Discussion and conclusions
The optimal management of retroperitoneal soft tissue sarcomas including the role of radiation therapy has been extensively debated in the past and remains still unclear. In contrast to extremity sarcoma, data from large randomized trials regarding the optimal treatment for retroperitoneal sarcoma is still lacking due to the rarity of this disease and even prospective phase I/II data is rare. Treatment recommendations are mainly based on retrospective single center series, usually covering small patients numbers treated over many years with various combinations of surgical approaches and radiation treatment modalities. Therefore, evidence gained in the much larger studies in extremity sarcoma patients is frequently transferred to guide treatment of retroperitoneal sarcomas. As margin status has been identified as an important prognostic factor in both extremity and retroperitoneal sarcomas [
7,
8] surgery with wide negative margins represents the cornerstone of curative intent approaches. However, the achievement of wide or even close negative margins is much more difficult in the retroperitoneal space than in the extremities, resulting in much higher local recurrence rates after surgery alone [
9]. Even with more aggressive surgical approaches using en bloc resections of adjacent uninvolved organs, local recurrence remains the dominant pattern of failure in retroperitoneal sarcomas [
10,
11]. Additional radiation therapy has clearly been shown to improve local control in extremity sarcomas irrespective of margin status with increasing benefits after close or positive margin resections [
7] and is therefore widely accepted as standard of care for these patients. As close or positive margins are more frequent in retroperitoneal sarcomas, this should theoretically lead to an even more pronounced benefit from additional radiation therapy, but only a minority of patients is currently treated with this combination approach [
12]. Postoperative external beam radiation therapy was the first RT modality, which has been investigated. But although local control seemed to be improved in many series compared to surgery alone [
11,
13,
14], concerns have been raised mainly by the difficulties in achieving adequate dose and target coverage in the postoperative setting. Delivery of efficient doses with generous margins as used in extremity sarcomas is often hampered by the presence of small bowel loops in the resection cavity which would result in unacceptable toxicity in many cases, while reduced doses or margins would compromise efficacy. Because of the known dose-effect relationship which favours doses ≥55 Gy [
15,
16], several institutions including ours investigated additional boosting techniques like intraoperative radiation therapy (IORT) or brachytherapy to overcome these limitations [
2,
3,
17‐
19]. The only randomized trial comparing different local treatment approaches in retroperitoneal sarcoma showed a clear benefit favouring a combination of postoperative EBRT (35–40 Gy) and IORT (20 Gy) compared to postoperative EBRT alone (50–55 Gy) in terms of local control (60% vs. 20%) and gastrointestinal toxicity, while neurological toxicity was markedly increased in the IORT arm [
3]. Several non-randomized single institution series have confirmed high rates of local control for this combination approach with acceptable neurological toxicities limiting the IORT dose to 15 Gy [
2,
17,
18,
20], but there still remains room for improvement regarding both local control and toxicity, especially if compared to extremity sarcomas. Preoperative radiation therapy offers several possible advantages compared to the postoperative approach. These include a possible sterilization of the operative field against seeding, a possible thickening of the often-present pseudocapsule easing resection and the avoidance of repopulation through treatment delays because of postoperative complications. However, the main advantage seems to be the more accurate target volume definition with the possibility of reduced safety margins and reduced toxicity especially to small bowel loops because of their displacement through the tumor itself. Furthermore, the improved oxygenation could increase radiosensitivity and lower the required dose as known from extremity sarcomas. Some of these advantages, especially regarding target coverage and reduction of dose to adjacent organs at risk, can be further exploited with the use of modern radiation techniques like IMRT, VMAT or Tomotherapy as shown in several planning studies [
4,
21‐
23], including the opportunity to reduce overall treatment time by an integrated boost concept.
