Prospective study of the relationship between the systemic inflammatory response, prognostic scoring systems and relapse-free and cancer-specific survival in patients undergoing potentially curative resection for renal cancer

Authors


Sara Ramsey, Department of Urology, Gartnavel General Hospital, Glasgow G12 0YN, UK.
e-mail: sara_l_ramsey@ntlworld.com

Abstract

OBJECTIVE

To examine the prognostic value of markers of systemic inflammatory response, together with established scoring systems, in predicting relapse-free and cancer-specific survival in patients with primary operable renal cancer, as there is increasing evidence that such markers provide prognostic information, in addition to scoring systems, in patients with metastatic renal cancer.

PATIENTS AND METHODS

In all, 83 patients undergoing potentially curative nephrectomy for localized renal cancer were recruited. The University of California Los Angeles Integrated Staging System (UISS), ‘Stage Size Grade Necrosis’ (SSIGN) and Kattan scores were constructed. The systemic inflammatory response was assessed by counting white cells, neutrophils, lymphocytes and platelets, and measuring albumin and C-reactive protein (CRP) concentrations.

RESULTS

On multivariate analysis of the significant individual covariates, T stage (hazard ratio 2.38, 95% confidence interval 1.06– 5.36, P = 0.037), necrosis (3.73, 1.26–11.05, P = 0.018) and CRP (4.31, 1.20–15.49, P = 0.025) were significant independent predictors of relapse-free survival. On multivariate analysis of significant scoring systems and CRP, only UISS (3.50, 1.66–7.40, P = 0.001), SSIGN (2.83, 1.19–6.72, P = 0.018) and CRP (4.14, 1.16–14.73, P = 0.028) were significant independent predictors of relapse-free survival.

CONCLUSION

Elevated circulating CRP levels appear to be better than other markers of the systemic inflammatory response, and independent of established scoring systems, in predicting relapse-free and cancer-specific survival in patients undergoing potentially curative nephrectomy for renal cancer.

Abbreviations
CSS

cancer-specific survival

RFS

relapse-free survival

UISS

University of California Los Angeles Integrated Staging System

SSIGN

‘Stage Size Grade Necrosis’

CRP

C-reactive protein

ECOG-PS

Eastern Cooperative Oncology Group performance status.

INTRODUCTION

RCC, although the 12th most common cause of death from cancer, is one of the most lethal urological cancers. Each year in the UK, there are ≈3500 new cases and ≈30% of patients present with metastases. Whilst patients with low-grade stage I tumours might have a 5-year survival rate of >90% after potentially curative resection, and those with stage IV disease have a median survival of 9 months, overall, only about half survive for 5 years [1].

As ≈30% of patients who have had a curative nephrectomy will subsequently develop metastatic disease, there has been long-standing interest in identifying those patients most likely to die from RCC. Ideally, a factor or combination of factors would clearly stratify patients who will remain disease-free and are ‘cured’ from those who will ultimately die from their cancer. Currently the TNM stage is the most widely used tool to predict the likely outcome. However, despite revisions, there is considerable ‘overlap’ in survival between the stages. Several scoring systems which include TNM stage and other host factors have been developed to improve the prediction of survival in primary operable RCC, including the University of California Los Angeles Integrated Staging System (UISS) [2], ‘Stage Size Grade Necrosis’ (SSIGN) [3] and Kattan [4] prognostic algorithms.

The systemic inflammatory response, as shown by alterations in circulating concentrations of acute-phase proteins, C-reactive protein (CRP) and albumin, has been shown to be a prognostic factor in various advanced cancers [5–8], including RCC [9], independent of scoring systems [10].

More recently, an elevated CRP level alone or combined with hypoalbuminaemia was shown to be associated with poorer cancer-specific survival (CSS), independent of pathological stage, when tumours were resected with curative intent [11–14]. Interestingly, elevated circulating level of CRP also provided additional prognostic value in patients with low- and intermediate-risk UISS scores who were treated with potentially curative nephrectomy for RCC [15].

However, there are other indicators of the systemic inflammatory response, e.g. leucocytosis [10,16], neutrophilia [10,17], lymphocytopenia [18,19] and thrombocytosis [20–22], which have been reported to have prognostic value in patients with advanced RCC. The aim of the present study was to prospectively compare the prognostic value of markers of the systemic inflammatory response and scoring systems in patients undergoing potentially curative resection for RCC.

