postoperative radiation boost does not improve local recurrence rates in extremity soft tissue...

RADIATION ONCOLOGY—ORIGINAL ARTICLE

Postoperative radiation boost does not improve localrecurrence rates in extremity soft tissue sarcomasVignesh K Alamanda,1 Yanna Song,2 Eric Shinohara,3 Herbert S Schwartz1 and Ginger E Holt1

Departments of 1Orthopaedics, 2Biostatistics and 3Radiation Oncology, Vanderbilt University Medical Center, Nashville, Tennessee, USA

VK Alamanda BS; Y Song MS; E ShinoharaMD; HS Schwartz MD; GE Holt MD.

CorrespondenceDr Ginger E Holt, Department of Orthopaedics

and Rehabilitation, Vanderbilt University

Medical Center, 1215 21st Avenue South,

Medical Center East, South Tower, Suite 4200,

Nashville, TN 37232-8774, USA.

Email: [email protected]

Conflict of interest: The authors have no

financial disclosures, conflicts of interest, or

funding sources to report.

Submitted 20 October 2013; accepted 29

March 2014.

doi:10.1111/1754-9485.12184

Abstract

Introduction: The standard of care for extremity soft tissue sarcomas continuesto be negative-margin limb salvage surgery. Radiotherapy is frequently usedas an adjunct to decrease local recurrence. No differences in survival havebeen found between preoperative and postoperative radiotherapy regimens.However, it is uncertain if the use of a postoperative boost in addition topreoperative radiotherapy reduces local recurrence rates.Methods: This retrospective review evaluated patients who received preop-erative radiotherapy (n = 49) and patients who received preoperative radio-therapy with a postoperative boost (n = 45). The primary endpoint analysedwas local recurrence, with distant metastasis and death due to sarcomaanalysed as secondary endpoints. Wilcoxon rank-sum test and either χ2 orFisher’s exact test were used to compare variables. Multivariable regressionanalyses were used to take into account potential confounders and identifyvariables that affected outcomes.Results: No differences in the proportion or rate of local recurrence, distantmetastasis or death due to sarcoma were observed between the two groups(P > 0.05). The two groups were similarly matched with respect to demo-graphics such as age, race and sex and tumour characteristics includingexcision status, tumour site, size, depth, grade, American Joint Committee onCancer stage, chemotherapy receipt and histological subtype (P > 0.05). Thepostoperative boost group had a larger proportion of patients with positivemicroscopic margins (62% vs 10%; P < 0.001).Conclusion: No differences in rates of local recurrence, distant metastasis ordeath due to sarcoma were found in patients who received both pre- andpostoperative radiotherapy when compared with those who received onlypreoperative radiotherapy.

Key words: local recurrence; soft tissue sarcoma; radiotherapy.

Introduction

Radiotherapy continues to serve as a cornerstone in thecare of patients with soft tissue sarcoma (STS).1 Coupledwith the evolution in limb salvage techniques, radio-therapy has produced improvements in local controlwhile providing patients with conservation of function.2,3

Studies have shown that radiotherapy alters a tumour’smicroenvironment and induces upregulation of severaltumour immune effectors.4 Typically, radiotherapy isadministered by external beam therapy either pre-operatively or postoperatively.5–7 Various studies haveshown that there are no significant differences in survival

between pre- and postoperative radiotherapy.7,8 In somecases, patients receiving preoperative radiation mightalso receive additional radiotherapy following surgerycalled a ‘boost’.9 Postoperative boosts are typicallyreserved for patients with close or positive microscopicmargins, such as those whose tumours might involveessential neurovascular bundles.10

Although radiotherapy is accepted as standard of carefor patients with STS, it is not without its downsides.Radiotherapy following surgery consists of a higher doseto a larger field and is also frequently associated withhigher rates of tissue oedema, fractures, and long-termfibrosis and joint stiffness.8,9,11 Preoperative radiotherapy

bs_bs_banner

Journal of Medical Imaging and Radiation Oncology 58 (2014) 633–640

© 2014 The Royal Australian and New Zealand College of Radiologists 633

mailto:[email protected]

is administered at a lower dose to a smaller field butalso has drawbacks, including higher rates of woundcomplications.11

In order to determine the necessity of a postoperativeboost in patients who receive preoperative radiotherapy,we aimed to evaluate if the addition of a postoperativeboost made a difference in outcomes – with the primaryendpoint being local recurrence and secondary endpointsbeing distant metastasis and disease-specific death.

