Comparative analysis for renal stereotactic body radiotherapy using Cyberknife, VMAT and proton therapy based treatment planning

Abstract Purpose We conducted this dosimetric analysis to evaluate the feasibility of a multi‐center stereotactic body radiation therapy (SBRT) trial for renal cell carcinoma (RCC) using different SBRT platforms. Materials/methods The computed tomography (CT) simulation images of 10 patients with unilateral RCC previously treated on a Phase 1 trial at Institution 1 were anonymized and shared with Institution 2 after IRB approval. Treatment planning was generated through five different platforms aiming a total dose of 48 Gy in three fractions. These platforms included: Cyberknife and volumetric modulated arc therapy (VMAT) at institution 1, and Cyberknife, VMAT, and pencil beam scanning (PBS) Proton Therapy at institution 2. Dose constraints were based on the Phase 1 approved trial. Results Compared to Cyberknife, VMAT and PBS plans provided overall an equivalent or superior coverage to the target volume, while limiting dose to the remaining kidney, contralateral kidney, liver, spinal cord, and bowel. Conclusion This dosimetric study supports the feasibility of a multi‐center trial for renal SBRT using PBS, VMAT and Cyberknife.


| INTRODUCTION
With an incidence of 62 700 new cases in 2016, kidney and renal pelvic cancers account for around 4% of newly diagnosed cancers in the USA. 1 Renal cell carcinoma (RCC) is the predominant and most lethal histology accounting for about 87% of these malignancies. 2,3 Historically, RCC has been labeled as a "radio-resistant tumor" and surgical nephrectomy was considered the cornerstone of treatment for RCC. A gradual shift in RCC treatment modalities began in the 1990s with the introduction of laparoscopic nephrectomy, high intensity focused ultrasound, cryoablation, radiofrequency ablation, and tyrosine kinase inhibitors but radiation remained rarely used. 2 Two main factors contributed to the underutilization of radiotherapy in treating RCCs: the high metastatic potential of the cancer and an inability to safely deliver high dose curative intent radiation to the primary tumor due to the anatomic proximity of the kidneys to other radio-sensitive structures such as small bowel. However, the successful use of both image-guided conventional radiotherapy and more recently stereotactic radio-surgery in the local treatment of extracranial and intracranial RCC metastases, respectively, challenged the role of radiation therapy in the management of RCC. While previous literature suggested that RCC is radio-resistant to small fraction sizes, 4 higher fraction dose delivered through Stereotactic Body Radiation Therapy (SBRT) can achieve promising rates of local control and acceptable toxicity. 5 SBRT allows for the accurate delivery of high dose radiation to specific extracranial targets while potentially avoiding toxic doses to adjacent structures. The use of SBRT in the treatment of local RCC was first reported by Qian et al. 5 who achieved a local control rate of 93% at a mean follow-up of 12 months. 6 A few years later, we reported our experience of a Phase I trial of SBRT using the Cyberknife platform and emphasizing the safety and efficacy in non-surgical RCC treatment. 7 SBRT can also be delivered with other platforms including volumetric modulated arc therapy (VMAT) or pencil beam scanning (PBS) proton therapy with each field covering the target uniformly. Dosimetric differences between these platforms have not been well studied in RCC and remain a major barrier for the implementation of large multi-institutional trials. Therefore, we conducted this study to assess the dosimetric feasibility of using non-robotic platforms for delivering curativeintent renal SBRT as a precursor for a future multi-institutional trial.

2.A | Patients selection
An institutional review board-approved phase I dose-escalation trial of SBRT using Cyberknife for primary treatment of non-surgical patients with localized RCC (NCT00458484) was initiated at our facility (Institution 1) since June 2006. The primary tumor was deemed to be resectable by an experienced urologic oncologist, but patients were referred to this phase I trial due to underlying medical conditions prohibiting surgical excision such as low probability of tolerating the general anesthesia, the surgery itself, or the postoperative recovery period. 7 At the time of the diagnostic biopsies, at least three fiducials markers were placed within and around the renal mass. 7 Within 1 week after fiducials insertion, computerized tomography (CT) simulation was acquired. Patients were treated to the primary tumor plus 0-3 mm margins with radiotherapy doses of 24, 32, 40, and 48 Gy in four fractions. Inclusion and exclusion criteria, radiation technique, dosimetric planning, and initial results were previously reported. 7 The institutional review board also approved the current dosimetric study. Among 19 patients with unilateral RCC treated according to the phase I trial with 48 Gy in four fractions, ten patients were randomly selected for this study. CT simulation images were then anonymized and shared with Institution 2.