PATIENTS AND METHODS

Patients with RCC who, on the basis of surgical findings and CT of the chest and abdomen before surgery, had a potentially curative resection between January 2001 and July 2005 in the North Glasgow NHS Trust, were prospectively included in the study. No patient had T4, nodal or metastatic disease, and all macroscopic tumour was removed at nephrectomy, with subsequent negative surgical margins. Patients were staged pathologically according to the 1997 TNM classification of renal tumours [23]. Tumours were graded according to criteria of Fuhrman et al.[24]. Clinical stage and performance status (Eastern Cooperative Oncology Group, ECOG-PS) were recorded before surgery. Routine laboratory measurements, including haemoglobin levels, white cell, neutrophil, lymphocyte and platelet counts, and albumin and CRP, were made before surgery. The Research Ethics Committee of North Glasgow NHS Trust approved the study.

A blood sample was collected before resection for the routine laboratory analyses; the limit of detection of the assay for CRP was <6 mg/L. The inter- and intra-assay variability of haemoglobin, white cell, neutrophil, lymphocyte and platelet counts, albumin and CRP were <10%. A CRP level of >10 mg/L was considered to indicate the presence of a systemic inflammatory response [5,25].

The UISS score was derived as previously described [2] (Table 1). Briefly, tumour stage, Fuhrman grade and ECOG-PS are combined to stratify patients into low, intermediate or high risk. The SSIGN score was also derived as previously described ([3]Table 2). Patients are assigned scores based on T stage, nodal disease, tumour size, nuclear grade, presence or absence of tumour necrosis and the presence or absence of metastases. Patients with scores of 0–2 were classified as low-risk, 3–5 as intermediate-, and ≥6 as high-risk. The Kattan score was also derived as previously described ([4], Table 2). A nomogram is used to assign points for presenting symptoms, tumour histology, tumour size and T stage, to derive a total score which can be used in the nomogram to estimate a 5-year recurrence-free survival (RFS). Patients with total scores of <70 were classified as low-, 70–100 as intermediate- and >100 as high-risk.

Table 1. 
The UISS scoring system
FeatureUISS
TNM stageNuclear gradeECOG PS
I1 or 20I
I1≥1II
I3 or 4Any 
IIAnyAny 
IIIAny0 
III1≥1 
III2–4≥1III
IV1, 20 
IV3, 40IV
IV1–3≥1 
IV4≥1V
Table 2. 
The SSIGN and Kattan scoring algorithms
FeatureSSIGNFeatureKattan
T stage
pT1a0pT1a 0
pT1b0pT1b 0
pT21pT220
pT3a2pT3a10
pT3b2pT3b30
pT3c2pT3c30
Fuhrman grade Symptoms 
10Incidental 0
20Localized 0
31Systemic40
43  
N Stage Histology: 
pNx0Chromophobe 0
pN00Papillary10
pN12Conventional40
pN22  
Tumour size, cm Tumour size, cm 
<50 210
≥52 420
 630
 840
1050
Histological tumour necrosis
Absent0  
Present2  
M Stage
pM00  
pM14  

values from groups of patients were compared using contingency-table analysis (chi-squared) as appropriate. Survival was analysed using the a Cox proportional-hazards model; deaths up to the end of March 2007 were included in the analysis. For multivariate survival analysis we used a stepwise backward procedure to derive a final model of the variables that had a significant independent relationship with survival. To remove a variable from the model, the corresponding P value had to be >0.10.

RESULTS

The characteristics of the 83 patients with RCC who had potentially curative resection are shown in Table 3; most were men, aged >60 years, had a good PS, had T stage I/II disease and no tumour necrosis. Most had haemoglobin, white cell, neutrophil, lymphocyte and platelet counts, and albumin and CRP levels, in the normal range before surgery.

