Clinicopathologic and molecular correlations of necrosis in the primary tumor of patients with renal cell carcinoma

Authors

  • John S. Lam M.D.,

    1. Department of Urology, David Geffen School of Medicine at University of California-Los Angeles, Los Angeles, California
    Search for more papers by this author
    • Dr. Lam and Dr. Shvarts contributed equally to the current study.

  • Oleg Shvarts M.D.,

    1. Department of Urology, David Geffen School of Medicine at University of California-Los Angeles, Los Angeles, California
    Search for more papers by this author
    • Dr. Lam and Dr. Shvarts contributed equally to the current study.

  • Jonathan W. Said M.D.,

    1. Department of Urology, David Geffen School of Medicine at University of California-Los Angeles, Los Angeles, California
    2. Department of Pathology and Laboratory Medicine, David Geffen School of Medicine at University of California-Los Angeles, Los Angeles, California
    Search for more papers by this author
  • Allan J. Pantuck M.D.,

    1. Department of Urology, David Geffen School of Medicine at University of California-Los Angeles, Los Angeles, California
    Search for more papers by this author
  • David B. Seligson M.D.,

    1. Department of Pathology and Laboratory Medicine, David Geffen School of Medicine at University of California-Los Angeles, Los Angeles, California
    Search for more papers by this author
  • Michael E. Aldridge B.S.,

    1. Department of Urology, David Geffen School of Medicine at University of California-Los Angeles, Los Angeles, California
    Search for more papers by this author
  • Matthew H. T. Bui M.D., Ph.D.,

    1. Department of Urology, David Geffen School of Medicine at University of California-Los Angeles, Los Angeles, California
    Search for more papers by this author
  • Xueli Liu Ph.D.,

    1. Department of Biostatistics and Human Genetics, David Geffen School of Medicine at University of California-Los Angeles, Los Angeles, California
    Search for more papers by this author
  • Steve Horvath Ph.D.,

    1. Department of Biostatistics and Human Genetics, David Geffen School of Medicine at University of California-Los Angeles, Los Angeles, California
    Search for more papers by this author
  • Robert A. Figlin M.D.,

    1. Department of Urology, David Geffen School of Medicine at University of California-Los Angeles, Los Angeles, California
    2. Department of Medicine, David Geffen School of Medicine at University of California-Los Angeles, Los Angeles, California
    Search for more papers by this author
  • Arie S. Belldegrun M.D.

    Corresponding author
    1. Department of Urology, David Geffen School of Medicine at University of California-Los Angeles, Los Angeles, California
    • Department of Urology, David Geffen School of Medicine at University of California-Los Angeles, CHS 66-118, P.O. Box 951738, 10833 Le Conte Avenue, Los Angeles, CA 90095-1738
    Search for more papers by this author
    • Fax: (310) 206-5343


Abstract

BACKGROUND

The presence of histologic necrosis in the primary tumor of patients with renal cell carcinoma (RCC) has been suggested to be an important predictor of survival. The authors investigated the relation of tumor necrosis to other clinicopathologic factors known to be important prognostic indicators for patients with RCC.

METHODS

The records of 311 patients undergoing treatment for RCC were evaluated for basic clinicopathologic information including TNM classification, nuclear grade, Eastern Cooperative Oncology Group (ECOG) performance status (PS), disease recurrence, and survival. The presence and extent of histologic necrosis of the primary tumors was recorded and correlated with clinicopathologic factors, carbonic anhydrase IX and Ki-67 expression, disease recurrence, and survival.

