Maximizing rectal dose sparing with hydrogel: A retrospective planning study

Abstract External beam radiation therapy for prostate cancer can result in urinary, sexual, and rectal side effects, often impairing quality of life. A polyethylene glycol‐based product, SpaceOAR© hydrogel (SOH), implanted into the connective tissue between the prostate gland and rectum can significantly reduce the dose received by the rectum and hence risk of rectal toxicity. The optimal way to manage the hydrogel and rectal structures for plan optimization is therefore of interest. In 13 patients, computerized tomography (CT) scans were taken pre‐ and post‐SpaceOAR© implant. A prescription of 60 Gy in 20 fractions was planned on both scans. Six treatment plans were produced per anonymized dataset using either a structure of rectum plus the hydrogel, termed composite rectum wall (CRW), or rectal wall (RW) as an inverse optimization structure and intensity modulated radiotherapy (IMRT) or volumetric modulated arc therapy (VMAT) as a treatment technique. Dose‐volume histogram metrics were compared between plans to determine which optimization structure and treatment technique offered the maximum rectal dose sparing. RW structures offered a statistically significant decrease in rectal dose over CRW structures, whereas the treatment technique (IMRT vs VMAT) did not significantly affect the rectal dose. There was improvement seen in bladder and penile bulb dose when VMAT was used as a treatment technique. Overall, treatment plans using the RW optimization structure offered the lowest rectal dose while VMAT treatment technique offered the lowest bladder and penile bulb dose.

the proximity of the rectum, bladder, and penile bulb/neurovascular bundles to the prostate. 2 The rectum is the dose-limiting organ in prostate cancer external beam irradiation due to its proximity to the prostate, with the anterior rectal wall often falling within the planning target volume. [3][4][5][6][7] In recent years a number of products have been developed to spare the rectum during radiotherapy. One such innovation is SpaceOAR© hydrogel (SOH), a polyethylene glycolbased product, that is injected between the rectum and the prostate to displace the prostate away from the rectum. The physical shift of the rectum allows a greater proportion of the organ to be spared high dose and, in a randomized trial, has resulted in reduced rectal toxicity and improved quality of life (QOL). [8][9][10][11][12][13] The SOH has been shown to reduce the rectal dose in patients receiving both volumetric modulated arc therapy (VMAT) 11,14,15 and intensity modulated radiotherapy (IMRT). 8,[10][11][12]16 Studies have compared VMAT and IMRT treatment techniques in external beam prostate cancer treatment, indicating similar results for prostate coverage. In many studies, the dose to organs at risk (OAR), including the rectum, bladder, and penile bulb, was decreased when using VMAT over IMRT. 17,18 However, one planning study shows an exception in which rectal dose was lower with application of IMRT 19 compared to VMAT. The insertion of SOH between the rectum and the prostate may alter dose between treatment techniques. Current SOH studies are split between treatment techniques. The generation of IMRT or VMAT plans involves an inverse planning optimization, through a series of dosimetric constraints on anatomical structures and regions within a set of radiotherapy planning computed tomography (CT) scans. In previous studies, different definitions of the rectal avoidance structure have been used during optimization to minimize the true rectal organ dose. To date, the rectum avoidance structure, which can be defined as either a solid form or a wall (i.e., rectal wall thickness of 3 mm, excluding the lumen) organ delineated from the anus or bottom of the ischial tuberosities to the rectosigmoid junction, has been commonly employed during optimization in SOH studies. 10,11,14,15 More recently, a fabricated structure, the composite rectal (CR), has been proposed. 16 This structure can be generated by combining the rectum with the hydrogel before extracting a wall structure (i.e., rectum + hydrogel, thickness of 3 mm excluding the lumen). The hydrogel is difficult to contour due to low contrast between rectum and SOH on CT scans, therefore the CR structure may offer a simpler alternative. Additionally, it has been suggested by te Velde et al. 16 that the CR may serve as an alternative rectal organ optimization structure. Optimization with each of these structures offers a varying degree of rectal dose reduction.
The aims of the present study were to firstly evaluate the ideal optimization structure rectal wall vs composite rectum wall (RW vs CRW) in the setting of SOH for hypofractionated EBRT and secondly, to test whether the VMAT technique offers additional rectal sparing compared to IMRT. In this regard, IMRT and VMAT treatment plans for 60 Gy in 20 fractions were generated using anonymized CT datasets from patients with implanted SOH, using RW and CRW in the optimization, and organ at risk (OAR) doses were compared. The treatment plans were examined to determine which combination of optimization structure (RW or CRW) and treatment technique (VMAT or IMRT) resulted in the lowest rectal dose distribution.

