Special Article: Proceedings of the 1992 Endocrine Pathology Society Meeting, Atlanta, GA, March 14, 1992
Adrenocortical Neoplasms in Childhood and Adolescence: Analysis of Prognostic Factors Including DNA Content
Claudia Zerbini, M.D., Harry P.W. Kozakewich, M.D., David S. Weinberg, M.D., Ph.D., Diane J. Mundt, Ph.D., James A. Edwards III, B.S., H.T. (A.S.C.P.), and Ernest E. Lack, M.D.
Abstract
Thirty-two adrenocortical neoplasms in children and adolescents were evaluated for prognos- tic factors including clinical and morphological parameters and DNA ploidy. The patients were segregated into two groups according to clinical outcome: group A, represented by patients with clinically benign neoplasms (n = 15), and group B, patients with clinically malignant tu- mors as evidenced by local recurrence, metastases, or fatal outcome (n = 17). Clinical and morphological parameters in these two groups were evaluated using appropriate statistical methods. Parameters with a significant predictive value in terms of prognosis were age (p = . 04), tumor size (p = . 0003), median tumor weight (p = . 0001), mitotic count (p = 0.04), and 25% tumor necrosis or more (p = . 03). Twenty-three cases were studied for DNA ploidy: 10 cases by image analysis and 13 by both image analysis and flow cytometry. By ploidy analysis, 17 of 23 cases-12 of 14 in group A and 5 of 9 in group B-were found to be aneuploid. Multiple aneuploid peaks were found in 5 of 23 cases-4 of 14 cases in group A and 1 of 9 cases in group B. In tumors studied by both image analysis and flow cytometry, there was no discrepancy between results of ploidy analysis. There was no statistically signifi- cant association demonstrated between clinical outcome and DNA ploidy pattern. DNA ploidy heterogeneity, characterized by multiple aneuploid populations of cells, was also de- tected in both benign and malignant neoplasms. Based on our results, aneuploidy is relatively frequent in pediatric adrenocortical tumors and does not appear to have predictive value for biological behavior. Endocr Pathol 3:116-128, 1992.
Department of Pathology, Children’s Hospital (CZ, HPWK, JAE), and Depart- ment of Pathology, Brigham and Women’s Hospital (DSW), Harvard Medical School, Boston, MA; De- partment of Biostatistics (DJM) and Department of Pathology (EEL), George- town University School of Medicine, Washington, DC; and formerly of the National Cancer Institute (EEL), Na- tional Institutes of Health, Bethesda, MD.
Address correspondence to Dr. Lack, Basic Science Bldg. Rm. 161, Georgetown University School of Medi- cine, 3900 Reservoir Road NW, Washington, DC 20007.
@ 1992 Blackwell Scientific Publications, Inc.
Adrenocortical neoplasms in children and adolescents are rare and account for only a small proportion of tumors observed in the first two decades of life. In general, it has proved difficult to develop criteria that distinguish benign and malignant adrenal tumors with certainty, although various re- searchers have suggested histological criteria for this purpose [14, 23, 29, 30]. Three separate studies have proposed histological criteria in an attempt to distinguish between clinically benign and malignant adrenocorti- cal tumors [14, 29, 30]. In the pediatric age group, however, the data regarding prognos-
tic features of adrenocortical tumors are scarce and more controversial [4, 20, 25]. It therefore seemed useful to combine the experience with these tumors from several institutions, focusing primarily on statistical analysis of pathological factors that might have prognostic value.
DNA content of adrenocortical neo- plasms in adults has been reported [1, 3, 6, 7, 13, 17-19, 24], but only a few studies have been done on tumors in the pediatric age group [26, 28]. In the present study, clinical and morphological features of DNA ploidy of 32 adrenocortical neoplasms from individuals
in the first two decades of life were statisti- cally analyzed for prognostic significance. Epidemiological, clinical, and pathological features of 30 tumors described herein were published recently [20]. A pertinent review of the literature related to DNA ploidy for adrenocortical neoplasms is provided.
Materials and Methods
Cases
The 32 cases of adrenocortical tumors were obtained from the National Cancer Institute, Bethesda, Maryland (n = 16), Children’s Hospital, Boston, Massachusetts (n = 15), and Georgetown University Medical Center, Georgetown, Washington, DC (n = 1). The age and sex of the patients, clinical data, size and weight of each tumor, and the original pathological diagnosis were obtained in all cases from the clinical records and pathology reports. A more detailed study of the clinico- pathological features, treatment, and fol- low-up of 30 of these patients is published elsewhere [20].
The patients were segregated into two groups based on clinical follow-up: Group A included patients who were alive (or who had died of other causes) without evidence of recurrent or metastatic adrenocortical tumor (n = 15). Included in group A were tumors that had originally been diagnosed as adreno- cortical adenomas (ACA, n = 5), adrenocorti- cal carcinomas (ACC, n = 9), and adrenocor- tical tumor with indeterminate malignant potential (ACI, n = 1). Group B consisted of patients with recurrence or documented metastases or fatal outcome due to tumor (n = 17).
The length of follow-up varied from 11 months to 17 years. Sex, age, laterality, the presence of an endocrine syndrome, weight and size of the tumor, and original pathologi- cal diagnosis (ACA or ACC) were evaluated in groups A and B using statistical methods.
Pathological Features
Using statistical analysis, the following param- eters were compared in groups A and B: growth pattern (alveolar or nesting, diffuse or solid, trabecular, or mixed), the presence of capsular invasion and vascular invasion, broad fibrous bands, nuclear pseudoinclusions, mi-
totic activity (including atypical mitoses), and microcalcifications, and the proportion of cells with compact, eosinophilic cytoplasm versus pale lipid-rich cells. Nuclear pleomor- phism was graded as 0 (absent), 1+ (mild), 2+ (moderate), or 3+ (marked). The mitotic rate was evaluated in 50 high-power fields (hpf) after selecting the areas with greatest mitotic activity. The amount of tumor necro- sis was visually estimated from examination of the histological slides and was expressed as a percentage of total area of tumor.
