The extent of tumour fat invasion affects survival in patients with renal cell carcinoma and venous tumour thrombosis

Authors


  • RB and MR contributed equally to this work.

Marco Roscigno, Ospedali Riuniti di Bergamo, Largo Barozzi 1, 24185, Bergamo, Italy. e-mail: roscigno.marco@gmail.com

Abstract

Study Type – Prognosis (case series)

Level of Evidence 4

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

Several studies have investigated the role of renal fat invasion in RCC with neoplastic thrombus. Our study confirmed that the simultaneous presence of TFI and VTT was associated with higher risk of CSM than the presence of VTT alone.

We demonstrated for the first time that patients with either PFI or RSFI had a two-fold increased risk of CSM compared with patients with VTT alone. Moreover, the combination of PFI and RSFI was associated with a three-fold risk of CSM.

OBJECTIVE

• To investigate the effect of presence and extent of tumour fat invasion (TFI) – perinephric invasion (PFI), renal sinus fat invasion (RSFI) or both PFI and RSFI – on cancer-specific mortality (CSM) in patients with renal cell carcinoma (RCC) and venous tumour thrombus (VTT).

METHODS

• We examined 184 consecutive patients with RCC with VTT treated with nephrectomy between 1987 and 2007. Associations with CSM were evaluated by univariable and multivariable Cox proportional hazard models.

RESULTS

• Median follow up was 21 months. The 5-year CSM-free survival estimates were 75%, 36% and 20% in patients with VTT without TFI, those with VTT with PFI or RSFI, and those with VTT with both PFI and RSFI, respectively (P < 0.001). In multivariable analyses, presence of either PFI or RSFI was associated with a two-fold increased risk of CSM, whereas presence of both PFI and RSFI was associated with a three-fold increased risk of CSM, relative to VTT-only cases.

• The inclusion of the variable describing the presence and extent of TFI in a base model including pT stage, Fuhrman grade and presence of nodal disease and metastatic disease significantly increased the accuracy in predicting CSM (+2.1%; P < 0.001) in patients with VTT.

CONCLUSIONS

• Patients affected by RCC with VTT and TFI have a higher risk of CSM relative to cases with VTT only. Patients with both PFI and RSFI showed increased CSM compared with patients with either PFI or RSFI.

• Our results suggest TFI should be accurately evaluated and included in routine pathological reports to provide better patient risk stratification.

Abbreviations
IVC

inferior vena cava

VTT

venous tumour thrombus

TFI

tumour fat invasion

PFI

perinephric fat invasion

RSFI

renal sinus fat invasion

CSM

cancer-specific mortality.

INTRODUCTION

Typically, RCC has a tendency to invade the venous system, with extension into the renal vein (23%) or inferior vena cava (IVC) (7%) [1]. However, the prognostic role of cranial tumour thrombus extension is controversial [2–8].

The new 2009 TNM staging system introduced the effect of IVC wall invasion and defined pT3c cases as patients having IVC venous tumour thrombus (VTT) above the diaphragm or patients with IVC VTT below the diaphragm but invasion of the IVC wall [3]. Moreover, RCC cases with VTT invading the renal vein were classified as pT3a, as were cases with tumour fat invasion (TFI) only [3].

Regarding prognostic factors, recent evidence has shown that the simultaneous presence of TFI and VTT is associated with a poor prognosis [4–9]. However, the novel 2009 TNM classification did not include TFI in VTT subclassification [3].

Recently, Roos et al. [9] analysed the impact of perinephric fat invasion (PFI) or renal sinus fat invasion (RSFI) alone and simultaneous PFI and RSFI on cancer-specific mortality (CSM) in patients with RCC and VTT. They confirmed that the simultaneous presence of TFI and VTT is associated with a poorer prognosis relative to the presence of VTT alone. Moreover, in the subgroups of TFI cases, the combination of PFI and RSFI did affect CSM, whereas the presence of PFI or RSFI alone was not associated with lower cancer-specific survival rates [9].

