- Email: [email protected]
The role of lymphovascular space invasion in renal cell carcinoma as a prognostic marker of survival after curative resection
Urologic Oncology: Seminars and Original Investigations 29 (2011) 738 –744
Original article
The role of lymphovascular space invasion in renal cell carcinoma as a prognostic marker of survival after curative resection Matthew D. Katz, M.D.a, Maria F. Serrano, M.D.b, Peter A. Humphrey, M.D., Ph.D.b, Robert L. Grubb III, M.D.a,d, Ted A. Skolarus, M.D.a, Feng Gao, M.D., Ph.D.c, Adam S. Kibel, M.D.a,d,* b
a Division of Urologic Surgery, Washington University School of Medicine, St. Louis, MO 63110, USA Division of Anatomic and Molecular Pathology, Department of Pathology and Immunology, Washington University School of Medicine, St. Louis, MO 63110, USA c Division of Biostatistics, Washington University School of Medicine, St. Louis, MO 63110, USA d Alvin J. Siteman Cancer Center, Washington University School of Medicine, St. Louis, MO 63110, USA
Received 20 June 2009; received in revised form 28 July 2009; accepted 29 July 2009
Abstract Objectives: Lymphovascular invasion (LVI) correlates with adverse outcomes in numerous malignancies. However, its role in predicting outcomes in RCC is unclear. Herein, we evaluated what effect LVI had on metastasis free survival (MFS), disease-specific survival (DSS), and overall survival (OS) in patients with RCC treated with surgical excision. Methods: Eight hundred forty-one consecutive patients who underwent partial or radical nephrectomy from 1989 to 2004 were identified. Pathologic and gross features examined were LVI, subtype, Fuhrman grade, stage, and size. Age and gender were also analyzed. Slides were re-reviewed by a single pathologist (MS). Variables with P ⬍ 0.1 on univariate analysis were incorporated in a Cox proportional hazards multivariate model. MFS, DSS, and OS were described for patients with and without LVI using the Kaplan-Meier method, and compared with the log-rank test. Results: LVI was seen on H and E stained slides in 91 patients (11%); 120 (14%) developed metastatic disease, 91 (11%) died of RCC, and 306 (36%) died during a median follow-up of 61 months. While on univariate analysis, LVI was strongly associated with decreased MFS, DSS, and OS (P ⬍ 0.0001), on multivariate analysis, LVI was no longer statistically significant for MFS, DSS, and OS with a HR of 0.976 (95% CI: 0.583–1.63; P ⫽ 0.93), 0.96 (95% CI: 0.542–1.69; P ⫽ 0.88), and 1.24 (95% CI: 0.869 –1.77; P ⫽ 0.24). Conclusions: We found LVI to be associated with worse MFS, DSS, and OS on univariate analysis, but not on multivariate analysis for patients with nonmetastatic RCC. In contrast to previously reported studies, LVI may not be an independent prognostic variable in patients with localized RCC. © 2011 Elsevier Inc. All rights reserved. Keywords: Kidney cancer; Lymphovascular invasion; Nephrectomy; Survival
1. Introduction There will be an estimated 57,760 new cases of renal cancer in the United States, and an estimated 12,980 deaths caused by this disease in 2009 [1]. With the widespread use of cross-sectional imaging for nonspecific abdominal complaints, the rate of incidental renal masses being detected is on the rise. However, even with the resection of localized renal cell carcinoma, up to a third of patients will go on to * Corresponding author. Tel.: ⫹1-314-362-8295; fax: ⫹1-314-454-5244. E-mail address: [email protected] (A.S. Kibel). 1078-1439/$ – see front matter © 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.urolonc.2009.07.034
develop metastatic disease [2– 4]. The prognostic factors of stage, grade, histologic subtype, and presence of sarcomatoid or rhabdoid differentiation are most commonly used to help delineate which patients are at highest risk of recurrent disease [5– 8]. The role of lymphovascular space invasion (LVI) in identifying patients with more aggressive biological forms of renal cell carcinoma has not been fully elucidated. The presence of LVI has been demonstrated to be associated with adverse outcomes in other urologic malignancies, including urothelial carcinoma and testicular germ cell tumors [9 –11]. However, there is conflicting evidence regarding
M.D. Katz et al. / Urologic Oncology: Seminars and Original Investigations 29 (2011) 738 –744
LVI and its importance in predicting worse outcomes in patients with RCC. While several studies have demonstrated the presence of LVI to correlate with decreased survival in patients with RCC [12–21,28], others have not [22–27]. Most of these studies have primarily focused on either disease-free or disease-specific survival (DSS), and there has only been 1 paper that has evaluated all 3 survival endpoints of metastasis-free survival (MFS), DSS, or overall survival (OS) [18]. Thus, in order to further delineate the significance of LVI within the primary tumor in patients treated with radical and partial nephrectomy, we investigated whether the presence of LVI correlated with the 3 survival endpoints of MFS, DSS, and OS.
2. Materials and methods 2.1. Study population After obtaining institutional board approval, we retrospectively reviewed the clinical and pathologic data of patients diagnosed with RCC from 1989 to 2004 treated with radical or partial nephrectomy with curative intent at our institution. Inclusion criteria were: (1) surgical resection of pathologically confirmed RCC with either partial or radical nephrectomy; (2) adequate tissue available for pathologic re-evaluation; and (3) adequate radiographic and or clinical follow-up data available. Exclusion criteria were: (1) pathologically confirmed urothelial carcinoma or any benign lesion; (2) cytoreductive nephrectomy done as part of treatment of known metastatic disease; (3) tissue unavailable for accurate re-evaluation; or (4) no available follow-up clinical or radiographic data. Based on these criteria, a total of 841 eligible patients were included within our analysis. 2.2. Data collection The Barnes-Jewish Hospital Oncology Data Services is an American College of Surgeons Commission on Cancer approved registry. Data on all malignancies diagnosed at our institution are extracted prospectively from charts by certified registrars, beginning at the time of diagnosis. Data collected by tumor registrars for all tumors include standard tumor registry data elements: date of diagnosis, demographic information, stage, grade and nodal spread at time of presentation, initial therapy information, follow-up information, subsequent therapy information, and outcome information. Data entry into the database was done by the tumor registrars. Retrospective chart review was performed to determine dates of recurrence, date and cause of death.
