ELSEVIER
Helsinki score-a novel model for prediction of metastases in adrenocortical carcinomas
Mirkka Pennanen MDa,b,*, Ilkka Heiskanen MD, PhDC, Timo Sane MD, PhDd, Satu Remes ScLa, Harri Mustonen PhDC, Caj Haglund MD, PhDc, e, 1, Johanna Arola MD, PhDa,b,1
aDepartment of Pathology, Haartman Institute, University of Helsinki, 00014 Helsinki, Finland
Department of Pathology, HUSLAB, 00029 Helsinki, Finland
“Department of Surgery, Helsinki University Central Hospital, 00029 Helsinki, Finland
ªDivision of Endocrinology, Department of Medicine, Helsinki University Central Hospital, 00029 Helsinki, Finland
·Research Programs Unit, Translational Cancer Biology, University of Helsinki, 00014 Helsinki, Finland
Received 14 August 2014; revised 6 November 2014; accepted 23 November 2014
Keywords:
Adrenocortical tumor; Histopathology; Weiss score; Proliferation; Carcinoma
Summary Histopathologic diagnosis of adrenocortical tumors is based on adverse features that indicate malignant potential. Proliferation index has served as a supplemental tool in assessing the malignant potential of adrenocortical tumors. None of the current histologic classification systems can sufficiently accurately predict tumors’ metastatic potential. We studied 177 consecutive adult patients with primary adrenocortical tumors operated on at Helsinki University Central Hospital between 1990 and 2003, all patients with a minimum follow-up of 5 years. We determined for each tumor the Weiss score and the Weiss revisited score by Aubert. Proliferation index was measured by computer-assisted image analysis. Each of the 9 Weiss criteria and the proliferation index were then used to establish a scoring system to predict the metastatic potential of adrenocortical tumors. Use of stepwise regression analysis led us to propose a calculation: 3 × mitotic rate (>5/50 high-power fields) + 5 x presence of necrosis + proliferation index in the most proliferative area of the tumor. Using a cutoff value of 8.5, the new scoring system was able to diagnose metastatic adrenocortical carcinoma with 100% sensitivity (confidence interval [CI], 76.8%-100%) and 99.4% specificity (CI, 96.6%-100%). The corresponding sensitivity of the Weiss system was 100% (CI, 76.8%-100%), and specificity, 90.2% (CI, 84.6%- 94.3%), with sensitivity of the Weiss revisited system at 100% (CI, 76.8%-100%) and specificity at 96.9% (CI, 93.0%-99.0%). The new Helsinki score thus was accurate in predicting the metastatic potential of adrenocortical tumors.
C 2014 Elsevier Inc. All rights reserved.
# Competing interests: The authors declare no conflict of interest.
** Funding/Support: This work was supported by the grants from the Finnish Cancer Foundation, Finska Läkaresällskapet, Helsinki University Hospital Research Fund, Medicinska Understödsföreningen Liv och Hälsa, and the Sigrid Jusélius Foundation.
* Corresponding author. Department of Pathology/Haartman Institute, P.O. Box 21, FIN-00014 University of Helsinki and HUSLAB, P.O. Box 400, 00029 HUS, Finland.
E-mail address: mirkka.pennanen@hus.fi (M. Pennanen).
1 Equal last authorship.
1. Introduction
The histopathologic diagnosis of adrenocortical tumors is based on adverse microscopic features indicating malignant potential. Among several histologic scoring systems and algorithms proposed to differentiate adrenocortical carcino- mas from adenomas [1-6], the Weiss scoring system is used the most widely. It includes 9 histopathologic criteria: the presence of any 3 of which indicates malignant potential [5,6]. Although very sensitive in diagnosing an adrenocor- tical carcinoma, the Weiss system suffers from interobserver variation. In 2002, Aubert et al [1] proposed a statistically modified scoring system based on the 5 most objective Weiss criteria. Immunohistochemical measurement of cell prolif- eration has also been useful in predicting malignancy of adrenocortical tumors [1,7-10], but it has not been included in any implemented scoring system.
Some low-score tumors, scored by either the Weiss or revisited Weiss system, do not metastasize. None of the currently available histologic or immunohistochemical (IHC) methods can accurately and reliably predict the metastatic potential of an adrenocortical tumor.
