Solitary solid renal mass: can we predict malignancy?

Authors


  • P.V. and S.A. contributed equally to this work.

  • W.K. is a recipient of a Research Scholar Award from the FRSQ.

Wassim Kassouf, Assistant Professor of Urology, McGill University Health Centre, Montreal General Hospital, 1650 Cedar Ave. L8-315, Montreal, Quebec, Canada H3G 1A4. e-mail: wassim.kassouf@muhc.mcgill.ca

Abstract

Study Type – Therapy (retrospective cohort)

Level of Evidence 3a

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

It is known that the majority (80%) of solid renal masses are malignant. Most of the literature suggests that smaller tumour size is associated with a higher incidence of benign disease.

We have confirmed that decreased tumour size is associated with benign disease, particularly for lesions <2 cm. Once radiologically apparent angiomyolipomas are excluded, the incidence of benign disease with small renal masses is lower than reported in the literature (11%). Furthermore, we have shown that small renal masses in females have a higher association with benign disease. Patient age and tumour location were not predictive of benign histology. We have also stratified our risks not only for benign disease but also for clinically indolent renal cancers to help physicians counsel patients with regard to managing these solid renal masses.

OBJECTIVE

  • • To determine the clinical predictors of benign disease in patients with solitary solid renal masses.

PATIENTS AND METHODS

  • • Pathology reports of patients who underwent radical or partial nephrectomy at two hospitals from 1998 to 2008 were reviewed.
  • • Only patients with solitary solid unilateral renal masses were included.
  • • Predictors of malignancy risk were assessed with univariate and multivariate logistic regression analysis.

RESULTS

  • • A total of 592 patients with a mean (sd) age of 60 (13) years were included, 38% of whom were women. Radical and partial nephrectomy was performed in 66% and 34% of patients, respectively.
  • • Renal masses were equally distributed on the right and left sides (49% vs 51%, P= 0.84). Masses were more commonly located in the upper and lower poles than in the mid pole (40.8% vs 38.7% vs 20.5%, respectively).
  • • The mean tumour size was larger in patients who underwent radical compared with partial nephrectomy (6.8 cm vs 2.9 cm, P < 0.001). The rate of benign disease in our overall population was 9.5%.
  • • On univariate and multivariate analysis, only a renal mass size <2 cm and female gender were predictive of benign disease. On further analysis the magnitude of this effect was found to be additive.

CONCLUSIONS

  • • Renal masses <2 cm and female gender were associated with a higher probability of benign disease.
  • • Patient age and tumour location were not predictive of benign disease.

INTRODUCTION

Kidney cancer is the seventh most common cancer diagnosis in men, and eighth most common in women in the USA [1]. In 2010, an estimated 8200 men will die from kidney cancer. Approximately 60% of kidney cancers are identified incidentally because of the increased use of abdominal imaging [2]. Contrast-enhancing kidney masses seen on tomography are considered to be malignant, and in most cases are surgically extirpated without previous tissue diagnosis. Renal masses are traditionally biopsied only when there is suspicion of lymphoma, abscess or extra-renal metastatic disease. Although most masses would be RCC on final pathology, a significant number of these masses may be benign so that patients undergo unnecessary surgery and are subject to its associated morbidity. Our aim was to identify a group of patients that have a higher likelihood of having benign histology, and would hence benefit from further diagnostic testing such as core biopsy, or active surveillance.

PATIENTS AND METHODS

We retrospectively reviewed the medical records, including pathology reports, of patients who underwent surgical management for solid renal masses that were suspected RCCs. After institutional ethics board approval, data were collected from one academic hospital and one community hospital affiliated with our institution. Inclusion criteria were patients with a solitary solid renal mass who underwent partial or radical nephrectomy, either laparoscopic or open, for suspicion of RCC during the period 1998–2008. Exclusion criteria were patients with multiple masses, bilateral masses, cystic masses, or if the solid renal mass was radiologically reported as angiomyolipoma owing to the presence of fat on the preoperative CT. We extracted demographic data for each patient such as age and gender. We also examined tumour characteristics: size of mass (widest diameter), location (upper, mid or lower pole), histology, tumour stage, Fuhrman grade, final margin status and presence of lymphovascular invasion.

Proportions were compared using the one-sample proportion test. Continuous variables were compared using the Student's t-test. The Cochran–Armitage test for trend was used to evaluate whether renal mass size was related to the risk of malignancy. Predictors of malignancy risk were assessed using univariate and multivariate logistic regression. Statistical analyses were performed using Stata v.10.1 (StataCorp, College Station, TX, USA).

