If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
The prevalence of hip prostheses is increasing. Prostate radiation delivery in the setting of hip prostheses is complicated by both imaging artifacts that interfere with volume delineation and dosimetric effects that must be addressed in the planning process. We hypothesized that with specialized planning, any photon-based definitive prostate radiotherapy approach may be utilized in patients with bilateral hip prostheses. Imaging data from sequential patients with prostate cancer and bilateral hip prostheses treated definitively with radiation were retrospectively reviewed. Bimodality imaging was used to define targets and organs at risk (OARs) along with specialized MRI sequences and/or orthopedic metal artifact reduction (OMAR) for MRI and CT artifact suppression, respectively. Multiple VMAT plans were generated for each set of patient images to include three fractionation schemes (conventional, hypofractionated, and SBRT), each with hip avoidance and with simulated normal hip. The ability to meet standard dose constraints was assessed for each plan type. Differences in target and OAR dosing between plans accounting for prosthetic hips via avoidance vs plans with simulated absence of prosthetic hip were also assessed. T-tests were used to compare dosimetric parameters. Ten patients with bilateral hip prostheses were identified, and 6 plans were created for each patient for a total of 60 radiation plans. Prosthetic hip avoidance did not result in failure to meet dose constraints for any patient. Hip avoidance resulted in minimal increases in high dose to the rectum and bladder (increases in mean V80%, V90%, and V95% ranged from 0.1% to 2.4%). Larger increases were seen at lower dose levels, with rectal V50% significantly increased in all three plan types with hip avoidance (conventional: 26.0% [standard deviation, SD 13.9] vs 16.9% [SD 10.2, p = 0.003]; hypofractionation: 26.4% [SD 13.3] vs 17.1% [SD 10.1, p = 0.002]; SBRT: 18.3% [SD 10.7] vs 10.5% [SD 6.9, p = 0.008]). Similarly, hip avoidance resulted in increases in bladder V50% to 31.7% (SD 16.8) vs 23.3% (SD 14.0, p = 0.001), 31.3% (SD 17.0) vs 23.3% (SD 13.8, p = 0.002), and 22.7% (SD 12.3) vs 16.5% (SD 12.6, p < 0.001) for conventional, hypofractionated, and SBRT plans, respectively. Hydrogel spacer resulted in reductions in rectal dose. For example, V70% for hip avoidance plans decreased with spacer presence to 8.3% (SD 6.7) vs 21.1% (SD 5.8, p = 0.021), 8.6% (SD 6.5) vs 21% (SD 5.7, p = 0.022), and 3.7% (SD 3.2) vs 15% (SD 8.2, p = 0.010) for conventional, hypofractionated, and SBRT plans, respectively. Any photon-based definitive prostate radiotherapy approach can be used with bimodality imaging for target and OAR definition and planning techniques to avoid dose attenuation effects of hip prostheses. Hydrogel spacer is a useful adjunct.
Therefore, trends in hip prosthesis placement and life expectancy will lead to increasing frequency at which those with hip prostheses will select radiation for management of prostate cancer. Hip prostheses present 2 major barriers to radiotherapy for prostate cancer. First, they interfere with target and organ at risk (OAR) delineation through artifact generation. Second, they complicate radiation planning due to radiation dose attenuation and scattering effects that can lead to dose heterogeneity.
Approaches have been developed to address both of these concerns.
With bilateral hip implants, the extent of radiation targets and OARs may not be clear on CT or MRI due to image artifacts. CT artifacts can be reduced through specialized imaging algorithms for metal artifact reduction (MAR) that may use dual energy CT, varied approaches to image reconstruction, and methods that directly identify artifacts and select appropriate corrections.
Several specialized metal artifact reduction sequences have been created for MRI; these include WARP (Siemens Healthcare, Munich, Germany), metal artifact reduction sequence (MARS), slice encoding for metal artifact correction (SEMAC), and multiacquisition with variable-resonance image combination (MAVRIC).
Bimodality imaging (CT and MRI) and use of sagittal, coronal, and axial images simultaneously may also aid structure delineation in this setting beyond benefits previously noted in normal prostate radiation planning.
Hip prostheses are made of high-Z materials, and their presence near a radiation target requires special attention during radiation planning to address attenuation and interface effects that can result in dose inhomogeneity.
TG 63 also devoted a substantial amount of effort to describing a large number of considerations in the event that treatment is delivered through a hip prosthesis to ensure that the nature of the implant and its impact on dose are well understood.
