First UK patient cohort treated with stereotactic ablative radiotherapy for primary kidney cancer

Abstract Aims Stereotactic ablative radiotherapy (SABR) for primary renal cell carcinoma (RCC) is a promising non‐invasive ablative treatment option. A prospective interventional clinical trial published showed that treatment was feasible and well tolerated. We present the first single‐institution UK cohort of patients with primary RCC receiving protocol‐based SABR with prospective follow‐up. We also present a protocol that could be used to facilitate more widespread use of the treatment. Materials and methods Nineteen biopsy‐proven primary RCC patients were treated with either 42 Gy in three fractions on alternate days or 26 Gy in a single fraction based on predefined eligibility criteria using either Linear Accelerator or CyberKnife platform. Prospective toxicity data using CTCAE V4.0 and outcome data such as estimated glomerular filtration rate (eGFR) and tumour response using CT thorax, abdomen and pelvis (CT‐TAP) were collected at 6 weeks, 3, 6, 12, 18 and 24 months post treatment. Results The 19 patients had a median age of 76 years (interquartile range [IQR] 64–82 years) and 47.4% were males, and they had a median tumour size of 4.5 cm (IQR 3.8–5.2 cm). Single and fractionated treatment was well tolerated and there were no significant acute side effects. The mean drop from baseline in eGFR at 6 months was 5.4 ml/min and that at 12 months was 8.7 ml/min. The overall local control rate at both 6 and 12 months was 94.4%. Overall survival at 6 and 12 months was 94.7% and 78.3%, respectively. After a median follow‐up of 17 months, three patients experienced a Grade 3 toxicity, which was resolved with conservative management. Conclusion SABR for primary RCC is a safe and feasible treatment for medically unfit patients, which can be delivered in most UK cancer centres using standard Linear Accelerator as well as CyberKnife platforms.


| INTRODUCTION
Kidney cancer incidence is increasing with approximately 13 100 new cases per year within the United Kingdom. Accounting for 4% of all cancer cases, kidney cancer is strongly related to age, with the highest incidence rates being in older men and women between 85 and 89 years. 1 The majority (56%) of patients are diagnosed with localised stage I and II diseases, compared with late stage (stages III and IV) with the 5-year overall survival rate being 85% for stage 1 disease and 75% for stage II disease. 1 Renal cell carcinoma (RCC) is traditionally considered to be a radio-resistant tumour with surgical excision being the current standard of care for primary kidney cancer. However, a significant proportion of patients are considered unsuitable for surgery due to concurrent co-morbidities and frailty.
Surgically unfit patients with primary RCCs may be offered invasive ablative procedures such as radiofrequency ablation (RFA) or cryoablation (CA). RFA and CA typically require general anaesthesia (GA) to allow the insertion of catheters, making them unsuitable for less fit patients. There is a substantial unmet clinical need for a non-invasive treatment option for patients where surgery is not possible. RFA and CA have poor control rates in tumours larger than 3-4 cm and those that are centrally located and have a high risk of complications in tumours near the hilum. 2,3 Consequently, for many patients, although they have localised disease, the current standard of care is watchful waiting until the development of the metastatic disease. Metastatic disease is incurable, and treatment may be poorly tolerated in older patients with co-morbidities. A prospective interventional clinical trial published by Siva et al. showed that a non-invasive treatment option, stereotactic ablative radiotherapy (SABR) for primary RCC, was feasible and well tolerated. 4 The treatment was shown to provide freedom from local progression, distant progression and overall survival rates at 2 years of 100%, 89% and 92%, respectively. 4 In this prospective case series, we present our experience from a single institution of treating localised RCCs using CyberKnife and Linear Accelerator platforms including early patient outcomes and initial toxicity. The cohort was used to assess the feasibility of the delivery of SABR using CyberKnife and Linear Accelerator platforms.

