The prognostic value of two different histopathological scoring systems for adrenocortical carcinomas
H P van’t Sant, N D Bouvy, G Kazemier, H J Bonjer, W C J Hop,1 R A Feelders,2 W W de Herder2 & R R de Krijger3
Departments of Surgery, 1 Epidemiology and Biostatistics, 2Endocrinology and 3Pathology, Josephine Nefkens Institute, Erasmus MC-University Medical Centre, Rotterdam, the Netherlands
Date of submission 24 August 2006 Accepted for publication 14 January 2007
van’t Sant H P, Bouvy N D, Kazemier G, Bonjer H J, Hop W C J, Feelders R A, de Herder W W & de Krijger R R (2007) Histopathology 51, 239-245
The prognostic value of two different histopathological scoring systems for adrenocortical carcinomas
Aims: To compare two different multiparameter histo- pathological scoring indices and determine their prog- nostic value in patients presenting with adrenocortical carcinoma (ACC).
Methods and results: Seventy-nine adrenal cortical tumours were divided into adenomas (n = 17), non- metastatic carcinomas (n = 24) and carcinomas with metastatic disease and/or local recurrence during follow-up (n = 19) or at time of presentation (n = 19). All cases were scored according to the Weiss revisited index (WRI) and the Van Slooten index (VSI). Both scoring indices yielded a significantly different score (P < 0.005) between adenomas and carcinomas. Non-metastasized carcinomas had a lower score with
both indices compared with carcinomas with metasta- ses at the time of presentation (VSI, P = 0.017; WRI, P = 0.019). The VSI also distinguished ACC that had metastasized at any time from those that had not (P = 0.015). Cancer-specific survival in patients with metastasized ACC correlated with the scores for both indices (VSI, P = 0.0078; WRI, P = 0.0025). Time from diagnosis of ACC to development of metastatic disease was correlated with the WRI (P = 0.036, r = - 0.350).
Conclusions: The VSI and the WRI have equal validity in the correct categorization of ACC and adenomas. Furthermore, both indices show a correlation with survival for metastasizing ACC.
Keywords: adrenocortical carcinoma, survival, Van Slooten index, Weiss revisited index Abbreviations: ACC, adrenocortical carcinoma; CT, computed tomography; IGF2, insulin-like growth factor; VSI, Van Slooten index; WRI, Weiss revisited index
Introduction
Adrenocortical carcinoma (ACC) is a rare but aggres- sive tumour with a poor prognosis. The yearly incidence is approximately 1-2 per million people and accounts for 0.2% of deaths due to cancer.1-5 Complete surgical resection of locally confined tumour is the treatment of choice for curative therapy of ACC.6,7 However, many patients present with locally advanced
Address for correspondence: R. R. de Krijger, Department of Pathology, Josephine Nefkens Institute, Erasmus MC, Rotterdam, PO Box 2040, 3000 CA, Rotterdam, the Netherlands. e-mail: r.dekrijger@erasmusmc.nl
or metastatic disease and approximately 20-30% of patients with radically resected tumours relapse.8 Malignancy risk rises with increasing size and weight of the tumour.1,9 A tumour > 50 mm in diameter on computed tomography (CT) should be resected surgi- cally to exclude malignancy.3 Some would propose that tumours weighing >50 g should be regarded as malignant.9 Mitotane (op’DDD) is considered the adju- vant treatment of choice for patients with radically resected ACC. Mitotane is also the therapy of choice in metastasized disease not amenable to surgery. How- ever, treatment should be carefully considered due to the severe toxic side-effects.1 For this reason, it is important to differentiate between ACCs with a high
@ 2007 The Authors. Journal compilation @ 2007 Blackwell Publishing Limited.
probability of rapid progression and those with a more indolent course of disease to avoid unnecessary morbidity.