DNA Ploidy Analysis
DNA ploidy was measured in 23 of the 32 tumors, 10 cases by image analysis only (Cases 7, 11, 12, 14, 16, 19, 20, 22, 24, 30) and 13 cases by both image analysis and flow cytometry (Cases 1-6, 8, 9, 13, 15, 21, 25, 29). Paraffin blocks of 13 different tumors were available. For image analysis, 5-um sections were cut from selected paraffin blocks and were stained for DNA by the Feulgen method (Cell Analysis System, Elm- hurst, IL). In 10 cases, hematoxylin and eosin-stained slides were destained and restained by the same Feulgen method (Cases 7, 11, 12, 14, 16, 19, 20, 22, 24, 30). DNA content of nuclei in Feulgen-stained sections was measured by image analysis using the CAS-200 System (Cell Analysis System) with the Quantitative Ploidy Analysis software (version 2.5). This software provides a filter function that automatically classifies nuclei into six classes according to nuclear area, shape, DNA content (in picograms), and optical density. The classes were previously defined in such a way that classes 1 and 2 corresponded to diploid and class 5 to tetraploid cells; classes 3 and 4 had DNA content between diploid and tetraploid, and class 6 represented hypertetraploid cells. In every case, at least 100 nuclei were measured. Only 1 of the 23 cases stained by the Feulgen method (Case 13) was not accepted by this filter function because all cells were in the hypodiploid range (suboptimal staining).
From the 13 cases in which paraffin blocks were available, nuclear suspensions were prepared from thick sections (50 um) by the method of Hedley and co-workers [12] as modified by Stephenson and associates [27]. Two additional 5-um hematoxylin and eo- sin-stained sections (before and after the
sections used for flow cytometry) were examined to confirm the histology. To each aliquot of 106 nuclei were added 50 µg/ml of propidium iodide (Sigma Chemical Co., St. Louis, MO) and 15 pl of ribonuclease A (final concentration: 500 µg/ml ribonuclease A, bovine pancreas type III-A [Sigma]). The nuclei were incubated at room temperature for 30 minutes. Immediately before analysis, the suspension was passed twice through a 26-gauge needle to disperse nuclear aggre- gates. Human peripheral blood lymphocytes served as an external diploid DNA standard to check instrument alignment and precision before each run. Flow cytometry of the stained samples was performed using a FACS Analyzer (Beckton-Dickinson Immunocy-
tometry Systems, San Jose, CA). Histograms were generated from 5,000 nuclei and were displayed as linear fluorescence using Con- sort 30 software (Beckton-Dickinson). A sample was defined as diploid if there was a single Go/G1 peak and as aneuploid if there was more than one distinct Go/G1 peak. The DNA index of aneuploid populations was calculated by dividing the mode of fluores- cence intensity (channel number) of each aneuploid Go/G1 peak by the modal fluores- cence intensity of the presumed diploid peak. Peaks having a DNA index close to 2.0 (1.80-2.10) and containing more than 15% of the total nuclei were considered tetraploid; all other nondiploid peaks were considered DNA aneuploid. The assignment of DNA
| Case | Age at Diagnosis | Sex | Endocrine Syndrome | Laterality | Tumor Size (cm) Weight (g) | Follow-up |
|---|---|---|---|---|---|---|
| Patients with pathological diagnosis of adrenocortical adenomas | ||||||
| 1 | 1 yr | F | Virilization | L | 9 cm/210 g | NED 4 yr |
| 2 | 1 yr | M | Virilization | R | 5 cm/18 g | NED 17 yr. Three years later, massive |
| 8 mo | liposarcoma of retroperitoneum. Died 1 yr later. | |||||
| 3 | 6 yr | F | Cushing's syndrome | R | 4 cm/23 g | NED 17 yr |
| 4 | 7 yr | M | Virilization | R | 7 cm/40 g | NED 5 yr |
| 5 | 17 yr | M | Cushing's syndrome | R | 3.5 cm/20 g | NED 11/2 yr |
| Patients with pathological diagnosis of adrenocortical tumor of indeterminate malignant potential | ||||||
| 6 | 9 yr | M | Nonfunctional | R | 12 g | NED 11 mo |
| Patients with pathological diagnosis of adrenocortical carcinomas | ||||||
| 7 | 6 mo | F | Virilization | R | Large as a clenched fist | NED 14 yr. Then left intrapleural li- posarcoma. NED-6 mo. later |
| 8 | 1 yr | F | Mixed virilization, Cushing's syndrome | R | 5 cm/44 g | NED 9 yr |
| 9 | 1 yr | F | Virilization | R | 12 cm/750 g | NED 17 yr |
| 2 mo | ||||||
| 10 | 2 yr | F | Virilization | R | 3 cm | NED 8 yr |
| 11 | 3 yr | M | Virilization | R | 2.4 cm | NED 4 yr |
| 12 | 3 yr | F | Virilization | R | 7.5 cm/110 g | Death 1 yr later, cause unknown, after |
| 8 mo | tonsillectomy. No clinical evidence of recurrence or metastasis | |||||
| 13 | 4 yr | F | Virilization | L | 6 cm | NED 8 yr |
| 14 | 14 yr | F | Nonfunctional | R | 122 g | NED 11/2 yr |
| 15 | 18 yr | M | Nonfunctional | L | 6 cm/120 g | NED 31/2 yr |
NED = no evidence of disease.