The aim of the present study was to investigate the effect of TFI on CSM in patients with RCC with VTT. Moreover, we tested whether the involvement of one (PFI or RSFI) or both (PFI and RSFI) renal fat compartments could be associated with further impairment of patient prognosis.

MATERIALS AND METHODS

After the approval of the local ethics committee, we retrospectively analysed 1652 patients with RCC treated with partial or radical nephrectomy at a single academic institution (1987–2007). We included 184 patients with pT3 RCC and VTT without direct invasion of the adrenal gland.

The level of the tumour thrombus was classified as limited to the renal vein, extending into the IVC below the diaphragm, or extending above the diaphragm. The cranial extent of the tumour thrombus was assessed by venocavography before 1995 and by MR angiography or CT angiography from 1995 to 2007. In cases of intrahepatic–suprahepatic IVC invasion, transoesophageal echocardiography was used to assess cardiac tumour involvement. The presence/absence of synchronous metastases at the time of referral for surgery was evaluated by abdominal CT or MR scans, chest radiographs or CT scans, and bone scans.

Surgical technique consisted of laparotomy for radical nephrectomy with ipsilateral adrenalectomy and retroperitoneal lymphadenectomy. The lymphadenectomy was performed according to the surgeon’s preference and on the basis of clinical and intraoperative findings. The lymphadenectomy template normally included hilar, precaval, retrocaval and interaortocaval nodes for the right kidney, and interaortocaval and para-aortic nodes for the left one.

All microscopic slides were retrospectively reviewed by a single uropathologist (MF) blinded to patient outcomes. All patients included in the current analysis had sinus fat tissue available for histological examination. Since 1987, according to a standardized sampling protocol, all nephrectomy specimens have been carefully examined. Specifically, description of tumour extension beyond the kidney into perinephric fat, renal sinus fat and renal vein, was always addressed. Moreover, at least one wedge of sinus fat was sampled in case of tumour mass far from sinus, or two or three wedges when the tumour was in close proximity, to confirm a possible microscopic invasion. PFI was defined as the extension of the tumour into the fat surrounding the renal capsule; RSFI was characterized by the spread of the tumour into the fat of the renal sinus.

After surgery, chest and abdominal imaging studies (abdominal ultrasound and chest radiographs or thoraco-abdominal CT) were performed every 4 months for the 1st year, every 6 months up to the 5th year, and annually thereafter. In addition, a bone scan was performed every year. Data on survival or cause of death were determined from the clinical files at our centre, information from patients’ relatives, or death certificates.

The survival rates were calculated from the date of surgery to the last available follow up (date of death in censused cases). Survival curves were estimated using Kaplan–Meier methods. Differences between the survival curves were determined by the log-rank test. Univariable and multivariable Cox regression analyses tested the association between all the available predictors (i.e. age, presence of synchronous metastases, lymph node involvement, Fuhrman grade, simultaneous presence of VTT and PFI, RSFI, or PFI and RSFI) and CSM.

Statistical analyses were performed with the software package SPSS version 13 (SPSS Inc., Chicago, IL, USA). Statistical significance was defined as a P value < 0.05.

RESULTS

The clinical and pathological characteristics of 184 patients with RCC and VTT are summarized in Table 1, according to the presence of TFI and its extent (PFI or RSFI, or PFI and RSFI). Nineteen of the 184 patients (10.4%) were pNx, 118 (64.1%) were pN0 and 47 (25.5%) were pN1). Median number of lymph nodes removed was eight (mean 10; range 2–58).

Table 1.  Patient characteristics and descriptive statistics
VariableAll patients (N= 184)No TFI (N= 48) n (%)PFI or RSFI only (N= 83)PFI and RSFI (N= 53)P value
  1. ECOG, Eastern Cooperative Oncology Group; TFI, tumour fat invasion; PFI, perinephric fat invasion; RSFI, renal sinus fat invasion. *One-way analysis of variance test;†Pearson chi-square test.

