Localised thoracic sarcomas: Outcome improvement over time at a single institution

European Journal of Cancer (2013) 49, 2689– 2697

Available at www.sciencedirect.com

journal homepage: www.ejcancer.com

Localised thoracic sarcomas: Outcome improvement over time at a single institution Leonardo Duranti a,⇑, Alessandro Gronchi b, Silvia Stacchiotti c, Marco Fiore b, Paolo G. Casali c, Paola Collini d, Giuseppe Pelosi d, Carlotta Galeone e,f, Ugo Pastorino a a

Division of Thoracic Surgery, Fondazione IRCCS Istituto Nazionale dei Tumori, Via Venezian 1, Milan, Italy Sarcoma Service, Department of Surgery, Fondazione IRCCS Istituto Nazionale dei Tumori, Via Venezian 1, Milan, Italy c Adult Mesenchymal Tumor Medical Oncology Unit, Department of Cancer Medicine, Fondazione IRCCS Istituto Nazionale dei Tumori, Via Venezian 1, Milan, Italy d Department of Diagnostic Pathology and Laboratory Medicine, Fondazione IRCCS Istituto Nazionale dei Tumori, Via Venezian 1, Milan, Italy e Department of Epidemiology, Mario Negri Institute for Pharmacological Research, Milan, Italy f Department of Clinical Sciences and Community Health, University of Milan, Italy b

Available online 14 May 2013

KEYWORDS Sarcoma Thoracic soft tissue sarcoma Surgery Chemotherapy Radiotherapy Prognosis Survival

Abstract Purpose: To assess changes in survival over time in patients affected by thoracic soft tissue sarcomas treated at a single institution. Patients and Methods: Patients with localised adult-type deep thoracic soft tissue sarcoma surgically treated at our institution between 1980 and 2012 were retrospectively reviewed. Patients were categorised into two groups according to timing of their first operation, i.e. surgery done before or after 31st December 2001 (so called ‘early years’ and ‘recent years’ groups, respectively), since a more extended surgery was used in the second interval. Overall survival (OS) and crude cumulative incidence (CCI) of local recurrence (LR) and distant metastases (DM) were calculated for each time period. Results: Three-hundred-thirty-seven patients were identified. Median follow-up was 4.7 years. Tumour size and rate of critical site involvement were larger in ‘recent years’, while the distribution of all other tumour- and patient-related factors was identical in the two periods. Despite this, OS and CCI of LR were significantly better in ‘recent years’ as compared to ‘early’ ones, the 5-year OS increasing from 58% to 72% and the CCI of LR dropping from 22% to 11%. CCI of DM was equal in the two periods.

⇑ Corresponding author: Address: Division of Thoracic Surgery, Fondazione IRCCS Istituto Nazionale dei Tumori, Via Venezian 1, 20133 Milan, Italy. Tel.: +39 0223902367; fax: +39 0223902907. E-mail address: [email protected] (L. Duranti).

