Histologic Criteria for Adrenocortical Proliferative Lesions Value of Mitotic Figure Variability
Alfredo Blanes, MD, PhD,1 and Salvador J. Diaz-Cano, MD, PhD, FRCPath2
Key Words: Adrenal cortex; Nodular hyperplasia; Adenoma; Carcinoma; Diagnostic criteria; Mitotic figure variability DOI: 10.1309/MCGUQ3R4A4WWN3LB
Abstract
This study compared 3 systems and a newly designed stepwise discriminant diagnostic system (SDDS) to assess accuracy, reproducibility, and reliability in adrenocortical nodular hyperplasia (ACNH; n = 82), adenoma (ACA; n = 78), and carcinoma (ACC; n = 32) (diagnoses according to World Health Organization criteria; median follow-up, 135 months). In cross-validations, we studied cortex appearance, growth pattern, cytoplasmic features, nuclear parameters, mitotic figure count (MFC), necrosis, invasion, and stromal-inflammatory reactions. The SDDS independent predictors were MFC/high- power field SD, anisokaryosis, cortex appearance, and stromal reaction, correctly classifying 100% of ACNH, 91% of ACA, and 88% of ACC cases. The Hough system correctly classified 78% of ACNH, 81% of ACA, and 84% of ACC cases; the Weiss and van Slooten systems correctly classified 100% of ACNH, 0% of ACA, and 92% of ACC cases. MFC variability is the most important adrenocortical malignancy criterion. Accurate malignancy diagnosis requires multiple variable evaluations, provided by SDDS (the most specific system) and the Weiss or van Slooten system (the most sensitive). SDDS is the most useful system for distinguishing tumors from ACNH (myxoid stroma and anisokaryosis).
Several algorithms are currently used to classify adreno- cortical tumors,1-3 but the morphologic distinction of different adrenocortical proliferative lesions remains controversial. It is assumed they represent a continuous spectrum, but the accu- rate identification of adrenocortical carcinoma (ACC) has important prognostic and therapeutic implications.4-11 These systems emphasize the identification of biologically aggres- sive lesions, resulting in nontested specificity, use different criteria to assess variables such as capsular invasion or mitot- ic figure counting, and do not address the diagnosis of adreno- cortical nodular hyperplasia (ACNH). In addition, no study has tested the reproducibility of these diagnostic systems in a single series.
This study compared the 3 most used systems with a newly designed stepwise discriminant diagnostic system (SDDS) for accuracy, reproducibility, reliability, and ease of application in a series of ACNH, adrenocortical adenoma (ACA), and ACC.
Materials and Methods
A series of consecutive bilateral ACNH (n = 82), ACA (n = 78), and ACC (n = 32) cases were selected from the pathol- ogy files of 3 referral university hospitals. All surgical speci- mens were routinely processed for histologic diagnosis and sampled with at least 1 block per centimeter of maximum lesion diameter. The cases were classified according to the World Health Organization and Armed Forces Institute of Pathology criteria12,13 and follow-up data (median, 135 months); lesions were classified malignant if metastases were present. The morphologic classification of adrenal cortical
hyperplasia took into account the nodule size, pattern of dis- tribution, color, and architectural pattern. The nodules com- press the adjacent cortex and reveal clear cells, along with clusters of compact cells. Cellular hypertrophy and nuclear pleomorphism may be seen in some nodules. Examination of the adjacent cortex in nodular hyperplasia shows evidence of hyperplasia.
The clinical syndrome was studied in each case by the standard clinical, laboratory (including basal hormone deter- mination, along with stimulatory and suppression tests, if nec- essary), and imaging methods.12-14
Histopathologic Analysis
All microscopic slides were reviewed and used for sys- tematic histologic evaluation by 2 independent observers (A.B. and S.J.D .- C.) who performed the evaluation in tumors and dominant adenomatous nodules of ACNH. In case of dis- agreement, the lesions were discussed during simultaneous inspection before final categorization. Reproducibility data were not recorded. Evaluation included the following:
1. The histologic pattern of the cortex adjacent to ACA and ACC was considered atrophic if architectural disarray and a thick, fibrous capsule were present in a thin cortex. The opposite features defined hyperplastic cortex, and the absence of those features was indicative of normal histologic features.
2. The lesion architecture was classified according to the growth pattern as alveolar, trabecular, nodular, insular, and diffuse or solid.
3. Necrosis was defined by dense eosinophilic cytoplasm and fragmented nuclei in a neutrophilic background. Its distribution pattern and extension (lesion percentage) were also surveyed. Lymphocytic satellitosis surrounding apoptotic cells was registered as peripolesis.
4. Tissue reaction, considering 2 main subtypes: desmoplastic (homogeneous hyaline connective tissue) and myxoid (basophilic loose extracellular matrix). When present, its extension was registered as lesion percentage.
5. Inflammation type and grade were separately considered.
6. Tumor capsule invasion (defined by tumor cell infiltration through the capsule).
7. Vascular and perineural invasion was evaluated as part of lesion extension.
8. Cytoplasmic features included staining (eosinophilic, basophilic, and pale), appearance (homogeneous, foamy, and granular), and the presence of globules or cytoplasmic pigment.
9. Nuclear features were independently analyzed as reported,15 including nuclear size variability, pleomorphism, chromatin distribution, and nucleolus.
Nuclear grade was considered high if 3 or more nuclear features were present.
10. Nuclear/cytoplasmic (N/C) ratio (low or high) was referred to that of the corresponding cellular type (foamy, clear, and eosinophilic).
11. The mitotic figure count (MFC) was performed in 50 consecutive high-power fields (HPF) (5.232 mm2) as reported.16,17 The MFC and cellularity estimations (N = >[n7/4]2), where n = number of cells intersected by the microscope field diameter, were registered in each HPF, expressing the results of the MFC per 50 HPF and per 1,000 cells and the SD per HPF and per 1,000 cells.17-21 Atypical mitoses were also counted in the same 50 HPF.