RESULTS

The presence of necrosis in the primary tumor of patients with RCC compared with patients with RCC without necrosis was associated with higher T classification (P < 0.0001), the presence of lymph node disease (P = 0.009), the presence of metastases (P < 0.0001), higher grade (P < 0.0001), greater mean tumor size (P < 0.0001), an ECOG PS score ≥ 1 (P = 0.007), higher University of California-Los Angeles Integrated Staging System (UISS) category (P < 0.0001), and higher Ki-67 expression (P < 0.0001). The extent of necrosis in the primary tumor was associated with the presence of lymph node disease (P = 0.009) and the presence of metastases (P < 0.0001), and correlated with higher T classification (σ = 0.31, P < 0.0001), poorer ECOG PS (σ = 0.18, P = 0.002), higher grade (σ = 0.33, P < 0.0001), greater tumor size (σ = 0.40, P < 0.0001), higher UISS category (σ = 0.37, P < 0.0001), and higher Ki-67 staining (σ = 0.32, P < 0.0001). Patients with the presence of necrosis in the primary tumor demonstrated a lower 5-year disease-specific survival compared with patients without necrosis in the primary tumor (36% vs. 75%; P < 0.0001). Multivariate analysis demonstrated that T classification (P < 0.0001), distant metastases (P < 0.0001), and ECOG PS (P < 0.0001) were independent predictors of DSS, whereas the presence of necrosis was not (P = 0.1100). Substratification into localized and metastatic disease demonstrated that the presence of necrosis was an independent predictor of survival in patients with localized (P = 0.025), but not metastatic (P = 0.44), disease. The extent of necrosis was not an independent predictor of survival (P > 0.05). Patients with the presence of necrosis in the primary tumor had a lower 5-year disease recurrence-free rate compared with patients without the presence of necrosis (62% vs. 92%, P < 0.0001).

CONCLUSIONS

The presence of necrosis in the primary tumor was associated with adverse prognostic factors such as high T classification, presence of lymph node disease and metastases, high grade, large tumor size, and poor ECOG PS. The extent of necrosis was found to be associated with the presence of lymph node disease and metastases and correlated with higher T classification, higher grade, greater tumor size, poorer ECOG PS, and higher UISS category. The presence of this histologic variant was an independent predictor of poor survival in patients with localized, but not metastatic, disease. In addition, Ki-67 expression served as a valuable surrogate marker for the presence of histologic tumor necrosis. Cancer 2005. © 2005 American Cancer Society.

Staging provides critical prognostic and therapeutic information in the management of patients with renal cell carcinoma (RCC). The systematic staging of patients provides a tool for clinicians to implement in counseling patients with regards to prognosis and anticipated life expectancy. In addition, the characterization of high-stage patients can help to identify those patients at high risk for local and metastatic disease recurrences after treatment. Such information is useful for selecting patients for adjuvant and experimental therapies. In addition, the establishment of various risk groups can help to create tailored surveillance regimens for patients based on risk of disease progression and disease recurrence.

Historically, anatomic stage has been the gold standard among prognostic factors. After the first anatomic staging system devised by Robson et al.,1 the TNM classification system was created to stratify patients according to anatomic factors such as size and local extent of the primary tumor, involvement of locoregional lymph nodes, and the presence of distant metastases.2 Although the TNM classification system has played a dominant role in staging, several other factors have been demonstrated to impact the prognosis of patients with RCC. Several studies have demonstrated that clinical factors such as symptomatic presentation and poor performance status (PS) impart a poor prognosis.3, 4 Other studies have focused on the histologic characteristics of tumors, demonstrating that tumor grade and histologic subtypes play an important role in determining outcomes for patients with RCC.5, 6

Tumor necrosis is a histologic feature that has presented significant controversy. The presence of tumor necrosis in the pathologic specimen may reflect tumor biology and may also provide additional useful prognostic information. It is believed to be present when tumors outgrow their blood supply and, therefore, histologic tumor necrosis has been proposed to be a sign of tumor aggressiveness that generally leads to poor clinical outcomes.7 Several studies have investigated the utility of tumor necrosis as a prognostic factor for patients with RCC and have reported conflicting results.8–12 Given this controversy, we evaluated the role of necrosis as a predictor of clinical prognosis for patients with RCC.

MATERIALS AND METHODS

Patient Selection

The medical records of 311 patients undergoing nephrectomy for localized or metastatic RCC at the University of California-Los Angeles (UCLA) Medical Center between 1989 and 2000 were reviewed retrospectively. All available RCC specimens from the UCLA tumor specimen archives were arrayed. The study cohort constituted a representative sample of all available patients treated with this condition during the study period. Clinicopathologic data including tumor size, anatomic stage, Eastern Cooperative Oncology Group (ECOG) PS,13 nuclear grade, UCLA Integrated Staging System (UISS) category,14 disease recurrence, and survival were collected for each patient. Anatomic stage was determined according to the 1997 TNM staging classification system2 and assessment of nuclear grade was carried out according to criteria first proposed by Fuhrman et al.5 Patients with an M classification of 1 or an N classification of > 0 were considered to have metastatic disease. Survival was determined from the time of surgical treatment to the time of last follow-up. This cohort exhibited a similar distribution of patient and tumor characteristics to a UCLA series of 643 consecutive patients from the same study period and database that validated the 1997 TNM classification system.15

