Renal Cell Carcinoma With Tumor Thrombus Extension: Biology, Role of Nephrectomy and Response to Immunotherapy

0022-5347/03/1693-0909/0 THE JOURNAL OF UROLOGY® Copyright © 2003 by AMERICAN UROLOGICAL ASSOCIATION

Vol. 169, 909 –916, March 2003 Printed in U.S.A.

DOI: 10.1097/01.ju.0000045706.35470.1e

RENAL CELL CARCINOMA WITH TUMOR THROMBUS EXTENSION: BIOLOGY, ROLE OF NEPHRECTOMY AND RESPONSE TO IMMUNOTHERAPY AMNON ZISMAN,* JEFF A. WIEDER, ALLAN J. PANTUCK, DEBBY H. CHAO, FREDERICK DOREY, JONATHAN W. SAID, BARBARA J. GITLITZ, JEAN B. DEKERNION, ROBERT A. FIGLIN AND ARIE S. BELLDEGRUN† From the Kidney Cancer Program, Division of Urologic Oncology, Department of Urology, Department of Medicine and Division of Hematology-Oncology, Department of Pathology and Laboratory Medicine, University of California School of Medicine, Los Angeles, California

ABSTRACT

Purpose: We outline the biology, prognosis and role of immunotherapy for renal cell carcinoma with gross venous tumor thrombus. Materials and Methods: A total of 207 patients with unilateral renal cell carcinoma and tumor thrombus into the renal vein (107) and inferior vena cava (100) who underwent nephrectomy and thrombectomy were compared with 607 without tumor thrombus. Results: At diagnosis 77 patients (37%) had N0M0 disease and 130 (63%) had lymph node (N⫹) or distant (M1) metastases. Compared with nontumor thrombus cases tumor thrombus was associated with more advanced stage, N⫹ (26% versus 12%), M1 (54% versus 31%) disease, higher grade and Eastern Cooperative Oncology Group performance status. In N0M0 cases with inferior vena caval tumor thrombus capsular penetration, collecting system invasion and extension into the hepatic vein were more important prognostic variables then the level of inferior vena caval thrombus. In patients with confined N0M0 tumors mean 2 and 5-year survival ⫾ SD was 83% ⫾ 8.8% and 72% ⫾ 10.7% in those with inferior vena caval tumor thrombus, and 90% ⫾ 9.4% and 68% ⫾ 16.1% in those with renal vein tumor thrombus, similar to the 93.4% ⫾ 1.7% and 81 ⫾ 3.1% rates, respectively, in those without thrombus who had no recurrence within 6 months after nephrectomy. Of patients with M1 disease in whom cytoreductive surgery was done those with and without thrombus showed a similar response to immunotherapy. When there was inferior vena caval and renal vein thrombus, mean 2-year survival was higher after nephrectomy and immunotherapy than after nephrectomy alone (41% ⫾ 9% and 52% ⫾ 7% versus 32% ⫾ 13% and 45% ⫾ 7%), immunotherapy alone (0% and 13% ⫾ 12%, respectively) and no treatment (0%). Conclusions: Renal cell carcinoma with tumor thrombus is associated with worse characteristics. Local tumor extension has greater prognostic importance than the level of inferior vena caval tumor thrombus. Survival is fair in patients with truly confined N0M0 disease and thrombus. The combination of surgery and immunotherapy has a role in thrombus cases. Our data provide the rationale for a prospective study of adjuvant immunotherapy after surgery in N0M0 cases with extensive tumor thrombus. KEY WORDS: kidney; carcinoma, renal cell; vena cava, inferior; thrombosis; renal veins

Renal cell carcinoma invades the venous system in 4% to 9% of cases.1– 4 Much of the literature suggests that tumor thrombus into the inferior vena cava is not necessarily associated with a worse prognosis and patients with tumor thrombus but no metastasis can achieve long-term survival after complete surgical extirpation.3, 5–11 However, the prognostic importance of the tumor thrombus level is controversial.3, 10, 12–15 With current surgical technology thrombus resection can be performed even when it extends into the right atrium. The role of nephrectomy and thrombectomy when there is capsular invasion,10 –12, 14 –17, regional lymph node involvement (N⫹) or distant metastasis (M1) 1, 3, 5, 11–16, 18 is not well defined. These factors may be important for determining the response to immunotherapy. Recent studies suggest that the combination of aggressive cytoreductive nephrectomy and immunotherapy may improve survival in

patients with M1 disease who have inferior vena caval tumor thrombus.19, 20 However, to our knowledge it is not known whether tumor thrombus extent impacts the response to immunotherapy. We outline the biology of renal cell carcinoma with gross tumor thrombus extension and define independent variables impacting survival. In addition, we determined whether patients with N0M0 disease and inferior vena caval involvement should be considered to be at increased risk for disease recurrence.

PATIENTS AND METHODS

Patient population. A total of 1,087 patients were treated for renal cell carcinoma in the kidney cancer program at our institution between 1989 and 2000. Of these 1,087 patients 264 (24%) had gross tumor thrombus extension into the venous system, including 130 (12%) into the renal vein and 134 (12%) into the inferior vena cava. We excluded 13 patients from study, including 10 (renal vein thrombus in 8 and inferior vena caval thrombus in 2) with bilateral renal cell car-

Accepted for publication October 4, 2002. * Financial interest and/or other relationship with Galil, Inc. † Financial interest and/or other relationship with Agensis, Chiron and Galil, Inc. 909