Here, we determined the diagnostic performance of the Weiss system, the revisited Weiss system by Aubert, and the proliferation index (PI), determined using computer-assisted image analysis, in a large consecutive material of benign and metastatic adrenocortical tumors. Our aim was to find an objective and accurate diagnostic system to predict the metastases of adrenocortical tumors.
2. Materials and methods
2.1. Clinical data and tumor material
We identified all adult patients who underwent surgery for a primary adrenocortical tumor at the Department of Surgery, Helsinki University Central Hospital, between 1990 and 2003. The series comprised 195 patients with 197 tumors. We excluded 20 patients who had died of causes other than the adrenocortical tumor in less than 5 years after resection of the tumor to ensure a minimum follow-up time of 5 years for each patient. The final study cohort thus included 175 patients with 177 tumors, as 2 patients had 2 separate tumors. Tumor specimens were stored in the archives of the Department of Pathology, University of Helsinki. Clinical data came from patient records. The criterion for clinically malignant disease was metastasis. Survival data and cause of death came from the Population Register Center and Statistics Finland. The study was approved by the Ethics Committee of the Helsinki University Central Hospital and the National Authority for Medicolegal Affairs.
2.2. Histopathology
Tumor tissues were fixed in 10% buffered neutral formalin and embedded in paraffin. The histopathologic diagnosis according to the Weiss criteria [5,6] for each tumor was reviewed from hematoxylin and eosin (HE)-stained sections by 2 pathologists (M. P. and J. A.). The Weiss system includes the following 9 histopathologic criteria: (1) high nuclear grade, (2) mitotic rate greater than 5 per 50 high-power fields, (3) presence of atypical mitoses, (4) clear lipid-rich cells comprising less than 25% of the tumor, (5) diffuse architecture, (6) necrosis, (7) invasion of venous structures, (8) invasion of sinusoidal structures, and (9) invasion of the capsule. Each tumor thus received a numeric score ranging from 0 to 9. The mitotic rate was assessed choosing 10 high-power fields in the area of greatest mitotic activity for 5 slides. If less than 5 slides were available for a given case, we evaluated a greater number of fields to have a total of 50 high-power fields for evaluation. Necrosis was regarded as present when occurring in at least confluent nests of cells. The Weiss revisited score (WRS) by Aubert was determined for each tumor as follows: 2 x cytoplasm + 2 × mitotic rate + abnormal mitoses + necrosis + capsular invasion. Thus each tumor received a score from 0 to 7.
2.3. Tissue microarray construction
Tissue microarray blocks were used for IHC stainings. Representative areas of each tumor were chosen from HE- stained slides. From histologically benign (Weiss 0-2) tumors, three 1-mm cores, and from histologically malignant (Weiss score [WS] 3-9) tumors, 6 cores were obtained with a semiautomatic tissue microarray instrument (Beecher Instru- ments, Silver Spring, MD).
2.4. Immunohistochemistry
Immunohistochemical stainings were done with mouse monoclonal Ki-67 antibody (clone MIB-1; Dako, Glostrup, Denmark) diluted 1:200 with the polymer detection kit EnVision (Dako) in a LabVision Autostainer (Thermo Scientific, Fremont, CA). The slides were counterstained with Mayer’s hematoxylin (Lillie’s Modification) (Dako) and mounted with Mountex (Histolab, Järfälla, Sweden).
2.5. Scoring of PI with ImmunoRatio
Assessment of PI was conducted with an Internet-based, free image analysis software called ImmunoRatio (Jorma Isola & Vilppu Tuominen, Institute of Biomedical Technol- ogy, University of Tampere, Finland) that can be used in quantitative assessment of any nuclei markers [11]. Image capture was performed with a Nikon Eclipse 80i light microscope (objective ×40) connected to a Digital Sight DS- 5 M (Nikon) digital camera and NIS-Elements F 3.0 image
capture software. From every core biopsy, 1 digital image (JPEG format, Resolution 1280 x 960) was captured from the most proliferative area of the tumor covering 50% of the core biopsy. With ImmunoRatio’s advanced mode setting, threshold values for hematoxylin (0) and 3,3’-diaminoben- zidine (-10) were adjusted, with no correction equation. A blank field image was taken to balance uneven illuminations in the final digital images. Image analysis settings were kept identical after light exposure (manual exposure, 10 millisec- onds; gain 1×), and thresholds were considered good. Of all images available for each tumor, the highest PI was recorded.
