A systematic review of stereotactic radiotherapy ablation for primary renal cell carcinoma

Authors

  • Shankar Siva,

    Corresponding author
    1. Division of Radiation Oncology and Cancer Imaging, Peter MacCallum Cancer Centre
    2. Sir Peter MacCallum Department of Oncology, University of Melbourne, Melbourne, Vic.
    Search for more papers by this author
  • Daniel Pham,

    1. Division of Radiation Oncology and Cancer Imaging, Peter MacCallum Cancer Centre
    Search for more papers by this author
  • Suki Gill,

    1. Division of Radiation Oncology and Cancer Imaging, Peter MacCallum Cancer Centre
    Search for more papers by this author
  • Niall M. Corcoran,

    1. Departments of Urology and Surgery, Royal Melbourne Hospital and the University of Melbourne, Parkville, Vic., Australia
    Search for more papers by this author
  • Farshad Foroudi

    1. Division of Radiation Oncology and Cancer Imaging, Peter MacCallum Cancer Centre
    2. Department of Pathology, University of Melbourne
    3. Sir Peter MacCallum Department of Oncology, University of Melbourne, Melbourne, Vic.
    Search for more papers by this author

Shankar Siva, Peter MacCallum Cancer Centre – Radiation Oncology, St Andrews Street East Melbourne, East Melbourne, Vic. 3002, Australia. e-mail: shankar.siva@petermac.org

Abstract

Study Type – Therapy (systematic review)

Level of Evidence 1a

What's known on the subject? and What does the study add?

At present, little is known about the role of stereotactic ablative body radiotherapy in the treatment of primary renal cell carcinoma. The published evidence to date totals 126 patients worldwide. The majority of evidence is retrospective in nature.

The present study adds context to the current literature by providing an overall summary of the evidence.

OBJECTIVE

  • • To critically assess the use of stereotactic ablative body radiotherapy (SABR) for the treatment of primary renal cell carcinoma with particular focus on local control and toxicity outcomes.

METHODS

  • • A systematic search on PubMed was performed in January 2012 independently by two radiation oncologists using structured search terms.
  • • Secondary manual searches were performed on citations in relevant publications and abstracts in major radiotherapy journals.
  • • Outcomes, techniques, biological doses and scientific rigour of the studies were analysed.

RESULTS

  • • In total 10 publications (seven retrospective and three prospective) were identified. A wide range of techniques, doses and dose fractionation schedules were found.
  • • A total of 126 patients were treated with between one and six fractions of SABR. Median or mean follow-up ranged from 9 to 57.5 months. A weighted local control was reported of 93.91% (range 84%–100%).
  • • The weighted rate of severe grade 3 or higher adverse events was 3.8% (range 0%–19%). The weighted rate of grade 1–2 minor adverse events was 21.4% (range 0%–93%). The most commonly employed fractionation schedule was 40 Gy delivered over five fractions.

CONCLUSIONS

  • • Current literature suggests that SABR for primary renal cell carcinoma can be delivered with promising rates of local control and acceptable toxicity.
  • • However, there was insufficient evidence to recommend a consensus view for dose fractionation or technique.
  • • This indicates the need for further prospective studies assessing the role of this technique in medically inoperable patients.
Abbreviations
RF

radiofrequency ablation

SABR

stereotactic ablative body radiotherapy

SRS

stereotactic radiosurgery

INTRODUCTION

RCC is the ninth most common cancer in Australia [1] and the eighth most common cancer in the UK [2]. It affects predominantly the older population with a median age at diagnosis of 65 years, with a slight male preponderance. According to the American Cancer Society kidney cancer incidence rates increased by 4.1% per year in men and 3.3% per year in women between 2004 and 2008 [3]. It is postulated that the increasing incidence of RCC is largely incidental and due to the more frequent use of abdominal imaging with ultrasound and CT. For those who require treatment, surgery, either as a radical nephrectomy or more preferably as a nephron-sparing partial nephrectomy, is the standard of care for primary disease. However, given the demographics of patients with RCC, many older patients have comorbidities which may preclude them from major surgery. Alternative strategies such as cryotherapy, radiofrequency ablation (RFA) and stereotactic radiosurgery have emerged as potentially curative treatment approaches for patients who refuse or are unsuitable for surgery.

