Image-guided percutaneous renal cryoablation: preoperative risk factors for recurrence and complications


Manish A. Vira, The Arthur Smith Institute for Urology, 450 Lakeville Road, New Hyde Park, NY 10040, USA. e-mail:


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

Given that percutaneous cryoablation (PCA) is a relatively new procedure, there are few studies published on this treatment with almost no long-term follow-up. The percutaneous approach, while not the first choice treatment for RCC, may be most appropriate for older patients with several comorbidities as it offers less invasive outpatient management of small renal masses (SRMs). It is therefore important to measure procedural outcomes noting rates of complications and reasons for treatment failure or recurrence.

To our knowledge, this is the first paper applying the R.E.N.A.L nephrometry scoring system to PCA of SRMs. The study adds insight into procedural outcomes from this treatment. Little has been published on this treatment strategy, but it has been increasingly considered for patients who are not candidates for traditional surgical approach. It is important to study and establish the outcomes of all treatments used by physicians. It is also necessary to understand treatment complications – how and why they occur – and seek reasons for treatment failure and recurrence. This allows physicians to choose the best management for each individual patient to improve outcomes.


  • • To investigate the value of the R.E.N.A.L nephrometry scoring system in predicting treatment success for image-guided percutaneous cryoablation (PCA).


  • • The study included 139 patients with renal masses treated with PCA.
  • • Preoperative computed tomography or magnetic resonance images were reviewed by a urology resident.
  • • The primary endpoint variable was incomplete treatment or tumour recurrence.
  • • R.E.N.A.L. scores were categorized into low (4–6), moderate (7–9), and high (10–12).
  • • Logistic regression analysis was conducted to predict tumour recurrence. Additional variables collected included age at surgery, American Society of Anesthesiologists score, lesion size, skin-to-tumour distance, skin-to-hilum distance, and number of treatment cryoprobes.


  • • At a median follow-up of 24 months, there were 10 tumour recurrences (six moderate and four high R.E.N.A.L. score categories). Nephrometry score and number of probes used were not associated with recurrence (odds ratio [OR] 1.02, P= 0.9 and P= 0.53, respectively).
  • • The tumour distances for patients with recurrence and no recurrence were 10.8 cm and 8.5 cm, respectively (P≤ 0.05), the skin-to-tumour distance was associated with treatment failure (OR 1.24, P= 0.015); for each unit increase in the mean value, patients were 1.5 times more likely to have a tumour recurrence (95% confidence interval [CI] 1.04–1.72).
  • • The model that best predicted complications included the number of probes used (P= 0.002) and R.E.N.A.L. score (OR 1.35, P= 0.027). For each additional probe used, patients were twice as likely to have complications (OR 1.98, 95% CI 1.28–3.05). With each unit increase in R.E.N.A.L. score, patients were 1.5 times more likely to experience a complication (OR 1.49, 95% CI 1.05–2.11).


  • • An increased skin-to-tumour distance is associated with a higher risk of treatment failure after PCA.
  • • Furthermore, an increase in both R.E.N.A.L nephrometry score and number of probes used was associated with an increased risk of complications after PCA.
  • • The R.E.N.A.L. nephrometry score as a measure of tumour complexity was not associated with tumour recurrence.

percutaneous cryoablation


odds ratio


small renal mass


American Society of Anesthesiologists


body mass index.


The current widespread use of cross-sectional imaging has increased the incidence of small renal masses (SRMs) [1], and the standardized classification of renal tumours has the potential to improve anatomic comparison of datasets. The R.E.N.A.L. nephrometry scoring system was developed as a standardized scoring system to quantify renal mass characteristics based on cross-sectional imaging [2]. Recent reports in the literature have validated the nephrometry scoring system as a measure of tumour complexity that also correlates with surgical outcomes of partial nephrectomy [3–7]. With the added emphasis on nephron-sparing approaches for the treatment of SRMs, ablative methods such as percutaneous cryoablation (PCA), are increasingly used and may be the preferred treatment choice in patients with multiple tumours or those with increased surgical risk secondary to medical comorbidities. To our knowledge, the nephrometry scoring system has not been validated for renal cortical tumours undergoing PCA.