2.6. Statistics
Results are given as mean, median and range, or number of patients and proportion of patients. The Mann-Whitney U test was used to analyze differences between patient groups in continuous and ordinal variables. Fisher exact test analyzed differences in dichotomous variables. Clopper-Pearson 95% confidence intervals (CIs) were calculated for sensitivity and specificity values. Differences in specificities between differ- ent scores were tested with the nonparametric paired McNemar test. Spearman p correlation coefficient was calculated between different scores and metastatic status. Survival rates were estimated by the Kaplan-Meier method. The log-rank test compared survival curves. Logistic regression and receiver operating characteristic (ROC) analyses served to model the new score developed here. Logistic regression with backward stepping was used to find the most influential variables, and ROC analysis, to find the optimal cutoff value by maximizing Youden index. P <. 05 was considered statistically significant. Two-tailed tests were used. Statistical calculations were performed with SPSS version 20 (SPSS, Chicago, IL) or with Statexact version 4 (Cambridge, MA).
3. Results
3.1. Clinical data
For clinical characteristics of the tumors, see Table 1. Follow-up time for all the patients ranged from 0.3 to 21.8 years (median, 10.8). Patients with local disease had follow- up times from 5.0 to 21.6 years (median, 11.2). The 14 patients with metastatic disease all succumbed to their disease (Table 2). Survival among these 14 patients ranged from 0.3 to 6.8 years (median, 2.0).
3.2. Weiss and Weiss revisited scorings
Of the total 177 tumors, 147 had WS 0 to 2, 158 tumors had WRS 0 to 2, 30 tumors had WS 3 to 9, and 19 tumors had WRS 3 to 7. Table 2 shows the histopathologic and clinical features of all tumors diagnosed as malignant according to the Weiss
| Table 1 Clinical and pathologic characteristics, local versus metastatic tumors | ||||
|---|---|---|---|---|
| All tumors | Local tumors | Metastatic P tumors | ||
| No. of tumors | 177 | 163 | 14 | |
| Age (y) | ||||
| Range | 24-82 | 24-82 | 32-76 | .9582 |
| Mean | 53.4 | 53.4 | 53.5 | |
| Median | 54 | 53 | 55.5 | |
| Sex ª | ||||
| Male | 58 | 52 | 6 | .5555 |
| Female | 117 | 109 | 8 | |
| Side | ||||
| Right | 93 | 85 | 8 | |
| Left | 84 | 78 | 6 | .7859 |
| Size (cm) | ||||
| Range | 0.5-28 | 0.5-28 | 4-22 | |
| Mean | 3.4 | 2.7 | 12.1 | <. 001 |
| Median | 2 | 2 | 12.5 | |
| Hormonal secretion | ||||
| Aldosterone | ||||
| n | 92 | 90 | 2 | |
| % | 52.0 | 55.2 | 14.3 | .0041 |
| Cortisol | ||||
| n | 68 | 61 | 7 | |
| % | 38.4 | 37.4 | 50.0 | .3977 |
| Androgens | ||||
| n | 11 | 7 | 4 | |
| % | 6.2 | 4.3 | 28.6 | .0060 |
| Inactive | ||||
| n | 20 | 15 | 5 | |
| % | 11.3 | 9.2 | 35.7 | .0114 |
| PI (%) b | ||||
| Range | 0-54.2 | 0-8.0 | 0.9-54.2 | <. 001 |
| Mean | 2.3 | 1.0 | 16.7 | |
| Median | 0.8 | 0.7 | 7.0 | |
| WS | ||||
| Range | 0-9 | 0-9 | 5-9 | <. 001 |
| Mean | 1.7 | 1.2 | 7.8 | |
| Median | 1 | 1 | 8 | |
| WRS | ||||
| Range | 0-7 | 0-7 | 4-7 | <. 001 |
| Mean | 0.9 | 0.5 | 6.3 | |
| Median | 0 | 0 | 7 | |
| Helsinki score | ||||
| Range | 0-62.2 | 0-11.3 | 8.9-62.2 | <. 001 |