Morphologic criteria proposed by Hough et al (Hough histologic system [HHS]), Weiss system (WS), and van Slooten et al (van Slooten system [VSS]) were evaluated as described Table 11.1-3 The systems were validated using an independent set of lesions (ACNH, 23; ACA, 37; ACC, 18) and calculating the percentage of cases correctly classified.
Statistical Analysis
Qualitative variables were compared by the Pearson x2 test (with Yates correction, if necessary) and the Fisher exact test, whereas quantitative variables were analyzed by nonpara- metric analysis of variance. Results were considered statisti- cally significant if the P value was less than .05 in 2-tailed dis- tribution assays.
Significant variables were introduced in several stepwise discriminant analyses using complete cases only, obtaining the corresponding discriminant function coefficients of each diagnostic group and the classification matrices. In addition, every discriminant analysis was validated by randomly sub- tracting 20 cases and then running new analyses with the remaining 172 cases. At least 10 cross-validation tests were performed for each diagnostic system. The multivariate analy- ses included the following: (1) Define a stepwise discriminant diagnostic system (SDDS) for ACNH, ACA, and ACC using the histologic features described in the preceding text. (2) Check the reproducibility of all diagnostic systems for the diagnosis of ACNH, ACA, and ACC.
Finally, contingency tables were used to calculate sensi- tivity, specificity, and predictive values for each diagnostic system in relation to ACC diagnosis, tumor diagnosis, and ACA diagnosis. Data are given as mean ± SD unless other- wise stated.
Results
The most frequent manifestations were Cushing syn- drome for ACNH and mass with no endocrine symptoms for ACA and ACC Figure 11; the simultaneous presence of
| Scoring System | Positive Variable if | ACNH (n = 82) | ACA (n = 78) | ACC (n = 32) |
|---|---|---|---|---|
| Hough histologic | ||||
| Growth pattern (0.92x) | Diffuse | 0 | 2 | 25 |
| Vascular invasion (0.92x) | Yes | 0 | 2 | 17 |
| Tumor necrosis (0.69x) | ≥2 HPF | 0 | 5 | 20 |
| Broad fibrous bands (1x) | >1 HPF | 0 | 5 | 20 |
| Capsular invasion (0.37x) | Yes | 0 | 7 | 20 |
| Mitotic index (0.60x) (100 HPF/10) | >1 | 0 | 5 | 30 |
| Pleomorphism (0.39x) | Moderate/severe | 17 | 59 | 30 |
| Weiss | ||||
| Fuhrman nuclear grade | III/IV | 9 | 17 | 27 |
| Mitotic index per 50 HPF | ≥5 | 0 | 2 | 25 |
| Atypical mitosis | Yes | 0 | 0 | 20 |
| Clear cytoplasm in cells | <25% | 17 | 32 | 32 |
| Growth pattern (diffuse ≥1/3) | Yes | 0 | 2 | 25 |
| Necrosis (confluent nests) | Yes | 0 | 5 | 20 |
| Sinusoid invasion | Yes | 0 | 2 | 12 |
| Venous invasion | Yes | 0 | 2 | 12 |
| Tumor capsule invasion | Yes | 0 | 7 | 20 |
| van Slooten | ||||
| Regressive changes (5.7x) | Moderate/extensive | 0 | 15 | 25 |
| Structural changes (1.6x) | Abnormal | 0 | 15 | 27 |
| Nuclear atypia (2.1x) | Moderate/severe | 22 | 56 | 27 |
| Nuclear hyperchromatism (2.6x) | Moderate/severe | 4 | 15 | 27 |
| Nucleolar structure (4.1x) | Abnormal | 0 | 2 | 0 |
| Mitotic index per 10 HPF (9.0x) | ≥2 | 0 | 0 | 20 |
| Capsule or capsular vessel invasion (3.3x) | Yes | 0 | 7 | 22 |
ACA, adrenocortical adenoma; ACC, adrenocortical carcinoma; ACNH, adrenocortical nodular hyperplasia; HPF, high-power field; 100 HPF/10, number of mitotic figures per 10 HPF after screening 100 HPF.
Unknown 21%
Conn 0%
Unknown 23%
Conn 15%
Unknown 6%
Conn 0%
Cushing 16%
Mass 16%
Cushing 19%
Virilization 16%
Mixed 0%
Cushing 63%
Virilization 3%
Mass 37%
Mass 56%
Mixed 6%
Virilization 0%
Mixed 3%
ACNH
ACA
ACC
endocrine symptoms and mass effect was suggestive of ACC. The only morphologic finding that correlated with the clinical manifestations was the presence of cortical atrophy in Cushing syndrome due to adrenocortical neoplasms.
Adrenocortical Nodular Hyperplasia
ACNHs were found in 69 females (84%) and 13 males (16%) with an age of 43.5 + 14.3 years. The adrenals weighed 32.2 ± 35.7 g and showed a maximum diameter of 10.4 ± 4.6 cm.
Histologic examination revealed a trabecular growth pat- tern in 60 cases (73%), alveolar in 17 cases (21%), and nodular
in 5 cases (6%). The cytoplasmic features of proliferating cells were foamy and clear (65 cases [79%]) or homogeneous and eosinophilic (17 cases [21%]). Eosinophilic globules were present in only 9 cases (11%) and pigment in 4 (5%).
No confluent necrosis was observed, but isolated cells showing apoptotic features and lymphocytic satellitosis were seen in 13 cases (16%). The inflammatory response was unre- markable, and there was no extraglandular extension.
Bland nuclear features predominated: low grade, 78 cases (95%); low anisokaryosis, 47 cases (57%); mild pleomor- phism, 73 cases (89%); fine chromatin distribution, 69 cases
(84%); inconspicuous nucleoli, 78 cases (95%); and low N/C ratio, 60 cases (73%). Only 4 cases (5%) showed nuclear pseudoinclusions, and 13 revealed multinucleated cells (16%). The MFC was low (0.44 ± 0.55 MF/50 HPF; 0.12 ± 0.22/1,000 cells) with no atypical MF.