2.A | Hydrogel implant
The anonymized CT datasets of thirteen prostate cancer patients who were implanted with 10 cc of SpaceOAR hydrogel between the prostate and the rectum were selected for this institutional research ethics board approved retrospective planning study. All patients receiving the SOH also had three to four gold fiducial markers implanted via a trans-perineal technique prior to gel placement. The CT datasets consisted of a pre-SOH and post-SOH planning CT scans for each patient. The pre-SOH planning CT scan was obtained with a comfortably full bladder and empty rectum 30 to 60 min prior to implantation of fiducial markers and SOH. Patients were given specific instructions to drink 750 ml of water within 15 min, 1 h prior to their pre-SOH CT scan and to perform a micro-enema 2 to 3 h prior to their appointment. One week later patients underwent a post-SOH planning CT scan as well as a pelvic MRI with the same bladder and bowel preparation instructions. The MR images were registered to the post-SOH planning CT images, using fusion to the gold fiducial markers, and used to assist with contouring the SOH, rectum, and prostate gland.

2.B | Structure of interest contours
A set of target and OARs for optimization and plan evaluation purposes were defined and peer-reviewed by a group of genitourinary radiation oncologists. Clinical target volume (CTV) was defined as the prostate gland and proximal 1 cm of seminal vesicles. The planning target volume (PTV) was defined as the CTV with margins of 7 mm in all directions except for a 5 mm margin in the posterior direction. Rectum was contoured as a solid organ from the rectosigmoid junction to the ischial tuberosities, and the cranial-caudal length was kept consistent from pre-to post-SOH. Composite rectum (CR) was defined as hydrogel plus rectum with manual editing to smooth jagged contours. RW and CRW structures were extracted using an inner wall margin of 3 mm (Fig. 1). Bladder was contoured with a bladder wall (BW) extracted using an inner wall margin of 3 mm. Femoral heads were contoured separately from the top of femoral head to the lesser trochanter. The penile bulb was contoured as the bulbous spongiosum below the GU diaphragm and proximal to the penile shaft.  Table 1. Most IMRT plans were created using five angle beam arrangements (0°, 50°, 100°, 260°, and 310°). Two additional beam angles (155°and 205°) were added to plans when hot spots in the subcutaneous tissues exceeded planning guidelines with a five beam arrangement.

2.C | Treatment plans
A plan optimization was deemed successful when the objectives listed in Table 2 were met using a plan normalization adjustment of less than ±0.5% following final dose calculation with AAA. Treatment plans which meet all OAR clinical objectives were difficult to produce with no plan normalization adjustment. In a clinical setting, plan normalization may vary up to 5% to meet required goals.
Restricting the plan normalization to ±0.5% limited its impact on the treatment plan comparison and ensured the correct balance between target coverage and OAR was achieved mainly during the optimization stage. The small adjustment to plan normalization limited the effect of plan normalization on treatment plans, creating a more difficult task for the planner to achieve the clinical goals set in Table 2. The clinical objectives, which are routinely utilized at BC Cancer -Victoria for prostate 60 Gy hypofractionated radiotherapy, were adapted from the PROFIT 20 and the CHHiP 21 study dosimetric objectives. The structures RW17.5 and BW17.5 in Table 2 were used solely to evaluate the quality of the plans and

2.D | Statistical analysis
The statistical testing was done using the nonparametric Wilcoxon signed rank test to compare different plan types as well as observe the change in volume between the pre-and post-SOH CT scans.
The tests were two-sided and considered significant at P < 0.01.

