The widespread use of abdominal cross-sectional imaging has contributed to a significant stage migration toward the diagnosis of incidental small renal masses (SRMs) and has prompted an evolution in therapy from radical nephrectomy to nephron-sparing surgery, thermal ablation (cryotherapy or radiofrequency ablation [RFA]), and observation.1–3 We congratulate Kunkle and Uzzo on attempting to provide a timely meta-analysis of existing published literature on cryoablation and RFA.4 Their article specifically comprises a subset of studies from their recent publication comparing surgery, thermal ablation, and observation for SRMs in an attempt to define respective oncologic outcomes.5 However, as in the earlier meta-analysis by the same group, although the authors acknowledge multiple shortcomings in the data, they nevertheless present a conclusion suggesting the superiority of cryoablation over RFA. This conclusion is potentially misleading; therefore, it is imperative to properly contextualize their observations and findings.
Most important, their conclusion is problematic in that it is a compilation of multiple single-institution series, all of which are plagued by their own inherent limitations and biases. These limitations are compounded collectively by a meta-analysis design. Zwahlen et al observed that the quality of component trials is crucially important and that, if raw materials are flawed, then findings from a review of the materials also may be compromised.6 When specifically considering cryoablation and RFA, such drawbacks include (but are not limited to) 1) differences in lesion selection, 2) bias for a particular ablative modality, 3) surgical approach (laparoscopic vs percutaneous), 4) type of anesthesia (general vs conscious sedation), and 5) follow-up regimen/criteria to assess treatment success or failure. Such a heterogeneous array of variables unfortunately contributes to disparate reports of therapeutic efficacy in the literature. We recently itemized these drawbacks for the authors in a letter to the editor of the Journal of Urology7 regarding the earlier meta-analysis.5 Because they are repeated in the current meta-analysis, it is important once again to emphasize the weaknesses of this meta-analysis for the readers of Cancer.
Kunkle and Uzzo report that cryoablation has lower retreatment rates and improved local control compared with RFA, suggesting that RFA is inferior to cryotherapy.4 The basis of this observation is misleading and warrants further discussion to understand the comparison better. The fundamental crux of the problem is that 1) an incomplete primary thermal ablation is not always clinically synonymous with a local tumor recurrence, and 2) outcomes observed between cryoablation and RFA cannot confidently distinguish between the efficacy of the ablation technology itself and the technique employed. The authors acknowledge these 2 critical points in their discussion, yet they fail to temper their conclusions based on these obvious limitations.
In particular, characterizing an incomplete primary ablation as a cancer recurrence (or even as a treatment failure) fails to appreciate the paradigm shift provided by ablative therapy for SRMs. This is a particularly key (and often misunderstood) point when considering complete primary ablation rates between different ablative techniques. Compared with cryotherapy, in which 77% of ablation procedures in the current meta-analysis were performed laparoscopically or as open procedures, 94% of RFA procedures were performed percutaneously, often as an outpatient procedure. Previously, we emphasized7 (and Kunkle and Uzzo briefly acknowledge) that percutaneous RFA under nonvisual image guidance (computed tomography or ultrasound) is easily repeatable and allows for caution by the treating physician to avoid over-aggressive therapy that can contribute to complications and unnecessary reduction in the glomerular filtration rate in patients who already have compromised renal reserves. Thus, physicians can be more judicious with needle placement and ablation zone dimension, knowing that an acute repeat ablation (the paradigm shift) for residual tumor is not because of true ‘local tumor progression’ or ‘recurrence’ and remains less morbid than surgery (laparoscopic or open). Most important, performing a repeat percutaneous cryoablation or RFA because of an initial incomplete treatment does not increase the risk of metastatic progression, as demonstrated even in the current meta-analysis. Conversely, a direct-vision open or laparoscopic approach for cryotherapy (or even RFA) is not easily repeatable and thus warrants a more aggressive strategy as well as an inpatient hospitalization. These observations and concepts are reinforced by evaluating the laparoscopic ablative experience reviewed in a contemporary multi-institutional series from centers of excellence. Matin et al demonstrated that the incomplete ablation rate for laparoscopic RFA is only 2.5% and is comparable to that for laparoscopic cryoablation (3.5%).8 Similarly, a recent series of percutaneous cryoablation (from an institution with extensive ablation experience) noted an initial ablation success rate of 83% which parallels that reported with most percutaneous RFA series.9 Clearly, pooling 2 distinct surgical techniques (laparoscopic vs percutaneous) when comparing cryotherapy and RFA significantly confounds the interpretation of local tumor recurrence rates, inflates ‘local tumor progression’ rates for percutaneous treatment (RFA), and is analogous to comparing ‘apples and oranges.’ It is unfortunate that Kunkle and Uzzo did not perform the necessary subset analyses comparing percutaneous RFA with percutaneous cryoablation and comparing laparoscopic RFA with laparoscopic cryoablation to address this issue better.