Evaluation of Tumor Necrosis

All available histologic sections from the primary tumor of each patient were individually reviewed by a single pathologist (J.W.S.) blinded to the clinical data. The presence or absence of necrosis and the extent of necrosis was evaluated and recorded for all 311 specimens. Estimation of the percentage of tumor necrosis was based on the macroscopic description of the tumor and histologic evaluation of all available tumor blocks. The tumors were sectioned at a minimum of 1-cm intervals so that the number of tissue blocks evaluated was partly determined by tumor size. The number of sections per specimen was variable, but included a minimum of 1 section per 1 cm of tumor. The extent of necrosis was graded according to the following scale: 0, no necrosis; 1, < 25% necrosis; 2, 25–50% necrosis; 3, 51–75% necrosis; and 4, 76–100% necrosis.

Tissue Array Construction

Archival tumor specimens from our cohort of 311 patients were obtained from the Department of Pathology at UCLA Medical Center. Briefly, 3 core tissue biopsy specimens, measuring 0.6 mm in greatest dimension, were taken from selected morphologically representative regions of each paraffin-embedded renal tumor and precisely arrayed using a custom-built instrument, as described previously.16 An additional core tissue biopsy specimen was taken from a morphologically normal-appearing region of each tumor. Sections of 4-μm thickness of each tissue array block were transferred to glass slides using the paraffin sectioning aid system, consisting of adhesive-coated slides PSA-CS4x, adhesive tape, and an ultraviolet lamp (Instrumedics Inc., Hackensack, NJ) to support the cohesion of 0.6-mm array elements. The quality control of histology and grade was assessed on hematoxylin-eosin–stained tissue array sections as well as on the array sections used for antibody analysis.

Immunohistochemistry

The presence of carbonic anhydrase IX (CA IX) and Ki-67 was assessed using a previously described methodology.17, 18 Immunohistochemical staining of tissue sections was performed using a peroxidase technique with antigen retrieval using heat treatment, as previously described using the Dako staining systems (Dako Corporation, Carpenteria, CA) or Vector ABC Elite (Vector Laboratories, Burlingame, CA).19 CA IX staining was performed using the mouse monoclonal antibody (MoAb) M75 at a 1:10,000 dilution and Ki-67 staining was performed using the mouse MoAb MIB-1 (Dako Corporation), at a dilution of 1:100 (final concentration 0.5 μg/mL). Semiquantitative assessment of antibody staining was performed by a single pathologist (D.B.S.) blinded to the clinicopathologic variables. Staining intensity was based on a 4-point scale from 0 to 3. The extent of staining was recorded as a percentage of the target tissue sample that had positive expression. Each spot was scored based on the staining intensity, the percentage of positive cells, and the percentage staining at maximal staining intensity. There were three spots per patient specimen. The overall score used for subsequent statistical analysis was the pooled mean from three spots of the same tumor specimen. A score of zero was given to tissue spots that had no evidence of specific immunostaining. For Ki-67, staining > 10% irrespective of intensity was considered high Ki-67 staining, whereas those ≤ 10% was considered low Ki-67 staining. For CA IX, staining > 85% irrespective of intensity was considered high CA IX staining, whereas those ≤ 85% was considered low CA IX staining

Statistical Methods

The Kruskal–Wallis test was used to compare ordinal variables between patients with and without the presence of histologic necrosis and with different extents of histologic necrosis. The Fisher exact test was used to analyze the dependence between categorical variables to examine possible differences between patients with and without the presence of histologic necrosis. The Spearman correlation coefficient was used to determine the correlation between quantitative and ordinal variables. Cancer-specific and disease recurrence-free survival was calculated using the Kaplan–Meier method. Deaths from causes other than RCC were censored. Log-rank tests were used to compare survival curves. Univariate and multivariate analysis based on the Cox regression model was performed to assess the independent associations of clinicopathologic variables with cancer-specific and disease recurrence-free survival.20 All P values were 2 sided and P < 0.05 was significant. All data were statistically analyzed using the freely available R statistical software package (available from URL: http://www.r-project.org ).