RESULTS

Baseline demographic data is shown in Table 1. Out of the 651 patients initially identified, 592 met the inclusion criteria and were included in our final analysis. Patients had a mean (sd) age of 60 (13) years and 38% were women. Approximately half the tumours were ≤4 cm in maximum diameter, and one third of those were <2 cm. Renal masses were equally distributed on the right and left sides (49% vs 51%, P= 0.84); however they were more commonly located in the upper and lower poles than in the mid pole (40.8% vs 38.7% vs 20.5%, respectively). The mean tumour size in patients who underwent radical compared with partial nephrectomy was 6.8 cm vs 2.9 cm (P < 0.001), respectively. The rate of benign disease in our overall population was 9.5%.

Table 1. Demographic data for 592 patients
  1. *Includes 11 unclassified RCC tumours, three mucinous tubular and spindle-cell carcinomas, one collecting-duct cancer, one angiosarcoma, one metastatic hemangiopericytoma, one tubulocystic RCC, one carcinoid tumour, and one liposarcoma. †American Joint Committee on Cancer 2009. ‡The absolute number of patients does not add up to 592 owing to missing information.

Population characteristic 
Gender, n (%) 
 Female226 (38)
 Male366 (62)
Mean (sd) age, years60 (13)
Histological type: malignant tumours, n (%)536
 Clear-cell400 (74.6)
 Papillary70 (13.1)
 Chromophobe46 (8.6)
 Other*20 (3.7)
Pathological stage: malignant tumour, n (%) 
 pT1a271 (45.8)
 pT1b100 (16.9)
 pT2a56 (9.5)
 pT2b19 (3.2)
 pT3a138 (23.3)
 pT3b6 (1)
 pT42 (<1)
Fuhrman grade, n (%) 
 134 (7.7)
 2250 (56.3)
 3131 (29.5)
 429 (6.5)
Mean (sd) tumour size, cm5.3 (3.5)
Tumour size, n (%) 
 0–4 cm293 (49.5)
 4.1–7 cm149 (25.2)
 7.1–10 cm98 (16.6)
 >10 cm52 (8.8)
Tumour location, n (%) 
 Upper189 (40.8)
 Lower179 (38.7)
 Mid95 (20.5)

Overall, 90.5% of patients had malignant tumours. Of these, 400 (74.6%) were clear-cell RCC, 70 (13.1%) were papillary, 46 (8.6%) were chromophobe, and 20 (3.7%) were miscellaneous other renal malignancies. Of malignant tumours, 64% were Fuhrman grade 1 or 2, and 36% were Fuhrman grades 3 or 4. Perinephric fat invasion was present in 85 (16%) specimens, and six (1%) specimens had vena cava invasion. Oncocytomas were the most common benign lesions occurring in 59% of patients with benign pathology, followed by angiomyolipoma in 25%. Other benign tumours identified were: schwannoma (n= 1), metanephric adenoma (n= 1), mesoblastic nephroma (n= 1), xanthogranulomatous pyelonephritis (n= 1), hydatid cyst (n= 1), cystic nephroma (n= 1), hamartoma (n= 1) and solitary fibrous tumour (n= 1).

Using the Cochran–Armitage test, initial analysis showed a significant association between smaller tumour size and decreased risk of malignancy (P= 0.005) (Table 2). This trend was even more significant when considering only the risk of high grade malignancy, i.e. Fuhrman score >2 (P < 0.001). We therefore performed a logistic regression analysis to better define this association. On univariate analysis (Table 3), tumours were stratified into four size groups: ≤2 cm, 2.1 to 3 cm, 3.1 to 4 cm and >4 cm. Tumour size ≤2 cm was associated with a decreased risk of malignancy (OR 0.39, P= 0.009) compared with larger tumours. Male gender conferred statistically greater odds of having a malignant tumour as compared to females (OR 2.16, P= 0.006). Patients were also stratified into four groups by age: <50, 50 to 60, 60 to 70, and >70. None of these groups differed significantly in terms of risk of malignancy. Tumour location (upper, middle or lower pole), or nature of the hospital (academic vs community), were also nonsignificant in determining risk of malignancy.

Table 2. Risk of malignancy by renal mass size
Renal mass sizeNo. of patients (%)Risk of malignancy (%)Risk of high grade (Fuhrman >2) malignancy (%)
  1. Cochran–Armitage test for trend: *P= 0.005 and **P < 0.001.