Some have attempted to characterize and account for the effects of hip prostheses on dose distribution to facilitate treatments that do not avoid dose entry through the prosthetic hip out of concern for potential increased dose to OARs near the prostate when hip avoidance is undertaken.
The work done thus far on treatment of patients with bilateral hip implants has generally assessed conventional radiation plans, but conventional radiation is becoming less common with increased adoption of moderate hypofractionation and, more recently, SBRT.
We hypothesized that with bimodality imaging, metal artifact reduction, and specialized radiation planning, any photon-based treatment approach may be utilized to achieve target and OAR goals for treatment of prostate cancer in patients with bilateral hip prostheses. Herein, we describe an approach to specialized prostate radiation treatment planning that allows accurate and safe prostate irradiation in patients with bilateral hip prostheses.
Methods and materials
Patients, data, and human subjects
Sequential patients with prostate cancer and bilateral hip prostheses who received prostate radiation via combination brachytherapy and/or external beam radiation (EBRT) 2019 to 2020 were reviewed retrospectively. Patients were treated at Ascension Providence Hospital or University of Michigan Hospital in MI. This study was conducted with approval from the Institutional Review Board (HUM00174718) at the University of Michigan.
Target and OAR delineation
Imaging artifacts due to hip prostheses were managed to facilitate definition of prostate, seminal vesicle, bladder, rectum, penile bulb, and urethral volumes. Imaging approaches included the following in all or a subset of patients: (1) bimodality imaging with both computed tomography (CT) and magnetic resonance imaging (MRI),
In addition to applying bimodality imaging and specialized sequences, the physician delineating volumes used axial, sagittal, and coronal images in the construction of volumes.
Patients with prostate cancer and bilateral hip prostheses who received prostate radiotherapy in 2019 to 2020 were reviewed retrospectively (Table 1). Volumetric modulated arc therapy (VMAT)-only plans for each subject were prepared by an experienced dosimetrist (E.S.) in Eclipse Treatment Planning System (Varian Medical Systems, Palo Alto, CA) using target coverage, contouring, and dose constraints (Tables S1 and S2) that varied based on plan type. The planning target volume (PTV) was defined as the prostate and proximal 1 cm of the seminal vesicles (SV) + 5 mm margin with prescriptions for conventional (79.2 Gy in 1.8 Gy/fraction) and hypofractionated (60 Gy in 2.0 Gy/fraction) radiation. The PTV expansion was 3 mm for stereotactic body radiotherapy (SBRT; 37.5 or 40 Gy in 7.5 or 8 Gy/fraction based on absence or presence of hydrogel spacer, respectively). No plans included lymph node treatment.
Additional planning details to account for hip prostheses included (1) the use of 3 arc VMAT, (2) avoidance sectors to prevent beam entry through bilateral hips + 5 mm margin, (3) ring optimization structures for integral dose dispersion, and (4) gradient optimization over the rectum. Two of the three VMAT arcs were 190 to 170 degrees while a third arc was 330 to 30 degrees. The ring structure began at 4.5 cm from the PTV and extended to the external surface with goal of ∼40% of max dose with adjustments made to adapt for the individual's anatomy. Gradient optimization over the posterior rectum consisted of structures at 5 mm and 1.5 cm posterior to the PTV with goals of 63% and 40% of maximum dose initially, respectively. Adjustments were made to gradient optimization structures as needed to account for anatomic variation. To assess the dosimetric effect of a decrease in rectal distension at the time of treatment compared to simulation, we simulated a 5 mm anterior shift of the isocenter and assessed coverage of the posterior PTV.
Prostheses were assumed to be titanium, so the CT number to electron density curve in the planning system was extended beyond the relative electron density of titanium. This allowed for the relative electron density of titanium to be set to the correct value of 3.74 for creation of plans that avoided hip prostheses. In the cases where full arc plans were created with the goal of simulating a patient without hip prostheses for comparison, the relative electron density of the implants was set to 1.17, reflecting the average femoral head density across 10 prostate cases.
Consideration was also given to density assumptions associated with artifact in the vicinity of the prosthetic hips, given that CT artifact associated with hip prostheses was noted on planning radiation scans. For artifact that corresponded anatomically to soft tissue, the density of these areas was overridden and was assigned a relative electron density of 1.00.
Data collection and analysis
Dose statistics were collected for each plan (see Table 2 and subsequent tables for listing of parameters). Percentage-based dosimetric parameters (percentage of structure volume receiving percentage of dose) were selected in most cases to facilitate comparison between plans with different dose targets. Summary statistics were generated for dose metrics. Paired T-tests were used to compare parameters between plans using avoidance of a hip implant vs those plans generated to simulate the absence of a hip implant. Additionally, T-tests were used to compare those with vs those without hydrogel spacers and plans with and without simulated prostate movement.
