- Email: [email protected]
Impact of Venous Tumour Thrombus Consistency (Solid vs Friable) on Cancer-specific Survival in Patients with Renal Cell Carcinoma
EUROPEAN UROLOGY 60 (2011) 358–365
available at www.sciencedirect.com journal homepage: www.europeanurology.com
Kidney Cancer
Impact of Venous Tumour Thrombus Consistency (Solid vs Friable) on Cancer-specific Survival in Patients with Renal Cell Carcinoma Roberto Bertini a, Marco Roscigno a,c, Massimo Freschi b, Elena Strada a,*, Diego Angiolilli a, Giovanni Petralia a, Rayan Matloob a, Francesco Sozzi a, Umberto Capitanio a, Luigi Filippo Da Pozzo c, Renzo Colombo a, Giorgio Guazzoni a, Anna Cremonini b, Francesco Montorsi a, Patrizio Rigatti a a
Department of Urology, Vita-Salute University San Raffaele, Milan, Italy
b
Department of Pathology, Vita-Salute University San Raffaele, Milan, Italy
c
USC Urologia, Ospedali Riuniti, Bergamo, Italy
Article info
Abstract
Article history: Accepted May 9, 2011 Published online ahead of print on May 24, 2011
Background: To our knowledge, the impact of venous tumour thrombus (VTT) consistency in patients affected by renal cell carcinoma (RCC) has never been addressed. Objective: To analyse the effect of VTT consistency on cancer-specific survival (CSS). Design, setting, and participants: We retrospectively analysed 174 consecutive patients with RCC and renal vein or inferior vena cava (IVC) VTT who underwent surgical treatment between 1989 and 2007 at our institute. Intervention: All patients underwent radical nephrectomy and thrombectomy. Measurements: Pathologic specimens were reviewed by a single uropathologist. In addition to traditional pathologic features, the morphologic aspect of the tumour thrombus was evaluated to distinguish solid from friable patterns. The prognostic role of thrombus consistency (solid vs friable) on CSS was assessed by means of Cox regression models. Results and limitations: The VTT was solid in 107 patients (61.5%) and friable in 67 patients (38.5%). The presence of a friable VTT increased the risk of having synchronous nodal or distant metastases, higher tumour grade, higher pathologic stage, and simultaneous perinephric fat invasion (all p < 0.05). The median follow-up was 24 mo. The median CSS was 33 mo; the median CSS was 8 mo in patients with a friable VTT and 55 mo in patients with a solid VTT ( p < 0.001). On multivariable analyses, the presence of a friable VTT was an independent predictor of CSS ( p = 0.02). The power of our conclusion may be somewhat limited by the relatively small study population and the retrospective nature of the study. Conclusions: In patients with RCC and VTT, the presence of a friable thrombus is an independent predictor of CSS. If our finding is confirmed by further studies, the consistency of the tumour thrombus should be introduced into routine pathologic reports to provide better patient risk stratification. # 2011 European Association of Urology. Published by Elsevier B.V. All rights reserved.
Keywords: Kidney Carcinoma Renal cell Thrombosis Nephrectomy Survival
* Corresponding author. Vita-Salute University San Raffaele, Via Olgettina 61, 20132, Milan, Italy. Tel. +39 2 26437267; Fax: +39 2 26437267. E-mail address: [email protected] (E. Strada). 0302-2838/$ – see back matter # 2011 European Association of Urology. Published by Elsevier B.V. All rights reserved.
doi:10.1016/j.eururo.2011.05.029
EUROPEAN UROLOGY 60 (2011) 358–365
1.
Introduction
359
Health Organisation 2004 classification and nuclear grading according to the Fuhrman classification. In addition to traditional pathologic features
The prevalence of venous thrombus ranges between 4%and 36% of newly diagnosed renal cell carcinoma (RCC) cases [1]. Controversies exist about the impact of the extension of a venous tumour thrombus (VTT) on cancer-specific survival (CSS). Specifically, although some authors have suggested that cephalic extension of VTT may represent a negative predictor of CSS [2–7], others showed no difference in survival according to thrombus level [8–10]. Moreover, some reports recently highlighted that the prognosis may be extremely poor in cases of simultaneous perinephric fat invasion (PFI) and VTT [11,12]. Conversely, although several papers described technical difficulties in the management of the tumour thrombus owing to its friable consistency, none analysed whether a solid or a friable tumour thrombus might be associated with a different prognosis [13,14]. To that end, after pathologically reevaluating all surgical specimens, we retrospectively evaluated our series of 174 consecutive patients with RCC and renal vein or inferior vena cava (IVC) tumour thrombosis to analyse the prognostic impact of VTT consistency on CSS.
(Table 1), the morphologic aspect of the tumour thrombus was evaluated to distinguish solid versus friable character. In a solid thrombus (Fig. 1), tumour growth was compact and cohesive, with a rounded linear profile and, sometimes, a partial endothelial lining simulating a pseudocapsule. In a friable thrombus, tumour cells were intermingled with abundant necrosis and fibrin and had a scalloped, irregular profile and fragmented aspect, sometimes with thin papillary features. Thrombus was classified as solid when at least 90% of the samples showed solid features. Moreover, we tested the concordance between surgical consistency (thrombus consistency as described by the surgeon during nephrectomy, available in the 61 patients with sub- or supradiaphragmatic IVC thrombus) and pathologic consistency.
2.4.
After surgery, oncologic follow-up required chest and abdominal imaging studies (abdominal ultrasound and chest x-rays or thoracoabdominal CT) every 4 mo for the first year, every 6 mo up to the fifth 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, from information from patients’ relatives, or from death certificates.
2.5.
2.
Follow-up
Statistical analysis
Material and methods The survival intervals were calculated from the date of surgery to the last
After obtaining approval from the San Raffaele Hospital ethics committee, we retrospectively analysed the database of 1468 consecutive RCC patients who had undergone partial or radical nephrectomy at our department (1989–2007). One hundred seventy-four patients (11.8%) with renal vein or IVC tumour thrombosis were included.
2.1.
Preoperative assessment
The level of the tumour thrombus was classified as limited to the renal vein, extending into the IVC below the diaphragm and extending above the diaphragm to the atrium. The cranial extent of the tumour thrombus was initially assessed by venacavography and later by magnetic resonance (MR) angiography or computed tomography (CT) angiography. Transesophageal echocardiography was used to assess cardiac tumour involvement in selected cases. The presence or absence of synchronous metastases at the time of referral for surgery was evaluated by abdominal CT or MR scans, chest x-rays or CT scans, and bone scans. CT or MR imaging of the brain was performed when clinically indicated.
2.2.
Therapeutic approach
Direct preoperative embolisation of the renal artery was performed in nine patients (5%). All patients underwent radical nephrectomy and thrombectomy through a midline abdominal or chevron approach. Median sternotomy or right anterior minithoracotomy [13] was performed to provide access to the right atrium. Cardiopulmonary bypass with deep hypothermic cardiac arrest was performed in 35 patients.
2.3.
Pathologic findings
All microscopic slides were retrospectively reassessed by a single uropathologist (MF) who had no knowledge of patient outcomes. Pathologic staging was performed according to the 2009 TNM classification [15], with histologic subtyping according to the World
available follow-up (in the case of cancer-related death, the date of death). Survival curves were estimated by means of Kaplan-Meier methods. Differences among the survival curves were determined by the log-rank test. Differences among categoric variables were evaluated by means of the x2 test. Differences among numeric variables were evaluated by means of the student t test. Multivariable Cox regression analyses (with the backward conditional method) addressed the association among age at diagnosis, the presence of synchronous metastases, lymph node involvement, Fuhrman grade, Eastern Cooperative Oncology Group performance status (ECOG PS; 0–1 vs 2–4), the level of the tumour thrombus (renal vein, thrombus extending into the IVC below the diaphragm [subdiaphragmatic IVC], or thrombus extending above the diaphragm to the atrium [supradiaphragmatic IVC/right atrium]), the presence of tumour necrosis, the presence of sarcomatoid features, the simultaneous presence of VTT and PFI, the presence of a friable or solid thrombus, and CSS. Statistical significance was defined as a p value <0.05. Statistical analyses were performed using MedCalc v.9.0.1.1 (MedCalc Software, Mariakerke, Belgium).