0959-8049/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ejca.2013.04.007

2690

L. Duranti et al. / European Journal of Cancer 49 (2013) 2689–2697

Conclusion: Reference institutions for sarcomas may have improved their outcome in the last years. Although biases of retrospective analyses need to be discounted, it is possible that optimal exploitation of a series of subtle improvements in sarcoma treatment may make a difference in results achievable today. Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction Thoracic soft tissue sarcomas (STS) are a heterogeneous group of rare tumours accounting for 0.3% of all adult malignancies.1 They can originate from different anatomical structures, i.e. chest wall, lung, pleura, trachea or bronchi, mediastinum, nerves, heart or vessels, and are made up by different histological subtypes.2 Surgery is the mainstay of therapy, but little is known about their specific prognostic factors and outcome. Published series are relatively few. In general, they are limited to specific sites of tumour origin3–10 and consist of limited numbers of patients recruited over long time spans. Advances in surgical techniques, especially in the field of vascular and chest wall reconstructions, as well as in preoperative imaging facilities, have recently allowed to obtain better resections even in most challenging presentations.11–14 Moreover, a multimodal approach has been increasingly used in the management of high-risk thoracic STS to improve tumour resectability and local control.15 We performed a retrospective analysis of thoracic STS patients surgically treated at our institution in a 30-year time span to investigate prognostic factors and the long-term outcome, focusing on major changes in survival over time. 2. Patients and methods We included all consecutive patients affected by localised thoracic STS surgically treated at the Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy, from January 1980 to March 2012. Patients presenting at our centre with second or further local recurrence, as well as with metastatic disease, were excluded. Bone sarcomas, heart and great vessels sarcoma, oesophageal sarcomas, breast sarcomas, and cases with STS arising from the skin or subcutaneous tissue were excluded. We also excluded small round cell sarcomas, and desmoid-type fibromatoses. Histopathological diagnosis was reviewed and reclassified according to the updated World Health Organisation (WHO) criteria.2 The French Fe´de´ration Nationale des Centres de Lutte Contre le Cancer (FNCLCC) grading system was applied to all cases.16 Resections were classified as macroscopically complete, in the absence of macroscopic tumour left behind or incomplete. All macroscopic complete resections were

further classified according to the closest surgical margin, which was microscopically categorised as positive (tumour within 1 mm from the inked surface) or negative (absence of tumour within 1 mm from the inked surface).17 The indication to radiation therapy was given by both the operating surgeon and the radiation oncologist, in principle when a higher risk of relapse was supposed to exist on clinical grounds. External beam radiation was used in all these cases, and doses ranged from 45 to 70 Gy (median 60 Gy). Chemotherapy was given at the discretion of the multidisciplinary institutional Sarcoma Board or as part of ongoing clinical trials. Anthracycline-based regimens were used, most often combined with ifosfamide. After surgery, all patients were regularly followed-up by contrast enhanced computed tomography (CT) scan or magnetic resonance imaging (MRI), in general every 4 months for the first 2 years, then every 6 months for the following 3 years, then yearly. In the last 10 years, we performed more extended surgical resections, as listed in Supplementary Table 1. Moreover, perioperative chemotherapy was more broadly used. We arbitrarily decided to split patients into two calendar groups, in order to explore whether these technical and strategic changes were associated to any improvement in the outcome over time. We compared patients operated in the last 10 years (from January 2002 on, i.e. ‘recent years’) with those operated before (from 1980 to December 2001, i.e. ‘early years’). This analysis was approved by the Institutional Ethics Committee. 2.1. Statistical analysis Continuous variables were presented as mean values ± standard deviation (SD) and median with inter-quartile range (IQR), and categorical variables as numbers and percentages, as listed in Table 1. Comparisons among groups for continuous variables were performed using a two-sided Student’s t-test, after checking that data were normally distributed, and twosided Wilcoxon’s rank-sum test was used otherwise, and for categorical variables using contingency table analysis with the Chi-square test. The primary end-points of the study were overall survival (OS), local recurrence (LR) and distant metastases (DM). For each end-point, the time to event

L. Duranti et al. / European Journal of Cancer 49 (2013) 2689–2697

2691

Table 1 Main patient and disease characteristics. Overall

p-Value*

Calendar period Early years

Recent years

No.

%

No.

%

No.

%

Total

337

–

183

–

154

–

–

Sex Female Male

134 203

39.8 60.2

73 110

39.9 60.1

61 93

39.6 60.4

0.96

Age (years) Mean (SD) Median (IQR)

50.8 (16.7) 52.9 (38.1–64.1)

Age (groups) <43 43–60 P61

112 115 110

Size (cm) Mean (SD) Median (IQR)

8.8 (5.2) 8.0 (5.0–11.0)