Methods

Following approval from our institutional review board,we conducted a retrospective cohort study at a majorsarcoma center to examine differences in outcomesbetween patients receiving preoperative radiotherapyand those receiving both preoperative radiotherapyand a postoperative boost. Patients undergoing surgicalresection of soft tissue sarcoma at our institutionbetween January 2001 and December 2008 (totaln = 397) who either received preoperative radiotherapyonly (n = 49) or a combination of both pre- and post-operative radiotherapy (n = 45) were identified based ona retrospective review and were considered for thestudy. Patients were excluded if they were younger than18 years of age, if they lacked adequate medical recordsand if they had a tumour with good prognosis andborderline malignancy, such as dermatofibrosarcomaprotuberans.12

Patient demographics and tumour characteristics col-lected included age at time of surgery, sex and race.Tumour characteristics consisted of size, depth (super-ficial or deep to the fascia of the underlying muscle), site(upper or lower extremity), grade (low, intermediate orhigh) and histological subtype. Staging of the patientswas also carried out per the guidelines recommendedby the American Joint Committee on Cancer (AJCC 6thedition).13 Margins were recorded as either positive ornegative for each patient following definitive surgery. Apositive margin was defined as the presence of malig-nant cells at the inked margin. Additionally, the excisionstatus was also noted for each patient; that is, if apatient presented for resection of STS prior to excision,he or she was categorised as a primary excision; if apatient presented after an unplanned excision else-where, he or she was categorised as a secondary exci-sion. The mean radiation dose administered was alsonoted for both groups. Additionally, the receipt ofchemotherapy was also recorded. Chemotherapy wasadministered to the patient at the discretion of the mul-tidisciplinary oncology team consistent with currentstandards of care. Most instances of chemotherapy con-sisted of anthracycline-based regimens. Reasons for apostoperative boost were also noted for those whoreceived them.

Disease status, death due to STS, distant metastasisand local relapse were recorded. The time period from

the index operation to each outcome listed above wasmeasured in all patients.

Patient demographics, tumour characteristics andprognostic outcomes were compared across groupsusing Wilcoxon rank-sum tests for continuous variablesand χ2 or Fisher’s exact tests. Survival curves fordisease-specific survival and metastasis-free survivalwere calculated and presented using the Kaplan–Meiermethod.14 The primary end point of the study was des-ignated as local recurrence. Death was treated as acensored observation for patients who died from a causenot directly related to their STS. Gray’s test was calcu-lated and used to compare the disease-specific hazardsof death and distant metastasis between the two groups(preoperative radiotherapy only vs preoperative radio-therapy plus a postoperative boost).15,16 In addition tounivariate comparisons, multivariable regression analy-ses were used to take into account potential confound-ers.17 This hazard ratio model examined postoperativeboost as the main independent predictor of outcomes(local recurrence, distant metastasis and sarcoma-specific death) after controlling for tumour size, marginstatus and the use of chemotherapy.

The statistical software program R (version 1.11.1,www.r-project.org) was used for all data analysis.Reported P values were two-sided, and a P value ofless than 0.05 was considered to indicate statisticalsignificance.

