To compare which radiotherapy technique is better in retroperitoneal sarcoma (RPS) rapid Arc (RA) or 3D-Conformal Radiation Therapy (3D-CRT).
Our study was on 10 patients with RPS diagnosed and treated at king Faisal Specialist Hospital & Research Center, Riyadh, Saudi Arabia, planned for pre or postoperative radiation therapy with prescribed dose of 45Gy in 25 fractions. In both techniques, we looked at planning target volume (PTV) coverage, dose homogeneity and organs at risk dose (stomach, bowel, liver, kidneys and spinal cord).
The PTV coverage, liver and stomach doses were similar in both plans however; RA significantly had better dose conformity (0.8 vs. 0.4, p = 0.034), dose homogeneity (1.08 vs. 1.3, p = 0.026), less bowel volume (V45 140cc vs. 243cc, p = 0.03) and lower Spinal cord dose (61% vs. 80%, p = 0.043).
Both plans achieved similar target coverage and organs at risk sparing however; RA showed statistically significant better dose homogeneity, bowel sparing volume and lower spinal cord dose in treating RPS by pre or postoperative radiation therapy.
Retroperitoneal sarcoma, Pre or postoperative, Conformal, Rapid arc
Retroperitoneal sarcoma (RPS) accounts to nearly 15% of all soft tissue sarcoma cases. The most common types are liposarcoma followed by leiomyosarcoma . At time of presentation, RPS has a large size due to late diagnosis. The most common symptoms are vague abdominal pain, weight loss and anorexia. The liver and lungs may represent the most common sites of metastasis .
Surgery is the mainstay treatment of RPS, however achieving a gross total resection is critical [3-8]. The incidence of local recurrence is high, so radiation therapy has a potential important role for RPS. There is no answer to the better timing of radiation therapy either pre or postoperative. Pre and intraoperative radiation are done by Pawlik, et al.  and Gronchi, et al.  who added concurrent chemotherapy. Other studies applied the same regimen with encouraging RT results [11-13].
To compare which radiotherapy technique is better in RPS, RA or 3D-CRT.
Our study was on 10 patients with RPS diagnosed and treated at king Faisal Specialist Hospital & Research Center, Riyadh, Saudi Arabia, 5 planned for preoperative and another 5 for postoperative radiation therapy with prescribed dose of 45Gy in 25 fractions. Dose constraints are shown in Table 1.
Table 1: Dose constraints. View Table 1
Planning CT scan was 4D performed with our departmental scanner (Philips Medical Systems, Cleveland, OH); with a slice thickness of 2 mm. Patients were placed supine with fully abducted arms with scanning from the upper thoracic to mid sacral spines. Fusion with preoperative CT, PET/CT or MRI scans was done to aid the delineation of gross target volume (GTV). GTV was expanded by 1.5 cm to create the clinical target volume (CTV), however in postoperative cases we have included the surgical clips as part of CTV then editing of organs at risk from CTV which then expanded by 1 cm to create the planning target volume (PTV). In both techniques, we looked at planning target volume coverage, dose homogeneity and organs at risk dose (stomach, bowel, liver, kidney, spinal cord).
Eclipse treatment planning system (Varian Medical Systems, Inc., Palo Alto, CA) was used along with the analytical anisotropic algorithm (AAA, Version 11.031) dose calculation algorithm. The plans were created with mixed 6 and 10 MV using 3-4 anterior, posterior and/or oblique fields.
Eclipse treatment planning system was used with optimization using progressive resolution optimizer (PRO) Version11.031. All plans generated using True Beam linacs with 120 leaf millennium MLC and KV imaging, 2 arcs (full and/or partial), Arc mode, 6 and 10 MV. Arcs had the same isocenter at the center of the PTV.
Dose-volume histogram statistics, dose conformity and dose homogeneity were analyzed to compare treatment plans. Both homogeneity (HI) [14-16] and conformity indices (CI)  were evaluated and calculated.
The planning target volumes, organs at risk, HI and CI endpoints were analyzed using non-parametric Wilcoxon signed rank test due to small sample size (SPSS, V19, USA), a probability value of < 0.05 considered to be statistically significant (two tailed).
PTV coverage was achieved and comparable in both plans, (Table 2).
Table 2: Preoperative cases. View Table 2
The CI and HI of RA were better and statistically significant than 3D plan (CI 0.8 vs. 0.4, p = 0.034) and (HI 1.08 vs. 1.3, p = 0.026).
We analyzed the dose parameters (mean and maximum doses) of the liver, stomach and duodenum, kidneys, spinal cord and bowels (V45, mean and maximum doses) for all patients (Table 2). All parameters were comparable in both plans especially liver, stomach and duodenum, and bowel doses; however, RA has statistically significant less bowel volume (V45 140cc vs. 243cc, p = 0.03) and lower Spinal cord doses (mean 36% vs. 55% and maximum 61% vs. 80% with p value = 0.03 & 0.043 consequently) (Table 2, Table 3 and Table 4).
Table 3: Postoperative cases. View Table 3
Table 4: Statistical results. View Table 4
The role of radiation therapy using both external beam and intraoperative radiation techniques for higher dose escalation to target volume either in naïve or recurrent RPS is still controversial with some studies achieved encouraging results [11-13,17-19]. Preoperative radiotherapy is preferred over postoperative one due to displacement of organs at risk especially bowel by the tumor itself with also better target coverage [20-22]. In our study we tried to look at which technique is better RA or 3D-CRT in RPS.
Using the same concept Paumier, et al. , Koshy, et al.  and Bossi, et al.  have been compared between 3D-CRT and IMRT of RPS, the first one  was for postoperative while the latter [24,25] was in the preoperative setting.
Regarding to the target coverage, it was identical in all the previous studies including our study except Koshy, et al.  who noticed increase of V95 (98.6% vs. 95.3%), PTV maximum and minimum doses (6% & 22%, P = 0.011 & P = 0.055) with IMRT arm.
Regarding to dose homogeneity, Paumier, et al. , Koshy, et al.  and Bossi, et al.  showed that CI was better in IMRT arm, similar to our results (CI 0.8 vs. 0.4, p = 0.034, HI 1.08 vs. 1.3, p = 0.026).
Regarding to organs at risk, Paumier, et al.  reported reduced bowel V50 and V40 five- and twofold, respectively with IMRT as in our study (V45 was 140cc vs. 243cc, p = 0.03), while Koshy, et al.  noted the lower small bowel volume receiving > 30Gy (63.5 to 43.1%, P = 0.043) with IMRT which was the same as by Bossi, et al. .
Paumier, et al.  reported the mean contralateral kidney dose increased from 1.5 (3D-CRT) to 4-4.4 Gy with IMRT, contrary to Bossi, et al.  who noticed that IMRT allows better sparing of the ipsilateral and contralateral kidney as well as in Koshy, et al.  while our results showed relative sparing of both kidneys by RA which was statistically insignificant, however bigger volume of the contralateral kidney received more doses in the RA arm.
Both plans achieved similar target coverage and organs at risk sparing however; RA showed statistically significant better dose homogeneity, dose conformity, bowel sparing volume and lower spinal cord dose in treating RPS by pre or postoperative radiation therapy.
All authors declare that there is no conflict of interest. For retrospective review of data with less than minimal risk to the patients, no consent was required by the ethics committee.