Total592 (100.0)536/592 (90.5)163/502 (32.5)
≤4 cm293 (49.5)260/293 (88.7)42/251 (16.7)
 ≤2 cm86 (14.5)71/86 (82.6)*10/81 (12.3)**
 2–3 cm112 (18.9)98/112 (87.5)*15/93 (16.1)**
 3–4 cm95 (16.1)91/95 (95.8)*17/77 (22.1)**
>4 cm299 (50.5)276/299 (92.3)*121/251 (48.2)**
 4–7 cm149 (25.2)135/149 (90.6)45/121 (37.2)
 7–10 cm98 (16.6)93/98 (94.9)49/86 (57.0)
 >10 cm52 (8.8)48/52 (92.3)27/44 (61.4)
Table 3. Variables associated with malignant renal mass (N= 592)
Variable n (%)Logistic regression analysis
UnivariateMultivariate
OR (95% CI) P OR (95% CI) P
  1. OR: odds ratio. P values <0.05 were considered significant and are marked with an asterisk.

Tumour size     
 ≤4 cm293 (49.5)    
  ≤2 cm86 (14.5)0.39 (0.20–0.79)0.009*0.48 (0.22–0.99)0.048*
  2–3 cm112 (18.9)0.58 (0.29–1.18)0.130.64 (0.31–1.31)0.22
  3–4 cm95 (16.1)1.90 (0.64–5.63)0.252.00 (0.67–6.00)0.22
 >4 cm299 (50.5)1.00 (reference)n/a1.00 (reference)n/a
Right side299 (50.5)0.95 (0.54–1.64)0.840.87 (0.49–1.53)0.62
Location     
 Upper pole189 (31.9)1.00 (reference)n/a1.00 (reference)n/a
 Mid pole95 (16.1)0.76 (0.32–1.84)0.550.80 (0.33–1.96)0.63
 Lower pole179 (30.2)0.87 (0.41–1.87)0.730.88 (0.41–1.93)0.76
 Unknown129 (21.8)0.49 (0.24–1.03)0.060.61 (0.28–1.34)0.22
Age     
 <50 years148 (25.0)0.86 (0.40–1.82)0.690.82 (0.38–1.78)0.61
 50–60 years134 (22.6)1.88 (0.74–4.81)0.191.86 (0.71–4.83)0.20
 60–70 years161 (27.2)0.78 (0.37–1.61)0.490.69 (0.33–1.48)0.34
 ≥70 years149 (25.2)1.00 (reference)n/a1.00 (reference)n/a
Male gender366 (61.8)2.16 (1.24–3.78)0.006*2.13 (1.20–3.78)0.01*
Underwent surgery at academic centre518 (87.5)1.00 (0.44–2.30)0.991.13 (0.48–2.69)0.78

On multivariate analysis, tumour size <2 cm and female gender were the only variables associated with a decreased risk of malignancy. We did not detect an interaction between significant variables in multivariate analysis (P= 0.42, data not shown). We then used a multivariate linear regression model to assess the probability of malignant disease by mass size and patient gender to determine whether these factors were additive or synergistic. Table 4 shows an additive relationship of the effect of mass size and gender on probability of malignant disease. Female patients with tumours ≤2 cm in size had a 76.8% probability of having a malignancy.

Table 4. Predicted probabilities of malignancy by gender and renal mass size using multivariate linear regression analysis
Renal mass sizeProbability of malignancy %
All Fuhrman GradesFuhrman grade >2
FemaleMaleFemaleMale
≤2 cm76.887.812.212.4
2–3 cm82.690.716.016.2
3–4 cm94.097.021.822.2
>4 cm88.894.247.948.4

Interestingly, if we change the endpoint to predicting clinically indolent disease, defined as benign or less aggressive disease (T1a with Fuhrman grade ≤2), only tumour size remains a significant predictor of clinically indolent disease. The risk of not having a clinically indolent disease decreased to 12.3% for tumours ≤2 cm (Tables 2,4).

DISCUSSION

Owing to the widespread use of abdominal imaging, incidentally found renal masses have become increasingly common. Given that a significant percentage of these will be benign, it is important to try to identify risk factors that can correctly predict benign disease. The present study shows that female gender and tumour size ≤2 cm were independent predictors of benign disease, and that the effect of these two variables were additive in lowering risk. To our knowledge, this is the first report to show such a relationship. Age and tumour location were not associated with benign disease or less aggressive tumours.

Our findings concur with previous studies. In a recent retrospective study, Zisman et al. [3] showed that female gender was associated with an increased risk of benign histology, with 21% having benign disease compared with 13% of men (P= 0.001). In another study of 488 cases, women were more likely than men to have benign mass (27.3% vs 14.5%, P < 0.001) [4]. In a study of Japanese patients undergoing partial nephrectomy for small renal masses, 26.1% of women had benign lesions compared with 5.4% of men (P < 0.001) [5].