2.B | Treatment planning and dosimetric variables
Using the anonymized CT images, treatment planning was performed using five different platforms with a prescription dose to the planning target volume (PTV) of 48 Gy in three fractions. These plat-

2.C | Statistical analysis
For each of the seven variables defined above (V 100% , V 14Gy , D 1cc -Bowel, D 1cc -Stomach, D 0.3cc -Cord, D 5% -Contralateral K., V 17Gy -Liver) we calculated the mean and the standard deviation (SD) across all 10 patients for each treatment planning platform. For statistical testing, V 100% across different platforms was considered paired as the planning measurement were applied to the same CT images.
We therefore performed a two tailed paired t-test with a confidence interval of 95% to compare V 100% , D 0.3cc -Cord, and D 1cc -Bowel of institution 1 Cyberknife to the other platforms and assess for target dose conformity. We used Bonferroni correction to account for multiple comparisons. D 1cc -Bowel and D 0.3cc -Cord in each institution were represented in value plots. All statistical analysis was done using Minitab â version 17.3.1 (Minitab Inc., State College, PA). Table 2 shows the mean and the SD of each of all seven variables.

| RESULTS
Tumor coverage was excellent while also sparing the ipsilateral kidney. The V 100% was greater than or equal to 97.4% for all the platforms and the V 14Gy ranged between 45.6% (VMAT -Institution 1) and 65.1% (Cyberknife -Institution 2). Mean V 100% was the lowest for Cyberknife at institution 1. The D 0.3cc -Cord constraint was satisfied for the five platforms [ Fig. 2(a) and Table 2]. For several cases, D 1cc -Bowel constraint was not achieved [ Fig. 2(b)]. The mean D 1cc -Bowel satisfied the dose constraint only for VMAT -Institution 2, while the other platforms had slightly higher means ranging between 1.76 Gy (Cyberknife -Institution 1) to 5.14 Gy (Cyberknife -Institution 2) above the dose constraint (Table 2). Table 3 show the Pvalues and the 95% confidence intervals of the paired t-test for V 100% , D 0.3cc -Cord, and D 1cc -Bowel. Using Bonferroni correction to account for multiple comparison, the P-value for V 100% was statisti-   The potential OAR toxicity associated with SBRT remains a limiting factor for adequate dose delivery. While VMAT in institution 1 was associated with higher D 0.3cc -Cord, VMAT and PBS planning at institution two resulted in lower D 0.3cc -Cord. With the OAR constraints used in our Phase I trial and for this study, no grade 3 or 4 toxicities were reported. 7 These constraints were developed initially for the institution 1 protocol prior to the initiation of the phase I trial. Since then, the International Radiosurgery Oncology Consortium for Kidney (IROCK) consortium has been adopted internationally to help guide uniform dose constraints. 9 These constraints are likely conservative and were strictly met for all the organs except the D 1cc -Bowel. In this feasibility study, no specific constraint optimization algorithm was used and the PTV expansion can be relaxed in cases were bowel toxicity is a concern. Regardless, the D 1cc -bowel show better overall dose sparing. 12 In the phase I protocol, image acquisition, target localization, and alignment correction were repeated during treatment delivery at intervals of 30-60 s. 7 To limit respiratory motion, 4D-CT images can be acquired with abdominal compression and plans could therefore incorporate an ITV. Contrary to gated radiotherapy, using an ITV would lead to an increase in the volume of the normal tissue irradiated. Given the large kidney size (3 9 6 9 12 cm 3 ), the relatively small respiratory-induced motion may mitigate the risk of non-gated treatment delivery, 13 especially that near complete sparing of a substantial volume of the kidney is usually associated with compensatory preservation of the renal function. 14 Compared to Cyberknife, VMAT and PBS have the advantage of decreased treatment times, 15 which decreases the risk of intrafraction movement, increases the number of patients treated per day, and provides more comfort to the patient. 15 Finally, difference in planning techniques between institutions could have been a confounder and thus this dosimetric study should be interpreted as an initial proof of concept that needs further support in multi-institutional settings.

CONFLI CT OF INTEREST
None.
T A B L E 3 Paired t-test for V 100% , D 0.3cc -Cord, and D 1cc -Bowel.