Stereotactic ablative body radiotherapy (SABR) is a novel, potentially curative treatment approach for inoperable primary RCC. ‘Radiosurgery’ was a term first coined by Swedish neurosurgeon Lars Leksell [4] in the 1950s to describe single-dose ablative radiotherapy delivered to brain lesions via stereotaxy. The term stereotaxis applies to the realization of tumour position via the use of coordinates derived from external surrogate markers or fiducials. These fiducials allow the determination of tumour coordinates in the sagittal, coronal and axial planes. This principle has been extrapolated to the stereotactic delivery of severely hypofractionated treatments to body targets. Lax et al. [5] developed the first stereotactic body frame to enclose the body from head to mid-femoral region with vacuum stabilization to provide high surface contact (Fig. 1). There is an established role for the use of stereotactic radiotherapy in the treatment of tumours in the brain [6], lung [2,7,8], liver [9] and spine [2,10]. Several reports of both single-fraction stereotactic radiosurgery (SRS) and fractionated SABR techniques have emerged in the treatment of primary kidney RCC. In contrast to RFA and cryotherapy, SABR and SRS are capable of treating both larger tumours and those adjacent to collecting vessels and ureteric ducts. Additionally, these novel techniques are non-invasive and delivered whilst the patient is fully awake.

Figure 1.

A representative 10-field SABR plan with a prescription dose of 42 Gy in three fractions (Peter MacCallum Cancer Centre). The right kidney RCC is contoured in red, with the radiotherapy planning target volume contoured in cyan. The dose colour wash depicts the 20 Gy (blue) to 46.5 Gy (red) dose spread.

BASIC PRINCIPLES OF STEREOTACTIC TREATMENT

Stereotactic techniques represent a revolutionary departure from conventional radiotherapy. Typical fractionation for curative intent treatment of epithelial tumours generally involves the use of 30 or more small doses of 1.8–2 Gy per fraction, delivered 5 days a week over many weeks. This is in order to maximize the therapeutic window between tumour cell kill and normal tissue repair in the context of relatively large and imprecise fields. In contrast, recent work with hypofractionated SABR/SRS has resulted in the use of fraction sizes of up to 25 Gy delivered to the upper abdomen [9,11]. The very large hypofractionated doses used in SABR can be given safely because (i) the treated volumes are small with tight margins and (ii) the technique employs a large number of beams (eight or more), which individually contribute a small dose along their path but together result in a much larger dose where they intersect and are summed at the locus of the cancer (Fig. 2). A specific challenge for SABR/SRS in the kidney is respiratory-synchronous organ motion. Common strategies employed to ameliorate respiratory induced uncertainty include respiratory gating of the beam delivery [13], with or without breath hold techniques [12], and tumour tracking of the beam using a robotic-arm-mounted linear accelerator [13]. Additional efforts to reduce total tumour excursion are commonplace and include abdominal compression or patient coaching for shallow breathing [14].

Figure 2.

A fully awake patient vacuum immobilized in the Elekta BodyFix® system (Medical Intelligence, Schwabmünchen, Germany) at the Peter MacCallum Cancer Centre.

A linear quadratic equation [15] is often used to estimate cell survival to radiation. A linear quadratic equation is used to calculate biological equivalent doses (BEDs) between the various fractionation schemes reported below, as this allows a direct comparison of effective doses for early effects to tumour and normal tissues. However, at very large doses per fraction, radiotherapy becomes tissue ablative and may not follow conventional radiobiological rules [16]; thus, the absolute values of the BED calculations reported here should be interpreted with caution.

The term ‘radiosurgery’ to describe single-fraction ablative radiotherapy is misleading as it does not involve surgery at all. It could be argued that ‘radioablation’ is a more appropriate term. Similarly, the definition of stereotaxis has been somewhat loosely applied throughout the literature. The need for fiducial surrogates for precise tumour localization has been largely replaced by image-guided radiotherapy using tumour visualization, although the term stereotactic is still often used. Image-guided radiotherapy involves the use of some method of direct or indirect radiological localization of the tumour in the treatment room immediately before or during treatment delivery [17]. This may be with volumetric soft tissue imaging devices attached to the treatment machine or with implanted radio-opaque fiducials that are detectable by orthogonal kilovoltage X-rays. In this review of the literature, we shall consider all definitions of stereotactic radiotherapy and radiosurgery published to date. A large proportion of published reports regarding SABR address the treatment of both primary and metastatic RCC lesions. This review will focus only on the reported outcomes of those patients treated for primary disease.

