Cryoablation or radiofrequency ablation of the small renal mass

A meta-analysis


  • David A. Kunkle MD,

    1. Department of Urologic Oncology, Fox Chase Cancer Center, Temple University School of Medicine, Philadelphia, Pennsylvania
    Search for more papers by this author
  • Robert G. Uzzo MD

    Corresponding author
    1. Department of Urologic Oncology, Fox Chase Cancer Center, Temple University School of Medicine, Philadelphia, Pennsylvania
    • Department of Urologic Oncology, Fox Chase Cancer Center, 333 Cottman Avenue, Philadelphia, PA 19111===

    Search for more papers by this author
    • Fax: (215) 214-1734

  • See editorial on pages 2623–6, this issue.



The incidence of renal cell carcinoma is rising because of incidental detection of small renal masses (SRMs). Although surgical resection remains the standard of care, cryoablation and radiofrequency ablation (RFA) have emerged as minimally invasive treatment alternatives. The authors of this report performed a comparative meta-analysis evaluating cryoablation and RFA as primary treatment for SRMs.


A search of the MEDLINE database was performed reviewing the world literature for clinically localized renal masses treated by cryoablation or RFA.


Forty-seven studies representing 1375 kidney lesions treated by cryoablation or RFA were analyzed. No differences were detected between ablation modalities with regard to mean patient age (P = .17), tumor size (P = .12), or duration of follow-up (P = .53). Pretreatment biopsy was performed more often for cryoablated lesions (82.3%) than for RFA (62.2%; P < .0001). Unknown pathology occurred at a significantly higher rate for SRMs that underwent RFA (40.4%) versus cryoablation (24.5%; P < .0001). Repeat ablation was performed more often after RFA (8.5% vs 1.3%; P < .0001), and the rates of local tumor progression were significantly higher for RFA (12.9% vs 5.2%; P < .0001) compared with cryoablation. The higher incidence of local tumor progression was found to be correlated significantly with treatment by RFA on univariate analysis (P = .001) and on multivariate regression analysis (P = .003). Metastasis was reported less frequently for cryoablation (1.0%) versus RFA (2.5%; P = .06). Cryoablation usually was performed laparoscopically (65%), whereas 94% of lesions that were treated with RFA were approached percutaneously.


Ablation of SRMs is a viable strategy based on short-term oncologic outcomes. Although extended oncologic efficacy remains to be established for ablation modalities, the current data suggest that cryoablation results in fewer retreatments and improved local tumor control, and it may be associated with a lower risk of metastatic progression compared with RFA. Cancer 2008. © 2008 American Cancer Society.

Cancer of the kidney is the third most common cancer of the urinary tract and accounts for 3.5% of all malignancies.1 With an estimated 51,190 new cases occurring in 2007 and 12,890 deaths attributable to the disease, renal cell carcinoma (RCC) is the most lethal of all genitourinary tumors.1

The clinical diagnosis of RCC is radiographic, and effective imaging of the kidneys can be achieved by ultrasound, computed tomography (CT), or magnetic resonance imaging (MRI).2 Because of the increased use of diagnostic imaging for the evaluation of patients with abdominal symptoms, incidentally discovered small renal masses (SRMs) are being diagnosed with greater frequency3 and may account for up to 66% of RCC diagnoses.4 Thus, an increased incidence of RCC over the last 30 years has been associated with stage migration3 and a concurrent rise in rates of surgical intervention.5 Unfortunately, despite earlier diagnosis and treatment, cancer-specific survival (CSS) and overall survival have not improved significantly.5

Surgical resection remains the standard of care for clinically localized RCC because of the favorable prognosis associated with surgery and the relative ineffectiveness of systemic therapy. Patients who undergo radical or partial nephrectomy for SRMs (tumors classified as pathologic T1a [pT1a], tumors ≤4 cm) exhibit 5-year CSS rates in excess of 95%.6, 7 Laparoscopic approaches to partial nephrectomy have produced similarly favorable early results.8

Recently, minimally invasive ablative technologies have emerged as potential treatment options for clinically localized RCC. Effective renal cryoablation has been achieved by open, laparoscopic, and percutaneous image-guided techniques.9 Percutaneous radiofrequency ablation (RFA) has been performed successfully under ultrasound, CT, or MRI guidance.10 Although these newer nephron-sparing techniques appear to be promising, the majority of published studies are single-institution series with relatively few numbers of patients. To our knowledge, only 1 published study to date has quantitatively compared the efficacy and outcomes of these 2 ablation modalities.11

Although treatment options for low-stage RCC have expanded in recent years, their proper application and their affect on the biology of SRMs has yet to be defined fully. The objective of the current study was to analyze the combined published data regarding ablation as primary treatment of SRMs. Herein, we review the published literature and perform a comparative meta-analysis evaluating cryoablation and RFA for the management of enhancing renal masses.