Differentiation between malignant and benign adrenocortical tumours can in most cases be achieved using multiple histological features. Several multiparameter systems have been published in an attempt to delineate malignancy. In 1984, Weiss et al. proposed a system to diagnose ACC based on nine histological parameters.10 In 2002 Aubert et al. simplified the Weiss system, eliminating four param- eters with limited discrimination, retaining only the most reliable criteria.11 For this new index they coined the term Weiss revisited index (WRI) consist- ing of five histological parameters. In 1985 Van Slooten et al. developed a system based on seven histopathological parameters, each combined with a numerical value, and currently known as the Van Slooten index (VSI).12
We examined a series of 79 adrenocortical tumours, divided into four groups with different clinical status, for two different histopathological scoring indices, i.e. the WRI10 and the VSI,12 with the aim of comparing their ability to distinguish benign from malignant adrenocortical tumours and to determine whether they can predict clinical outcome in ACC patients.
Materials and methods
PATIENTS AND TUMOURS
In the period 1985-2004, 79 patients were selected, representing all patients with an ACC in the central Dutch pathology database for whom clinical data, slides and paraffin blocks were available, as well as a control group of adrenocortical adenomas from the files of the Department of Pathology, Erasmus MC, Rotter- dam. All slides were reviewed by the authors to confirm the adrenocortical nature of the lesions.
Patients were divided into four groups (Figure 1). The first group (A) consisted of 17 patients diagnosed with adrenocortical adenoma according to standard histological criteria (mean follow-up 71 months; range 43-107). The second group (B) consisted of 24 patients diagnosed with ACC who remained free of metastasis and recurrence during follow-up (mean follow-up 52 months; range 4-140). The third group (C) consis- ted of 19 patients diagnosed with ACC and metastasis or tumour recurrence during follow-up (mean follow- up 50 months; range 3-146). The fourth group (D) consisted of 19 patients diagnosed with ACC and metastasis at time of presentation (mean follow-up 36 months; range 2-163). All ACCs were diagnosed on
100
A/B
80
Survival (%)
60
C
40
20
D
0
0
2
4
6
8
10
Years
the basis of the VSI and/or WRI, combined with their clinical course.
Time and localization of metastasis, time of local recurrence and length of survival were recorded. Survival was defined as cancer-specific survival count- ing only deaths due to ACC. Metastasis was defined as distant tumour dissemination confirmed by CT or magnetic resonance imaging, and local invasive growth into adjacent organs was confirmed microscop- ically. Functional status was based on the presence of clinical and/or biochemical adrenocortical hypersecre- tion, i.e. cortisol, androgens or oestrogens. Non-func- tioning tumours were those without evidence of clinical or biochemical adrenocortical hypersecretion.
HISTOLOGY
Two independent observers without prior knowledge of the clinical course or initial tumour diagnosis reviewed the slides of primary adrenocortical tumours. All cases were assessed according to the histopathological criteria of the WRI and VSI.10-12 The WRI consists of the following parameters: mitotic index, clear cytoplasm <25%, abnormal mitoses, necrosis and capsular inva- sion. Each criterion was scored 0 when absent and 1 when present, except for mitotic index and clear cyto- plasm <25%, which are scored 2 when present. Thus, each tumour was scored from 0 to 7 according to the total number of criteria present. The threshold for malig- nancy in this scoring index was a total score of ≥ 3.
The VSI consists of seven parameters, each with a different weight indicated between parentheses:
(i) regressive changes such as necrosis, haemorrhage, fibrosis or calcification (5.7); (ii) structural changes (1.6); (iii) nuclear atypia (2.1); (iv) nuclear hyper- chromasia (2.6); (v) abnormal structure of nucleoli (4.1); (vi) mitotic activity >2/10 high-power fields (x40 objective) (9); (vii) capsule and/or blood vessel invasion (3.3). The total score was based on the addition of parameters, if present, yielding a score between 0 and 28.4 for each individual tumour. The threshold for malignancy using the VSI was > 8.