3.
Results
The clinical and pathologic characteristics of the 174 patients are summarised in Table 1 according to the consistency of the tumour thrombus (solid vs friable). Seventy-nine patients were pN0/NxM0, whereas 17 and 78 patients were pN+M0 and pN(any)M+, respectively. Of these last two subgroups, 31 patients underwent immunotherapy (with the majority receiving an interleukin-2–based regimen). Nonetheless, immunotherapy did not positively affect patient outcomes in this subset population ( p = 0.4). The median follow-up was 24 mo (mean: 40 mo; range: 2–220 mo). One hundred seventeen patients (67%) had died, 111 (63%) from RCC. Forty patients (23%) were alive and disease free, whereas 17 (10%) patients were alive but
360
EUROPEAN UROLOGY 60 (2011) 358–365
Table 1 – Patient characteristics and descriptive statistics Variable Age, yr Mean (median) Range ECOG PS, no. (%) 0–1 2 Cranial thrombus extent, no. (%) Renal vein Subdiaphragmatic IVC Supradiaphragmatic IVC/right atrium Pathologic grade, no. (%) 2 3 4 Pathologic stage, no. (%) pT3a pT3b pT3c pT4 Nodal status, no. (%) pN0/pNx pN+ Distant metastases, no. (%) No Yes Simultaneous PFI, no. (%) No Yes Sarcomatoid differentiation, no. (%) No Yes Tumour necrosis, no. (%) No Yes
All patients (n = 174)
Friable VTT (n = 67)
Solid VTT (n = 107)
62.5 (63.1) 26–83
60.1 (61.8) 35–83
64.4 (65.2) 26–78
118 (67.8) 56 (32.2)
41 (61.2) 26 (38.8)
77 (72.0) 30 (28.0)
113 (64.9) 45 (25.9) 16 (9.2)
38 (56.7) 20 (29.8) 9 (13.5)
75 (70.1) 25 (23.4) 7 (6.5)
35 (20.1) 94 (54.0) 45 (25.9)
10 (14.9) 32 (47.8) 25 (37.3)
25 (23.4) 62 (57.9) 20 (18.7)
97 37 22 18
(55.7) (21.2) (12.7) (10.4)
28 (41.8) 16(23.9) 11 (16.4) 12 (17.9)
69 (64.5) 21(19.6) 11 (10.3) 6 (5.6)
126 (72.4) 48 (27.6)
42 (62.7) 25 (37.3)
84 (78.5) 23 (21.5)
97 (55.7) 77 (44.3)
29 (43.3) 38 (56.7)
68 (63.5) 39 (36.5)
41 (23.6) 133 (76.4)
9 (13.4) 58 (86.6)
32 (29.9) 75 (70.1)
157 (90.3) 17 (9.7)
60 (89.6) 7 (10.4)
97 (90.6) 10 (9.4)
29 (16.7) 145 (83.3)
5 (7.5 62 (92.5)
24 (22.4) 83 (77.6)
p value 0.324*
0.138§
0.160§
0.013§
0.016§
0.045§
0.009§
0.020§
0.971§
0.005§
ECOG PS = Eastern Cooperative Oncology Group performance status; PFI = perinephric fat invasion; IVC = inferior vena cava; VTT = venous tumour thrombus. Student t test. § x2 test. *
presenting distant metastases or local recurrence. The median follow-up of the 40 patients still alive and disease free was 55 mo (range: 3–220 mo). With respect to pathologic stage, 97 patients presented pT3a disease (according to the new 2009 TNM), whereas 37, 22, and 18 patients presented with pT3b, pT3c, and pT4 disease, respectively. The tumour thrombus was solid in 107 patients (61.5%). Of those patients, 69, 21, 11, and 6 presented with pT3a, pT3b, pT3c, and pT4 disease, respectively. The thrombus was friable in 67 patients (38.5%), with 28, 16, 11, and 12 patients having pT3a, pT3b, pT3c, and pT4 disease, respectively. The presence of friable thrombus was significantly associated with higher pathologic stage ( p = 0.016). Moreover, the presence of a friable tumour thrombus was strongly associated with the presence of synchronous nodal or distant metastases, higher tumour grade and stage, tumour necrosis, and simultaneous PFI (x2 test: all p < 0.05). In subdiaphragmatic and supradiaphragmatic IVC thrombus, surgical and pathologic friable thrombus was recorded in 34 and 29 cases, respectively. Moreover, excellent concordance was demonstrated between surgical and pathologic evaluation of thrombus consistency (k = 0.83).
Finally, to analyse the risk of a subjective interpretation of the thrombus consistency, a second pathologic revision of the available specimens was performed. We submitted 60 specimens (35%) to pathologic evaluation by a second, independent pathologist. In six cases (10%), thrombus consistency differed from initial evaluation. Specifically, in 2 of 20 specimens (10%), solid thrombus was finally classified as friable. Similarly, in 4 of 40 specimens (10%), friable thrombus was finally classified as solid. Good concordance was demonstrated between the two pathologists (k = 0.78). 3.1.
Impact of thrombus consistency on cancer-specific
mortality
Median overall survival (OS) was 24 mo. Median OS was 8 mo for patients with friable thrombus and 44 mo for patients with solid thrombus. The actuarial 5-yr OS estimate was 15% in patients with friable thrombus and 41% in patients with solid thrombus. The median CSS was 32 mo, and median recurrence-free survival (RFS) was 10 mo. The median CSS in pN0/NxM0 patients, pN+M0 patients, and pN(any)M+ patients was 114, 32, and 11 mo, respectively. The actuarial 5-yr CSS estimate was 62%, 18%, and 15%, respectively.
361
EUROPEAN UROLOGY 60 (2011) 358–365
[()TD$FIG]
and the presence of a friable thrombus ( p < 0.001; Fig. 2) were associated with patient outcomes (Table 2). In contrast, the level of tumour thrombus did not affect CSS in analyses coded for three different levels (renal vein, subdiaphragmatic IVC thrombus, and supradiaphragmatic IVC/right atrium thrombus; p = 0.2) or in analyses of renal vein tumour thrombosis versus IVC/right atrium involvement ( p = 0.5).
On univariable analysis, Fuhrman grade (G2, G3, or G4; p < 0.001), the presence of synchronous metastases ( p < 0.001), lymph node involvement ( p < 0.001), ECOG PS (0–1 vs 2–4; p = 0.045), the simultaneous presence of PFI and VTT ( p < 0.001), pathologic stage (pT3a, pT3b, pT3c, or pT4; p < 0.001), the presence of sarcomatoid features ( p < 0.001), the presence of tumour necrosis ( p = 0.009),
100
Survivval probabiliity (%)
[()TD$FIG]
Fig. 1 – Microscopic view of solid and friable tumour thrombi.
80
60
Tumor thrombus Friable Solid
40
20
0 0
50
150
100
200
250
Months Patients at risk
1-yr
3-yrs
5-yrs
Friable
27
18
8
Solid
87
68
36
Log-rank
P<0.001
Fig. 2 – Probability of cancer-specific survival in the overall population according to the presence of solid or friable tumour thrombus (n = 174).