Site Soft tissues Chest wall Mediastinum Lung/pleura

236 52 32 17

70.0 15.4 9.5 5.0

133 30 11 9

72.7 16.4 6.0 4.9

103 22 21 8

66.9 14.3 13.6 5.2

0.12

Grading G1 G2 G3

75 56 206

22.3 16.6 61.1

37 33 113

20.2 18.0 61.8

38 23 93

24.7 14.9 60.4

0.53

Histological subtypes UPS Syno MPNST Lipo SFT Leio Other

134 46 40 38 33 31 15

39.8 13.6 11.9 11.3 9.8 9.2 4.4

68 22 25 18 17 20 13

37.2 12.0 13.7 9.8 9.3 10.9 7.1

66 24 15 20 16 11 2

42.9 15.6 9.7 13.0 10.4 7.1 1.3

0.09

Presentation Primary Recurrence

306 31

90.8 9.2

167 16

91.3 8.7

139 15

90.3 9.7

0.75

Margins R0 R1 R2

261 64 12

77.4 19.0 3.6

140 32 11

76.5 17.5 6.0

121 32 1

78.6 20.8 0.6

0.03

195 134

59.3 40.7

123 54

69.5 30.5

72 80

47.4 52.6

<0.001

160 169

48.6 51.4

90 87

50.8 49.2

70 82

46.0 54.0

0.39

49.2 (17.2) 51.9 (35.7–63.6) 33.2 34.1 32.6

66 59 58

52.7 (15.9) 54.5 (39.8–64.9) 36.1 32.2 31.7

7.8 (4.7) 7.0 (4.0–10.0)

46 56 52

0.12

29.9 36.4 33.8

9.9 (5.4) 8.0 (6.0–12.0)

0.47 <0.001

**

Chemotherapy No Yes

**

Radiotherapy No Yes

SD: standard deviation; IQR: inter-quartile range; G1: grade 1; G2; grade 2; G3: grade 3; UPS: undifferentiated pleomorphic sarcoma; Syno: synovial sarcoma; MPNST: malignant peripheral nerve sheath tumour; Lipo: liposarcoma; SFT: Solitary Fibrous Tumour; Leio: leiomyosarcoma. **The sum does not add up to the total because of some missing values. * Comparison between calendar periods, obtained using the chi-square or Student’s t-test or Wilcoxon rank-sum test, as appropriate.

occurrence was computed from the date of surgery to the date when the event was recorded, or was censored at the date of last follow-up assessment in event-free patients.

We tested the proportional hazard assumption by including time dependent effects in the model (i.e. a covariate for interaction of the predictor and the logarithm of survival time), and no violation was found.

2692

L. Duranti et al. / European Journal of Cancer 49 (2013) 2689–2697

Hazard ratios (HR) of all end-points considered and the corresponding 95% confidence intervals (CIs) were estimated using Cox proportional hazard models including terms for age and sex18, as reported in Table 2. In the multivariate models we also included the following variables: tumour size (quartiles), site (soft tissue, chest wall, pleura/lung, mediastinum), FNCLCC grading (G1, G2, G3), histotypes, phase (primary/recurrence) as covariates related to the tumours, and margins of resection (R0, R1, R2), chemotherapy (no/yes), radiotherapy (no/yes), calendar time (surgery before/after 31st December 2001) as covariates related to the treatment. Survival curves were estimated using the Kaplan–Meier method and were compared by the log-rank test.19 Crude cumulative incidence (CCI) curves for LR and DM, whichever occurred first (and which included death as the competing event), were calculated in a competitive risks framework and comparisons between curves were carried out by means of the Gray test.20,21 Concurrent LR and DM were computed as DM. All tests were two-sided and a p-value of less than 0.05 was taken as statistically significant. Statistical analyses were performed using SAS 9.1 (SAS Institute, Cary, NC) and the figures were obtained using STATA 11.0 (StataCorp LP, College Station, TX) statistical software. 3. Results From January 1980, 337 patients were identified, 183 were in the ‘early’ and 154 in the ‘recent years’. The median follow-up (FU) for the alive patients was 4.7 years (IQR: 1.3–10.2). 11.1 (IQR: 10.0–15.9) in the ‘early’ and 3.3 years (IQR: 1.1–8.9) in the ‘recent years’. Clinical characteristics of patients, overall and by time of surgery, are reported in Table 1 3.1. Overall survival At the time of the present analysis, 135 patients died, 114 affected by primary and 21 by recurrent tumour. The OS for primary versus locally relapsed tumour was 65% (95% CI: 59–71%) and 55% (95% CI: 48–61%) versus 37% (95% CI: 19–56%) and 25% (95% CI: 9–45%) at 5 and 10 years, respectively, (p < 0.001) (Fig. 1, panel A). Thirty-seven in 76 patients operated on with positive margins and 98 in 261 with negative ones died. Five and 10 year OS for patients operated with positive versus negative margins was 51% (95% CI: 34–65%) and 51% (95% CI: 34–65%) versus 69% (95% CI: 62–75%) and 56% (95% CI: 48–63%), respectively. OS of patients who underwent incomplete resection was 18% (95% CI: 3–44%) at 5 and not evaluable (NE) at 10 years, respectively (p < 0.001) (Fig. 1, panel B).