Results

Analysis of the 94 patients treated for soft tissuesarcoma of the extremity at our center who met ourinclusion and exclusion criteria revealed that 49 patientsunderwent preoperative radiotherapy and the remaining45 underwent both pre- and postoperative radiotherapy.Both groups were similarly matched in terms of demo-graphics including age, sex and race (P > 0.05). In com-paring excision status, both groups were again similarlymatched. No significant differences in tumour character-istics were noted between the two groups with regard totumour site, size, depth, grade, AJCC stage or histologi-cal subtype (P > 0.05). Both groups had a high propor-tion of high-grade and deep tumours. The use ofchemotherapy was also not significantly differentbetween the two groups. As expected, the radiation dosewas different between the two groups, with the preop-erative radiotherapy group receiving a median dose of50 Gy and the pre- and postoperative radiotherapygroup receiving 64.8 Gy (P < 0.001). The postoperativeboost group had a significantly larger proportion ofpatients with a positive margin (P < 0.001). Twenty-eight of the 45 patients (62%) in the postoperative boostgroup had a positive margin, while only five of the 49patients (10%) in the preoperative radiotherapy-onlygroup had a positive margin. The mean follow-up timefor the group that received preoperative radiotherapy

VK Alamanda et al.

© 2014 The Royal Australian and New Zealand College of Radiologists634

http://www.r-project.org

only was 3.64 years; it was 4.12 years for the group thatreceived both pre- and postoperative radiotherapy(Table 1).

The 45 patients who received a postoperative boostwere analysed for the indications that prompted postop-erative radiotherapy. The majority of such cases weredue to positive margins (n = 29). These instances ofpositive margins were further subclassified based on thereason that prompted the positive margin. Fifteen of the29 instances were due to tumour encasement of essen-tial neurovascular bundles. Other reasons for a positivemargin included tumour involvement of multiple sur-rounding structures (n = 3), wound breakdown followingpreoperative radiotherapy or prior inappropriate excisionand subsequent contamination of the wound with malig-nant cells (n = 3), avoidance of radical surgery such asamputation for large tumours (n = 2) and palliation fol-lowing a debulking procedure (n = 1). Twelve instancesof close but negative margins were also reported asreasons for a postoperative boost. In such patients, themean distance between the inked margins and the pres-ence of tumour cells was 3.7 mm (interquartile range:1.0–5.25 mm). Other reasons for a postoperative boostincluded good response to preoperative radiation wherethe tumour had a greater than 80% necrosis at time ofresection (n = 4) and clinically aggressive disease such

as tumours that were large-sized and high-grade (n = 3)(Table 2).

Local recurrence

Of the patients receiving only preoperative radiotherapy,five out of the 49 patients (10%) developed local recur-rence of their STS. In the group that received thepostoperative boost, eight of the 45 patients (18%)developed local recurrence. This difference was notfound to be statistically significant (P = 0.29). Differ-ences in the cumulative incidence for local recurrence-free survival were also statistically insignificant betweenthe two groups (P = 0.672). By the end of 72 monthsfrom surgery, the estimated cumulative incidence forlocal recurrence-free survival was 84.5% for the preop-erative radiotherapy group and 74.0% for the pre- andpostoperative radiotherapy group (Fig. 1).

Distant metastasis

In analysing distant metastasis, 20 of the 49 patients(41%) in the preoperative radiotherapy group developedevidence of distant disease. In the group receiving bothpre- and postoperative radiotherapy, 19 of the 45patients (42%) developed distant metastasis. This

Fig. 1. Local recurrence curves for patients

receiving preoperative versus pre- and postop-

erative radiotherapy.

Postoperative boost for STS


Table 1. Patient demographics and tumour characteristics

Variable Preoperative

radiotherapy (N = 49)

Pre- and postoperative

radiotherapy (N = 45)

P value

Age (years), median (IQR) 59 (46–70) 54 (47–67) 0.34

Sex, n (%) 0.40

Male 32 (65%) 33 (73%)

Female 17 (35%) 12 (27%)

Race,n (%) 0.83

White 43 (88%) 39 (87%)

Non-white 4 (8%) 3 (7%)

Others 2 (4%) 3 (7%)

Excision status, n (%) 0.40

Primary excision 38 (78%) 38 (84%)

Secondary excision 11 (22%) 7 (16%)

Site, n (%) 1.00

Upper extremity 12 (24%) 11 (24%)

Lower extremity 37 (76%) 34 (76%)

Size, (cm), median (IQ Range) 12 (10–19) 16 (9–21) 0.30

Depth,n (%) 0.35

Superficial 2 (4%) 4 (9%)