| MATERIALS AND METHODS
Patients were identified at a single institution by a urology multidisciplinary team (MDT) with the following eligibility criteria: presence of biopsy-proven RCC of less than 6 cm maximum diameter, non-surgical candidates and consensus at the urology MDT that radical treatment was appropriate. The decision to treat the RCC with SABR was because the patient was not fit for the GA needed for surgery or CA/RFA or was unsuitable for CA/RFA due to the location of the tumour. Majority of the patients who had >4 cm tumour had period of surveillance confirming progressive disease.
Although SABR technique can be used for tumours larger than 6 cm, for the purpose of keeping the cohort similar, we chose to keep maximum tumour size up to 6 cm.
Exclusion criteria included metastatic disease and factors precluding abdominal radiotherapy such as inflammatory bowel disease.
Patients were subsequently discussed at the SABR MDT to confirm they were technically treatable and to decide on treatment platform.
Patients were treated as part of a prospective service evaluation with approval from the Novel Therapeutics Committee with peerreviewed approval of the protocol and funding from the hospital charity.
After patients gave their informed consent, the following data were collected: baseline Charlson co-morbidity index, quality of life (QOL) assessment using Equation 5D and EORTC QLQ-30 questionnaires, estimated glomerular filtration rate (eGFR), CT thorax, abdomen and pelvis (CT-TAP) and dimercaptosuccinic acid (DMSA) split renal function, where clinically appropriate.
Face-to-face or telephone follow-up data were collected at 2 weeks, 6 weeks, 3 months and 6 months, then 6 monthly thereafter.
CT-TAP (with intravenous [IV] contrast where appropriate) was performed post treatment at 6, 12 and 24 months. Toxicity using CTCAE V4.0 and QOL assessments were collected at 6 weeks, 3 months and 6 months, then 6 monthly thereafter for 3-5 years. Some imaging was delayed during the coronavirus disease 2019 (COVID-19) pandemic.
The primary endpoint was local control at 12 months and the secondary endpoints were safety, QOL, maintenance of renal function, response rate and overall survival.

| Radiotherapy planning and motion management
The decision on treatment platform and technique, standard Linear Accelerator, using a volumetric modulated arc therapy (VMAT) technique, or on CyberKnife, was made at the SABR MDT. Patients trea- For patients treated with CyberKnife, fiducials were inserted under local anaesthesia and CT guidance by a trained radiologist. Two sets of gold fiducials (two fiducials in each needle) were inserted with a separation distance of 2 cm. Fiducial placement occurred at least 10 days prior to planning CT scans to mitigate the effect of migration.
Patients were scanned in a supine position in an inhale and exhale breath-hold with 1-mm scan thickness. For the exhale breath-hold scan, IV contrast was given and this scan was used as the primary dataset.
For the VMAT treatments, patients were scanned with IV contrast in exhale breath-hold, using a slice thickness of 2 mm, followed by a 4DCT scan to capture the extent of tumour motion. Patients were scanned in a supine position and immobilised using a vac bag, knee fix and with their arms above the head on a wing board. Abdominal compression was used for further immobilisation where appropriate.

| Definition of target volumes and OARs
For patients treated with the VMAT technique, the 3D exhale breathhold scan was used as the primary scan for contouring and planning.
The fused 4DCT was used to generate an ITV (interval target volume) using the 4D scan. A 5-mm margin was then added in all directions to the ITV to create the PTV (planning target volume). For CyberKnife treatments, the exhale-phase CT scan was used for contouring and identifying fiducial position. The CTV (clinical target volume) was the same as GTV (gross tumour volume) and a 3-mm margin was added in all directions to create the PTV.
Small bowel, large bowel, duodenum, stomach, liver, spleen, contralateral kidney and spinal cord were outlined as OARs on the primary planning scan. Outlining volumes were reviewed by two clinical oncologists with SABR and uro-oncology experience.

