SEARCH

SEARCH BY CITATION

Keywords:

  • renal tumours;
  • percutaneous image-guided ablation;
  • percutaneous cryoablation

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Methods
  5. Statistical Analysis
  6. Results
  7. Discussion
  8. Conflict of Interest
  9. References

Objective

  • To evaluate the technical and oncological efficacy of an image-guided cryoablation programme for renal tumours.

Patients and Methods

  • A prospective analysis of technical and radiological outcomes was undertaken after treatment of 171 consecutive tumours in 147 patients.
  • Oncological efficacy in a subset of 125 tumours in 104 patients with >6 months' radiological follow-up and a further subset of 62 patients with solitary, biopsy-proven renal carcinoma was also analysed.
  • Factors influencing technical success, as determined by imaging follow-up, and complication rates were statistically analysed using a statistics software package and logistic regression analyses.

Results

  • No variables were found to predict subtotal treatment, although gender (P = 0.08), tumour size of >4 cm (P = 0.09) and central location of tumour (P = 0.07) approached significance.
  • Upper pole location was the single variable that was found to predict complications (P = 0.006).
  • Among the 104 patients (125 tumours), radiologically assessed at ≥6 months and with a mean radiological follow-up of 20.1 months, we found a single case of unexpected late local recurrence.

Conclusion

  • Percutaneous image-guided cryoablation, at a mean of 20.1 months' follow-up, appears to provide a safe and effective treatment option with a low complication rate.
  • Anteriorly sited tumours should not be considered a contraindication for percutaneous image-guided cryoablation.

Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Methods
  5. Statistical Analysis
  6. Results
  7. Discussion
  8. Conflict of Interest
  9. References

In recent years there has been a significant increase in the incidental detection of small renal tumours during cross-sectional imaging, often undergone for unrelated clinical symptoms [1]. Whilst nephron-sparing procedures, such as open and laparoscopic partial nephrectomy, have become the clinical standard for the treatment of T1a (<4 cm) renal lesions, these procedures remain onerous for patients, a substantial proportion of whom will have baseline renal impairment at the time of tumour detection [2]. These procedures also incur a small but significant risk of haemorrhage, urinary leakage and warm ischaemic injury, with the likelihood of contributing further to diminished renal function [3].

There is evidence that many smaller tumours can behave in a relatively indolent fashion with mean growth rates of 3 mm/year [4]; however, T1a renal tumours express a range of behaviours and some have the potential for malignant dissemination. Surveillance imaging of smaller tumours in older patients has been advocated but is not without its own costs and psychological impact on the patient. An effective, minimally invasive technique for these tumours is therefore a highly attractive therapeutic option. The present study describes the technical and radiological outcomes of the first consecutive 147 patients who took part in a percutaneous cryoablation programme between May 2007 and May 2012. To ensure absolute clarity about oncological outcomes, a subset analysis of solitary, sporadic, biopsy-proven RCCs with ≥6 months radiological follow-up was also performed.

Patients and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Methods
  5. Statistical Analysis
  6. Results
  7. Discussion
  8. Conflict of Interest
  9. References

This prospective study was approved by our institutional Research Ethics Panel and the requirement for informed consent waived. Procedural details refer to all tumours treated, including occasional cases, in the setting of concurrent metastases and with previous contralateral renal malignancy. The whole cohort includes patients with multiple tumours, as well as those with solitary, sporadic RCCs and no previous history of renal malignancy.

The population in the technical outcomes part of the study consisted of all patients treated by image-guided, percutaneous renal cryoablation between May 2007 and May 2012. Patients were discussed at a urological multidisciplinary team meeting, and were selected for ablation dependent upon significant renal impairment (usually grade 2–4, i.e. estimated GFR 15–89 mL/min/1.73 m2), in the event of solitary kidneys or in the setting of significant comorbidities, where partial nephrectomy might have represented an undue surgical risk. The radiological imaging was reviewed by an interventional radiologist (D.J.B. or T.J.B.) who deemed whether the tumour was amenable to image-guided cryoablation.