Note: Group A: Patients alive (or dead of other causes) without evidence of recurrence or metastatic adrenocortical neoplasm.
| Case | Age at Diagnosis | Sex | Endocrine Syndrome | Laterality | Tumor Size (cm) Weight (g) | Follow-up |
|---|---|---|---|---|---|---|
| 16 | 10 mo | F | Nonfunctional | R | 10 cm | NED for 12 yr 4 mo; then devel- oped spinal chondrosarcoma and died 16 mo later with metastases. No clinical evidence of adreno- cortical carcinoma |
| 17 | 14 mo | M | Virilization, probably feminization as well with gynecomastia | R | 6,000 g | DOD 6 yr |
| 18 | 5 yr | M | Mixed virilization, Cushing's syndrome | L | 11 cm/560 g | DOD 11 mo |
| 19 | 5 yr | M | Virilization | L | - | DOD 9 mo |
| 20 | 61/2 yr | M | Virilization | L | 6.5 cm | DOD 3 yr, 5 mo |
| 21 | 61/2 yr | M | Virilization | L | 15 cm/475 g | DOD (intraoperative death due to massive tumor embolization to right heart). Metastases |
| 22 | 8 yr | M | Virilization | R | 17 cm | DOD 8 mo. Also paraventricular cerebral astrocytoma |
| 23 | 9 yr | F | Mixed virilization, Cushing's syndrome | L | 12 cm/340 g | DOD 1 yr 4 mo |
| 24 | 10 yr | F | Nonfunctional but increased urinary 17-ketosteroids | L | 22 cm/1,428 g | DOD 5 mo |
| 25 | 101/2 yr | F | Virilization | R | 5 cm | DOD 1 wk after admission. Hyper- pyrexia and hypertensive crisis with seizures. Metastases |
| 26 | 13 yr | M | Virilization | L | - | DOD 11 mo |
| 27 | 14 yr | F | Mixed virilization, Cushing's syndrome | R | 9 cm/230 g | DOD 9 mo |
| 28 | 16 yr | F | Mixed virilization, Cushing's syndrome | R | 10 cm | DOD 1 yr 8 mo |
| 29 | 16 yr | F | Virilization | R | 15.5 cm/1,250 g | DOD 1 yr 2 mo |
| 30 | 17 yr | F | Virilization | L | - | DOD 2 mo |
| 31 | 18 yr | F | Nonfunctional | R | Massive retro- peritoneal mass | DOD 2 yr |
| 32 | 19 yr | f | Mixed virilization, Cushing's syndrome | L | 10 cm | DOD 1 yr |
NED = no evidence of disease; DOD = dead of disease.
Note: Group B: Patients with documented metastases or fatal outcome due to adrenocortical neoplasm.
ploidy was made without knowledge of clinical data, histological diagnosis, or outcome. Acceptable histograms were obtained from 6 of 13 cases in which flow cytometric study was performed (Cases 3, 5, 13, 15, 21, 29).
Statistical Analysis
Various prognostic factors in Groups A and B were evaluated for significance using appropri- ate statistical methods. For comparisons
involving categorical variables, Fisher’s exact test was used to determine the p value corresponding to the likehood of obtaining the observed distribution of the variable of interest by group. Comparisons involving continuous variables were evaluated using the Student’s t test; variables that showed a highly skewed distribution (tumor size and weight) were compared using a nonparametric me- dian test. As the sample size was small and data were missing for a number of patients,
| Feature | Group Aª (n = 15) | Group B (n = 17) |
|---|---|---|
| Average age | 5.9 yr | 10.3 yr |
| Average size | 5.8 cm | 11.9 cm |
| Median weight | 44 g | 560 g |
| Initially classified as | ||
| Adrenocortical adenomas (no.) | 5/15 | 0/17 |
| Virilization | 3/5 | - |
| Cushing's syndrome | 2/5 | - |
| Average weight | 62 g (range, 18-210 g) | - |
| Initially classified as | ||
| Adrenocortical carcinomas (no.) | 9/15 | 17/17 |
| Virilization | 6/9 | 9/17b |
| Mixed virilization/Cushing's syndrome | 1/9 | 5/17 |
| Nonfunctional | 2/9 | 3/17 |
| Average weight | 229 g (range, 44-750 g) | 1,469 g (range, 230-6,000 g) |
ªOne case was classified as adrenocortical tumor, indeterminate (12 g).
bOne case probably mixed with feminization.
multivariate modeling to predict outcome was not possible.
Results
Clinical Findings
The clinical features, including size and weight of tumor and follow-up for patients in
No Evidence of Disease
= Female
= Male
OO
0000
00
O
O
O
O
O
o
O
Dead of Disease
o
-
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
Age (years)
groups A and B, are presented in Tables 1 and 2, respectively. The main findings in both groups are summarized in Table 3. The distribution of the patients by age and sex in groups A and B is presented in Fig. 1. The average and median ages of the 32 patients (19 girls, 13 boys) were 8.3 years and 6.7 years, respectively, at the time of diagnosis (range, 6 months to 19 years). Most patients were white, the exceptions being 1 black, 1 Filipino, and 1 Portuguese. Two patients (Cases 11 and 19) were members of a kindred with the cancer family (Li-Fraumeni) syn- drome. Twenty tumors (63%) arose in the right adrenal gland and 12 (37%) in the left. Nearly all children presented because of signs or symptoms related to endocrine hormone overproduction. Virilization alone was present in 18 patients (56%), and a mixed endocrine syndrome consisting of virilization and Cush- ing’s syndrome was present in 6 (19%). Four of the 6 patients without clinical manifesta- tions of endocrine dysfunction presented with sudden onset of abdominal or flank pain and, in 1 case, the tumor was discovered at the time of laparotomy and resection of an adenocarcinoma of the cecum (Case 14). One patient presented with headaches secondary to hypertension (Case 6). An upper abdomi- nal or flank mass was palpable in 10 patients. Statistically, there was no prognostic signifi- cance with regard to type of clinical endocrine syndrome.