| Mean | 3.1 | 1.2 | 24.2 | |
| Median | 0.8 | 0.7 | 15.0 | |
| Abbreviations: PI, proliferation index; WS, Weiss score; WRS, Weiss revisited score. | ||||
| a One man and 1 woman each had 2 separate tumors. | ||||
| b PI was determined for 174 tumors, 160 local and 14 metastatic. | ||||
system. Of the 11 tumors with WS 3, 10 were benign according to WRS (2), and 1, malignant (WRS 3). None of these tumors metastasized. Only 1 WS 4 tumor (patient 12 in Table 2) was malignant also according to WRS (4) and did not metastasize. Among the 18 tumors with WS 5 to 9, a discrepancy between WS and WRS concerned only 1: a WS 5 tumor scored WRS 2,
| Table 2 | Comparison of Helsinki score, WS, and WRS of all tumors histologically malignant according to their WS | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Helsinki score | Proliferation index (%) | Mitotic rate 0/3 | Necrosis 0/5 | WS | WRS | Tumor size (cm) | Local/metastatic disease | Survival (follow-up) | |
| Helsinki score <8.5 | |||||||||
| P 1 | 0.3 | 0.3 | 0 | 0 | 3 | 2 ª | 3.5 | L | Alive (15.3 y) |
| P 2 | 0.8 | 0.8 | 0 | 0 | 3 | 2ª | 3.5 | L | Alive (19.3 y) |
| P 3 | 0.9 | 0.9 | 0 | 0 | 3 | 2 ª | 3.0 | L | Alive (16.3 y) |
| P 4 | 1.0 | 1.0 | 0 | 0 | 3 | 2 ª | 3.0 | L | Alive (8.3 y) |
| P 5 | 1.1 | 1.1 | 0 | 0 | 3 | 2 ª | 2.5 | L | Alive (12.8 y) |
| P 6 | 1.4 | 1.4 | 0 | 0 | 3 | 2ª | 2.5 | L | (15.6 y)b |
| P 7 | 1.9 | 1.9 | 0 | 0 | 3 | 2 ª | 7.0 | L | Alive (10.5 y) |
| P 8 | 2.4 | 2.4 | 0 | 0 | 3 | 2 ª | 2.0 | L | Alive (20.3 y) |
| P 9 | 3.1 | 3.1 | 0 | 0 | 5 | 2ª | 5.5 | L | Alive (7.9 y) |
| P 10 | 4.1 | 4.1 | 0 | 0 | 3 | 3 | 1.0 | L | Alive (8.1 y) |
| P 11 | 4.8 | 4.8 | 0 | 0 | 3 | 2 ª | 2.0 | L | Alive (11.2 y) |
| P 12 | 5.5 | 2.5 | 3 | 0 | 4 | 4 | 7.0 | L | Alive (12.0 y) |
| P 13 | 8.0 | 8.0 | 0 | 0 | 3 | 2ª | 1.0 | L | Alive (8.9 y) |
| P 14 | 8.0 | 0 | 3 | 5 | 9 | 7 | 13.0 | L | Alive (11.5 y) |
| P 15 | 8.3 | 0.3 | 3 | 5 | 8 | 6 | 12.5 | L | Alive (17.5 y) |
| Helsinki score >8.5 | |||||||||
| P 16 | 8.9 | 0.9 | 3 | 5 | 9 | 7 | 9.0 | M | 1.0 y |
| P 17 | 9.1 | 4.1 | 0 | 5 | 5 | 4 | 12.0 | M | 1.2 y |
| P 18 | 9.3 | 6.3 | 3 | 0 | 7 | 5 | 4.0 | M | 3.1 y |
| P 19 | 9.3 | 1.3 | 3 | 5 | 9 | 7 | 7.5 | M | 6.6 y |
| P 20 | 9.6 | 1.6 | 3 | 5 | 6 | 5 | 8.0 | M | 5.3 y |
| P 21 | 11.2 | 3.2 | 3 | 5 | 9 | 7 | 22.0 | M | 4.7 y |
| P 22 | 11.3 | 3.3 | 3 | 5 | 9 | 7 | 28.0 | L | Alive (7.8 y) |
| P 23 | 14.3 | 6.3 | 3 | 5 | 7 | 6 | 4.5 | M | 5.4 y |
| P 24 | 15.7 | 7.7 | 3 | 5 | 9 | 7 | 16.0 | M | 1.4 y |
| P 25 | 16.2 | 8.2 | 3 | 5 | 6 | 6 | 17.0 | M | 6.8 y |
| P 26 | 32.0 | 24 | 3 | 5 | 8 | 6 | 13.0 | M | 2.0 y |
| P 27 | 33.8 | 25.8 | 3 | 5 | 9 | 7 | 15.0 | M | 0.3 y |
| P 28 | 47.3 | 39.3 | 3 | 5 | 8 | 7 | 12.0 | M | 0.6 y |
| P 29 | 59.5 | 51.5 | 3 | 5 | 9 | 7 | 15.0 | M | 2.1 y |
| P 30 | 62.2 | 54.2 | 3 | 5 | 8 | 7 | 14.0 | M | 1.2 y |
| Abbreviations: WS, Weiss score; WRS, Weiss revisited score; L, local; M, metastatic. a Histologically benign tumors according to WRS. b Death due to other cause than adrenocortical carcinoma. | |||||||||
and this tumor did not metastasize. Altogether, 19 tumors were diagnosed as histologically malignant according to both WS and WRS; 14 of these tumors metastasized.