| Dose fractionation
A dose of 26 Gy in one fraction or 42 Gy in three fractions, on alternate days, was prescribed. Patients with tumours > 5 cm or those that were in proximity to OARs were treated with three fractions. Treatment plans were produced with the aim to prescribe 100% of the prescription dose to at least 95% of PTV. For PTVs in proximity or overlapping OARs, a PTV_prescribe volume was created to ensure mandatory OAR constraints were achieved; the treatment dose was then prescribed to this structure.

| Treatment planning
For CyberKnife patients, treatment plans were calculated on Precision 2.0.0.1 (Accuray Inc, Sunnyvale, CA, USA) with a ray-tracing algorithm and a 1-mm dose grid and using the IRIS variable aperture collimator.
VMAT plans were created using RayStation 6.0.0.24 (RaySearch Laboratories AB, Stockholm, Sweden) using 6FFF beams with 1-2 arcs and, where possible, 180 arcs avoiding entrance through the contralateral side. A 2-mm dose grid was used and plans were calculated using a collapsed cone algorithm.
Plans were created to meet the mandatory OAR constraints listed in Table 1. The dose constraints were amalgamated and modified based on constraints taken from the UK SABR Consortium guidelines, 5 TG101 report 6 and the FASTRACK (Focal Ablative STereotactic Radiosurgery for Cancers of the Kidney) trial. 7 Conformity indices, as defined in the UK SABR Consortium guidelines, 5 were also used. to be <1 mm. VMAT plans were delivered on the Elekta Versa machines with XVI images taken pre-and mid-way through the treatment. Corrections were made for set-up errors > 2 mm.

| Dosimetric analysis
Analysis of dosimetric data and delineated structures was performed using in-house analysis software and Python 3.6. Doses to critical structures were assessed against the constraints in Table 1. Target doses were assessed using the PTV D95% and D2% as well as the GTV mean dose.
BED (biologically effective dose) values for the tumour were calculated using the α/β ratios 2.6 and 6.9 Gy from a study of two cell lines. 8 These relatively low values for the α/β ratio are a key motivation for SABR, although the absolute values of α and β are as small as may be expected for a highly radio-resistant tumour.

| Statistical analysis
Statistical analyses were carried out using Stata Version 17.0. Descriptive statistics, such as medians and interquartile ranges (IQRs) for continuous variables and frequency and percentages for categorical data, were used to summarise patient characteristics and their treatment.
Tumour response was assessed using the RECIST (Response Evaluation Criteria in Solid Tumours) criteria for those with measurements available at the required time point.
The time-to-event outcomes of time to local control failure and overall survival were calculated from the date of the first SABR treatment until the date of local control failure, or death, respectively, or censored at the last follow-up date to be alive and event free. These were summarised using Kaplan-Meier methods. Frequency of adverse events and their grade were summarised to assess safety.

| Relationship between change in eGFR and the delivered dose
In addition, we explored the relationship between the change in eGFR and the delivered dose. To explore this relationship, the DVH (dose volume histogram) for healthy kidney (Kidney-GTV) was converted to BED 9 : where n is the number of fractions, d is the dose per fraction and the α/β ratio was set to 3 Gy for a late effect. The generalised EUD (equivalent uniform dose) was used to condense the DVH into a single number. Using a volume parameter of 1, the correlation between the gEUD and mean change in eGFR post treatment was calculated.  This was due to the deaths prior to the time point or being unable to book the CT scan due to the COVID-19 pandemic. At 6 months, 3 patients had a partial response and 12 patients had stable disease.
The median percentage reduction in tumour size from baseline to 6 months was 16.67% (IQR 2.7-28.0). At 12 months, five patients had a partial response and nine patients had stable disease. The median percentage reduction in tumour size from baseline to 12 months was 26.5% (IQR 7.9-34.5).
One patient had local progression and two patients had metastatic disease during the follow-up period. The overall rate of local control at 6 and 12 months was the same: 94.4% (95% confidence interval [CI]: 66.6% to 99.2%; Figure 2A).
The overall survival rate at 6 months was 94.7% (95% CI: 68.1% to 99.2%) and the overall survival at 12 months was 78.3% (95% CI: 51.9% to 91.3%; Figure 2B). Seven patients died in total, of which six patients died of other causes not related to SABR or primary RCC (cardiac arrest, sepsis, heart failure or left ventricular failure; and two patients died of pneumonia).
A related concern is the magnitude of any relationship between the change in eGFR and the delivered radiation dose. Using a volume parameter of 1, the correlation between the gEUD and mean change in eGFR post treatment was 0.46 and was borderline statistically significant (p = 0.05). Setting the volume parameter equal to 1 is based on the assumption that the kidney is a parallel organ. Figure 3 shows