Adrenocortical Adenoma
ACAs were found in 64 females (82%) and 14 males (18%) with an age of 45.5 ± 14.9 years. The adrenals weighed 47.9 ± 92.6 g, and the tumor measured 5.5 ± 3.0 cm in maxi- mum diameter.
Histologically, the peritumoral gland was atrophic in 29 cases (37%), hyperplastic in 10 cases (13%), and unremark- able in 39 cases (50%). The growth pattern of the tumors was alveolar in 29 cases (37%), trabecular in 27 (35%), nodular in 20 (26%), and insular in 2 (3%). Cytologically, tumors composed of foamy, clear cells predominated in this series (46 cases [59%]), and only 20 cases (26%) contained eosinophilic globules. Twelve cases (15%) had pigment-con- taining cells.
Small necrotic foci were present in 5 cases (6%) and iso- lated apoptotic cells in 25 cases (32%). No or minimal inflam- matory response was found, predominantly lymphocytic (51 cases [65%]). Myxoid stromal reaction was observed in 29 cases (37%) and desmoplastic tissue response in 5 cases (6%). Only 2 cases (3%) showed occasional intravascular cells through wall discontinuity and capsular disruption, but no perineural or lymphatic involvement was seen.
Nuclear features were low grade (63 cases [81%]), prominent anisokaryosis (76 cases [97%]), mild pleomor- phism (56 cases [72%]), fine chromatin distribution (46 cases [59%]), inconspicuous nucleoli (63 cases [81%]), and a low N/C ratio (39 cases [50%]). Nuclear pseudoinclusions were observed in 18 cases (23%) and multinucleated cells in 37 cases (47%). The MFC was low (0.90 ± 0.09 MF/50 HPF; 0.24 ± 0.36/1,000 cells) with no atypical MF.
Adrenocortical Carcinoma
ACCs were found in 23 females (72%) and 9 males (28%) with an age of 48.3 ± 19.2 years. The adrenals weighed 470.0 ± 590.8 g, and the tumor measured 12.0 ± 6.2 cm in maximum diameter.
Microscopically, the peritumoral gland was atrophic in 3 cases (9%), hyperplastic in 3 (9%), and unremarkable in 26 (82%). The growth patterns of the tumors were diffuse in 18 cases (56%), insular in 7 (22%), nodular in 5 (16%), and tra- becular in 2 (6%). Cytologically, all tumors predominantly revealed eosinophilic cells, and 7 cases (22%) also showed eosinophilic globules. No pigment was seen.
Tumor cell necrosis was frequently observed (21 cases [66%]), with a confluent (14 cases [44%]) or perithelial pat- tern (7 cases [22%]). In addition, apoptotic cells were present
in 25 cases (78%), associated with lymphocytic satellitosis in only 11 cases (34%). The inflammatory response was minimal in 23 cases (72%), predominantly lymphocytic in 16 cases (50%). A stromal reaction was observed in 21 cases (66%) and was desmoplastic in 19 (59%) and myxoid in 2 (6%). Local staging showed extracapsular extension in 16 cases (50%) and blood vessel invasion (capsular blood in 14 cases [44%] and extracapsular in 2 [6%]). Only 1 case showed perineural inva- sion, but no lymphatic invasion was observed.
Nuclear features included high grade in 30 cases (94%), prominent anisokaryosis in 32 (100%), severe pleomorphism in 30 (94%), coarse chromatin in 25 (78%), inconspicuous nucleoli in 25 (78%), and a low N/C ratio in 21 (66%). Nuclear pseudoinclusions were present in 7 cases (22%) and multinucleated cells in 11 (34%). The MFC was high (15.58 ± 0.53 MF/50 HPF; 3.52 ± 1.96/1,000 cells) with atypical MFs in 21 cases (66%; indexing 1.83 ± 2.98/50 HPF and 0.36 ± 0.61/1,000 cells). Fewer than 15% of ACC cases (4/32) had an SDMF/HPF of less than 0.24, whereas more than 95% of the benign lesions (154/160) had an SDMF/HPF of less than 0.29.
SDDS Definition
ACC was mainly defined by unremarkable peritumoral gland, diffuse or insular tumor growth pattern, eosinophilic cyto- plasm in tumor cells, high nuclear grade (highly variable nuclear size and severe pleomorphism), and high SDMF/HPH, along with confluent necrosis and desmoplastic reaction ITable 21.
The SDDS independent predictors were SDMF/HPF, peri- tumoral gland pattern, anisokaryosis, and stromal reaction (presence and type) [Image 11, variables that were selected together in 70% or more of forthcoming cross-validation tests. Considering all variables as equally important, the probability of finding any of them (P) was .931 (P5 = 0.7; therefore, P = . 931). The percentage of correctly classified cases was always more than 88% “Table 31 and more than 92% when applying the initial set of criteria to the independ- ent set of cases Table 41.
Anisokaryosis and peritumoral adrenal pattern were always selected in the cross-validation tests, whereas SDME/HPE and the stromal reaction (presence and type) were selected by 80% and 70% of cross-validations, respectively. If the MFC offered no diagnostic information, the pondered nuclear grade became the first alternative.