Only 1 child (Case 25) had no surgical intervention, and 3 had exploratory laparot- omy and biopsy only. Twelve patients re- ceived chemotherapy, and 1 patient received adjuvant radiotherapy. There were 4 deaths unrelated to the adrenocortical tumor (Cases 2, 7, 12, 16). Four patients developed second primary tumors (Cases 2, 7, 16, 22).
The only clinical feature that was signifi- cantly different statistically between both groups was age (p = . 04) (Table 4). The average tumor size (p = . 0003) and median value of tumor weight (p = . 0001) were found to be statistically significant in groups A and B (see Table 4). The clinically malignant tumors tended to occur in children older than 5 years (see Fig. 1) and tended to be larger. Only 1 tumor in group A was greater than 10 cm in diameter and weighed more than 250 g, whereas all but 3 tumors in group B were larger than 10 cm, and all except 1 weighed more than 250 g.
| Variable | Average | Group A | Group B | P Values |
|---|---|---|---|---|
| Age (yr) | 8.3 | 5.9 | 10.3 | .04 |
| Tumor size, avg. (cm) | 8.6 | 5.8 | 11.9 | .0003 |
| Tumor weight, median (g) | 210 | 44 | 560 | .0001 |
| Mitotic count | 20 | 11 | 31 | .04 |
| Tumor necrosis ≥25% (visually estimated) | 41% | 20% | 59% | .03 |
Histopathological Features
The frequency of occurrence of the major histopathological features in groups A and B are presented in Table 5. With regard to growth pattern, although there was some overlap in both groups, a broad anastomosing trabecular pattern was more prevalent in the group B tumors, although it was found not to be statistically significant. This pattern was rather characteristic, with delicate microvascu- lature and broad anastomosing columns of cells, often 10 to 20 cells wide; in some fields, depending on plane of section, there were free-floating islands or nests of tumor. This pattern is very similar to the “sinusoidal invasion” referred to by Weiss [30]. Vascular invasion was more commonly noted in group B tumors (24% versus 7%). Capsular inva- sion, seen either histologically as irregular tongues of tumor extending into or through
| Feature | Group A (n = 15) | Group B (n = 17) |
|---|---|---|
| No. initially classified as ad- renocortical carcinoma | 9/15 (60%) | 17/17 (100%) |
| No. with mitotic figures | 10/15 (67%) | 16/17 (94%) |
| Avg. no. mitosis (range) | 11/50 hpf (1-61) | 31/50 hpf (1-88) |
| Atypical mitosis | 4/15 (27%) | 5/17 (29%) |
| Tumor necrosis ≥25% (no.) | 3/15 (20%) | 10/17 (59%) |
| 7/15 | 15/17 | |
| Vascular invasion | 1/15 (7%) | 4/17 (24%) |
| Capsular invasion | 6/15 (37%) | 6/17 (35%) |
| Broad fibrous bands | 4/15 (27%) | 7/17 (41%) |
| Calcifications | 5/15 (33%) | 11/17 (65%) |
| hpf = high-power field. |
the capsule or clinically as local invasion noted intraoperatively, was present with nearly equal frequency in group A and group B tumors. Broad fibrous bands and calcifica- tions were detected more frequently in group B (41% and 65%, respectively) compared with group A (27% and 33%, respectively). Cells with compact eosinophilic cytoplasm predominated over pale-staining, lipid-rich cells in both groups. Nuclear pseudoinclu- sions were present in 10 tumors (31%), appearing equally in group A and B tumors, though within individual tumors, they oc- curred in only a minority of cells. Mitotic figures were observed in 67% of group A tumors and nearly all (94%) of group B tumors; when present, mitoses were fewer in group A (mean of 11 per 50 hpf) compared with group B (mean of 31 per 50 hpf). Atypical mitoses were present with nearly equal frequency in groups A (27%) and B (29%). Nuclear pleomorphism was encoun- tered in both groups and was not a useful prognostic indicator. Although foci of necro- sis were observed histologically in nearly half of group A tumors, significant necrosis (≥25%) occurred in 20% of group A compared with 59% of group B tumors. Among the histopathological parameters, only mean mitotic count (p = . 04) and tumor necrosis (≥25%) (p = . 03) were statistically different between the two groups (see Table 4).
DNA Ploidy
Reliable histograms were obtained in 22 of 23 cases by image analysis and in 6 of 13 cases by flow cytometry. In 5 cases (Cases 3, 5, 15, 21, 29), both image analysis and flow cytometric DNA histograms were obtained. In only 1 case (case 13) was only flow cytometric data available. The results are presented in Table 6.
Seventeen of 23 tumors (74%) were aneuploid-12 of 14 (86%) in group A and 5 of 9 (56%) in group B. The difference between the two groups was not statistically significant (p = . 16). In Case 5 (group A), a small questionable aneuploid peak (10% of total number of cells) was detected by image analysis. Heterogeneity of DNA content, characterized by multiple aneuploid peaks in the same sample, was found in 5 of 23 tumors (22%) in which ploidy analysis was per- formed-4 of 14 in group A and 1 of 9 in
group B-and again the difference between groups A and B was not statistically significant (p = . 21).
In 1 tumor (Case 14), flow cytometric analysis was performed on two separate blocks, and different aneuploid populations were detected (Fig. 2). In Case 21, two different blocks were evaluated by flow cytometry and both were diploid. In all other cases, one selected block or slide was evalu- ated for DNA ploidy.