362
EUROPEAN UROLOGY 60 (2011) 358–365
Table 2 – Clinicopathologic features and cancer-specific survival Variable ECOG PS 0–1 2 Cranial thrombus extent Renal vein Subdiaphragmatic IVC Supradiaphragmatic IVC/right atrium Pathologic grade 2 3 4 Pathologic stage pT3a pT3b pT3c pT4 Nodal status pN0/pNx pN+ Distant metastases No Yes Venous tumour thrombus Friable Solid Simultaneous PFI No Yes Sarcomatoid differentiation No Yes Tumour necrosis No Yes
No. of patients (n = 174), (%)
Actuarial 5-yr CSS, %
Median CSS, mo
118 (67.8) 56 (32.2)
41.2 25.9
42 16
113 (64.9) 45 (25.9) 16 (9.2)
37.7 33.6 45.2
28 19 44
35 (20.1) 94 (54.0) 45 (25.9)
59.7 39.4 16.5
76 40 10
97 37 22 18
(55.7) (21.2) (12.7) (10.4)
42.0 38.5 41.2 0
39 42 44 6
126 (72.4) 48 (27.6)
47.3 11.0
46 8
97 (55.7) 77 (44.3)
55.5 15.3
88 12
67 (38.5) 107 (61.5)
21.0 47.2
8 55
41 (23.6) 133 (76.4)
72.8 25.9
131 19
157 (90.3) 17 (9.7)
41.0 7.4
40 6
29 (16.7) 145 (83.3)
60.5 31.1
131 26
p value (log-rank test) 0.045
0.229
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
0.009
CSS = cancer-specific survival; ECOG PS = Eastern Cooperative Oncology Group performance status; PFI = perinephric fat invasion; IVC = inferior vena cava.
At multivariable analysis, after adjusting for the other clinical and pathologic variables, the presence of a friable tumour thrombus was independently associated with a lower CSS ( p = 0.02). The other independent predictors of CSS were the presence of synchronous metastases ( p < 0.001), lymph node involvement ( p < 0.001), concomitant PFI and VTT ( p = 0.003), and the presence of sarcomatoid features ( p = 0.001). ECOG PS, pathologic stage, tumour necrosis, and tumour grade did not achieve the status of independent predictors of CSS (Table 3). The presence of friable thrombus was independently associated with lower OS ( p = 0.001). The other independent predictors of OS were the presence of synchronous metastases ( p < 0.001), lymph node involvement ( p < 0.001), concomitant PFI and VTT ( p = 0.004), and the presence of sarcomatoid features ( p = 0.043). The analyses were repeated in the subgroup of pN0/ pNxM0 patients (n = 79). The existence of concomitant PFI and VTT and the presence of a friable thrombus were independent predictors of CSS (Table 4). Furthermore, the independent prognostic role of thrombus consistency was confirmed in the subgroup of patients with VTT limited to the renal vein (n = 113; p = 0.04; hazard ratio: 0.56). Finally, the presence of a friable thrombus was the second-most informative predictor (area under the curve
[AUC]: 62%) after the variable depicting the presence of a synchronous metastasis (AUC: 68%). Moreover, inclusion of the variable ‘‘friable thrombus’’ in a base model including pathologic stage, Fuhrman grade, and the presence of nodal disease and metastatic disease significantly increased the accuracy in predicting CSS (+2.5%; p < 0.001). 4.
Discussion
The treatment of RCC with vascular invasion is challenging and controversial. Several series reported a 5-yr survival rate ranging from 25% to 57% [1–9]. Nonetheless, careful patient selection for surgical treatment has been advocated, especially in metastatic disease at diagnosis [14,16]. Although some RCC with VTT can be cured surgically, recurrence developed in a significant proportion of patients. Recently, Kirkali and Van Poppel [17] highlighted the importance of patient selection in patients with high risk of recurrence who may benefit from adjuvant treatment with targeted agents. Several investigators have combined various prognostic factors to obtain a system that enables outcome prediction. The University of California, Los Angeles Integrated Staging System incorporates TNM stage, tumour grade, and ECOG PS [18]. Finally, neither tumour grade nor ECOG PS were independent predictors of CSS in
363
EUROPEAN UROLOGY 60 (2011) 358–365
Table 3 – Univariable and multivariable Cox regression analyses predicting cancer-specific survival in the overall population (n = 174)* Variables
Distant metastases Yes vs no Nodal status PN+ vs pN0/pNx Venous tumor thrombus Solid vs friable Sarcomatoid differentiation Yes vs no Simultaneous PFI Yes vs no Age Continuous variable Tumour grade G3 vs G2 G4 vs G2 ECOG PS 0–1 vs 2–4 Pathologic stage (TNM 2009) pT3b vs pT3a pT3c vs pT3a pT4 vs pT3a Tumour necrosis Yes vs no
Univariable analyses
Multivariable analyses (reduced model)
HR
p value
HR
p value
3.86
<0.001
2.28
<0.001
1.73
<0.001
1.50
<0.001
0.42
<0.001
0.62
0.021
2.67
<0.001
2.57
0.001
2.85
<0.001
2.22
0.003
1.01
0.128 <0.001 0.125 <0.001
– –
– –
–
– –
1.63 1.41 1.46
0.045 <0.001 0.148 0.017 <0.001
–
2.16
0.009
–
1.37 2.68 0.688
–
CSS = cancer-specific survival; RCC = renal cell carcinoma; HR = hazard ratio; PFI = perinephric fat invasion; ECOG PS = Eastern Cooperative Oncology Group performance status. * Multivariable results were included only for variables included in the final reduced model after backward conditional selection.
Table 4 – Univariable and multivariable Cox analyses on cancer-specific survival in patients with pN0/Nx M0 disease (n = 79)* Variables
Venous tumour thrombus Solid vs friable Simultaneous PFI Yes vs no Age Continuous variable Tumour grade G3 vs G2 G4 vs G2 ECOG PS 0–1 vs 2–4 Pathologic stage (TNM 2009) pT3b vs pT3a pT3c vs pT3a pT4 vs pT3a Tumour necrosis Yes vs no
Univariable analyses
Multivariable analyses
HR
p value
HR
p value
0.43
0.023
0.48
0.047
2.65
0.025
2.45
0.039
1.01
0.239 0.142 0.861 0.124
– –
– –
– –
– –
0.89 0.73 4.49
0.765 0.164 0.812 0.559 0.043
1.79
0.029
–
–
0.91 2.03 1.16
CSS = cancer-specific survival; RCC = renal cell carcinoma; HR = hazard ratio; PFI = perinephric fat invasion; ECOG PS = Eastern Cooperative Oncology Group performance status. * Multivariable results were included only for variables included in the final reduced model after backward conditional selection.