910

SURGERY AND IMMUNOTHERAPY FOR RENAL CELL CARCINOMA WITH TUMOR THROMBUS

cinoma and 3 (renal vein thrombus in 1 and inferior vena caval thrombus in 2) with missing data. The remaining 251 patients with tumor thrombus comprise our principal study population. Of these 251 patients 44 (renal vein thrombus in 14 and inferior vena caval thrombus in 30) did not undergo surgery because they refused surgery, the condition was unresectable at exploration (10) or they were not considered surgical candidates because of widespread metastatic disease or Eastern Cooperative Oncology Group performance status (ECOG) greater than 1. The remaining 207 patients (renal vein thrombus in 107 and inferior vena caval thrombus in 100) underwent unilateral radical nephrectomy and tumor thrombus extraction. We compared 607 nontumor thrombus cases with thrombus cases (pathological stages pT1 in 272, pT2 in 124, pT3a in 202 and pT4 in 9). Histology and grade. Tumor histology type and grade were determined according to the 1997 WHO classification21 and Fuhrman’s criteria.22 Tumor stage was assigned based on the 1997 American Joint Committee on Cancer TNM staging system.23 N0M0 tumors were further stratified as confined or unconfined. A confined primary tumor was defined as an N0M0 tumor confined within the renal capsule with negative surgical margins and no invasion into the renal calices or pelvis. When tumor thrombus was present, confined disease was defined as a confined primary tumor without hepatic vein extension. In this report patients are designated using the extent of tumor thrombus (none, renal vein or inferior vena caval thrombus) and NM stage. For example, thrombus, N⫹M1 indicates gross venous tumor thrombus with metastasis to regional nodes and distant metastasis. Inferior vena caval thrombus, N0M0 confined indicates inferior vena caval thrombus with stage N0M0 confined disease, whereas inferior vena caval N0M0 unconfined indicates unconfined disease. When confinement status is not indicated, for example inferior vena caval N0M0, all confined and unconfined N0M0 cases are included. ECOG performance status and the University of California, Los Angeles Integrated Staging System (UISS)24 were determined for each patient. The UISS, which predicts patient survival based on 1997 TNM stage, Fuhrman’s grade and ECOG performance status, stratifies patients into 5 risk groups. Survival and prognosis. In N0M0 cases failure was defined as the development of local or systemic renal cell carcinoma after complete surgical extirpation. Survival was calculated from the date of nephrectomy or from the initial renal cell carcinoma consultation at our institution in nonsurgical cases. The survival end point was the date of the last followup or death. The disease-free interval (time from nephrectomy to failure), failure site and response to immunotherapy were analyzed. Multivariate analysis was performed in N0M0 and in M⫹ cases to test whether tumor thrombus influenced the prognosis more than the UISS. Thereafter, univariate and multivariate analyses were performed for groups in which thrombus extension had an independent impact on survival. The variables tested were patient age, sex, T stage, distant metastasis (M1), metastasis to regional nodes (N⫹), tumor grade, ECOG performance status, incidentally detected tumor, tumor side, tumor thrombus extension into the hepatic veins, tumor penetration or adhesion to the vein wall, capsular penetration, caliceal or renal pelvic invasion, nephrectomy, positive surgical margins, caudal venous extension toward the iliac veins, cephalad thrombus level (subdiaphragmatic, supradiaphragmatic or atrial), immunotherapy, metastasis only to the lung and histological type. Immunotherapy in patients with tumor thrombus. A total of 123 patients with thrombus, ECOG performance status 1 or less and no medical contraindications underwent immunotherapy. Of these 123 patients 106 (86%) received interleukin-2 based therapy. Immunotherapy was adminis-

tered in 23 of the 33 N0M0 failed (70%), 91 of the 150 M1 (61%) and 9 of the 16 N⫹M0 cases. We defined the response to immunotherapy as complete response— the complete disappearance of all clinically detectable disease, partial response—a 50% or greater decrease in the tumor burden without any new lesions, stable disease—less than a 25% increase or less than a 50% decrease in tumor burden and progressive disease—a 25% or greater increase in the tumor burden. No evidence of disease was defined as the complete resection of evaluable metastasis before immunotherapy with a durable disease-free interval after immunotherapy. Responses to immunotherapy were grouped as major— complete response and no evidence of disease, minor—partial response and stable disease, and nonresponders—progressive disease. Two patients lost to followup after immunotherapy were presumed to have had progression. Some statistical tests were performed only in cases of metastasis at diagnosis and an ECOG performance status of 1 or less to minimize selection bias and exclude those in which palliative surgery was done. Statistics. Survival was assessed by the product-time method of Kaplan-Meier. Comparisons of groups were performed using the log rank method to assess the significance of Kaplan-Meier curves and the Cox regression model. Entry and removal criteria were 0.1 and 0.15, respectively. The chi-square, Student t and Kruskal-Wallis tests for ranked data were performed as indicated. Data were analyzed using commercially available software. RESULTS