STATISTICAL ANALYSIS
The Mann-Whitney U-test was used for comparing the distributions of the two scoring indices between the four groups of patients. Survival curves were calculated using the Kaplan-Meier method13 and comparisons between groups of patients were made using a log rank test. Cox regression was used to evaluate both scoring indices simultaneously with respect to survival, adjust- ing for age and gender. Correlation coefficients given are Spearman’s.
Results
CLINICAL DATA
Clinical data for the four groups are shown in Table 1. In groups C (n = 19) and D (n = 19), 22 patients (57.8%) developed pulmonary metastases, 19 (50%) developed liver metastases, five (13.2%) developed bone metastases and 14 (36.8%) developed metastases elsewhere (e.g. brain, lymph nodes, kidney, peritoneal cavity and skin).
CLINICAL FOLLOW-UP AND SURVIVAL
Table 2 shows the outcome of both scoring indices comparing adrenocortical adenomas (group A) with non-metastasized (group B) and metastasized ACC (groups C and D). The P-values indicate the statistical significance of the result of the WRI and VSI for any of the comparisons made. Median survival for group C was 39 months and for group D 21 months, which is displayed in a Kaplan-Meier curve in Figure 1. We found a significant association when comparing the outcome of the histopathological scoring indices with time of survival for all patient groups diagnosed with ACC (groups B, C and D) for both the VSI (P = 0.0025) (Figure 2) and the WRI (P = 0.0078) (Figure 3). We correlated the time to development of metastasis and/or local recurrence with the outcome of both scoring indices for group C and found a P-value of
0.036 for the WRI and 0.443 for the VSI. Matching relationship coefficients were -0.350 and -0.132, respectively. Age and gender were not significant factors on analysis of survival. Cox regression, adjust- ing for age and gender, showed that neither scoring index gave additional significant predictive value for survival regarding the other index (WRI, P = 0.17; VSI, P = 0.23). On analysing tumour diameter, aden- omas were significantly smaller than carcinomas (P < 0.001). No significant correlation was found between cancer-specific survival and tumour diameter (P = 0.607).
Discussion
Distinction of ACC from adrenocortical adenoma on histopathological grounds is fraught with difficulty in at least a proportion of adrenocortical tumours. Fur- thermore, prediction of prognosis with resultant choice of therapy in those tumours labelled ACC is uncertain in many cases. Therefore, in this study we investigated the value of two previously published histopathology scoring indices, the WRI and the VSI, for their ability to distinguish adenomas from ACC and to predict sur- vival. We have shown that the two indices have equal performance in the classification of adrenocortical tumours into the adenoma or carcinoma group. In addition, both indices can be used as a parameter for the prediction of survival of ACC patients.
Over the years, many attempts have been made to distinguish adrenocortical adenomas from ACC, including histopathology using a variety of criteria, immunohistochemistry and, more recently, molecular techniques. With regard to histopathology, two sem- inal studies were performed in the last century by Weiss et al. in 1984 and by van Slooten et al. in 1985, the former having been adapted by Aubert et al.10-12 When comparing the parameters included in the WRI with those in the VSI, the following differences can be seen: the WRI includes the presence of cells with clear cytoplasm constituting ≤ 25% of the tumour and atypical mitoses as parameters, whereas the VSI does not. In contrast, the VSI uses hyperchromasia, nuclear atypia, structural changes and structure of nucleoli as parameters, whereas the WRI does not. Aubert et al. included only tumour capsular invasion in their scoring index, whereas van Slooten et al. also evaluated venous and/or capsular invasion as a single param- eter. As only two parameters correspond in the two scoring indices, both indices, to a large extent, assess different features of adrenocortical tumours. Despite this, we have convincingly shown that the two indices perform equally well with regard to the correct