Data Sources

A search of the MEDLINE database was performed through October 2007 using the National Center for Biotechnology Information PubMed Internet site to review the world literature regarding the treatment of suspected renal malignancies by means of renal cryoablation or RFA.

Study Selection

This meta-analysis was limited to series that analyzed clinically localized, sporadic renal tumors that were managed by either open, laparoscopic, and percutaneous cryoablation or RFA. Series that included only patients with hereditary or metastatic RCC were excluded. Data taken from series that reported ablation of both sporadic and hereditary renal lesions were censored to include only those that reported sporadic RCC. In addition, series that were purely technical and did not assess tumor recurrence or other oncologic endpoints were excluded. Prospective and retrospective series were included in the study, although single case reports were excluded. Multi-institutional series as well as single-institution experiences were analyzed, provided that other inclusion criteria were met. In the case of multiple series from an institution or overlapping patient cohorts with potentially redundant data, only the most recent series or the series with the largest study population was selected to avoid double counting of patients. In total, 47 studies from 45 institutions met inclusion criteria and were analyzed.

Data Analysis

Mean data pertaining to patient age, tumor size, and length of follow-up were extracted from published series. The incidence with which preablation biopsy was performed was determined from each series, and the results were collated. Pathologic data were categorized as histologically confirmed RCC, benign lesions, or unknown/indeterminate histology. The oncologic outcomes that were evaluated included local tumor progression or distant metastases. For both cryoablation and RFA, local tumor progression was defined as radiographic or pathologic evidence of residual disease after initial treatment, regardless of the time to recurrence, in accordance with the recommendations of the Working Group on Image-Guided Tumor Ablation.12 In addition, data regarding tumors that were subjected to repeat ablation sessions were extracted for analysis.

Differences in mean patient age, tumor size, and follow-up were weighted by differences in study sample size and analyzed using the Mann-Whitney test. Data for each series could not be weighted by inverse standard errors, which may have increased the efficiency of analysis, because >40% of the studies did not include variances or standard errors with their descriptive statistics. Many authors reported ranges instead of variances as a measure of variability.

Differences in malignancy rates and pathologic reporting between ablation modalities were investigated using the Fisher exact test. The rate of preablation biopsy, the use of repeat ablation, the incidence of local tumor progression, and the development of metastatic disease were analyzed similarly by using the Fisher exact test. We also investigated models of local and metastatic tumor progression using univariate and multivariate linear regression analyses. The multivariate analysis included data pertaining to treatment modality, sample size in each study, incidence of histologically confirmed RCC, incidence of unknown/indeterminate pathology, mean patient age, mean tumor size, and mean follow-up time as covariates.


Forty-seven series representing 1375 renal tumors that were treated at 45 institutions met inclusion criteria and were analyzed. The institutions that contributed published data are listed in Table 1. Table 2 lists patient and tumor characteristics for the studies that were included.

Table 1. Institutions With Published Series of Renal Tumor Ablation Included in the Current Meta-analysis Listed by Ablation Modality
  1. MCP indicates Medical College of Pennsylvania; NSW, New South Wales.