Presentation. Table 1 lists the characteristics of the study population. Of the 207 surgical patients with tumor thrombus 77 (37%) had nonmetastatic renal cell carcinoma (N0M0), whereas 130 (63%) had metastasis to regional lymph nodes (N⫹) or distant sites (M1). The lung was the most prevalent location of metastasis alone in 33 of 111 cases (30%) and combined with other sites in 59 (53%). Only extrapulmonary metastases were present in 19 cases (17%). Of the 100 patients with inferior vena caval thrombus 79 had infradiaphragmatic and 21 had supradiaphragmatic extension, including 14 (67%) with extension into the right atrium. Nine patients (infradiaphragmatic extension in 6 and supradiaphragmatic extension in 3) had tumor thrombus extension into the hepatic veins and 14 (infradiaphragmatic extension in 11 and supradiaphragmatic extension in 3) had tumor thrombus extending toward the iliac veins. Extent of tumor thrombus. Clinical and Pathological Differences: Patients with thrombus had a higher ECOG performance status and were less likely to undergo surgical treatment. The frequency of nonoperative treatment increased as thrombus extent increased, including 2.1% in nonthrombus, 11% in renal vein (p ⫽ 0.00001) and 22% in inferior vena caval (p ⫽ 0.01) cases. Compared with patients without thrombus those with thrombus had larger tumors (p ⬍0.05), extension into the perinephric fat (renal vein and inferior vena caval thrombus p ⫽ 0.0001 and 0.02), invasion into the renal pelvis or major calices (p ⫽ 0.004 and 0.0001), involvement of adjacent organs (p ⫽ 0.001 and 0.0002, respectively), more advanced 1997 TNM stage (p ⬍0.008), N⫹ disease (12% versus 26%, p ⫽ 0.0002), distant metastasis (31% versus 54%, p ⫽ 0.0001), higher Fuhrman grade (p ⬍0.0001), higher ECOG performance status (p ⬍0.003) and higher UISS category (p ⬍0.0001). These trends were significant when the renal vein and inferior vena caval subgroups were separately compared with nonthrombus cases. The rate of positive surgical margins was similar in thrombus and nonthrombus cases. Sarcomatoid, collecting duct and unclassified tumors as a group were more common in the renal vein and inferior vena caval subgroups (p ⫽ 0.008 and 0.002, respectively). The distribution of histological type did not differ in renal vein and inferior vena caval cases (p ⫽ 0.27). However, pa-

911

SURGERY AND IMMUNOTHERAPY FOR RENAL CELL CARCINOMA WITH TUMOR THROMBUS TABLE 1. Study population characteristics Surgical Renal Vein

Nonthrombus

Renal Vein

Inferior Vena Cava

607 30.9 6.7 ⫾ 3.7

107 34 9.1 ⫾ 4.0

100 25.7 10.2 ⫾ 3.8

13 19.1 6.7 ⫾ 4.0

14 All pts. died 9.1 ⫾ 3.5

30 2.4 12.5 ⫾ 4.5

210 ⫾ 35

52 ⫾ 52 79 21 7 32 ⫾ 32† 3⫾3 14 26 40 ⫾ 40 6⫾6 3⫾3 31 ⫾ 31.0 17 ⫾ 17.0 3

8† ⫾ 62

8 ⫾ 1.3 61 ⫾ 10.1 24 ⫾ 4.0 263 130 393 ⫾ 65 8 ⫾ 1.3 14 ⫾ 2.3 140 ⫾ 23.1 50 ⫾ 8.2 0

61 ⫾ 57 107 0 11 ⫾ 10.3 22 ⫾ 21† 0 21 16 37 ⫾ 35 4 ⫾ 3.7 3 ⫾ 2.8 43 ⫾ 40.2 20 ⫾ 18.7 0

0*

5 ⫾ 36* 14 0 1

12† ⫾ 40* 16 14 7*

5 ⫾ 39 0 0 5 ⫾ 38.5 3 ⫾ 23.1 0

0 0 2 ⫾ 14.3 6 ⫾ 42.9 6 ⫾ 42.9 0

237 (39.0) 79 (13.0) 85 (14.0) 206 (33.9)

0 0 40 (37) 67 (63)

0 0 46 (46) 54 (54)

4 (30.8) 0 1 (7.7) 8 (61.5)

428 (73) 55 (9.4) 18 (3.1) 46 (7.8) 30 (5.1) 3 (0.5) 8 (1.4) 19 (3.1)

78 (77) 2 (2) 0 7 (6.9) 11 (11) 1 (1) 3 (2.9) 5 (4.7)

67 (68) 5 (5.1) 2 (2) 9 (9.1) 9 (9.1) 0 7 (7.1) 1 (1)

5 (39) 0 8 (61) 0 0 0

267 (44.0) 293 (48.3) 21 (3.5) 0 26 (4.3)

28 (28) 64 (64) 7 (7) 1 (1) 7

90 (14.8) 304 (50.1) 160 (26.3) 21 (3.5) 32 (5.3) 130 (21.4) 222 (36.6) 54 (8.9) 157 (25.9) 11 (1.8) 33 (5.4)

Nonthrombus No. pts. Median survivor followup (mos.) Mean tumor size ⫾ SD (cm.) No. tumor extent (%): T3a (invades perinephric fat) T3b (thrombus below diaphragm) T3c (thrombus above diaphragm) T4 Caliceal invasion Pos. surgical margin N0M0 confined N0M0 unconfined N0M0 (all) N1M0 N2M0 N0M1 N⫹M1 NxM0 No. 1997 TNM stage (%): I II III IV No. histological category (%): Clear cell Chromophil (papillary) Chromophobe Mixed Sarcomatoid Collecting duct Unclassified Missing data No. ECOG performance status (%): 0 1 2 3 Missing data No. Fuhrman’s grade (%): 1 2 3 4 Missing data No. UISS (%): I II III IV V Missing data * On imaging.