One-hundred-six in 183 patients operated in the ‘early years’ and 29 in 154 in the ‘recent years’ died. The OS for patients operated in the ‘early years’ and in the ‘recent years’ was 58% (95% CI: 50–65%) and 48% (95% CI: 40–55%) versus 72% (95% CI: 60–81%) and NE at 5 and 10 years, respectively (p = 0.03) (Fig. 1, panel C). Tumour size was the other significant variable for OS at multivariate analysis (Table 2). 3.2. Local recurrence Sixty-seven patients (20%) developed LR after surgery at our institution. Forty-four in 306 (14%) patients operated for a primary tumour, and 9 in 31 (29%) operated for a recurrent tumour, developed LR as their first event. The corresponding figures of CCI of LR at 5 years were 14% (95% CI: 11–19%) for patients operated for a primary tumour and 31% (95% CI: 14–49%) for those operated for a recurrent tumour (p = 0.02) (Fig. 2, panel A). Twelve in 64 (19%) patients operated on with positive margins and 39 in 261 (15%) with negative margins developed LR as primary event. Five year CCI of LR of patients operated with positive versus negative margins was 24% (95% CI: 12–37%) versus 14% (95% CI: 10–19%) (p = 0.04) (Fig. 2, panel B). Fifty-three in 183 (23%) patients operated in the ‘early years’ and 12 in 154 (8%) in the ‘recent years’ developed LR as their first event. The CCI of LR for patients operated in the ‘early’ and ‘recent years’ was 22% (95% CI: 16–28%) versus 11% (95% CI: 6–18%) at 5 years (p = 0.01) (Fig. 2, panel C). The other significant variable correlating with LR was radiation therapy (Table 2). 3.3. Distant metastases Eighty-four patients (25%) developed DM (39 to the lung, 20 to the liver, 10 to the bone, 15 to other sites including a combination of the above). Forty-seven in 291 (16%) patients operated for a primary tumour and 4 in 30 (13%) for a recurrent one developed DM as primary event. The corresponding figures of CCI of DM at 5 years were 17% (95% CI: 13–23%) for patients operated for a primary tumour and 16% (95% CI: 5–32%) for those affected by recurrent ones (p = 0.80) (Supplementary Fig. 1, panel A). Eleven in 60 (18%) patients operated on with positive margins and 38 in 249 (15%) with negative ones developed DM as primary event. Five year CCI of DM of patients operated with positive versus negative margins was 20% (95% CI: 11–33%) versus 17% (95% CI: 12–21%) (p = 0.58) (Supplementary Fig. 1, panel B). Twenty-five in 170 (15%) patients in the ‘early years’ and 26 in 151 (17%) in ‘recent years’ developed DM as

L. Duranti et al. / European Journal of Cancer 49 (2013) 2689–2697

2693

Table 2 Results from the Cox proportional hazard models on the three end-points analysed. OS