our analysis, although these findings may be due to the relatively small number of patients included or to the retrospective nature of our analyses. Our study confirms the prognostic role of PFI concomitant with VTT in patients with RCC, as previously reported in other studies [11,12,19]. The impact of concomitant PFI was further emphasised by the fact that the simultaneous
presence of both venous thrombus and PFI was associated with lower CSS, even in the subgroup of nonmetastatic patients. Moreover, sarcomatoid features were independently associated with poor prognosis in our series, as previously demonstrated by Klatte et al [9]. In contrast, tumour necrosis, the prognostic role of which is still controversial
364
EUROPEAN UROLOGY 60 (2011) 358–365
[19,20], did not achieve independent predictor status. To our knowledge, no previous reports have analysed the impact of thrombus consistency on CSS, and this is why we tested the hypothesis that this new parameter may help in identifying high-risk patients. All surgeons are aware that VTT may present with different characteristics and may be solid/spongeous or friable. This last feature may cause some technical difficulties in the management of tumour thrombosis, especially when it invades the IVC up to the right atrium. We tried to standardise the pathologic definition of tumour thrombus consistency as a solid thrombus that presents with compact and cohesive tumour growth and a pseudocapsulelike structure sometimes surrounding the VTT. In a friable thrombus, the pseudocapsule is lacking; tumour cells are intermingled with abundant necrosis and fibrin; and the thrombus has a scalloped, irregular profile and a fragmented aspect. The presence of a friable thrombus significantly affected patient prognosis at univariable logrank analysis. Moreover, a friable thrombus was an independent predictor of CSS at multivariable Cox regression analysis after adjusting for the other clinical and pathologic confounders. The negative impact of a friable thrombus was confirmed both in the overall population and in nonmetastatic patients. These findings suggest that a friable thrombus might be associated with a high risk of tumour haematogenous spread and, in clinically nonmetastatic patients, could provide a higher risk of systemic progression. If our finding is confirmed by other retrospective or prospective studies, the consistency of the tumour thrombus should be introduced into routine pathologic reports to provide further information that may help in patient selection for adjuvant targeted therapy. Thrombus consistency is the second-most informative predictor of CSS and may improve the identification of patients with a high risk of progression and death. Moreover, if the effect of thrombus consistency on CSS is validated, enhanced understanding of the biologic factors responsible for the friable versus solid phenotype will be crucial in identifying potential novel adjuvant therapies. Our study is not devoid of limitations. The power of our conclusion may be limited somewhat by the relatively small study population and the retrospective nature of the study. Nonetheless, all pathologic specimens were reviewed by a single expert uropathologist, thus eliminating the possibility of intra- and interobserver variability and allowing uniform application of the definition of all pathologic features. Good concordance was demonstrated between the two pathologists who performed pathologic evaluation of the thrombus consistency. This result supports the reproducibility of our proposed pathologic definition of the tumour thrombus consistency. It could be asserted that the friable features of the tumour thrombus might be the result of thrombus manipulation during surgery rather than a real expression of biologic aggressiveness. Nonetheless, we found reproducible results in the subgroup of renal vein VTT, which usually received less manipulation during radical nephrectomy. Moreover, the interrater reliability, using the
k statistic, showed good concordance between pathologic and surgical features, as described by the surgeons during cavotomy and thrombectomy in patients with IVC thrombus. 5.
Conclusions
We reported for the first time that in patients with RCC and VTT, the presence of a friable thrombus is an independent predictor of CSS. Synchronous metastases, nodal involvement, simultaneous presence of PFI and VTT, and presence of sarcomatoid features are also predictors of CSS. The presence of a friable thrombus could help in identifying patients at high risk for progression who may benefit from adjuvant treatment with targeted agents. More research is needed to better clarify the prognostic role of thrombus consistency for patient outcomes and to suggest its introduction into routine pathologic reports. Author contributions: Elena Strada had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Bertini. Acquisition of data: Strada, Angiolilli, Petralia, Sozzi, Matloob. Analysis and interpretation of data: Roscigno, Bertini, Da Pozzo. Drafting of the manuscript: Roscigno, Bertini, Strada. Critical revision of the manuscript for important intellectual content: Montorsi, Colombo, Rigatti, Guazzoni. Statistical analysis: Capitanio. Obtaining funding: None. Administrative, technical, or material support: Freschi, Cremonini. Supervision: Montorsi. Other (specify): None. Financial disclosures: I certify that all conflicts of interest, including specific financial interests and relationships and affiliations relevant to the subject matter or materials discussed in the manuscript (eg, employment/ affiliation, grants or funding, consultancies, honoraria, stock ownership or options, expert testimony, royalties, or patents filed, received, or pending), are the following: None. Funding/Support and role of the sponsor: None.
References [1] Pouliot F, Shuch B, LaRochelle JC, Pantuck A, Belldegrun AS. Contemporary management of renal tumors with renal tumor thrombus. J Urol 2010;184:833–41. [2] Moinzadeh A, Libertino JA. Prognostic significance of tumor thrombus level in patients with renal cell carcinoma and venous tumor thrombus extension. Is all T3b the same? J Urol 2004;171:598–601. [3] Glazer AA, Novick AC. Long-term followup after surgical treatment for renal cell carcinoma extending into the right atrium. J Urol 1996; 155:448–50. [4] Novara G, Ficarra V, Antonelli A, et al. Validation of the 2009 TNM version in a large multi-institutional cohort of patients treated for renal cell carcinoma: are further improvements needed? Eur Urol 2010;58:588–95, Corrigendum. Eur Urol 2011;59:182. [5] Al Otaibi M, Abou Youssif T, Alkhaldi A, et al. Renal cell carcinoma with inferior vena caval extention: impact of tumour extent on surgical outcome. BJU Int 2009;104:1467–70.
EUROPEAN UROLOGY 60 (2011) 358–365
365
[6] Wagner B, Patard J-J, Me´jean A, et al. Prognostic value of renal vein
[13] Bertini R, Roscigno M, Lapenna E, et al. Radical nephrocapsulectomy
and inferior vena cava involvement in renal cell carcinoma. Eur Urol
and caval thrombectomy with extracorporeal circulation and deep
2009;55:452–60.
hypothermic circulatory arrest in right anterior minithoracotomy:
[7] Haferkamp A, Bastian PJ, Jakobi H, et al. Renal cell carcinoma with tumor thrombus extension into the inferior vena cava: prospective long-term follow-up. J Urol 2007;177:1703–8. [8] Lambert EH, Pierorazio PM, Shabsigh A, et al. Prognostic risk stratification and clinical outcomes in patients undergoing surgical treatment for renal cell carcinoma with vascular tumor thrombus. Urology 2007;69:1054–8.
a minimally invasive approach. Urology 2008;71:957–61. [14] Bissada NK, Yakout HH, Babanouri A, et al. Long-term experience with management of renal cell carcinoma involving the inferior vena cava. Urology 2003;61:89–92. [15] Edge SB, Byrd DR, Carducci MA, et al. AJCC cancer staging manual. ed 7. New York, NY: Springer; 2009. [16] Montie JE, Pontes JE, Novick AC, et al. Resection of inferior vena cava
[9] Klatte T, Pantuck AJ, Riggs SB, et al. Prognostic factors for renal cell
tumor thrombi from renal cell carcinoma. Am Surg 1991;57:56–61.
carcinoma with tumor thrombus extension. J Urol 2007;178:
[17] Kirkali Z, Van Poppel H. A critical analysis of surgery for kidney
1189–95.
cancer with vena cava invasion. Eur Urol 2007;52:658–62.
[10] Staehler G, Brkovic D. The role of radical surgery for renal cell
[18] Zisman A, Pantuck AJ, Dorey F, et al. Mathematical model to predict
carcinoma with extension into the vena cava. J Urol 2000;163:
individual survival for patients with renal cell carcinoma. J Clin
1671–5.
Oncol 2002;20:1368–74.
[11] Thompson RH, Cheville JC, Lohse CM, et al. Reclassification of
[19] Ficarra V, Galfano A, Guille´ F, et al. A new staging system for locally
patients with pT3 and pT4 renal cell carcinoma improves prognos-
advanced (pT3-4) renal cell carcinoma: a multicenter European
tic accuracy. Cancer 2005;104:53–60.
study including 2000 patients. J Urol 2007;178:418–24.
[12] Margulis V, Tamboli P, Matin SF, Meisner M, Swanson DA, Wood CG.