Nonsurgical Inferior Vena Cava

1⫾3 0 1 ⫾ 3.3 7 ⫾ 23.3 20 ⫾ 66.7 0

0 0 0 14 (100)

0 0 1 (3) 29 (97)

0

3 (33) 0 1 (11) 1 (11) 2 (22) 0 2 (22) 5 (36)

6 1 0 1 1 0 3 18 (60)

26 (27) 68 (70) 3 (3) 0 3

1 (7.7) 5 (38.5) 7 (53.8) 0 0

3 (21) 8 (57) 2 (14) 1 (7) 0

0 15 (52) 7 (24) 7 (24) 1

4 (4) 36 (37) 51 (52) 7 (7) 9

1 (1) 36 (38) 52 (54) 7 (7) 4

1 (7.7) 1 (7.7) 4 (30.8) 0 7 (53.8)

1 (14) 0 3 (43) 3 (43) 7

1 (14) 0 3 (43) 3 (43) 23

0 19 (20) 16 (17) 54 (57) 6 (6) 12

0 17 (18) 32 (34) 44 (46) 2 (2) 5

1 (7.7) 1 (7.7) 0 4 (30.8) 0 7 (53.8)

0 0 1 (14) 3 (43) 3 (43) 7

0 0 0 4 (57) 3 (43) 23

tients with renal vein thrombus were more likely to have confined disease than those with inferior vena caval thrombus (p ⫽ 0.05). The anatomical distribution of metastases was not influenced by thrombus extent, that is no thrombus, renal vein or inferior vena caval thrombus (p ⫽ 0.015). Survival and Recurrence: Table 2 shows disease-free survival. Nonthrombus cases had a significantly longer freedom from failure interval than thrombus cases (p ⫽ 0.02, fig. 1). In N0M0 confined cases disease specific survival decreased as thrombus extent increased, as shown by the progressive decrease in the 2-year mean survival rate of 95% ⫾ 1.4% for no thrombus to 84% ⫾ 8.3% for renal vein to 76% ⫾ 12.3% for inferior vena caval cases. In the presence of regional lymph node metastasis and/or distant metastasis thrombus extent no longer impacted survival (fig. 2). On multivariate analysis performed separately in inferior vena caval N0M0 cases thrombus (hazard coefficient 0.64, p ⫽ 0.013) and the UISS (hazard coefficient 0.72, p ⫽ 0.0001) were independent covariables associated with survival, whereas in renal vein N0M0 N⫹ or M1 renal vein and inferior vena caval cases thrombus did not attain co-significance with the UISS. The anatomical distribution of metastases was not influenced by thrombus extent. A similar metastatic pattern was observed

in nonthrombus, renal vein and inferior vena caval thrombus cases (p ⫽ 0.15). Confined versus unconfined tumors. Overall survival in

TABLE 2. Two and 5-year disease specific survival subgroups of surgical patients with inferior vena caval and renal vein tumor thrombus Mean % Survival ⫾ SD 2-Yr. Nonthrombus N0M0 confined 95 ⫾ 1.4 Inferior vena cava: N0M0 confined 76 ⫾ 12.3 N0M0 unconfined 60 ⫾ 11.1 N⫹M0 55 ⫾ 18.6 N0M1 43 ⫾ 9.6 N⫹M1 33 ⫾ 21.8 Renal vein: N0M0 confined 84 ⫾ 8.3 N0M0 unconfined 77 ⫾ 12 N⫹M0 60 ⫾ 21.9 N0M1 44 ⫾ 7.8 N⫹M1 22 ⫾ 10.5 * Last death at 14 months with 2 of 17 patients alive after † Fewer than 10 patients at risk at 60 months of followup.

5-Yr. 88 ⫾ 2.4 56 ⫾ 15.2 35 ⫾ 12.9 27 ⫾ 21.4 21 ⫾ 8.3 33 ⫾ 21.8* 72 ⫾ 10.8 48 ⫾ 18.1 0⫹ 21 ⫾ 7.9† 14 ⫾ 9.2† 5 years.

912

SURGERY AND IMMUNOTHERAPY FOR RENAL CELL CARCINOMA WITH TUMOR THROMBUS

FIG. 1. Freedom from failure in patients with N0M0 disease without tumor thrombus (Non-TT), with renal vein thrombus (RV) and with inferior vena caval thrombus (IVC). NS, not statistically significant.

FIG. 2. Kaplan-Meier overall survival estimates for patients with N0M0 disease who had confined tumor without tumor thrombus (Non-TTN0M0-C), confined tumor with inferior vena caval thrombus (IVCN0M0-C), unconfined tumor with inferior vena caval thrombus (IVCN0M0-nC), and inferior vena caval thrombus with regional lymph node metastasis (IVCN⫹M0), distant metastasis (IVCN0M1), and regional lymph node and distant metastasis (IVCN⫹M1). NS, not statistically significant.

patients without thrombus who had N0M0 confined disease was better than in those with inferior vena caval N0M0 confined disease but this difference did not achieve significance (p ⫽ 0.17, fig. 2). Of patients with inferior vena caval thrombus N0M0 confined disease was associated with significantly better overall survival than N0M0 unconfined tumors. Overall survival in patients with inferior vena caval thrombus and N0M0 unconfined tumors was similar to that in those with inferior vena caval N⫹M0, N0M1 and N⫹M1 disease (fig. 2). Thus, patients with inferior vena caval thrombus had 2 tiers of overall survival from highest to lowest, including N0M0 confined greater than unconfined or metastatic (fig. 2). In patients with renal vein thrombus there was no survival difference for N0M0 confined and N0M0 unconfined disease. However, confined and unconfined renal vein N0M0 tumors were associated with longer survival than renal vein N⫹M0 or N0M1 disease (p ⫽ 0.007, fig. 3). In turn the latter patients had longer survival than those with renal vein N⫹M1 cancer (p ⫽ 0.01). Thus, patients with renal vein tumor thrombus had 3 tiers of overall survival from highest to lowest, including confined or unconfined N0M0 greater than lymph node invasion or distant metastasis greater than lymph node invasion and distant metastasis. On univariate and multivariate analyses performed in the entire study group the significant variables were ECOG performance status (mean Cox coefficient ⫾ SE 0.98 ⫾ 0.118, 95% CI 0.746 to 1.21, p ⫽ 0.0001), 1997 T stage (0.65 ⫾ 0.099, 95% CI 0.45 to 0.84, p ⫽ 0.0001) and Fuhrman grade (0.61 ⫾ 0.108, 95% CI 0.4 to 0.82, p ⫽ 0.0001). Inferior vena caval thrombus was not significant (Cox coefficient ⫺0.22 ⫾ 0.185, 95% CI 0.58 to 0.15, p ⫽ 0.24). Therefore, univariate and multivariate analyses were performed in inferior vena caval N0M0 cases using the relevant variables listed. Tumor extension into perinephric fat (hazard ratio 2.7, 95% CI 1.2 to 5.9, p ⫽ 0.01), invasion into the collecting system (hazard ratio 2.7, 95% CI 1.3 to 5.7, p ⫽ 0.01) and thrombus extension into the hepatic veins (hazard ratio 18.9, 95% CI 4.1 to 87.8, p ⫽ 0.0001) were independent co-variables in addition to the UISS. Simulation based on the Cox hazard model showed the negative impact of these variables (table 3). Multivariate analysis also revealed that the level of the inferior vena caval thrombus (subdiaphragmatic, supradiaphragmatic or atrial) was not an independent variable that impacted prognosis. Metastasis. In patients with renal vein tumor thrombus lymph node or distant metastasis significantly decreased overall survival compared with renal vein N0M0 disease (median survival 14 and 84 months, respectively, p ⫽ 0.0001,