While female gender appears to be uniformly protective from malignancy in the literature, the effect of tumour size on malignancy is less clear. In the present study, tumour size <2 cm was associated with a decreased risk of malignancy and there was a significant trend for smaller tumours to be less malignant. This is consistent with other reports [6,7]. One study reported a significant correlation between benign histological findings and tumour diameter (P= 0.017) [6]. Another report found that for every 1 cm increase in tumour size, the odds of having RCC as opposed to a benign tumour went up by 17% (P < 0.001) [7]. However, other studies found no significant association between tumour size alone and malignancy risk [3,4,8,9].

The present study showed an overall rate of malignancy of 90.5%. This is substantially higher than the ∼80–85% malignancy rate reported by others [4,10,11]. The high malignancy rate in the present study could be attributable to the fact that we excluded tumours diagnosed as angiomyolipomas on preoperative CT scans. This was purposely done to reflect real-life practice where smaller lesions, which are reported as angiomyolipoma owing to the presence of gross fat on CT, are typically followed without intervention. Furthermore, being a referral centre, urologists at our institution are likely to see a high proportion of complex cases, which are more likely be malignant.

Based on the present results, 23.2% of women with small tumours ≤2 cm will have benign disease on pathology. This finding implies that ∼1 in 4 would have been exposed to potential morbidity from surgery unnecessarily. One can speculate whether these patients may benefit from further investigation, such as a core needle biopsy. Significant strides have been made recently in improving the accuracy and safety of biopsies. Older techniques, which employed fine needle aspiration had suboptimum sensitivities whereas newer techniques, employing core biopsies have shown low false-negative rates [12,13]; therefore, greater consideration could be given to obtaining tissue diagnosis before surgery for women with small tumours.

An alternative approach to upfront surgery is active surveillance of small tumours. We, and others, have shown that selected patients with small renal masses can be safely followed [14,15]. In a recent meta-analysis, Chawla et al.[16] examined the outcome of 234 observed small renal masses after a mean duration of 34 months. The mean lesion size at presentation was 2.6 cm with a mean rate of growth of 0.28 cm/year. There were only three documented reports of metastases, one in a patient with a rapidly growing tumour and one in a patient with a 9-cm initial sized tumour. Specific information regarding the third patient who developed metastatic disease is absent from the original report. In a recent article by Jewett et al. [17], 178 patients were followed prospectively with biopsies and serial imaging to determine the rate of progression and metastases in patients with small renal masses (≤4 cm maximum tumour diameter). Out of the 127 patients included in the growth analysis, 25 progressed locally and two developed metastases. The rates of growth of benign and malignant lesions were not significantly different. These articles lend further support to the hypothesis that small renal masses grow slowly with rare metastases. Combining this information with results from the present study suggests that active surveillance may be even more advocated for elderly women with small renal masses (especially those ≤2 cm) as they have the highest probability of harbouring benign or clinically indolent disease. Lastly, unlike the authors of other studies, we did not find an association between age and incidence of benign disease or less aggressive tumours.

The strengths of the present study include its large sample size and its multifaceted nature, encompassing multiple surgeons in both academic and community practices. In addition, the present analysis included clinically relevant, readily available baseline patient and tumour characteristics. It also reflected real-life practice in that it excluded lesions that are reported as angiomyolipoma owing to the presence of gross fat on CT as patients with such lesions are typically followed without intervention.

The present study was limited by being retrospective with inherent associated biases. In addition, we used pathological specimen dimensions to quantify tumour size. This probably underestimated the actual size of the masses in vivo, because formalin fixation causes tumours to shrink. This fact is noteworthy because in previous studies looking at non-small cell lung cancer, formalin had been shown to cause sufficient shrinkage such that there was stage migration in a minority of tumours [18]. The malignancy rates we obtained, based on dimensions from pathological specimens, could theoretically be different from those from CT-based measurements; however, as Kurta et al. [19] have shown, the difference between pathological and CT sizing of renal masses is, at most, a few mm, and hence not clinically significant. We believe the present results would therefore also be applicable to renal masses measured by tomography. Finally, whether specific preoperative imaging characteristics (such as degree of enhancement, presence of central scar, etc.) can predict malignant histology requires further evaluation.

In conclusion, small incidentally found renal masses pose a challenge to urologists because many of these tumours are either benign or have less aggressive biology. We have identified a subgroup of patients, namely females with tumours <2 cm who have significantly reduced odds of having malignant tumours. Those odds decrease further when we exclude tumours of less aggressive biology. The option of renal biopsy and active surveillance could therefore be more strongly advocated for this patient population.

CONFLICT OF INTEREST

None declared. Source of funding: Le fonds de la recherche en Santé du Québec.

Ancillary