PATIENTS AND METHODS

A systematic literature search was performed using PubMed for the period from January 1995 to February 2012. The search employed the following structured search terms: ‘kidney neoplasms [MeSH] AND (radiotherapy[Title] OR radiosurgery[Title] OR stereotactic[Title] OR radiation[Title]) NOT (metastases[Title] OR metastasis[Title]) NOT brain[Title]’. This yielded 161 publications. Papers were subsequently independently reviewed by two radiation oncologists (SS and FF) for relevance. A secondary search was performed using recovered paper citations and an abstract search of the major North American and European radiotherapy journals: the International Journal of Radiation Oncology, Biology and Physics and Radiotherapy and Oncology. Where multiple publications existed for a single institution, outcome data from only the most recent or most relevant papers were included. However, previously published information regarding treatment techniques were considered if this gave useful additional information. Where publications reported outcomes for both primary and secondary RCC, only the former patients were considered. In total, 10 clinical studies and two preclinical studies were identified. Relevant clinical information concerning tumour characteristics, treatment techniques, duration of patient follow-up, treatment-related toxicities and radiobiological information is tabulated chronologically in Table 1. Two-year overall survival and local control data, when not reported, were extrapolated from reported time points assuming a constant hazard. Weighted crude local control rate and mean 2-year local control rate were calculated using the average of each of the categories with respect to the contribution of the number of patients within each study. The BED of each treatment was calculated using the α/β estimates of the two common RCC cell lines Caki-1 and A498 as investigated by Ning et al. [18].

Table 1. Summary of studies assessing SABR in primary RCC, tabulated in chronological order
Author, year of publicationPatientsStudy designFollow-up (median or mean)Dose/fractionationOutcome – crude local control (%)Estimated 2-year local controlOverall survival, medianBED estimated by RCC typeToxicity
Caki-1 (α/β= 6.1)A498 (α/β= 2.6)
  • *

    Additional information through personal correspondence.

Qian et al. 2003 [44]20Retrospective128 Gy × 593%86 Not reported86.4163.1Not reported
Beitler et al. 2004 [45]9Retrospective26.78 Gy × 5, 7 Gy × 6100%1004 of 9 patients alive84.6–86.4155.1–163.133% grade 1–2, nil grade 3+
Wersäll et al. 2005 [24]8Retrospective378 Gy × 5, 10 Gy × 4, 15 Gy × 3100%100Median survival 58+ months86.4–142.8163.1–304.620% grade 1–2, 19% grade 3, nil grade 4+
Gilson et al. 2006 [46]33Retrospective17Median 8 Gy × 594%92Not reported86.4163.1Not reported
Svedman et al. 2006 [47]5Prospective phase II528 Gy × 4, 10 Gy × 4, 15 Gy × 2, 15 Gy × 380%91Median survival 32 months69.1–142.8130.4–304.689% grade 1–2, 4% grade 3
Teh et al. 2007 [41]2Retrospective924–48 Gy in 3–6 fractions100%100Not reported51.8–103.797.8–195.7Not reported
Svedman et al. 2008 [40]7Retrospective3910 Gy × 3 or 10 Gy × 486%91Not reported73.48–98.0145.4–193.858% grade 1–2, nil else
Nomiya et al. 2008 [23]10Retrospective57.5Median 4.5 Gy × 16100%1005-year overall survival 74%119196.910% grade 4 toxicity, no other toxicities > grade 1
Kaplan et al. 2009 [21]12Prospective phase INot reportedMax 13 Gy × 384%Not reportedNot reported112.5234Nil
Ponsky and Vricella [22], ongoing*20Prospective phase INot reportedMax 16 Gy in 3 fractionsNot reportedNot reportedNot reported159.3343.4Nil

RESULTS: PRECLINICAL EVIDENCE

Stereotactic radioablative approaches have been evaluated in both porcine and murine models in the literature. Walsh et al. [19] injected nude mice with the A498 line of human RCC cells and subsequently irradiated 12 mice with seven mice observed as unirradiated controls. This group used a fractionated SABR approach, with three fractions of 16 Gy delivered weekly. All except one mouse completed all 3 weeks of treatment, and this mouse died of unknown cause without evidence of radiation toxicity. Whilst tumour volume in control mice progressively increased, the irradiated tumours decreased to <30% of the initial volume. Tumours in mice sacrificed 4 weeks after treatment demonstrated no mitotic activity.