The population in the oncological outcome analysis part of this study consisted of a subset of patients who had a solitary, biopsy-proven RCC, no metastatic disease at the time of initial treatment, and radiological follow-up of ≥6 months. This subset was chosen because the currently reported experience of renal ablation procedures is often confused by the presence of a diverse population of treated tumours, sometimes not all of proven malignancy, and occasionally by patients with multifocal disease or who have previously undergone nephrectomy for contralateral malignancy. Table 1 shows brief demographic details of this group of 62 patients, alongside a contemporary group of 42 patients, some with multiple tumours (where a single representative tumour was biopsied, some with tumours other than RCCs) and a further 43 patients who had not yet reached 6 months' radiological follow-up. Comorbidity played a part in the selection of patients for ablative therapy, and mean Charlson index score was 7.15 (patients aged ≥ 80 years were given an age score of 5) with a median of 7 and a range of 2–14.

Table 1. Baseline characteristics of 147 study participants,153 procedures and 171 tumours
 Cases of histologically confirmed solitary RCCs with ≥6 months follow-upCases with ≥6 months follow-up with multiple tumours, tumours not RCCs or without histologyCases with <6 months follow-up
  1. *Including one tumour in right moiety of a horseshoe kidney.

Total participants (=147)624243
Mean age, years67.667.066.0
Males, n (%)41 (66)30 (72)30 (70)
Pre-existing metastases, n032
Procedures, n624843
Mean (median) radiological follow-up, months17.7 (16.0)23.2 (20.2)1.6 (0.6)
Radiological follow-up range, months6.3–40.56.4–58.10.43–5.95
Tumours, n626346
Mean size, cm3.322.923.49
Size range1.5–6.21.1–6.71.7–5.5
≤4 cm, n (%)48 (77)53 (84)25 (58)
4–7 cm, n (%)14 (23)10 (16)18 (39)
Right, n (%)26 (42)29* (56)29 (63)
Left, n (%)36 (58)34 (44)17 (37)
Anterior, n (%)29 (47)27 (43)18 (39)
Posterior, n (%)33 (53)36 (57)28 (61)
Upper Pole, n (%)12 (19)15 (24)11 (24)
Interpolar, n (%)31 (50)23 (36)24 (52)
Lower pole, n (%)19 (31)25 (40)11 (24)
Exophytic, n (%)31 (50)43 (68)32 (70)
Intracortical, n (%)31 (50)20 (32)14 (30)

On referral, tumours were assessed by recent imaging (<4 months) and again by contrast-enhanced CT assessment at the time of the cryoablation procedure. The longest transverse tumour diameter was recorded and the tumour was classified as anterior or posterior (with respect to the longitudinal greater curve of the kidney) and in thirds through the kidney parenchyma as upper, interpolar or lower pole by location. Tumours were then classified as exophytic or central according to whether > or <50% of their volume lay within the substance of the renal cortex, respectively. Tumours were biopsied (at least two 18-G cores) after placement of cryoprobes and before treatment. Biopsies were taken from only a single tumour in the case of patients with multiple similar tumours, or not taken on rare occasions when the tumour was too small to biopsy reliably after cryoprobe placement (usually tumours <18 mm in diameter).

All patients underwent outpatient consultation and anaesthetic assessment before cryoablation. All patients had normal coagulation status at the time of ablation with an international normalized ratio ≤1.4 and a platelet count >90 × 109/L. All procedures were performed by an interventional radiologist under prone-oblique general anaesthesia in the CT scanner with occasional additional ultrasonography guidance where necessary. A urologist was not present at any procedure.

All treatments were performed using the Galil ‘Seednet Gold’ cryoablation system (Galil Medical, Arden Hills, Minneapolis, MN, USA). A mean of 4 (and up to 8) 17-G (1.47 mm diameter) sealed argon cryoprobes were placed around the tumours, generating temperatures as low as −110°C. The resultant ‘tissue-lethal’, composite iceball was radiologically aimed to incorporate the target tumour with at least a 5-mm treatment margin. This ensured that the lethal isotherm of ∼−20°c overlay the tumour margin. The amount and volume of the therapeutic iceball was controlled by the rate of delivery of argon to the probe tip and thawing was achieved with a similar delivery of helium gas through the closed circuit yielding tip temperatures of ∼42°C. The tumours were exposed to a double freeze–thaw cycle (10 min freeze–8 min thaw–10 min freeze) which maximised the tissue lethality of the iceball. We used CT (GE LightSpeed, Milwaukee, WI, USA) to intermittently monitor cryoprobe placements and to visualise the size and location of the evolving iceball using a 2.5-mm slice thickness and a standard CT technique (120 kV peak and ∼240 mAs; Fig. 1).