In diagnosing metastatic adrenocortical carcinoma, the sensitivity of both the Weiss and Weiss revisited system systems was 100% (CI, 76.8%-100%); specificity, 90.2% (CI, 84.6%-94.3%) and 96.9% (CI, 93.0%-99.0%), respec- tively. The Weiss and Weiss revisited systems showed a strong correlation, r = 0.768 (P < . 001).
3.3. Proliferation
The PI of the tumors was assessed from Ki-67 stainings by ImmunoRatio image analysis software and was deter- mined for 174 tumors. In 3 local tumors, the PI could not be determined for technical reasons; 2 tumors displayed false positivity in the pseudocolor images due to pigmentation, and for 1 tumor, after IHC staining, all 3 cores were missing.
The PI of 160 local tumors ranged from 0% to 8.0% (average, 1.0%; median, 0.7%; interquartile range, 1.0). In 14 metastatic tumors, the PI ranged from 0.9% to 54.2% (average, 16.7%; median, 7.0%; interquartile range, 22.5%). The difference in PI between local and metastatic tumors was significant (P <. 01). A cutoff value of 4% was considered optimal. Using this cutoff, PI diagnosed metastatic carcinoma with 71.4% sensitivity (CI, 42.0%-90.4%) and 96.9% specificity (CI, 92.4%-98.8%).
3.4. The Helsinki score
We used logistic regression with backward stepping to find the most influential variables of the nine Weiss criteria plus PI in our material. This led us to a scoring model, which we called the Helsinki score: 3.28 x mitotic rate + 4.42 x necrosis + 1.02, which was simplified to 3 x mitotic rate + 5 x necrosis + PI (Fig. 1). ROC analysis was used to determine the optimal cutoff value. The Helsinki score, with a cutoff
Helsinki score
A
B
C
D
3 x mitotic rate
5 x necrosis
PI %
value of 8.5, diagnosed a metastatic adrenocortical carcino- ma with 100% sensitivity (CI, 76.8%-100%) and 99.4% specificity (CI, 96.6%-100%). Only 1 tumor earned a false- positive result by the new score; this tumor (P 22 in Table 2) was considered malignant with high scores according to all scoring systems: WS 9, WRS 7, and Helsinki score 11.3. The tumor was large (28 cm) and locally invasive but was excised radically. This patient received mitotane treatment after the operation and was alive at the end of the follow-up period, 8 years after surgery. Four tumors were malignant according to both the Weiss system and Weiss revisited system, but not according to the Helsinki scoring system (P 10, 12, 14, and 15 in Table 2). None of these tumors have metastasized during follow-up ranging from 8.1 to 17.5 years.
The Helsinki score also correlated with survival. We categorized the tumors according to their Helsinki score into 3 groups: (1) benign tumors with Helsinki score 0 to 8.5, (2) malignant tumors with Helsinki score 8.5 to 17, and (3) malignant tumors with Helsinki score greater than 17. For survival curves, see Fig. 2. Differences in survival among the 3 groups were significant: group 1 versus group 2, P <. 01; group 1 versus group 3, P < . 01; group 2 versus group 3, P = . 010.
4. Discussion
Here, we introduce a novel Helsinki score that can diagnose metastatic adrenocortical carcinoma with 100% sensitivity and 99.4% specificity. The new score uses 2 of the Weiss criteria, mitotic index, and necrosis, plus PI determined as a continuous parameter, measured by computer-assisted image analysis.
The Helsinki score predicts metastases in adrenocortical tumors, the main cause of a fatal outcome. In many studies, the diagnostic performance of a histologic classification system
has been compared to that of the Weiss system or other scoring systems [1,12,13]. In studies assessing the diagnostic performance of proliferation, malignancy has often been defined by the histologic criteria of Weiss [8,9,14]. None of the currently used classifications can, however, predict metastases quite accurately, thus challenging clinical decision making.