The radioablative effect of stereotactic irradiation was also evaluated in a porcine model in an earlier study by Ponsky et al. [20]. In this model, eight pigs and 16 kidneys were irradiated with single-fraction SRS in a dose escalation study from 24 Gy to 40 Gy. A hypothetical renal ‘target’ was irradiated whilst the pigs were anaesthetized, and subsequently the kidneys were harvested at 4, 6 and 8 weeks after treatment. This group demonstrated a sharp demarcation of fibrosis at a gross and microscopic level between the renal ‘targets’ and surrounding normal parenchyma. After 8 weeks, the ‘targets’ showed complete fibrosis with relative sparing of surrounding renal tissue. At the time of kidney harvest there was no gross evidence of injury to the surrounding organs or body wall, renal blood vessels or the collecting system.

RESULTS: CLINICAL EVIDENCE

A total of 10 published studies were reviewed and 126 patients were identified treated for primary RCC (Table 1). Three studies were prospective in nature, whilst the other seven were retrospective. Kaplan et al. [21] and Ponksy and Vricella [22] used a robotic arm-held linear accelerator system in their respective phase I studies, and Nomiya et al. [23] reported the use of a heavy carbon-ion particle accelerator, whilst all other studies used conventional gantry-operated linear accelerators. None of the groups restricted or excluded tumours based on proximity to collecting vessels or renal vasculature. Reporting institutions cited no size restrictions. There were no reports of pathological confirmation of tumour response through post-treatment biopsy, which is consistent with routine clinical practice after radiation therapy.

Of the 10 published clinical studies, techniques and dose fractionation schedules varied widely. Three, four and five fraction approaches were most commonly reported. The most commonly employed fractionation schedule was 40 Gy delivered over five fractions. The median or mean follow-up of reported series ranged between 9 and 57.5 months. Crude local control was commonly reported, with individual reports ranging from 84% to 100%. The crude weighted local control rate and 2-year estimated weighted local control rate were 93.1% and 92.9% respectively. Overall survival was inconsistently reported in these series. The most common reported toxicities were fatigue and nausea, followed by radiation dermatitis and enteritis. Rates of severe toxicity (grade 3+) were very low, although in one study a 19% rate was recorded [24]. Nomiya et al. [23] reported one late grade 4 skin toxicity. The weighted rate of severe toxicity was 3.8%, and the weighted rate of minor toxicity (grade 1–2) was 21.4%.

DISCUSSION

A common preconception is that RCC is resistant to radiotherapy. This notion was challenged in the neoadjuvant setting through an early report of improved disease-free survival in one surgeon's series in the 1960s [25]. However, subsequent clinical trials throughout the 1970s and 1980s failed to demonstrate any increase in survival with external beam radiation delivered in the neoadjuvant or adjuvant setting [26–29]. A meta-analysis of seven randomized controlled trials and 735 patients treated with postoperative radiotherapy was recently published [30]. Pooled results from these trials showed a significant reduction of locoregional failure in patients treated with postoperative radiotherapy (P < 0.001) but no difference in overall survival (P= 0.29). Currently the role of conventional radiotherapy for primary RCC is limited, possibly to patients with poor pathological features who are at high risk of tumour recurrence [31].

Research from Stanford University suggests that the inherent radioresistance of RCC may only exist at low doses per fraction. Ning et al. [18] performed clonogenic survival assays with two human RCC cell lines (Caki-1 and A498). The cells were irradiated with 0–15 Gy and surviving fractions were calculated. At doses used in conventional radiotherapy (1.8–2 Gy per fraction) there was only a small proportion of cell kill observed in either cell line. However, at doses above 6 Gy per fraction an exponential rate of cell kill was noted, suggesting that radiotherapy may be profoundly effective in RCC when used at high doses per fraction. Considerable evidence for this effectiveness can be demonstrated in the well established stereotactic literature for RCC metastases to the brain, in which multiple studies consistently report local control rates of 90% or higher [32–37].

When considering the dose fractionation schedules employed by the groups assessed in this literature review, a very high BED is certainly achieved through SABR. BED calculations are commonly used to compare dose fractionation schedules in radiotherapy [38]. The BED is based an the α/β ratio, which expresses a tumour or tissue sensitivity to a specified dose [15]. The BEDs of the different schedules used have been calculated in Table 1 by using the estimated α/β ratios of the Caki-1 and A498 RCC cell lines (6.9 and 2.6 respectively) [18]. The estimated BED ranged widely in the studies between 86.4 Gy and 159.3 Gy for the Caki-1 cell line. Similarly, the estimated BED ranged between 163.1 Gy and 343.4 Gy for the A498 cell line. In contrast, a hypothetical conventionally fractionated long course of radiotherapy (e.g. 60 Gy in 30 fractions delivered over 6 weeks) would only achieve a BED of 77.4 Gy and 106.1 Gy for the Caki-1 and A498 cell lines respectively. It is clear that the BEDs achieved with SABR and SRS techniques are an order of magnitude greater than can be achieved with conventional radiotherapy techniques. The caveat to these comparisons is that BED calculations may not be reliable at the severely hypofractionated doses in the SABR and SRS techniques, given that preclinical models did not account for such high doses per fraction.

