British Journal of Urology (1992), 70, 125-1 34 01992 British Journal of Urology
Renal Carcinoma in Patients Undergoing Nephrectomy : Analysis of Survival and Prognostic Factors A. P. S E N E , L. HUNT, R. F.T. McMAHON and R. N. P. CARROLL Department of Urology and Faculty of Medicine Computational Group, Manchester Royal Infirmary; Department of Pathology, University of Manchester, Manchester
Summary-A series of 155 patients who underwent nephrectomy for renal carcinoma between 1965 and 1985 at Manchester Royal Infirmary were analysed for survival in relationship to presenting features, surgical staging and histopathology. Univariate and multivariate analyses were carried out. Five-year survival estimates for stage 1 disease were 81%, for stage 2 disease 65%, for stage 3 disease 39% and for stage 4 disease 6%.An erythrocyte sedimentation rate (ESR) greater than 30 mm/h was associated with worse survival and a history of hypertension was associated with better survival. Renal vein invasion alone was related t o worse survival. Perinephric fat invasion was also associated with worse survival and this association in the multivariate analysis was more significant than expected, suggesting that the principles of radical surgery should be observed. The presence of granular cells a s opposed to clear cells worsened survival. Patients with papillary tumours had a better survival than those with solid tumours.
Renal carcinoma has an incidence of approximately 4/100,000 (Ritchie and Chisholm, 1983) and accounts for 3% of all malignancies. It was described as early as 1826 and yet its natural history remains unpredictable and poorly understood. In 1883 Grawitz described the condition when he noted the striking gross resemblance of the yellow renal tumour to the adrenal cortex, and he suggested that it might be derived from adrenal rests. It was, however, later confirmed by light and electron microscopy that it arises from renal tubular epithelium. Knowledge of features of the disease which relate accurately to prognosis and survival could be helpful in management. Although systemic therapy has in the past been disappointing, new trials may provide effective adjuvant therapy in the near future. It would be important to be able to select patients who would benefit from such treatment, which may itself have significant side effects.
Accepted for publication 6 January 1992
In this study, 155 patients who underwent nephrectomy for renal carcinoma and had a minimum of 5 years’ potential follow-up were analysed statistically to identify features of prognostic significance. We assessed symptoms and signs, surgical staging and histopathology and attempted to relate them to the outcome.
Patients and Methods The records of 155 patients who underwent nephrectomy for renal carcinoma between 1965 and 1985 were examined. Clinical features at presentation, surgical staging, histopathology and outcome were recorded. When patients died at home or at another hospital, the mode of death was ascertained from the General Practitioners’ records or from the other hospital’s records. Of the 155 patients, 152 had unilateral tumours and the classification of Robson et al. (1969) was used to stage these patients: Stage l-tumour is confined to renal parenchyma and the true capsule is not breached.
125
126 Stage 2-tumour invades perinephric fat but is contained within Gerota’s fascia. Stage 3-tumour involves renal vein, inferior vena cava or lymph nodes. Stage 4-distant metastatic disease or local direct spread is present. This classification is a modification of the first staging system introduced by Flocks and Kadesky in 1958. In addition, the Stage 3 patients were subdivided into Stage 3A (venous invasion alone), Stage 3B (venous invasion and perinephric fat invasion) and Stage 3C (lymph node involvement). Most patients underwent radical nephrectomy (defined as early ligation of the renal artery and vein and removal of the entire perinephric fat without incising Gerota’s fascia). Surgery was accomplished through a flank incision (n = 58) or by an anterior transperitoneal approach (n = 97). The operative records indicate that the procedure was modified according to the surgical findings. Visibly enlarged nodes were removed but formal retroperitoneal lymphadenectomy was not performed. Peri-operative deaths were defined as deaths occurring within 1 month of surgery. Histopathology The tissue was fixed in 10% neutral buffered formalin and was routinely embedded in paraffin. Haematoxylin and eosin stained sections were examined (mean 3 sections, range 1-11) and analysed according to the parameters defined by Fuhrman et al. (1982). Nuclear grade was determined on the highest grade/most malignant area regardless of quantity. Grade 1 tumours contained small uniform nuclei, approximately 10 ym in dimension, with absent or inconspicuous nucleoli (Fig. 1A). Grade 2 tumours had larger nuclei with irregular outlines, approximately 15 ym in diameter, and nucleoli when examined under high power (Fig. 1B). Grade3 tumours had cells which were larger again, with irregular nuclei up to 20 pm in diameter and containing nucleoli which were visible on low power microscopy (Fig. 1C). Grade 4 tumours had all of the characteristics of grade 3 lesions but with the addition of bizarre multilobed nuclei and clumped chromatin (Fig. 1D). Cell type was divided into 3 cateogories-clear cell (Fig. lA), granular (dark) cell (Fig. 1B) or mixed. Where more than 75% of the tumour was of either clear or granular type, the tumour was assigned to that group. If the minority component
BRITISH JOURNAL OF UROLOGY