Other Systems
The HHS limits to diagnose ACNH, ACA, and ACC Figure 21 meant that ACNH must not fulfill any criterion, although in this series, 4 ACNH cases showed moderate to severe pleomorphism (Table 1). Similarly, ACA criteria would include moderate to severe pleomorphism, capsular invasion, and/or any of the following: tumor necrosis, broad fibrous bands, or mitotic index of more than 1/10 HPF, in
ITable 2 Independent Predictive Variables Selected by Stepwise Discriminant Analysis and Cross-Validation Tests in 192 Samples*
| ACNH (n = 82) | ACA (n = 78) | ACC (n = 32) | P | |
|---|---|---|---|---|
| Peritumoral gland | <. 0001 | |||
| Atrophic | — | 29 (15.1) | 3 (1.6) | |
| Normal | — | 39 (20.3) | 26 (13.5) | |
| Hypertrophic | — | 10 (5.2) | 3 (1.6) | |
| Cytoplasmic staining | .0001 | |||
| Eosinophilic | 22 (11.5) | 32 (16.7) | 32 (16.7) | |
| Basophilic | 0 (0.0) | 0 (0.0) | 0 (0.0) | |
| Pale, finely vacuolated | 60 (31.3) | 46 (24.0) | 0 (0.0) | |
| Cytoplasmic appearance | .0001 | |||
| Homogeneous | 18 (9.4) | 16 (8.3) | 27 (14.1) | |
| Foamy | 64 (33.3) | 52 (27.1) | 0 (0.0) | |
| Granular | 0 (0.0) | 10 (5.2) | 5 (2.6) | |
| Anisokaryosis | <. 0001 | |||
| Mild | 47 (24.5) | 2 (1.0) | 0 (0.0) | |
| Intense | 35 (18.2) | 76 (39.6) | 32 (16.7) | |
| Nuclear grade | <. 0001 | |||
| Low | 78 (40.6) | 63 (32.8) | 2 (1.0) | |
| High | 4 (2.1) | 15 (7.8) | 30 (15.6) | |
| Atypical mitosis | <. 0001 | |||
| No | 82 (42.7) | 78 (40.6) | 12 (6.3) | |
| Yes | 0 (0.0) | 0 (0.0) | 20 (10.4) | |
| Necrosis | <. 0001 | |||
| No | 82 (42.7) | 73 (38.0) | 12 (6.3) | |
| Yes | 0 (0.0) | 5 (2.6) | 20 (10.4) | |
| Capsular invasion | .0002 | |||
| No | 82 (42.7) | 76 (39.6) | 17 (8.9) | |
| Yes | 0 (0.0) | 2 (1.0) | 15 (7.8) | |
| Stromal reaction | .0002 | |||
| No | 82 (42.7) | 49 (25.5) | 10 (5.2) | |
| Yes | 0 (0.0) | 29 (15.1) | 22 (11.5) | |
| Type of stromal reaction | <. 0001 | |||
| None | 82 (42.7) | 49 (25.3) | 10 (5.2) | |
| Myxoid | 0 (0.0) | 25 (13.0) | 2 (1.0) | |
| Desmoplastic | 0 (0.0) | 4 (2.1) | 20 (10.4) | |
| SDME/HPE (50 HPF) (Mean ± SD) | 0.05 ± 0.08 | 0.09 ± 0.10 | 0.53 ± 0.29 | <. 0001 |
| Necrosis extension (%) (Mean ± SD) | 0.00 ± 0.00 | 0.97 ± 4.55 | 16.67 ± 17.23 | <. 0001 |
ACA, adrenocortical adenoma; ACC, adrenocortical carcinoma; ACNH, adrenocortical nodular hyperplasia; MF/HPF, mitotic figure/high-power field. Data are given as number (percentage of 192) unless otherwise indicated.
decreasing importance. Although any histologic criterion was found in ACA, cases without nuclear pleomorphism or fulfilling more than 4 criteria were unlikely (Table 1). ACC was diagnosed if the lesion showed moderate to severe pleo- morphism, a mitotic index of more than 1/10 HPF, a diffuse growth pattern, and any of the remaining features (especial- ly necrosis, broad fibrous bands, and capsular invasion), in decreasing order of relevance (Table 1). No Hough criterion was ACC-specific. The percentages of cases correctly classi- fied by the HHS are shown in Table 51 and Table 61.
WS and VSS discriminant function coefficients provid- ed a narrow range for ACA diagnosis (Figure 2), explaining the ACNH diagnosis assigned to all benign lesions. Lack of clear cytoplasm and Fuhrman III/IV nuclear grade were the most frequent WS criteria in ACA; the others were found in fewer than 10% (Table 1). VSS characterized ACA by mod- erate to severe nuclear atypia, moderate to severe hyperchro- matism, and moderate to extensive regressive changes; tumor capsule invasion was found in tumors with no regressive
changes (Table 1). ACC-specific criteria were the presence of atypical mitosis and 2 or more MF/10 HPF, whereas more than 5 MF/50 HPF and intravascular tumor cells were also found in ACA (Table 1). ACC fulfilled several criteria (WS scores > 3 and VSS scores > 1.49), including moderate to severe pleomorphism and hyperchromatism, structural changes (diffuse growth pattern), and high mitotic indices (Table 1). VSS regressive changes were frequent findings, whereas WS always revealed no clear cells. The percentages of cases correctly classified by the WS and VSS are shown in Table 71 and Table 81.
Reproducibility, Sensitivity, and Specificity Analyses
The sensitivity, specificity, and positive and negative predictive values are shown in ITable 91. Discriminant func- tions achieved high reproducibility with diagnostic agree- ment in 159/192 cases (82.8%); WS and VSS discriminant functions did not distinguish ACNH from ACA, restricting the reproducibility for these systems to the benign-malignant
A
B
C
ITable 3 Classification Matrix for the SDDS
| Final Diagnosis Classification | SDDS Classification Percentage Correct ACNH ACA ACC Total | ||||
|---|---|---|---|---|---|
| Nodular hyperplasia | 100 | 82 | 0 | 0 | 82 |
| Adenoma | 91 | 7 | 71 | 0 | 78 |
| Carcinoma | 88 | 0 | 4 | 28 | 32 |
| Total | 94.3 | 89 | 75 | 28 | 192 |
ACA, adrenocortical adenoma; ACC, adrenocortical carcinoma; ACNH, adrenocortical nodular hyperplasia; SDDS, stepwise discriminant diagnostic system.
ITable 4 Classification Matrix for the SDDS Classification in the Validation Set of Cases
| Final Diagnosis Classification | Percentage Correct | SDDS Classification ACNH ACA ACC Total | |||
|---|---|---|---|---|---|
| Nodular hyperplasia | 100 | 23 | 0 | 0 | 23 |
| Adenoma | 92 | 3 | 34 | 0 | 37 |
| Carcinoma | 94 | 0 | 1 | 17 | 18 |
| Total | 95 | 26 | 35 | 17 | 78 |
ACA, adrenocortical adenoma; ACC, adrenocortical carcinoma; ACNH, adrenocortical nodular hyperplasia; SDDS, stepwise discriminant diagnostic system.