In our present study, we therefore combined the theoretically advantageous techniques of preoperative intensity-modulated RT with an integrated boost concept, surgery and IORT with the aim to achieve maximal local control. Unfortunately, we faced the same difficulties as many other groups investigating this rare disease, namely poor accrual over a long time span. We therefore decided to perform this unplanned interim analysis to evaluate whether the initial aims of the study are still reasonably achievable. With a median follow-up of 33 months, we found encouraging local control and overall survival rates (estimated 5-year LC 72%, and 5-year OS 74%), especially given the unfavourable patient selection (median tumor volume 1146 ccm and surgery with negative margins possible in only 22% of the patients). Moreover, we found a change in the pattern of failure, with distant metastasis being more prevalent than local recurrences. However, two recurrences were observed shortly after 5 years of follow-up, indicating the need for longer follow up to draw definitive conclusions. Nevertheless, our preliminary results seem to compare favourably with other groups using similar approaches. Pawlik et al. [
1] reported a combined analysis of two prospective trials including 72 patients with high grade retroperitoneal sarcomas treated with preoperative radiation therapy and surgery. Preoperative radiation was completed in 89% of the patients and 57 proceeded to surgery. Gross total resection was achieved in 54 patients of whom 32 received an additional boost via IORT or brachytherapy. With a median follow-up of 40 months, they observed 2- and 5-year local control rates of 79% and 60% and a 5-year overall survival rate of 61% after gross total resection. Gronchi et al. [
24] reported a prospective trial of 83 patients, who were treated by preoperative radiation therapy combined with chemotherapy. Radiation therapy was completed as planned in 73 and an additional IORT boost was given in 14 patients. 79 patients (95%) underwent surgery. With a median f/u of 58 months, 5-year local control, distant control and overall survival rates were 63%, 74% and 59%, respectively. Smith et al. [
25] published the long term analysis of a prospective trial which investigated preoperative radiation in 40 patients combined with selectively applied postoperative brachytherapy. After a median follow-up of 106 months, 5-year overall survival was reported to be 70% with a crude local control rate of 68%. Tzeng et al. [
26] reported a small prospective trial investigating dose escalated preoperative IMRT with integrated boost in 16 patients. Gross total resection was achieved in 14 patients, resulting in a 2-year local control rate of 80% after a median f/u of 28 months. These results were also supported by several retrospective series which reported 5-year local control rates of 63-68% and 5-year overall survival rates of 64-72% with similar approaches [
27‐
29], see Table
7. In summary, preoperative radiation therapy followed by surgery with or without additional boost consistently reached high local control and overall survival rates, which seem to be superior to the results of surgery alone or combinations of surgery with postoperative radiation, although formal high level evidence is still lacking. Therefore two phase III trials (ACOSOG Z9031, EORTC 62092) have been designed to evaluate preoperative radiation. While the first trial has already been closed due to poor accrual [
27], the results of the ongoing EORTC trial are eagerly awaited and will hopefully clarify the role of preoperative radiation therapy.
Table 7
Series with preoperative radiation therapy
Pawlik1
| 2006 | Pro.,comb. | 72 | 40 | 89% | 79% | 44% | 60%* | 61%* |
Gronchi24
| 2014 | Pro. | 83 | 58 | 88% | 95% | 17% | 63% | 59% |
Smith25
| 2014 | Pro.,subgr. | 40 | 106 | 100% | 100% | 48% | 63% (cr) | 70% |
Tzeng26
| 2006 | Pro. | 16 | 28 | 100% | 88% | 0% (d) | 80% (2 yr) | n.s. |
McBride27
| 2013 | Retro. | 33 | 33 | 100% | 100% | 30% | 63% (3 yr) | 64% (3 yr) |
Sweeting28
| 2013 | Retro. | 18 | 43 | 94% | 100% | 100% | 64% | 72% |
Alford29
| 2012 | Retro. | 24 | 28 | 100% | 75% | 0% | 68% | 54% |
Present data | 2014 | Pro.,interim | 27 | 33 | 93% | 96% | 85% | 72% | 74% |
Besides oncological outcome, every additional therapy comes along with toxicity. Therefore, the possible benefits in terms of local control or overall survival have to be weighed against side effects. In our study, mild gastrointestinal and hematological toxicity was common during preoperative radiation therapy, but only 4 patients (15%) suffered from grade III acute toxicity, which seems quite acceptable given the large radiation fields (median tumor size 15 cm, median PTV ~ 2400 ccm). Jones et al. [
30] described GI/pelvic grade I toxicity in 34% and grade II in 47% with no grade III side effects in their prospective trial on preoperative radiation. Pisters et al. [
31] reported a dose escalation trial with simultaneously given doxorubicin and described high rates of acute gastrointestinal grade III-IV toxicity (18%) and hematological toxicity (27%) in patients treated at the 50.4 Gy dose level, although at least parts of the side effects might be attributable to chemotherapy. Gronchi et al. [
24] also reported higher grade III/IV toxicity rates, but again this trial included simultaneously applied chemotherapy and therefore is difficult to compare. Caudle et al. [
32] found acute toxicity in 43% of 14 patients treated with preoperative radiation. From the description, it can be estimated that the rate of severe acute side effects (grade 3 or higher) was 21%, although toxicity was not formally graded. Zlotecki et al. [
33] compared patients with preoperative and postoperative radiation and found significantly decreased severe acute side effects after preoperative radiation therapy (36% vs 80%).