| Group A (n = 17) | Group B (n = 24) | Group C (n = 19) | Group D (n = 19) | |
|---|---|---|---|---|
| Age | ||||
| Mean | 55.8 | 54.2 | 50.2 | 45.0 |
| Range | 38-91 | 37-76 | 31-78 | 16-74 |
| Sex | ||||
| Male | 2 | 9 | 7 | 8 |
| Female | 15 | 15 | 12 | 11 |
| Functioning Cushing | 15 | 6 | 5 | 5 |
| Virilization | 0 | 1 | 2 | 1 |
| Cushing/ virilization | 0 | 0 | 2 | 5 |
| Feminization | 0 | 0 | 1 | 0 |
| Non-functioning | 2 | 17 | 9 | 8 |
| Tumour location | ||||
| Right | 5 | 9 | 8 | 9 |
| Left | 11 | 14 | 11 | 10 |
| Bilateral | 1 | 1 | 0 | 0 |
| Tumour size (mm) | ||||
| Mean | 47 | 136 | 113 | 130 |
| Range | 30-80 | 70-260 | 50-210 | 60-180 |
| Tumour weight (g) | ||||
| Mean | 61 | 857 | 612 | 664 |
| Range | 11-131 | 68-3672 | 106-1830 | 66-1500 |
| Recurrence | 0 | 0 | 7 | 3 |
| Metastases | 0 | 0 | 17 | 19 |
| Time to metastases (months) | ||||
| Mean | – | – | 29 | 0 |
| Range | – | – | 2-96 | 0 |
| Follow-up (months) | 71 | 52 | 50 | 36 |
| Mean | ||||
| Range | 43-107 | 4-140 | 3-146 | 2-163 |
classification of adrenocortical tumours into the benign or malignant category. Apart from the WRI and the VSI, we also scored all tumours according to the original Weiss index, which, in addition to the five WRI features, includes nuclear pleomorphism, diffuse growth, venous and sinusoidal invasion. In the original Weiss scoring index all criteria have an equal contri-
bution and malignancy is considered when three or more features are present. Since our data between the original Weiss scoring index and the WRI were largely similar, these data are not shown.
Interestingly, in two cases the calculated histopatho- logical score did not reach the cut-off point for malignancy, although metastasis did occur. One case
| Weiss revisited index | Van Slooten index | |||||
|---|---|---|---|---|---|---|
| Group | N | Median | Range | N | Median | Range |
| A | 17 | 0 | 0-2 | 17 | 0 | 0-8.1 |
| B | 24 | 5 | 3-7 | 24 | 21.7 | 12.0-26.8 |
| C | 19 | 5 | 3-7 | 19 | 24.3 | 7.0-28.4 |
| D | 19 | 6 | 2-7 | 19 | 24.3 | 12.7-28.4 |
| Groups | Weiss revisited index, P-value | Van Slooten index, P-value |
|---|---|---|
| A versus C + D | < 0.005 | < 0.005 |
| A versus B | < 0.005 | < 0.005 |
| C versus B | 0.436 | 0.086 |
| D versus B | 0.019 | 0.017 |
| C versus D | 0.206 | 0.647 |
| B versus C + D | 0.066 | 0.015 |
Group A, Adrenocortical adenoma; group B, ACC without metastasis and/ or local recurrence; group C, ACC with metastasis and/or local recurrence during follow-up; group D, ACC with metastasis at time of presentation.
100
80
1
Survival (%)
60
2
40
20
3
0
0
2
4
6
8
10
Years
had a score of 7, which is below the cut-off point for malignancy using the VSI and a score of 3 using the WRI. However, this patient developed pulmonary metastasis after 9 months’ follow-up. The second case
100
80
Survival (%)
1
60
2
3
40
4
20
5
0
0
2
4
6
8
10
Years
had a score of 2, which is below the cut-off point for malignancy using the WRI and a score of 16.8 calculated using the VSI. However, this patient pre- sented with liver metastasis. Therefore, we recommend
application of both scoring indices in case either of these yields an equivocal result, because our data suggest that malignant behaviour should be seriously entertained if either of the histopathological scores is above the threshold for that particular scoring index.