Age (years)    0.324*
 Mean (median)62.5 (63.1)60.1 (61.8)62.4 (63.1)64.4 (65.2)
 Range26–8335–8332–7926–78
Tumour dimensions (cm)    0.004*
 Mean (median)10.2 (10.3)8.5 (8.7)10.3 (10.6) 11.3 (11.5)
 Range3.2–233.2–20.03.5–23.06.0–22.0
ECOG performance status, n (%)    0.415
 0–1125 (67.9)34 (70.8)56 (67.5)35 (64.8)
 ≥259 (32.1)14 (29.2)27 (32.5)18 (35.2)
Pathological stage (TNM 2009), n (%)    0.058
 pT3a 116 (63.0)35 (72.9)55 (66.3)26 (49.1)
 pT3b27 (14.7)7 (14.6)12 (14.4)8 (15.1)
 pT3c41 (22.3)6 (12.5)16 (19.3)19 (35.8)
Pathological grade, n (%)    0.011
 238 (20.6)14 (29.2)18 (21.7)6 (11.3)
 399 (53.8)27 (56.3)47 (56.6)25 (47.2)
 447 (25.6)7 (14.5)18 (21.7)22 (41.5)
Nodal status (TNM 2009), n (%)    0.029
 pN0 118 (64.1)33 (72.9)55 (66.3)30 (56.7)
 pN+47 (25.5)7 (14.5)21 (25.3)19(35.8)
 pNx19 (10.4)8 (16.6)7 (8.4)4 (7.5)
Distant metastases, n (%)    0.032
 No109 (59.2)35 (72.9)49 (59.0)25 (47.2)
 Yes75 (40.7)13 (27.1)34 (41.0)28 (52.8)

Clear cell RCC was diagnosed in 172 patients, papillary RCC type 2 was found in 11 patients and chromophobe RCC was found in one patient.

The median follow up was 21 months (range 2–220 months). Of the 131 patients (71%) who died, 109 (59%) died from RCC. Thirty-six patients (20%) were alive and disease-free and 17 (9%) patients were alive but presenting distant metastases or local recurrence. The median follow up of the 55 living patients was 45 months (mean 64 months; range 15–220 months). Forty-eight patients did not present TFI, 83 patients showed either PFI or RSFI, and 53 patients presented both PFI and RSFI.

Eleven patients presented pT3c renal cell carcinoma with subdiaphragmatic VTT and IVC wall invasion. Two of these patients presented no TFI, four had either PFI or RSFI, and five had both PFI and RSFI. Median CSM-free survival in these patients was 11 months (range 2–170 months).

The presence of either PFI or RSFI, and invasion of both renal fat compartments were significantly associated with the presence of synchronous nodal or distant metastases, higher tumour grade, and greater tumour dimensions, compared with patients with no TFI (all P < 0.05) (Table 1).

The median CSM-free survival in patients with VTT without TFI, and in patients with VTT and TFI was 99 and 23 months, respectively (P < 0.001; Fig. 1). The median CSM-free survival in patients with VTT without TFI, VTT with either PFI or RSFI, and VTT with both PFI and RSFI was 99, 42 and 10 months, respectively (P < 0.001). The 5-year CSM-free survival estimates were 75%, 36% and 20%, respectively (P < 0.001; Fig. 2). The 5-year CSM-free survival was significantly different between patients with VTT without TFI and those with VTT and either PFI or RSFI (P= 0.002), and between patients with VTT without TFI and those with VTT and both PFI and RSFI (P < 0.001). Furthermore, significantly different survival was found between patients with VTT and either PFI or RSFI and those with both PFI and RSFI (P= 0.022). No survival differences were found between patients with VTT and PFI and those with VTT and RSFI.

Figure 1.

Probability of cancer-specific mortality (CSM)- free survival according to the absence or presence of tumour fat invasion (TFI) (overall population; N= 184).

Figure 2.

Probability of cancer-specific mortality (CSM) -free survival according to the absence of tumour fat invasion (TFI), presence of either perinephric (PFI) or renal sinus (RSFI) fat invasion, or both PFI and RSFI (overall population; N= 184).