A
7
6
6.44
5
ACC
4
3
3.42
1.95 2
ACA
1.89
0.30
1
0
0.42
0.52
0.4
ACNH
-1
0.09
-0.33
0.85
2
ACNH
ACA
ACC
B
4.5
4.0
3.99
3.5
ACC
3.0
2.5
2.75
1.93
2.0
1.89
1.5
ACA
1.51
1.05
1.0
-1
1.1
0.5
ACNH
0.31
0
ACNH
ACA
ACC
C
35
30
32.37
25
20
15
ACC
16.35
10
CA A
10.74
1.49
5
3.4
0.73
3.7
1.04
0
ACNH
0.33
-5
-1.94
3.34
-10
ACNH
ACA
ACC
distinction. ACNH cases were always classified as benign, regardless of the diagnostic system. However, HHS misclas- sified 18 ACNH cases (22%) as ACA; the diagnosis was con- cordant in the remaining cases. HHS discriminant function correctly classified 63 ACAs (81%, including 11 cases inter- preted as intermediate) but misclassified 12 as ACNH (15%) and 3 as ACC (4%). SDDS misclassified 7 ACAs as ACNH
(9%) from the same group misclassified by HHS and 4 ACCs as ACA (13%) ITable 101. Most ACCs were repro- ducibly diagnosed: 27 cases (84%) were correctly classified with all systems, and only 4 ACCs were misclassified as ACA by SDDS (Table 10). The relevance of this diagnostic system is also supported by the follow-up data of malignant and benign lesions with atypical features ITable 111.
ITable 5 Classification Matrix for the Hough Histologic Score
| Final Diagnosis Classification | Percentage Correct | Hough Histologic Score ACNH ACA ACC Total | |||
|---|---|---|---|---|---|
| Nodular hyperplasia | 78 | 64 | 18 | 0 | 82 |
| Adenoma | 81 | 12 | 63 | 3 | 78 |
| Carcinoma | 84 | 0 | 5 | 27 | 32 |
| Total | 80.2 | 76 | 86 | 30 | 192 |
ACA, adrenocortical adenoma; ACC, adrenocortical carcinoma; ACNH, adrenocortical nodular hyperplasia.
ITable 61
| Final Diagnosis Classification | Percentage Correct | Hough Histologic Score ACNH ACA ACC Total | |||
|---|---|---|---|---|---|
| Nodular hyperplasia | 78 | 18 | 5 | 0 | 23 |
| Adenoma | 81 | 6 | 30 | 1 | 37 |
| Carcinoma | 83 | 0 | 3 | 15 | 18 |
| Total | 81 | 24 | 38 | 16 | 78 |
ACA, adrenocortical adenoma; ACC, adrenocortical carcinoma; ACNH, adrenocortical nodular hyperplasia.
ITable 7 Classification Matrix for the Weiss and van Slooten Systems
| Final Diagnosis Classification | Percentage Correct | Weiss and van Slooten Systems ACNH ACA ACC Total | |||
|---|---|---|---|---|---|
| Nodular hyperplasia | 100 | 82 | 0 | 0 | 82 |
| Adenoma | 0 | 73 | 0 | 5 | 78 |
| Carcinoma | 94 | 2 | 0 | 30 | 32 |
| Total | 58.3 | 157 | 0 | 35 | 192 |
ACA, adrenocortical adenoma; ACC, adrenocortical carcinoma; ACNH, adrenocortical nodular hyperplasia.
ITable 81 Classification Matrix for the Weiss and van Slooten Systems in the Validation Set of Cases
| Final Diagnosis Classification | Percentage Correct | Weiss and van Slooten Systems ACNH ACA ACC Total | |||
|---|---|---|---|---|---|
| Nodular hyperplasia | 100 | 23 | 0 | 0 | 23 |
| Adenoma | 0 | 35 | 0 | 2 | 37 |
| Carcinoma | 94 | 1 | 0 | 17 | 18 |
| Total | 51 | 59 | 0 | 19 | 78 |
ACA, adrenocortical adenoma; ACC, adrenocortical carcinoma; ACNH, adrenocortical nodular hyperplasia.
Discussion
The diagnosis of adrenocortical proliferative lesions requires a combined and systematic evaluation. The single most important criterion of malignancy is intratumor MF vari- ability, whereas the absence of positive histologic findings defines ACNH (with the exception of some nuclear atypia). Except for high and variable MFCs, any histologic feature can be present alone in ACA. The morphologic findings used to classify adrenocortical proliferative lesions were not predictive of the clinical syndrome.
The diagnostic usefulness of MF variability (SDMF/HPF) has been hardly described for ACC diagnosis, although MFC is the most important diagnostic variable.2,3,10,11,22-24 DNA and prolif- eration heterogeneity are hallmarks of malignant transformation in the adrenal cortex,18 and, for the first time, this variability has been successfully incorporated in a diagnostic system as
SDMF/HPF. SDMF/HPF was the most powerful malignancy pre- dictor, with an SDMF/HPF of 0.30 or more (1 MF/HPF in 5 HPF, or at least 3 MF/50 HPF, 2 of them in a single HPF, screening 50 HPF) highly suggestive of malignancy (Image 1). Owing to the adrenocortical low proliferation index, accurate MF evaluation requires screening several HPF, but only the MFC has been used diagnostically.1-3 Although MFC and SD, MF/HPF are directly correlated,18 the data variability was much higher for MFC, resulting in bigger group overlaps and less reliable classifications.18 The diagnostic specificity increases by reducing the shadow areas, explaining the selection of SDMF/HPF for the SDDS and our proposal of SDMF/HPF evalua- tion in 50 HPF. Only an MFC of 2/10 HPF or more and the presence of atypical mitosis were ACC-specific2,3 but were less sensitive (present in 63% of ACCs) and associated with other non-ACC-specific criteria, such as high nuclear grade, anisokaryosis, or desmoplastic reaction (Image 1). Only the
ITable 91 Sensitivity, Specificity, and Predictive Values of Adrenocortical Diagnostic Systems*
| Sensitivity | Specificity | Positive Predictive Value | Negative Predictive Value | |
|---|---|---|---|---|
| SDDS | ||||
| Benign/ACC | 88 (28/32) | 100.0 (160/160) | 100 (28/28) | 97.6 (160/164) |
| ACNH/tumor | 93.6 (103/110) | 100 (82/82) | 100.0 (103/103) | 92 (82/89) |
| Others/ACA | 91 (71/78) | 96.5 (110/114) | 95 (71/75) | 94.0 (110/117) |
| HHS | ||||
| Benign/ACC | 84 (27/32) | 98.1 (157/160) | 90 (27/30) | 96.9 (157/162) |
| ACNH/tumor | 89.1 (98/110) | 78 (64/82) | 84.5 (98/116) | 84 (64/76) |
| Others/ACA | 81 (63/78) | 79.8 (91/114) | 73 (63/86) | 85.8 (91/106) |
| WS and VSS | ||||
| Benign/ACC | 94 (30/32) | 96.9 (155/160) | 86 (30/35) | 98.7 (155/157) |
| ACNH/tumor | 31.8 (35/110) | 100 (82/82) | 100 (35/35) | 52.2 (82/157) |
ACA, adrenocortical adenoma; ACC, adrenocortical carcinoma; ACNH, adrenocortical nodular hyperplasia; HHS, Hough histologic score; SDDS, stepwise discriminant diagnostic system; VSS, Van Slooten system; WS, Weiss system.