Apart from standard histopathology, the role of immunohistochemistry in the subclassification of adrenocortical neoplasms has been investigated in many studies. For example, expression of p53 protein, MIB1 antigen and proliferating cell nuclear antigen show differences in adenomas and carcinomas, but the sensitivity and specificity are insufficient to be used in clinical practice. 14-20
More recently, complementary DNA microarray techniques have been used to investigate the differ- ences in transcriptional profiles of malignant and benign adrenocortical neoplasms. Slater et al. revealed many genes with more than fourfold differences in expression between ACC, normal tissue and cortical adenoma. They also found significant up-regulation of the insulin-like growth factor (IGF2) gene in ACC compared with adenomas and normal tissue.21 These results are supported by other authors, who have revealed multiple genes displaying significant differen- tial expression between malignant, benign and normal tissue.22,23 Fraipont et al. have identified two clusters of genes (IGF2 cluster and steroidogenesis cluster) whose combined levels of expression are good predictors of malignancy and appear equal to the Weiss index. The combination of these two clusters identifies a sub- population of tumours that has a higher probability of metastatic recurrence (60%).24
Once a diagnosis of ACC has been made, presently still largely on the basis of histopathology, or on clinical grounds if metastasis is present, prognosis is guarded, although large variations in outcome are noted (5-year survival 16-60%).25,26 However, new data emerging from the above-mentioned gene expression profiling studies may lead to new treatment strategies, often based on small molecule inhibitors, necessitating pre- diction of prognosis.27 One major study has investigated the prognostic value of a large number of clinical, histopathological and immunohistochemical cell cycle- related factors. It found six significant predictors of disease-specific survival, including distant metastasis at time of presentation, venous, capsular and adjacent organ invasion, tumour necrosis, mitotic rate, atypical mitoses and mdm-2 overexpression. The authors also conducted a multimolecule analysis including all cell cycle proteins and found one phenotype that could distinguish normal from malignant adrenocortical tis- sue, but this does not surpass conventional microscopy and is therefore of little clinical use.8 Our study was
limited to the analysis of histopathological criteria that were included in the VSI and WRI, and subsequently to the total score of those criteria. Despite this limitation, we are able to draw conclusions comparable to those of the study by Stojadinovic et al. showing that both VSI and WRI are correlated with survival in patients with ACC.8 Also, both VSI and WRI were significantly higher in patients with metastasized tumours at diagnosis than in non-metastasizing ACC. Finally, the VSI, but not the WRI, was significantly higher in ACCs that metastasized at any time than in those that did not. In contrast, the WRI, but not the VSI, was correlated with time from diagnosis of ACC to development of metastatic disease or local recurrence. It should be noted, however, that the differences between VSI and WRI were not very large, with strong trends even in those cases where statistical significance was not reached. We propose that such differences are attributable to variation in those histopathological criteria previously mentioned.
References
1. Allolio B, Hahner S, Weismann D. Fassnacht. Management of adrenocortical carcinoma. Clin. Endocrinol. 2004; 60; 273- 287.
2. Meyer A, Niemann U, Behrend M. Experience with the surgical treatment of adrenal cortical carcinoma. Eur. J. Surg. Oncol. 2004; 30; 444-449.
3. Ng L, Libertino JM. Adrenocortical carcinoma: diagnosis, eval- uation and treatment. J. Urol. 2003; 169; 5-11.
4. Icard P, Charpenay C, Andreassian B et al. Adrenocortical carcinomas: surgical trends and results of a 253-patient series from the French Association of Endocrine Surgeons Study Group. World J. Surg. 2001; 25; 891-897.
5. Schteingart DE. Current perspective in the diagnosis and treatment of adrenocortical carcinoma. Rev. Endocr. Metab. Disord. 2001; 2; 323-333.
6. Luton JP, Cerdas S, Billaud L et al. Clinical features of adreno- cortical carcinoma, prognostic factors, and the effect of mitotane therapy. N. Engl. J. Med. 1990; 322; 1195-1201.
7. Schulick RD, Brennan MF. Adrenocortical carcinoma. World J. Urol. 1999; 17; 26-34.
8. Stojadinovic A, Ghossein RA, Hoos A et al. Adrenocortical carcinoma: clinical, morphologic, and molecular characteriza- tion. J. Clin. Oncol. 2002; 20; 941-950.