Table 3.  The baseline characteristics and CSS in 83 patients undergoing potentially curative nephrectomy; univariate analysis
VariablePatients n/nHazard ratio (95% CI)
RFSPCSSP
Age group (≤60/>60 year)40/430.98 (0.37–2.61)0.964 1.03 (0.30–3.55)0.965
Sex (male/female)50/331.71 (0.64–4.59)0.282 1.67 (0.48–5.81)0.418
ECOG PS (0/≥1)69/142.54 (0.88–7.34)0.085 3.73 (1.04–13.37)0.043
T stage (I/II/III/IV)37/11/33/23.17 (1.50–6.73)0.003 4.86 (1.42–16.60)0.012
Tumour size (<5/≥5 cm)33/503.34 (1.21–9.18)0.020 2.44 (0.87–6.86)0.091
Fuhrman grade (1/2/3/4)10/27/26/142.65 (1.43–4.93)0.002 2.38 (1.10–5.15)0.028
Tumour necrosis (absent/present)55/286.19 (2.11–18.15)<0.00118.26 (2.28–146.25)0.006
Haemoglobin (≥12/<12 g/dL)64/192.14 (0.77–5.91)0.142 2.85 (0.80–10.14)0.106
White cells/L (<8.5/8.5–11/>11 × 109)64/11/80.84 (0.35–2.02)0.697 1.01 (0.36–2.84)0.979
Neutrophils/L (<7.5/≥7.5 × 109)73/100.54 (0.07–4.13)0.556 1.05 (0.13–8.38)0.966
Lymphocytes/L (>3.0/1–3/<1.0 × 109) 2/70/111.13 (0.31–4.08)0.852 0.97 (0.16–5.82)0.978
Platelets/L (<400/≥400 × 103)74/94.17 (1.32–13.16)0.01514.22 (3.39–59.73)<0.001
Albumin (≥35/<35 g/L)81/25.02 (0.65–38.55)0.121 7.46 (0.93–60.09)0.059
CRP (≤10/>10 mg/L)49/347.09 (2.02–24.89)0.00215.13 (1.91–120.12)0.010
UISS (I/II/III/IV/V)22/51/5/2/15.07 (2.67–9.61)<0.001 6.21 (2.79–13.82)<0.001
SSIGN (0–2/3–5/≥6)28/27/224.81 (2.07–11.20)<0.001 5.43 (1.66–17.74)0.005
Kattan (<70/70–100/>100)32/27/242.10 (1.10–4.02)0.025 1.53 (0.71–3.30)0.280

The minimum follow-up was 19 months and the median follow-up of the survivors was 38 months. During this period, 16 patients relapsed and 10 died from their cancer; a further four died from intercurrent disease. On univariate survival analysis, PS (P < 0.10), T stage (P < 0.005), tumour size (P < 0.05), grade (P < 0.005), necrosis (P < 0.001), platelet count (P < 0.05), and CRP level (P < 0.01) were significantly associated with RFS. On multivariate analysis of these significant covariates, T stage (hazard ratio 2.38, 95% CI 1.06–5.36, P = 0.037), necrosis (3.73, 1.26–11.05, P = 0.018) and CRP (4.31, 1.20–15.49, P = 0.025) were significant independent predictors of RFS.

The scoring systems UISS (P < 0.001), SSIGN (P < 0.001) and Kattan (P < 0.05) were also significantly associated with RFS. On multivariate analysis of these significant covariates and CRP, the UISS (3.50, 1.66–7.40, P = 0.001), SSIGN (2.83, 1.19–6.72, P = 0.018) and CRP (4.14, 1.16–14.73, P = 0.028) were significant independent predictors of RFS.

On univariate survival analysis, PS (P < 0.05), T stage (P < 0.05), tumour size (P < 0.10), grade (P < 0.05), necrosis (P < 0.01), platelet count (P < 0.001), albumin (P < 0.10) and CRP levels (P ≤ 0.01) were significantly associated with CSS. On multivariate analysis of these significant covariates, T stage (5.90, 1.20–28.92, P = 0.029), necrosis (14.09, 1.73–114.88, P = 0.014) and CRP (19.20, 1.88–195.94, P = 0.013) were significant independent predictors of CSS.

The scoring systems UISS (P < 0.001) and SSIGN (P < 0.01) were also significantly associated with CSS. On multivariate analysis of these significant covariates and CRP, UISS (6.62, 2.07–21.19, P = 0.002), SSIGN (3.49, 1.08–11.24, P = 0.036) and CRP (14.89, 1.67–132.52, P = 0.016) were significant independent predictors of CSS.

The relationship between the presence of a systemic inflammatory response and the clinicopathological characteristics are shown in Table 4. An elevated CRP level was associated with a greater proportion of patients with poorer PS (P < 0.05), advanced tumour stage (P < 0.01), increased tumour size (P < 0.05) and Fuhrman grade (P < 0.05). An elevated CRP level was also associated with increased platelet count (P < 0.10), UISS (P < 0.01), SSIGN (P < 0.05) and Kattan (P < 0.05) scores.