Deep 46 (96%) 41 (91%)

Grade,n (%) 0.81

I 5 (10%) 3 (7%)

II 6 (12%) 5 (11%)

III 38 (78%) 37 (82%)

AJCC stage,n (%) 0.95

Stage I 4 (8%) 5 (11%)

Stage II 5 (10%) 4 (9%)

Stage III 18 (37%) 15 (33%)

Stage IV 22 (45%) 21 (47%)

Radiotherapy dose (Gy), median (IQR) 50 (50–50.2) 64.8 (50.4–70) <0.001

Chemotherapy,n (%) 0.084

Yes 18 (37%) 9 (20%)

No 31 (63%) 35 (80%)

Microscopic margins,n (%) <0.001

Negative 44 (90%) 17 (38%)

Positive 5 (10%) 28 (62%)

Histology type,n (%) 0.69

Malignant fibrous histiocytoma 25 (51%) 22 (49%)

Liposarcoma 11 (22%) 12 (27%)

Leiomyosarcoma 5 (10%) 2 (4%)

Malignant peripheral nerve sheath tumour 1 (2%) 1 (2%)

Rhabdomyosarcoma 1 (2%) 1 (2%)

Synovial sarcoma 1 (2%) 3 (7%)

Vascular sarcoma 0 (0%) 1 (2%)

Fibrosarcoma 0 (0%) 1 (2%)

Others 5 (10%) 2 (4%)

Survival status,n (%) 0.36

Alive 23 (47%) 25 (56%)

Dead due to sarcoma 20 (41%) 18 (40%)

Dead due to other causes 6 (12%) 2 (4%)

Distant metastasis,n (%) 0.89

Yes 20 (41%) 19 (42%)

No 29 (59%) 26 (58%)

Local recurrence,n (%) 0.29

Yes 5 (10%) 8 (18%)

No 44 (90%) 37 (82%)

AJCC, American Joint Committee on Cancer; IQR, interquartile range.

VK Alamanda et al.


difference was not statistically significant (P = 0.89). Inaddition, differences in the cumulative incidence ofdistant metastasis-free survival between the two groupswere also statistically insignificant (P = 0.68). By theend of 72 months from surgery, the estimated cumula-tive incidence for distant metastasis-free survival was60.4% for the preoperative radiotherapy group and54.1% for the pre- and postoperative radiotherapygroup (Fig. 2).

Sarcoma-specific survival

Of the patients receiving only preoperative radiotherapy,a total of 20 patients (41%) died due to their sarcoma.In the group receiving both pre- and postoperative radio-therapy, 18 of the 49 patients (40%) died due to theirsarcoma. This difference was not statistically signifi-cant (P = 0.36). In addition, differences in the cumula-tive incidence of disease-specific survival were also

Table 2. Indications for a postoperative boost

Reasons for postoperative boost among 45 patients* Frequency

Instances of positive microscopic margins 29

Tumour involvement of essential neurovascular bundle 15

Positive margins not otherwise specified 5

Tumour involvement of multiple surrounding structures 3

Wound contamination with tumour cells due to either preoperative radiotherapy

and/or prior incomplete excision

3

Avoiding the need for a radical surgery (i.e. amputation) for large-sized tumours 2

Palliation following debulking procedure 1

Close microscopic margins (mean 3.7 mm; IQR 1.0–5.25 mm) 12

Good response to preoperative radiation (i.e. > 80% tumour necrosis at time of resection) 4

Clinically aggressive tumour (i.e. high grade, large size) 3

*Some patients were classified in more than one category. IQR, interquartile range.

Fig. 2. Metastasis curves for patients receiving

preoperative versus pre- and postoperative

radiotherapy.



statistically insignificant (P = 0.75). By the end of 72months from surgery, the estimated cumulative inci-dence for sarcoma-free survival was 63.3% for thepreoperative radiotherapy group and 64.5% for the pre-and postoperative radiotherapy group (Fig. 3).