9. Fajardo R, Montalvo J, Velazquez D et al. Correlation between radiologic and pathologic dimensions of adrenal masses. World J. Surg. 2004; 28; 494-497.
10. Weiss LM. Comparative histologic study of 43 metastasizing and nonmetastasizing adrenocortical tumours. Am. J. Surg. Pathol. 1984; 8; 163-169.
11. Aubert S, Wacrenier A, Leroy X et al. Weiss system revisited: a clinicopathologic and immunohistochemical study of 49 adreno- cortical tumours. Am. J. Surg. Pathol. 2002; 26; 1612-1619.
12. van Slooten H, Schaberg A, Smeenk D, Moolenaar JM. Morpho- logic characteristics of benign and malignant adrenocortical tumours. Cancer 1985; 55; 766-773.
13. Kaplan EL. Nonparametric estimation from incomplete observa- tions. J. Am. Stat. Assoc. 1958; 53; 457-481.
14. Martins AC, Cologna AJ, Tucci S Jr, Suaid HJ, Falconi RA. Clinical features and immunoexpression of p53, MIB-1 and proliferating cell nuclear antigen in adrenal neoplasms. J. Urol. 2005; 173; 2138-2142.
15. Wachenfeld C, Beuschlein F, Zwermann O et al. Discerning malignancy in adrenocortical tumours: are molecular markers useful? Eur. J. Endocrinol. 2001; 145; 335-341.
16. McNicol AM, Struthers AJ, Nolan CE, Hermans J, Haak HR. Proliferation in adrenocortical tumours: correlation with clinical outcome and p53 status. Endocr. Pathol. 1997; 8; 29-36.
17. Terzolo M, Boccuzzi A, Bovio S et al. Immunohistochemical assessment of Ki-67 in the differential diagnosis of adrenocortical tumours. Urology 2001; 57; 176-182.
18. Sasano H, Suzuki T, Moriya T. Discerning malignancy in resected adrenocortical neoplasms. Endocr. Pathol. 2001; 12; 397-406.
19. Barzon L, Chilosi M, Fallo F et al. Molecular analysis of CDKNIC and TP53 in sporadic adrenal tumours. Eur. J. Endocrinol. 2001; 145; 207-212.
20. McNical AM, Nolan CE, Struthers AJ, Farquharson MA, Hermans J, Haak HR. Expression of p53 in adrenocortical tumours: clinicopathological correlations. J. Pathol. 1997; 181; 146-152.
21. Slater EP, Diehl SM, Langer P et al. Analysis by cDNA micro- arrays of gene expression patterns of human adrenocortical tumours. Eur. J. Endocrinol. 2006; 154; 587-598.
22. Giordano TJ, Thomas DG, Kuick R. et al. Distinct transcriptional profiles of adrenocortical tumours uncovered by DNA microarray analysis. Am. J. Pathol. 2003; 162; 521-531.
23. Velazquez-Fernandez D, Laurell C, Geli J et al. Expression profiling of adrenocortical neoplasms suggests a molecular signature of malignancy. Surgery 2005; 138; 1087-1094.
24. de Fraipont F, El Atifi M, Cherradi N et al. Gene expression profiling of human adrenocortical tumours using complementary deoxyribonucleic acid microarrays identifies several candidate genes as markers of malignancy. J. Clin. Endocrinol. Metab. 2005; 90; 1819-1829.
25. Thompson GB, Young WF Jr. Adrenal incidentaloma. Curr. Opin. Oncol. 2003; 15; 84-90.
26. Vassilopoulou-Sellin R, Schultz PN. Adrenocortical carcinoma. Clinical outcome at the end of the 20th century. Cancer 2001; 92; 1113-1121.
27. Kirschner LS. Review: Emerging treatment strategies for adreno- cortical carcinoma: a new hope. J. Clin. Endocrinol. Metab. 2006; 91; 14-21.