Accurate Risk Assessment of Patients with Pathologic T3aN0M0 Renal Cell Carcinoma

To develop a more precise risk-stratification system by investigating the prognostic impact of tumor growth within fatty tissues surrounding the kidney and/or renal vein. We conducted a retrospective review of the medical records of 211 patients with a pathologic diagnosis of T3aN0M0RCC among 4,483 renal cell carcinoma (RCC) patients from February 1988 to December 2015 according to the number of T3a pathologies—extrarenal fat invasion (EFI) and/or renal venous invasion (RVI). During a mean follow-up duration of 38.8 months, the patients with both pathologies (EFI + RVI) had lower recurrence free survival (RFS) rate than those with only a single pathology (p = 0.001). Using multivariable Cox regression analysis, the presence of both factors was shown to be an independent predictor of RFS (HR = 1.964, p = 0.032); cancer specific survival rate was not different among patients with EFI and/or RVI. Patients with pathologic T3aN0M0 RCC presenting with both EFI and RVI were at an increased risk of recurrence following nephrectomy. Therefore, pathologic T3a RCC could be sub-divided into those with favorable and unfavorable disease according to presence of EFI and/or RVI pathologies.

ance with the seventh AJCC TNM staging 13 and Fuhrman grading system 14 , respectively. Histological subtypes of RCC were assigned in accordance with the WHO classification system 15 . We evaluated the following clinical and pathologic features: age at surgery, gender, body mass index, Eastern Cooperative Oncology Group performance (ECOG) performance status, surgery type (partial nephrectomy [PN] or radical nephrectomy [RN]), tumor location, tumor size, histologic subtype (clear cell, papillary, chromophobe and unclassified), Fuhrman nuclear grade, capsular invasion, presence of necrosis, angiolymphatic invasion, perineural invasion and sarcomatoid differentiation. Additionally, T3a tumors were classified as: i) isolated EFI, isolated RVI, or iii) EFI and RVI.
Follow up protocol. After nephrectomy, all patients were followed up on according to the SNU protocol.
Symptom assessments, routine blood tests, chest x-rays and kidney/bladder Computed tomography (CT) scans were conducted every 3 months for the first 2 years. Following year 2, the same battery of tests were conducted every 6 months for 2 years followed by annual follow ups as necessary.
Statistical analysis. Comparisons were conducted across the T3a patient classifications (ie, EFI, RVI, and EFI + RVI). Another group analysis was performed according to the number of cause of pT3a (ie, isolated EFI or RVI, and EFI + RVI). Analyses were performed using SPSS for Windows, version 22.0 (SPSS Inc., Chicago,IL, USA). Fisher's exact and Pearson chi-squared tests were performed to compare clinical and pathological features among groups. Differences in variables with a continuous distribution across dichotomous categories were assessed using the Mann -Whitney U -test. A P value of <0.05 was considered statistically significant.
Recurrence was defined as radiologically verified metastasis or local disease recurrence during the study period. Recurrence-free survival (RFS) was defined as the interval between the primary surgery and the last follow-up visit and the absence of disease recurrence or progression. Cancer-specific survival (CSS) was defined as the interval between the primary surgery and the last follow-up visit or RCC related death. The RFS and CSS in each group were estimated using Kaplan-Meier methods and the log-rank test was applied to compare survival curves. Univariate and Multivariate Cox proportional hazard models were used to verify predictors of RFS and CSS.

Baseline characteristics of pT3a RCC patients. The clinical and pathological characteristics of 211
T3aN0M0 RCC patients who underwent nephrectomy are summarized in Table 1 Recurrence-free survival. During a mean follow-up of 38.8 months (median 26.0 months), radiologic recurrence was observed in 42 patients (19.9%). The site of recurrence was recorded as renal fossa (28/42, 66.6%), lymph node (8/42, 19.0%) and lung (4/42. 9.5%). RFS was significantly lower in EFI + RVI group than either EFI only or RVI only groups (Fig. 1A, log rank test = 0.008 or 0.010, respectively). According to the number of parameters leading to T3a classification, which Group B (EFI + RVI) had a significantly lower RFS rate than Group A (EFI only + RVI only) (Fig. 1B, p = 0.002). Seven-year RFS rates were calculated as 57.8% in Group A and 26.4% in Group B.
As shown in Table 2, sex, tumor size, tumor necrosis, perineural invasion, angiolymphatic invasion, sarcomatoid differentiation and having both EFI and RVI (Group B) were significant factors in univariate Cox proportional hazard analysis and predictive of tumor recurrence. Using a multivariate analysis, sex, tumor size, tumor necrosis, and having both EFI and RVI (Group B) were shown to be significant factors.
Cancer-specific survival. During follow-up, 23 patients died due to RCC. CSS free survival was lower in the EFI + RVI group than either EFI-only or RVI-only group, however this difference was not statistically significant ( Fig. 2, log rank test = 0.625, 0.882, respectively). Using a multivariate Cox proportional hazard analysis, tumor size was shown to be a significant factor in determining CSS among T3a RCC patients who underwent nephrectomy (HR = 1.133, 95% CI: 1.040-1.234, p = 0.004) ( Table 3). Another factors (positive surgical margin and presence of tumor necrosis) were significant factors to predict cancer specific survival (HR = 19.864, p < 0.001 and HR = 4.238, p = 0.025, respectively).