In univariable analyses, Fuhrman grade (G2 vs G3 vs G4, P= 0.022), the presence of synchronous metastases (P < 0.001), lymph node involvement (pN0/Nx vs pN+; P < 0.001), presence of sarcomatoid features (P < 0.001) and simultaneous presence of VTT and TFI (no TFI, either PFI or RSFI, or PFI and RSFI; P < 0.001) were all associated with CSM (Table 2). In contrast, pathological stage (2009 TNM staging system) and Eastern Cooperative Oncology Group Performance Status (ECOG PS 1–2 vs 3–4) did not affect CSM.

Table 2.  Univariable and multivariable Cox analyses on CSM in the overall population (N = 184)
VariablesUnivariable analysisMultivariable analyses
HRP valueHRP value
  1. CSM, cancer-specific mortality; RCC, renal cell carcinoma; HR, hazard ratio; TFI, tumour fat invasion; PFI, perinephric fat invasion; RSFI, renal sinus fat invasion.

Distant metastases    
Yes vs No3.86<0.0013.18<0.001
Nodal status    
pN+ vs pN0/pNx2.73<0.0011.840.012
Tumour fat invasion <0.001 0.001
PFI/RSFI only vs No TFI2.42 2.140.008
PFI and RSFI vs No TFI3.29 3.15<0.001
Pathological tumour dimensions    
Continuous variable1.07<0.0011.070.041
Pathological stage (TNM 2009) 0.198 0.389
pT3b vs pT3a1.630.1481.190.504
pT3c vs pT3a1.410.2271.570.184
Sarcomatoid differentiation    
Yes vs No2.33<0.0012.010.025
Tumour grade 0.022 0.579
G3 vs G21.440.1241.200.498
G4 vs G21.790.0161.430.296

In multivariable analyses, after adjusting for all the available clinical and pathological variables, the simultaneous presence of VTT and TFI was independently associated with a higher risk of CSM (P= 0.001). Specifically, the presence of either PFI or RSFI was associated with a two-fold increased risk of CSM relative to VTT-only cases, and the presence of both PFI and RSFI was associated with a three-fold increased risk of CSM relative to VTT-only cases. The other independent predictors of CSM risk were the presence of synchronous metastases (P < 0.001), lymph node involvement (P= 0.012) and sarcomatoid features (P= 0.025). Fuhrman grade and 2009 pT stage did not achieve the status of independent predictors of CSM (Table 2).

The analyses were repeated in the subgroup of pN0/pNxM0 patients (n= 101). The 5-year CSM-free survival estimates were 81%, 50% and 30%, respectively (P= 0.006; Fig. 3). The 5-year CSM-free survival was statistically different between patients with VTT without TFI and those with VTT and either PFI or RSFI (P= 0.042) and between patients with VTT without TFI and those with VTT and both PFI and RSFI (P= 0.001). Furthermore, significantly different survival was found between patients with VTT and either PFI or RSFI and those with both PFI and RSFI (P= 0.031).

Figure 3.

Probability of cancer-specific mortality (CSM) -free survival according to the absence of tumour fat invasion (TFI), presence of either perinephric (PFI) or renal sinus (RSFI) fat invasion, or both PFI and RSFI (pN0/pNxM0 patients; n= 101).

The simultaneous presence of TFI and VTT achieved the independent predictor status of CSM (P= 0.014). Specifically, the presence of either PFI or RSFI was associated with a 2.3-fold increased risk of CSM relative to VTT-only cases, whereas the presence of both PFI and RSFI was associated with a 3.3-fold increased risk of CSM relative to VTT-only cases. The other independent predictor was the presence of sarcomatoid differentiation (P= 0.006; hazard ratio 3.8).

Finally, the simultaneous presence of VTT and TFI was the second most informative predictor (area under the curve = 62%) after the variable depicting the presence of a synchronous metastasis (area under the curve = 68%). Moreover, inclusion of the variable describing the presence and extent of TFI in a base model including pT stage, Fuhrman grade and presence of nodal disease and metastatic disease significantly increased the accuracy in predicting CSM (+ 2.1%; P < 0.001) in patients with VTT.