Institutions with published results on renal tumor cryoablation 
 Brigham and Women's Hospital, Boston, MassSilverman 200534
 Cedars-Sinai Medical Center, Los Angeles, CalifColon & Fuchs 200335
 Cleveland Clinic, Cleveland, OhioGill 200523
 Columbia University, New York, NYWeld 2007,36 Hruby 200637
 Geisinger Medical Center, Danville, PaDavol 200638
 Johns Hopkins University, Baltimore, MdRodriguez 2000,39 Gupta 200540
 Mayo Clinic, Rochester, MinnAtwell 200741
 MCP-Hahneman, Philadelphia, PaRukstalis 200142
 New York University, New York, NYO'Malley 200743
 Northwestern University, Chicago, IllNadler 200344
 San Raffaele Hospital, Milan, ItalyCestaru 200445
 Southern Illinois University, Springfield, IllSchwartz 200646
 Thomas Jefferson University, Philadelphia, PaLee 200347
 Jikei University School of Medicine, Tokyo, JapanMiki 200648
 University Hospital, Basel, SwitzerlandWyler 200749
 University of California, Irvine, CalifWeld 200736
 University of California, Los Angeles, CalifGore 200550
 University of Massachusetts, Amherst, MassSukvernab 200534
 University of Mississippi, Jackson, MissShingleton & Sewell 200151
 University of Virginia, Charlottesville, VaBassignani 200452
 University of Wisconsin, Madison, WisBandi 200753
 Viborg Sygehus, Viborg, DenmarkLund 200754
 Washington University, St Louis, MoWeld 200736
Institutions with published results from renal tumor radiofrequency ablation 
 Aachen University of Technology, Aachen, GermanyMahnken 200555
 Brown University, Providence, RIMayo-Smith 200356
 Case Western Reserve, Cleveland, OhioLewin 200457
 Cleveland Clinic, Cleveland, OhioHegarty 200611
 Duke University, Durham, NCWeizer 200558
 Eberhard-Karls University, Tubingen, GermanyBoss 200559
 General Hospital of Vienna, Vienna, AustriaMemarsadeghi 200560
 Hull and East Yorkshire Hospitals, Kingston Upon Hull, United KingdomRoy-Choudhury 200361
 Institut Gustave Roussy, Villejuif, Francede Baere 200262
 Johns Hopkins University, Baltimore, MdVarkarakis 200563
 Kochi University, Kochi, JapanKarashima 200664
 Kyoto Prefectural University Medicine, Kyoto, JapanUkimura 200465
 Madigan Army Medical Center, Tacoma, WashUribe 200666
 Massachusetts General Hospital, Boston, MassMcDougal 200567
 Mayo Clinic, Rochester, MinnFarrell 200368
 M. D. Anderson Cancer Center, Houston, TexAhrar 200569
 Medical University of Lodz, Lodz, PolandSalgierski 200670
 Taipei Veterans General Hospital, Taipei, TaiwanChiou 200571
 University of Pennsylvania, Philadelphia, PaClark 200672
 University of Texas Southwestern, Dallas, TexMatsumoto 200573
 University of Turin, Turin, ItalyVeltri 200674
 Wake Forest University, Winston-Salem, NCZagoria 2004,22 Zagoria 200775
 Westmead Hospital, NSW, AustraliaSabharwal & Vladica 200676
 Wilford Hall Medical Center, Lackland Air Force Base, San Antonio, TexArzola 200677
Table 2. Patient/Tumor Characteristics and Outcomes According to Ablation Modality
VariableAll Ablated LesionsCryoablationRFAP
  1. RFA indicates radiofrequency ablation; RCC, renal cell carcinoma.

No. of series472225
No. of institutions452324
No. of lesions1375600775
Mean patient age, y67.266.367.8.17
Mean tumor size, cm2.642.582.69.12
Rate of preablation biopsy, %7182.362.2<.0001
Mean duration of follow-up, mo18.722.515.8.53
Incidence of unknown or indeterminate pathology, %33.524.540.4<.0001
Incidence of RCC with known pathology, %8171.790<.0001
Rate of repeat ablation, %<.0001
Rate of local tumor progression, %<.0001
Rate of metastatic progression, %1.812.5.06

Patient and Tumor Characteristics

The mean patient age weighted by sample size for all ablated lesions was 67.2 years. The mean tumor size for all renal lesions that underwent ablation was 2.64 cm, and the mean duration of reported follow-up after ablation was 18.7 months. No statistically significant differences were observed between ablation modalities with regard to patient age (P = .17), tumor size (P = .12), or duration of postablation follow-up (P = .53). Reported approaches to renal cryoablation included laparoscopy (64.8%), percutaneous (23.2%), and open surgery (12%). Percutaneous renal tumor RFA was described for 93.7% of lesions, and laparoscopy was used for 6.3% (Fig. 1).

Figure 1.

Reported approaches to kidney tumor ablation. RFA indicates radiofrequency ablation.