Clinical evidence to date supports the assertion that high tumour control rates with minimal associated toxicities can be achieved using high dose per fraction stereotactic techniques. Individual reports of crude local control ranged from 84% to 100%. The crude weighted local control rate and 2-year estimated weighted local control rate were 93.1% and 92.9% respectively. These results are comparable with other ablative techniques, with a recent meta-analysis indicating that local tumour progression was 5.2% after renal cryoablation and 12.9% after RFA [39]. Survival, when reported, was difficult to interpret in the context of the patient demographics. Most patients were medically inoperable with multiple comorbidities and most patient deaths occurred with the renal disease controlled. Treatment delivery was well tolerated with minimal severe toxicities. The weighted severe (grade 3+) toxicity rate in these series was 3.81% with no study reporting a patient death.

There have been no reports of clinically relevant renal dysfunction secondary to SABR. The potential for nephron injury in humans kidneys after SABR has been most eloquently researched in a series of seven patients with only one functioning kidney published by Svedman et al. [40]. These patients were subsequently treated with SABR in the remaining kidney, with a maximum reported follow-up of 6 years. Five of the seven patients had no change in renal function. In one patient serum creatinine increased from 120 µmol/L to 150–160 µmol/L after 2 years of follow-up. In the other patient, creatinine levels increased by 20% during the 6-year follow-up period, without medical or dialysis intervention required. None of the patients developed hypertension over the 6-year follow-up period. In another series by Teh et al. [41] asymptomatic elevations in serum creatinine were observed in patients 52 months after renal SABR. In light of this the QUANTEC group consensus guidelines for dose constraints in radiotherapy recommended no dose constraint be applied for the kidney during SABR [42]. The QUANTEC group suggested that ‘one hypothesis [is] that nearly complete sparing of a substantial volume of normal kidney is associated with preservation of renal function’[42]. An alternative hypothesis is that late renal nephropathy may manifest at a later date (up to 10+ years after SABR); however, current data are not yet mature enough to support this hypothesis and it may not be relevant in the medically inoperable patient cohort.

There was insufficient evidence in this review to support that any one delivery system was clinically superior to another. The conventional gantry-operated linear accelerator was most commonly used in these reports and has the advantage of being readily accessible in most radiotherapy centres. The Cyberknife® (Accuray, Sunnyvale, CA, USA) is a robotic arm-held linear accelerator, allowing many degrees of freedom to deliver over 100 beamlets, which permits the high dose regions to conform tightly to the target area. However, the system is expensive, may require several hours to deliver each dose of radiation (limiting patient throughput) and usually requires the insertion of radio-opaque fiducials to assist tumour tracking. The use of a heavy carbon-ion particle accelerator (reported by Nomiya et al. [23]) represents a highly specialized solution to radiotherapy delivery, which takes advantage of the distinct dosimetric advantages of heavy particle radiation over conventional X-ray therapy. However, this system is not readily available for medical use outside Japan. Irrespective of the system used to deliver SABR or SRS similar clinical results were achieved.

Although SABR and SRS techniques are still an emerging modality for the treatment of primary RCC, high reported rates of tumour control and low toxicity are promising. In particular, stereotactic radiotherapy techniques are non-invasive and are not limited to smaller tumours or to those in close proximity to hilar structures. They are delivered without anaesthesia whilst the patient is fully awake. These attributes are particularly attractive when considering alternative modalities for the treatment of inoperable primary RCC, such as cryotherapy or RFA. At present, there is insufficient evidence to recommend a consensus view for the optimal dose/fractionation, technique or delivery system. There is only limited scientifically rigorous prospective evidence at this time. Further well designed prospective trials are required to validate the available literature.

ACKNOWLEDGEMENTS

The authors gratefully acknowledge philanthropic research support through the CASS Foundation Science and Medicine Grant.

CONFLICT OF INTEREST

None declared.

Ancillary