on 69 cases. Am. J. Surg. Pathol. 32, 72–77, 2008a. Deyrup AT, Tighiouart M, Montag AG, Weiss SW : Epithelioid hemangioendothelioma of soft tissue: a proposal for risk stratification based on 49 cases. Am. J. Surg. Pathol. 32, 924–927, 2008b.
Guadagnolo BA, Zagars GK, Araujo D, Ravi V, Shellenberger TD, Sturgis EM : Outcomes after definitive treatment for
SC
cutaneous angiosarcoma of the face and scalp. Head Neck 33, 661–667, 2011.
Holden CA, Spittle MF, Jones EW : Angiosarcoma of the face and scalp, prognosis and treatment. Cancer 59, 1046–
M AN U
1057, 1987.
Jones EW : Malignant angioendothelioma of the skin. Br. J. Dermatol. 76, 21–39, 1964. Käcker C, Marx A, Mössinger K, Svehla F, Schneider U, Hogendoorn PCW, Nielsen OS, Küffer S, Sauer C, Fisher C, Hallermann C, Hartmann JT, Blay JY, Mechtersheimer G, Hohenberger P, Ströbel P : High frequency of MYC
TE D
gene amplification is a common feature of radiation-induced sarcomas. Further results from EORTC STBSG TL 01/01. Genes. Chromosomes Cancer 52, 93–98, 2013. Mark RJ, Poen JC, Tran LM, Fu YS, Juillard GF : Angiosarcoma. A report of 67 patients and a review of the literature. Cancer 77, 2400–2406, 1996.
EP
Mitelman F, Johansson B, Mertens : Mitelman Database of Chromosome Aberrations and Gene Fusions in Cancer
AC C
[WWW Document]. URL http://cgap.nci.nih.gov/Chromosomes/Mitelman (accessed 2.18.14), 2014. Morgan MB, Swann M, Somach S, Eng W, Smoller B : Cutaneous angiosarcoma: a case series with prognostic correlation. J. Am. Acad. Dermatol. 50, 867–874, 2004. Ogawa K, Takahashi K, Asato Y, Yamamoto Y, Taira K, Matori S, Iraha S, Yagi N, Yogi A, Haranaga S, Fujita J, Uezato H, Murayama S : Treatment and prognosis of angiosarcoma of the scalp and face: a retrospective analysis of 48 patients. Br. J. Radiol. 85, e1127–1133, 2012. Perez MC, Padhya TA, Messina JL, Jackson RS, Gonzalez RJ, Bui MM, Letson GD, Cruse CW, Lavey RS, Cheong D, Forster MR, Fulp WJ, Sondak VK, Zager JS : Cutaneous angiosarcoma: a single-institution experience. Ann. Surg. Oncol. 20, 3391–3397, 2013.
ACCEPTED MANUSCRIPT
Rouhani P, Fletcher CDM, Devesa SS, Toro JR : Cutaneous soft tissue sarcoma incidence patterns in the U.S. : an analysis of 12,114 cases. Cancer 113, 616–627, 2008. Stewart FW, Treves N : Lymphangiosarcoma in postmastectomy lymphoedema; a report of six cases in elephantiasis chirurgica. Cancer 1, 64–81, 1948. Wollina U, Koch A, Hansel G, Schönlebe J, Kittner T, Pabst F, Haroske G, Nowak A : A 10-year analysis of cutaneous
RI PT
mesenchymal tumours (sarcomas and related entities) in a skin cancer center. Int. J. Dermatol. 52, 1189– 1197, 2013.
Yeang MS, Tay K, Ong WS, Thiagarajan A, Tan DSW, Ha TC, Teo PTH, Soo KC, Tan HK, Iyer NG : Outcomes and
case-control study. Ann. Surg. Oncol. 20, 3066–3075, 2013.
SC
prognostic factors of post-irradiation and de novo sarcomas of the head and neck: a histologically matched
Young RJ, Brown NJ, Reed MW, Hughes D, Woll PJ : Angiosarcoma. Lancet Oncol. 11, 983–991, 2010.
AC C
EP
TE D
M AN U
Zahm SH, Fraumeni JF Jr : The epidemiology of soft tissue sarcoma. Semin. Oncol. 24, 504–514, 1997.
ACCEPTED MANUSCRIPT
Tables
Table 1 : Study population and considered variables with regard to DSS.
Variable
Disease specific survival (DSS) Alive
Dead
RI PT
n
Sign. (p-value)
Clinical Parameters
(0.072)
SC
Gender 50
29
21
Female
30
23
7
80
Age
M AN U
Male
70.0
73.7
SD [years]
16.4
9.5
TE D
Mean [years]
Age groups 70 years
(0.815)
AC C
Localisation of Primary Tumour Face
45
30
15
Scalp
35
22
13
(0.034)*
Extent of primary tumour One region
45
34
11
Two or more regions
35
18
17
Anatomical tumour site
(0.210)
(0.327)
Lower third of face
2
1
1
Midface incl. nose
27
18
9
Upper third of face
10
6
4
Ear / periauricular
7
7
0
> 1 region : face or scalp
9
7
2
> 1 region : face and scalp
3
2
1
RI PT
ACCEPTED MANUSCRIPT 22 11 11
Scalp
(0.018)*
Tumour size 36
18
18
> 5 cm
43
33
10
M AN U
Resection status ( pR)
SC
= 1 cm
Yes No
AC C
Locoregional metastasis
50
32
18
29
19
10
EP
Safety margins
TE D
R0
(0.545)
(0.001)**
27
11
16
48
38
10
( 70 years
(0.189)
(0.013)*
23
90
84
248.0
57
66
51
93.3
EP
1 region : face or scalp
9
80
48
59.3
> 1 region : face and scalp
3
60
38
36.0
36
64
61
> 5 cm
43
82
61
Resection status ( pR)
159.2
35
97
97
265.2
R1
43
53
28
69.9
AC C
>= 1 cm
79
66
173.4
29
64
56
98.4
(