739
2.3. Pathology All hematoxylin and eosin stained (H and E) slides were re-reviewed by a single pathologist (MS) who was blinded to clinical outcome data. Tumors were classified according to the 2004 World Health Organization classification system; TNM stage was classified according to the 2002 American Joint Committee on Cancer (AJCC) staging system. Tumors were graded according to the conventional Fuhrman grading system as 1, 2, 3, or 4. All tumors were assessed for the presence of LVI. Primary tumor size was recorded from the surgical pathology reports. 2.4. Evaluation for lymphovascular invasion LVI was considered present if any tumor cells were seen within the luminal space lined by endothelial cells by visual inspection on H and E stained slides from each tumor. The identification of unequivocal endothelial lining was mandatory. No distinction was made between lymphatic and vascular invasion. Renal vein involvement was defined as tumor cells in a large vein with a muscular wall in the renal sinus. This was not categorized as LVI. A mean of 5 slides
Table 1 Baseline demographic, clinical, and tumor features (n ⫽ 841) Variables
Categories
n
% Patients
Sex
● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●
530 311 404 334 103 361 189 219 490 85 166 28 118 471 215 37 641 133 37 28 67 6 17 2 4 23 42 23 13 10 42 26
63% 37% 48% 40% 12% 43% 25% 28% 64% 10% 22% 4% 14% 56% 26% 4% 76% 16% 4% 3% 11% 5% 50% 7% 3% 5% 20% 62% 3% 12% 25% 93%
Age
Tumor size
AJCC stage
Fuhrman grade
Subtype
Lymphovascular invasion by subtype
Lymphovascular invasion by grade
Lymphovascular invasion by stage
Male Female ⱕ60 years old 61–75 years old ⱖ76 years old ⱕ4 cm 4.1–6.9 cm ⬎7 cm I II III IV 1 2 3 4 Clear cell Papillary Unclassified Chromophobe Clear cell Papillary Unclassified Chromophobe I II III IV I II III IV
740
M.D. Katz et al. / Urologic Oncology: Seminars and Original Investigations 29 (2011) 738 –744
with tumor were reviewed per case, and all tumor slides from each case were used for evaluation of pathologic parameters.
analyses were performed using SAS software version 9.1 (SAS Institute, Cary, NC). A P value under 0.05 was taken to indicate significance, and all statistical tests were twosided.
2.5. Statistical analysis Patients with bilateral tumors were included as one tumor (tumor first operated on used) in the analysis. The “zero time” for each patient was defined as the time of pathologic diagnosis starting with the date of surgery. MFS was defined as time (in months) from the date of diagnosis to the date of recurrence or when the patient died of disease, whichever occurred first. Those patients who were alive and diseasefree as well as those who died of other causes were censored at the time of last contact. DSS was defined as time to the date the patient died of disease while OS was defined as the time to death due to any cause. The distribution of MFS, DSS, and OS was described using the Kaplan-Meier method, and the difference in each of the outcomes among different subgroups of patients was compared with the log-rank test. Cox proportional hazards regression models were used to quantify the effect of LVI, while adjusting for other demographic and clinical factors, such as the histologic subtype, grade, size, stage, and age. The proportional hazard assumption was examined by plotting the survival time against scaled Schoenfeld residuals from the multivariate Cox model. The magnitude of the effect was quantified by the hazard ratio (HR) with 95% confidence intervals. Statistical
3. Results The clinical and pathologic features for the 841 patients are summarized in Table 1. Median follow-up was 61 months (range, 1 to 209 months). Median age at diagnosis was 61 years of age. There were 233 partial and 622 radical nephrectomies performed. A total of 121 (14%) patients developed metastatic disease, 94 (11%) patients died of RCC, and 306 (36%) died of any cause. LVI was seen in 91 patients (11%). Forty-one patients out of the 91 (45%) with LVI progressed to metastatic disease. On univariate analysis (Table 2), the variables of LVI, Fuhrman grade, histologic subtype, tumor size, and TNM stage were all associated with MFS, DSS, and OS. In addition, age was significantly associated with OS but not with MFS or DSS (P ⫽ 0.81 and P ⫽ 0.87, respectively). Fig. 1 shows the Kaplan-Meier survival curves demonstrating MFS, DSS, and OS when stratifying patients with and without LVI present. All of the variables associated with MFS, DSS, and OS with P ⬍ 0.1 on univariate analysis were put into a multivariate model. Within the multivariate analysis, only stage
Table 2 Univariate analysis demonstrating effect of presence/absence of LVI, Fuhrman grade, stage, subtype, size, and age in terms of DSS, MFS, and OS MFS
Lymphovascular invasion Absent Present Sex Female Male Fuhrman grade 1 2 3 4 TNM stage I II III IV Histologic subtype Clear cell Chromophobe* Papillary I Papillary II Unclassified Size Age
DSS
OS
HR (95% CI)
P value
HR (95% CI)
P value
HR (95% CI)
P value
1.00 (ref) 2.86 (1.78–4.60)
⬍0.0001
1.00 (ref) 3.12 (1.88–5.19)
⬍0.0001
1.00 (ref) 2.27 (1.65–3.12)
⬍0.0001
1.00 (ref) 1.12 (0.76–1.66)
0.56
1.00 (ref) 1.17 (0.75–1.82)
0.48
1.00 (ref) 1.10 (0.86–1.40)
0.46
1.00 (ref) 2.58 (0.93–7.16) 6.46 (2.33–17.9) 11.0 (3.51–34.7) 1.00 (ref) 2.06 (1.06–4.0) 6.07 (3.88–9.49) 23.3 (12.9–42.1) 1.00 (ref) — 1.00 (0.51–2.0) 0.65 (0.29–1.49) 2.70 (1.48–4.95) 1.21 (1.17–1.27) 1.00 (0.98–1.01)
⬍0.0001
⬍0.0001
0.0007
⬍0.0001 0.81
* Chromophobe subtype did not have any events for MFS or DSS.
1.00 (ref) 1.85 (0.66–5.21) 5.02 (1.80–14.0) 10.2 (3.20–32.5) 1.00 (ref) 3.46 (1.72–6.96) 8.36 (4.88–14.3) 32.1 (15.9–64.9) 1.00 (ref) — 0.78 (0.34–1.80) 0.66 (0.27–1.64) 2.55 (1.31–4.96) 1.24 (1.19–1.30) 1.00 (0.99–1.02)
⬍0.0001
⬍0.0001
0.004
⬍0.0001 0.87
1.00 (ref) 1.10 (0.74–1.64) 1.42 (0.93–2.17) 3.14 (1.80–5.48) 1.00 (ref) 1.19 (0.81–1.76) 2.63 (2.02–3.44) 6.33 (3.93–10.2) 1.00 (ref) 0.37 (0.14–1.00) 1.18 (0.80–1.76) 1.12 (0.74–1.70) 1.63 (1.02–2.61) 1.10 (1.07–1.13) 1.03 (1.02–1.04)
0.0002
⬍0.0001
0.03
⬍0.0001 ⬍0.0001
M.D. Katz et al. / Urologic Oncology: Seminars and Original Investigations 29 (2011) 738 –744
741
Fig. 1. Kaplan-Meier curves representing the effect of presence vs. absence of lymphovascular space invasion on metastasis-free survival (a), disease-specific survival (b), and overall survival (c) in patients treated with surgical resection for localized RCC.
and size remained statistically significant, while grade, subtype, and LVI failed to remain independent predictors of either MFS or DSS. Stage, subtype, size, and age all remained independent predictors of decreased OS, while LVI and grade did not (Table 3). Subset analysis performed on the 641 patients with clear cell histology failed to demonstrate an association between LVI and MFS, DSS, or OS controlling for stage, grade and size of tumor (P ⫽ 0.66, 0.71, and 0.14, respectively). Subset analysis was also performed on the 575 patients who had stages I and II disease. This also failed to demonstrate an association between LVI and MFS, DSS, or OS controlling for stage, grade and size of tumor (P ⫽ 0.46, 0.12, and 0.22, respectively).