Percutaneous cryoablation for SRMs has been shown to provide intermediate local oncological control of 70–100% for up to 2 years [8]. Features of the procedure and the mass itself may predict treatment failure and complications [9]; therefore, we sought to evaluate the association of the R.E.N.A.L nephrometry scoring system with treatment success and complications of PCA. Furthermore, in addition to the variables specified in the scoring system, we included secondary measures of skin-to-tumour distance, skin-to-hilum distance, and number of treatment probes as independent variables. Specifically, we aimed to correlate renal tumour nephrometry score, cross-sectional imaging measures and number of treatment probes with post-procedural rates of tumour recurrence and complications.


After receiving institutional review board approval we reviewed records of all patients who underwent PCA at our institution between December 2005 and May 2011.

The R.E.N.A.L. nephrometry scores were calculated for all patients with available preoperative CT or MRI. Images were reviewed and scoring was performed by a single urology resident. As previously described, points were assigned according to the tumour diameter (1 point if ≤4 cm, 2 points if >4 and <7 cm, 3 points if ≥7 cm), exophytic properties (1 if ≥50% exophytic, 2 if <50% exophytic, 3 if completely endophytic), nearness to the collecting system (1 if ≥7 mm, 2 if <7 mm but >4 mm, 3 if ≤4 mm), and polar location (1 if completely upper or lower pole, 2 if crosses the polar line, 3 if entirely interpolar or crosses axial renal midline). Additionally, a descriptive suffix was assigned for anterior (a) vs posterior (p) location, or for hilar tumours (h), which were defined as those that touched the main renal artery or vein.

In addition to the other predictive variables collected, the skin-to-tumour and skin-to-hilum distances were measured and analysed. These measures were taken at 0, 45, and 90° on cross-sectional imaging (Fig. 1). The final calculated distances comprised the means of these three measurements.

Figure 1.

CT measurement of skin-to-tumour distance.

Demographics, peri-, and postoperative data were analysed to determine the presence of complications and recurrences.


The primary endpoint outcome variables were peri-operative complications and tumour recurrence during the follow-up period. R.E.N.A.L. scores were categorized into low (4–6), moderate (7–9), and high (10–12) as previously described by Kutikov et al. [2]. Descriptive statistics (mean, sd, median, interquartile range, minimum, maximum, frequencies and percentages) were used to describe age, gender, tumour size, polarity, pathology and American Society of Anesthesiologists (ASA) score. SAS statistical software was used for statistical analysis. The frequency distribution of complications and R.E.N.A.L. categorical scores and recurrences were compared using the chi-squared test. Pearson correlations were performed between number of probes and lesion size and between body mass index (BMI) and skin-to-tumour and skin-to-hilar distances. Logistic regression analysis was conducted to predict complications and recurrence of tumours. Backward elimination logistic regression was used and the following variables were entered as possible predictors of recurrence and complications: R.E.N.A.L. score, lesion size (cm), mean skin-to-tumour distance, mean skin-to-hilar distance, ASA, age at surgery, and number of probes used. A predetermined P value of <0.05 was considered to indicate statistical significance.


All PCA procedures were performed in a hospital CT suite under general anaesthesia. The Endocare cryoablation system (HealthTronics® Inc., Austin, TX, USA) was used for all procedures. The PCA technique used has been described previously [10].

Under initial local anaesthesia, the cryoprobes were placed into the target renal tumour under CT guidance. Probe position was intermittently controlled with CT imaging. The total number of probes placed was based on tumour size with the goal of establishing an ice ball that would comprehend the tumours in their entirety for complete ablation.

Before initiating the ablation process, an 18-gauge core biopsy was obtained from the tumour under CT guidance. In the case of tumours adjacent to segments of bowel, a 5-F catheter was placed along the margin of the tumour and sterile 0.9% normal saline was injected, displacing the bowel segment away from the targeted renal tumour. Sufficient fluid was introduced to provide a safe distance, at least 2 cm from the tumour edge. Each tumour received a two-cycle treatment of freezing and thawing. To monitor ice ball growth, non-contrast CT images were obtained approximately every 2 min during the freezing portions of the cycle. There was no standardized freezing time in our protocol. The length of freezing time was based upon the growth of the ice ball relative to the tumour. Because complete cell death occurs ∼3 mm inside the edge of the ice ball, the goal was to extend the ice ball 5 mm beyond the tumour margin during both freezing portions of the ablation treatment.

Technical success was defined as an extension of the ice ball beyond the tumour margin and post-ablation images showing no contrast enhancement in the area of the original tumour within 6 months of the initial procedure.


