Hyper-Fractionated Radiotherapy for Soft Tissue Sarcoma: Results of the Second Study of Hyper-Fractionated Radiotherapy

Purpose and Method. Hyper-fractionated radiotherapy for treatment of soft tissue sarcomas is designed to deliver a higher total dose of radiation without an increase in late normal tissue damage. In a previous study at the Royal Marsden Hospital, a total dose of 75 Gy using twice daily 1.25 Gy fractions resulted in a higher incidence of late damage than conventional radiotherapy using 2 Gy daily fractions treating to a total of 60 Gy. The current trial therefore used a lower dose per fraction of 1.2 Gy and lower total dose of 72 Gy, with 60 fractions given over a period of 6 weeks. Subjects. A total of 37 patients (22 males and 15 females) with a median age of 56 years (range 19–88 years) were treated. Results. Of eight patients treated pre-operatively, six showed a partial response and in two the tumour was static. The maximum acute toxicities were grade 1 in eight, grade 2 in 14 and grade 3 in 15 patients. Late toxicities of the skin were graded 1 in 10 and grade 2 in nine patients. Five patients complained of pain in the irradiated bone and soft tissues, which was of moderate severity (grade 2). Stiffness was graded 2 in three patients and severe (grade 3) in one.Three patients had moderate and one patient had severe lymphoedema following treatment. The 5-year recurrence-free survival probability of patients treated radically was 76%. Following excision of local recurrences the study group had a disease-free survival probability of 86% at 5 years. Discussion. The regime is well tolerated with comparable local control and late complication rates to standard daily fractionated therapy.The potential benefit of this regime needs to be defined in a prospective randomized trial.


Introduction
Radiotherapy is an essential com ponent of treatm ent for all soft tissue sarcom as (STS) except those w hich have b een w id ely excise d o r are o f low g rad e. Conservative surgery and post-operative radiotherapy is the preferred procedure for lim b and lim b-g irdle sarcom as, w ith 5-year local control rates of 80± 90% rep o r ted in d iffe r en t ser ies. 1± 4 P re-o p era tive rad io th erap y m ay b e u sed ( w ith o r w itho u t chemotherapy) in tum ours not initially am enable to lim b conser ving surger y and around a 30% improvem en t in the rate o f lim b conser vation has been reported. 5± 7 D espite best efforts at conser vative surgery and conventional radiotherapy, m ost studies report local recurrence rates of up to 20% at 5 years; late local failure can continue to occur beyond that period. 1,2 Tum o u r s o f h igh er g rad e a n d larg er siz e are particularly at risk of local failure. 2,8,9 Local recurrence is observed in about one ® fth of patients w ith m etastases, 2 although it is uncertain whether the risk of distant m etastases is increased following local recurrence. 2,9 Around 60% of local recurrences can be controlled with further surgery but am putation m ay be necessar y. 2,5 Studies aim ed at further im proving local control of extremity STS are hence of param ount im portance. It app ears that conven tion al rad iotherapy has reached the lim it of its perform ance and it is im perative that newer m odalities/techniques be tested in an attempt to im prove results. C onservative surger y and brachytherapy is repor ted successfu lly to con trol locally advanced sarcom as of the lim b with good preservation of function. 10 Better target precision and tum our oxygenation are postulated advantages with this approach but n o sig ni® cant im provem ent in overall results has been observed with brachytherapy over conventional external beam radiotherapy. 10,11 Im proved results are reported with the addition of in fu sion al chem o therap y to rad io therap y in the pre-operative setting, although at the expense of additional toxicity. 12± 14 T he initial prom ise of higher relative biological effectiveness (RB E) with neutron therapy in the treatm ent of these tumours has also been offset by a decreased therapeutic ratio through increased late dam age. 15± 17 T he possib le bene® t of radiotherapy dose escalatio n in th e treatm en t of S T S ha s n o t b een prospectively tested. M ost radiotherapy schedules lim it the total dose of post-operative radiotherapy to 60± 66 G y in view of the unacceptably high incidence of late com plications with doses above this level. It was in this setting that altered radiotherapy fractionation was evaluated in the post-operative treatm ent of extrem ity ST S.