4. Discussion Renal cell carcinoma has an unpredictable course and represents a wide spectrum of potential biologic behavior. Factors such as TNM stage, histologic subtype, sarcomatoid, and rhabdoid differentiation, primary tumor size, and grade have all been shown to correlate with risk of progression and relapse after surgical resection for renal cell carcinoma [5– 8]. While others have found that LVI is an independent risk factor for all 3 clinical endpoints (MFS, DSS, OS), our study, the largest to date, fails to demonstrate that LVI provides any additional clinical information. Lymphovascular space invasion has proven to be important in better defining aggressive disease in many different
742
M.D. Katz et al. / Urologic Oncology: Seminars and Original Investigations 29 (2011) 738 –744
Table 3 Multivariate analysis demonstrating effect of presence/absence of LVI, Fuhrman grade, stage, subtype, and size in terms of DSS, MFS, and OS MFS HR (95% CI) Lymphovascular invasion Absent Present Fuhrman Grade 1 2 3 4 TNM Stage I II III IV Histologic subtype Clear cell Chromophobe* Papillary I Papillary II Unclassified Size
1.00 (ref) 0.98 (0.58–1.63) 1.00 (ref) 1.87 (0.66–5.24) 2.98 (1.04–8.54) 2.89 (0.87–9.60) 1.00 (ref) 1.05 (0.49–2.23) 2.88 (1.62–5.10) 7.40 (3.43–16.0) 1.00 (ref) — 2.12 (1.04–4.34) 0.81 (0.35–1.86) 1.67 (0.88–3.15) 1.13 (1.06–1.20)
DSS P value 0.93
0.07
⬍0.0001
0.14
0.0001
HR (95% CI) 1.00 (ref) 0.96 (0.54–1.69) 1.00 (ref) 1.35 (0.47–3.89) 1.92 (0.65–5.68) 2.51 (0.74–8.58) 1.00 (ref) 1.62 (0.72–3.63) 3.53 (1.78–7.03) 8.88 (3.61–21.8) 1.00 (ref) — 1.93 (0.80–4.64) 1.13 (0.48–2.84) 1.55 (0.76–3.15) 1.15 (1.07–1.23)
OS P value 0.88
0.26
⬍0.0001
0.49
⬍0.0001
HR (95% CI) 1.00 (ref) 1.24 (0.87–1.77) 1.00 (ref) 0.86 (0.57–1.29) 0.84 (0.53–1.34) 1.46 (0.81–2.63) 1.00 (ref) 0.89 (0.55–1.42) 1.58 (1.11–7.34)
1.00 (ref) 0.44 (0.16–1.21) 1.58 (1.05–2.38) 1.44 (0.94–2.20) 1.25 (0.76–2.04) 1.08 (1.03–1.13)
P value 0.24
0.12
⬍0.0001
0.04
0.001
* Chromophobe subtype did not have any events for MFS or DSS.
urologic and nonurologic malignancies. However, there has been conflicting evidence regarding the presence of LVI in RCC correlating with adverse patient outcome (Table 4) [12–28]. Of all the studies that demonstrate LVI to be associated with decreased survival on multivariate analysis, Lang et al. demonstrated the longest follow-up with a median time of 183 months. In that study, the authors found microscopic venous invasion to be an independent predictor of survival for patients with nonmetastatic RCC in terms of disease-specific and OS, but not metastases-free survival within their multivariate model [18]. Sorbellini et al. examined the largest population of patients (n ⫽ 833) and found that only LVI (P ⫽ 0.012) and Fuhrman grade (P ⫽ 0.002) remained independent predictors for freedom from recurrence within their multivariate analysis. In contrast to our study population, Sorbellini et al. only examined clear cell histology and Lang et al. predominantly clear cell histology [28]. However, subset analysis in our population of patients with clear cell carcinoma demonstrated no association between LVI and MFS, DSS, or OS. Previous studies that failed to find an association tended to have a smaller cohort size [22–27]. The largest negative study by Ishimura et al. examined only 171 patients with a median follow-up of 42 months and found that only pathologic stage maintained significance on multivariate analysis when controlling for gender, size, mode of detection, and grade. Interestingly, Ishimura et al. examined a subset of patients with T1 and T2 disease and found a positive association between LVI and recurrence-free survival [24]. Subset analysis of this cohort of patients in our population did not reveal any association between LVI and recurrence. In our study, lymphovascular space invasion was str-
ongly associated with adverse outcomes within the univariate analysis. However, within the multivariate analysis, LVI lost significance for all 3 survival endpoints of MFS, DSS, and OS. The established predictors of survival (stage and size) were more strongly associated with survival than LVI. With a total of 841 patients and a median follow-up of 61 months, to our knowledge, our study represents the largest cohort of patients studied. Although LVI failed to remain statistically significant within our multivariate model, it was noted to be present more commonly as tumors increased in Fuhrman grade and stage (Table 1). Patients found to have LVI within their primary tumor were noted to develop metastatic disease nearly half of the time (45%). There are several possible limitations of this study. First, we used only visual inspection to define LVI, while other studies have incorporated immunohistochemical analysis using either antifactor VIII antibodies or anti-CD34 antibodies in cases negative for LVI by examination of H and E stained sections [18,20]. This added measure may increase the detection rate of LVI. In addition, immunohistochemical staining may provide insight if the route of metastasis is lymphatic (using antibody D2-40) or hematogenous. However, since this staining is not routine practice, it is unlikely to be implemented into general practice without stronger evidence that LVI assessment is useful. Second, we confined our analysis to include patients with localized RCC. It is possible that with inclusion of patients with metastatic disease, LVI might have remained significant in terms of predicting worse survival. However, since survival in this population of patients is poor, an additional prognostic indicator is of limited clinical utility. Third, our follow-up is 61 months. While this is longer than all but a few of the
M.D. Katz et al. / Urologic Oncology: Seminars and Original Investigations 29 (2011) 738 –744 Table 4 Previous studies evaluating the effect of LVI on survival within a multivariate analysis Study [ref] (year published)
n
Survival endpoint
P value for LVI
Van Poppel et al. [21] (1997) Sevinc et al. [26] (2000) Griffins et al. [19] (2002) Ishimura et al. [24] (2004) Ishimura et al. [25] (2004) Goncalves et al. [16] (2004) Yildiz et al. [17] (2004) Lang et al. [18] (2004) Sorbellini et al. [28] (2005) Madbouly et al. [14] (2007) Dall’Oglio et al. [15] (2007) Dall’Oglio et al. [13] (2007) Dall’Oglio et al. [12] (2007) Yildiz et al. [17] (2007) Horiguchi et al. [23] (2008) Present series
180
MFS
⬍0.00001
41
MFS
NS
176
MFS
157
MFS ⫹ DSS
NS for both
171
MFS ⫹ DSS
NS for both
95
DSS
⬍0.05
48
DSS
⬍0.003
255
MFS, DSS, OS
NS; 0.0016; 0.015
833
MFS
0.012
48
MFS
0.007
230
MFS ⫹ DSS
0.041
230
MFS ⫹ DSS
0.001; 0.007
230
MFS ⫹ DSS
0.015; 0.002
54
DSS
NS
106
DSS
NS
841
MFS, DSS, OS
NS for all 3
0.057
n ⫽ sample size; MFS ⫽ metastasis-free survival; DSS ⫽ disease specific survival; OS ⫽ overall survival; NS ⫽ not significant.
previous studies, it is possible that with longer follow-up LVI could become significant on multivariate analysis.
