Anil Kurup, Department of Radiology, Mayo Clinic, 200 First St SW, Rochester, MN 55902, USA. e-mail: firstname.lastname@example.org
Study Type – Prognosis (case series)
Level of Evidence 4
What's known on the subject? and What does the study add?
Growth of small renal masses has been shown to be variable, with mean growth rates ranging from 1 to 5 mm/year in several series. Small numbers of oncocytomas in these series show similar but variable growth rates of 0.5–5.2 cm/year with no statistically significant difference compared with RCCs.
This study gathered growth rate data from the largest pool (n= 33) of pathologically proven oncocytomas to date, using current imaging techniques. Primarily, it confirms the variable growth of renal oncocytomas over an extended period of time with mean (sd) growth of 2.9 (2.6) mm/year over a mean of 36 months. Observed oncocytoma growth was equivalent to that previously established for RCCs, and small renal masses generally, under surveillance. Secondarily, we found considerable similarity in tumour measurements between observers, suggesting that even low growth rates may be reproducibly documented.
• To describe the growth rates of oncocytomas before treatment with surgical resection or percutaneous ablation.
PATIENTS AND METHODS
• This single-institution retrospective study included 33 consecutive, pathologically proven renal oncocytomas with serial contrast-enhanced computed tomography scans spanning at least 1 year before intervention, from 2000 to 2009.
• Tumours were measured by two radiologists, and growth rates and interobserver variability were calculated.
• The mean (range) pre-procedural imaging surveillance period was 36 (12–124) months (median 33 months).
• The mean (sd) oncocytoma size was 17 (11) mm (range 4–47 mm, median 15 mm) in maximum transverse diameter on initial imaging and grew to a mean (sd) of 26 (5) mm (range 10–83 mm, median 23 mm) by the time of treatment.
• Overall, the mean (sd) and median growth rates were 2.9 (2.6) mm/year and 2.7 mm/year, respectively (range −1.2–10.9 mm/year). After weighting by the se of each tumour's growth rate, the mean (sd) change was 2.1 (1.2) mm/year.
• The mean (range) interobserver variability for each tumour measurement was 1 (0–7) mm with an intraclass correlation coefficient of 0.99.
• Renal oncocytomas grow at a rate similar to reported growth rates of renal cell carcinoma.
• As the observation of growth does not distinguish between benign and malignant renal tumours, growth of small renal masses under active surveillance should be carefully considered before a switch is made to intervention.
Partly because of increased use of cross-sectional imaging, the incidental detection of small renal masses is increasing and presents a challenge in the management of renal tumours. Most benign tumours, such as unilocular cysts and fat-containing angiomyolipomas, can be appropriately characterized with cross-sectional imaging. Unfortunately, reliable non-invasive differentiation of oncocytoma from RCC is not possible . Imaging of oncocytoma with CT typically demonstrates a well-circumscribed, homogeneously enhancing mass without haemorrhage, calcification or necrosis, occasionally with a central, stellate scar. This appearance is not specific and has considerable overlap with RCC, as shown in one study in which radiologists retrospectively reviewing preoperative CT scans of renal tumours gave the correct diagnosis of oncocytoma in only 7 of 49 (12%) observations .
Moreover, histological differentiation between benign and malignant primary renal tumours has been fraught with issues related to high false-negative and non-diagnostic biopsy rates [3,4]. One study comparing surgically excised tissue with biopsy specimens obtained in the operating room immediately after nephrectomy or partial nephrectomy for renal tumours in 103 patients reported diagnostic accuracy of 76 and 80% with non-diagnostic rates of 11 and 17% . More recently, preoperative renal mass biopsy has been gaining acceptance, with some authors reporting accuracy of >90%, although results are not uniform and the practice of biopsy guiding treatment decisions is variable .
At the same time, active surveillance has become an accepted alternative treatment strategy for small renal masses according to the AUA guidelines . Given this interest in active surveillance, it is important to determine whether growth can distinguish between benign and malignant renal masses and may serve as an appropriate and reproducible endpoint in deciding upon future intervention. The present study was performed primarily to determine the growth rate of renal oncocytomas during a period of observation before therapy. Secondarily, the interobserver variability in measurement of these tumours was assessed.
PATIENTS AND METHODS
The present study was approved by the institutional review board and was compliant with the Health Insurance Portability and Accountability Act. Patients with small renal masses detected by imaging at our institution are evaluated by a urologist and informed of all treatment options, including observation, partial nephrectomy and percutaneous ablation. The treatment plan offered to the patient is based on patient comorbidities, previous surgery, accessibility of the renal lesion to surgical or percutaneous approaches, and patient preference. Patients who are placed on an active surveillance programme are followed with CT or MRI every 6–12 months, at the urologist's discretion. The typical surveillance schedule is 6 months, 12 months, and yearly thereafter.