[20] Klatte T, Said JW, de Martino M, et al. Presence of tumor necrosis is
Redefining pT3 renal cell carcinoma in the modern era. A proposal
not a significant predictor of survival in clear cell renal cell carci-
for a revision of the current TNM primary tumor classification
noma: higher prognostic accuracy of extent based rather than
system. Cancer 2007;109:2439–44.
presence/absence classification. J Urol 2009;181:1558–64.
available at www.sciencedirect.com journal homepage: www.europeanurology.com
Kidney Cancer
Impact of Venous Tumour Thrombus Consistency (Solid vs Friable) on Cancer-specific Survival in Patients with Renal Cell Carcinoma Roberto Bertini a, Marco Roscigno a,c, Massimo Freschi b, Elena Strada a,*, Diego Angiolilli a, Giovanni Petralia a, Rayan Matloob a, Francesco Sozzi a, Umberto Capitanio a, Luigi Filippo Da Pozzo c, Renzo Colombo a, Giorgio Guazzoni a, Anna Cremonini b, Francesco Montorsi a, Patrizio Rigatti a a
Department of Urology, Vita-Salute University San Raffaele, Milan, Italy
b
Department of Pathology, Vita-Salute University San Raffaele, Milan, Italy
c
USC Urologia, Ospedali Riuniti, Bergamo, Italy
Article info
Abstract
Article history: Accepted May 9, 2011 Published online ahead of print on May 24, 2011
Background: To our knowledge, the impact of venous tumour thrombus (VTT) consistency in patients affected by renal cell carcinoma (RCC) has never been addressed. Objective: To analyse the effect of VTT consistency on cancer-specific survival (CSS). Design, setting, and participants: We retrospectively analysed 174 consecutive patients with RCC and renal vein or inferior vena cava (IVC) VTT who underwent surgical treatment between 1989 and 2007 at our institute. Intervention: All patients underwent radical nephrectomy and thrombectomy. Measurements: Pathologic specimens were reviewed by a single uropathologist. In addition to traditional pathologic features, the morphologic aspect of the tumour thrombus was evaluated to distinguish solid from friable patterns. The prognostic role of thrombus consistency (solid vs friable) on CSS was assessed by means of Cox regression models. Results and limitations: The VTT was solid in 107 patients (61.5%) and friable in 67 patients (38.5%). The presence of a friable VTT increased the risk of having synchronous nodal or distant metastases, higher tumour grade, higher pathologic stage, and simultaneous perinephric fat invasion (all p < 0.05). The median follow-up was 24 mo. The median CSS was 33 mo; the median CSS was 8 mo in patients with a friable VTT and 55 mo in patients with a solid VTT ( p < 0.001). On multivariable analyses, the presence of a friable VTT was an independent predictor of CSS ( p = 0.02). The power of our conclusion may be somewhat limited by the relatively small study population and the retrospective nature of the study. Conclusions: In patients with RCC and VTT, the presence of a friable thrombus is an independent predictor of CSS. If our finding is confirmed by further studies, the consistency of the tumour thrombus should be introduced into routine pathologic reports to provide better patient risk stratification. # 2011 European Association of Urology. Published by Elsevier B.V. All rights reserved.
Keywords: Kidney Carcinoma Renal cell Thrombosis Nephrectomy Survival
* Corresponding author. Vita-Salute University San Raffaele, Via Olgettina 61, 20132, Milan, Italy. Tel. +39 2 26437267; Fax: +39 2 26437267. E-mail address: [email protected] (E. Strada). 0302-2838/$ – see back matter # 2011 European Association of Urology. Published by Elsevier B.V. All rights reserved.
doi:10.1016/j.eururo.2011.05.029
EUROPEAN UROLOGY 60 (2011) 358–365
1.
Introduction
359
Health Organisation 2004 classification and nuclear grading according to the Fuhrman classification. In addition to traditional pathologic features
The prevalence of venous thrombus ranges between 4%and 36% of newly diagnosed renal cell carcinoma (RCC) cases [1]. Controversies exist about the impact of the extension of a venous tumour thrombus (VTT) on cancer-specific survival (CSS). Specifically, although some authors have suggested that cephalic extension of VTT may represent a negative predictor of CSS [2–7], others showed no difference in survival according to thrombus level [8–10]. Moreover, some reports recently highlighted that the prognosis may be extremely poor in cases of simultaneous perinephric fat invasion (PFI) and VTT [11,12]. Conversely, although several papers described technical difficulties in the management of the tumour thrombus owing to its friable consistency, none analysed whether a solid or a friable tumour thrombus might be associated with a different prognosis [13,14]. To that end, after pathologically reevaluating all surgical specimens, we retrospectively evaluated our series of 174 consecutive patients with RCC and renal vein or inferior vena cava (IVC) tumour thrombosis to analyse the prognostic impact of VTT consistency on CSS.
(Table 1), the morphologic aspect of the tumour thrombus was evaluated to distinguish solid versus friable character. In a solid thrombus (Fig. 1), tumour growth was compact and cohesive, with a rounded linear profile and, sometimes, a partial endothelial lining simulating a pseudocapsule. In a friable thrombus, tumour cells were intermingled with abundant necrosis and fibrin and had a scalloped, irregular profile and fragmented aspect, sometimes with thin papillary features. Thrombus was classified as solid when at least 90% of the samples showed solid features. Moreover, we tested the concordance between surgical consistency (thrombus consistency as described by the surgeon during nephrectomy, available in the 61 patients with sub- or supradiaphragmatic IVC thrombus) and pathologic consistency.
2.4.
After surgery, oncologic follow-up required chest and abdominal imaging studies (abdominal ultrasound and chest x-rays or thoracoabdominal CT) every 4 mo for the first year, every 6 mo up to the fifth 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, from information from patients’ relatives, or from death certificates.
2.5.
2.
Follow-up
Statistical analysis
Material and methods The survival intervals were calculated from the date of surgery to the last
After obtaining approval from the San Raffaele Hospital ethics committee, we retrospectively analysed the database of 1468 consecutive RCC patients who had undergone partial or radical nephrectomy at our department (1989–2007). One hundred seventy-four patients (11.8%) with renal vein or IVC tumour thrombosis were included.
2.1.
Preoperative assessment
The level of the tumour thrombus was classified as limited to the renal vein, extending into the IVC below the diaphragm and extending above the diaphragm to the atrium. The cranial extent of the tumour thrombus was initially assessed by venacavography and later by magnetic resonance (MR) angiography or computed tomography (CT) angiography. Transesophageal echocardiography was used to assess cardiac tumour involvement in selected cases. The presence or absence of synchronous metastases at the time of referral for surgery was evaluated by abdominal CT or MR scans, chest x-rays or CT scans, and bone scans. CT or MR imaging of the brain was performed when clinically indicated.
2.2.
Therapeutic approach
Direct preoperative embolisation of the renal artery was performed in nine patients (5%). All patients underwent radical nephrectomy and thrombectomy through a midline abdominal or chevron approach. Median sternotomy or right anterior minithoracotomy [13] was performed to provide access to the right atrium. Cardiopulmonary bypass with deep hypothermic cardiac arrest was performed in 35 patients.
2.3.
Pathologic findings
All microscopic slides were retrospectively reassessed by a single uropathologist (MF) who had no knowledge of patient outcomes. Pathologic staging was performed according to the 2009 TNM classification [15], with histologic subtyping according to the World
available follow-up (in the case of cancer-related death, the date of death). Survival curves were estimated by means of Kaplan-Meier methods. Differences among the survival curves were determined by the log-rank test. Differences among categoric variables were evaluated by means of the x2 test. Differences among numeric variables were evaluated by means of the student t test. Multivariable Cox regression analyses (with the backward conditional method) addressed the association among age at diagnosis, the presence of synchronous metastases, lymph node involvement, Fuhrman grade, Eastern Cooperative Oncology Group performance status (ECOG PS; 0–1 vs 2–4), the level of the tumour thrombus (renal vein, thrombus extending into the IVC below the diaphragm [subdiaphragmatic IVC], or thrombus extending above the diaphragm to the atrium [supradiaphragmatic IVC/right atrium]), the presence of tumour necrosis, the presence of sarcomatoid features, the simultaneous presence of VTT and PFI, the presence of a friable or solid thrombus, and CSS. Statistical significance was defined as a p value <0.05. Statistical analyses were performed using MedCalc v.9.0.1.1 (MedCalc Software, Mariakerke, Belgium).