5. Conclusions Based on the results of our study, LVI does not appear to be an independent prognosticator of decreased survival for patients with localized RCC and, therefore, has limited clinical utility in identification of high risk RCC patients. While on univariate analysis, it was strongly predictive of metastatic disease, disease specific and OS, LVI lost its significance within our multivariate model. Based on our data and the conflicting data in the literature, LVI cannot be recommended as a marker of more aggressive disease in patients with localized RCC.
References [1] Jemal A, Siegel R, Ward E, et al. Cancer Statistics, 2009. CA Cancer J Clin 2009;59:225– 49.
743
[2] Janzen NK, Kim HL, Figlin RA, et al. Surveillance after radical or partial nephrectomy for localized renal cell carcinoma and management of recurrent disease. Urol Clin North Am 2003;30: 843–52. [3] Campbell SC, Novick AC, Bukowski RM. Renal Tumors. In: Wein AJ, Kavoussi LR, Novick AC, et al., editors. Campbell-Walsh Urology, 9th ed. Philadelphia: W. B. Saunders 2007. p. 1567– 637. [4] Bukowski RM. Natural history and therapy of metastatic renal cell carcinoma: The role of interleukin-2. Cancer 1997;80:1198 –220. [5] Amin MB, Amin MB, Tamboli P, et al. Prognostic impact of histologic subtyping of adult renal epithelial neoplasms: An experience of 405 cases. Am J Surg Pathol 2002;26:281–91. [6] Bretheau D, Lechevallier E, de Fromont M, et al. Prognostic value of nuclear grade in renal cell carcinoma. Cancer 1995;76:2543–9. [7] Gokden N, Nappi O, Swanson PE, et al. Renal cell carcinoma with rhabdoid features. Am J Surg Pathol 2000;24:1329 –38. [8] Cheville JC, Lohse CM, Zincke H, et al. Comparisons of outcome and prognostic features among histologic subtypes of renal cell carcinoma. Am J Surg Pathol 2003;27:612–24. [9] Lin W, Hu F, Chung S, et al. The role of lymphovascular invasion in predicting the prognosis of clinically localized upper tract urothelial carcinoma (pT1-3cNOMO). J Urol 2008;180:879 – 84. [10] Heidenreich A, Sesterhenn I, Mostofi F, et al. Prognostic risk factors that identify patients with clinical stage I nonseminomatous germ cell tumors at low risk and high risk for metastasis. Cancer 1998;83:1002–11. [11] Lotan Y, Gupta A, Shariat S, et al. Lymphovascular invasion is independently associated with overall survival, cause-specific survival, and local and distant recurrence in patients with negative lymph nodes at radical cystectomy. J Clin Oncol 2005;23:6533–9. [12] Dall’Oglio M, Antunes A, Sarkis A, et al. Microvascular tumor invasion in renal cell carcinoma: The most important prognostic factor. BJU Int 2007;100:552–5. [13] Dall’Oglio M, Filho L, Antunes A, et al. Microvascular tumor invasion, tumor size, and Fuhrman grade: A pathological triad for prognostic evaluation of renal cell carcinoma. J Urol 2007;178: 425– 8. [14] Madbouly K, Al-Qahtani S, Ghazwani Y, et al. Microvascular tumor invasion: Prognostic significance in low-stage renal cell carcinoma. Urology 2007;69:670 – 4. [15] Dall’Oglio M, Arap M, Antunes A, et al. Impact of clinicopathological parameters in patients treated for renal cell carcinoma. J Urol 2007;177:1687–91. [16] Goncalves P, Srougi M, Dall’Oglio M, et al. Low clinical stage renal cell carcinoma: Relevance of microvascular tumor invasion as a prognostic parameter. J Urol 2004;172:470 – 4. [17] Yildiz E, Gokce G, Kilicarslan H, et al. Prognostic value of the expression of Ki-67, CD44, and vascular endothelial growth factor, and microvessel invasion, in renal cell carcinoma. BJU Int 2004;93: 1087–93. [18] Lang H, Lindner V, Letourneux H, et al. Prognostic value of microscopic venous invasion in renal cell carcinoma: Long-term follow-up. Eur Urol 2004;46:331–5. [19] Griffiths D, Verghese A, Golash A, et al. Contribution of grade, vascular invasion, and age to outcome in clinically localized renal cell carcinoma. BJU Int 2002;90:26 –31. [20] Dekel Y, Koren R, Kugel V, et al. Significance of angiogenesis and microvascular invasion in renal cell carcinoma. Pathol Oncol Res 2002;8:129 –32. [21] Van Poppel H, Vandendriessche H, Boel K, et al. Microscopic vascular invasion is the most relevant prognosticator after radical nephrectomy for clinically nonmetastatic renal cell carcinoma. J Urol 1997;158:45–9. [22] Yildiz E, Ayan S, Goze F, et al. Relation of microvessel density with microvascular invasion, metastasis, and prognosis in renal cell carcinoma. BJU Int 2007;101:758 – 64.
744
M.D. Katz et al. / Urologic Oncology: Seminars and Original Investigations 29 (2011) 738 –744
[23] Horiguchi A, Keiichi I, Sumitomo M, et al. Intratumoral lymphatics and lymphatic invasion are associated with tumor aggressiveness and poor prognosis in renal cell carcinoma. Urology 2008;71:928 –32. [24] Ishimura T, Sakai I, Hara I, et al. Microscopic venous invasion in renal cell carcinoma as a predictor of recurrence after radical surgery. Int J Urol 2004;11:264 – 8. [25] Ishimura T, Sakai I, Harada K, et al. Clinicopathological features of recurrence after radical surgery for nonmetastatic renal cell carcinoma. Int J Clin Oncol 2004;9:369 –72.
[26] Sevinc M, Kirkali Z, Yorukoglu K, et al. Prognostic significance of microvascular invasion in localized renal cell carcinoma. Eur Urol 2000;38:728 –33. [27] Lang H, Lindner V, Saussine C, et al. Microscopic venous invasion: A prognostic factor in renal cell carcinoma. Eur Urol 2000; 38:600 –5. [28] Sorbellini M, Kattan M, Snyder M, et al. A postoperative prognostic nomogram predicting recurrence for patients with conventional clear cell carcinoma. J Urol 2005;173:48 –51.