Multivariable regression analysis

Regression analysis found that after adjusting for tumoursize, use of chemotherapy and margin status, there wasno statistical evidence that the addition of the postop-erative boost to preoperative radiotherapy decreasedrates of local recurrence, distant metastasis or sarcoma-specific death. In analysing local recurrence, it was noted

that positive margins approached statistical significancein increasing rates of local recurrence by a factor of 3.91(P = 0.07). In analysing distant metastasis, it was foundthat every 1-cm increase in size increased the rate ofdistant metastasis by a factor of 1.47 (confidence inter-val 1.00, 2.17; P = 0.0489). Similarly, for every 1-cmincrease in size, the rate of sarcoma-specific deathincreased by a factor of 1.05 (confidence interval 1.01,1.09; P = 0.0240) (Table 3).

Discussion

Radiotherapy is frequently used as an adjunct in thetreatment of soft tissue sarcomas to achieve better local

Fig. 3. Sarcoma-specific survival curves for

patients receiving preoperative versus pre- and

postoperative radiotherapy.

Table 3. Multivariable regression analysis of local recurrence, distant metastasis and sarcoma-specific death

Variable Local recurrence Distant metastasis Sarcoma-specific death

P-value Hazard ratio (2.5%, 97.5%) P-value Hazard ratio (2.5%, 97.5%) P-value Hazard ratio (2.5%, 97.5%)

Size 0.4118 0.68 (0.28, 1.69) 0.0489 1.47 (1.00, 2.17) 0.0240 1.05 (1.01, 1.09)

Radiotherapy (pre- and post-op

vs pre-op only)

0.6548 0.73 (0.18, 2.94) 0.3243 0.65 (0.28, 1.52) 0.8700 0.94 (0.44, 2.01)

Chemotherapy (yes vs no) 0.4585 1.59 (0.47, 5.38) 0.1398 1.70 (0.84, 3.44) 0.2600 1.53 (0.73, 3.17)

Margin status (positive vs

negative)

0.0702 3.91 (0.89, 17.11) 0.2392 1.67 (0.71, 3.94) 0.7100 0.86 (0.38, 1.94)

VK Alamanda et al.


control than limb-sparing surgery alone can provide.3,18

By inhibiting cancer cell proliferation through a combi-nation of anti-inflammatory effects and promotion oftumour cell death,19 radiotherapy serves as an importantweapon in achieving control of the disease. Recent evi-dence continues to suggest the vital role that radio-therapy plays in modulating the immune responsethrough the upregulation of immune effectors and thedownregulation of immune suppressors.4

Radiotherapy is often administered postoperatively fol-lowing definitive surgery and preoperative radiotherapy.This postoperative boost aims to eradicate residualmalignant disease. As in both this study and others, apostoperative boost is frequently administered if positiveor close margins are reported following surgery or if thetumour involves essential neurovascular bundles.9,20,21 Inthis study, there were no differences in the rates of localrecurrence, distant metastasis and death due to sarcomabetween the group that received preoperative radio-therapy only and those who received a postoperativeboost in addition to the preoperative radiotherapy.

Additionally, multivariable regression analysis foundthat even after accounting for margin status followingdefinitive surgery, the addition of a postoperative boostfor patients receiving preoperative radiotherapy had aninsignificant effect on all three outcomes (local recur-rence, distant metastasis, sarcoma-specific death) meas-ured. Similarly, the addition of a postoperative boost didnot affect outcomes even when tumour size and thereceipt of chemotherapy were taken into account.

Positive margins continue to be the single greatestpredictor of local recurrence.22–24 However, positivemargins are not always avoidable. This is particularlyapplicable in cases where the tumour involves criticalstructures such as the neurovascular bundle. If negativemargins are not obtainable or if close margins arereported post-surgery, then the question arises of to whatlevel to accept a close margin and how best to manage thepatient postoperatively. In this group of patients, theconventional option would have been to administer addi-tional radiotherapy postoperatively or proceed with aradical surgery such as an amputation. However, if addi-tional radiation postoperatively results in no additionalbenefits with respect to local recurrence9 and amputationonly provides a marginal benefit in reducing local recur-rence and does not improve survival10 despite signifi-cantly decreasing the patient’s functional abilities, itseems that close monitoring with periodical imaging inthese high-risk patients is likely the best option aftersurgery.