*The sensitivity, specificity, and predictive values are given as percentages with the numbers used for calculation in parentheses.
Blanes and Diaz-Cano / HISTOLOGIC CRITERIA IN ADRENOCORTICAL LESIONS
ITable 101 Misclassified Adrenocortical Tumors
| Case No. | HHS | WS | VSS | SDDS |
|---|---|---|---|---|
| ACA-6 | Benign | Benign | Malignant | ACNH |
| ACA-13 | Undetermined | Malignant | Benign | ACNH |
| ACA-20 | Undetermined | Benign | Malignant | ACNH |
| ACA-37 | Undetermined | Malignant | Benign | ACNH |
| ACA-38 | Benign | Benign | Malignant | ACNH |
| ACA-51 | Benign | Malignant | Benign | ACNH |
| ACA-68 | Benign | Benign | Malignant | ACNH |
| ACC-2 | Benign | Malignant | Malignant | ACA |
| ACC-8 | Benign | Benign | Benign | ACA |
| ACC-17 | Benign | Malignant | Malignant | ACA |
| ACC-25 | Benign | Benign | Benign | ACA |
ACA, adrenocortical adenoma; ACC, adrenocortical carcinoma; ACNH, adrenocortical nodular hyperplasia; HHS, Hough histologic score; SDDS, stepwise discriminant diagnostic system; VSS, Van Slooten system; WS, Weiss system.
ITable 111 Clinicopathologic Features of ACAs With Angioinvasion or Misclassified as Malignant and ACC of the Adrenal Cortex
| Case No. | Stage* | Size (cm) | Extra-adrenal Invasion | Metastasis at Diagnosis | Survival | ||
|---|---|---|---|---|---|---|---|
| Lymph Node | Distant | Disease Free (mo) | Status | ||||
| ACA-20+ | I | 5 | No | No | No | 180 | ANED |
| ACA-23+ | I | 3 | No | No | No | 179 | ANED |
| ACA-51+ | I | 5 | No | No | No | 138 | ANED |
| ACA-66+ | I | 2.5 | No | No | No | 194 | ANED |
| ACA-6+ | 8 | No | No | No | 210 | ANED | |
| ACA-13+ | 9 | No | No | No | 163 | ANED | |
| ACA-37+ | 6 | No | No | No | 119 | ANED | |
| ACA-38+ | 6.7 | No | No | No | 125 | ANED | |
| ACA-68+ | 10 | No | No | No | 152 | ANED | |
| ACC-2$ | I | 4 | No | No | No | 67 | AWED |
| ACC-175 | I | 4 | No | No | No | 79 | AWED |
| ACC-26 | I | 5 | No | No | No | 74 | AWED |
| ACC-32 | I | 5 | No | No | No | 66 | AWED |
| ACC-5 | 9 | No | No | No | 50 | AWED | |
| ACC-10 | 12 | No | No | No | 80 | AWED | |
| ACC-11 | 15 | No | No | No | 49 | AWED | |
| ACC-20 | 7 | No | No | No | 15 | AWED | |
| ACC-21 | 15 | No | No | No | 67 | DOD | |
| ACC-27 | 6 | No | No | No | 69 | AWED | |
| ACC-30 | 6 | No | No | No | 70 | AWED | |
| ACC-1 | 15 | Yes | No | No | 10 | DOD | |
| ACC-4 | 11 | Yes | No | No | 20 | DOD | |
| ACC-6 | 5 | Yes | No | No | 22 | AWED | |
| ACC-7 | 12 | Yes | No | No | 13 | DOD | |
| ACC-8$ | 4.2 | No | Yes | No | 20 | AWED | |
| ACC-9 | 21 | Yes | No | No | 12 | DOD | |
| ACC-13 | III | 10 | Yes | No | No | 21 | AWED |
| ACC-14 | III | 9 | Yes | No | No | 11 | DOD |
| ACC-15 | III | 8 | Yes | No | No | 11 | DOD |
| ACC-16 | III | 12 | Yes | No | No | 13 | DOD |
| ACC-29 | III | 14 | Yes | No | No | 12 | DOD |
| ACC-31 | III | 11 | Yes | No | No | 21 | DOD |
| ACC-3 | IV | 20 | No | Yes | Yes | 3 | DOD |
| ACC-12 | IV | 23 | Yes | Yes | Yes | 3 | DOD |
| ACC-18 | IV | 20 | Yes | Yes | Yes | 5 | DOD |
| ACC-19 | IV | 19 | No | Yes | Yes | 3 | DOD |
| ACC-22 | IV | 19 | No | Yes | Yes | 2 | DOD |
| ACC-23 | IV | 24 | Yes | Yes | Yes | 1 | DOD |
| ACC-24 | IV | 20 | Yes | Yes | Yes | 5 | DOD |
| ACC-25$ | IV | 4.5 | No | Yes | Yes | 23 | DOD |
| ACC-28 | IV | 15 | Yes | Yes | Yes | 4 | DOD |
ACA, adrenocortical adenoma; ACC, adrenocortical carcinoma; ANED, alive, no evidence of disease; AWED, alive with evidence of disease (metastasis); DOD, died of disease; SDDS, stepwise discriminant diagnostic system.