RESULTS

Patient and tumor characteristics are shown in Table 1. Of 311 patients, 187 (60%) presented with localized disease and 123 (40%) presented with metastatic disease. The median age of the patients was 62 years (range, 27–89 years). The male-to-female ratio was 2:1. The median follow-up for patients who died from an RCC-related cause was 14 months (range, 1.3–101 months), whereas the median follow-up for all survivors was 45 months (range, 0.3–117 months). Of the tumor specimens evaluated, 270 (87%) were clear cell carcinomas, 30 (10%) were papillary carcinomas, 7 (2%) were chromophobes, and 4 (1%) were collecting duct carcinomas. The number of patients that presented with T1, T2, T3, and T4 disease was 115 (37%), 42 (14%), 136 (44%), and 17 (6%), respectively. Lymph node disease was present in 37 (12.2%) patients. The number of patients with Fuhrman Grade 1, 2, 3, and 4 tumors was 40 (13%), 146 (48%), 107 (35%), and 12 (4%), respectively. The number of patients with ECOG PS scores of 0, 1, 2, and 3 was 120 (39%), 175 (57%), 12 (4%), and 1 (1%), respectively. Distribution of patients according to UISS category I, II, III, IV, and V was 78 (27%), 67 (23%), 23 (8%), 106 (37%), and 10 (3.5%) patients, respectively.

Table 1. Patient and Tumor Characteristics (n = 311)
CharacteristicsNo. of patients (%)
  1. ECOG-PS: Eastern Cooperative Oncology Group performance status; UISS: University of California-Los Angeles Integrated Staging System.

Median age (yrs) (range)62 (27–89)
Male to female ratio2:1
Localized disease187 (60)
Metastatic disease123 (40)
Histologic subtype 
 Clear cell270 (87)
 Papillary30 (10)
 Chromophobe7 (2)
 Collecting duct4 (1)
T classification 
 1115 (37)
 242 (14)
 3136 (44)
 417 (6)
Lymph node positive disease37 (12)
Grade 
 140 (13)
 2146 (48)
 3107 (35)
 412 (4)
ECOG-PS score 
 0120 (39)
 1175 (57)
 212 (4)
 31 (1)
UISS category 
 178 (27)
 267 (23)
 323 (8)
 4106 (37)
 510 (4)

The prevalence of histologic necrosis according to clinicopathologic characteristics is shown in Table 2. Necrosis was present in 54% of all specimens evaluated. The presence of histologic necrosis in the primary tumor was significantly higher in patients with metastatic disease compared with patients with localized disease (71% vs. 42%, P < 0.0001). Similarly, necrosis was more likely to be present in tumor specimens with associated locoregional lymph node involvement than tumor specimens without lymph node involvement (76% vs. 51%, P = 0.009). Evaluation according to T classification demonstrated that necrosis was present in 34%, 57%, 68%, and 77% of T1, T2, T3, and T4 tumor specimens, respectively (P < 0.0001) and in 28%, 45%, 72%, and 100% of patients with Grade 1, 2, 3, and 4 tumors, respectively (P < 0.0001). Patients with an ECOG PS score of 0 were less likely to have necrosis present in their primary tumor compared with patients with an ECOG PS score ≥ 1 (43% vs. 62%, P = 0.007). Stratification according to UISS categories demonstrated that necrosis was present in 27%, 49%, 74%, 71%, and 100% of the primary tumors of patients with UISS categories I–V, respectively (P < 0.0001). The mean tumor size of tumor specimens with necrosis was significantly larger than tumor specimens without necrosis (8.6 ± 3.7 cm vs. 5.7 ± 3.5 cm, P < 0.0001). The presence of histologic necrosis was not found to be associated with CA IX staining for patients with either metastatic or localized disease (P > 0.05), whereas the presence of necrosis was found to be significantly associated with high Ki-67 staining (P < 0.0001).

Table 2. Prevalence of Necrosis According to Clinicopathologic Characteristics
CharacteristicsNo. of patients (%)P valuea
  • ECOG-PS: Eastern Cooperative Oncology Group-performance status; UISS: University of California-Los Angeles Integrated Staging System.

  • a

    P value was derived from a Fisher exact or Kruskal–Wallis test.