FIG. 3. Kaplan-Meier overall survival estimates in patients with tumor thrombus and no (N0M0) or any (N⫹ and M1) metastasis. RV, renal vein. IVC, inferior vena cava.

913

SURGERY AND IMMUNOTHERAPY FOR RENAL CELL CARCINOMA WITH TUMOR THROMBUS

TABLE 3. Effect of tumor extension into the perinephric fat, collecting system and hepatic veins on predicted 2 and 5-year disease specific survival in patients with N0M0 disease by Cox hazard model % Disease Specific Survival

Extension

Nonthrombus N0M0 confined Inferior vena caval thrombus N0M0:

Into Calix/Renal Pelvis

Through Renal Capsule

Into Hepatic Veins

2-Yr.

5-Yr.

Neg.

Neg.

Not applicable

95

88

Neg. Pos. Neg. Pos. Neg. Neg. Pos.

Neg. Neg. Pos. Pos. Neg. Pos. Neg.

Neg. Neg. Neg. Neg. Pos. Pos. Pos.

51 25 30 8 0 0 0

Pos.

Pos.

Pos.

83 68 51 20 7 4 Less than 1 0

fig. 3). However, in patients with inferior vena caval thrombus median overall survival was similar for N⫹, M1 and N0M0 disease (21 and 28 months, respectively, p ⫽ 0.11, fig. 3). Recurrence within 6 months. Tumor recurred in 105 of the 393 nonthrombus N0M0 (27%), 17 of the 37 renal vein N0M0 (46%) and 16 of the 40 inferior vena caval N0M0 (40%) cases. There was no difference in freedom from failure in renal vein and inferior vena caval cases. Of the 33 patients with recurrent tumor thrombus there was failure only in the lungs in 6 (18%), in the lungs and other organs in 4 (12%), and outside of the lungs in 23 (70%). Recurrence 6 months or less after nephrectomy was observed in 4 renal vein N0M0 (11%), 5 inferior vena caval N0M0 (13%) and 17 nonthrombus N0M0 (4%) cases (p ⫽ 0.03, table 4). Of patients with N0M0 confined tumors 4 of 35 (11%) had recurrence within 6 months of surgery compared with only 6 of 263 (2.3%) without thrombus (p ⬍0.05). Furthermore, when there was recurrence 6 months or less after excision all 4 patients with tumor thrombus died compared with 3 of the 6 (50%) without thrombus at 5, 6 and 21 months, respectively. Of cases of thrombus the proportion of failures within 6 months of nephrectomy was similar for N0M0 confined (4 of 35 or 11%) and N0M0 unconfined (4 of 40 or 10%) disease. Mean 2 and 5-year overall survival ⫾ SD in patients with confined disease without failure less than 6 months after nephrectomy was 83% ⫾ 8.8% and 72% ⫾ 10.7% for inferior vena caval N0M0 confined disease, 90% ⫾ 9.4% and 68 ⫾ 16.1% for renal vein N0M0 confined disease (p ⫽ 0.931), and 93.4% ⫾ 1.7% and 81% ⫾ 3.1% for nonthrombus N0M0 confined disease, respectively (versus renal vein thrombus p ⫽ 0.015). Immunotherapy. N0M0 Disease With Recurrence: Immunotherapy was administered in 100 patients with N0M0 disease with recurrence after surgery, including 77 of 105 without thrombus (75%), 11 of 17 with renal vein thrombus (65%) and 12 of 16 with inferior vena caval thrombus (75%) (p ⬎0.5). The mean interval ⫾ SD from surgery to immunotherapy was 40 ⫾ 34.6, 44 ⫾ 38.8 and 15 ⫾ 12.6 months, respectively. Patients with inferior vena caval thrombus had a significantly shorter time to immunotherapy than those without thrombus (p ⫽ 0.001). Information on the best re-

