From the Oncologic Pediatric Surgery Division, Erasto Gaertner Hospital, Liga Paranaense de Combate ao Câncer, Curitiba; Division of Pediatric Endocri- nology and Center for Molecular Genet- ics and Childhood Cancer Research (CEGEMPAC), Hospital de Clinicas, Fed- eral University of Paraná, Curitiba, Paraná; the Biostatistics Division, Hemocentro, and The Instituto Domin- gos Boldrini, Universidade Estadual de Campinas, Campinas; Department of Pediatric Oncology, Universidade Fed- eral de São Paulo, Escola Paulista de Medicina; Department of Pediatrics, São Paulo University Medical School, Sao Paulo, Brazil; the Departments of Pathology, the Department of Hematol- ogy-Oncology, and the International Outreach Program, St Jude Children’s Research Hospital; and the Department of Pediatrics, University of Tennessee College of Medicine, Memphis, TN.
Submitted August 12, 2003; accepted December 12, 2003.
Supported in part by grants CA-21765 and CA-71907 from the National Insti- tutes of Health (US Department of Health and Human Services), a Center of Excellence grant from the State of Tennessee, the Conselho Nacional de Pesquisa (CNPq) of Brazil, and the American Lebanese Syrian Associated Charities (ALSAC).
Authors’ disclosures of potential con- flicts of interest are found at the end of this article.
Address reprint requests to Raul C. Ribeiro, MD, Department of Hematology- Oncology, St Jude Children’s Research Hospital, 332 N Lauderdale St, Mem- phis, TN 38105-2794; e-mail: raul.ribeiro@stjude.org.
@ 2004 by American Society of Clinical Oncology
0732-183X/04/2205-838/$20.00
DOI: 10.1200/JCO.2004.08.085
Clinical and Outcome Characteristics of Children With Adrenocortical Tumors: A Report From the International Pediatric Adrenocortical Tumor Registry
E. Michalkiewicz, R. Sandrini, B. Figueiredo, E.C.M. Miranda, E. Caran, A.G. Oliveira-Filho, R. Marques, M.A.D. Pianovski, L. Lacerda, L.M. Cristofani, J. Jenkins, C. Rodriguez-Galindo, and R.C. Ribeiro
ABSTRACT
Purpose
We created a registry for pediatric adrenocortical tumors (ACTs), which are rare and are not well characterized. We provide a descriptive analysis of 254 patients registered on the International Pediatric Adrenocortical Tumor Registry.
Patients and Methods
Between January 1990 and December 2001, 254 patients younger than 20 years of age with newly diagnosed or previously treated ACTs were registered. A histologic diagnosis of ACT was required, although central review was not mandatory. Follow-up information was periodically requested from the referring physician. Treatment was chosen by the primary physician.
Results
The overall female-male ratio was 1.6:1, but it varied widely among age groups. The most common presenting sign (84.2%) was virilization. Cushing’s syndrome without virilization was uncommon (5.5%). Tumors were completely resected in 83% of patients. Patients with disseminated or residual disease received mitotane, cisplatin, etoposide, and/or doxorubicin, and rarely, radiation therapy. At a median follow-up of 2 years and 5 months, 157 patients (61.8%) survived without evidence of disease and 97 patients (38.2%) had died. The 5-year event-free survival estimate was 54.2% (95% CI, 48.2% to 60.2%). In a multivariate analysis, disease stage, presenting signs of endocrine dysfunction, and age were independently associated with prognosis.
Conclusion
Childhood ACTs occur predominantly in females and almost always causes clinical signs. Complete resection is required for cure. Residual or metastatic disease carries a poor prognosis. Our results demonstrate the feasibility of a disease-specific database for obtaining meaningful clinical and outcome information.
J Clin Oncol 22:838-845. @ 2004 by American Society of Clinical Oncology
INTRODUCTION
Adrenocortical tumors (ACTs) are rare in children and adolescents. The Surveil- lance Epidemiology End Results data indi- cate about 14 new patient cases per year in individuals younger than age 20 years in the United States; ACTs represent 1.3% of all carcinomas in this age group [1]. Be- cause of the rarity of pediatric ACTs, no single pediatric oncology center has ac- quired extensive experience with this tu- mor. Most reported series describe only a
few patients observed over a period of sev- eral years [2].