The institutional surgical pathology database was searched for all renal oncocytomas resected from 2000 to 2009. Of this population, only those with contrast-enhanced CT imaging spanning at least 1 year before surgery were included in the analysis. A total of 318 oncocytomas were resected from 325 patients during the period 2000–2009. Of these, 19 oncocytomas had enhanced CT imaging for at least 1 year before surgery in 14 patients.
The radiology departmental ablation database was also searched for renal tumours diagnosed as oncocytomas by biopsy obtained before percutaneous ablation from 2000 to 2009. A total of 527 renal tumours were treated with percutaneous ablation in 447 patients from 2000 to 2009. As routine practice, these tumours were biopsied at the time of ablation; biopsy consisted of one to three 18-gauge x 2-cm core biopsies obtained under real-time ultrasonography guidance immediately before ablation. This yielded 57 tumours diagnosed as oncocytomas in 52 patients. Tumours diagnosed as oncocytic neoplasm not otherwise specified were excluded. Of these oncocytomas, 14 were seen in 12 patients with at least 1 year of follow-up with enhanced CT imaging before ablation. Of the 12 patients treated with percutaneous ablation, indications for ablation included previous renal surgery (n= 8), medical comorbidity (n= 3) and patient preference (n= 1).
The tumour diameters were measured by two attending radiologists (A.N.K. and T.D.A.) with subspecialty training in abdominal radiology, blinded to each other's measurements. The greatest diameter of each tumour was measured on axial enhanced CT images on an Advantage Workstation (GE Healthcare, Waukesha, WI, USA) with sizes recorded to the nearest mm. CT was performed on a variety of CT scanners of different manufacturers with variable imaging parameters, including detector number, slice thickness, amount and rate of iodinated contrast material injected, and timing of the CT acquisition. When multiphasic imaging was present, measurements were made on the phase of enhancement that best depicted the tumour based on each reader's discretion.
Agreement in oncocytoma size assessment for the two reviewers was assessed using the intra-class correlation coefficient (ICC), ICC(2,1), as all oncocytomas were assessed by the reviewers, and the reviewers were assumed to be a random subset of possible reviewers. According to Landis and Koch an ICC of 0.8–1.0 is ‘almost perfect’, with perfect being 1.0 . Agreement was also assessed over periods of follow-up to understand whether there was a temporal difference in measurement agreement; none was found.
Since agreement of the two reviewers was very nearly perfect, the mean of the two reviewers' size assessments was used to estimate the change in oncocytoma size during the follow-up period. Assuming a linear trend in size over time, the slope of the regression line was estimated for each of the 33 oncocytomas. The simple mean (sd) and median of the estimated slopes were reported. However, since the number of measurements made in the follow-up period varied, the precision of the tumours' slope estimates also varied. Therefore, the mean slope by using a weighted regression of the individual slopes was also reported, where the weight is the variance taken from the se of the slope estimates of the tumour. The associations of initial tumour size, treatment type, and follow-up duration with the change in oncocytoma size over time were assessed using weighted linear regression on the slopes.
Among the 12 patients with 14 oncocytomas diagnosed by biopsy and treated with percutaneous ablation, there were 10 males and two females with a mean (range) age of 72 (54–83) years (median age 75 years). The mean (range) pre-ablation imaging surveillance period for oncocytomas treated with ablation was 42 (16–78) months (median 39 months). Five oncocytomas were treated with radiofrequency ablation, and nine were treated with cryoablation.
Among the 14 patients with 19 oncocytomas treated surgically, there were nine males and five females with a mean (range) age of 67 (51–85) years (median 69 years). The mean (range) preoperative imaging surveillance period for tumours treated surgically was 31 (12–124) months (median 17 months). Five oncocytomas were resected with laparoscopic radical nephrectomy (in a patient with end-stage renal disease after renal transplantation), 12 with open partial nephrectomy, and two with laparoscopic partial nephrectomy.