Henley DJ, van Heerden JA, Grant CS, et al. Adrenal cortical carcinoma: a continuing challenge. Surgery. 1983;94:926-931. The same staging criteria were applied to ACA in terms of size and extra-adrenal extension.
ACA misclassified as malignant by at least 1 system (SDDS classified them as benign).
* ACA with angioinvasion.
$ ACC classified as benign by SDDS.
HHS (broad fibrous bands) and VSS (as part of the regressive changes) have considered the desmoplastic reaction.1,3
Our discriminant study revealed no major diagnostic use- fulness of local staging criteria. Defined by tumor cell nests within the fibrous capsule,1-3 capsular invasion was present in 6% of ACAs and 69% of ACCs. However, capsular invasion became more specific when invasion through the capsule was required (Table 2). Vascular invasion was normally associated with tumor capsule invasion, resulting in more specificity for malignancy if the vascular invasion was demonstrated in venous structures (2/78 ACAs and 12/32 ACCs).
SDDS defines ACNH by the absence of positive histolog- ic features and, occasionally, some nuclear atypia. ACNH and ACA have shown overlapping histopathologic features,12,13 especially for multinodular glands with a dominant nodule.25 The presence of a myxoid stromal reaction and anisokaryosis (Image 1) help distinguish ACNH from ACA, and their pres- ence strongly suggests ACA. It is normally associated with dilated blood vessels, internal hemorrhage, edema, and atypical nuclei.26 This reaction probably expresses a failure of the blood vessel to support tumor growth as proven by its presence in the biggest ACNH nodules, but only ACC has revealed a statistical- ly significant increase of the vascular area.27 Alternatively, perivascular stromal cells are key elements of epithelial cell growth by secretion of stimulatory factors or lack of inhibitory factors in endocrine gland nodules and tumors.28,29 The stromal reaction has been previously evaluated as regressive changes (including hemorrhage and fibrosis) in the VSS3 and as fibrous bands in the HHS1; WS does not evaluate it.2 Atypical nuclei have been demonstrated DNA aneuploid and positive with in situ end labeling,26,30 suggesting they represent regressive changes that help distinguish ACA from ACNH only.26,31 These features have not been assessed in other systems.
There is no consensus on the evaluation of diagnostic variables like MFC, capsular invasion, regressive changes, stromal reaction, nuclear features, and local extension. These differences suggest that pathologists should become familiar with 1 or 2 diagnostic systems, especially for these relatively rare adrenocortical lesions. Our results revealed SDDS as the most specific, whereas the WS and VSS are the most sensitive for malignancy. For tumor diagnosis, the most sensitive was SDDS, and the WS, VSS, and SDDS were equally specific. These results recommend using SDDS and either WS or VSS when evaluating a potentially malignant adrenocortical lesion. The WS and VSS were developed to differentiate benign from malignant neoplasms, but they failed to distinguish ACNH from ACA. For this distinction, SDDS provided the best results, revealing only 7 ACAs (9%) misclassified as ACNH. These misclassified ACAs revealed a polyclonal pattern by human androgen receptor assay test in 6/7 females), regressive kinetics (low proliferation/relatively high apoptosis), and atypi- cal nuclear features (5/6 cases) (data not shown). These clonal
and kinetic patterns are more consistent with hyperplastic rather than neoplastic conditions, as SDDS has classified,26 providing additional biologic meaning to the SDDS classification.
How should these results be applied for the diagnosis of adrenocortical lesions? We proposed the sequential application of statistically significant criteria that appear in IFigure 31, prioritizing by specificity.
MF/HPF variability is the most important malignancy cri- terion for adrenocortical lesions, but a systematic evaluation
MF counting
SDMF/HPF ≤0.2 ≤2 MF/50 HPF
0.2 < SDMF/HPF ≤0.3 2-5 MF/50 HPF
SDMF/HPF >0.3 >5 MF/50 HPF
High nuclear grade
ACC
No
Yes
Anisokaryosis
Anisokaryosis
Diffuse growth pattern
Yes
No
No
Yes
No
Yes
Adjacent adrenal
ACNH
Tissue reaction
ACC
Hyperplastic-
Hyperplastic+
Myxoid
Desmoplastic
ACA
ACNH
ACA
ACC
MF/HPF
Blanes and Diaz-Cano / HISTOLOGIC CRITERIA IN ADRENOCORTICAL LESIONS
of multiple histologic parameters is required for accuracy. At least 2 diagnostic systems should be used to corroborate a malignant diagnosis: SDDS is the most specific, whereas WS and VSS are the most sensitive. SDDS is the most useful sys- tem to distinguish tumors from ACNH based on the analysis of myxoid stromal reaction and anisokaryosis.
From the Departments of Pathology, 1 University Hospital, Málaga, Spain; and 2 King’s College Hospital and King’s College London School of Medicine, London, England.
Presented in part at the Annual Meeting of the United States and Canadian Academy of Pathology, Washington, DC, March 1996; the Pathological Society of Great Britain and Ireland, Cambridge, England, January 1999; and the International Congress of the International Academy of Pathology, Amsterdam, the Netherlands, September 2002.
Address reprint requests to Dr Diaz-Cano: Dept of Histopathology, King’s College Hospital, Denmark Hill, London SE5 9RS, England.
References
1. Hough AJ, Hollifield JW, Page DL, et al. Prognostic factors in adrenal cortical tumors: a mathematical analysis of clinical and morphologic data. Am J Clin Pathol. 1979;72:390-399.