Original article
The role of lymphovascular space invasion in renal cell carcinoma as a prognostic marker of survival after curative resection Matthew D. Katz, M.D.a, Maria F. Serrano, M.D.b, Peter A. Humphrey, M.D., Ph.D.b, Robert L. Grubb III, M.D.a,d, Ted A. Skolarus, M.D.a, Feng Gao, M.D., Ph.D.c, Adam S. Kibel, M.D.a,d,* b
a Division of Urologic Surgery, Washington University School of Medicine, St. Louis, MO 63110, USA Division of Anatomic and Molecular Pathology, Department of Pathology and Immunology, Washington University School of Medicine, St. Louis, MO 63110, USA c Division of Biostatistics, Washington University School of Medicine, St. Louis, MO 63110, USA d Alvin J. Siteman Cancer Center, Washington University School of Medicine, St. Louis, MO 63110, USA
Received 20 June 2009; received in revised form 28 July 2009; accepted 29 July 2009
Abstract Objectives: Lymphovascular invasion (LVI) correlates with adverse outcomes in numerous malignancies. However, its role in predicting outcomes in RCC is unclear. Herein, we evaluated what effect LVI had on metastasis free survival (MFS), disease-specific survival (DSS), and overall survival (OS) in patients with RCC treated with surgical excision. Methods: Eight hundred forty-one consecutive patients who underwent partial or radical nephrectomy from 1989 to 2004 were identified. Pathologic and gross features examined were LVI, subtype, Fuhrman grade, stage, and size. Age and gender were also analyzed. Slides were re-reviewed by a single pathologist (MS). Variables with P ⬍ 0.1 on univariate analysis were incorporated in a Cox proportional hazards multivariate model. MFS, DSS, and OS were described for patients with and without LVI using the Kaplan-Meier method, and compared with the log-rank test. Results: LVI was seen on H and E stained slides in 91 patients (11%); 120 (14%) developed metastatic disease, 91 (11%) died of RCC, and 306 (36%) died during a median follow-up of 61 months. While on univariate analysis, LVI was strongly associated with decreased MFS, DSS, and OS (P ⬍ 0.0001), on multivariate analysis, LVI was no longer statistically significant for MFS, DSS, and OS with a HR of 0.976 (95% CI: 0.583–1.63; P ⫽ 0.93), 0.96 (95% CI: 0.542–1.69; P ⫽ 0.88), and 1.24 (95% CI: 0.869 –1.77; P ⫽ 0.24). Conclusions: We found LVI to be associated with worse MFS, DSS, and OS on univariate analysis, but not on multivariate analysis for patients with nonmetastatic RCC. In contrast to previously reported studies, LVI may not be an independent prognostic variable in patients with localized RCC. © 2011 Elsevier Inc. All rights reserved. Keywords: Kidney cancer; Lymphovascular invasion; Nephrectomy; Survival
1. Introduction There will be an estimated 57,760 new cases of renal cancer in the United States, and an estimated 12,980 deaths caused by this disease in 2009 [1]. With the widespread use of cross-sectional imaging for nonspecific abdominal complaints, the rate of incidental renal masses being detected is on the rise. However, even with the resection of localized renal cell carcinoma, up to a third of patients will go on to * Corresponding author. Tel.: ⫹1-314-362-8295; fax: ⫹1-314-454-5244. E-mail address: [email protected] (A.S. Kibel). 1078-1439/$ – see front matter © 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.urolonc.2009.07.034
develop metastatic disease [2– 4]. The prognostic factors of stage, grade, histologic subtype, and presence of sarcomatoid or rhabdoid differentiation are most commonly used to help delineate which patients are at highest risk of recurrent disease [5– 8]. The role of lymphovascular space invasion (LVI) in identifying patients with more aggressive biological forms of renal cell carcinoma has not been fully elucidated. The presence of LVI has been demonstrated to be associated with adverse outcomes in other urologic malignancies, including urothelial carcinoma and testicular germ cell tumors [9 –11]. However, there is conflicting evidence regarding
M.D. Katz et al. / Urologic Oncology: Seminars and Original Investigations 29 (2011) 738 –744
LVI and its importance in predicting worse outcomes in patients with RCC. While several studies have demonstrated the presence of LVI to correlate with decreased survival in patients with RCC [12–21,28], others have not [22–27]. Most of these studies have primarily focused on either disease-free or disease-specific survival (DSS), and there has only been 1 paper that has evaluated all 3 survival endpoints of metastasis-free survival (MFS), DSS, or overall survival (OS) [18]. Thus, in order to further delineate the significance of LVI within the primary tumor in patients treated with radical and partial nephrectomy, we investigated whether the presence of LVI correlated with the 3 survival endpoints of MFS, DSS, and OS.
2. Materials and methods 2.1. Study population After obtaining institutional board approval, we retrospectively reviewed the clinical and pathologic data of patients diagnosed with RCC from 1989 to 2004 treated with radical or partial nephrectomy with curative intent at our institution. Inclusion criteria were: (1) surgical resection of pathologically confirmed RCC with either partial or radical nephrectomy; (2) adequate tissue available for pathologic re-evaluation; and (3) adequate radiographic and or clinical follow-up data available. Exclusion criteria were: (1) pathologically confirmed urothelial carcinoma or any benign lesion; (2) cytoreductive nephrectomy done as part of treatment of known metastatic disease; (3) tissue unavailable for accurate re-evaluation; or (4) no available follow-up clinical or radiographic data. Based on these criteria, a total of 841 eligible patients were included within our analysis. 2.2. Data collection The Barnes-Jewish Hospital Oncology Data Services is an American College of Surgeons Commission on Cancer approved registry. Data on all malignancies diagnosed at our institution are extracted prospectively from charts by certified registrars, beginning at the time of diagnosis. Data collected by tumor registrars for all tumors include standard tumor registry data elements: date of diagnosis, demographic information, stage, grade and nodal spread at time of presentation, initial therapy information, follow-up information, subsequent therapy information, and outcome information. Data entry into the database was done by the tumor registrars. Retrospective chart review was performed to determine dates of recurrence, date and cause of death.