0

sponse to immunotherapy was available on 95 of 100 cases. Patients with inferior vena caval thrombus in whom treatment failed had a significantly lower response to immunotherapy than those with renal vein thrombus and no thrombus. There was a major response in 5 nonthrombus (7%), 1 renal vein (10%) and 0 inferior vena caval cases, and a minor response in 38 (52%), 6 (60%) and 2 (17%), respectively. Furthermore, no response was more common in patients with inferior vena caval tumor thrombus, including 30 (41%), 3 (30%) and 10 (83%), respectively. M1 Disease at Presentation: Of patients with M1 disease and tumor thrombus who underwent surgery and immunotherapy 12 of 48 with inferior vena caval (25%) and 21 of 63 with renal vein (33%) tumors had metastasis only to the lungs. When metastases were confined to the lungs, median survival was longer after immunotherapy in these 2 subgroups, including 26 months with 0.228 deaths per year of followup for renal vein thrombus, M1 disease, lung only cases, 10 months with 0.456 deaths per year (RR 2) for renal vein thrombus, M1 disease, nonlung only cases, 36 months with 0.214 deaths per year for inferior vena caval, M1 disease, lung only cases and 13 months with 0.588 deaths per year (RR 2.8) for inferior vena caval, M1, nonlung only cases. Patients without tumor thrombus in whom metastases were limited to the lungs had a 2-fold higher rate of complete response to immunotherapy than those with extrapulmonary disease (p ⫽ 0.03). The same trend was noted in patients with tumor thrombus but it did not attain statistical significance due to small sample size (p ⫽ 0.09; table 5). To define further the impact of surgery, extent of tumor thrombus and pulmonary metastasis on the response to immunotherapy in M1 cases the Kruskal-Wallis test was performed. Patients with M1 disease and tumor thrombus who underwent surgery had a higher response to immunotherapy than those treated nonoperatively (p ⫽ 0.0052, table 5). Surgical patients with and without tumor thrombus had a similar response to immunotherapy (p ⫽ 0.17), even when the renal vein and inferior vena caval subgroups were analyzed separately (p ⫽ 0.14). However, patients with tumor thrombus who underwent surgery had a significantly better response to immunotherapy than those with thrombus who

TABLE 4. Recurrence rate in certain patient subsets No. Recurrences/Total No. Pts. (%) Nonthrombus All N0M0 105/393 (27) N0M0, failure within 6 mos. 17/393 (4) N0M0 confined, failure within 6 mos. 6/263 (2) N⫹M0 10/22 (45) * Renal vein versus nonthrombus Yates correction p ⫽ 0.01. † Inferior vena cava versus, nonthrombus Yates correction p ⫽ 0.05.

Renal Vein

Inferior Vena Cava

17/37 (46) 4/37 (11) 3/21 (14)* 6/7 (86)

16/40 (40) 5/40 (13) 2/14 (14)† 5/9 (56)

p Value

0.015 0.03 0.0006 Not significant

914

SURGERY AND IMMUNOTHERAPY FOR RENAL CELL CARCINOMA WITH TUMOR THROMBUS TABLE 5. Best response to immunotherapy in patients with M1 disease at presentation No. Pts.

Surgical: Renal vein Inferior vena cava Nonsurgical: Renal vein Inferior vena cava Surgical: All thrombus All nonthrombus Thrombus, lung metastasis only Nonthrombus, lung metastasis only

No. Response (%) Major

Minor

None

Not Evaluable

Totals

63 48

4 (10) 4 (12)

23 (55) 11 (33)

14 (33) 18 (55)

1 (2) 0

42 (67) 33 (69)

12 27

0 0

1 (13) 2 (25)

7 (88) 6 (75)

34 (45) 73 (60) 13 (56) 28 (60)

32 (43) 42 (34) 6 (26) 11 (23)

111 190 111 190

8 (11) 8 (7) 4 (17) 6 (13)

were treated nonoperatively (p ⫽ 0.02). The same observation was noted when the comparison was restricted to patients with M1 disease at diagnosis and an ECOG performance status of 1 or less (p ⫽ 0.02). Comparison of therapies. Of the patients with M1 disease and inferior vena caval thrombus those who underwent nephrectomy and thrombectomy had longer survival than those treated nonoperatively (fig. 4, A). The combination of immunotherapy and cytoreductive nephrectomy improved median overall survival over nephrectomy and thrombectomy alone (3 months longer), although the difference did not achieve significance (fig. 4, A). Survival in patients with inferior vena caval thrombus who received only immunotherapy was similar to that in those who underwent no treatment (fig. 4, A). In cases of renal vein thrombosis associated with M1 disease the greatest survival was achieved with surgery and immunotherapy (fig. 4, B). When renal vein and inferior vena caval cases were grouped together, survival was significantly longer in M1 cases with tumor thrombus in which nephrectomy with thrombectomy and immunotherapy were performed than in cases in which single modality (p ⫽ 0.001) or no (p ⫽ 0.0001) treatment was done. DISCUSSION

Analysis of our experience reveals that renal cell carcinoma with tumor thrombus is more likely to be larger, higher grade and higher stage, and be more frequently associated with metastases and a worse prognosis. Patients with tumor thrombus have poorer performance status, are less likely to undergo surgical treatment and have a higher recurrence rate. Thus, cancer with tumor thrombus is biologically more aggressive and carries a worse clinical outcome. Tumors with renal vein and inferior vena caval tumor thrombus have comparable clinical and pathological characteristics, suggesting similar intrinsic behavior. Inferior vena caval and renal vein tumors also have a similar distribution of histological types, rate of metastasis at diagnosis and anatomical site of metastases. Clinically N0M0 cases with renal vein or inferior vena caval thrombus have a comparable ECOG performance status, recurrence rate (46% and 40%, respectively) and freedom from failure as well as a similar response to immunotherapy (fig. 1). Differences in renal vein and inferior vena caval tumors exist and are probably related to lead time differences. A tumor thrombus observed for a period begins as a renal vein thrombus and progresses to involve the inferior vena cava. Patients with renal vein thrombus were more likely to have cancer confined to the renal capsule and more likely to be surgical candidates than those with inferior vena caval thrombus. When diagnosed early, disease specific survival of patients with N0M0 disease tends to be higher in those with renal vein than with inferior vena caval thrombus. Therefore, it appears that tumors with a renal vein and inferior vena caval thrombus have similar tumor biology but earlier diagnosis resulted in an overall poorer outcome. The overall and