Initial imaging was obtained within 24 h of the procedure to document technical success and to assess potential complications. In patients with adequate renal function and no allergy to iodinated contrast, CT both with and without i.v. contrast was obtained immediately after cryoprobe removal at the end of the procedure.

Additional follow-up CT or MRI imaging was obtained at 3, 6 and 12 months after ablation and yearly thereafter. When renal function and absence of allergy allowed, i.v. contrast medium was administered. We defined local tumour recurrence as new enhancement in the ablated tumour or a progression in tumour size >6 months after the procedure.

Intraoperative as well as postoperative complications within 30 days of initial procedure were recorded. Complications were classified according to the modified Clavien grading system [11].


A total of 162 patients were treated with image-guided PCA for a renal mass during this period. We performed a retrospective review and identified 139 patients with available preoperative CT or MRI. Patient demographic and clinical characteristics are listed in Table 1.

Table 1. Patient demographics and clinical data
No. of patients139
Mean (range) age, years70 (37–91)
Gender [male: female]79:21
ASA score, % 
 I, II46
Mean (range) BMI, kg/m227.5 (15–43)
Mean (range) lesion size, cm2.4 (1–6.5)
Mean (range)R.E.N.A.L. nephrometry score6.5 (4–11)
 Low, %51
 Moderate, %40
 High, %9
Mean (range) skin-to-tumour distance, cm8.6 (4–20)
Mean (range) skin-to-hilum distance, cm11 (7–19)
Biopsy, n/N (%) 
 Hispathology, %90/139 (65)
  Benign (oncocytoma)11
Recurrence, n (%)10 (7)

At a median follow-up of 24 months, there were 10 tumour recurrences (six moderate and four high R.E.N.A.L. score categories). There were no significant differences among nephrometry subcategories in patients with recurrences (odds ratio [OR] 1.02, P= 0.922). Additionally, tumour size (P> 0.6) and number of probes used (P= 0.53) were not significantly associated with tumour recurrence; however, our analysis did show that skin-to-tumour distance significantly correlated with risk of recurrence (OR 1.25, P= 0.015). For each cm increase in distance, patients were 1.5 times more likely to have a tumour recurrence (OR 1.67; P= 0.02, 95% CI 1.04–1.72). The mean skin-to-tumour distances for patients with recurrence and no recurrence were 10.8 cm and 8.5 cm, respectively (P< 0.05).

Complications are shown in Table 2. Overall, there were 18 (12.9%) complications, which were mostly Clavien grade I and II. There were significant differences in the observed complication rates among the low- (0%), moderate- (35%), and high-complexity (100%) groups (P< 0.001). Nephrometry score was associated with a higher risk of complications (OR 1.35, P< 0.027). Additionally, tumour size was associated with a higher risk of peri-operative complications (OR 2.45, P< 0.001). As a dichotomized set, 12 (27%) patients with large tumours (>4 cm) had complications while only seven (6%) with small tumours had complications. As expected, the number of probes used significantly correlated with mass size (P< 0.001). The model that best predicted complications included the number of probes used (P= 0.002) and R.E.N.A.L. score (P= 0.03). For each additional probe used, patients were twice as likely to have complications (OR 1.98, 95% CI 1.28–3.05) and with each unit increase in R.E.N.A.L. score, patients were 1.5 times more likely to experience a complication (OR 1.49, 95% CI 1.05–2.11). Skin-to-tumour distance was not associated with complications (OR 0.91, P= 0.377).

Table 2. Complications of PCA, classified according to the modified Clavien system
Clavien gradeComplication n (%)
1Perinephric haematoma9 (6.4)
Atrial fibrillation1 (0.7)
Pneumothorax, conservative management5 (3.6)
2Intestinal perforation, conservative management3 (2.2)
Total 18 (12.9)

Finally, we assessed whether an increase in BMI was associated with complications. BMI data were available for 85 patients. The mean (sd) BMI was 27.3 (6.87) and the median was 26.4. BMI was not associated with complications (OR 1.03, P= 0.562).