was greater than 25%, the tumour was classified as of mixed type. The tumour patterns included (i) solid forms composed of monotonous sheets of cells (Figs 1A and B), (ii) papillary where cells were present lining fibrovascular stalks (Fig. 1C) and (iii) spindle where areas of spindle cell or sarcomatoid differentiation were identified (Fig. 1D). The papillary pattern was assigned only to those tumours where more than 50% of the tumour contained definite fibrovascular stalks. Histological demonstration of invasion of the renal vein at the hilar resection margin, renal capsular spread, pelvicaliceal invasion and lymph node metastases formed the basis of the surgical staging used to classify the 152 patients with unilateral disease. Statistical methods Details from each patient were put on to coding forms for statistical analysis using BMDP (Dixon, 1985) and SAS User’s Guide (1985) implemented on the University of Manchester’s mainframe computer. Associations between surgical stage and presenting features and histopathological features were assessed using chi-squared tests, after appropriate recombination of cells with small expected frequencies. Univariate survival analysis was carried out by calculating Kaplan-Meier survival estimates and comparing estimates between subgroups of patients using the Logrank test (Peto et al., 1977). The end-point for analysis was death due to renal carcinoma or post-operative death. Deaths from causes other than renal carcinoma were censored in the analysis. Finally, multivariate survival analysis was carried out using the Cox Proportional Hazards regression model. In this analysis any non-significant variables were deleted in a backward-stepwise manner, using the likelihood-ratio test and a 5% level of significance.
Results Presentingfeatures Of 155 patients who underwent nephrectomy for renal carcinoma there were 102 males and 53 females (ratio approx 2: 1). The age range was 19 to 86 years (median 59). There were 3 patients with bilateral carcinomas and 152 with unilateral disease, 85 on the left and 67 on the right. The weight of the nephrectomy specimens ranged from 90 to 2300 g (median 361). Of 152 patients undergoing nephrectomy for unilateral disease, 69 (45%) had
RENAL CARCINOMA IN PATIENTS UNDERGOING NEPHRECTOMY
127
Fig. 1 (A) Classical renal clear cell carcinoma of solid pattern with nuclear grade 1. (H and E x 400). (B) Solid renal granular cell carcinoma with nuclear grade 2. (H and E x 400). (C) Papillary renal granular cell carcinoma with nuclear grade 3. (H and E x 400). (D) Granular renal carcinoma with a solid/spindle cell pattern, nuclear grade 4 and abundant hyaline globules. (H and E x 400).
Stage 1 disease, 15 (10%) Stage 2, 37 (24%) Stage 3 and 31 (20%) Stage4 disease. Of the 37 patients with Stage 3 disease, 17 were in Stage 3A (venous invasion only), 11 were in Stage 3B (venous and perinephric fat invasion) and 9 were in Stage 3C (nodal involvement). Of the 28 patients with venous invasion 20 had macroscopic and 8 had microscopic involvement. Table 1 illustrates the incidence of clinical features at presentation for all 152 patients with unilateral disease and according to surgical staging. As expected, haematuria, abdominal mass and pain were the commonest features but the classical triad was present in only 12%.Symptomsof malaise, weight loss or anorexia were present in 30%; 15%
of patients were on antihypertensive treatment at presentation and this had to be continued after surgery. The chi-squared tests indicated significant associations between the presence of malaise, weight loss or anorexia, ESR > 30 mm/h or anaemia (Hb < 10 g/dl in females, < 12 g/dl in males) and increasing surgical stage. Polycythaemia (Hb> 18 g/dl in males, > 16.5 g/dl in females) at presentation occurred in only 3 patients. Hypercalcaemia (calcium > 2.6 mmol/l) without bony metastases occurred in only 2 patients. Abnormal liver function tests were present in only 5 patients but the investigation results were located in only 20 patients. Five patients presented with metastases without haematuria, abdominal pain or an abdom-
128
BRITISH JOURNAL OF UROLOGY
Table 1 Features at Presentation of 152 Patients with Unilateral Renal Carcinoma
Haematuria
Classical triad (%)
Pyrexia (%)
Malaise, weight loss or anorexia (%)
Hypertension
7(10) 0 (0) 7(19) 3 (10) 17 (1 1)
Stage
(%)
Mass (%) Pain (%)
1 (n = 69) 2 (n= 15) 3 (n= 37) 4(n=31) All stages (n= 152) Significance
50(73) 8 (53) 30(81) 16(52) 104 (68)
27 (39) 8 (53) 17(46) 19(61) 71 (47)
31 (45) 8 (53) 16(43) 15 (48) 70 (46)
O(0) 6(16) 7(23) 18 (12)
P
Renal Carcinoma in Patients Undergoing Nephrectomy : Analysis of Survival and Prognostic Factors A. P. S E N E , L. HUNT, R. F.T. McMAHON and R. N. P. CARROLL Department of Urology and Faculty of Medicine Computational Group, Manchester Royal Infirmary; Department of Pathology, University of Manchester, Manchester
Summary-A series of 155 patients who underwent nephrectomy for renal carcinoma between 1965 and 1985 at Manchester Royal Infirmary were analysed for survival in relationship to presenting features, surgical staging and histopathology. Univariate and multivariate analyses were carried out. Five-year survival estimates for stage 1 disease were 81%, for stage 2 disease 65%, for stage 3 disease 39% and for stage 4 disease 6%.An erythrocyte sedimentation rate (ESR) greater than 30 mm/h was associated with worse survival and a history of hypertension was associated with better survival. Renal vein invasion alone was related t o worse survival. Perinephric fat invasion was also associated with worse survival and this association in the multivariate analysis was more significant than expected, suggesting that the principles of radical surgery should be observed. The presence of granular cells a s opposed to clear cells worsened survival. Patients with papillary tumours had a better survival than those with solid tumours.