figure

Figure 1. A, 38-mm right lower polar RCC in a 78-year old male. B, Intraprocedural CT demonstrates parallel cryoprobe placement and consumption of the tumour by the expanding iceball (black arrow). Note contrast-tinted hydrodissection fluid displacing the adjacent bowel (white arrow). C, 12-month follow-up CT demonstrates no residual enhancement and early involutional change.

Download figure to PowerPoint

Technical success was defined by imaging follow-up ∼2 weeks after the procedure. In the vast majority of patients, this was performed using diligent CT densitometry using non-contrast and post-contrast images in the late arterial and portal venous phases to define the non-enhancing ablation zone. In those with significant renal impairment (usually estimated GFR < 30 mL/min/1.73 m2) late arterial phase subtraction MRI was used. Imaging follow-up was then performed at 3 months (where there was uncertainty about the completeness of treatment at the initial post-procedural CT) and then at 6 months, 12 months, 18 months, and annually thereafter (Fig. 1).

Where thermally sensitive structures, such as the bowel and pancreas, were closely approximated, an 18-G needle was inserted into the interposed, retroperitoneal space and 2% contrast-tinted, normal saline was injected under CT guidance to displace adjacent viscera to a distance of ∼15–20 mm from the target tumour margin. These measures were carried out to create an adequate safety margin and provide an operating space large enough to enable thorough and effective ablation.

All complications were defined using Clavien–Dindo criteria and are described below [5]. A significant or major complication rate was defined as complications that required medical intervention, i.e. ≥Clavien–Dindo grade 2.

Statistical Analysis

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Methods
  5. Statistical Analysis
  6. Results
  7. Discussion
  8. Conflict of Interest
  9. References

Data were first entered into the IBM SPSS 19 statistics package for analysis. Initially simple frequency tables were compiled for demographics and tumour characteristics, including tumour size, anterior/posterior tumour position, polar location of tumour within the kidney and whether the tumour was exophytic or central. Maximum likelihood logistic regression analyses were then performed at the tumour level for technical success of treatment, and at the procedural level (grouping together multiple tumours treated in one session) for complications. Data were then transferred to Stata 11 and these analyses were repeated using exact logistic regression analysis, which allows for small numbers of events. The computational intensity of exact logistic regression analysis necessitated the non-reporting of some coefficients, specifically those for age and gender. A separate exact logistic regression analysis examined the relationship of numbers of probes used and technical success, controlling for age, gender and tumour size.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Methods
  5. Statistical Analysis
  6. Results
  7. Discussion
  8. Conflict of Interest
  9. References

In total, 171 tumours were treated in 147 patients during 153 cryoablation procedures (137 performed by D.J.B., 16 by T.J.B.; Table 1). Of these, 97 were histologically confirmed RCCs, of which 87 were solitary tumours and 62 were both solitary and had ≥6 months radiological follow-up. Twenty-three of the tumours were oncocytomas, 15 (11.1% of those biopsied) proved to be indeterminate on biopsy (non-diagnostic), and in 36 tumours (usually in cases with multiple tumours) no biopsy was performed. Tumour locations and characteristics are shown in Table 1. Four patients underwent more than one procedure for separate tumours and eight patients had multiple tumours treated during a single ablation session.

Thirteen tumours in 12 patients treated in 12 treatment sessions had initial treatments that were technically incomplete. In 11 of these, the subtotal treatment was apparent at immediate post-procedural follow-up. The 13 tumours included nine histologically confirmed RCCs and one oncocytoma. In three cases no biopsy was performed.