T he technique of hyp er-frac tionated radiotherapy can be used to deliver a higher total dose of radiation without an increase in late norm al tissue dam age. 18 This approach is proven to be of bene® t in prim ary tum ours arising in the head and neck, bladder, lung and in gliom as 19,20 but there are few reports on its efficacy in the treatm ent of ST S. Trials of com bined chem otherapy and hyp er-frac tionated radiotherapy in paediatric Ewing's and rhabdomyosarcom a have show n prom ising results 21,22 and radiotherapy using m u ltip le d aily fractio n s w a s effe ctive in th e pre-operative setting in com bination with radiosensitizing agents. 23,24 These observations formed the basis of a pilot study of hyper-frac tionated radiotherapy in the treatment of extrem ity STS at the Royal M arsden Hospital. T he ® rst study of hyp er-frac tionation was designed to determine the feasibility of irradiating large volum es to a dose of 75 G y using tw ice daily 1.25 G y fractions w ith a m inim um interval of 6 hours between fractions. 25 Assum ing on a /b value of 1.36 for late responding tissu es, the expected late effects with hyper-fractionated regim e were equal to conventional irradiation using 2 Gy fractions and a total of 60 G y. However, it app eared that the late dam age caused by this dose greater than w ith conventional irradiation, suggestin g higher a /b valu es fo r late-resp ond in g tissues.
In order to reduce the incidence and severity of late com plications, we designed the current trial of hyper-frac tionated radiotherapy in extrem ity STS, assu m ing a a /b of 3 for late reaction. By using a lower dose per fraction of 1.2 G y and lower total dose of 72 Gy, with 60 fractions being given over a period of 6 weeks, it was expected that the late toxicity would be equivalent to conventional fractionation. Assum ing an a /b of 10 for tumour, there would be a 11% increase in the effective dose for tum our control.

Materials and m ethods
This study was conducted on patients with ST S of the extremity referred to the Sarcom a Unit of the Royal Marsden Hospital between 1990 and 1995. All patients were initially assessed by the m ultidisciplinary team com prising surgeon, radiotherapist and m edical oncologist. Initial investigations in cluded com puted tomography/m agnetic resonance im ag ing (C T/M RI) scans of the prim ary site and C T scan of the ch est. Patients referred w ith n o histolog ical con® rm ation of diagnosis underwent tru-cut biopsy of tum our. All histology specim ens were reviewed by the sam e pathologist and graded as high, interm ediate or low.
Surgical excision of the tumour was the preferred m ode of treatm ent whenever possible and patients who achieved wide excision of a low grade tumour were not offered po st-operative radiotherapy. A ll patients w ith interm ediate or high grade tum ours and those in whom tumour was not excised with wide m argins, received post-operative radiotherapy. T his included patients who had`intra-capsular' or marginal' resection of tum ours by the Enneking's classi® cation. 6 Pre-operative radiotherapy was given to patients whose tumours were not initially amenable to lim b-conserving surgery. Palliative radiotherapy was given for patients with poor perform ance status or docum ented m etastatic disease. N o patient received chemotherapy as part of their prim ary treatm ent.
A l l p a ti en ts r e q u i r in g r a d i o t h e r a p y w e r e c o n s id e r ed fo r tre a tm e n t u s in g th e h y p e rfra ction ated protocol. Presen ce of bowel or n eural tissue in the targ et vo lum e w as the o nly criterio n for in elig ibility, as the stu dy desig n did not con sider the a /b ratios and sen sitivities of these str ucture s w hich are d ifferent to soft-tissu e. D urin g this tim e period , the large m ajority of patients were treated conventionally. O n ly those w ho were able to atten d for tw ice daily radio therapy could be inclu ded in the stud y, w ith m ost bein g ad m itted to the w ard . W ritten infor m ed con sent was m an dator y.