Lefevre et al [3] analyzed the clinical characteristics and treatment outcomes of 42 children treated in several French hospi- tals over a 22-year period. They found a median age of 3.9 years, a predominance of girls, presenting signs of virilization in more than 90% of patients, and long-term sur- vival of approximately 50%. Tumor size was the most important prognostic factor. The prevalence of increased predisposition to cancer (Li-Fraumeni syndrome) or other as-
Childhood Adrenocortical Tumors
| Table 1. Modified Disease Staging System Used for All Patients in the Registry | |
|---|---|
| Stage | Description |
| I | Tumor completely excised with negative margins, tumor weight ≤ 200 g, and absence of metastasis |
| II | Tumor completely excised with negative margins, tumor weight > 200 g, and absence of metastasis |
| III | Residual* or inoperable tumor |
| IV | Hematogenous metastasis at presentation |
| *Residual tumor is defined as the presence of microscopic or gross tumor after surgical resection. | |
sociated constitutional genetic syndromes was not re- ported. Similar clinical characteristics, treatment outcomes, and prognostic factors were observed among children with ACTs treated at a single institution in southern Brazil [4,5]. In a subsequent study, the Brazilian investigators found that the majority of these patients carried a germline p53 muta- tion [6]. In another study, Liou and Kay [7] summarized the clinical and outcome data of 412 patients from several pub- lished series, mostly from the United States. Again, the findings did not differ substantially from those previously described, suggesting that despite disparate geographic, ethnic, and genetic differences among patients, childhood ACTs seem to have a consistent clinical and biologic course [2]. To better define the clinical manifestations, prognostic factors, treatment, and outcome of childhood ACTs, we created a disease-specific database. Here we report the find- ings derived from 254 registered patient cases.
PATIENTS AND METHODS
Eligibility Criteria and Data Collection
Patients younger than 20 years with newly diagnosed or pre- viously treated ACTs were registered and included in the study. All patient cases were reported by the patient’s primary physician. Patients and their families were never contacted directly. All infor- mation was entered into a customized electronic database admin- istered by a data manager. Data were reviewed by an International Pediatric Adrenocortical Tumor Registry (IPACTR) physician for consistency before entry. The referring physician was contacted directly if clarification was necessary. Central review of tumor histology was offered to the referring physician but was not a requirement for registration. Over the study period, the histology of approximately 80% of the patient cases was reviewed at St Jude Children’s Research Hospital (Memphis, TN). Because of the dif- ficulty of consistently discriminating adenomas from carcinomas in pediatric ACTs, patients with a diagnosis of either adenoma or carcinoma were included in the database. There was no central review of imaging studies, but all patients had imaging evaluation of the chest and abdomen and underwent extensive endocrine studies after surgery. Informed consent was obtained from all patients by the participating physicians.
The extent of disease was retrospectively defined as localized or advanced on the basis of a modified staging system proposed previously (Table 1) [5,8]. Assignment to the localized disease (stage I or II) category required complete tumor resection and no evidence of disease elsewhere. Assignment to the advanced disease (stage III or IV) category required the presence of local residual or
metastatic disease. When tumor weight was not provided, it was estimated from tumor measurements obtained from computed tomography scans, ultrasound examinations, or pathology reports by using the following formula: weight = 1.342 . (volume)0.8563 (Fig 1). In 30 patients, the tumor weight could not be determined because tumor volume was not available. Twelve of these 30 pa- tients had distant tumor metastasis (stage IV), seven patients had local residual tumor (stage III), and the remaining 11 patients were classified arbitrarily on the basis of the largest tumor diameter measured by computed tomography or ultrasound examinations, or data reported by the pathologist. Tumors for which the largest diameter was less than 6 cm were classified as stage I disease.
On the basis of clinical data, tumors were classified as functional (causing virilization, Cushing’s syndrome, Conn’s syndrome, feminization, or mixed effects) or nonfunctional (no clinical signs of hormone excess). Information on tumor spillage during surgery and the presence of intravenous tumor thrombus was retrospectively obtained for the subset of pa- tients treated in Curitiba, Brazil, directly from the patients’ medical records.
Treatment
Patients were treated according to the primary physician’s preference, usually in consultation with an IPACTR physician. In general, for localized disease, surgery alone was recommended by IPACTR physicians. Patients with residual disease after surgery or with metastatic disease (stage III or IV) received a variety of
2000
1800
y = 1.342x0.8563
1600
Tumor Weight (g)
1400
1200
1000
800
600
400
200
0
1000
2000
3000
4000
5000
Tumor Volume (cm3)
chemotherapy regimens, most of which included cisplatin, etopo- side, doxorubicin, and mitotane [1,1-dichloro-2-(O-chlorophe- nyl)-2-(p-chlorophenyl)-ethane]. For recurrent disease, complete resection was attempted when feasible and was followed almost always by chemotherapy. Radiation therapy rarely was used.