Table 4.  The relationship between the presence of a systemic inflammatory response and clinicopathological characteristics in patients with operable RCC
VariableCRP level, mg/L (n patients) P
≤10 (49)>10 (34)
Age group (≤60/>60 years)23/2617/170.785
Sex (male/female)32/1718/160.261
ECOG-PS (0/≥1)45/424/100.012
T Stage (I/II/III/IV)29/4/16/0 8/7/17/20.003
Tumour size (<5/≥5 cm)25/24 8/260.012
Fuhrman grade (1/2/3/4) 7/19/14/5 3/8/12/90.032
Tumour necrosis (absent/present)35/1420/140.235
Haemoglobin (≥12/<12 g/dL)42/722/120.026
White cells/L (<8.5/8.5–11/>11 × 109)37/8/427/3/40.983
Neutrophils/L (<7.5/≥7.5 × 109)43/630/40.948
Lymphocytes/L (>3.0/1.0–3.0/<1.0 × 109) 2/40/7 0/30/40.855
Platelets/L (<400/≥400 × 103)46/328/60.099
Albumin (≥35/<35 g/L)48/133/10.794
UISS (I/II/III/IV/V)19/27/1/1/0 3/24/4/1/10.002
SSIGN (0–2/3–5/≥6)21/15/9 7/12/130.015
Kattan (<70/70–100/>100)24/15/10 8/12/140.012

DISCUSSION

The present prospective study evaluated the prognostic value of several markers of the systemic inflammatory response, including leucocytosis, neutrophilia, lymphocytopenia, thrombocytosis, hypoalbuminaemia and an elevated CRP levels, together with clinicopathological criteria and current scoring systems in patients with primary operable RCC. Of these markers, only an elevated CRP level had prognostic value for RFS and CSS independent of the clinicopathological criteria and current scoring systems.

We previously reported that an elevated CRP level had additional prognostic value to that of the UISS score [15]; in the present study we confirmed this finding and established that an elevated CRP level had additional prognostic value to the SSIGN and Kattan scoring systems. The individual clinicopathological criteria, tumour stage, tumour necrosis and CRP all had independent prognostic value for RFS or CSS. Whereas tumour stage is incorporated in the both the UISS and SSIGN scoring systems, and tumour necrosis is only incorporated the SSIGN scoring system, CRP is not incorporated in any of the current validated scoring systems. Based on the present results it would be important to prospectively examine whether a new scoring system based on tumour stage, tumour necrosis and CRP had better prognostic value to the UISS and SSIGN scoring systems in patients undergoing potentially curative resection for RCC. This would be consistent with recent work incorporating an elevated CRP level into a prognostic score for patients with metastatic RCC [16,17], and support its inclusion into prognostic scoring systems for patients with primary operable RCC. Moreover, Karakiewicz et al.[26] recently confirmed, in 313 patients with a median follow-up of ≈20 years (compared with the present 83 patients and median 3 years of follow-up) the utility of the incorporating of CRP into the UISS scoring systems.

Of the haematological markers of the systemic inflammatory response, thrombocytosis has been most frequently identified as having independent prognostic value in patients with either advanced [20–22] or operable [27–29] RCC. However, such studies to date in patients with primary operable disease have been predominantly retrospective and not included a comparison with current prognostic scoring systems or CRP levels. On the basis that both thrombocytosis and an elevated CRP level reflect increased interleukin-6 concentrations, Ito et al.[30] reported that although correlated, CRP had better prognostic value than the platelet count in patients who had a radical nephrectomy for RCC. These results are consistent with those of the present study.

The basis of the better prognostic value of an elevated CRP level than the platelet count in patients with RCC is not clear. It might be related to its close direct relationship with circulating interleukin-6 and -10 concentrations [31]. These cytokines probably reflect T-lymphocytic/macrophage activation [32] and are pivotal in promoting a continuing Th2 cytokine response, that might be of primary importance in promoting tumour progression in renal cancer [33,34]. Also, the present results suggest that increased vascular endothelial growth factor levels associated with thrombocytosis [35] are of secondary importance.

Irrespective, an elevated CRP level before surgery might be a useful therapeutic target and warrants further clinical investigation. It remains to be established what treatment might be given to patients at high risk of recurrence after potentially curative resection for RCC. To date, no adjuvant treatment has consistently provided an improvement in survival in such patients. However, the results of the multicentre HYDRA trial, comparing observation to 8 weeks of intensive adjuvant chemo-immunotherapy, and that has recently completed accrual, might provide a stimulus for additional treatment of high-risk patients identified using existing scoring systems and CRP levels.

In summary, an elevated circulating CRP level appears to be better than other markers of the systemic inflammatory response, and independent of established scoring systems, in predicting RFS and CSS in patients undergoing potentially curative nephrectomy for RCC.

CONFLICT OF INTEREST

None declared.

Ancillary