An associated diploid peak was also present in 15 of 17 aneuploid tumors, interpreted as residual normal tissue. In the 5 cases in which both image analysis and flow
| Case | Ploidy | Other Interesting Findings |
|---|---|---|
| Group A | ||
| 1 | Aneuploid [2.14] | Associated diploid peak |
| 2 | Aneuploid [1.48] | Associated diploid peak |
| 3 | Aneuploid [1.43] | Associated diploid peak |
| 4 | Diploid | |
| 5 | Diploid | |
| 6 | Aneuploid | Multiple aneuploid peaks [1.73, 2.23] |
| 7 | Aneuploid [1.42] | |
| 8 | Aneuploid [1.38] | Associated diploid peak |
| 9 | Aneuploid | Multiple aneuploid peaks [1.57, 1.81] |
| 11 | Aneuploid [1.23] | Associated diploid peak |
| 12 | Aneuploid | Multiple aneuploid peaks [1.75, 4.26] |
| 13 | Aneuploid [1.20] | Associated diploid peak |
| 14 | Aneuploid | Multiple aneuploid peaks [1.37, 1.63] |
| 15 | Aneuploid [1.67] | Associated diploid peak |
| Group B | ||
| 16 | Aneuploid [1.28] | Associated diploid peak |
| 19 | Diploid | |
| 20 | Diploid | |
| 21 | Diploid | |
| 22 | Aneuploid | Multiple aneuploid peaks [1.55, 2.44, 2.69] |
| 24 | Aneuploid [1.59] | Associated diploid peak |
| 25 | Aneuploid [1.28] | Associated diploid peak |
| 29 | Diploid | |
| 30 | Aneuploid [1.34] | |
| [ ] DNA index by image analysis for Case 13, for which there was only flow cytometric data. | ||
cytometry were performed, similar results were obtained.
No statistically significant correlation was found between DNA ploidy and age of patient, weight and size of tumor, mitotic count, necrosis, and even nuclear pleomor- phism. Examples of DNA histograms from diploid tumors (Figs. 3, 4), an aneuploid tumor (Fig. 5), and a tumor with multiple aneuploid peaks (Fig. 6) are presented.
Discussion
Clinical and Histopathological Features
In the past decade, our knowledge of adrenocortical neoplasms has been expanded greatly. With the recent advent of computed tomography and magnetic resonance imaging techniques, these tumors are diagnosed ear- lier and are frequently smaller, increasing the need for accurate diagnosis and pathological indicators of prognosis. Recently, many ad- vances toward this end have been developed, and at least one set of clinicopathological [14] and two sets of pathological [29, 30] criteria have been proposed for distinguishing benign adrenocortical tumors from those with malig- nant potential in adults.
Because of the rarity of adrenocortical neoplasms in the pediatric group, however, it has been difficult to draw definitive conclu- sions regarding the value of gross or micro- scopical pathology for predicting biological behavior. In addition, the limited experience of most pathologists with adrenocortical tumors in children may contribute to a tendency to overdiagnose adrenocortical tu- mors as carcinoma in this younger group. Ribeiro and colleagues [25], using three classifications [14, 29, 30] in 33 pediatric adrenocortical tumors, found that only 4 patients (12%) did not meet the criteria for malignancy using any of these classification systems. Of these 4 patients, 3 were surviving disease-free for 8+ years, 5+ years, and 6+ months, respectively, and 1 patient was lost to follow-up. Of the 29 patients whose tumors were classified as malignant, 14 (48%) were alive without evidence of disease [25].
Cagle and co-workers [4], in a study of 23 adrenocortical neoplasms in children (17 benign and 6 malignant), concluded that the tumors in the pediatric group were more likely to be benign than had previously been
A
300
B.
1.53
20
2
1.06
Cell Number
16
1.20
Cell Count
12
1.80
1,73
8
1.94
4
O
O
100
200
0
4
DNA Mass (pg)
8
12
16
0
Channel Number
C.
200
D
20
2
Cell Number
1,37
Cell Count
16
1,63
1.87
12
8
4
O
%
0
4
8
12
16
Channel Number
100
200
DNA Mass (pg)
believed. Comparing the pediatric tumors with 42 adult tumors (29 benign and 13 malignant), these investigators noted that the clinically benign childhood tumors were significantly more likely to have mitoses, necrosis, broad fibrous bands, and moderate to severe nuclear pleomorphism compared with clinically benign adrenocortical tumors in adults. Our study, which is restricted to childhood adrenocortical tumors, indicates that an increased number of mitoses and extensive tumor necrosis (≥25%), are statisti- cally more common in clinically malignant tumors (p < . 05). Vascular invasion, broad fibrous bands, and calcifications were also more common in the malignant group of tumors, but differences were found not to be statistically significant.
Mean weight and median size of tumor were both found to be statistically different in malignant and benign adrenocortical tumors in our study. As exceptions, we have noted 1 patient in group A (Case 9) with a very large tumor in (12 cm, 750 g) who did not have any evidence of disease 17 years later and 1 patient in group B (Case 25) with a relatively small primary tumor (5 cm) who died with
A
6
*
L
P
B
B
1000
Cell Number
O
O
Channel Number
100
200
C
100
2
4
Cell Count
80
60
40
20
…
0
0
8
16
24
32
DNA Mass (pg)
multiple metastases 1 week after admission. The prognostic data related to tumor size and weight in the literature are controversial in children. Tumor weight is considered by Cagle and co-workers [4] to be the only reliable predictor of malignant behavior in the pediatric group, with 500 g being the critical weight. Ribeiro’s group [25] found that tumor weight exceeding 100 g and tumor size greater than 200 cm3 were associated with unfavorable outcome. Hayles and colleagues [11] pointed out that the likelihood of metastases best correlated with the relation-
A
B
500
Cell Number
0
O
Channel Number
100
200
C
60
2
4
Cell Count
48
36
24
12
0
0
8
16
24
32
DNA Mass (pg)
ship of adrenal tumor weight to body weight. Lefevre and associates [21] described several tumors weighing 180 g or less that were clinically malignant. Nonetheless, Hum- phrey and co-workers [16], in an overview of 72 ACC, 30 ACA, and 12 indeterminate tumors, were cautious in advocating tumor weight as the sole criterion for malignancy.