739
2.3. Pathology All hematoxylin and eosin stained (H and E) slides were re-reviewed by a single pathologist (MS) who was blinded to clinical outcome data. Tumors were classified according to the 2004 World Health Organization classification system; TNM stage was classified according to the 2002 American Joint Committee on Cancer (AJCC) staging system. Tumors were graded according to the conventional Fuhrman grading system as 1, 2, 3, or 4. All tumors were assessed for the presence of LVI. Primary tumor size was recorded from the surgical pathology reports. 2.4. Evaluation for lymphovascular invasion LVI was considered present if any tumor cells were seen within the luminal space lined by endothelial cells by visual inspection on H and E stained slides from each tumor. The identification of unequivocal endothelial lining was mandatory. No distinction was made between lymphatic and vascular invasion. Renal vein involvement was defined as tumor cells in a large vein with a muscular wall in the renal sinus. This was not categorized as LVI. A mean of 5 slides
Table 1 Baseline demographic, clinical, and tumor features (n ⫽ 841) Variables
Categories
n
% Patients
Sex
● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●
530 311 404 334 103 361 189 219 490 85 166 28 118 471 215 37 641 133 37 28 67 6 17 2 4 23 42 23 13 10 42 26
63% 37% 48% 40% 12% 43% 25% 28% 64% 10% 22% 4% 14% 56% 26% 4% 76% 16% 4% 3% 11% 5% 50% 7% 3% 5% 20% 62% 3% 12% 25% 93%
Age
Tumor size
AJCC stage
Fuhrman grade
Subtype
Lymphovascular invasion by subtype
Lymphovascular invasion by grade
Lymphovascular invasion by stage
Male Female ⱕ60 years old 61–75 years old ⱖ76 years old ⱕ4 cm 4.1–6.9 cm ⬎7 cm I II III IV 1 2 3 4 Clear cell Papillary Unclassified Chromophobe Clear cell Papillary Unclassified Chromophobe I II III IV I II III IV
740
M.D. Katz et al. / Urologic Oncology: Seminars and Original Investigations 29 (2011) 738 –744
with tumor were reviewed per case, and all tumor slides from each case were used for evaluation of pathologic parameters.
analyses were performed using SAS software version 9.1 (SAS Institute, Cary, NC). A P value under 0.05 was taken to indicate significance, and all statistical tests were twosided.
2.5. Statistical analysis Patients with bilateral tumors were included as one tumor (tumor first operated on used) in the analysis. The “zero time” for each patient was defined as the time of pathologic diagnosis starting with the date of surgery. MFS was defined as time (in months) from the date of diagnosis to the date of recurrence or when the patient died of disease, whichever occurred first. Those patients who were alive and diseasefree as well as those who died of other causes were censored at the time of last contact. DSS was defined as time to the date the patient died of disease while OS was defined as the time to death due to any cause. The distribution of MFS, DSS, and OS was described using the Kaplan-Meier method, and the difference in each of the outcomes among different subgroups of patients was compared with the log-rank test. Cox proportional hazards regression models were used to quantify the effect of LVI, while adjusting for other demographic and clinical factors, such as the histologic subtype, grade, size, stage, and age. The proportional hazard assumption was examined by plotting the survival time against scaled Schoenfeld residuals from the multivariate Cox model. The magnitude of the effect was quantified by the hazard ratio (HR) with 95% confidence intervals. Statistical
3. Results The clinical and pathologic features for the 841 patients are summarized in Table 1. Median follow-up was 61 months (range, 1 to 209 months). Median age at diagnosis was 61 years of age. There were 233 partial and 622 radical nephrectomies performed. A total of 121 (14%) patients developed metastatic disease, 94 (11%) patients died of RCC, and 306 (36%) died of any cause. LVI was seen in 91 patients (11%). Forty-one patients out of the 91 (45%) with LVI progressed to metastatic disease. On univariate analysis (Table 2), the variables of LVI, Fuhrman grade, histologic subtype, tumor size, and TNM stage were all associated with MFS, DSS, and OS. In addition, age was significantly associated with OS but not with MFS or DSS (P ⫽ 0.81 and P ⫽ 0.87, respectively). Fig. 1 shows the Kaplan-Meier survival curves demonstrating MFS, DSS, and OS when stratifying patients with and without LVI present. All of the variables associated with MFS, DSS, and OS with P ⬍ 0.1 on univariate analysis were put into a multivariate model. Within the multivariate analysis, only stage
Table 2 Univariate analysis demonstrating effect of presence/absence of LVI, Fuhrman grade, stage, subtype, size, and age in terms of DSS, MFS, and OS MFS
Lymphovascular invasion Absent Present Sex Female Male Fuhrman grade 1 2 3 4 TNM stage I II III IV Histologic subtype Clear cell Chromophobe* Papillary I Papillary II Unclassified Size Age
DSS
OS
HR (95% CI)
P value
HR (95% CI)
P value
HR (95% CI)
P value
1.00 (ref) 2.86 (1.78–4.60)
⬍0.0001
1.00 (ref) 3.12 (1.88–5.19)
⬍0.0001
1.00 (ref) 2.27 (1.65–3.12)
⬍0.0001
1.00 (ref) 1.12 (0.76–1.66)
0.56
1.00 (ref) 1.17 (0.75–1.82)
0.48
1.00 (ref) 1.10 (0.86–1.40)
0.46
1.00 (ref) 2.58 (0.93–7.16) 6.46 (2.33–17.9) 11.0 (3.51–34.7) 1.00 (ref) 2.06 (1.06–4.0) 6.07 (3.88–9.49) 23.3 (12.9–42.1) 1.00 (ref) — 1.00 (0.51–2.0) 0.65 (0.29–1.49) 2.70 (1.48–4.95) 1.21 (1.17–1.27) 1.00 (0.98–1.01)
⬍0.0001
⬍0.0001
0.0007
⬍0.0001 0.81
* Chromophobe subtype did not have any events for MFS or DSS.
1.00 (ref) 1.85 (0.66–5.21) 5.02 (1.80–14.0) 10.2 (3.20–32.5) 1.00 (ref) 3.46 (1.72–6.96) 8.36 (4.88–14.3) 32.1 (15.9–64.9) 1.00 (ref) — 0.78 (0.34–1.80) 0.66 (0.27–1.64) 2.55 (1.31–4.96) 1.24 (1.19–1.30) 1.00 (0.99–1.02)
⬍0.0001
⬍0.0001
0.004
⬍0.0001 0.87
1.00 (ref) 1.10 (0.74–1.64) 1.42 (0.93–2.17) 3.14 (1.80–5.48) 1.00 (ref) 1.19 (0.81–1.76) 2.63 (2.02–3.44) 6.33 (3.93–10.2) 1.00 (ref) 0.37 (0.14–1.00) 1.18 (0.80–1.76) 1.12 (0.74–1.70) 1.63 (1.02–2.61) 1.10 (1.07–1.13) 1.03 (1.02–1.04)
0.0002
⬍0.0001
0.03
⬍0.0001 ⬍0.0001
M.D. Katz et al. / Urologic Oncology: Seminars and Original Investigations 29 (2011) 738 –744
741
Fig. 1. Kaplan-Meier curves representing the effect of presence vs. absence of lymphovascular space invasion on metastasis-free survival (a), disease-specific survival (b), and overall survival (c) in patients treated with surgical resection for localized RCC.
and size remained statistically significant, while grade, subtype, and LVI failed to remain independent predictors of either MFS or DSS. Stage, subtype, size, and age all remained independent predictors of decreased OS, while LVI and grade did not (Table 3). Subset analysis performed on the 641 patients with clear cell histology failed to demonstrate an association between LVI and MFS, DSS, or OS controlling for stage, grade and size of tumor (P ⫽ 0.66, 0.71, and 0.14, respectively). Subset analysis was also performed on the 575 patients who had stages I and II disease. This also failed to demonstrate an association between LVI and MFS, DSS, or OS controlling for stage, grade and size of tumor (P ⫽ 0.46, 0.12, and 0.22, respectively).