LR

DMs

HR§

95% CI

HR§

95% CI

HR§

95% CI

1 1.21

Ref. 0.82–1.77

1 1.51

Ref. 0.85–2.69

1 1.03

Ref. 0.55–1.93

1 1.09 1.56

Ref. 0.67–1.78 0.93–2.62

1 1.21 2.01

Ref. 0.56–2.63 0.92–4.40

1 0.73 0.51

Ref 0.34–1.52 0.23–1.14

Calendar period Early years Recent years

1 0.44

Ref. 0.28–0.70

1 0.40

Ref. 0.20–0.82

1 0.89

Ref. 0.45–1.76

Size (cm) First quartile (<5.0) Second quartile (5.0–7.9) Third quartile (8.0–10.9) Fourth quartile (P11.0)

1 3.67 3.86 6.68

Ref. 1.83–7.38 1.84–8.10 3.16–14.13

1 1.44 0.97 2.49

Ref. 0.61–3.39 0.35–2.71 0.96–6.42

1 1.23 0.72 0.63

Ref. 0.48–3.18 0.24–2.13 0.21–1.90

Site Soft tissues Chest wall Mediastinum Lung/pleura

1 0.89 1.91 1.63

Ref. 0.50–1.57 0.95–3.81 0.72–3.72

1 0.84 0.42 0.64

Ref. 0.31–2.25 0.10–1.80 0.16–2.51

1 0.71 0.42 0.68

Ref. 0.26–1.97 0.12–1.48 0.08–5.79

Grading G1 G2 G3

1 0.86 1.89

Ref. 0.35–2.10 0.87–4.13

1 1.81 2.29

Ref. 0.49–6.64 0.67–7.85

1 2.94 21.06

Ref. 0.25–34.41 2.73–170.6

Histological subtypes UPS Syno MPNST Lipo SFT Leio Other

1 1.77 1.22 0.99 0.66 1.47 1.21

Ref. 0.97–3.22 0.67–2.24 0.47–2.10 0.23–1.92 0.79–2.74 0.53–2.75

1 2.27 1.60 1.07 1.06 2.03 –

Ref. 0.86–5.95 0.66–3.86 0.35–3.28 0.20–5.76 0.77–5.36 –

1 1.21 0.62 2.13 1.27 – –

Ref. 0.51–2.86 0.20–1.92 0.67–6.79 0.54–2.98 – –

Presentation Primary Recurrence

1 2.24

Ref. 1.22–4.11

1 5.49

Ref. 1.96–15.35

1 1.50

Ref. 0.48–4.73

Margins R0 R1 R2

1 1.90 3.88

Ref. 1.19–3.03 1.83–8.23

1 3.09 –

Ref. 1.49–6.40 –

1 0.57 –

Ref. 0.23–1.38 –

Chemotherapy No Yes

1 1.39

Ref. 0.90–2.15

1 1.25

Ref. 0.59–2.65

1 1.30

Ref. 0.60–2.82

Radiotherapy No Yes

1 0.79

Ref. 0.53–1.17

1 0.36

Ref. 0.19–0.68

1 1.07

Ref. 0.60–2.03

Sex Male Female Age (years) <43 43–60 P61

Abbreviations: OS, overall survival; LR, local relapse; DMs, distant metastases; HR, hazard ratio; CI, confidence interval; G1: grade 1; G2; grade 2; G3: grade 3; UPS: undifferentiated pleomorphic sarcoma; Syno: synovial sarcoma; MPNST: malignant peripheral nerve sheath tumour; Lipo: liposarcoma; SFT: Solitary Fibrous Tumour; Leio: leiomyosarcoma. § Model adjusted for age, sex and mutually adjusted for all the covariates listed below.

their first event. The CCI of DM for patients operated in the ‘early’ and ‘recent years’ was 15% (95% CI: 10–21%) versus 20% (95% CI: 14–28%) at 5 years (p = 0.26) (Supplementary Fig. 1, panel C). The only significant variable for DM at the multivariable analysis was histological grade (Table 2).