Renal carcinoma has an incidence of approximately 4/100,000 (Ritchie and Chisholm, 1983) and accounts for 3% of all malignancies. It was described as early as 1826 and yet its natural history remains unpredictable and poorly understood. In 1883 Grawitz described the condition when he noted the striking gross resemblance of the yellow renal tumour to the adrenal cortex, and he suggested that it might be derived from adrenal rests. It was, however, later confirmed by light and electron microscopy that it arises from renal tubular epithelium. Knowledge of features of the disease which relate accurately to prognosis and survival could be helpful in management. Although systemic therapy has in the past been disappointing, new trials may provide effective adjuvant therapy in the near future. It would be important to be able to select patients who would benefit from such treatment, which may itself have significant side effects.
Accepted for publication 6 January 1992
In this study, 155 patients who underwent nephrectomy for renal carcinoma and had a minimum of 5 years’ potential follow-up were analysed statistically to identify features of prognostic significance. We assessed symptoms and signs, surgical staging and histopathology and attempted to relate them to the outcome.
Patients and Methods The records of 155 patients who underwent nephrectomy for renal carcinoma between 1965 and 1985 were examined. Clinical features at presentation, surgical staging, histopathology and outcome were recorded. When patients died at home or at another hospital, the mode of death was ascertained from the General Practitioners’ records or from the other hospital’s records. Of the 155 patients, 152 had unilateral tumours and the classification of Robson et al. (1969) was used to stage these patients: Stage l-tumour is confined to renal parenchyma and the true capsule is not breached.
125
126 Stage 2-tumour invades perinephric fat but is contained within Gerota’s fascia. Stage 3-tumour involves renal vein, inferior vena cava or lymph nodes. Stage 4-distant metastatic disease or local direct spread is present. This classification is a modification of the first staging system introduced by Flocks and Kadesky in 1958. In addition, the Stage 3 patients were subdivided into Stage 3A (venous invasion alone), Stage 3B (venous invasion and perinephric fat invasion) and Stage 3C (lymph node involvement). Most patients underwent radical nephrectomy (defined as early ligation of the renal artery and vein and removal of the entire perinephric fat without incising Gerota’s fascia). Surgery was accomplished through a flank incision (n = 58) or by an anterior transperitoneal approach (n = 97). The operative records indicate that the procedure was modified according to the surgical findings. Visibly enlarged nodes were removed but formal retroperitoneal lymphadenectomy was not performed. Peri-operative deaths were defined as deaths occurring within 1 month of surgery. Histopathology The tissue was fixed in 10% neutral buffered formalin and was routinely embedded in paraffin. Haematoxylin and eosin stained sections were examined (mean 3 sections, range 1-11) and analysed according to the parameters defined by Fuhrman et al. (1982). Nuclear grade was determined on the highest grade/most malignant area regardless of quantity. Grade 1 tumours contained small uniform nuclei, approximately 10 ym in dimension, with absent or inconspicuous nucleoli (Fig. 1A). Grade 2 tumours had larger nuclei with irregular outlines, approximately 15 ym in diameter, and nucleoli when examined under high power (Fig. 1B). Grade3 tumours had cells which were larger again, with irregular nuclei up to 20 pm in diameter and containing nucleoli which were visible on low power microscopy (Fig. 1C). Grade 4 tumours had all of the characteristics of grade 3 lesions but with the addition of bizarre multilobed nuclei and clumped chromatin (Fig. 1D). Cell type was divided into 3 cateogories-clear cell (Fig. lA), granular (dark) cell (Fig. 1B) or mixed. Where more than 75% of the tumour was of either clear or granular type, the tumour was assigned to that group. If the minority component
BRITISH JOURNAL OF UROLOGY