Statistical Analysis
Analysis was based on the status of the registered patients as of December 1, 2001. All data were analyzed using descrip- tive statistical methods, and the proportions of patients within each group of characteristics were compared using the x2 or Fisher’s exact test, when appropriate. Continuous variables were compared using the Mann-Whitney test. Complete remis- sion was defined as complete (macroscopic and microscopic) tumor resection, absence of distant metastasis, and normaliza- tion of all laboratory values.
Survival curves were created using Kaplan-Meier [9] product limit estimates, and groups were compared using the log-rank test. The duration of overall survival was measured from the date of diagnosis to the date of death as a result of any cause. The duration of event-free survival (EFS) was measured from the date of com- plete response to the date of either local or distant relapse or death as a result of any cause. Complete response was defined as absence of any evidence of disease, either after initial surgery (patients with stage I and II disease) or after chemotherapy and/or definitive surgery (patients with stage III or IV disease). Failure to enter remission was considered an event at zero time. Patients who were lost to follow-up and had active disease at the last visit were considered to have had an adverse event at the time of the last visit. Data from all other patients were censored at the date of the last follow-up. Cox regression analyses and hazard function for the risk of death and relapse were used to identify factors associated with EFS rates in patients with limited disease (disease stage I or II); the forward stepwise method (Wald) was used to analyze the relation of survival to the following variables present at diagnosis: age, sex, presence of a clinical endocrine syndrome, interval be- tween symptom onset and diagnosis, blood pressure, disease ex- tent (stage I or II), complete tumor resection, tumor weight, tumor spillage, and presence of intravascular thrombosis. The level of statistical significance was prospectively determined to be P < . 05. All analyses were performed by using SPSS software, version 10.1 (SPSS Inc, Chicago, IL).
RESULTS
Between January 1990 and December 2001, 259 patients were registered in the IPACTR. Five patients were excluded: four because a copy of the pathology report could not be obtained and one because of an incorrect diagnosis. Data on the remaining 254 patients were analyzed. Most of the patients (79.5%) were from southern Brazil. Thirteen percent were from the United States, and 7.5% were from nine other countries. Two-hundred twenty-eight tumors were histologically classified as carcinoma, and the re- maining 26 were classified as adenoma. The median in- terval between the first symptoms and diagnosis was 5 months (range, 0 to 90 months).
The median age of the 156 girls and 98 boys was 3.2 years (range, 0 to 19 years). Approximately 60% of patients
| Table 2. Selected Clinical Features of 254 Patients With Adrenocortical Tumors | ||
|---|---|---|
| Feature | Value | % |
| Age, years | ||
| < 4 | 153 | 59.8 |
| 4-12 | 65 | 30.0 |
| 13-20 | 36 | 10.2 |
| Sex | ||
| Female | 156 | 61.4 |
| Male | 98 | 38.6 |
| Clinical presentation | ||
| Virilization | 140 | 55.1 |
| Cushing's syndrome | 14 | 5.5 |
| Mixed* | 74 | 29.2 |
| Nonfunctional tumor | 26 | 10.2 |
| Female-male ratio, by age | ||
| < 4 years | 1.7:1 | |
| 4-12 years | 0.8:1 | |
| 13-20 years | 6.2:1 | |
| *Indicates clinical and/or laboratory evidence of abnormal production of more than one hormone, including aldosterone (Conn's syndrome) or estrogen (feminization). | ||
were younger than 4 years of age, and only 14% were older than 13 years of age. One child had a diagnosis of ACT during the neonatal period; another 18 patients had a his- tory compatible with onset of overproduction of adrenal hormones during the first few months of life. The overall female-male ratio was 1.6:1, but it varied widely across age groups. Girls predominated in the age groups 0 to 3 years (ratio, 1.7:1) and ≥ 13 years (ratio, 6.2:1). In the 4- to 12-year-old group, neither sex predominated (Table 2).