3.
Results
The clinical and pathologic characteristics of the 174 patients are summarised in Table 1 according to the consistency of the tumour thrombus (solid vs friable). Seventy-nine patients were pN0/NxM0, whereas 17 and 78 patients were pN+M0 and pN(any)M+, respectively. Of these last two subgroups, 31 patients underwent immunotherapy (with the majority receiving an interleukin-2–based regimen). Nonetheless, immunotherapy did not positively affect patient outcomes in this subset population ( p = 0.4). The median follow-up was 24 mo (mean: 40 mo; range: 2–220 mo). One hundred seventeen patients (67%) had died, 111 (63%) from RCC. Forty patients (23%) were alive and disease free, whereas 17 (10%) patients were alive but
360
EUROPEAN UROLOGY 60 (2011) 358–365
Table 1 – Patient characteristics and descriptive statistics Variable Age, yr Mean (median) Range ECOG PS, no. (%) 0–1 2 Cranial thrombus extent, no. (%) Renal vein Subdiaphragmatic IVC Supradiaphragmatic IVC/right atrium Pathologic grade, no. (%) 2 3 4 Pathologic stage, no. (%) pT3a pT3b pT3c pT4 Nodal status, no. (%) pN0/pNx pN+ Distant metastases, no. (%) No Yes Simultaneous PFI, no. (%) No Yes Sarcomatoid differentiation, no. (%) No Yes Tumour necrosis, no. (%) No Yes
All patients (n = 174)
Friable VTT (n = 67)
Solid VTT (n = 107)
62.5 (63.1) 26–83
60.1 (61.8) 35–83
64.4 (65.2) 26–78
118 (67.8) 56 (32.2)
41 (61.2) 26 (38.8)
77 (72.0) 30 (28.0)
113 (64.9) 45 (25.9) 16 (9.2)
38 (56.7) 20 (29.8) 9 (13.5)
75 (70.1) 25 (23.4) 7 (6.5)
35 (20.1) 94 (54.0) 45 (25.9)
10 (14.9) 32 (47.8) 25 (37.3)
25 (23.4) 62 (57.9) 20 (18.7)
97 37 22 18
(55.7) (21.2) (12.7) (10.4)
28 (41.8) 16(23.9) 11 (16.4) 12 (17.9)
69 (64.5) 21(19.6) 11 (10.3) 6 (5.6)
126 (72.4) 48 (27.6)
42 (62.7) 25 (37.3)
84 (78.5) 23 (21.5)
97 (55.7) 77 (44.3)
29 (43.3) 38 (56.7)
68 (63.5) 39 (36.5)
41 (23.6) 133 (76.4)
9 (13.4) 58 (86.6)
32 (29.9) 75 (70.1)
157 (90.3) 17 (9.7)
60 (89.6) 7 (10.4)
97 (90.6) 10 (9.4)
29 (16.7) 145 (83.3)
5 (7.5 62 (92.5)
24 (22.4) 83 (77.6)
p value 0.324*
0.138§
0.160§
0.013§
0.016§
0.045§
0.009§
0.020§
0.971§
0.005§
ECOG PS = Eastern Cooperative Oncology Group performance status; PFI = perinephric fat invasion; IVC = inferior vena cava; VTT = venous tumour thrombus. Student t test. § x2 test. *
presenting distant metastases or local recurrence. The median follow-up of the 40 patients still alive and disease free was 55 mo (range: 3–220 mo). With respect to pathologic stage, 97 patients presented pT3a disease (according to the new 2009 TNM), whereas 37, 22, and 18 patients presented with pT3b, pT3c, and pT4 disease, respectively. The tumour thrombus was solid in 107 patients (61.5%). Of those patients, 69, 21, 11, and 6 presented with pT3a, pT3b, pT3c, and pT4 disease, respectively. The thrombus was friable in 67 patients (38.5%), with 28, 16, 11, and 12 patients having pT3a, pT3b, pT3c, and pT4 disease, respectively. The presence of friable thrombus was significantly associated with higher pathologic stage ( p = 0.016). Moreover, the presence of a friable tumour thrombus was strongly associated with the presence of synchronous nodal or distant metastases, higher tumour grade and stage, tumour necrosis, and simultaneous PFI (x2 test: all p < 0.05). In subdiaphragmatic and supradiaphragmatic IVC thrombus, surgical and pathologic friable thrombus was recorded in 34 and 29 cases, respectively. Moreover, excellent concordance was demonstrated between surgical and pathologic evaluation of thrombus consistency (k = 0.83).
Finally, to analyse the risk of a subjective interpretation of the thrombus consistency, a second pathologic revision of the available specimens was performed. We submitted 60 specimens (35%) to pathologic evaluation by a second, independent pathologist. In six cases (10%), thrombus consistency differed from initial evaluation. Specifically, in 2 of 20 specimens (10%), solid thrombus was finally classified as friable. Similarly, in 4 of 40 specimens (10%), friable thrombus was finally classified as solid. Good concordance was demonstrated between the two pathologists (k = 0.78). 3.1.
Impact of thrombus consistency on cancer-specific
mortality
Median overall survival (OS) was 24 mo. Median OS was 8 mo for patients with friable thrombus and 44 mo for patients with solid thrombus. The actuarial 5-yr OS estimate was 15% in patients with friable thrombus and 41% in patients with solid thrombus. The median CSS was 32 mo, and median recurrence-free survival (RFS) was 10 mo. The median CSS in pN0/NxM0 patients, pN+M0 patients, and pN(any)M+ patients was 114, 32, and 11 mo, respectively. The actuarial 5-yr CSS estimate was 62%, 18%, and 15%, respectively.
361
EUROPEAN UROLOGY 60 (2011) 358–365
[()TD$FIG]
and the presence of a friable thrombus ( p < 0.001; Fig. 2) were associated with patient outcomes (Table 2). In contrast, the level of tumour thrombus did not affect CSS in analyses coded for three different levels (renal vein, subdiaphragmatic IVC thrombus, and supradiaphragmatic IVC/right atrium thrombus; p = 0.2) or in analyses of renal vein tumour thrombosis versus IVC/right atrium involvement ( p = 0.5).
On univariable analysis, Fuhrman grade (G2, G3, or G4; p < 0.001), the presence of synchronous metastases ( p < 0.001), lymph node involvement ( p < 0.001), ECOG PS (0–1 vs 2–4; p = 0.045), the simultaneous presence of PFI and VTT ( p < 0.001), pathologic stage (pT3a, pT3b, pT3c, or pT4; p < 0.001), the presence of sarcomatoid features ( p < 0.001), the presence of tumour necrosis ( p = 0.009),
100
Survivval probabiliity (%)
[()TD$FIG]
Fig. 1 – Microscopic view of solid and friable tumour thrombi.
80
60
Tumor thrombus Friable Solid
40
20
0 0
50
150
100
200
250
Months Patients at risk
1-yr
3-yrs
5-yrs
Friable
27
18
8
Solid
87
68
36
Log-rank
P<0.001
Fig. 2 – Probability of cancer-specific survival in the overall population according to the presence of solid or friable tumour thrombus (n = 174).