Overall prevalence168 (54) 
Presence of metastatic disease <0.0001
 No74 (42) 
 Yes90 (71) 
Lymph node status 0.009
 Lymph node negative disease136 (51) 
 Lymph node positive disease28 (76) 
Histologic subtype 0.017
 Clear cell152 (56) 
 Papillary16 (53) 
 Chromophobe0 (0) 
 Collecting duct1 (25) 
T classification <0.0001
 139 (34) 
 224 (57) 
 392 (68) 
 413 (77) 
Grade <0.0001
 111 (28) 
 266 (45) 
 377 (72) 
 412 (100) 
ECOG-PS score 0.007
 051 (43) 
 ≥1117 (62) 
UISS category <0.0001
 121 (27) 
 233 (49) 
 317 (74) 
 475 (71) 
 510 (100) 

The extent of histologic necrosis was also correlated with established clinicohistologic prognostic indicators. The extent of necrosis in the primary tumor specimen was found to be significantly correlated with advanced T classification (σ = 0.31, P < 0.0001), poorer ECOG PS (σ = 0.18, P = 0.002), higher grade (σ = 0.33, P < 0.0001), larger tumor size (σ = 0.40, P < 0.0001), and higher UISS category (σ = 0.37, P < 0.0001). The extent of necrosis in the primary tumor specimen also was found to be associated with the presence of lymph node disease (P = 0.009) and the presence of metastases (P < 0.0001). The extent of histologic necrosis was not correlated with CA IX staining for patients with either metastatic or localized disease (P > 0.05), but was significantly correlated with high Ki-67 staining (σ = 0.32, P < 0.0001).

Kaplan–Meier estimates for all patients with RCC demonstrated that patients whose primary tumor specimens did not manifest histologic necrosis had significantly higher 5-year disease-specific survival (DSS) compared with patients with tumor specimens that manifested necrosis (75% vs. 36%, P < 0.0001) (Fig. 1A). For patients with localized RCC, Kaplan–Meier estimates also demonstrated a significantly higher 5-year DSS for patients without the presence of necrosis compared with patients with the presence of necrosis in the primary tumor specimen (86% vs. 67%, P = 0.0022) (Fig. 1B). However, Kaplan–Meier estimates did not demonstrate a significant difference in 5-year DSS between patients with or without necrosis in the primary tumor specimens for patients with metastatic RCC (19% vs. 35% vs. P = 0.0770) (Fig. 1C). The analysis was also subsequently limited to patients with clear cell RCC and the results were similar to those obtained for all patients with RCC.

Figure 1.

Disease-specific survival for patients with renal cell carcinoma. Kaplan–Meier estimates according to the presence of histologic necrosis in the primary tumor. (A) All patients (P < 0.0001). (B) Patients with localized disease (P = 0.0022). (C) Patients with metastatic disease (P = 0.077).

In univariate Cox proportional hazards analysis for all patients with RCC, the presence of necrosis, greater extent of necrosis, advanced T classification, the presence of distant metastases, higher grade, and poorer ECOG PS were all negative predictors of DSS (P < 0.0001) (Table 3). In multivariate Cox proportional hazards analysis for all patients with RCC, the presence of necrosis was analyzed with T classification, the presence of distant metastases, grade, and ECOG PS for their impact on DSS (Table 3). For all patients with RCC, T classification (hazard ratios [HR], 1.71; P < 0.0001, 95% confidence interval [95% CI], 1.33–2.19), the presence of distant metastases (HR, 3.33; P < 0.0001, 95% CI, 2.15–5.16), and ECOG PS (HR, 1.55; P = 0.0029, 95% CI, 1.16–2.07) were retained as independent predictors of survival, whereas the presence of necrosis (P = 0.11) and grade were not (P = 0.17). However, when multivariate analysis was limited to patients with localized RCC, patients with histologic tumor necrosis had a 2.71-fold higher risk of death compared with patients without necrosis (P = 0.025; 95% CI, 1.13–6.47). T classification (HR, 2.10; P = 0.0019; 95% CI, 1.32–3.36) was also retained as an independent predictor of survival, whereas grade (P = 0.4) and ECOG PS (P = 0.33) were not retained as independent predictors (Table 3). For patients with metastatic disease, T classification (HR, 1.50; P = 0.013; 95% CI, 1.09–2.06) and ECOG PS (HR, 1.48; P = 0.021; 95% CI, 0.74–1.98) were retained as independent predictors of survival, whereas grade (P = 0.49) and the presence of necrosis (P = 0.44) were not (Table 3). The extent of necrosis was not an independent predictor of survival for all patients (P = 0.059), patients with localized disease (P = 0.29), or patients with metastatic disease (P = 0.11). Univariate and multivariate analyses were also subsequently limited to only patients with clear cell RCC and the results were similar to those obtained for all patients with RCC.