4. Discussion Renal cell carcinoma has an unpredictable course and represents a wide spectrum of potential biologic behavior. Factors such as TNM stage, histologic subtype, sarcomatoid, and rhabdoid differentiation, primary tumor size, and grade have all been shown to correlate with risk of progression and relapse after surgical resection for renal cell carcinoma [5– 8]. While others have found that LVI is an independent risk factor for all 3 clinical endpoints (MFS, DSS, OS), our study, the largest to date, fails to demonstrate that LVI provides any additional clinical information. Lymphovascular space invasion has proven to be important in better defining aggressive disease in many different
742
M.D. Katz et al. / Urologic Oncology: Seminars and Original Investigations 29 (2011) 738 –744
Table 3 Multivariate analysis demonstrating effect of presence/absence of LVI, Fuhrman grade, stage, subtype, and size in terms of DSS, MFS, and OS MFS HR (95% CI) Lymphovascular invasion Absent Present Fuhrman Grade 1 2 3 4 TNM Stage I II III IV Histologic subtype Clear cell Chromophobe* Papillary I Papillary II Unclassified Size
1.00 (ref) 0.98 (0.58–1.63) 1.00 (ref) 1.87 (0.66–5.24) 2.98 (1.04–8.54) 2.89 (0.87–9.60) 1.00 (ref) 1.05 (0.49–2.23) 2.88 (1.62–5.10) 7.40 (3.43–16.0) 1.00 (ref) — 2.12 (1.04–4.34) 0.81 (0.35–1.86) 1.67 (0.88–3.15) 1.13 (1.06–1.20)
DSS P value 0.93
0.07
⬍0.0001
0.14
0.0001
HR (95% CI) 1.00 (ref) 0.96 (0.54–1.69) 1.00 (ref) 1.35 (0.47–3.89) 1.92 (0.65–5.68) 2.51 (0.74–8.58) 1.00 (ref) 1.62 (0.72–3.63) 3.53 (1.78–7.03) 8.88 (3.61–21.8) 1.00 (ref) — 1.93 (0.80–4.64) 1.13 (0.48–2.84) 1.55 (0.76–3.15) 1.15 (1.07–1.23)
OS P value 0.88
0.26
⬍0.0001
0.49
⬍0.0001
HR (95% CI) 1.00 (ref) 1.24 (0.87–1.77) 1.00 (ref) 0.86 (0.57–1.29) 0.84 (0.53–1.34) 1.46 (0.81–2.63) 1.00 (ref) 0.89 (0.55–1.42) 1.58 (1.11–7.34)
1.00 (ref) 0.44 (0.16–1.21) 1.58 (1.05–2.38) 1.44 (0.94–2.20) 1.25 (0.76–2.04) 1.08 (1.03–1.13)
P value 0.24
0.12
⬍0.0001
0.04
0.001
* Chromophobe subtype did not have any events for MFS or DSS.
urologic and nonurologic malignancies. However, there has been conflicting evidence regarding the presence of LVI in RCC correlating with adverse patient outcome (Table 4) [12–28]. Of all the studies that demonstrate LVI to be associated with decreased survival on multivariate analysis, Lang et al. demonstrated the longest follow-up with a median time of 183 months. In that study, the authors found microscopic venous invasion to be an independent predictor of survival for patients with nonmetastatic RCC in terms of disease-specific and OS, but not metastases-free survival within their multivariate model [18]. Sorbellini et al. examined the largest population of patients (n ⫽ 833) and found that only LVI (P ⫽ 0.012) and Fuhrman grade (P ⫽ 0.002) remained independent predictors for freedom from recurrence within their multivariate analysis. In contrast to our study population, Sorbellini et al. only examined clear cell histology and Lang et al. predominantly clear cell histology [28]. However, subset analysis in our population of patients with clear cell carcinoma demonstrated no association between LVI and MFS, DSS, or OS. Previous studies that failed to find an association tended to have a smaller cohort size [22–27]. The largest negative study by Ishimura et al. examined only 171 patients with a median follow-up of 42 months and found that only pathologic stage maintained significance on multivariate analysis when controlling for gender, size, mode of detection, and grade. Interestingly, Ishimura et al. examined a subset of patients with T1 and T2 disease and found a positive association between LVI and recurrence-free survival [24]. Subset analysis of this cohort of patients in our population did not reveal any association between LVI and recurrence. In our study, lymphovascular space invasion was str-
ongly associated with adverse outcomes within the univariate analysis. However, within the multivariate analysis, LVI lost significance for all 3 survival endpoints of MFS, DSS, and OS. The established predictors of survival (stage and size) were more strongly associated with survival than LVI. With a total of 841 patients and a median follow-up of 61 months, to our knowledge, our study represents the largest cohort of patients studied. Although LVI failed to remain statistically significant within our multivariate model, it was noted to be present more commonly as tumors increased in Fuhrman grade and stage (Table 1). Patients found to have LVI within their primary tumor were noted to develop metastatic disease nearly half of the time (45%). There are several possible limitations of this study. First, we used only visual inspection to define LVI, while other studies have incorporated immunohistochemical analysis using either antifactor VIII antibodies or anti-CD34 antibodies in cases negative for LVI by examination of H and E stained sections [18,20]. This added measure may increase the detection rate of LVI. In addition, immunohistochemical staining may provide insight if the route of metastasis is lymphatic (using antibody D2-40) or hematogenous. However, since this staining is not routine practice, it is unlikely to be implemented into general practice without stronger evidence that LVI assessment is useful. Second, we confined our analysis to include patients with localized RCC. It is possible that with inclusion of patients with metastatic disease, LVI might have remained significant in terms of predicting worse survival. However, since survival in this population of patients is poor, an additional prognostic indicator is of limited clinical utility. Third, our follow-up is 61 months. While this is longer than all but a few of the
M.D. Katz et al. / Urologic Oncology: Seminars and Original Investigations 29 (2011) 738 –744 Table 4 Previous studies evaluating the effect of LVI on survival within a multivariate analysis Study [ref] (year published)
n
Survival endpoint
P value for LVI
Van Poppel et al. [21] (1997) Sevinc et al. [26] (2000) Griffins et al. [19] (2002) Ishimura et al. [24] (2004) Ishimura et al. [25] (2004) Goncalves et al. [16] (2004) Yildiz et al. [17] (2004) Lang et al. [18] (2004) Sorbellini et al. [28] (2005) Madbouly et al. [14] (2007) Dall’Oglio et al. [15] (2007) Dall’Oglio et al. [13] (2007) Dall’Oglio et al. [12] (2007) Yildiz et al. [17] (2007) Horiguchi et al. [23] (2008) Present series
180
MFS
⬍0.00001
41
MFS
NS
176
MFS
157
MFS ⫹ DSS
NS for both
171
MFS ⫹ DSS
NS for both
95
DSS
⬍0.05
48
DSS
⬍0.003
255
MFS, DSS, OS
NS; 0.0016; 0.015
833
MFS
0.012
48
MFS
0.007
230
MFS ⫹ DSS
0.041
230
MFS ⫹ DSS
0.001; 0.007
230
MFS ⫹ DSS
0.015; 0.002
54
DSS
NS
106
DSS
NS
841
MFS, DSS, OS
NS for all 3
0.057
n ⫽ sample size; MFS ⫽ metastasis-free survival; DSS ⫽ disease specific survival; OS ⫽ overall survival; NS ⫽ not significant.
previous studies, it is possible that with longer follow-up LVI could become significant on multivariate analysis.