Some of the Weiss criteria, such as sinusoidal invasion and diffuse growth, are subjective with large interobserver variation [1,15]. To make the diagnostic procedure of adrenocortical tumors more simple and objective, Aubert et al [1] introduced a modification, based on the 5 most objective Weiss criteria. Their revised Weiss system correlated signifi- cantly with the original Weiss system (r= 0.98). The result was validated by van’t Sant et al [12]. However, the systems have a
100%
-0-8.5
Cumulative Survival (%)
80%
— >8.5-17
→17
60%
40%
20%
0%
0
5
10
15
20
Time (Years)
| Time (years): | 0 | 1 | 2 | 5 | 7 | |
|---|---|---|---|---|---|---|
| Cumulative survival | >8.5 - 17 | 90 % | 70 % | 50 % | 10 % | |
| 95% CI | 47 - 99 % | 33 - 89 % | 18 -75 % | 1 -36 % | ||
| >17 | 60 % | 20 % | ||||
| 13 - 88 % | 1-58 % | |||||
| Number at risk | 0 - 8.5 | 159 | 159 | 159 | 158 | 154 |
| >8.5- 17 | 10 | 9 | 7 | 5 | ||
| >17 | 5 | 3 | 1 | 0 | 0 |
In group 0-8.5, number at risk at 10, 15 and 20 years were 99, 46 and 9, respectively.
Fig. 2 Survival according to Helsinki score.
common problem: although sensitive in diagnosing adrenocor- tical carcinomas, a number of borderline cases cannot be reliably classified, and specificity could be improved. In our 177 cases, the sensitivity of both the Weiss and Weiss revisited systems in diagnosing adrenocortical carcinoma was 100%. The specificity of the Weiss system was 90.2%, and that of the Weiss revisited system, 96.9%. These scoring systems showed a high correlation (r = 0.768).
Our suggested Helsinki score is simple, objective, sensitive, and specific in diagnosing adrenocortical tumors with metastatic potential. It includes only 2 of the histologic parameters of the Weiss system: mitotic rate greater than 5/50 high-power fields and presence of necrosis. These criteria are reasonably objective to assess [1]. In addition, the Helsinki score includes an accurate PI assessed from Ki-67 stainings by computer-assisted image analysis software ImmunoRatio. The new score was able to diagnose metastatic adrenocortical carcinoma with 100% sensitivity and with a higher (99.4%) specificity than the Weiss and Weiss revisited systems. Of 15 tumors diagnosed as carcinomas according to the Weiss system, all had a Helsinki score below the cutoff value of 8.5, and none of these tumors has metastasized during the follow-up of 7.9 to 20.3 years.
In adrenocortical tumors, proliferative activity correlates with malignancy, with histologic features of malignancy as well as with clinical outcome [1,7-10]. Because of too low sensitivity, however, PI has served only as an additional tool in predicting the behavior of adrenocortical tumors; it has not been included in any implemented diagnostic protocol. Using a cutoff value of 4%, the sensitivity of the PI in diagnosing metastatic carcinoma was 71.4%, and the specificity, 96.9% in our study. Because PI showed high specificity, whereas histologic scoring systems are sensitive but have lower specificity, we hypothesized that PI could add specificity to a scoring system.
In adrenocortical tumors, mitotic rate is known to show a strong correlation with malignancy and with patient survival [2,6,16,17]. Although mitotic frequency and PI both describe the proliferative capacity of the tumor, they do not show a direct correlation. In fact, some metastatic adrenocortical tumors have a surprisingly low PI. Indeed, in the Helsinki score, PI and mitotic frequency supplemented each other.
A major advantage in our study is large cohort size and comprehensive follow-up data on all patients. A limitation is the small number of metastatic tumors (14 tumors) because adrenocortical tumors are rare. The method used to develop the score also adapts to the data, and no attempts were possible to verify the validity of the model. The model, therefore, presents only the current data and can be regarded only as a suggestion for further research.
The use of immunohistochemistry can cause interlabora- tory variations due to differences in staining protocols and in assessment methods. Conventional manual assessment of the PI has major interobserver and intraobserver variations [18]. In this study, however, the computer-assisted image analysis software ImmunoRatio was used to produce accurate PI measurements.