Regarding the importance of age, it has been reported that the survival rate for adolescent patients diagnosed as having ACC was worse than for the infantile group [16].
Our results agree with this observation, the survival rate for patients older than 5 years who have a pathological diagnosis of ACC (n = 15) being 13%, whereas it is 70% for patients 5 years or younger. One patient whose neoplasm was diagnosed at age 14 months (Case 17) survived 6 years before dying of metastatic ACC. In a fairly recent study of ACC in adults [31], mitotic rate was found to be useful in distinguishing low- grade (<20 mitoses per 50 hpf) from high-grade ACC (> 20 mitoses per 50 hpf). A
A
B
400
Cell Number
1.65
0
0
Channel Number
100
200
C
2
20
1.09
Cell Count
16
1.43
12
8
4
0
0
4
8
12
16
DNA Mass ( pg)
A
5
B
20
2
4
1
8
1.73
Cell Count
16
2.23
4.61
12
4.08
8
4
0
0
16
32
48
64
DNA Mass (pg)
similar trend was present in our study, although the association was not statistically significant.
Several other studies have used statistical methods to evaluate clinical and pathological parameters that have predictive value for disease-free and overall survival, mainly in adults. Our findings, like those of Cagle and associates [4], show that clinically benign adrenocortical neoplasms from patients in the first two decades of life demonstrated a greater frequency of adverse histopathological features compared with tumors in adults. One important reason for this may be the rarity of aldosterone-producing adenomas in childhood and adolescence. These tumors are not infrequent in adults, are typically small (i.e., <2 cm in diameter), and have minimal malignant potential. Weiss’s series [30], with an end point similar to ours (i.e., metastases or recurrence), includes 17 examples of benign aldosterone-producing adrenocortical neoplasms (40% of the 43 tumors analyzed); the median diameter and weight were only 2 cm and 10 g, respectively. The absence of
these small, nonmetastasizing, largely innocu- ous adrenocortical neoplasms in the pediatric group leaves a more select group of neo- plasms that are statistically larger and have more worrisome histopathology.
DNA Ploidy
The assessment of DNA ploidy in adrenocor- tical tumors has been the subject of a number of studies since 1985 [1, 3, 6, 7, 13, 15, 17-19, 24, 26, 28]. Two approaches to cellular DNA measurements have been used: flow cytome- try and static image cytometry. In our retrospective study, the earliest case dating back to 1932, the paraffin blocks were generally more successfully analyzed using Feulgen staining and image analysis than flow cytometry. Only 1 of 23 cases could not be evaluated by image analysis; 6 of 13 cases in which the paraffin blocks were available resulted in reliable DNA histograms by flow cytometry. Furthermore, in 10 cases in which the paraffin blocks were not available, the hematoxylin and eosin-stained slides were destained and restained by the Feulgen method, resulting in DNA histograms of excellent quality by image analysis. Fallenius and co-workers [9], in a methodological investigation, found that there were no quantitative errors in DNA measurements introduced by destaining archival material, indicating that there is no major difference in Feulgen stainability of cells in old specimens when compared with fresh specimens.
A summary of the literature on DNA ploidy in adrenocortical tumors, mostly in adults, is presented in Table 7. These studies are small, reflecting the rarity of these tumors, and only Taylor and colleagues [28] studied exclusively pediatric tumors in a series of 10 patients. Flow cytometry in paraffin blocks was used in every study except that by Klein and associates [18], who used fresh tissue. The results and criteria to define malignancy were highly variable. Therefore, in Table 7 we have expressed the results based on the clinical outcome (i.e., recurrence, metastasis, or death from disease). The incidence of aneuploidy varied from 19 to 58%. Camuto and co-workers [6] reported that 96% of tumors were aneuploid, but only clinically malignant tumors were evaluated. On the other hand, a high incidence of aneuploidy was found in the pediatric group [28].
| Cases with Follow-Up Total | Aneuploid | ||||||
|---|---|---|---|---|---|---|---|
| Study | Age Group | Time of Follow-Up | No. (%) | Clinically Benign | Indeter- minateb | Clinically Malignant | |
| Klein 1985 [18] | ?Adults | 7/7 | 6 mo-2 yr | 4/7 (57) | 0/3 | 1/1 | 3/3 |
| Bowlby | 3-70 yr (2 child) | 22/22 | 7 mo-21/2 yr | 5/22 (23) | 0/16 | 2º/3 | 3/3 |
| 1986 [3] | |||||||
| Amberson 1987 [1] | 2.5-88 yr (mean, 50 yr) | 36/38 | > 2 yr | 9/36 (19) | 3/29 | 2/2 | 4/5 |
| Hosaka 1987 | 1-75 yr (mean, 46.4 | 52/52 | 0 mo-28 yr (2 child) | 30/52 (58) | - | - | - |
| [13]d | yr) | ||||||
| Joensuu | 2-78 yr | 17/17 | 1-11 yr | 9/17 (53) | 9/17 | - | - |
| 1988 [17]e | (mean, 51 yr) | . | |||||
| Rainwater 1989 [24] | Adults | 5/6 | 7 mo-28 yr | 9/26 (35) | 1/3 | - | 2/2 |
| Cibas 1989 [7] | 5-78 yr (mean, 51 yr) | 32/43 | 6 mo-12 yr 8 mo | 5/43 (35) | 8/23 | - | 5/9 |
| Camuto 1991 [6]f | Mainly adults but age unspecified | 22/22 | - | 21/22 (96) | - | - | 21/22 |
| Taylor 1987 [28] | Pediatric | 10/10 | 5 mo-18 yr | 5/10 (50) | 1/6 | - | 4/4 |
aAll determined by flow cytometry.
bCases considered malignant by pathology but without clinical evidence of malignancy (local recurrence, metastasis, dead of disease).