Approximately 90% of the children had clinical evi- dence of an endocrine syndrome. Virilization, alone or in combination with signs of overproduction of other adrenal hormones, was the most common clinical presentation (84.3%). Isolated Cushing’s syndrome was rare. Tumors that did not cause clinical manifestations of hormone excess (nonfunctional tumors) were observed in 10.2% of pa- tients. Isolated Cushing’s syndrome (median age, 12.6 years) and nonfunctional tumors (median age, 5.7 years) tended to occur in the older children (P < . 001; data not shown), without sex differences. Twelve patients were reg- istered as having increased secretion of aldosterone, either alone (Conn’s syndrome; n = 2) or together with overse- cretion of other adrenocortical hormones (n = 10). The latter 10 patients had increased levels of androgens, glu- cocorticoids, and/or mineralocorticoids and were included in the mixed subgroup for analysis, although not all patients received a comprehensive laboratory hormone profile (Ta- ble 2). About two thirds of the patients had limited disease: 112 patients (44.1%) had stage I disease and 80 patients (31.5%) had stage II disease. Among patients with advanced disease, 25 patients (9.8%) had stage III and 37 patients
1.0
0.8
Cumulative Survival
0.6
Overall Survival
Event-Free Survival
0.4
0.2
n = 254 patients
0
2
4
6
8
10
12
14
16
18
Survival Time in Years
(14.6%) had stage IV disease. Sites of metastasis were noted in 34 of these 37 patients. Eighteen patients had pulmonary metastasis, 11 patients had hepatic metastasis, and five pa- tients had involvement of both liver and lungs.
Tumor spillage during surgery was reported in 18 of 86 patients (21%) with stage I or II disease for which this information was available, whereas IV tumor thrombus was found in nine of 64 such patients (14%).
In addition to surgery, seven of 116 patients with stage I disease received adjuvant treatment (five patients, mito- tane; one patient, local radiotherapy; and one patient, com- bination chemotherapy). Of 80 patients with stage II dis- ease, 16 received adjuvant treatment (13 patients, mitotane; two patients, combination chemotherapy; and one patient, local radiation). All patients with stage III or IV disease were treated with intensive chemotherapy.
At a median follow-up of 2 years and 5 months (range, 5 days to 22 years), 157 patients (61.8%) remained alive and 97 patients (38.2%) had died. Death was associated with disease progression in 92 of these patients (including 22 patients who had active disease at the time of the last fol- low-up visit). Five patients died as a result of causes unre- lated to tumor progression (two died as a result of infection, one died as a result of hypertensive complication, one died as a result of massive hemorrhage during surgery, and one died as a result of an unspecified complication). The 5-year EFS and overall survival estimates were 54.2% (95% CI, 48.2% to 60.2%) and 54.7% (95% CI, 48.7% to 60.7%), respectively (Fig 2).
Several clinical features, including age (Fig 3), sex, clin- ical syndrome, interval between first symptoms and diag-
1.0
0.8
0-3 years (n=153)
Cumulative Survival
0.6
0.4
4-12 years (n=65)
13-20 years (n=36)
0.2
P <. 0001
0
2
4
6
8
10
Survival Time in Years
nosis, blood pressure, disease stage (Fig 4), tumor spillage during surgery, IV tumor thrombus, and tumor weight were examined for association with outcome. Because con- sistent histologic criteria were not used to classify pediatric adrenocortical tumors as benign (adenoma) or malignant (carcinoma), these data were not examined as prognostic factors. However, among 26 patients reported to have adre- nocortical adenoma, only one has experienced relapse.