0 0 1 (1) 0 0 2 (4)

8 (67) 8 (30) 75 (68) 123 (65) 23 (17) 47 (24)

median survival of patients with confined N0M0 disease and inferior vena caval thrombus are in accordance with the rates in previous reports3, 8, 11, 12 These studies indicate that when the primary tumor is confined to the kidney, inferior vena caval tumor thrombus does not impact survival. Our data confirm these observations in confined tumors without failure within 6 months after nephrectomy, whereas overall survival was lower in the overall population of patients with confined N0M0 disease when tumor thrombus was present. Of the patients with inferior vena caval tumor thrombus those with unconfined tumors had significantly worse overall survival than those with confined disease. Multivariate analysis of inferior vena caval N0M0 cases revealed that the determinants of an unconfined tumor, including capsular penetration, invasion into the collecting system and thrombus extension into the hepatic veins, were independently prognostic co-variables in these cases together with UISS. However, the level of the inferior vena caval thrombus (subdiaphragmatic, supradiaphragmatic or atrial), invasion into the inferior vena caval wall and positive surgical margins status did not influence the prognosis over the UISS. Our study supports previous reports stressing the importance of local invasiveness1, 13, 17 rather than tumor thrombus level,16 inferior vena caval wall invasion3 or the completeness of resection.3 N0M0 cases with failure 6 months or less after surgery may represent a subgroup with occult metastasis at surgery that are not detected by currently available staging modalities. There was recurrence within 6 months in 11% of confined N0M0 disease cases with tumor thrombus in our series. When compared with renal vein thrombus cases, tumors associated with inferior vena caval thrombus had a similar recurrence rate 6 months or less after surgery. Significantly more patients with tumor thrombus had recurrence within 6 months than those without thrombus, suggesting that tumor thrombus is associated with a higher rate of micrometastasis. Thus, micrometastatic spread is an early event that occurs during the initiation of tumor thrombus. The future development of kidney cancer specific tumor markers may help identify these early failures. Patients with tumor thrombus had a higher rate of metastasis at diagnosis than those without tumor thrombus (63% versus 25%). Metastasis was most common in the lungs. Of the 207 patients with tumor thrombus 74 (36%) presented with nonlymph node metastasis (N0M1), whereas only 16 (8%) presented with lymph node metastasis (N⫹M0). Therefore, most metastases in patents with tumor thrombus probably develop early and hematogenously rather than through the lymphatics. The overall survival of patients with inferior vena caval thrombus in our series was similar regardless of metastatic status. This observation suggests that patients with inferior vena caval tumor thrombus should be regarded as being at high risk with a higher potential for metastatic spread. Patients with inferior vena caval thrombus and N0M0

SURGERY AND IMMUNOTHERAPY FOR RENAL CELL CARCINOMA WITH TUMOR THROMBUS

915

FIG. 4. Kaplan-Meier overall survival estimates for patients with metastasis, tumor thrombus and ECOG performance status 1 or less at diagnosis who underwent nephrectomy and immunotherapy (Nx⫹IMT⫹), nephrectomy only (Nx⫹IMT-), immunotherapy only (Nx-IMT⫹) and no treatment (Nx-IMT-). A, inferior vena caval (IVC) thrombus. B, renal vein (RV) thrombus. NS, not statistically significant.

disease with recurrence after surgery had a lower rate of response to immunotherapy and a higher rate of progression than their renal vein and nontumor thrombus counterparts. Patients with M1 disease and thrombus who underwent surgery had a significantly better response to immunotherapy than those treated nonoperatively. Immunotherapy after cytoreductive nephrectomy was associated with a similar response rate in thrombus and nonthrombus cases. In patients with renal vein thrombus and metastasis at presentation surgical extirpation and immunotherapy achieved the longest survival (fig. 4, B). In patients with inferior vena caval thrombus and metastasis at presentation the longest survival was achieved by surgical extirpation with or without immunotherapy (fig. 4, A). Our data suggest that the combination of cytoreductive surgery and immunotherapy has an important role in patients with renal vein thrombus and a potential role in those with inferior vena caval thrombus. In the current study 32 M1 cases with tumor thrombus and an ECOG performance status of 1 or less (nonpalliative) were managed by cytoreductive nephrectomy and tumor thrombus extraction followed by immunotherapy. These patients had a mean survival of 16 months, a 2-year survival of 41% and a

5-year survival of 24%. Quek et al reported comparable survival in a similar group of patients,9 whereas those with M1 disease who underwent only nephrectomy had a median survival of only 5 to 10 months.3, 12 Thus, it appears that cytoreductive surgery and immunotherapy in patients with M1 disease and tumor thrombus result in prolonged survival, especially in those with tumor thrombus and lung only metastasis, as previously reported.20 CONCLUSIONS

Renal cell carcinoma with tumor thrombus represents more aggressive disease. Local tumor extension has greater prognostic importance than the level of inferior vena caval tumor thrombus. Patients with thrombus and truly confined N0M0 disease have excellent survival. The combination of surgery and immunotherapy improves survival in M1 cases with thrombus, suggesting that cytoreductive surgery and immunotherapy have an important role in tumor thrombus cases. Although this series is the largest in recent literature to detail the outcome in patients with localized or metastatic

916

SURGERY AND IMMUNOTHERAPY FOR RENAL CELL CARCINOMA WITH TUMOR THROMBUS

renal tumors and gross venous extension, it is limited in being retrospective in nature and influenced by selection and referral biases. Thus, a prospective study to analyze the role of adjuvant immunotherapy in this select group of patients is needed. Incorporating molecular markers in the future would further enhance the subclassification of this group of patients and identify those most likely to benefit from various therapies. REFERENCES