In addition, as a further caution regarding the reliability of meta-analyses that rely on disparate retrospective studies, the impact of seemingly subtle differences in percutaneous ablation techniques on clinical results also must be emphasized. One such variable that illustrates this point is the choice of anesthesia; namely, general sedation versus conscious sedation. General anesthesia optimizes patient tolerance, allows for greater control of respiratory motion during probe placement, and theoretically improves the accuracy of targeting the lesion. In fact, emerging data10, 11 suggest that percutaneous ablation performed under general anesthesia has a higher ‘initial ablation success’ rate than the same procedure performed under conscious sedation. Specifically, both the Mayo Clinic experience10 with percutaneous cryoablation and the combined University of Texas Southwestern Medical Center and University of Miami percutaneous RFA experience indicated a successful initial ablation rate of 97% (unpublished data). These initial ablation rates are comparable with data from laparoscopic ablation series8 and are significantly higher than the 80% primary ablation success rate reported by a high-volume center with extensive experience in conscious sedation (though their long-term oncologic outcomes were excellent).11, 12 Once again, pooling such differences and interpreting an incomplete ablation as ‘local tumor progression’ yields numerous inaccuracies that are compounded by a meta-analysis.
In their discussion Kunkle and Uzzo emphasize 1 retrospective study that compares the results from laparoscopic cryotherapy and the results from percutaneous RFA and highlight it as an example supportive of the findings of their meta-analysis. In that study, Hegarty et al13 reported radiographic evidence of disease persistence in 1.8% of lesions after cryoablation and in 11.1% of lesions after RFA. Unfortunately, that study highlights many concerns we have described already and cautions authors and readers alike concerning the judicious interpretation of data in the literature. Namely, reviewing patient demographics and tumor characteristics from that article (see Tables 1 and 2 in Hegarty et al13) reveals that RFA was used in a patient cohort with 7 times more solitary renal remnants (23% vs 3%), 6 times as many central tumors (37% vs 6%), and 15% fewer normal contralateral kidneys (45% vs 60%) compared with the cryotherapy group. Such observations suggest a significant selection bias for the ablative modality. It is likely that RFA was offered to patients who were poor candidates for laparoscopic cryoablation and that the treating physician would be more judicious with needle placement and ablation size in such a percutaneous RFA group.
Short of repeating this meta-analysis and controlling for percutaneous versus laparoscopic technique, urologist versus radiologist performing the procedure, and conscious sedation versus general anesthesia as well as tumor entry criteria, we propose a more accurate conclusion from the analysis: Differences in ablation technique and patient selection may influence tumor retreatment rates and disease recurrence, precluding any conclusion on the superiority of one ablation technology over the other. Prospective randomization is required.
Finally, Kunkle and Uzzo appropriately emphasize in their discussion that the published experience with surveillance of SRMs suggests an over-treatment bias given the indolent nature of certain SRMs. Nevertheless, we all are aware that many patients, and even referring physicians, will insist on treatment. In these patients, for whom any treatment (partial nephrectomy or ablation) admittedly may not have an impact on the biologic potential and natural history of the tumor, offering an approach that is percutaneous, efficacious, cost-effective, minimally morbid, and maximally preserves renal function is desirable. It appears that both percutaneous cryoablation and RFA can achieve this in appropriately selected tumors.10, 14
In summary, the nonurologic oncology reader who is not familiar with renal ablation techniques and the literature can be misled by the current meta-analysis and may draw incorrect conclusions regarding the promise of cryoablation and RFA. A better representation of the oncologic efficacy of thermal ablation would be obtained from prospective series with uniform patient selection, lesion targeting, and surgical technique strategies.7 Although these studies currently are lacking, a meta-analysis (for all the reasons outlined above) is not a reliable substitute. Finally, incomplete primary percutaneous ablation (whether cryoablation or RFA) should not necessarily always be viewed as a ‘local cancer recurrence’ or a ‘treatment failure.’ It simply reflects a different approach to treating SRMs whereby a minimally morbid, outpatient, second potential treatment is accepted as a reasonable alternative compared with potentially over-aggressive primary surgery. Such an approach has produced excellent intermediate oncologic outcomes12, 15, 16 and low complication rates16 with excellent preservation of renal function outcomes.14