Efforts to minimise the additional radiation exposurefrom postoperative boosts should be made if suchtherapy fails to result in significant improvements tosurvival and local recurrence rates. This is particularlyapplicable to a select group of patients for whom thepostoperative boost may contribute to additional mor-bidity in the immediate postoperative period. For

example, the use of chemotherapy has been to knownto complicate the wound healing process followingsurgery.25,26 When combined with postoperative radio-therapy, chemotherapy can further negatively affect thewound healing process after tumour resection. Thus, theimpetus to decrease the total dose of radiation is evengreater in such patients, for whom the additional radia-tion can complicate the wound healing process in theimmediate postoperative period while not affordingincreases in survival benefits.

Study limitations

This study primarily explores the role of postoperativeradiotherapy in addition to preoperative radiotherapy inaffecting outcomes at a single institution; it is inherentlynot representative of findings elsewhere. Additionally,due to the limited number of patients with positivemargins, a separate analysis following stratification ofpatients by microscopic tumour clearance (i.e. R0 or R1resection) could not be performed with adequate power.Thus, a beta error cannot be ruled out with certainty.However, as Al Yami et al.9 have demonstrated, theaddition of a postoperative boost in patients with R1resections has failed to provide an advantage in reducingrates of local recurrence. Furthermore, as evidenced bythe results from this study, multivariable analysis per-formed using the present population dataset shows thatthe addition of a postoperative boost fails to improverates of local recurrence, distant metastasis and deathdue to sarcoma even after accounting for tumour size,receipt of chemotherapy and margin status at time ofdefinitive resection. Despite the above-mentioned limi-tations, the relationships and conclusions postulatedremain highly plausible.

Conclusion

While radiotherapy is an important cornerstone in thetreatment of soft tissue sarcoma, its use does not comewithout harmful side-effects, ranging from poor woundhealing to long-term fibrosis and joint stiffness. Thus,efforts to balance the beneficial effects of radiotherapywith its negative side-effect profile must be judiciouslybalanced to optimise the patient’s treatment plans. Asevidenced in this study, the addition of postoperativeboost to preoperative radiotherapy does not affect ratesof local recurrence, distant metastasis and sarcoma-specific death. Thus, the use of a postoperative boostshould be reserved for a select group of patients, such asthose with gross residual tumour following definitiveresection, in order to optimally implement this double-edged sword.

References

1. Sadoski C, Suit HD, Rosenberg A et al. Preoperativeradiation, surgical margins, and local control of



extremity sarcomas of soft tissues. J Surg Oncol1993; 52: 223–30.

2. Jebsen NL, Trovik CS, Bauer HC et al. Radiotherapyto improve local control regardless of surgical marginand malignancy grade in extremity and trunk wallsoft tissue sarcoma: a Scandinavian sarcoma groupstudy. Int J Radiat Oncol Biol Phys 2008; 71:1196–203.

3. Wood WC, Suit HD, Mankin HJ et al. Radiation andconservative surgery in the treatment of soft tissuesarcoma. Am J Surg 1984; 147: 537–41.

4. Sharma A, Bode B, Moch H et al. Radiotherapyof human sarcoma promotes an intratumoralimmune effector signature. Clin Cancer Res 2013;19: 4843–53.

5. Rosenberg SA, Tepper J, Glatstein E et al. Thetreatment of soft-tissue sarcomas of the extremities:prospective randomized evaluations of (1)limb-sparing surgery plus radiation therapy comparedwith amputation and (2) the role of adjuvantchemotherapy. Ann Surg 1982; 196: 305–15.

6. Kepka L, Suit HD, Goldberg SI et al. Results ofradiation therapy performed after unplanned surgery(without re-excision) for soft tissue sarcomas. J SurgOncol 2005; 92: 39–45.