3.4. Post-operative outcome The mean post-operative stay was 8.3 days (range 2– 119), 7 days (range 4–119) in ‘early years’ and 14.5 days (range 2–47) in ‘recent years’. There were no intraoperative deaths. The overall post-operative mortality

2694

L. Duranti et al. / European Journal of Cancer 49 (2013) 2689–2697

Fig. 1. Comparative overall survival between primary and local recurrences (Panel A), the margins of resections (Panel B) and calendar period of the first operation (Panel C).

was 0.6% (two patients died for pulmonary embolism), 0.5% in ‘early years’ and 0.6% in ‘recent’ ones. The overall post-operative morbidity was 9.9% (34 post-operative events, 19 after soft tissue resections and 15 after extended surgery), 9.57% in ‘early years’ and 10.3% in ‘recent’ ones. Complications after major surgical procedures are detailed in Supplementary Table 2.

4. Discussion In this series of 337 patients with a localised thoracic STS treated over a 30 year time span, overall survival at 10 years was 52%, with 57% patients being continuously disease-free at 10 years. Patients operated in the last 10 years (i.e. ‘recent years’) had a significantly better outcome as compared to those operated in the first 20 years (i.e. ‘early years’). Other significant prognosticators were the disease phase at presentation, tumour size and surgical marginal status. Radiation therapy

was associated to a better local control, with only a trend towards a better survival. This is a retrospective series, thus suffering from all biases they may entail. On the other hand, this is one of the largest published on localised thoracic STS, i.e. a rare subset of sarcoma patients. Patients of the last decade had a better outcome in spite of larger tumour sizes and a higher rate of involvement of critical sites, while the distribution of other main tumour- and patient-related factors was similar. In particular, there was no difference in the percentage of high-grade tumours, nor in tumour histologies. A Will Rogers phenomenon22 could be advocated to explain such an outcome improvement. In this case, however, we would expect to observe an overall improvement even in the CCI of DM, which did not occur. By and large, an absolute improvement of more than 10% in both LR and OS is of interest. The time point separating the two periods was chosen based on an institutional shift at that time towards more extended procedures, especially for

L. Duranti et al. / European Journal of Cancer 49 (2013) 2689–2697

2695

Fig. 2. Crude cumulative incidence of local recurrences between primary and local recurrences (Panel A), the margins of resections (Panel B) and calendar period (Panel C).

tumours located at difficult sites, thanks to the availability of better reconstructive skills (Fig. 3). Indeed, there was a more frequent use of reconstructive surgery in the last period (from <12% in the first periods to >36% of cases in the second). This was accompanied by a dramatic decrease in the number of incomplete resections carried out in the last decade. As a result, local control improved significantly and ultimately it also affected the outcome. In fact, lack of local control for sarcomas of critical sites may become a leading cause of death23,24 On the other side, no breakthrough in STS management has occurred in this time interval as to justify such a difference. Another major difference among the two periods was the use of chemotherapy. In the overall multivariable analysis this was not associated to a better outcome for any of the end-points analysed. We are aware that adjuvant, or neo-adjuvant, chemotherapy has not definitively proved its efficacy in localised STS and is not viewed as a standard. However, an updated meta-analysis of randomised trials25 showed a statistically significant advantage in terms of both relapse free survival and OS, which was confirmed even after the inclusion

of a recently published negative trial.26 In keeping with the findings of the meta-analysis, the contribution of chemotherapy may have added to local control. The benefit shown is similar to what suggested by the proposed comparison between the results of a recent trial on neoadiuvant chemotherapy27 and what foreseen by the currently available prognostic nomograms for STS.28,29 Therefore we cannot exclude that a broader use of chemotherapy may have improved the prognosis of some patients. Radiation therapy proved to be associated to a better local control on the multivariate analysis. This is in line with what was found in small but convincing randomised trials on extremity STS.30,31 The administration of RT was almost equivalent in the two periods. This is consistent with the hypothesis that better surgery and possibly the administration of adjuvant chemotherapy contributed to improve local control over time. Of note, morbidity in the recent years was not worse than in the early years, in spite of the increased number of more extended procedures. In fact, median post-operative hospital stay was longer in recent years, but no differences in complications rate or mortality were observed.