Patients were im m obilized using a perspex cast and treatment was planned using a CT scan, for 5/6 M V photons in two phases. Phase I volum e included the whole m uscle com partment of the lim b, usually with opposing ® elds, angled if necessar y, with or without compensators. For phase II a m uch sm aller volum e encom passin g the initial extent of the primary tum our with a 2 cm m arg in was used. C are was taken to leave a corridor of norm al tissue un-irradiated in all patients and to spare the palm , sole, heel, Achilles tendon, toes and joints w henever possible. 26± 28 T he p r in ciples an d tech n iq u es o f th re e-d im en sio n al conform al radiotherapy have been described previously. 29 A phase I dose of 60 Gy followed by a phase II dose of 12 Gy in 1.2 G y fractions treating twice a day was given in 60 fractions over 6 weeks post-operatively. Alternatively, pre-operative radiotherapy was given to the phase I volum e to 60 G y over 5 weeks following which patients were considered for surgery; tum ours which were still not am enable to lim b-conser ving surger y were treated further using radiotherapy to a phase II volum e, to a total dose of 72 Gy. Palliative irradiation was also given to a total dose of 72 G y in two phases. There was always a m inim um gap of 6 h between the two daily fractions.
Patien ts w er e m o n ito red w eek ly d u r in g radiotherapy: skin erythem a, desquam ation, oedema and ulceration were recorded. Toxicity was graded according to the RTO G radiation m orbidity scoring criteria (Table 1).
Tum our response was assessed 4 weeks after the en d o f p h ase I treatm en t in p atien ts treated pre-operatively. Tum ours w ith less than 50% reduction in the product of two perpendicular dim ensions were assesse d as having`static disease' . A reduction in dim ensions of 50% or m ore but short of com plete response was designated a`partial response' . Patients with`static disease' were suitable for surgery if the tum our had become technically operable.
Patien ts w ere review ed at 3-m on thly in ter vals d ur in g the ® rst 2 year s an d at lon ger inter vals thereafter, for evaluating disease status and scoring late toxicity. RTO G/EO RT C late radiation morbidity scoring scheme was used to study the late effects on skin, sub-cutaneous tissue, bone and joint; the N CIC late lim b oed em a sc ale was u sed to score posttreatment oedem a ( Table 2). Follow -up CT /M RI scans of the lim b and chest were perform ed w henever clinically indicated.
Recurrence-fre e survival was calculated using the Kaplan± M eier product lim it m ethod. Prognostic variables for radiation reactions were studied using the Chi-squ ared test. (Table 3) Thirty-seven patients underwent treatm ent using the hyperfractionated regim e. T here were 22 m ales and 15 fem ales with a m edian age of 56 years (range 19± 88 years). T high was the com m onest site of presentation, in 15 patients. T he other comm on sites of tum our were lower leg (n=10) upper arm , and foot (n=3 each). Leiomyosarcom a (n=14) and m alignant ® brous histiocytom a (n=12) were the m ost frequent histological subtypes; in two patients the type was not speci® ed (NO S). In 10 patients the tumour was of interm ediate grade and in 27 patients high grade. Twenty-nine patients were treated post-operatively and eight pre-operatively. Post-operative radiotherapy was given for high grade tum ours in 23 patients and for close m arg ins of excision in six patients.

Patient characteristics
T hirty patients were treated with radical intent: 27 post-operatively an d thre e pre-operative ly. Seven patients were treated with palliative intent because of docum ented m etastatic disease but to the sam e dose. T h is in clu d ed ® ve pre-o perative an d tw o po stoperative radiotherapy.
Tum our response and local control (Table 4) O f the eigh t patients treated w ith pre -operative radiotherapy, six patients showed partial response with m ore than 50% reduction in tum or size and in two  Thirty-® ve patients com pleted the planned course of treatm ent without interruption. Treatment was stopped in two patients treated post-operatively at 58.8 and 64.8 G y due to break-dow n of the surgical ap. Patients in this study had a m edian follow up of 44 m onths (range 2± 114). T he 5-year recurrence-fre e sur vival probability in patients treated with radical intent using pre-or post-operative hyperfractionated rad io therap y, w as 76% . T h e 5-ye ar d ise ase -fr ee sur vival after surgical salvage was 86%.