Ten patients with bilateral hip prostheses were included in this study retrospectively (Table 1). These individuals represented a mix of those with (70%) and without hydrogel spacer (30%), a range of gland sizes (SBRT PTV mean 88.1 cc; Range: 44.4 to 185.0 cc), and a range of pelvis widths (assessed by hip to hip distance mean 11.3 cm, range: 9.7 to 12.4 cm; Table 1). Six plans were created for each individual. These plans were derived using three different fractionation schemes: conventional fractionation (79.2 Gy), hypofractionation (60 Gy), and SBRT (37.5 or 40 Gy). For each fractionation scheme, we generated 2 plans: one with hip avoidance VMAT, and one without hip avoidance to simulate a patient without hip prostheses for each patient (60 plans total).
Table 1Patient population and plan summary details
Number of patients
MRI with WARP series
Number of plans
Number of arcs
SBRT PTV mean (range) in cc
Conventional and hypofractionated PTV mean (range) in cc
Hip-hip distance mean (range) in cm
Note: Number of plans includes plans developed accounting for hip prostheses with avoidance as well as plans developed to simulate absence of hip prostheses.
Bimodality imaging with MRI and CT was used for all patients. OMAR was used for CT artifact management in 8/10 patients (Fig. 1), and WARP sequence was used for management of MRI artifact in 3/10 patients (Fig. 2). At least one form of specialized imaging artifact management (either via CT, MRI, or both) was used in every case.
Adherence to dose constraints
All conventional and SBRT plans met dose constraints, and all hypofractionated radiation plans with hip avoidance met dose constraints save for two (Table S1). In one hypofractionated plan that failed to meet all constraints, the patient had a small bladder, which led to failure to meet bladder constraints for both the plan that avoided the hip implants and the plan that simulated absence of implants. In the other hypofractionated plan that failed to meet all constraints, rectal constraints were not met due to substantial PTV overlap with the rectum. In this case, both the plan that avoided the hip implants and the plan that simulated the absence of hip implants failed to meet rectal constraints.
Impact of hip implant and implant avoidance on dose parameters
A large number of dose parameters were assessed for each radiation plan (Tables 2 to 4). These are presented using percentages to allow comparison across modalities that use differing PTV expansions (3 mm for SBRT vs 5 mm for hypofractionated and conventional plans) and different dosing schemas as reviewed above. In all treatment approaches, higher max dose to PTV was delivered in the setting of hip implant avoidance, though in all cases this increased dose amounted to less than 1 Gy (Tables 2, 3, 4). D99% was not different between plans avoiding hip implants and those that simulated implant absence (Tables 2 to 4).
Table 2Impact of implant avoidance on conventional radiation plans
Simulated implant absence
Note: Values in columns not marked as “Parameter” or “P value” are formatted as mean (standard deviation).
Regarding rectal dosing, small increases in dose were seen at higher dose levels in the setting of hip avoidance vs. simulated absence of prosthetic hips for V80%, V90%, and V95% (differences ranged from 0.2% to 1.9% for the three treatment approaches). Larger increases were noted for lower dose levels. The V50% values were significantly higher in hypofractionated or conventional plans with hip avoidance. In these plans, V50% increased from 17.1% (SD 10.1) and 16.9% (SD 10.2) to 26.4% (SD 13.3; p = 0.002) and 26.0% (SD 13.9; p = 0.003) for hypofractionated and conventional plans, respectively, when hips were avoided. For SBRT plans, V50% approximately doubled to 18.3% (SD 10.7) from 10.5% (SD 6.9; p = 0.008).
Regarding bladder dosing, trends were similar to those observed with the rectum. In all plans, higher dose was delivered to the bladder when avoiding hip implants. There were small changes seen in V80%, V90%, and V95% (differences ranged from 0.3% to 2.4%) for all plan types. Lower dose parameter values increased to a greater degree, with V50% values increasing the most with hip avoidance. For conventional and hypofractionated plans, V50% bladder values increased to 31.7% (SD 16.8) and 31.3% (SD 17.0) from 23.3% (SD 14.0, p = 0.001) and 23.3% (SD 13.8, p = 0.002), respectively. For SBRT plans, the V50% with hip avoidance was 22.7% (SD 12.3) vs 16.5% (SD 12.6, p < 0.001).