2696

L. Duranti et al. / European Journal of Cancer 49 (2013) 2689–2697

Fig. 3. Pre-operative computed tomography (CT) of recurrent low-grade sarcoma extending from the apex of the chest to the diaphragm (a), with posterior extra-thoracic growth at magnetic resonance imaging (MRI) (b); post-operative CT after en-bloc resection of the left lung, pericardium, diaphragm and chest wall, with rib-like prosthesis (c); posterior view of chest wall reconstruction by three-dimensional CT (d).

A wider use of plastic reconstructive surgery may have contributed to improve the post-operative outcome even after extended procedures. Notably, 46% of patients who underwent major surgical procedures are still alive and disease-free at a median of 2 years. The initial approach is therefore critical. In fact chance of getting cured is dramatically reduced by more than a half in patients developing LR. Referral of these patients to tertiary centres should be strongly encouraged to avoid the risk of suboptimal approaches and subsequent loco-regional failures and deaths.32 In conclusion optimising the surgical approach to thoracic STS resulted in a better local control and survival in the last 10 years. Possible contributors were the improvement of reconstructive skills and the administration of complementary therapies, in spite of an overall increase in tumour size within our most recent case mix. Further improvements at tertiary centres may come from the development of new therapies to complement the optimal surgical approach. While waiting for these therapies every effort should be made for referral of these patients to specialised centres, in order to offer to all of them the highest chances of cure.

Conflict of interest statement None declared. Acknowledgements We gratefully acknowledge Morelli Maria and Errigo Priscilla for their great contribution in research of clinical data and in assistance for publication. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/ 10.1016/j.ejca.2013.04.007. References 1. Jemal A, Siegel R, Xu J, Ward E. Cancer statistics, 2010. CA Cancer J Clin 2010;60:277–300. 2. Fletcher CDM, Bridge JA, Hogendoorn PCW, Mertens F, editors. WHO classification of tumours of soft tissue and bone. Lyon: IARC; 2013. 3. Macchiarini P, Ostertag H. Uncommon primary mediastinal tumours. Lancet Oncol 2004;5(2):107–18 [Review].

L. Duranti et al. / European Journal of Cancer 49 (2013) 2689–2697 4. Re´gnard JF, Icard P, Guibert L, de Montpreville VT, Magdeleinat P, Levasseur P. Prognostic factors and results after surgical treatment of primary sarcomas of the lung. Ann Thorac Surg 1999;68(1):227–31. 5. Burt M, Ihde JK, Hajdu SI, et al. Source: Department of Surgery, Memorial Sloan-Kettering Cancer Center, New York, NY 10021, USA. Primary sarcomas of the mediastinum: results of therapy. J Thorac Cardiovasc Surg 1998;115(3):671–80. 6. Kachroo P, Pak PS, Sandha HS, et al. Single-Institution, Multidisciplinary Experience with Surgical Resection of Primary Chest Wall Sarcomas. J Thorac Oncol. 2012;7(1):151–6. 7. Janssen JP, Mulder JJ, Wagenaar SS, Elbers HR, van den Bosch JM. Primary sarcoma of the lung: a clinical study with long-term follow-up. Ann Thorac Surg 1994;58(4):1151–5. 8. Salas S, Bui B, Stoeckle E, et al. Soft tissue sarcomas of the trunk wall (STS-TW): a study of 343 patients from the French Sarcoma Group (FSG) database. Ann Oncol 2009;20(6):1127–35 [Epub 2009 Jan 29]. 9. Mitchell JD. Solitary fibrous tumor of the pleura. Review. Semin Thorac Cardiovasc Surg 2003;15(3):305–9. 10. Wouters MW, van Geel AN, Nieuwenhuis L, et al. Outcome after surgical resections of recurrent chest wall sarcomas. J Clin Oncol 2008;26(31):5113–8. 11. Girotti P, Leo F, Bravi F, et al. The “rib-like” technique for surgical treatment of sternal tumors: lessons learned from 101 consecutive cases. Ann Thorac Surg 2011;92(4):1208–15 [discussion 1215–6]. 12. Leo F, Bellini R, Conti B, Delledonne V, Tavecchio L, Pastorino U. Superior vena cava resection in thoracic malignancies: does prosthetic replacement pose a higher risk? Eur J Cardiothorac Surg 2010;37(4):764–9. 13. Incarbone M, Pastorino U. Surgical treatment of chest wall tumors. World J Surg 2001;25(2):218–30 [Review]. 14. Mansour KA, Thourani VH, Losken A, et al. Chest wall resections and reconstruction: a 25-year experience. Ann Thorac Surg 2002;73:1720–6. 15. Gronchi A, Miceli R, Colombo C, et al. Primary extremity soft tissue sarcoma: outcome improvement over time at a single institution. Ann Oncol 2011;22(7):1675–81. 16. Trojani M, Contesso G, Coindre JM, et al. Soft tissue sarcomas of adults: study of pathological prognostic variables and definition of a histopathologic grading system. Int J Cancer 1984;33:37–42. 17. Gronchi A, Casali PG, Mariani L, et al. Status of surgical margins and prognosis in adult soft tissue sarcomas of the extremities: a series of patients treated at a single institution. J Clin Oncol 2005;23(1):96–104. 18. Cox DR. Regression models and life tables (with discussion). J Roy Stat Soc B 1972;34:187–220.