7. Suit HD, Mankin HJ, Wood WC et al. Preoperative,intraoperative, and postoperative radiation in thetreatment of primary soft tissue sarcoma. Cancer1985; 55: 2659–67.

8. O’Sullivan B, Davis AM, Turcotte R et al. Preoperativeversus postoperative radiotherapy in soft-tissuesarcoma of the limbs: a randomised trial. Lancet2002; 359: 2235–41.

9. Al Yami A, Griffin AM, Ferguson PC et al. Positivesurgical margins in soft tissue sarcoma treated withpreoperative radiation: is a postoperative boostnecessary? Int J Radiat Oncol Biol Phys 2010; 77:1191–7.

10. Alamanda VK, Crosby SN, Archer KR et al.Amputation for extremity soft tissue sarcoma doesnot increase overall survival: a retrospective cohortstudy. Eur J Surg Oncol 2012; 38: 1178–83.

11. Baldini EH, Lapidus MR, Wang Q et al. Predictors formajor wound complications following preoperativeradiotherapy and surgery for soft-tissue sarcoma ofthe extremities and trunk: importance of tumorproximity to skin surface. Ann Surg Oncol 2013; 20:1494–9.

12. Bowne WB, Antonescu CR, Leung DH et al.Dermatofibrosarcoma protuberans: a clinicopathologicanalysis of patients treated and followed at a singleinstitution. Cancer 2000; 88: 2711–20.

13. Edge SB. AJCC Cancer Staging Manual, 7th edn.Springer, New York & London, 2009.

14. Kaplan EL, Meier P. Nonparametric estimation fromincomplete observations. J Am Stat Assoc 1958; 53:457–81.

15. Gooley TA, Leisenring W, Crowley J et al. Estimationof failure probabilities in the presence of competingrisks: new representations of old estimators. StatMed 1999; 18: 695–706.

16. Gray RJ. A class of K-sample tests for comparing thecumulative incidence of a competing risk. Annals ofStatistics 1988; 16: 1141–54.

17. Fine JP, Gray RJ. A proportional hazards model forthe subdistribution of a competing risk. J Am StatAssoc 1999; 94: 496–509.

18. Wilson AN, Davis A, Bell RS et al. Local control ofsoft tissue sarcoma of the extremity: the experienceof a multidisciplinary sarcoma group with definitivesurgery and radiotherapy. Eur J Cancer 1994; 30A:746–51.

19. Rubner Y, Wunderlich R, Ruhle PF et al. Howdoes ionizing irradiation contribute to the inductionof anti-tumor immunity? Front Oncol 2012; 2:75.

20. Kim YB, Shin KH, Seong J et al. Clinical significanceof margin status in postoperative radiotherapy forextremity and truncal soft-tissue sarcoma. Int JRadiat Oncol Biol Phys 2008; 70: 139–44.

21. Kandel R, Coakley N, Werier J et al. Surgical marginsand handling of soft-tissue sarcoma in extremities: aclinical practice guideline. Curr Oncol 2013; 20:e247–54.

22. Alamanda VK, Crosby SN, Archer KR et al. Predictorsand clinical significance of local recurrence inextremity soft tissue sarcoma. Acta Oncol 2013; 52:793–802.

23. Cantin J, McNeer GP, Chu FC et al. The problem oflocal recurrence after treatment of soft tissuesarcoma. Ann Surg 1968; 168: 47–53.

24. Brennan MF. The enigma of local recurrence. AnnSurg Oncol 1997; 4: 1–12.

25. Chmell MJ, Schwartz HS. Analysis of variablesaffecting wound healing after musculoskeletalsarcoma resections. J Surg Oncol 1996; 61:185–9.

26. Edmonson JH, Petersen IA, Shives TC et al.Chemotherapy, irradiation, and surgery forfunction-preserving therapy of primary extremity softtissue sarcomas: initial treatment with ifosfamide,mitomycin, doxorubicin, and cisplatin plusgranulocyte macrophage-colony-stimulating factor.Cancer 2002; 94: 786–92.

VK Alamanda et al.


postoperative radiation boost does not improve local recurrence rates in extremity soft tissue...

Documents