Femoral head and penile bulb dosing were also assessed. With respect to the femoral heads, entry avoidance had indeed been achieved with very low doses delivered to the hips when the implant was avoided. There were no differences in penile bulb dosing, and this structure received very little dose. Only V20% is reported for penile bulb because parameters used to document higher dose demonstrated lack of dose at those levels (Tables 2 to 4).
Impact of hydrogel spacer
Though our patient set included only 3 individuals without hydrogel spacer for comparison with 7 individuals who had hydrogel spacer placed, comparisons were attempted and are presented in the supplementary appendix (Tables S3 to 5). The presence of a hydrogel spacer resulted in less dose to the rectum in all plan types regardless of whether or not implant avoidance was being used. Differences were significant for conventional plans with implant avoidance for V80% with mean 5.1% (SD 4.9) for those with spacers vs 15.5% (SD 4.9) without spacer (p = 0.015). Similarly, rectal V60% mean was 13.1% (SD 8.4) for those with spacers vs. 28.3% (SD 7.8) for those without spacers (p = 0.029). Similar significant changes were seen with hypofractionation and SBRT plans with hip avoidance.
Impact of movement
The effect of intrafraction prostate movement on hip avoidance plans was assessed in the plan type with the smallest PTV expansion, SBRT. When a 5 mm posterior prostate movement was simulated, the mean change in PTV D99 (Gy) was −7.3 Gy (Range: −10.3, −4.4) for plans that avoided the hips vs −9.4 Gy (Range: −13.1, −6.9) for plans that simulated absence of hip implants (p = 0.006). This suggests that plans that used hip implant avoidance were more robust to movement than those that simulated absence of hip implant (Table S6).
Hip implants present a challenge to the delivery of prostate radiotherapy. Due to demographic changes and shifts in orthopedic surgical practice, the frequency with which patients with bilateral hip implants have indications for prostate radiation will continue to increase. Though delivery of radiation to patients with bilateral hip prostheses has been described in the literature in a limited fashion for conventional radiation dose and fractionation, the advent of hypofractionation and SBRT has led many radiation oncologists to use conventional radiation less frequently. We have dosimetrically examined delivery of hypofractionated radiation and SBRT in the setting of bilateral hip prostheses to consider safety and plan robustness, and our results demonstrate that all photon-based approaches to external beam radiation should be safe and robust to both setup error and intrafraction prostate movement in patients with bilateral hip implants.
Appropriately designed radiation plans for management of prostate cancer in the setting of hip prostheses through hip avoidance should not present great risks to patients, regardless of fractionation scheme utilized (conventional, hypofractionation, or SBRT). We demonstrated this through two approaches. First the use of hip prosthesis avoidance did not result in failure to meet any of our dose constraints, which are either published or have to our knowledge been used to treat many patients with minimal toxicity (Tables S1 and S2). There were 2 instances of failure to meet constraints, but these were the result of patient features and were not the result of prosthesis avoidance. Second, though there were statistically significant increases in dose delivery to the rectum and bladder, these increases were largely at lower dose levels and amounted to only a couple of percentage points at higher dose levels, suggesting minimal clinical importance, despite statistical significance. Therefore, it is unlikely that altered fractionation plans would result in failure to appropriately treat disease or lead to OAR compromise.
Though the plans described in this report are very likely to be safe, there are 2 techniques that can be considered to increase safety in special circumstances, brachytherapy and hydrogel spacer placement. In patients who have a narrow pelvis, a particularly small bladder, or a small rectum, there could be a greater concern for toxicity. Brachytherapy provides the benefit of allowing delivery of 0% (if brachytherapy monotherapy) or ∼50% (if combination brachytherapy) of the dose externally. Additionally, brachytherapy relies on ultrasound, which is not impacted by hip implant presence. The use of hydrogel spacer in our patients resulted in a general reduction in rectal dose in all plans (Tables S3 to S5). We recommend hydrogel spacer use in hip implant cases where possible given that it provides an extra margin for safety and careful consideration of brachytherapy in appropriate patients for further mitigation of toxicity risk.
There are two general approaches to radiation delivery in patients with bilateral hip prostheses. The radiation planning approach utilized in this report focused on avoidance of hip implants, as suggested by TG 63 (Fig. 3, Fig. 4). Others have taken a different approach that involves accounting for the impact of the hip implant on dosimetry to facilitate the use of complete arcs without avoidance of beam entry through the hips.