Early reactions
T he m axim um acute toxicity was only grade 1 in eight patients and 14 patients had grade 2 toxicity. G rade 3 toxicity was recorded in 15 patients, but none developed ulceration or necrosis (grade 4 toxicity). N ine patients had breakdown of their grafts while on rad iotherapy not graded as grade 4 radiation toxicity. Breakdown of surgical scars/grafts could not be entirely attributed to radiotherapy treatm ent, and graft breakdown occurred before the onset of grade 2 or 3 toxicity in six patients. G raft breakdow n was com plicated by infection in three patients and, in the other six, the breakdown was m inor, healing with conservative m anagement.  F igure 1 shows the time in weeks at which toxicities developed. Acute reactions generally developed during the second week of treatment and m ore than 60% of grade 1 toxicities occurred by the end of week 3. G rade 2 toxicity peaked around the ® fth week of rad iotherapy an d co ntin ued to develop throughout the duration of treatment. Progression to grade 3 toxicity was com m on around week 6. Acute radiation reactions were the same within the phase I and II radiation ® elds and there was no increase in incidence or severity of acute toxicity within the phase II volum e.

Late reactions
Late radiation m orbidity was assessed in 32 patients. Assessm ent was not possible in three patients as they died of m etastatic disease before the developm ent of late changes. O ne patient u nderwent am putation fo llo w in g p ar tia l resp o n se to p re-o p erative radiotherapy and one patient was lost to follow-up.
L ate m orbidity was m ost com m only observed in the skin and sub-cutaneous tissues. Ten patients had only grade 1 m orbidity of the skin w ith pigm entation and nine patients grade 2 m orbidity w ith m oderate telangectasia. In the sub-cutaneous tissue, induration was recorded as slight (grade 1) in 11, m oderate (grade 2) in nine and severe (grade 3) in one patients.
F ive patients com plained of pain in the irradiated bone or soft tissues, which was of m oderate severity (grade 2). Five patients had m ild (grade 1) stiffn ess of the irradiated joint. Stiffness was graded 2 in three patients and severe (grade 3) in one. Lym phoedema was slight (grade 1) in 13 patients, m oderate (grade 2) in three and considerable (grade 3) in one.
Incidence of grade 2 or 3 late radiation m orbidity of the sub-cutaneous tissue, bone or joint were higher in patients treated with phase II ® eld areas greater than 250 cm 2 , althou gh this w as n ot statistically sig n i® ca n t (g iven th e sm all n u m b er o f even ts recorded). Figures 2-4 show an ulcerated STS treated w ith wide excision and hyperfractionated radiotherapy, with excellent results.

Discussion
In the p re vio u s stu d y o f hy p er-f ra ctio n ated radiotherapy at the Royal M arsden H ospital, a total dose of 75 Gy in twice daily 1.25 G y fractions was tested. Assum ing an a /b ratio for late dam age of 1.36, it was expected that this regim e would give equivalent late toxicity com pared to 60 Gy in 30 fractions. 26 T he late effects from that study were greater than w ith conven tion al frac tion atio n an d in ou r current study an 11% increase in the therapeutic ratio was expected for the sam e late effects, assu m ing a higher a /b ratio of 3 for late-responding tissues.
C om parison of results w ith that of our previous study show s a slightly reduced incidence in grade 2± 4 acute radiation m orbidity (43 vs 48% ).T he new group of patients have a m inim um follow-up period of 33 months, with a m edian of 44 m onths, enabling scoring of m ost late radiation reactions. A shorter median follow-up period in the previous study would suggest that the actual incidence of late toxicity could be even higher than reported. The incidence of grade 2 or 3 induration is 27% in this study contrasted with 53% in the previous study. An overall local control probab ility of 86% at 5 years in this study com pares F igure 1. Graph showing the week of onset of acute radiation tox icity. favo u ra b ly w ith r esu lts u sin g co n ve n tio n al rad io th era p y. 1,2,30 T h is su gges ts th at hyp erfractionated radiotherapy using tw ice daily 1.2 G y fractions is well tolerated with good local control rates and no increase in the incidence of late toxicity.