One patient with a single tumour was excluded from the analysis of technical success as this tumour was not evaluable: the patient, with a solitary kidney, was treated with radiological evidence of segmental vein invasion as an adjunct to concurrent tyrosine kinase inhibitor therapy. Tables 2 and 3 show results of maximum likelihood and exact logistic regression analyses of technical success in the remaining 170 tumours. No variables significantly predicted primary subtotal treatment in either model, although coefficients were similar in both models. The analysis of the relationship of numbers of probes used and technical success, controlling for age, gender and tumour size, gave an odds ratio of 2.01 for four or more probes used, but this was nonsignificant (95% CI 0.31–18.60; P = 0.68).

Table 2. Results of maximum likelihood logistic regression analysis of variables predicting technical success in the treatment of 170 tumours
Patient/tumour variableOdds ratio95% CI for odds ratioSignificance
LowerUpper
Age (continuous)0.990.951.040.75

Gender: Male = 1

Female = 0

2.960.909.760.08

Tumour ≥4 cm. = 1

Tumour < 4 cm. = 0

0.330.091.180.09

Anterior position = 1

Other position = 0

0.930.283.060.91

Upper pole = 1

Else = 0

1.130.274.770.87

Central = 1

Exophytic = 0

0.320.091.080.07
Table 3. Results of exact logistic regression analysis of variables predicting technical success in treatment of 170 tumours (coefficients for age and gender not shown)
Patient/tumour factorOdds ratio95% CI for odds ratioSignificance
LowerUpper

Tumour ≥ 4 cm = 1

Tumour < 4 cm = 0

0.370.091.590.21

Anterior position = 1

Other position = 0

0.880.233.441.00

Upper pole = 1

Else = 0

1.180.267.581.00

Central = 1

Exophytic = 0

0.340.081.260.12

Table 4 shows the procedure-based analysis by tertiles of date of treatment, showing that the majority of subtotal treatments occurred in the first tertile of the treatment programme (chi-squared test for trend = 6.60; dof = 1; P = 0.01).

Table 4. Analysis of proportion of 153 treatment procedure sessions resulting in subtotal treatment, by tertile of treatment
 Tertile 1 (procedures 1–51)Tertile 2 (procedures 52–102)Tertile 3 (procedures 103–153)
  1. Chi-squared test for trend = 6.60; degrees of freedom = 1; P = 0.01.

Technical success434850
Subtotal treatment831
Total515151

During cryoablation, adjacent structures prone to thermal injury were displaced using 2% contrast-tinted, normal saline hydrodissection in 75 of the 153 procedures (49%). Warm, pressure-bagged pyeloperfusion (retrogradely via a ureteric catheter with subsequent placement of a temporary ureteric stent) was used to reduce the risk of PUJ injury in six of 153 procedures (3.9%).

In total there were 16 complications related to the 153 procedures, seven of which (4.6%) met or exceeded Clavien–Dindo grade 2 criteria (Table 5). One patient developed post-procedural haemorrhage requiring arterial embolisation, whilst another developed a perirenal haematoma warranting two-unit blood transfusion. Pelvicalyceal injuries were observed in three patients (debris colic, minor pelvicalyceal leak and low grade PUJ stenosis), all of whom required temporary ureteric stenting. A further two patients developed intraprocedural pneumothoraces requiring chest drain insertion. The nine complications that did not require medical intervention included: four small pneumothoraces that resolved on observation and did not require aspiration or chest drain insertion; three patients who developed urinary retention; one patient who sustained persistent flank paraesthesia; and a further patient who had to be readmitted to hospital for urinary sepsis.

Table 5. Complications of treatment in 147 patients
ComplicationNumber
Medically significant (>Clavier–Dindo Grade 2)
Pelvicalyceal injury3
Post-procedural haemorrhage1
Perirenal haematoma1
Pneumothorax requiring chest drain insertion2
Other complications
Pneumothorax not requiring chest drain insertion4
Urinary retention3
Persistant pain or parasthesia1
Other (re-admitted for urinary sepsis)1
Total16

Tables 6 and 7 show results of maximum likelihood and exact logistic regression analyses of the variables predicting complications. The single variable found to predict complications in both analyses was that of one or more tumours treated in a session being in the upper pole of a kidney, perhaps reflecting the fact that the largest single group of complications was small self-limiting pneumothoraces, which did not require medical intervention.