was greater than 25%, the tumour was classified as of mixed type. The tumour patterns included (i) solid forms composed of monotonous sheets of cells (Figs 1A and B), (ii) papillary where cells were present lining fibrovascular stalks (Fig. 1C) and (iii) spindle where areas of spindle cell or sarcomatoid differentiation were identified (Fig. 1D). The papillary pattern was assigned only to those tumours where more than 50% of the tumour contained definite fibrovascular stalks. Histological demonstration of invasion of the renal vein at the hilar resection margin, renal capsular spread, pelvicaliceal invasion and lymph node metastases formed the basis of the surgical staging used to classify the 152 patients with unilateral disease. Statistical methods Details from each patient were put on to coding forms for statistical analysis using BMDP (Dixon, 1985) and SAS User’s Guide (1985) implemented on the University of Manchester’s mainframe computer. Associations between surgical stage and presenting features and histopathological features were assessed using chi-squared tests, after appropriate recombination of cells with small expected frequencies. Univariate survival analysis was carried out by calculating Kaplan-Meier survival estimates and comparing estimates between subgroups of patients using the Logrank test (Peto et al., 1977). The end-point for analysis was death due to renal carcinoma or post-operative death. Deaths from causes other than renal carcinoma were censored in the analysis. Finally, multivariate survival analysis was carried out using the Cox Proportional Hazards regression model. In this analysis any non-significant variables were deleted in a backward-stepwise manner, using the likelihood-ratio test and a 5% level of significance.
Results Presentingfeatures Of 155 patients who underwent nephrectomy for renal carcinoma there were 102 males and 53 females (ratio approx 2: 1). The age range was 19 to 86 years (median 59). There were 3 patients with bilateral carcinomas and 152 with unilateral disease, 85 on the left and 67 on the right. The weight of the nephrectomy specimens ranged from 90 to 2300 g (median 361). Of 152 patients undergoing nephrectomy for unilateral disease, 69 (45%) had
RENAL CARCINOMA IN PATIENTS UNDERGOING NEPHRECTOMY
127
Fig. 1 (A) Classical renal clear cell carcinoma of solid pattern with nuclear grade 1. (H and E x 400). (B) Solid renal granular cell carcinoma with nuclear grade 2. (H and E x 400). (C) Papillary renal granular cell carcinoma with nuclear grade 3. (H and E x 400). (D) Granular renal carcinoma with a solid/spindle cell pattern, nuclear grade 4 and abundant hyaline globules. (H and E x 400).
Stage 1 disease, 15 (10%) Stage 2, 37 (24%) Stage 3 and 31 (20%) Stage4 disease. Of the 37 patients with Stage 3 disease, 17 were in Stage 3A (venous invasion only), 11 were in Stage 3B (venous and perinephric fat invasion) and 9 were in Stage 3C (nodal involvement). Of the 28 patients with venous invasion 20 had macroscopic and 8 had microscopic involvement. Table 1 illustrates the incidence of clinical features at presentation for all 152 patients with unilateral disease and according to surgical staging. As expected, haematuria, abdominal mass and pain were the commonest features but the classical triad was present in only 12%.Symptomsof malaise, weight loss or anorexia were present in 30%; 15%
of patients were on antihypertensive treatment at presentation and this had to be continued after surgery. The chi-squared tests indicated significant associations between the presence of malaise, weight loss or anorexia, ESR > 30 mm/h or anaemia (Hb < 10 g/dl in females, < 12 g/dl in males) and increasing surgical stage. Polycythaemia (Hb> 18 g/dl in males, > 16.5 g/dl in females) at presentation occurred in only 3 patients. Hypercalcaemia (calcium > 2.6 mmol/l) without bony metastases occurred in only 2 patients. Abnormal liver function tests were present in only 5 patients but the investigation results were located in only 20 patients. Five patients presented with metastases without haematuria, abdominal pain or an abdom-
128
BRITISH JOURNAL OF UROLOGY
Table 1 Features at Presentation of 152 Patients with Unilateral Renal Carcinoma
Haematuria
Classical triad (%)
Pyrexia (%)
Malaise, weight loss or anorexia (%)
Hypertension
7(10) 0 (0) 7(19) 3 (10) 17 (1 1)
Stage
(%)
Mass (%) Pain (%)
1 (n = 69) 2 (n= 15) 3 (n= 37) 4(n=31) All stages (n= 152) Significance
50(73) 8 (53) 30(81) 16(52) 104 (68)
27 (39) 8 (53) 17(46) 19(61) 71 (47)
31 (45) 8 (53) 16(43) 15 (48) 70 (46)
O(0) 6(16) 7(23) 18 (12)
P