5. Conclusions Based on the results of our study, LVI does not appear to be an independent prognosticator of decreased survival for patients with localized RCC and, therefore, has limited clinical utility in identification of high risk RCC patients. While on univariate analysis, it was strongly predictive of metastatic disease, disease specific and OS, LVI lost its significance within our multivariate model. Based on our data and the conflicting data in the literature, LVI cannot be recommended as a marker of more aggressive disease in patients with localized RCC.
References [1] Jemal A, Siegel R, Ward E, et al. Cancer Statistics, 2009. CA Cancer J Clin 2009;59:225– 49.
743
[2] Janzen NK, Kim HL, Figlin RA, et al. Surveillance after radical or partial nephrectomy for localized renal cell carcinoma and management of recurrent disease. Urol Clin North Am 2003;30: 843–52. [3] Campbell SC, Novick AC, Bukowski RM. Renal Tumors. In: Wein AJ, Kavoussi LR, Novick AC, et al., editors. Campbell-Walsh Urology, 9th ed. Philadelphia: W. B. Saunders 2007. p. 1567– 637. [4] Bukowski RM. Natural history and therapy of metastatic renal cell carcinoma: The role of interleukin-2. Cancer 1997;80:1198 –220. [5] Amin MB, Amin MB, Tamboli P, et al. Prognostic impact of histologic subtyping of adult renal epithelial neoplasms: An experience of 405 cases. Am J Surg Pathol 2002;26:281–91. [6] Bretheau D, Lechevallier E, de Fromont M, et al. Prognostic value of nuclear grade in renal cell carcinoma. Cancer 1995;76:2543–9. [7] Gokden N, Nappi O, Swanson PE, et al. Renal cell carcinoma with rhabdoid features. Am J Surg Pathol 2000;24:1329 –38. [8] Cheville JC, Lohse CM, Zincke H, et al. Comparisons of outcome and prognostic features among histologic subtypes of renal cell carcinoma. Am J Surg Pathol 2003;27:612–24. [9] Lin W, Hu F, Chung S, et al. The role of lymphovascular invasion in predicting the prognosis of clinically localized upper tract urothelial carcinoma (pT1-3cNOMO). J Urol 2008;180:879 – 84. [10] Heidenreich A, Sesterhenn I, Mostofi F, et al. Prognostic risk factors that identify patients with clinical stage I nonseminomatous germ cell tumors at low risk and high risk for metastasis. Cancer 1998;83:1002–11. [11] Lotan Y, Gupta A, Shariat S, et al. Lymphovascular invasion is independently associated with overall survival, cause-specific survival, and local and distant recurrence in patients with negative lymph nodes at radical cystectomy. J Clin Oncol 2005;23:6533–9. [12] Dall’Oglio M, Antunes A, Sarkis A, et al. Microvascular tumor invasion in renal cell carcinoma: The most important prognostic factor. BJU Int 2007;100:552–5. [13] Dall’Oglio M, Filho L, Antunes A, et al. Microvascular tumor invasion, tumor size, and Fuhrman grade: A pathological triad for prognostic evaluation of renal cell carcinoma. J Urol 2007;178: 425– 8. [14] Madbouly K, Al-Qahtani S, Ghazwani Y, et al. Microvascular tumor invasion: Prognostic significance in low-stage renal cell carcinoma. Urology 2007;69:670 – 4. [15] Dall’Oglio M, Arap M, Antunes A, et al. Impact of clinicopathological parameters in patients treated for renal cell carcinoma. J Urol 2007;177:1687–91. [16] Goncalves P, Srougi M, Dall’Oglio M, et al. Low clinical stage renal cell carcinoma: Relevance of microvascular tumor invasion as a prognostic parameter. J Urol 2004;172:470 – 4. [17] Yildiz E, Gokce G, Kilicarslan H, et al. Prognostic value of the expression of Ki-67, CD44, and vascular endothelial growth factor, and microvessel invasion, in renal cell carcinoma. BJU Int 2004;93: 1087–93. [18] Lang H, Lindner V, Letourneux H, et al. Prognostic value of microscopic venous invasion in renal cell carcinoma: Long-term follow-up. Eur Urol 2004;46:331–5. [19] Griffiths D, Verghese A, Golash A, et al. Contribution of grade, vascular invasion, and age to outcome in clinically localized renal cell carcinoma. BJU Int 2002;90:26 –31. [20] Dekel Y, Koren R, Kugel V, et al. Significance of angiogenesis and microvascular invasion in renal cell carcinoma. Pathol Oncol Res 2002;8:129 –32. [21] Van Poppel H, Vandendriessche H, Boel K, et al. Microscopic vascular invasion is the most relevant prognosticator after radical nephrectomy for clinically nonmetastatic renal cell carcinoma. J Urol 1997;158:45–9. [22] Yildiz E, Ayan S, Goze F, et al. Relation of microvessel density with microvascular invasion, metastasis, and prognosis in renal cell carcinoma. BJU Int 2007;101:758 – 64.
744
M.D. Katz et al. / Urologic Oncology: Seminars and Original Investigations 29 (2011) 738 –744
[23] Horiguchi A, Keiichi I, Sumitomo M, et al. Intratumoral lymphatics and lymphatic invasion are associated with tumor aggressiveness and poor prognosis in renal cell carcinoma. Urology 2008;71:928 –32. [24] Ishimura T, Sakai I, Hara I, et al. Microscopic venous invasion in renal cell carcinoma as a predictor of recurrence after radical surgery. Int J Urol 2004;11:264 – 8. [25] Ishimura T, Sakai I, Harada K, et al. Clinicopathological features of recurrence after radical surgery for nonmetastatic renal cell carcinoma. Int J Clin Oncol 2004;9:369 –72.
[26] Sevinc M, Kirkali Z, Yorukoglu K, et al. Prognostic significance of microvascular invasion in localized renal cell carcinoma. Eur Urol 2000;38:728 –33. [27] Lang H, Lindner V, Saussine C, et al. Microscopic venous invasion: A prognostic factor in renal cell carcinoma. Eur Urol 2000; 38:600 –5. [28] Sorbellini M, Kattan M, Snyder M, et al. A postoperative prognostic nomogram predicting recurrence for patients with conventional clear cell carcinoma. J Urol 2005;173:48 –51.