Table 6. Results of maximum likelihood logistic regression analysis of variables predicting incidence of any complication in 153 procedures
Patient/tumour factorOdds ratio95% CI for odds ratioSignificance
LowerUpper
Age (continuous)1.030.981.090.22

Gender: Male = 1

Female = 0

1.070.333.520.91

Tumour ≥ 4 cm. = 1

Tumour < 4 cm. = 0

1.140.343.800.83

Anterior position = 1

Other position = 0

1.060.353.210.91

Upper pole = 1

Else = 0

4.701.5614.190.006

Central = 1

Exophytic = 0

1.670.545.130.37
Table 7. Results of exact logistic regression analysis of variables predicting incidence of any complication in 153 procedures (coefficients for age and gender not shown)
Patient/tumour factorOdds ratio95% CI for odds ratioSignificance
LowerUpper

Tumour ≥ 4 cm. = 1

Tumour < 4 cm. = 0

1.150.294.101.00

Anterior position = 1

Other position = 0

1.150.333.931.00

Upper pole = 1

Else = 0

4.291.2914.920.02

Central = 1

Exophytic = 0

1.610.455.560.55

Turning to solitary RCCs with ≥6 months' radiological follow-up we have, to date, observed only one case of unexpected late local recurrence of disease at the margin of a tumour which had been deemed completely treated. This recurrence was documented 12 months after the cryoablation procedure and a further ablation treatment achieved treatment completion (Fig. 2). The median inpatient hospital stay was 1 day (mean 1.5 days). Forty seven of 62 (76%) procedures for solitary RCC required only a single day in hospital.

figure

Figure 2. A, Avidly enhancing 40-mm left interpolar RCC in a 67-year old male. B, Follow-up late arterial phase CT 10 days after cryoablation demonstrates apparent complete tumour ablation. C, Follow-up late arterial phase CT 12 months after cryoablation demonstrates local enhancing nodular recurrence (white arrow) within previous tumour ablation zone. D, Follow-up CT 6 months after targeted repeat cryoablation shows no ongoing tumour enhancement (18 months' total follow-up of initial ablation).

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Methods
  5. Statistical Analysis
  6. Results
  7. Discussion
  8. Conflict of Interest
  9. References

Major developments in radiological imaging techniques over the last few decades have driven the increasing detection of small and usually subclinical renal tumours. This, in turn, has invoked a management problem, particularly in the setting of possibly indolent renal tumours of uncertain malignant potential. Standard surgical resection, with its attendant morbidity, has been justified in the case of larger renal tumours where, for example, T1b tumours (4–7 cm) may harbour higher grade (Fuhrman grade 3–4) disease in up to 39% of cases [6], yet 3–4 cm tumours may also harbour similar higher grade disease in 14–25% of cases [7]. The biopsy outcomes in the present study reflect expected malignancy rates for a mean tumour size of 31 mm. There are likely to be a number of false-negative biopsies for malignancy amongst the 15 biopsies (11.1%) that yielded necrotic/inflammatory material, emphasising the need to follow up patients in this indeterminate group.

Approximately ≥70% of renal tumours are <4 cm in diameter at presentation and often occur in older patients with baseline chronic renal impairment at the time of tumour detection. Active surveillance has been proposed in elderly patients with significant comorbidities [8] but a management dilemma remains in the older, yet reasonably fit, patient who may be subjected to repeated imaging and a small but indeterminate risk of metastatic progression, as high as 6% in some series [9]. An effective, low morbidity and minimally invasive therapeutic option would be extremely useful in the management of these patients and there is accruing evidence of the efficacy of image-guided ablation in this role [10]. Thermal ablation techniques for smaller renal tumours have been advocated for some time with meta-analysis of retrospective, non-randomised case series suggesting that radiofrequency ablation has a significantly higher risk of residual disease than cryoablation [8]. Accurate treatment dosimetry using radiofrequency ablation , i.e. determination of a tumour-lethal thermal dose, remains problematic with radiofrequency ablation, largely owing to gas obscuration of the target tumour during treatment. Cryoablation has the clear merit of physically demonstrating its tissue effects through the radiologically demonstrable iceball that is induced around the clustered cryoprobes and this permits the ablation to be carefully assessed during the course of the procedure [11, 12]. The long-term cytocidal effect of cryoablation has been confirmed [13, 14] and there is accruing evidence that this technique can be effectively deployed percutaneously [15-17].