362
EUROPEAN UROLOGY 60 (2011) 358–365
Table 2 – Clinicopathologic features and cancer-specific survival Variable ECOG PS 0–1 2 Cranial thrombus extent Renal vein Subdiaphragmatic IVC Supradiaphragmatic IVC/right atrium Pathologic grade 2 3 4 Pathologic stage pT3a pT3b pT3c pT4 Nodal status pN0/pNx pN+ Distant metastases No Yes Venous tumour thrombus Friable Solid Simultaneous PFI No Yes Sarcomatoid differentiation No Yes Tumour necrosis No Yes
No. of patients (n = 174), (%)
Actuarial 5-yr CSS, %
Median CSS, mo
118 (67.8) 56 (32.2)
41.2 25.9
42 16
113 (64.9) 45 (25.9) 16 (9.2)
37.7 33.6 45.2
28 19 44
35 (20.1) 94 (54.0) 45 (25.9)
59.7 39.4 16.5
76 40 10
97 37 22 18
(55.7) (21.2) (12.7) (10.4)
42.0 38.5 41.2 0
39 42 44 6
126 (72.4) 48 (27.6)
47.3 11.0
46 8
97 (55.7) 77 (44.3)
55.5 15.3
88 12
67 (38.5) 107 (61.5)
21.0 47.2
8 55
41 (23.6) 133 (76.4)
72.8 25.9
131 19
157 (90.3) 17 (9.7)
41.0 7.4
40 6
29 (16.7) 145 (83.3)
60.5 31.1
131 26
p value (log-rank test) 0.045
0.229
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
0.009
CSS = cancer-specific survival; ECOG PS = Eastern Cooperative Oncology Group performance status; PFI = perinephric fat invasion; IVC = inferior vena cava.
At multivariable analysis, after adjusting for the other clinical and pathologic variables, the presence of a friable tumour thrombus was independently associated with a lower CSS ( p = 0.02). The other independent predictors of CSS were the presence of synchronous metastases ( p < 0.001), lymph node involvement ( p < 0.001), concomitant PFI and VTT ( p = 0.003), and the presence of sarcomatoid features ( p = 0.001). ECOG PS, pathologic stage, tumour necrosis, and tumour grade did not achieve the status of independent predictors of CSS (Table 3). The presence of friable thrombus was independently associated with lower OS ( p = 0.001). The other independent predictors of OS were the presence of synchronous metastases ( p < 0.001), lymph node involvement ( p < 0.001), concomitant PFI and VTT ( p = 0.004), and the presence of sarcomatoid features ( p = 0.043). The analyses were repeated in the subgroup of pN0/ pNxM0 patients (n = 79). The existence of concomitant PFI and VTT and the presence of a friable thrombus were independent predictors of CSS (Table 4). Furthermore, the independent prognostic role of thrombus consistency was confirmed in the subgroup of patients with VTT limited to the renal vein (n = 113; p = 0.04; hazard ratio: 0.56). Finally, the presence of a friable thrombus was the second-most informative predictor (area under the curve
[AUC]: 62%) after the variable depicting the presence of a synchronous metastasis (AUC: 68%). Moreover, inclusion of the variable ‘‘friable thrombus’’ in a base model including pathologic stage, Fuhrman grade, and the presence of nodal disease and metastatic disease significantly increased the accuracy in predicting CSS (+2.5%; p < 0.001). 4.
Discussion
The treatment of RCC with vascular invasion is challenging and controversial. Several series reported a 5-yr survival rate ranging from 25% to 57% [1–9]. Nonetheless, careful patient selection for surgical treatment has been advocated, especially in metastatic disease at diagnosis [14,16]. Although some RCC with VTT can be cured surgically, recurrence developed in a significant proportion of patients. Recently, Kirkali and Van Poppel [17] highlighted the importance of patient selection in patients with high risk of recurrence who may benefit from adjuvant treatment with targeted agents. Several investigators have combined various prognostic factors to obtain a system that enables outcome prediction. The University of California, Los Angeles Integrated Staging System incorporates TNM stage, tumour grade, and ECOG PS [18]. Finally, neither tumour grade nor ECOG PS were independent predictors of CSS in
363
EUROPEAN UROLOGY 60 (2011) 358–365
Table 3 – Univariable and multivariable Cox regression analyses predicting cancer-specific survival in the overall population (n = 174)* Variables
Distant metastases Yes vs no Nodal status PN+ vs pN0/pNx Venous tumor thrombus Solid vs friable Sarcomatoid differentiation Yes vs no Simultaneous PFI Yes vs no Age Continuous variable Tumour grade G3 vs G2 G4 vs G2 ECOG PS 0–1 vs 2–4 Pathologic stage (TNM 2009) pT3b vs pT3a pT3c vs pT3a pT4 vs pT3a Tumour necrosis Yes vs no
Univariable analyses
Multivariable analyses (reduced model)
HR
p value
HR
p value
3.86
<0.001
2.28
<0.001
1.73
<0.001
1.50
<0.001
0.42
<0.001
0.62
0.021
2.67
<0.001
2.57
0.001
2.85
<0.001
2.22
0.003
1.01
0.128 <0.001 0.125 <0.001
– –
– –
–
– –
1.63 1.41 1.46
0.045 <0.001 0.148 0.017 <0.001
–
2.16
0.009
–
1.37 2.68 0.688
–
CSS = cancer-specific survival; RCC = renal cell carcinoma; HR = hazard ratio; PFI = perinephric fat invasion; ECOG PS = Eastern Cooperative Oncology Group performance status. * Multivariable results were included only for variables included in the final reduced model after backward conditional selection.
Table 4 – Univariable and multivariable Cox analyses on cancer-specific survival in patients with pN0/Nx M0 disease (n = 79)* Variables
Venous tumour thrombus Solid vs friable Simultaneous PFI Yes vs no Age Continuous variable Tumour grade G3 vs G2 G4 vs G2 ECOG PS 0–1 vs 2–4 Pathologic stage (TNM 2009) pT3b vs pT3a pT3c vs pT3a pT4 vs pT3a Tumour necrosis Yes vs no
Univariable analyses
Multivariable analyses
HR
p value
HR
p value
0.43
0.023
0.48
0.047
2.65
0.025
2.45
0.039
1.01
0.239 0.142 0.861 0.124
– –
– –
– –
– –
0.89 0.73 4.49
0.765 0.164 0.812 0.559 0.043
1.79
0.029
–
–
0.91 2.03 1.16
CSS = cancer-specific survival; RCC = renal cell carcinoma; HR = hazard ratio; PFI = perinephric fat invasion; ECOG PS = Eastern Cooperative Oncology Group performance status. * Multivariable results were included only for variables included in the final reduced model after backward conditional selection.