Discussion
In this analysis, we reviewed 211 T3aN0M0 RCC patients who underwent surgical treatment, and stratified patients according to the cause of their RCC classification as T3a; the 173 EFI (82.0%) and 84 RVI (39.8%) patients had similar RFS rates during a median follow-up of 26.0 months. However, patients with both EFI and RVI (n = 47, 23.3%) had a significantly lower RFS rate than patients who had only one factor. Despite of no significance to CSS, the patients who had EFI and RVI had more aggressive feature among T3a RCC patients.
Da Costa et al. 10  According to the current TNM staging system, EFI and RVI can each lead to inclusion in same stage -(T3a) regardless of whether one or both of these pathologies (EFI or RVI) are present. However, several previous studies have shown that patients with EFI and RVI had poorer outcomes than those with only one (EFI or RVI) 10,[16][17][18] . A recent European multi-center study 17 , validated the latest edition of the TNM (2009), evaluating 503 cases of T3a stage RCC and noting significantly lower survival rates in patients with concomitant EFI and RVI vs those with the single pathologies. Margulis et al. 18 confirmed these results by analyzing 419 patients treated for pT3a stage RCC; poorer survival rates were observed in patients with EFI + RVI compared to patients with either pathology alone.
Our results indicate a significant difference in RFS among groups A (single pathology) and B (both pathologies), however we did not observe a significant difference among the CSS of the two groups The difference between our study and previous studies is that we excluded patients who had lymph node invasion (LNI) 10,16 . The previous study showed that patients who had EFI and RVI had higher rates of LNI. According to the Da Costa group 10 17 , in a large multi-institutional study, showed that an RVI-only group had a better prognosis than the fat-invasion group. In contrast, Jung et al. 20 reported that a group of RVI-only patients had lower 2-year and 5-year survival rates than the fat invasion group. Park et al. 16 showed that RVI-only patients had significantly lower RFS and CSS than EFI among T3aN0M0 patients after nephrectomy. Therefore, this issue is still controversial and future prospective studies may be required to confirm the results of these studies.
Another interesting issue in this study was shown in multivaratie analysis to predict RFS and CSS. Regardless of the number of factors of pT3a, tumor diameter and tumor necrosis were significant factors to RFS and CSS, respectively. A previous reports also showed in their study among pT3a RCC patients the importance of tumor size in pT3a, HR to recurrence was 1.125 (p < 0.001) 21 and HR to CSS was 2.506 (p = 0.011) 2 , it was same result in our study. And another study about tumor necrosis was similar results to our study, they emphasized the significance of tumor necrosis to prognosis in RCC patients 22,23 . There were several limitations, in our study. Firstly, our study was a retrospective chart review. Secondly, we could not differentiate between EFI into perinephric fat invasion and sinus fat invasion. However, most studies which compare perinephric fat invasion and sinus fat invasion have reported similar results between two groups. Mouracade et al. 24 showed, in their analysis of 143 T3a RCC patients, sinus fat invasion and perinephric fat invasion had similar prognostic values. Poon et al. 25 reported that sinus fat invasion was not a predictor of CSS. Therefore, our inability to stratify patients with EFI into perinephric fat invasion and sinus fat invasion groups could be significant. Despite a relatively long follow up duration (mean of roughly 40 months), only 10% patients died from RCC, therefore CSS analysis was limited by the small sample size. More studies with a longer follow up and more patients from multiple centers could overcome this limitation.
In conclusion, among T3aN0M0 RCC patients, those who had both EFI and RVI had poor PFS rates than patients with only one pathology (ie, EFI or RVI). Therefore, following surgery, T3a patients with EFI and RVI should be more closely followed up than other T3a RCC patients.  Table 3. Uni-and Multi-variable Cox-regression analyses on cancer specific survival. HR; hazard ratio, CI; confidence interval, ECOG; Eastern Cooperative Oncology Group, RN; radical nephrectomy, PN; partial nephrectomy.