1. Hoehn, W. and Hermanek, P.: Invasion of veins in renal cell carcinoma—frequency, correlation and prognosis. Eur Urol, 9: 276, 1983 2. Casanova, G. A. and Zingg, E. J.: Inferior vena caval tumor extension in renal cell carcinoma. Urol Int, 47: 216, 1991 3. Hatcher, P. A., Anderson, E. E., Paulson, D. F., Carson, C. C. and Robertson, J. E.: Surgical management and prognosis of renal cell carcinoma invading the vena cava. J Urol, 145: 20, 1991 4. Pagano, F., Dal Bianco, M., Artibani, W., Pappagallo, G. and Prayer-Galetti, T.: Renal cell carcinoma with extension into the inferior vena cava: problems in diagnosis, staging and treatment. Eur Urol, 22: 200, 1992 5. Mattos, R. M. and Libertino, J. A.: Survival of patients with renal cell carcinoma invading the inferior vena cava. Semin Urol Oncol, 14: 223, 1996 6. Skinner, D. G., Colvin, R. B., Vermillion, C. D., Pfister, R. C. and Leadbetter, W. F.: Diagnosis and management of renal cell carcinoma. A clinical and pathologic study of 309 cases. Cancer, 28: 1165, 1971 7. Skinner, D. G., Pfister, R. F. and Colvin, R.: Extension of renal cell carcinoma into the vena cava: the rationale for aggressive surgical management. J Urol, 107: 711, 1972 8. Skinner, D. G., Pritchett, T. R., Lieskovsky, G., Boyd, S. D. and Stiles, Q. R.: Vena caval involvement by renal cell carcinoma. Surgical resection provides meaningful long-term survival. Ann Surg, 210: 387, 1989 9. Quek, M. L., Stein, J. P. and Skinner, D. G.: Surgical approaches to venous tumor thrombus. Semin Urol Oncol, 19: 88, 2001 10. Glazer, A. A. and Novick, A. C.: Long-term followup after surgical treatment for renal cell carcinoma extending into the right atrium. J Urol, 155: 448, 1996 11. Libertino, J. A., Zinman, L. and Watkins, E., Jr.: Long-term results of resection of renal cell cancer with extension into inferior vena cava. J Urol, 137: 21, 1987 12. Ljungberg, B., Stenling, R., Osterdahl, B., Farrelly, E., Aberg, T. and Roos, G.: Vein invasion in renal cell carcinoma: impact on metastatic behavior and survival. J Urol, 154: 1681, 1995

13. Tongaonkar, H. B., Dandekar, N. P., Dalal, A. V., Kulkarni, J. N. and Kamat, M. R.: Renal cell carcinoma extending to the renal vein and inferior vena cava: results of surgical treatment and prognostic factors. J Surg Oncol, 59: 94, 1995 14. Staehler, G. and Brkovic, D.: The role of radical surgery for renal cell carcinoma with extension into the vena cava. J Urol, 163: 1671, 2000 15. Kuczyk, M. A., Bokemeyer, C., Kohn, G., Stief, C. G., Machtens, S., Truss, M. et al: Prognostic relevance of intracaval neoplastic extension for patients with renal cell cancer. Br J Urol, 80: 18, 1997 16. Sosa, R. E., Muecke, E. C., Vaughan, E. D., Jr. and McCarron, J. P., Jr.: Renal cell carcinoma extending into the inferior vena cava: the prognostic significance of the level of vena caval involvement. J Urol, 132: 1097, 1984 17. Heney, N. M. and Nocks, B. N.: The influence of perinephric fat involvement on survival in patients with renal cell carcinoma extending into the inferior vena cava. J Urol, 128: 18, 1982 18. Emmott, R. C., Hayne, L. R., Katz, I. L., Scribner, R. G. and Tawes, R. L., Jr.: Prognosis of renal cell carcinoma with vena caval and renal vein involvement. An update. Am J Surg, 154: 49, 1987 19. Slaton, J. W., Balbay, M. D., Levy, D. A., Pisters, L. L., Nesbitt, J. C., Swanson, D. A. et al: Nephrectomy and vena caval thrombectomy in patients with metastatic renal cell carcinoma. Urology, 50: 673, 1997 20. Naitoh, J., Kaplan, A., Dorey, F., Figlin, R. and Belldegrun, A.: Metastatic renal cell carcinoma with concurrent inferior vena caval invasion: long-term survival after combination therapy with radical nephrectomy, vena caval thrombectomy and postoperative immunotherapy. J Urol, 162: 46, 1999 21. Storkel, S., Eble, J. N., Adlakha, K., Amin, M., Blute, M. L., Bostwick, D. G. et al: Classification of renal cell carcinoma: Workgroup No. 1. Union Internationale Contre le Cancer (UICC) and the American Joint Committee on Cancer (AJCC). Cancer, 80: 987, 1997 22. Fuhrman, S., Lasky, L. C. and Limas, C.: Prognostic significance of morphologic parameters in renal cell carcinoma. Am J Surg Pathol, 6: 655, 1982 23. Guinan, P., Sobin, L., Algaba, F., Badellino, F., Kameyama, S., MacLennan, G. et al: TNM staging of renal cell carcinoma: Workgroup No. 3—Union International Contre le Cancer (UICC) and the American Joint Committee on Cancer (AJCC). Cancer, 80: 992, 1997 24. Zisman, A., Pantuck, A. J., Dorey, F., Said, J. W., Shvarts, O., Quintana, D. et al: Improved prognostication of RCC using an Integrated Staging System. J Clin Oncol, 19: 1649, 2001