After validating the diagnostic performance of the Weiss and Weiss revisited systems, their comparison with our new scoring system called the Helsinki score leads us to introduce it for accurate prediction of the metastatic potential of adrenocortical tumors.
Acknowledgments
The authors thank Eija Heiliö, Elina Aspiala, and Päivi Peltomäki for technical support.
References
[1] Aubert S, Wacrenier A, Leroy X, et al. Weiss system revisited: a clinicopathologic and immunohistochemical study of 49 adrenocortical tumors. Am J Surg Pathol 2002;26:1612-9.
[2] Blanes A, Diaz-Cano SJ. Histologic criteria for adrenocortical proliferative lesions: value of mitotic figure variability. Am J Clin Pathol 2007;127:398-408.
[3] Hough AJ, Hollifield JW, Page DL, Hartmann WH. Prognostic factors in adrenal cortical tumors. A mathematical analysis of clinical and morphologic data. Am J Clin Pathol 1979;72:390-9.
[4] van Slooten H, Schaberg A, Smeenk D, Moolenaar AJ. Morphologic characteristics of benign and malignant adrenocortical tumors. Cancer 1985;55:766-73.
[5] Weiss LM. Comparative histologic study of 43 metastasizing and nonmetastasizing adrenocortical tumors. Am J Surg Pathol 1984;8:163-9.
[6] Weiss LM, Medeiros LJ, Vickery Jr AL. Pathologic features of prognostic significance in adrenocortical carcinoma. Am J Surg Pathol 1989;13:202-6.
[7] Morimoto R, Satoh F, Murakami O, et al. Immunohistochemistry of a proliferation marker Ki67/MIB1 in adrenocortical carcinomas: Ki67/ MIB1 labeling index is a predictor for recurrence of adrenocortical carcinomas. Endocr J 2008;55:49-55.
[8] Schmitt A, Saremaslani P, Schmid S, et al. IGFII and MIB1 immunohistochemistry is helpful for the differentiation of benign from malignant adrenocortical tumours. Histopathology 2006;49:298-307.
[9] Soon PS, Gill AJ, Benn DE, et al. Microarray gene expression and immunohistochemistry analyses of adrenocortical tumors identify IGF2 and Ki-67 as useful in differentiating carcinomas from adenomas. Endocr Relat Cancer 2009;16:573-83.
[10] Wachenfeld C, Beuschlein F, Zwermann O, et al. Discerning malignancy in adrenocortical tumors: are molecular markers useful? Eur J Endocrinol 2001;145:335-41.
[11] Tuominen VJ, Ruotoistenmaki S, Viitanen A, Jumppanen L, Isola J. ImmunoRatio: a publicly available web application for quantitative image analysis of estrogen receptor (ER), progesterone receptor (PR), and Ki-67. Breast Cancer Res 2010;12:R56.
[12] van’t Sant HP, Bouvy ND, Kazemier G, et al. The prognostic value of two different histopathological scoring systems for adrenocortical carcinomas. Histopathology 2007;51:239-45.
[13] Volante M, Bollito E, Sperone P, et al. Clinicopathological study of a series of 92 adrenocortical carcinomas: from a proposal of simplified diagnostic algorithm to prognostic stratification. Histopathology 2009;55:535-43.
[14] Terzolo M, Boccuzzi A, Bovio S, et al. Immunohistochemical assessment of Ki-67 in the differential diagnosis of adrenocortical tumors. Urology 2001;57:176-82.
[15] Tissier F, Aubert S, Leteurtre E, et al. Adrenocortical tumors: improving the practice of the Weiss system through virtual microscopy: a National Program of the French Network INCa- COMETE. Am J Surg Pathol 2012;36:1194-201.
[16] Stojadinovic A, Ghossein RA, Hoos A, et al. Adrenocortical carcinoma: clinical, morphologic, and molecular characterization. J Clin Oncol 2002;20:941-50.
Predicting metastases in adrenocortical carcinoma
[17] Miller BS, Gauger PG, Hammer GD, Giordano TJ, Doherty GM. Proposal for modification of the ENSAT staging system for adrenocortical carcinoma using tumor grade. Langenbecks Arch Surg 2010;395:955-61.
[18] Mengel M, von Wasielewski R, Wiese B, Rudiger T, Muller- Hermelink HK, Kreipe H. Inter-laboratory and inter-observer reproducibility of immunohistochemical assessment of the Ki-67 labelling index in a large multi-centre trial. J Pathol 2002;198:292-9.