“These 2 aneuploid cases were the only 2 children without any clinical evidence of malignancy at 4 yr and 5 yr 11 mo of follow-up.
dAll cases considered malignant according to pathological criteria, but data do not permit reclassification according to follow-up. Of the 2 children (1 and 2 years old), one (with Cushing’s syndrome) had diploid DNA and the other (with virilization) was aneuploid. These patients were alive for 27 yr and 8 yr, respectively.
“Only clinically benign adrenocortical tumors.
fOnly clinically malignant adrenocortical tumors.
Joensuu and Klemi [17] studied only clini- cally benign endocrine tumors of various types and detected aneuploidy in 53% of the adrenocortical tumors although, similar to other studies in the literature [1, 6, 17, 18, 24], there is no precise breakdown as to the exact number of pediatric tumors analyzed.
All studies except two [3, 18] detected aneuploidy in clinically benign tumors. How- ever, in both series, some aneuploid cases diagnosed as malignant by morphological criteria never presented clinical features of malignancy. Interestingly, in the study by Bowlby and associates [3], the two examples of aneuploid but clinically benign tumors
(classified as ACC by histology) were the only pediatric cases in their report. Similarly, Hosaka and co-workers [13], in the largest DNA ploidy study in the literature, found that the only patient with a virilizing ACC who survived was a 2-year-old child with an aneuploid tumor. The only other child in this report had a diploid neoplasm and was alive 27 years later [13].
In our 23 patients in whom DNA ploidy studies have been performed, 17 were found to be aneuploid (74%)-12 of 14 tumors in group A (87%) and 5 of 9 (56%) in group B. Our higher incidence of aneuploidy, com- pared to that quoted in the literature, might
reflect a true higher incidence of aneuploidy in the pediatric group or it may reflect a greater sensitivity of image analysis for detecting minor aneuploid or near-diploid populations of cells [2, 8, 9, 22, 32]. To the best of our knowledge, all studies of DNA ploidy of adrenocortical tumors in the litera- ture have used flow cytometry. Image analysis has been used to evaluate adrenal cytomegaly in nonneoplastic glands [5, 10].
Multiploid histograms, found in 5 of 23 of cases in our study (22%), have been reported in other studies as well [1, 3, 6, 7, 13]. This DNA heterogeneity is evidence that adrenocortical tumors, whether clinically benign or malignant, may arise from multiple clones of cells. In all but 2 cases in which one or multiple aneuploid stem lines were de- tected, one well-defined diploid population of adrenocortical cells was also found. This population may represent a diploid stem line of neoplastic cells or residual entrapped adrenal cells. Though overall there was no statistical correlation between nuclear pleo- morphism (scored 0 to 3+ by visual estima- tion) and ploidy as measured by image analysis or flow cytometry, there was good concordance between these two features in individual cases. The large macronuclear cells commonly present in adrenocortical tumors exhibited a wide range of increased DNA content, without constituting a clonal aneu- ploid stem line. Although accurate measure- ment of the S-phase fraction was not techni- cally possible because of the complexity of the histograms with multiple aneuploid peaks, in several cases we noted that the presence of macronuclei was associated with increased S phase and a wide range of hypertetraploid cells. Interestingly, Camuto and associates [5] reported similar findings studying DNA ploidy by flow cytometry in 2 cases of fetal adrenocortical cytomegaly.
Comments
In summary, accurate classification of some adrenocortical tumors in children remains a difficult diagnostic problem. It appears that some of the data related to prognosis in adults are not applicable for pediatric patients. In statistical analysis of our 32 cases, age at diagnosis, tumor size, median tumor weight, mitotic rate, and tumor necrosis were signifi-
cantly (p < . 05) related to clinical outcome. DNA aneuploidy seems to occur fairly frequently, particularly as compared with reported series in adults and, in our experi- ence, it was not related statistically to clinical outcome. This finding probably reflects the high frequency of aneuploidy in endocrine tumors in general, without any connotation of worse outcome.
Acknowledgment
This work was supported in part by the F.A.P.E.S.P. grant no. 90/0766-0.
References
1. Amberson JB, Vaughan ED, Gray GF, Naus GJ. Flow cytometric analysis of nuclear DNA from adrenocortical neoplasms. A retrospec- tive study using paraffin-embedded tissue. Cancer 59:2091-2095, 1987.
2. Bauer TW, Tubbs RR, Edinger MG, Suit PF, Gephard GN, Levin HS. A prospective com- parison of DNA quantitation by image and flow cytometry. Am J Clin Pathol 93: 322-326, 1990.
3. Bowlby LS, DeBault LE, Abraham SR. Flow cytometric analysis of adrenal cortical tumor DNA. Relationship between cellular DNA and histopathologic classification. Cancer 58: 1499-1505, 1986.
4. Cagle PT, Hough AJ, Pyshertj, Page DL, Kirland RT, Holcombe JH, Hawkins EP. Comparison of adrenal cortical tumors in children and adults. Cancer 57:2235-2237, 1986.
5. Camuto PM, Wolman SR, Perle MA, Grecco MA. Flow cytometry of fetal adrenal glands with adrenocortical cytomegaly. Pediatr Pathol 9:551-558, 1989.