Preoperative radiation did not compromise the general ability for surgery in our study, as all patients proceeded to surgery and all except one received gross total resection. However, we observed a considerable rate (33%) of severe postoperative complications, including 2 patients (7%) who finally died after multiple-interventions in the prolonged postoperative period (30 day postoperative mortality rate was 0%). Jones et al. [
30] reported severe postoperative complications in 41% of their patients treated with preoperative radiation and selectively applied brachytherapy. They further described a 30 day mortality rate of 2%, but two additional patients (4%) died during the following 18 months due to anastomotic leakage and duodenal perforation. Gronchi et al. [
24] observed 21% major postoperative complications after preoperative chemoradiation and Alford et al. [
29] found 44% severe postoperative complications after preoperative radiation without additional boosting techniques. These results raise the question, if preoperative radiation increases the postoperative complication rate. However, this question is difficult to answer, as data from prospective series using surgery only is rare. Strauss et al. [
34] described a 30 day mortality rate of 3% after surgery alone. Lewis et al. [
35] reported a 30 day mortality rate of 4% in a large series of patients treated with surgery only or additional radiation and chemotherapy. Bonvalot et al. reported a mortality rate of 3% with only 18% of their patients requiring an invasive therapeutic procedure and 12% re-operation rate in a series treated with so called “aggressive frontline surgery” [
36]. About one third of the patients had received preoperative radiation, but unfortunately the complication rate was not reported separately for the groups with or without radiation treatment. Zlotecki et al. [
33] compared pre- and postoperative radiation therapy and observed a significant difference in severe postoperative morbidity (20% vs 53%) favouring the preoperative approach. Finally, Bartlett et al. [
37] compared preoperative radiation with surgery alone in a large retrospective series and found no significant differences, neither in overall morbidity and mortality nor in specific side effects. Given the extended surgical approach with contiguous organ resection in our patients, the postoperative complication rate seems acceptable and at least not distinctly increased by preoperative radiation. However, regarding the differences in mortality rates covering different postoperative time spans and the wide range of “severe” postoperative complications reported in the (mainly retrospective) literature after surgery with or without radiation therapy, further clarification is needed by standardized reporting of toxicity data from prospective trials.
Severe late toxicity was uncommon in our study, as only 1 patient (6%) showed grade III late toxicity at one year and none of the patients at two years. Jones et al. [
30] reported a 10% severe late toxicity rate at 18 months and Smith et al. [
25] described severe late toxicity in 11% of the patients after 18 months in an updated report of the same trial with mature follow-up. Alford et al. [
29] observed late toxicities of any grade in 46% of their patients. Interestingly, they described neurological lower limb side effects in 21%, although no additional boost via IORT has been performed, indicating that at least parts of the neuropathy effects usually attributed to IORT treatments might be caused by other reasons. However, the low rate of neuropathy might been explained by the lower IORT doses used in our trial compared to many other series. Based on the early findings reported by Shaw et al., which described a significant association between increased IORT dose and neuropathy rate [
20], the IORT dose was limited to 10–12 Gy in our study protocol. Interestingly, Haddock et al. described a significant reduction of neuropathy with intraoperative doses of less than 12.5 Gy in a recently updated series of rectal cancer treated with IORT [
38], indicating that our dose constraint seemed reasonable.
Clearly our analysis has some limitations mainly because of its nature of an unplanned interim analysis due to slow accrual. This naturally results in a relatively small number of patients with short follow-up, thus limiting the ability to draw definitive conclusions. Nevertheless, it represents prospectively collected data for the combination of preoperative radiation therapy using IMRT followed by aggressive surgery with contiguous organ resection and intraoperative radiation therapy in this rare disease of retroperitoneal sarcoma and therefore adds valuable information to the small existing body of evidence for this combination approach.
In summary, combination of preoperative IMRT, surgery and IORT resulted in promising 5-year local control and overall survival rates with low rates of acute and late toxicity and acceptable postoperative complications. Long term follow-up seems mandatory given the observation of late recurrences. Accrual of patients will be continued with extended follow-up.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
FR drafted the manuscript, participated in data aquisition and statistical analysis and supervised preoperative and intraoperative radiation therapy. AU and IA participated in data acquisition and treated the patients surgically. GH and MU participated in data acquisition and intraoperative radiation treatment. LSE participated in preoperative radiation treatment, data acquisition, statistical analysis and preparation of the manuscript. PEH participated in conducting the study, preparing the manuscript and reviewed the manuscript critically. FWH assisted in study conception, preparation of the study protocol and participated in conducting the trial regarding the aspects of radiation physics. DSE planned the study, prepared the study protocol, conducted the correspondence with the legal authorities and participated in study conduction. AN assisted in planning of the study, preparation of the study protocol and participated in conducting the study. RK supervised intraoperative radiation therapy, and participated in data acquisition. GM participated in conducting the trial regarding the aspects of pathology. JD participated in study conception and design. MB participated in manuscript preparation, conduction of the study and reviewed the manuscript critically. All authors read and approved the final manuscript.