Pathologic Data

Although 494 of 600 lesions (82.3%) that were treated by cryoablation underwent a preablation biopsy, only 482 of 775 (62.2%) of the lesions that were those managed by RFA were biopsied (P < .0001). Available pathologic data were classified as malignant, benign, or unknown/indeterminate histology for each ablation modality (Fig. 2). Overall, 53.9% of all ablated lesions were pathologically confirmed RCC, whereas 12.7% were benign lesions, and 33.5% had unknown or indeterminate pathology. Unknown or indeterminate pathology was reported for 147 of 600 lesions (24.5%) that were managed by cryoablation and for 313 of 775 lesions (40.4%) that were managed by RFA, a difference that was statistically significant (P < .0001). The incidence of malignancy among only those reported lesions with known histology was determined for cryoablation (325 of 453 lesions; 71.7%) and RFA (416 of 462 lesions; 90%; P < .0001).

Figure 2.

Pathologic data for renal lesions treated by cryoablation and radiofrequency ablation (RFA).

Local Tumor Progression and Metastasis

Local tumor progression was reported in 31 of 600 lesions (5.2%) after renal cryoablation and in 100 of 775 lesions (12.9%) after RFA, a difference that was highly significant (P < .0001). Of 600 cryoablated lesions, 8 lesions (1.3%) subsequently were treated by repeat ablation; by comparison, 66 of 775 lesions (8.5%) that were treated by RFA were subjected to at least 2 ablation sessions (P < .0001). Progression to metastatic disease was described in 6 of 600 lesions (1%) that underwent cryoablation and in 19 of 775 lesions (2.5%) that underwent RFA (P = .06). Overall, metastasis developed in 25 of 1375 reported lesions (1.8%) that underwent ablation.

Univariate and multivariate models analyzing local tumor progression and metastatic disease were performed using univariate and multivariate linear regressions (Table 2). Forty-three of 47 studies (91%) included complete information and were included in the regression analysis. The incidence of local tumor progression was correlated significantly with ablation modality in the univariate (P = .001) and multivariate (P = .003) regression analyses. The incidence of malignant pathology, incidence of unknown pathology, mean patient age, and mean tumor size were not associated statistically with local recurrence in either the univariate or multivariate analyses. Detection of local tumor progression was associated marginally with duration of follow-up in the multivariate analysis (P = .076). When regression analyses were performed to evaluate the incidence of metastases, no significant differences were observed for lesions regardless of ablation modality. Table 3 depicts the incidence of metastases, which was not associated significantly with any of the tested variables in univariate or multivariate regression analyses.

Table 3. Multiple Linear Regression Analysis of the Incidence of Local Tumor Progression or the Development of Metastatic Disease*
Local Tumor ProgressionMetastatic Progression
  • RCC indicates renal cell carcinoma.

  • *

    Forty-two studies had complete information available for the multivariate analyses.

Ablation modality.
Sample size of study.
Incidence of histologically confirmed RCC.
Incidence of unknown or indeterminate pathology.
Mean patient age.
Mean tumor size.
Mean follow-up time.


Surgical resection is considered the standard of care for clinically localized RCC because of the favorable prognosis associated with surgery. The 5-year CSS for patients after nephrectomy ranges from 97% and 87% for pT1a and pT1b tumors, respectively, to only 20% for pT4 disease.6 Similarly, patients who undergo partial nephrectomy demonstrate 5- and 10-year rates CSS of 92% and 80%, respectively, across all pathologic stages and 96% and 90%, respectively, for tumors <4 cm.7 Similarly, early data on laparoscopic partial nephrectomy are favorable, with a 100% CSS rate reported at a 3-year median follow-up.8 The importance of effective treatment for localized RCC is accentuated by the finding that systemic treatments have demonstrated limited success as treatment for metastatic disease and in an adjuvant setting.13, 14

Cryoablation was applied first to SRMs in 1995,15 and its mechanism of tumor destruction involves the use of rapid freeze-and-thaw cycles.9 Initial extracellular ice formation causes movement of intracellular water, alterations of intracellular pH, and protein denaturation.16 Ice formation also results in mechanical disruption of cell membranes.16 Hours and days after cryotherapy, delayed tissue necrosis occurs as injury to the local microvasculature causes diminished tissue perfusion and delayed cell death.16, 17 Effective renal cryoablation has been achieved by open, laparoscopic, and percutaneous image-guided techniques.9 The procedure can be monitored by means of a thermocouple or ultrasound in real time to confirm extension of the ice ball beyond the margins of the renal tumor.9, 18 Our analysis of the published literature demonstrates that, currently, >75% of reported renal cryoablation treatments have been performed through an open or laparoscopic approach, whereas percutaneous approaches have been used much less frequently.