2697

19. Kaplan EL, Maier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958;53:457–81. 20. Marubini E, Valsecchi MG. Analysing survival data for clinical trials and observational studies. Chichester, United Kingdom: John Wiley & Sons Ltd; 1995. 21. Gray RJ. A class of K-sample tests for comparing the cumulative incidence of a competiting risk. Ann Stat 1988;16:1141–54. 22. Feinstein AR, Sosin DM, Wells CK. The Will Rogers phenomenon. Stage migration and new diagnostic techniques as a source of misleading statistics for survival in cancer. N Engl J Med 1985;312(25):1604–8. 23. Gronchi A, Lo Vullo S, Colombo C, et al. Extremity soft tissue sarcoma in a series of patients treated at a single institution: local control directly impacts survival. Ann Surg 2010;251(3):506–11. 24. Gronchi A, Miceli R, Colombo C, et al. Frontline extended surgery is associated with improved survival in retroperitoneal low-intermediate grade soft tissue sarcomas. Ann Oncol 2012;23(4):1067–73. 25. Pervaiz N, Colterjohn N, Farrokhiar F, et al. A systematic metaanalysis of randomized controlled trials for adjuvant chemotherapy for localized resectable soft tissue sarcoma. Cancer 2008;113(3):573–81. 26. Woll PJ, Reichardt P, Le Cesne A, et al. Adjuvant chemotherapy with doxorubicin, ifosfamide, and lenograstim for resected softtissue sarcoma (EORTC 62931): a multicentre randomised controlled trial. Lancet Oncol 2012;13:1045–54. 27. Gronchi A, Frustaci S, Mercuri M, et al. Short, full-dose adjuvant chemotherapy in high-risk adult soft tissue sarcomas: a randomized clinical trial from the Italian Sarcoma Group and the Spanish Sarcoma Group. J Clin Oncol 2012;30:850–6. 28. Kattan MW, Leung DH, Brennan MF. Postoperative nomogram for 12-year sarcoma specific death. J Clin Oncol 2002;20:791–6. 29. Mariani L, Miceli R, Kattan MW, et al. Validation and adaption of a nomogram for predicting the survival of patients with extremity soft tissue sarcoma using a three-grade system. Cancer 2005;103:402–8. 30. Pisters PWT, Harrison LB, Leung DHY, et al. Long-term results of a prospective randomized trial of adjuvant brachytherapy in soft tissue sarcoma. J Clin Oncol 1996;14:859–68. 31. Yang JC, Chang AE, Sindelar WF, et al. Randomized prospective study of the benefit of adjuvant radiation therapy in the treatment of soft tissue sarcomas of the extremity. J Clin Oncol 1998;16:197–203. 32. Derbel O, Cropet C, Meeus P, et al. Adhesion to clinical practice guidelines (CPG’S) and role on survival for soft tissue sarcoma patients. Analysis of a population based cohort from Rhone-Alpes region. Ann Oncol 2012;23(Suppl. 9):ix478.