our analysis, although these findings may be due to the relatively small number of patients included or to the retrospective nature of our analyses. Our study confirms the prognostic role of PFI concomitant with VTT in patients with RCC, as previously reported in other studies [11,12,19]. The impact of concomitant PFI was further emphasised by the fact that the simultaneous
presence of both venous thrombus and PFI was associated with lower CSS, even in the subgroup of nonmetastatic patients. Moreover, sarcomatoid features were independently associated with poor prognosis in our series, as previously demonstrated by Klatte et al [9]. In contrast, tumour necrosis, the prognostic role of which is still controversial
364
EUROPEAN UROLOGY 60 (2011) 358–365
[19,20], did not achieve independent predictor status. To our knowledge, no previous reports have analysed the impact of thrombus consistency on CSS, and this is why we tested the hypothesis that this new parameter may help in identifying high-risk patients. All surgeons are aware that VTT may present with different characteristics and may be solid/spongeous or friable. This last feature may cause some technical difficulties in the management of tumour thrombosis, especially when it invades the IVC up to the right atrium. We tried to standardise the pathologic definition of tumour thrombus consistency as a solid thrombus that presents with compact and cohesive tumour growth and a pseudocapsulelike structure sometimes surrounding the VTT. In a friable thrombus, the pseudocapsule is lacking; tumour cells are intermingled with abundant necrosis and fibrin; and the thrombus has a scalloped, irregular profile and a fragmented aspect. The presence of a friable thrombus significantly affected patient prognosis at univariable logrank analysis. Moreover, a friable thrombus was an independent predictor of CSS at multivariable Cox regression analysis after adjusting for the other clinical and pathologic confounders. The negative impact of a friable thrombus was confirmed both in the overall population and in nonmetastatic patients. These findings suggest that a friable thrombus might be associated with a high risk of tumour haematogenous spread and, in clinically nonmetastatic patients, could provide a higher risk of systemic progression. If our finding is confirmed by other retrospective or prospective studies, the consistency of the tumour thrombus should be introduced into routine pathologic reports to provide further information that may help in patient selection for adjuvant targeted therapy. Thrombus consistency is the second-most informative predictor of CSS and may improve the identification of patients with a high risk of progression and death. Moreover, if the effect of thrombus consistency on CSS is validated, enhanced understanding of the biologic factors responsible for the friable versus solid phenotype will be crucial in identifying potential novel adjuvant therapies. Our study is not devoid of limitations. The power of our conclusion may be limited somewhat by the relatively small study population and the retrospective nature of the study. Nonetheless, all pathologic specimens were reviewed by a single expert uropathologist, thus eliminating the possibility of intra- and interobserver variability and allowing uniform application of the definition of all pathologic features. Good concordance was demonstrated between the two pathologists who performed pathologic evaluation of the thrombus consistency. This result supports the reproducibility of our proposed pathologic definition of the tumour thrombus consistency. It could be asserted that the friable features of the tumour thrombus might be the result of thrombus manipulation during surgery rather than a real expression of biologic aggressiveness. Nonetheless, we found reproducible results in the subgroup of renal vein VTT, which usually received less manipulation during radical nephrectomy. Moreover, the interrater reliability, using the
k statistic, showed good concordance between pathologic and surgical features, as described by the surgeons during cavotomy and thrombectomy in patients with IVC thrombus. 5.
Conclusions
We reported for the first time that in patients with RCC and VTT, the presence of a friable thrombus is an independent predictor of CSS. Synchronous metastases, nodal involvement, simultaneous presence of PFI and VTT, and presence of sarcomatoid features are also predictors of CSS. The presence of a friable thrombus could help in identifying patients at high risk for progression who may benefit from adjuvant treatment with targeted agents. More research is needed to better clarify the prognostic role of thrombus consistency for patient outcomes and to suggest its introduction into routine pathologic reports. Author contributions: Elena Strada had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Bertini. Acquisition of data: Strada, Angiolilli, Petralia, Sozzi, Matloob. Analysis and interpretation of data: Roscigno, Bertini, Da Pozzo. Drafting of the manuscript: Roscigno, Bertini, Strada. Critical revision of the manuscript for important intellectual content: Montorsi, Colombo, Rigatti, Guazzoni. Statistical analysis: Capitanio. Obtaining funding: None. Administrative, technical, or material support: Freschi, Cremonini. Supervision: Montorsi. Other (specify): None. Financial disclosures: I certify that all conflicts of interest, including specific financial interests and relationships and affiliations relevant to the subject matter or materials discussed in the manuscript (eg, employment/ affiliation, grants or funding, consultancies, honoraria, stock ownership or options, expert testimony, royalties, or patents filed, received, or pending), are the following: None. Funding/Support and role of the sponsor: None.
References [1] Pouliot F, Shuch B, LaRochelle JC, Pantuck A, Belldegrun AS. Contemporary management of renal tumors with renal tumor thrombus. J Urol 2010;184:833–41. [2] Moinzadeh A, Libertino JA. Prognostic significance of tumor thrombus level in patients with renal cell carcinoma and venous tumor thrombus extension. Is all T3b the same? J Urol 2004;171:598–601. [3] Glazer AA, Novick AC. Long-term followup after surgical treatment for renal cell carcinoma extending into the right atrium. J Urol 1996; 155:448–50. [4] Novara G, Ficarra V, Antonelli A, et al. Validation of the 2009 TNM version in a large multi-institutional cohort of patients treated for renal cell carcinoma: are further improvements needed? Eur Urol 2010;58:588–95, Corrigendum. Eur Urol 2011;59:182. [5] Al Otaibi M, Abou Youssif T, Alkhaldi A, et al. Renal cell carcinoma with inferior vena caval extention: impact of tumour extent on surgical outcome. BJU Int 2009;104:1467–70.
EUROPEAN UROLOGY 60 (2011) 358–365
365
[6] Wagner B, Patard J-J, Me´jean A, et al. Prognostic value of renal vein
[13] Bertini R, Roscigno M, Lapenna E, et al. Radical nephrocapsulectomy
and inferior vena cava involvement in renal cell carcinoma. Eur Urol
and caval thrombectomy with extracorporeal circulation and deep
2009;55:452–60.
hypothermic circulatory arrest in right anterior minithoracotomy:
[7] Haferkamp A, Bastian PJ, Jakobi H, et al. Renal cell carcinoma with tumor thrombus extension into the inferior vena cava: prospective long-term follow-up. J Urol 2007;177:1703–8. [8] Lambert EH, Pierorazio PM, Shabsigh A, et al. Prognostic risk stratification and clinical outcomes in patients undergoing surgical treatment for renal cell carcinoma with vascular tumor thrombus. Urology 2007;69:1054–8.
a minimally invasive approach. Urology 2008;71:957–61. [14] Bissada NK, Yakout HH, Babanouri A, et al. Long-term experience with management of renal cell carcinoma involving the inferior vena cava. Urology 2003;61:89–92. [15] Edge SB, Byrd DR, Carducci MA, et al. AJCC cancer staging manual. ed 7. New York, NY: Springer; 2009. [16] Montie JE, Pontes JE, Novick AC, et al. Resection of inferior vena cava
[9] Klatte T, Pantuck AJ, Riggs SB, et al. Prognostic factors for renal cell
tumor thrombi from renal cell carcinoma. Am Surg 1991;57:56–61.
carcinoma with tumor thrombus extension. J Urol 2007;178:
[17] Kirkali Z, Van Poppel H. A critical analysis of surgery for kidney
1189–95.
cancer with vena cava invasion. Eur Urol 2007;52:658–62.
[10] Staehler G, Brkovic D. The role of radical surgery for renal cell
[18] Zisman A, Pantuck AJ, Dorey F, et al. Mathematical model to predict
carcinoma with extension into the vena cava. J Urol 2000;163:
individual survival for patients with renal cell carcinoma. J Clin
1671–5.
Oncol 2002;20:1368–74.
[11] Thompson RH, Cheville JC, Lohse CM, et al. Reclassification of
[19] Ficarra V, Galfano A, Guille´ F, et al. A new staging system for locally
patients with pT3 and pT4 renal cell carcinoma improves prognos-
advanced (pT3-4) renal cell carcinoma: a multicenter European
tic accuracy. Cancer 2005;104:53–60.
study including 2000 patients. J Urol 2007;178:418–24.
[12] Margulis V, Tamboli P, Matin SF, Meisner M, Swanson DA, Wood CG.
[20] Klatte T, Said JW, de Martino M, et al. Presence of tumor necrosis is
Redefining pT3 renal cell carcinoma in the modern era. A proposal
not a significant predictor of survival in clear cell renal cell carci-
for a revision of the current TNM primary tumor classification
noma: higher prognostic accuracy of extent based rather than
system. Cancer 2007;109:2439–44.
presence/absence classification. J Urol 2009;181:1558–64.