RFA of an exophytic renal mass before open radical nephrectomy was described first in 1997,19 and the first report of RFA as sole treatment for a renal tumor was published in 1999.20 Although RFA has been applied using open, laparoscopic, or percutaneous approaches under ultrasound, CT, or MRI guidance,9, 10 the current literature describes percutaneous access in approximately 94% of patients who underwent renal RFA. Tumor coagulation by RFA occurs as radiofrequency waves are converted to heat, resulting in thermal tissue damage.9 High-frequency current applied to target tissues results in ionic agitation, thereby heating the tissues and resulting in denaturation of proteins and disruption of cell membranes.21 This process occurs over 4 to 6 minutes at temperatures >50°C and almost immediately >60°C.21 Because temperatures >105°C result in tissue vaporization and ineffective ablation, optimal RFA is performed at temperatures from 50°C to 100°C throughout the tumor.21 Vascular parenchyma, however, may act as a heat sink during RFA; therefore, exophytic tumors may be ablated better than central tumors in close proximity to the renal vasculature.22

In addition to their minimally invasive nature, another primary benefit of ablative therapy for renal tumors is the potential for preservation of renal function. However, to our knowledge, few studies to date have examined the effects of kidney ablation on renal function. Gill et al examined 56 patients with 3-year follow-up after renal cryoablation and reported preoperative and postoperative serum creatinine levels of 1.2 mg/dL and 1.4 mg/dL, respectively.23 In 10 patients who had solitary kidneys in that series, the mean preoperative and postoperative serum creatinine levels were 2.2 mg/dL and 2.6 mg/dL, respectively, and 13 patients with baseline renal insufficiency demonstrated levels of 3 mg/dL and 2.7 mg/dL, respectively.23 In another series of 14 patients who underwent cryoablation in a solitary kidney, no adverse effect on renal function was noted, although 3 lesions required repeat treatment for incomplete ablation.24 A series that examined the effects of RFA on 16 patients with a solitary kidney demonstrated a decrease in the mean glomerular filtration rate from 54.2 mL per minute per 1.73 m2 preoperatively to 47.5 mL per minute per 1.73 m2 at last follow-up.25 Similarly, Jacobsohn et al studied 16 patients who underwent RFA in a solitary kidney and demonstrated a 13.3% change in creatinine clearance within 1 week after ablation and a 9.1% change at a mean follow-up of 15.3 months with 1 patient developing chronic renal failure.26 Thus, although a few studies have examined the functional impact of ablation, none have directly compared the ability of cryoablation and RFA to preserve renal function. The ability to spare nephrons maximally remains a careful balance against the possibility of insufficient tumor destruction.

Radiographic follow-up after cryoablation or RFA is currently the most common means of assessing treatment effect.12 Enhancement on postcontrast imaging is considered evidence of incompletely treated disease.27 Some centers have performed biopsy after ablation to assess for viable disease, whereas others have relied solely on radiographic evaluation.12, 28, 29 Although the presence of macroscopically viable disease may be detectable on follow-up imaging immediately after ablation, microscopic disease may require a longer duration of surveillance to become apparent. This may explain recent data suggesting that viable tumor may be present on postablation biopsy despite lack of radiographic enhancement.23 The Working Group on Image-Guided Tumor Ablation has used the term local tumor progression to indicate incomplete tumor destruction regardless of the time it takes for enduring disease to become evident clinically.12 Therefore, in our meta-analysis, we considered any lesion with evidence of local disease persistence after ablation as local tumor progression, regardless of the time to reappearance. However, it should be noted that many local recurrences after ablation have been retreated successfully by subsequent ablation. Thus, the ultimate rate of treatment failure after salvage ablation remains to be fully defined. Furthermore, it is likely that ablation techniques have undergone refinement with increased experience; therefore, published series may not truly reflect contemporary results.