Table 3. Cox Regression Analyses for Disease-Specific Survival for all Patients with RCC
CharacteristicsHazard ratio95% CIP value
  1. RCC: renal cell carcinoma; 95% CI: 95% confidence interval; ECOG-PS: Eastern Cooperative Oncology Group-performance status.

All patients   
Univariate   
 Presence of necrosis2.962.01–4.35<0.0001
 Extent of necrosis1.441.24–1.69<0.0001
 T classification2.461.99–3.02<0.0001
 Distant metastases6.514.42–9.58<0.0001
 Grade2.221.74–2.84<0.0001
 ECOG-PS2.632.05–3.38<0.0001
Multivariate   
 Presence of necrosis1.410.93–2.150.11
 T classification1.711.33–2.19<0.0001
 Distant metastases3.332.15–5.16<0.0001
 Grade1.210.92–1.580.17
 ECOG-PS1.551.16–2.070.0029
Localized patients   
 Univariate   
  Presence of necrosis3.241.53–6.890.0022
  Extent of necrosis1.391.03–1.880.029
  T classification2.311.58–3.38<0.0001
  Grade1.270.73–2.200.0099
  ECOG-PS2.771.42–5.380.0027
 Multivariate   
  Presence of necrosis2.711.13–6.470.025
  T classification2.101.32–3.360.0019
  Grade1.210.92–1.580.4
  ECOG-PS1.460.69–3.120.33
Metastatic patients   
 Univariate   
  Presence of necrosis1.540.95–2.480.077
  Extent of necrosis1.351.06–1.730.016
  T classification1.611.19–2.170.002
  Grade1.361.00–1.850.049
  ECOG-PS1.521.09–2.110.012
 Multivariate   
  Presence of necrosis1.210.74–1.980.44
  T classification1.501.09–2.060.013
  Grade1.120.81–1.550.49
  ECOG-PS1.480.74–1.980.021

An analysis of the impact of histologic tumor necrosis on disease recurrence after nephrectomy for patients with localized disease also was performed. Patients without the presence of necrosis in their primary tumor specimen demonstrated a 5-year disease recurrence-free survival of 92% compared with 62% for patients with necrosis present in their tumor specimens (P < 0.0001) (Fig. 2). Univariate analysis demonstrated that the presence of necrosis (P = 0.0002), greater extent of necrosis (P = 0.013), advanced T classification (P < 0.0001), higher grade (P = 0.015), and poorer ECOG PS (P < 0.0001) were positive predictors of disease recurrence after nephrectomy. Multivariate analysis evaluating T classification, grade, ECOG PS, and the presence of tumor necrosis as predictors of tumor recurrence demonstrated that patients with histologic tumor necrosis were 4 times as likely to have a tumor recurrence compared with patients without necrosis present in the primary tumor specimen (P = 0.0068; 95% CI, 1.48–11.50) (Table 4). T classification (HR, 2.05; P = 0.0064; 95% CI, 1.22–3.44) and ECOG PS (HR, 2.40; P = 0.037; 95% CI, 1.06–5.44) were also retained as independent predictors of disease recurrence, whereas grade (P = 0.41) was not (Table 4).

Figure 2.

Disease recurrence-free survival for patients with localized renal cell carcinoma. Kaplan–Meier estimates according to the presence of histologic necrosis in the primary tumor (P < 0.0001).

Table 4. Cox Regression Analyses for Disease Recurrence Free Survival for all Patients with Localized RCC
CharacteristicsHazard ratio95% CIP value
  1. RCC: renal cell carcinoma; 95% CI: 95% confidence interval; ECOG-PS: Eastern Cooperative Oncology Group-performance status

Univariate   
 Presence of necrosis5.632.27–13.90.0002
 Extent of necrosis1.481.09–2.020.013
 T classification2.391.59–3.59<0.0001
 Grade2.051.15–3.640.015
 ECOG-PS4.432.16–9.10<0.0001
Multivariate   
 Presence of necrosis4.121.48–11.500.0068
 T classification2.051.22–3.440.0064
 Grade1.270.72–2.250.41
 ECOG-PS2.401.06–5.440.037