Combining the ablation and surgical cases, a total of 33 oncocytomas were included in the analysis in 25 patients (one patient underwent both ablative and surgical procedures for oncocytomas in the study period). Including synchronous and metachronous tumours, nine patients had bilateral oncocytomas. Six of the 25 patients also had synchronous or metachronous RCCs. None of the patients had a known hereditary disease associated with renal tumours. There were 18 males and seven females with mean (range) age of 71 (51–85) years (median 73 years). The mean (range) pre-procedural imaging surveillance period was 36 (12–124) months (median 33 months). A total of 149 CT scans were reviewed with a mean of 4.5 and median of 6 CT scans per tumour and a mean of 6 and median of 4 CT scans per patient.
The mean (sd) size of the oncocytomas was 17 (11) mm (range 4–47 mm, median 15 mm) on initial imaging and they grew to a mean (sd) size of 26 (15) mm (range 10–83 mm, median 23 mm) by the time of treatment (Table 1). The slope estimates of the growth curves were variable, ranging from −1.3 to 10.9 (Fig. 1). The mean (sd) estimated growth rate, weighted for the precision of the slope estimate of the growth curve for each oncocytoma was 2.1 (1.2) mm/year. The mean (sd) and median growth rates, not weighted for this precision, were 2.9 (2.6) mm/year and 2.7 mm/year, respectively, and varied from −1.2 to 10.9 mm/year.
Table 1. Growth of renal oncocytomas treated by percutaneous ablation or surgery between 2000 and 2009
Mean (median; range) initial size, mm
12 (11; 4–20)
21 (17; 7–47)
17 (15; 4–47)
Mean (median; range) final size, mm
22 (21; 12–43)
30 (23; 10–83)
26 (23; 10–83)
Mean (median; range) total growth, mm
10 (9; 1–27)
9 (5; −2–48)
9 (7; −2–48)
Mean (median; range) follow-up, months
42 (39; 16–78)
31 (17; 12–124)
36 (33; 12–124)
Mean (median; range) growth rate, mm/year
2.8 (2.2; 0.3–7.4)
3.0 (2.8; −1.2–10.9)
2.9 (2.7; −1.2–10.9)
Insubstantial growth was observed in eight (24%) of the 33 oncocytomas. Specifically, four (12%) showed growth of <1 mm/year, two (6%) showed no growth, and two (6%) showed negative growth.
Assuming linear growth over the follow-up period and weighting the estimates, the mean slope of the growth curves was 2.6 for excised tumours vs 1.9 for ablated tumours, a difference which was not significant (P= 0.19). The final sizes of the oncocytomas treated with ablation (mean 22 (7.9) mm, median 20 mm) were smaller than those treated surgically (mean 30 (18.6) mm, median 22 mm), but this difference was not significant (P= 0.37; Wilcoxon rank-sum test).
The oncocytoma growth rates were not significantly associated with size on initial CT, with a linear regression parameter estimate of 0.08 (0.05) mm/year (P= 0.10 [Fig. 2]). In other words, a 1 mm increase in initial size had an increased growth rate of 0.08 mm/year with se of 0.05. Similarly, the oncocytoma growth rates did not vary significantly by duration of follow-up, with a linear regression parameter estimate of 0.02 (0.18) mm/year (P= 0.89 [Fig. 3]).
Oncocytoma growth was not associated with patient age (per 10 years; P= 0.57), gender (P= 0.09), multifocality (P= 0.82), or history of RCC (P= 0.45).
Each radiologist made a total of 149 tumour measurements. The median (range) absolute value in size difference as measured by the two observers was 1 (0–7) mm (Fig. 4). For 42 of the 149 (28%) measurements the radiologists reported the same size, and in 108 (72%) measurements the radiologists differed by, at most, 1 mm. The overall assessment of interobserver agreement in the 149 measurements yielded an ICC of 0.99. The absolute difference in tumour measurement between the two observers did not significantly vary by mean tumour size (r=−0.08, P= 0.36).
Growth of small renal masses, including RCCs, has previously been shown to be variable, with mean growth rates ranging from 1 to 5 mm/year in several series [8–15]. These growth rates are important when one considers the greater role of expectant management of small renal masses, particularly in patients with medical comorbidities. Given that the development of metastatic disease has not been reported in the absence of tumour growth , the presence of tumour growth may serve as a trigger for active intervention.