Our own consecutive experience of percutaneous renal tumour cryoablation, from the initiation of a treatment programme, has yielded a single-session complete tumour treatment success of 92.4% (157/170) with a further nine tumours successfully treated after repeat ablation, thus yielding an overall technical success rate of 97.6% (166/170). The overall successful technical outcome in our cohort of patients corresponds to that published in the literature [10]. In the remaining four cases a multidisciplinary team or patient decision was made not to repeat the image-guided ablation treatment. One patient had a marginal crescent of persistent tumour enhancement adjacent to the renal pelvis. This was deemed problematic for further ablation by the multidisciplinary team and the patient proceeded to nephrectomy, which confirmed residual viable disease. A second young patient presented for ablation with two targetable new foci of disease in a solitary kidney, having already undergone contralateral nephrectomy for clear-cell disease 3 years earlier, along with subsequently resected metastatic liver disease. At follow-up, new focal disease was found within the remnant kidney as well as aggressive peri-pelvic invasive disease and new liver metastases. A decision was made not to pursue further ablation or resection. A further patient died from metastatic RCC after a subtotal cryoablation procedure. In retrospect this patient, who had recently undergone resection for a contralateral 47-mm tumour and had a concurrent 18-mm tumour for percutaneous cryoablation, had evidence of early metastatic disease at the time of initial treatment, with small (<1 cm) pulmonary nodules, which progressed in size and volume at subsequent follow-up imaging. In the fourth patient there was apparent persistent enhancing disease after two cryoablation procedures but the patient declined further interventions. This went on to become substantive local recurrence and the patient proceeded to salvage partial nephrectomy.

Of 104 patients with 125 renal tumours (both solitary and multiple) at a mean radiological follow-up of 20.1 months, of which 62 were documented solitary RCCs and another 10 were RCCs in multi-tumour presentations, we have documented only a single case of unexpected late local recurrence which occurred at 12 months. This has been successfully re-treated by cryoablation with no evidence of disease recurrence. Notably, the chronological analysis of subtotal treatments by tertile (Table 4) clearly shows a decreasing trend over time, testifying to a definite procedural learning curve.

There has previously been a widespread perception in the literature that both anterior and larger renal tumours (>4 cm) are problematic for percutaneous ablation therapy in terms of increased rates of subtotal treatment or complication, but our experience supports data emerging from recent studies that have shown that ablation of anteriorly sited tumours [15] and tumours > 4 cm [16] is technically feasible with good treatment outcomes. Multivariate analysis of our cohort of patients did not show a significant difference in outcome for anteriorly located tumours, although there was a trend towards initial subtotal treatment in tumours > 4 cm in diameter (P = 0.09) and those situated centrally (P = 0.07). This highlights the need for diligent technique and careful imaging follow-up in these treatment groups.

The medically significant (Clavien–Dindo Grade ≥ 2, requiring medical intervention) complication rate was 4.6%. Post-procedural arterial embolisation, two-unit blood transfusion and temporary chest drain were required in three cases. Temporary ureteric stenting was required for sloughed material causing transient ureteric obstruction, a temporary pelvicalyceal leak and a low-grade PUJ stenosis. It is difficult to anticipate these adverse events but with hindsight, we would advocate pyeloperfusion and/or ureteric stenting for patients in whom ablative therapy might encroach upon the PUJ.

A limitation of the present study is the mean radiological follow-up at 20.1 months for all tumours with ≥6 months' follow-up and 17.7 months for solitary, biopsy-proven RCCs. This can be particularly telling in the case of relatively indolent renal tumours which may take some time to manifest small foci of local disease recurrence. We aim to keep this patient cohort under close radiological follow-up for at least 5 years.

In summary, percutaneous, image-guided cryoablation of renal tumours appears to provide an effective and minimally invasive treatment option for the management of incidental renal tumours with a low complication rate. In our experience, tumours in an anterior location do not represent an undue risk for subtotal treatment or complications. The long-term efficacy of the treatment remains to be determined by careful follow-up.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Methods
  5. Statistical Analysis
  6. Results
  7. Discussion
  8. Conflict of Interest
  9. References