6. Camuto P, Schinella R, Gilchrist K, Citrin D, Frederickson G. Adrenal cortical carcinoma; flow cytometric study of 22 cases, an ECOG study. Urology 37:380-384, 1991.
7. Cibas ES, Medeiros LJ, Weinberg DS, Gelb AB, Weiss LM. Cellular DNA profiles of benign and malignant adrenocortical tumors. Am J Surg Pathol 13:202-206, 1989.
8. Dawson AE, Norton JE, Weinberg DS. Comparative assessment of proliferation and DNA content in breast carcinoma by image analysis and flow cytometry. Am J Pathol 136:1115-1124, 1990.
9. Fallenius AG, Auer GU, Carstensen JM. Prognostic significance of DNA measure-
ments in 409 consecutive breast cancer pa- tients. Cancer 62:331-341, 1988.
10. Favara BE, Steele E, Grant JH, Steele P. Adrenal cytomegaly: quantitative assessment by image analysis. Pediatr Pathol 11:521-536, 1991.
11. Hayles AB, Hahn JR, Sprague RG, Bahn RC, Pristley JT. Hormone-secreting tumors of the adrenal cortex in children. Pediatrics 37: 19-25, 1966.
12. Hedley DW, Friedlaender ML, Taylor IW, Musgrove EA. Method for analysis of cellular DNA content of paraffin-embedded patholog- ical material using flow cytometry. J His- tochem Cytochem 31:1333-1335, 1983.
13. Hosaka Y, Rainwater LM, Grant CS, Young WF Jr, Farrow GM, van Heerden JA, Lieber MM. Adrenocortical carcinoma: nuclear deox- yribonucleic acid ploidy studied by flow cytometry. Surgery 102:1027-1034, 1987.
14. 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 72:390-399, 1979.
15. Hough AJ Jr. Flow cytometry and adrenal cortical tumors. J Urol 134:931-932, 1985.
16. Humphrey GB, Pysher T, Holcombe J, Grass M, Chan H, Cushing B, D’Angio GJ, Schein P, Lerner LR, Carney JA, Raney BL. Overview on the management of adrenal cortical carci- noma. In GB Humphrey et al, eds, Adrenal and endocrine tumors in children. Boston: Martinus Nijhoff, 349-358, 1983.
17. Joensuu H, Klemi PJ. DNA aneuploidy in adenomas of endocrine organs. Am J Pathol 132:145-151, 1988.
18. Klein FA, Kay S, Ratliff JE, White FKH, Newsome HH. Flow cytometric determina- tions of ploidy and proliferation patterns of adrenal neoplasms: an adjunct to histologic classification. J Urol 134:862-866, 1985.
19. Klein FA, Miller NL, Hackler RH. Flow cytometry in feminizing adrenocortical carci- noma. J Urol 134:933-935, 1985.
20. Lack EE, Mulvihill JJ, Travis WD, Koza- kewich HPW. Adrenal cortical neoplasms in the pediatric and adolescent age group: clinico- pathologic study of 30 cases with emphasis on epidemiological and prognostic factors. Pathol Annu 27(1):1-53, 1992.
21. Lefevre M, Gerard-Marchant R, Gubler JP, Chaussain JL, Lemerle J. Adrenal cortical carcinoma in children: 42 patients treated from 1958 to 1980 at Villejuif. In GB
Humphrey et al, eds, Adrenal and endocrine tumors in children. Boston: Martinus Nijhoff, 265-276, 1983.
22. McFadden PW, Clowry LJ, Daehnert K, Hause LL, Koethe SM. Image analysis confir- mation of DNA aneuploidy in flow cytometric DNA distributions having a wide coefficient of variation of the GO/G1 peak. Am J Clin Pathol 93:637-642, 1990.
23. Medeiros LJ, Weiss LM. New developments in the pathologic diagnosis of adrenal cortical neoplasms. A review. Am J Clin Pathol 97:73-83, 1992.
24. Rainwater LM, Young WF Jr, Farrow FM, Grant CS, van Heerden JA, Lieber MH. Flow cytometric analysis of deoxyribonucleic acid ploidy in benign and malignant aldosterone- producing neoplasms of the adrenal gland. Surg Gynecol Obstet 168:491-496, 1989.
25. Ribeiro RC, Neto RS, Schell MJ, Lacerda L, Sampaio GA, Cat I. Adrenocortical carcinoma in children: a study of 40 cases. J Clin Oncol 8:67-74, 1990.
26. Saracco S, Abramowsky C, Taylor S, Silver- man RA, Berman BW. Spontaneous regressing adrenocortical carcinoma in a newborn. A case report with DNA ploidy analysis. Cancer 62:507-511, 1988.
27. Stephenson RA, Gay H, Fair WR, Melamed MR. Effect of section thickness on quality of flow cytometric DNA content determination in paraffin-embedded tissues. Cytometry 7:41-44, 1986.
28. Taylor SR, Roederer M, Murphy RF. Flow cytometric DNA analysis of adrenocortical tumors in children. Cancer 59:2059-2063, 1987.
29. Van Slooten H, Schaberg A, Smeenk D, Moolenaar AJ. Morphologic characteristics of benign and malignant adrenal cortical tumors. Cancer 55:766-773, 1985.
30. Weiss LM. Comparative histologic study of 43 metastasizing and nonmetastasizing adrenal cortical tumors. Am J Surg Pathol 8:163-169, 1984.
31. Weiss LM, Medeiros LJ, Vickery AL Jr. Pathologic features of prognostic significance in adrenal cortical carcinoma. Am J Surg Pathol 13:202-206, 1989.
32. Zerbini C, Weinberg DS, Hollister KA, Perez-Atayde AR. DNA ploidy in the liver of children with hereditary tyrosinemia: correla- tion with histopathologic features. Am J Pathol 140:1111-1129, 1992.