In an important meta-analysis of the natural history of small renal masses, the mean (range) growth rate of all 234 renal tumours under surveillance was shown to be 2.8 (0.9–8.6) mm/year at a mean follow-up of 34 months . Progression to metastatic disease developed in three patients (1% of tumours); two of the three tumours showed interval enlargement (one with rapid growth of 12 mm/year from 2 to 8 cm and the other with slow growth of 2 mm/year from 8.8 cm over 111 months) with no data given regarding the third tumour. Of the 76 tumours with pathological correlation, there was no significant difference in the growth rate of nine oncocytomas compared with 67 RCCs. However, the small number of oncocytomas yielded a highly variable growth rate of 0.5 (6.7) mm/year. Siu et al.  found a mean growth rate of 5.2 mm/year in six pathologically confirmed oncocytomas, compared with 7.1 mm/year for 10 RCCs. In a study reviewing 15 oncocytomas proven by percutaneous biopsy, growth of the six tumours ultimately treated surgically was 2.4 (2.1) mm/year, significantly greater than the remaining nine oncocytomas that were observed (0.7 [0.5] mm/year) .
In an assessment of growth kinetics of renal masses under active surveillance, Crispen et al.  showed a mean (sd) growth rate of 2.85 (4.1) mm/year, with subset analysis showing no significant difference (P= 0.629) between pathologically proven benign lesions (1.3 mm/year, n= 9, including seven oncocytomas and two angiomyolipomas), low-grade malignancies (3.4 mm/year, n= 40), and high-grade malignancies (5.3 mm/year, n= 12).
Thus, the growth rate of malignant renal masses has been shown to be not significantly different from renal oncocytomas. The present study adds to this finding with growth rate data from the largest pool of pathologically proven oncocytomas to date. It confirms the variable growth of renal oncocytomas over an extended period of time. Observed oncocytoma growth was equivalent to that previously established for RCCs, and small renal masses generally, under surveillance. By contrast to the previous studies, we had the opportunity to review a relatively large number of tumours using current imaging techniques. In addition to showing growth of many oncocytomas, we also showed the lack of growth in some tumours; however, the absence of growth of a renal mass has previously been shown not to confirm benignity .
The present data suggest that oncocytoma growth rates do not vary significantly by tumour size, as previously observed in small renal masses generally [8,10,13,15]. However, the small number of tumours limits the power of this observation, as there was a slight trend toward increased growth rate with larger initial tumour size. This observation differs from the association found by Crispen et al. , between tumour volume and growth rates, with smaller tumours showing faster growth than larger tumours. Growth rates of the oncocytomas in the present study also did not correlate with the length of observation before intervention, which suggests a linear growth pattern, although the majority of tumours were small (<3 cm) and observed for <3 years.
When one uses tumour enlargement as a criterion for predicting behaviour, accuracy in tumour measurement is necessary to document significant growth with confidence. A previous study by Punnen et al. , showed measurements of 29 renal masses <4 cm in 21 patients to vary by 3.1 mm for multiple readers and 2.3 mm for a single reader. By contrast, the tumour measurements in the present study were remarkably similar between the two subspecialty radiologists, suggesting that growth may be reliably documented using standard technique on clinical workstations. The present study suggests that growth, even as low as the 2–3 mm/year seen in small renal masses, may be reproducibly documented.
Several limitations of the present study should be noted. First, the CT scans used for tumour measurement were performed with variable imaging techniques, including variations in scanner detector number, slice thickness and timing of the contrast bolus. Such technical differences could account for some perceived change in size based on imaging in the absence of actual tumour enlargement. Likewise, these differences could account for the few cases of negative growth. While the heterogeneity of CT scans may limit the precision of the measurements, the miniscule interobserver variability is encouraging. Moreover, the variability in CT scan parameters match that faced in clinical practice during active surveillance. Second, percutaneous biopsy, which was used to diagnose the oncocytomas treated by percutaneous ablation, is subject to sampling error with one study showing RCC ultimately existing in 18% of tumours that were shown to be oncocytomas on percutaneous biopsy and another series of 138 patients with oncocytoma at our institution finding that 10% of oncocytomas had coexisting RCC [21,22]. Therefore, at our institution, percutaneous needle biopsy is frequently followed by the ablation procedure at the same session. Finally, and implicit in any study investigating tumour growth with pathological proof, is the selection bias towards tumours that were managed in an active vs passive fashion. One could hypothesize that potential oncocytomas may have been observed but without growth, pathology was never confirmed. Thus, our results could potentially overestimate the growth rate of oncocytomas generally. However, this seems unlikely as the growth rates in the present study are similar to those observed in small renal masses without pathological proof and selected small pools of oncocytomas in previous series studying the natural history of small renal masses.
In conclusion, among renal masses treated either with percutaneous thermal ablation or surgically, oncocytomas grow at a rate similar to reported growth rates of RCC. Thus, renal mass growth alone cannot be used to predict malignancy. Whether the presence of growth itself or some growth threshold should prompt intervention during active surveillance requires further study.