DISCUSSION

Historically, histologic tumor necrosis has been associated with more aggressive tumor behavior. Brinker et al.21 suggested that more aggressive tumors demonstrated necrosis and that these tumors tended to be associated with higher stage, grade, size, and extrarenal extension. Furthermore, tumors with necrosis progressed in 40% of patients. A more recent study by Leibovitch et al.8 evaluated 173 patients undergoing nephrectomy for RCC and found that 17.9% of these patients had evidence of histologic tumor necrosis. The authors of that study demonstrated that the presence of necrosis was associated with larger tumors, as well as with more perirenal involvement and venous extension. Our data confirm the association of histologic necrosis with aggressive tumor features, demonstrating a statistically significant correlation between this histologic variant and more advanced stage, higher grade, larger size, and poorer PS. In addition, we found that the extent of histologic necrosis helps to further gauge the aggressiveness of these tumors.

The role of histologic tumor necrosis as a prognostic determinant of survival for patients with RCC has been more controversial. An early study in 1974 first demonstrated that the presence of histologic necrosis was an adverse prognostic indicator.9 Roosen et al.10 subsequently verified this finding whereas Delahunt et al.11 found that necrosis in the tumor specimen had no association with survival. Leibovitch et al.,8 as well, demonstrated that this histologic variable was not shown to be a good predictor of survival as DSS did not differ for patients without and with necrotic tumors (68% vs. 66%, respectively).

Recently, several studies from the Mayo Clinic have demonstrated the importance of histologic tumor necrosis as an independent predictor of survival in patients with clear cell RCC. Evaluating > 1800 patients with clear cell RCC, Frank et al.12 found that necrosis was an independent predictor of survival according to both univariate and multivariate analyses with risk ratios of 4.51 and 1.92, respectively. Necrosis was such an important prognostic factor in that study that the histologic variant was included, along with TNM classification, tumor size, and nuclear grade, in the creation of a comprehensive prognostic tool called the stage, size, grade, and necrosis score that predicts DSS for patients with clear cell RCC. However, 84.2% of the patients in the Frank et al. study had localized disease. The authors further studied patients with localized clear cell RCC and demonstrated that necrosis also helps to predict disease progression and disease recurrence for this patient population.22, 23

The current study data confirm the importance of histologic necrosis as a prognostic indicator for patients with RCC. We found that patients with histologic evidence of necrosis in their primary tumor specimens demonstrated a 5-year DSS rate of 36% versus 75% for patients without this histologic variant (P < 0.0001). Multivariate analysis, however, demonstrated that the poor survival of these patients could not be attributed to the presence of necrosis, but rather to the other prognostic factors with which it was associated, such as advanced stage, grade, and ECOG PS score. This finding was in stark contrast to the data presented by Frank et al.12 from the Mayo Clinic. We attribute this difference to the finding that 40% of our patient population presented with metastatic disease compared with the 15% of patients in the study by Frank et al. To test our assertion, we separately analyzed the impact of histologic necrosis on the prognosis of patients with localized and metastatic disease. As demonstrated by Frank et al.,12 histologic necrosis was found to be an independent predictor of survival for patients with localized disease (HR, 2.71; P = 0.025; 95% CI, 1.13–6.47). We similarly demonstrated the importance of necrosis as an independent predictor of disease recurrence in this localized patient population (HR, 4.12; P = 0.0068; 95% CI, 1.48–11.50), reporting 5-year disease recurrence rates of 38% and 8% for tumors with and without necrosis, respectively. In contrast, however, for patients presenting with metastatic disease, tumor necrosis had no independent impact on DSS (P = 0.44).

Previous studies have demonstrated a differential impact of necrosis as a prognostic indicator on patients with clear cell RCC versus patients with other pathologic variants such as papillary RCC. Cheville et al.24 found that although histologic tumor necrosis was significantly associated with cancer-related death from clear cell RCC, it had no association with death from papillary RCC. To exclude this possible negative influence of patients with papillary RCC, we further limited our study population to only include patients with clear cell RCC. A multivariate analysis of these patients similarly demonstrated that necrosis was an independent predictor of survival for patients with localized (HR, 2.51; P = 0.044; 95% CI, 1.03–6.13), but not metastatic (P = 0.49), clear cell RCC.