DISCUSSION

The treatment of RCC with vascular extension is challenging and controversial. Several series reported 5-year survival rates ranging from 25% to 57%[10–12]. Nonetheless, careful patient selection for surgical treatment has been advocated, especially in metastatic disease at diagnosis [12,13].

In our series, both nodal metastasis and distant metastasis at diagnosis emerged as independent predictors of CSM. In contrast, pT stage, according to the novel 2009 pT classification, did not achieve independent predictor status.

Our results confirmed results reported by Klatte et al. [5] and Lambert et al. [14] who found tumour thrombus level to be a poor predictor of survival outcomes. Conversely, we confirmed data from Margulis et al. [8] that showed that the presence of venous thrombosis, but not its extent (renal vein, subdiaphragmatic or supradiaphragmatic) is independently associated with cancer-specific survival within the pT3N0/NxM0 cohort of patients with RCC.

Several studies have investigated the role of renal fat invasion in RCC with neoplastic thrombus [5–9,15]. Leibovich et al. [15] reported that patients with perinephric or sinus fat invasion and neoplastic thrombus have a two-fold higher risk of death from RCC compared with patients with neoplastic thrombus alone, after adjusting for regional lymph node involvement and distant synchronous metastases. Margulis et al. [8] found a similar risk of death from RCC for patients with PFI alone and for those with neoplastic thrombus alone, whereas patients harbouring both features were significantly more likely to die from RCC (P < 0.001). The importance of the prognostic role of renal fat invasion was emphasized by Klatte et al. [5] who proposed a reclassification of pT3b-c based on renal fat involvement. Roos et al. [9] evaluated a cohort of 118 pT3b/c N0M0 patients, 38 of whom had renal fat invasion (10, 9 and 19 with PFI, RSFI, and both PFI and RSFI, respectively). TFI was an independent predictor of CSM. The combination of PFI and RSFI was associated with higher risk of CSM, whereas PFI or RSFI alone was not associated with reduced survival.

Our data confirm that TFI affects CSM in patients with RCC and VTT, independently from nodal and synchronous distant metastases. Furthermore, patients with either PFI or RSFI had a two-fold increased risk of CSM. Moreover, the combination of PFI and RSFI was associated with a three-fold risk of CSM.

The prognostic impact of TFI was further validated in the subgroup of non-metastatic patients, in whom the extent of the tumour in one or both of the fat compartments might be associated with a higher risk of an already micrometastatic disease.

Finally, the simultaneous presence of VTT and TFI was the second most informative predictor after the variable depicting the presence of a synchronous metastasis and inclusion of the variable describing the presence and extent of TFI in a base model including pT stage, Fuhrman grade and presence of nodal disease and metastatic disease significantly increased the accuracy in predicting CSM in patients with VTT.

Our findings support the concept that an accurate evaluation of the extent of TFI should always be included in routine pathology reports to provide better risk stratification of patients with RCC and VTT. Moreover, our data may support the importance of revising and integrating the TNM system for pT3 RCC, including TFI assessment, as proposed by Margulis et al. [8] and Klatte et al. [5]. Indeed, both studies proposed that T3 RCC with either TFI or VTT should be classified as T3a, whereas T3 RCC with VTT and simultaneous TFI should be classified as T3b.

Our study has some limitations. The power of our conclusion may be somewhat limited by the relatively small study population. Nonetheless, all pathology specimens were reviewed by the same expert uropathologist, eliminating the possibility of interobserver variability and allowing uniform application of the definition of all pathological features. Nonetheless, the retrospective nature of our study did not allow a full assessment of the renal sinus specimen, which is the main limitation: it is possible that the incidence of renal sinus fat invasion may be slightly underestimated.

Our study indicates that the simultaneous presence of TFI and VTT was associated with higher risk of CSM than the presence of VTT alone. To our knowledge, we showed for the first time that patients with either PFI or RSFI had a two-fold increased risk of CSM and that the combination of PFI and RSFI was associated with a three-fold risk of CSM. Our results further support the need for an accurate evaluation of the presence and extent of TFI in patients with RCC and VTT, to provide better risk stratification.

CONFLICT OF INTEREST

None declared.

Ancillary