| DISCUSSION
There has been increasing interest in the delivery of SABR to pri- Charlson co-morbidity score However, doses were inconsistent in these studies, and we have aimed to align treatment planning and doses with the largest prospective cohort. 4 In our cohort, only one patient had a solitary kidney with a large RCC. Due to location and size of the tumour, it was felt that CA was not appropriate. Nephrectomy would have deemed the patient anephric, requiring unavoidable dialysis, an option that the patient refused.
The option of SABR has had significant impact on the QOL for this patient as well as saving of the cost of the dialysis for the NHS. Demonstrating SABR would be a promising non-invasive treatment especially for solitary kidneys. 15 Radiological response assessment following SABR has been challenging especially for renal cancers but has not been unique. 4 There is a paucity of literature on the response of healthy kidney to partial irradiation, which was discussed in the QUANTEC (Quantitative Analysis of Normal Tissue Effects in the Clinic) article on kidney toxicity. 24 A commonly used value of the volume parameter is 0.7, which dates to Emami et al. 25 In this case, we have a highly heterogeneous kidney dose with small volumes receiving the treatment dose with the contralateral kidney limited to 10 Gy to 33% of the volume or 18.6 Gy to 33% for one and three fractions, respectively. to a maximum of 0.46 for volume parameters around 1-5, before falling rapidly. It seems highly unlikely that a naturally parallel organ such as the kidney would have a volume parameter much greater than 3.
The majority of toxicity in our cohort was Grade 1 and most com-  11,13,27 There was no treatmentrelated mortality. Data on QOL have also been collected and will be published with additional late toxicity information after a further follow-up period.
The OAR constraints used for this cohort have been adapted and combined from various sources. [5][6][7] The use of different definitions of maximum dose to a structure and different volume parameters within the publications has made the process challenging and it is acknowledged that further work needs to be carried out. The allowed values for stomach and bowel were higher than routinely used in the United Kingdom. However, in this cohort, there were only two patient plans that exceeded the UK consensus 5 : one in the case of large bowel, where 0.5 cc = 42.9 Gy, and one for small bowel, where 5 cc = 21.4 Gy (UK SABR Consortium constraints for these parameters are 28.2 and 17.7 Gy, respectively). Neither of these patients reported more than Grade 2 toxicity.
Our prospective cohort has limitations, including singleinstitutional study, small sample size and short follow-up. In time with longer follow-up, we hope to update our results.
Although we recognise that, in this cohort, 37% of patients died, none of them died either due to the treatment received or due to treatment-related toxicity or metastatic disease, giving hope that if confirmed in future studies with longer follow-up, this treatment may be an option for patients who do not have surgical option and have less co-morbidities than reflected in this cohort.
There is an urgent need for a wider collection of prospective data to firmly establish this treatment technique. With an aging population and increased detection of the renal cancers, there is an unmet need for non-invasive effective treatment for primary RCC.

| CONCLUSION
Our study has demonstrated that SABR for RCC is safe, feasible and deliverable in a UK institution with previous SABR experience, using both one-and three-fraction regimes. It is our hope that this information will contribute to the development of a multicentre clinical trial or national guideline to be used in the UK clinical oncology community.