Several possibilities can explain why necrosis appears to be more significant for localized rather than metastatic tumors. Tumors with evidence of histologic necrosis are usually more aggressive, expanding to the extent that they outgrow their blood supplies, cause hypoxia, and create the necrosis visible histologically.7 It is not surprising, then, that these aggressive tumors often lead to distant metastasis.21 The presence of distant metastases, in turn, imparts a poor prognosis to patients who initially present with localized disease. In contrast, for patients already presenting with metastatic disease, any impact of primary tumor necrosis on local and distant spread is negligible given that the prognoses of all patients in this population are equally impacted by the presence of already existent metastases. As a result, the presence of histologic necrosis is not an independent predictor of survival in patients with metastatic disease.

The presence of histologic necrosis has been correlated with staining for CA IX in numerous tumor types.25, 26 Tissue specimens from these different tumors were stained with CA IX and demonstrated the presence of this marker in perinecrotic areas.26, 27 This staining pattern has been explained by the finding that, in most tissue specimens, CA IX is regulated by the transcription factor, hypoxia inducible factor (HIF), which is upregulated by the presence of hypoxia.27 When tumors are exposed to hypoxia as a result of rapid growth beyond existing blood supplies, the furthest extent of the tumors become necrotic.7 In the more proximal, perinecrotic areas, the hypoxic conditions stimulate the production of HIF, which further stimulates increased production of CA IX. CA IX subsequently alters the extracellular milieu of the cancer cells, optimizing conditions for growth and invasion.27 As a result, staining of tissue specimens from carcinomas of the breast and head and neck have demonstrated high CA IX expression adjacent to areas of necrosis.25, 26 We evaluated this relation between CA IX and necrosis. However, analysis of the tumor specimens in our patient population failed to demonstrate a correlation between CA IX expression and necrosis (P > 0.05). Similarly, no correlation was observed between these two factors when the analysis was restricted to patients with either localized or metastatic disease. This finding is attributed to defects in the von Hippel-Lindau (VHL) genes in most RCC tumors. In normal tissue specimens, the expression of these genes restricts the upregulation of HIF.28 In RCC tissue specimens, however, a large percentage of patients have some defect in the VHL genes, which eliminates the regulatory function of this gene, allowing the constitutive production of HIF and CA IX in these tumors.29 As a result, unlike in other cancers in which CA IX production occurs in direct response to hypoxia and necrosis, CA IX production in RCC tumors is constitutive and does not correlate with the presence of necrosis.

The results of the current study also demonstrate a significant correlation between the presence of histologic tumor necrosis and Ki-67 staining in RCC tumors. This relation can be explained by the underlying tumor characteristics that govern the expression of these markers. Ki-67 is a nuclear antigen present in the G1, S, G2, and M phases of all cycling human cells.30 The expression of Ki-67 has long been established as a marker for the proliferative capacity of tumors.31 This marker also has been demonstrated to be an independent predictor of survival in patients with RCC.32 Because tumors possess areas of histologic necrosis as a result of rapid, hypoxia-producing growth, the cells composing these tumors should, in turn, have high proliferative activity.7 The highly proliferative cells composing necrotic tumors, in turn, manifest high levels of Ki-67 staining. As a result, Ki-67 expression and histologic necrosis are significantly related in RCC tumors. We believe that this relation is extremely important in the context of accurately and reproducibly predicting prognosis in this patient population. The identification of the presence and extent of tumor necrosis is highly dependent on the pathologist who performs the evaluation. As such, the accuracy and reproducibility of using histologic necrosis within the context of a prognostic algorithm is yet to be determined. In contrast, staining with Ki-67 is a procedure familiar to most pathologists that yields a definitive result not requiring interpretation. Given this increased accuracy and reproducibility as well as the significant relation between Ki-67 and necrosis, we believe that Ki-67 can be employed as a reliable surrogate for histologic tumor necrosis.

Our findings demonstrate the prognostic impact of histologic tumor necrosis in the primary tumor specimens of patients with RCC. Manifesting the aggressive nature of localized tumors, the presence and extent of histologic tumor necrosis is associated and correlated with more advanced stage, higher grade, larger tumor size, and poorer PS, but is not associated with increased expression of CA IX in patients with RCC. Although patients with evidence of histologic tumor necrosis demonstrate substantially lower survival overall and increased disease recurrence rates, the presence of necrosis itself appears to be an independent predictor of survival in patients with localized, but not metastatic, RCC. Given its relation to necrosis and the ease and reproducibility with which it can be employed, Ki-67 staining can be used reliably as a surrogate marker for histologic tumor necrosis in the context of a prognostic algorithm.

Ancillary