2. Weiss LM. Comparative histologic study of 43 metastasizing and nonmetastasizing adrenocortical tumors. Am J Surg Pathol. 1984;8:163-169.
3. van Slooten H, Schaberg A, Smeenk D, et al. Morphologic characteristics of benign and malignant adrenocortical tumors. Cancer. 1985;55:766-773.
4. Didolkar MS, Bescher RA, Elias EG, et al. Natural history of adrenal cortical carcinoma: a clinicopathologic study of 42 patients. Cancer. 1981;47:2153-2161.
5. Hajjar RA, Hickey RC, Samaan NA. Adrenal cortical carcinoma: a study of 32 patients. Cancer. 1975;35:549-554.
6. King DR, Lack EE. Adrenal cortical carcinoma: a clinical and pathologic study of 49 cases. Cancer. 1979;44:239-244.
7. Luton JP, Cerdas S, Billaud L, et al. Clinical features of adrenocortical carcinoma, prognostic factors, and the effect of mitotane therapy. N Engl J Med. 1990;322:1195-1201.
8. Richie JP, Gittes RF. Carcinoma of the adrenal cortex. Cancer. 1980;45(7 suppl):1957-1964.
9. Tang CK, Gray GF. Adrenocortical neoplasms: prognosis and morphology. Urology. 1975;5:691-695.
10. Weiss LM, Medeiros LJ, Vickery AL Jr. Pathologic features of prognostic significance in adrenocortical carcinoma. Am J Surg Pathol. 1989;13:202-206.
11. Stojadinovic A, Ghossein RA, Hoos A, et al. Adrenocortical carcinoma: clinical, morphologic, and molecular characterization. J Clin Oncol. 2002;20:941-950.
12. DeLellis RA, Lloyd RV, Heitz PU, et al, eds. Pathology and Genetics of Tumours of Endocrine Organs. Lyon, France: IARC Press; 2004. World Health Organization Classification of Tumours.
13. Lloyd RV, Douglas BR, Young WRJ. Endocrine Diseases. Washington, DC: American Registry of Pathology and the Armed Forces Institute of Pathology; 2002. Atlas of Nontumor Pathology; First series, Fascicle 1.
14. Rockall AG, Babar SA, Sohaib SA, et al. CT and MR imaging of the adrenal glands in ACTH-independent Cushing syndrome. Radiographics. 2004;24:435-452.
15. Pozo L, Naase M, Cerio R, et al. Critical analysis of histologic criteria for grading atypical (dysplastic) melanocytic nevi. Am J Clin Pathol. 2001;115:194-204.
16. van Diest PJ, Baak JP, Matze-Cok P, et al. Reproducibility of mitosis counting in 2,469 breast cancer specimens: results from the Multicenter Morphometric Mammary Carcinoma Project. Hum Pathol. 1992;23:603-607.
17. Blanes A, Rubio J, Martinez A, et al. Kinetic profiles by topographic compartments in muscle-invasive transitional cell carcinomas of the bladder: role of TP53 and NF1 genes. Am J Clin Pathol. 2002;118:93-100.
18. Blanes A, Diaz-Cano SJ. DNA and kinetic heterogeneity during the clonal evolution of adrenocortical proliferative lesions. Hum Pathol. 2006;37:1295-1303.
19. Harjacek M, Diaz-Cano S, Alman BA, et al. Prominent expression of mRNA for proinflammatory cytokines in synovium in patients with juvenile rheumatoid arthritis or chronic Lyme arthritis. J Rheumatol. 2000;27:497-503.
20. Koch M, de Miguel M, Höfler H, et al. Kinetic profiles of intraepithelial and invasive prostatic neoplasias: the key role of down-regulated apoptosis in tumor progression. Virchows Arch. 2000;436:413-420.
21. Harjacek M, Diaz-Cano S, De Miguel M, et al. Expression of galectins-1 and -3 correlates with defective mononuclear cell apoptosis in patients with juvenile idiopathic arthritis. J Rheumatol. 2001;28:1914-1922.
22. DeLellis RA. Does the evaluation of proliferative activity predict malignancy of prognosis in endocrine tumors? Hum Pathol. 1995;26:131-134.
23. DeLellis RA. Proliferation markers in neuroendocrine tumors: useful or useless? A critical reappraisal. Verh Dtsch Ges Pathol. 1997;81:53-61.
24. Vargas MP, Vargas HI, Kleiner DE, et al. Adrenocortical neoplasms: role of prognostic markers MIB-1, p53, and RB. Am J Surg Pathol. 1997;21:556-562.
25. Lack EE. Tumors of the Adrenal Gland and Extra-adrenal Paraganglia. Washington, DC: Armed Forces Institute of Pathology; 1997. Atlas of Tumor Pathology; Third series, Fascicle 19.
26. Diaz-Cano SJ, de Miguel M, Blanes A, et al. Clonality as expression of distinctive cell kinetics patterns in nodular hyperplasias and adenomas of the adrenal cortex. Am J Pathol. 2000;156:311-319.
27. Diaz-Cano SJ, de Miguel M, Blanes A, et al. Contribution of the microvessel network to the clonal and kinetic profiles of adrenal cortical proliferative lesions. Hum Pathol. 2001;32:1232-1239.
28. Thomas GA, Williams D, Williams ED. The clonal origin of thyroid nodules and adenomas. Am J Pathol. 1989;134:141-147.
29. Diaz-Cano SJ, de Miguel M, Blanes A, et al. Clonal patterns in phaechromocytomas and MEN-2A adrenal medullary hyperplasias: histologic and kinetic correlates. J Pathol. 2000;192:221-228.
30. Diaz-Cano SJ, de Miguel M, Galera-Davidson H, et al. Proliferative and apoptotic markers in the malignancy diagnosis of adrenal cortical tumors [abstract]. Pathol Int. 1996;46:15A.
31. Diaz-Cano S, Gonzalez-Campora R, Rios-Martin JJ, et al. Nuclear DNA patterns in adrenal cortex proliferative lesions. Virchows Arch A Pathol Anat Histopathol. 1993;423:323-328.