| RESULTS
Volume statistics for structures in the pre-and post-SOH CT datasets are summarized in Table 3. The majority of structures showed no significant difference between the CT scans with the exception of the PTV which showed a significant difference between pre-and post-SOH volumes (P = 0.006). The composite rectum volume was also statistically different from the summed individual rectum and SOH volumes (P = 0.001) due to smoothing of the edges of the composite rectum structure.
Comparisons between all six treatment plans (two pre-and four post-SOH plans) are shown in Table 4 while

| DISCUSSION
SOH has been incorporated into the radiotherapeutic management of prostate cancer in numerous cancer centers as a result of the proven benefits in reducing rectal toxicity and improving QOL. 9 The PROFIT protocol has been adopted clinically by many Canadian centers. This protocol evaluates rectal doses using a RW structure as opposed to the whole organ. As such RW endpoints depicted in Table 2 were used in the present study as planning goals for the rectum. ing were achieved. Gain in rectal sparing of 25% was considered clinically relevant as this gain was seen in RV70 Gy between use of three dimensional conformal radiotherapy (3D-CRT) and IMRT for prostate cancer treatment with RV70 Gy being linked to high rectal toxicity. 24 (Fig. 2). This trend is The P-values from plan to plan comparisons of the rectal metrics.  The post-SOH CTV volumes were found to be 4.7% larger compared to pre-SOH volumes after an average per patient ratio. The small, but almost statistically significant, difference in the CTV median volume was likely due to a combination of prostate edema from fiducial marker insertion, CTV contouring variation, and differences in the prostate appearance on the pre-and post-SOH CT scans. The prostate edema effect has been well-documented after brachytherapy seed implant. 25 Intra-and inter-observer variability of around 10%- Mean rectal dose volumes reported for the RW optimization structure in this study are slightly lower than those reported in other dosimetric studies and more than 50% lower than van Gysen et al.
for the same 80%-90% dose range. However, PTV volumes in this current study were 30% lower on average while rectum volumes were 15% higher, likely leading to lower PTV overlap with the rectum and therefore lower rectal dose volumes.
Bladder dose was evaluated using V60 Gy, V55 Gy, V50 Gy, V46 Gy, and V37 Gy metrics. Each treatment plan passed the clinical objectives, BWV46 Gy < 20% and BWV37 Gy < 40%. There were no statistically significant changes in bladder dose pre-to post-SOH in VMAT or IMRT treatment plans consistent (P > 0.1) with other studies. 8,10,14 However, VMAT treatment plans resulted in lower bladder dose compared to IMRT treatment plans (P < 0.007). Similarly, mean penile bulb dose was found to have no statistically significant difference between pre-and post-SOH plans (P > 0.24) while VMAT treatment techniques yielded lower PB dose with an improvement ranging from 0.5%-1.0% (P < 0.04). Other studies reported higher pre-and post-SOH mean PB dose while reporting a decrease in mean PB dose from pre-to post-8 which is inconsistent with the results of this study.
Finally, the gradient measure (GM) and the conformity index (CI) indices were useful indicators of plan quality in addition to the OAR dose (Table 6). In Eclipse, GM was defined as the difference between the equivalent sphere radii of the prescription and 50% isodose lines while CI was defined as the volume enclosed by the prescription isodose surface divided by the target volume. VMAT plans had a statistically significant (P < 0.002) decrease in GM compared to IMRT while no change in CI was seen between plans. Although VMAT post-SOH treatment plans offered lower bladder dose, mean penile bulb dose and lower GM, IMRT treatment plans were created two times more quickly and as such there must be consideration made to the planning time required.

| CONCLUSION S
Rectal dose sparing greater than 25% was achieved in most post-SpaceOAR© Hydrogel treatment plans generated in this planning study. The use of SpaceOAR© Hydrogel significantly reduced rectal dose regardless of optimization structure or treatment technique employed. The rectal wall optimization structure offered a statistically significant reduction in rectal dose compared to the CRW.
There was no difference in rectal dose when using VMAT and IMRT treatment techniques, but VMAT offered lower bladder dose, mean penile bulb dose and gradient measure.

CONFLI CT OF INTEREST
No conflict of interest. IMRT: intensity modulated radiotherapy; RW: rectal wall; VMAT: volumetric modulated arc therapy.