For decades, small renal cell carcinomas (RCCs) have been incidentally detected more frequently by abdominal imaging [1]. Radical nephrectomy is a standard treatment option with no size criteria; recently, partial nephrectomy has been recommended for treating clinically localised small RCC if technically feasible [2]. Recently, RCC surveillance might be an alternative to surgery if there is no progression or metastasis [3]. Active surveillance of small renal masses has been reported to be acceptable for short-term oncological outcomes [3,4].

Although most small renal masses are known to have low malignant potential, certain patients under surveillance have developed progression or metastasis [5,6]. Furthermore, most studies of small renal mass did not use a total cancer cohort, meaning the cohort consisted of malignant and benign masses, and the ratio of benign masses was approximately 20%-30% [5,6]. This finding is due to a low accuracy rate of the imaging, but a recent study noted that preoperative multiphasic computed tomography (CT) imaging has an accuracy for identifying malignancy of approximately 90% [7], Another researcher also reported that perfusion CT significantly increased the accuracy of predicting RCC to 95% [8]. The more radiologic imaging has evolved, the more small RCCs have been detected accurately.

Therefore, if the surveillance of small RCC increases and is validated in selected patients [9], determining the risk of small renal masses will be more important to enable physicians to decrease overtreatment and avoid missing aggressive tumours. Looking at the literature, few data exist for risk stratification for small RCCs. Factors that affect RCC prognosis can be summarised in various ways; the TNM staging classification system is known to be the most reliable prognostic factor based on a systematic review [5]. Above the TNM staging, advanced age is supposed to be one of considering factors to decide early treatment or delayed intervention [10].

However, some of the small masses may have adverse final pathology like T3a. Thus, we investigated the prognostic significance of clinical features in patients with a small RCC by analysing postoperative pathologic outcomes.

Table 1 shows the demographics for all of the patients in this study. On final pathology, 3.5% of all patients were categorised into the PUG. All the PUG patients had pathologic T3a (unpublished data). The PUG consisted of older patients with larger tumour sizes than the NUG patients (all p<0.05). There was no difference in histology between the groups (p=0.538).

Table 2 presents the results of the logistic regression analysis. In the univariate logistic regression analysis, age was found to be an independent predictor for the PUG (OR, 1.04; 95% CI, 1.02-1.06; p<0.001). Hypertension (OR, 1.55; 95% CI, 1.04-2.65; p=0.035) and larger tumour size (≥30 mm) (OR, 1.96; 95% CI, 1.23-3.13; p=0.005) were also statistically significant predictors for the PUG. In the multivariate logistic regression analysis encompassing all variables, age (OR, 1.04; 95% CI, 1.02-1.06; p=0.001) and larger tumour size (≥30 mm) (OR, 1.94; 95% CI, 1.21-3.11; p=0.001) were observed to be significantly correlated with the PUG.

We stratified the study patients by the number of risk factors, consisting of old age (≥60 years) and larger tumour size (≥30 mm). We analysed the clinical importance of the risk factors using the PUG criteria; 8.4% (24 of 286) of the patients with 2 risk factors were in the PUG, compared to 1.8% (15 of 838) of the patients with no risk factor (Table 3).

The mean and median (interquartile range) follow-up duration after surgery for the 2,085 patients included in the present study were 40.9±33.5 months and 34.0 months (13.0-62.0 months), respectively. During follow-up, recurrence occurred in 2.9% of all subjects, with 11% (8 of 73) in the PUG and 2.6% (53 of 2,012) in the NUG. The Kaplan-Meier analysis demonstrated that patients in the PUG had a worse recurrence-free survival than the other group (p<0.001) (Fig. 1). In multivariate Cox regression analysis encompassing clincopathological variables, upstaging was observed to be an independent predictor for worse recurrence-free survival (hazard ratio [HR], 3.40; 95% confidence interval [CI], 1.57-7.34; p=0.002) along with tumour size≥30 mm (HR, 1.86; 95% CI, 1.11-3.12; p=0.018) (Table 4).

Delayed intervention for RCC is especially suited for the patients with a low malignant potential tumour. Currently, many studies for renal mass surveillance have been based of the patient's preference, age, comorbidities, and imaging study [4-6]. However, they did not focus on how to distinguish patients based on malignant potential. The growth rate on imaging studies is the most common way to assess malignant potential [14]. Deferred treatments made obtaining sequential images possible. Stratification in patients with RCC still remains a challenge. Although CT imaging showed no differences in size estimation compared to surgical specimens with small RCC, CT could not precisely estimate the actual malignant potential [15]. Percutaneous renal biopsy also could not outperform a multiparametric imaging study [16]. Therefore, the recent American Urological Association guidelines state that RCC surveillance should be accepted as an option after giving information for tumour progression to patients [17].

To successfully selection for early intervention or delayed intervention, risk stratification for small RCCs must be essential. In this multicener-based study on the relationship between clinical features and postoperative pathologic outcomes for RCC, we investigated the prognostic importance of age and tumour size for pathologic T3a upstaging following nephrectomy. There was no difference in cell histology between the PUG and the NUG. Furthermore, we found that postoperative pathological upstaging may be a potential surrogate for recurrence-free survival after nephrectomy. Our findings were also observed in previously published data by other researchers. Smaldone et al. [5] have reported that increased age and initial greatest tumour dimension were associated with progression after metastasis in their systematic review (all p<0.001). Komai et al. [18] observed that young patients with RCC had a better cancer-specific survival, implying low malignant potential. The authors explained that nonclear cell type RCC is more prominent in younger patients. Our results also corroborated the theory that the mean age of the clear cell group was significantly higher than that of the nonclear cell group (55.1 vs. 53.3, p=0.028). However, cell histology did not show additional impacts on the recurrence-free survival in present analysis. Oh et al. [19] have shown that recurrence-free survival after nephrectomy significantly correlates with tumour size. Nguyen and Gill [20] reported that the risk of metastasis depends on tumour size and that a 3-cm threshold was a useful cutoff value for their epidemiologic study. A larger tumour size (≥3 cm) was also a significant predictor for the PUG in our study. A threshold of 3 cm could be an acceptable cutoff to distinguish which tumours have a more aggressive potential.

In the present study, the primary end point was to define which clinical features correlate with pathological upstaging after nephrectomy. We had supposed that a pT3 tumour had more malignant potential than a pathologic T1 tumour. The TNM stage classification system is widely recommended for clinical use. Many clinical nomograms for RCC have introduced the pT stage as the most powerful prognostic factor [21,22]. We also found that the PUG had worse progression-free survival according to Kaplan-Meier and multivariate analyses. There have been many clinical trials for adjuvant therapy after nephrectomy. Those trials used the pT stage for eligibility for study inclusion [23].

Our study has some limitations. First, a potential limitation is the relative retrospective design from several centers and it is possible influence of selection bias on the results. Moreover, we focused on RCC only, which is also a prone to selection bias. Second, only subjects of the same race were included in the present study; racial differences in tumour biology were overlooked. Third, central radiologic and pathology review probably enhances the quality of data, thus lack of central review dose appear to be a limitation. We believe that our results should be validated with further investigation. However, there was no consistent protocol for small RCCs. This study suggests that patient age and tumour size must be considered for evaluating indolent aggressiveness before treatment. Patients with 2 risk factors had a high risk of having pT3a ranging from 1.8% to 8.4% compared to patients with no risk factor.

In conclusion, age and tumour size could be useful in determining the malignant potential in patients with clinically localised small RCC and appropriate candidates for early or delayed treatment. For older patients with a relatively large RCC tumour, early treatment should be considered to be the first option.