T here is som e evidence to suggest that soft tissue sa rco m as sh o w a d ose resp o n se to ir rad iatio n , although this has not been proven. In a group of patients with advanced STS treated with radiotherapy alone, Tepper reported better local control in patients receiving 64 Gy or m ore. 9 Sim ilar results were also reported by Slater et al. w ith better local control of tum our in patients receiving m ore that 65 Gy. 31 In a retrospective analysis by Tanabe and colleagues, a higher rate of local recurrence was seen in patients treated with pre-operative radiotherapy. 32  In o u r exp er ien ce, patien ts treated w ith pre-operative radiotherapy to doses over 60 Gy had up to 80% response rates. 34 The incidence of late m orbidity is greater with use of a higher total dose of radiotherapy and higher dose per fraction. In the series by Slater, m ajor com plications were increased in patients treated to 70 G y or m ore. 31 In a previous study perform ed at our centre, the degree of ® brosis was related to total dose of radiotherapy in patients treated for STS of the extremity. 35 Selch reported a signi® cantly higher rate of toxicity in patients treated pre-operatively w ith high d o se p er fractio n rad io th era py an d co n cu r ren t chemotherapy. 36 H ence a total dose of 60± 66 Gy in 2 G y fractions is seldom exceeded in the treatm ent of these tum ours, especially as the volum es treated are large and inevitably include a substantial am ount of adjacent norm al tissue. H yper-fractionated radiotherapy exploits the differences in fractionation response between tum our and norm al tissue, by virtue of their different a /b ratios. W here the tum our has an a /b ratio exceeding that of late responding tissue, this technique could be used to deliver a higher total dose of radiation without an increase in late norm al tissue dam age. 18 Although studies on experim ental tum ours suggests a low a /b ratio in sarcom as 37 hyper-fractionation has not been extensively tested in the clinical setting. T here are only few published trials o f hyp erfractionated radiotherapy in soft tissue sarcom as.
Tr ials o f co m b in ed ch em o th era p y a n d hyp erfractionated radiotherapy in paediatric Ew ing's and rh a b d o m yo sa rco m a h ave sh o w n p ro m is in g results. 38,39 In a study by M andell et al., alternating chemotherapy and hyper-fractionated radiotherapy were used in gross residual or m etastatic paediatric rhabdom yosarcom a. Patients were treated to a total dose of 54 Gy in 1.5 G y fractions, treating tw ice a day with lesser acute toxicity and fewer treatm ent in ter r u p tio n s co m p ar ed w ith co n cu r r en t chem otherapy and conventional radiotherapy. 38,40 D unst et al. reported the use of hyp er-frac tionated radiotherapy in com bination with chem otherapy in the treatment of paediatric Ewing's sarcom a. 39 W illet et al. com pared conventional radiotherapy to tw ice daily hyperfractionated radiotherapy (1.8± 2 G y fractions separated by 4 h) in ST S of borderline resectability. Signi® cantly better histological response was noted in the group treated using the hyp erfractionated regim e. T his suggested that hyperfractio n ated rad io th erap y m ay b e u sefu l in the conser vative surgical excision of STS of border-line resectability. 23 Hyper-frac tionated radiotherapy has been used w ith radio-sensitizing agents in the treatm ent of advanced ST S. In a group of 36 patients with unresectable ST S at different sites treated with hyper-frac tionated radiotherapy in 1.5 Gy twice daily fractions to a dose of 70± 75 G y in com bination with Iododeoxyuridine, G offm an et al. reported 60% local control, with m oderate toxicity. 41 It is im po r tan t to test th e effic acy of hyperfra ctio n ated rad io th e rap y p ro sp ectively w ith conventional fractionation. Accurate com pariso n of late tox icity b etw een co nven tio n al an d hyp erfractionated regim es is essential possibly to escalate the dose of radiation in the hyper-frac tionated arm . Historical controls are of lim ited use in this regard and radiobiological predictions require clinical validation.
We propose to conduct a multi-arm trial, comparing hyper-frac tionated with conventional radiotherapy in the post-operative treatment of ST S. Patients with macroscopic residue following initial surgery are random ized for re-excision, followed by conventional or hyp er -f ra ctio n ated ra d io th era p y. Pa tien ts w ith com plete excision or m icroscopic residual disease are random ized directly into the radiotherapy arm s. T his trial would help to assess the role of hyper-fractionated radiotherapy and surgical re-excision in local control of ST S.

Conclusion
It is fe a si b le to d el iv e r h y p e r -f ra ct io n a te d rad iotherapy to a total d ose of 72 G y usin g tw ice daily 1.2 G y fractions over 6 weeks in the treatm ent of ST S of the extrem ity. T he regim e is well tolerated w ith com parable local control rates and late m orbid ity com pared w ith stan dard irradiation. T he a d va n ta g e o f th is re g i m e o ve r c o n ve n ti o n a l rad iotherapy n eeds to be con® rm ed by a prospective rand om iz ed trial.