To our knowledge, only 1 published study has compared the efficacy and outcomes of these 2 ablation modalities quantitatively. In 2006, investigators at the Cleveland Clinic retrospectively compared their experience in laparoscopic cryoablation of 179 renal lesions with 81 tumors treated by RFA at that institution.11 Radiographic evidence of disease persistence was identified in 1.8% of lesions after cryoablation and in 11.1% of lesions after RFA.11 Thus, the rates of local tumor progression reported in that institutional series were similar to the rates determined in the current meta-analysis. The CSS rate was 98% after cryoablation and 100% after RFA at a median 1-year follow-up in this cohort.11

When comparing the rates of local disease persistence among treatment modalities, it is important to consider that nonuniform criteria may have been used to define recurrence after ablation. Although the majority of series used contrast-enhanced imaging to determine treatment effect, the definition of ablative success has been called into question by studies that have demonstrated viable tumor on postablation biopsy despite a lack of enhancement on imaging.29 Perhaps the true rate of local disease progression could be determined more accurately if biopsy were included routinely in postablation surveillance protocols.

The context in which the technical success of renal ablation is evaluated must include consideration of the emerging body of data regarding the observation or active surveillance of SRMs in elderly or infirm populations. Although published series addressing the natural history of small renal tumors under active surveillance report some variability in the clinical behavior of observed SRMs (growth rates of 0.09-0.86 cm per year), a meta-analysis of clinically localized tumors determined an overall median growth rate of 0.28 cm per year for observed lesions across multiple published series.2 Moreover, it has been reported that from 26% to 33% of enhancing renal masses demonstrate zero net growth when observed over a median of 29 months.30 It is noteworthy that only 1% of lesions reported in the active surveillance literature have demonstrated progression to metastatic disease in the absence of treatment.2 This information raises the possibility of an over-treatment bias for SRMs and suggests that treatment may not have an impact on the biologic potential of many lesions.31 Thus, the indolent nature of certain SRMs must be considered when analyzing the treatment efficacy of ablative technologies.

In this meta-analysis, we demonstrate that cryoablation and RFA are technically viable approaches to the management of the SRMs. However, our data indicate that patients who had lesions treated by cryoablation had a significantly lower incidence of local tumor progression after initial treatment—an association that remained significant on univariate and multivariate analyses. Although the incidence of progression to metastatic disease was low regardless of ablation modality, lesions treated by RFA exhibited a greater incidence of metastatic progression, whereas the difference was only marginally significant and was not apparent on regression analysis. It is not evident from these data whether this difference in the ability to produce complete tumor destruction is a function of inherent differences in the ablation technologies themselves or, rather, whether the laparoscopic approach that is used more frequently for cryoablation provides a greater propensity for effective tumor treatment than percutaneous approaches. Indeed, lesions treated by cryoablation most often were approached laparoscopically (65%), whereas 94% of lesions treated by RFA were approached percutaneously. This information further reflects the finding that urologists have used cryoablation technology preferentially, whereas renal RFA has been performed chiefly in the setting of interventional radiology suites.

Minimally invasive ablation treatment options generally have been performed selectively on older patients with smaller tumors; thus, a selection bias potentially may confound data related to their use. In addition, series of ablated lesions tend to include shorter posttreatment follow-up compared with published series of tumors managed by surgical excision or active surveillance.2, 32 Furthermore, the absence of known pathology remains a confounding factor when attempting to compare oncologic outcomes for lesions treated by cryoablation or RFA. This category of tumors with unknown pathology certainly includes a number of histologically benign lesions; thus, measures of treatment efficacy may be overestimated.9 Although the weaknesses of meta-analyses are well recognized,31, 33 these data underscore the need for long-term, prospective, randomized trials to determine the proper application and biologic implications for ablation of SRMs. Furthermore, this study emphasizes the importance of thorough consideration of all treatment options for patients with SRMs, including proper discussion of risks, benefits, and expected outcomes after surgical excision, ablation, and active surveillance.


The current data illustrate that ablative technologies are viable strategies for SRMs masses based on short-term oncologic outcomes. However, extended oncologic efficacy remains to be established for ablation modalities. Among ablation modalities, data suggest that cryoablation may result in significantly lower rates of local tumor progression compared with RFA, although no statistical differences were detected in the incidence of progression to metastatic disease.