Much of the literature on the R.E.N.A.L. nephrometry scoring system has thus far focused on peri-operative outcomes during and after partial nephrectomy. To our knowledge, this is the first report validating the scoring system for patients undergoing PCA. Owing to its lower morbidity and its minimally invasive nature, PCA has become an increasingly popular option for managing SRMs. This enthusiasm is supported by reports of successful functional and intermediate oncological outcomes [12–14]. In their comparison of PCA and laparoscopic cryoablation for the treatment of solid renal masses, Hinshaw et al. [15] reported that PCA had a 100% disease-specific survival and a 10% rate of residual/recurrent disease, which compares favorably with laparoscopic cryoablation. Similar findings have been established in more recent series comparing PCA with laparoscopic cryotherapy [13,16,17].

Previous reports in the literature report a complication rate of up to 25% after PCA with major complications of up to 6% [9,18,19]. Potential complications after PCA include post-procedural haemorrhage, flank pain, adjacent organ injury, and other minor complications such as short-term ileus, nausea/vomiting and haematuria [18]. Case reports of bowel, splenic, pleural injury and renal intestinal fistula have been reported after PCA [20–23]. Our findings suggest that an increase in both the R.E.N.A.L. nephrometry score and number of probes used are associated with an increased risk of post-PCA complications. The number of cryoablation probes used has previously been shown to increase the risk of post-procedural complications [9], but this is the first report correlating tumour complexity with complication rates after PCA. Multivariate analysis demonstrates that a R.E.N.A.L nephrometry score of ≥9 (P= 0.02) is an independent predictor of complications. Atwell et al. [20], in the largest series in the literature (n= 573), concluded that a higher complication rate was associated with tumour size and central location. The authors found that larger and more central tumours required more cryoprobes which they hypothesized ultimately contributed to the increased incidence of post-procedural complications. Similarly, this present study showed that an increase in R.E.N.A.L. score (tumours larger in size and anatomically closer to the hilum) and use of additional cryoprobes, were predictive of adverse procedural outcomes.

With respect to tumour recurrence, while the size of the tumour, the R.E.N.A.L nephrometry score and number or probes used was not predictive, skin-to-tumour distance was significantly associated with a probability of recurrence. Overall, at a median follow-up of 24 months, 10 patients experienced a recurrence for a total treatment failure rate of 7.2%. This compares favourably with other reports in the literature. Of 116 patients treated with PCA for single renal masses, Duffey et al. [22]reported 2- and 5-year recurrence-free survival rates of 83 and 77% in biopsy-proven RCC. Atwell et al. [24] examined their tumour recurrence in a series of 91 patients and noted a successful ablation rate of 96% with an overall local control rate of 95% at a mean follow-up of 26 months.

In the present series, skin-to-tumour distance, and not the R.E.N.A.L. nephrometry score, was found to be the only variable predictive of tumour recurrence after PCA. For each cm increase in the mean skin-to-tumour distance patients experienced a 1.5-fold increase in tumour recurrence (95% CI 1.04–1.72). The reasons for this finding are not entirely clear but may be related to increased inaccuracy in proper probe placement or perhaps tumours that were more difficult to treat. In their series, Strom et al. [16] noted a higher recurrence rate with upper pole lesions or endophytic lesions, i.e. perhaps lesions that were more difficult to reach. To our knowledge, the finding that skin-to-tumour distance is associated with a higher rate of treatment failure has not previously been reported. Clearly, this variable needs to be validated in other cohorts of patients to confirm our findings.

The present series has certain notable limitations. First, the analysis was performed retrospectively introducing recall bias, especially with respect to peri-procedural complications. Certainly, a prospective evaluation would probably more accurately capture all complications associated with PCA. Given that the majority of tumours were <4 cm producing low numbers of high renal scores, this may have contributed to additional bias. The current cohort is from a single institution and therefore our findings may not be generalized to all patients undergoing PCA. We were not able to review directly the follow-up imaging studies after PCA; rather we relied on a radiologist reporting whether or not there was measurable enhancement in the treatment bed to define recurrence. Finally, our follow-up is intermediate and may represent incomplete evidence of all patients who have had a tumour recurrence.

In conclusion, with its acceptable functional and intermediate oncological results, PCA has become an increasingly recommended approach to the treatment of SRMs. The present results suggest that rather than tumour size or complexity, the skin-to-tumour distance may be the most important variable in determining the risk of treatment failure. Furthermore, the results indicate that tumour complexity and the number of probes used are important determinants of the risk of perioperative complications. These results may assist urologists in identifying the best patients for PCA and in counselling patients as to the risk of the procedure.


None declared.