Though we acknowledge that careful work to individually assess each implant's impact on dose distribution can facilitate delivery of quality radiation plans, we have concerns that this approach is labor intensive and may not be as robust to variation inherent in real world treatment settings. Both small prostatic movements during treatment delivery and setup errors due to imaging artifact on daily setup imaging are likely to have more profound effects on dose delivery when radiation is delivered through a prosthesis than they would with plans practicing avoidance of the hip prosthesis. Smaller PTV margins used with prostate SBRT would likely amplify concerns around any coverage uncertainty. Because of uncertainties about treatment delivery through the hip prosthesis, our practice has been to avoid hip prostheses when delivering radiation.
Extra steps were taken in radiation planning to further address some of the concerns raised in the previous paragraph related to inter fraction position change and intrafraction prostate movement. We included an additional 0.5 cm margin around each hip implant to assist with management of day to day (inter fraction) variation in the position of the prostate relative to the hips (Video S1). This extra margin on the hips allows for plans to be delivered easily even if there is an anterior or posterior shift of 0.5 cm without beam passage through the hips. With regard to intrafraction movement, we found that plans with hip implant avoidance are more robust to prostate movement in the AP dimension given the greater distribution of dose in the AP dimension when hips are avoided (Fig. 4, Video S1, Table S6). Whenever a patient with bilateral hip prostheses is being treated, special attention should be paid to the prostate in addition to the hips. If the prostate appears to have shifted anteriorly or posteriorly such that this motion would potentially lead to beam entry through the hips, a re-simulation should be considered, particularly when fewer fractions are planned.
Delivery of external beam radiation in the setting of bilateral hip implants requires special consideration with regard to image guidance. In cases where kV imaging is utilized, orthogonal imaging may be impossible due to the presence of implants. It is possible to instead perform kV imaging anterior-posterior and at 290°. For shorter duration plans, daily cone beam CT can be used with acceptance of limitations from artifact. It is important to recognize that effective image guidance in the setting of bilateral implants may require that non-standard approaches be utilized.
There are several potential limitations of this study. This study is based upon in silico planning and does not include actual patient outcomes data, so it does not present hard toxicity outcomes. It is important to note that others who have presented toxicity data using similar approaches to the one described in this report have found low levels of toxicity.
Here, we have noted adherence to appropriate dose constraints. The 10 patients presented here constitute a small sample of potential anatomic variation, and as a result, it is entirely possible that one would encounter variants that present novel challenges not encountered in this set; however, this set of 10 patients with bilateral hip implants is large compared to the bulk of previously published articles on radiation delivery to patients with bilateral hip implants. Specifically, many prior studies have included mostly unilateral hip prosthesis or have examined very small or singular sets of patient imaging data.
A strength of this presentation is the comparison between plans developed with hip implants against those prepared as if no hip implant was present in an effort to quantify the overall impact of this approach. Despite our reassuring findings and those of others, we encourage the involvement of a medical physicist early in the planning process for any bilateral hip implant case to address unforeseen complicating factors.
The techniques described in this report provide a methodology for delivery of any photon-based primary prostate radiation approach to an individual with bilateral hip implants. However, these techniques can be applied in settings beyond the delivery of primary definitive prostate radiotherapy. In our experience, dose constraints can be met with pelvic nodal treatment in combination with prostate radiotherapy, though this is beyond the scope of this manuscript. These concepts would also benefit approaches for radiation delivery in gynecologic malignancies and in the delivery of post-prostatectomy radiation. Future studies should explore these treatment scenarios in the context of bilateral hip implants.
With specialized planning, including hip avoidance with margin and gradient optimization, the cost of bilateral hip prostheses on radiation delivery to the prostate and proximal seminal vesicles is limited to small increases in mid-range rectal and bladder dose. Any photon-based radiation approach can be used safely. Hydrogel spacer is a useful adjunct.
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Conflicts of interest
The authors of this manuscript have no conflicts of interest to report.
This data has not been previously published or presented elsewhere.
Table S1. Organ at risk dose constraints for conventional, hypofractionated, and SBRT plans.
Table S2. Target dose coverage goals for conventional, hypofractionated, and SBRT plans.
Table S3. Impact of hydrogel spacer presence on conventional radiation plans with and without hip implant avoidance.
Table S4. Impact of hydrogel spacer presence on hypofractioanted radiation plans with and without hip implant avoidance.
Table S5. Impact of hydrogel spacer presence on SBRT plans with and without hip implant avoidance.
Table S6. D99% (Gy) for SBRT plans with hip implant avoidance vs those without implant avoidance (PTV = CTV + 3 mm).
Video S1. Video explanation of the contouring process for hip avoidance and impact of hip margin on dose distribution after movement.
The authors would like to thank graphic designer George Hixson for his assistance.