In patients with localized disease, age between 0 and 3 years, virilization alone, normal blood pressure, disease
1.0
Stage I (n=112)
0.8
Cumulative Survival
0.6
0.4
Stage II (n=80)
Stage III (n=25)
0.2
Stage IV (n=37)
P <. 0001
0
2
4
6
8
10
Survival Time in Years
| Table 3. Relation of Estimated 2- and 5-Year EFS to Selected Features of Patients With Localized Disease (stage I and II) at Diagnosis | ||||||
|---|---|---|---|---|---|---|
| Feature | No. of Patients | 2-Year EFS | 5-Year EFS | P* | ||
| % | 95% CI | % | 95% CI | |||
| Age, years | 192 | <. 0001 | ||||
| 0-3 years | 133 | 88.0 | 82.0 to 94.0 | 85.6 | 79.0 to 92.2 | |
| 4-12 years | 40 | 64.5 | 47.9 to 81.1 | 59.9 | 42.1 to 77.7 | |
| 13-20 years | 19 | 38.1 | 11.1 to 65.1 | 38.1 | 11.1 to 65.1 | |
| Sex | 192 | NS | ||||
| Female | 117 | 77.2 | 68.8 to 85.6 | 75.5 | 66.7 to 84.3 | |
| Male | 75 | 82.2 | 73.0 to 91.4 | 78.1 | 67.7 to 88.5 | |
| Interval between first signs and diagnosis, months | 192 | NS | ||||
| 6 | 122 | 78.4 | 70.4 to 86.4 | 72.1 | 63.1 to 81.1 | |
| 12 | 38 | 85.6 | 73.8 to 97.4 | 85.6 | 73.8 to 97.4 | |
| 18 | 14 | 92.3 | 85.0 to 99.6 | 92.3 | 85.0 to 99.6 | |
| 24 | 8 | 68.5 | 49.9 to 87.1 | 68.5 | 49.9 to 87.1 | |
| > 24 | 10 | 77.7 | 50.1 to 91.5 | 77.7 | 50.1 to 91.5 | |
| Clinical syndrome | 192 | <. 0001 | ||||
| Virilizing | 116 | 89.4 | 83.4 to 95.4 | 88.0 | 81.6 to 94.4 | |
| Cushing's | 7 | 53.3 | 48.5 to 58.1 | 53.3 | 48.5 to 58.1 | |
| Mixed | 54 | 62.7 | 48.3 to 77.1 | 56.4 | 41.0 to 71.8 | |
| Nonfunctional | 15 | 66.6 | 42.4 to 90.8 | 66.6 | 42.4 to 90.8 | |
| Blood pressure | 192 | .04 | ||||
| Normal | 114 | 82.9 | 75.9 to 89.9 | 82.9 | 75.9 to 89.9 | |
| Hypertension | 78 | 76.8 | 66.2 to 87.4 | 66.0 | 53.2 to 78.8 | |
| Stage | 192 | <. 0001 | ||||
| I | 112 | 91.1 | 85.4 to 96.6 | 91.1 | 85.4 to 96.6 | |
| II | 80 | 60.0 | 47.8 to 72.2 | 52.7 | 39.1 to 66.3 | |
| Spillage during surgeryt | 86 | <. 0001 | ||||
| No spillage | 68 | 93.4 | 87.2 to 99.6 | 87.4 | 78.4 to 96.4 | |
| Spillage | 18 | 55.0 | 31.4 to 78.6 | 47.1 | 22.1 to 72.1 | |
| Intravenous tumor thrombust | 64 | NS | ||||
| No thrombus | 55 | 86.3 | 76.9 to 95.7 | 81.3 | 69.9 to 92.7 | |
| Thrombus | 9 | 53.3 | 19.3 to 87.3 | 53.3 | 19.3 to 87.3 | |
| Tumor weight, gt | 182 | <. 0001 | ||||
| ≤ 200 | 99 | 90.2 | 84.0 to 96.4 | 87.6 | 80.6 to 94.6 | |
| > 200 | 83 | 57.1 | 42.5 to 71.7 | 47.6 | 31.8 to 63.4 | |
Abbreviations: EFS, event-free survival; NS, not significant.
*Log-rank test.
tInformation available for a subset of the patients.
stage I, absence of spillage during surgery, and tumor weight ≤ 200 g were associated with a greater probability of EFS (Table 3). In a Cox regression model analysis, only disease stage I, virilization alone, and age 0 to 3 years were independently associated with a greater probability of EFS (Table 4). When patients with adenoma were excluded, the overall results of the Cox regression analysis did not change (data not shown).
| Regression Analysis (multivariate) in Patients With Localized Disease | |||
|---|---|---|---|
| Variable | Adjusted Odds Ratio | 95% CI | P |
| Stage I | 4.1 | 1.79 to 9.31 | <. 001 |
| Virilization alone | 3.7 | 1.73 to 7.7 | <. 001 |
| Age < 4 years | 2.6 | 1.21 to 5.48 | .01 |
Because of the small number and dismal outcome of patients with stage III or IV disease, prognostic factor anal- ysis was not performed in this group.
DISCUSSION
Our analysis of the IPACTR data suggests that pediatric ACTs are heterogeneous. Patients who have a diagnosis within the first 3 years of life tend to be girls, whereas an equal sex distribution is noted in the age interval 4 to 12 years. In adolescents, as in young children, there is a pre- dominance of females. This age-dependent sex distribution is also seen in the Surveillance Epidemiology End Results data [1]. Girls and boys in the 0- to 3-year and 4- to 12-year age groups did not differ in presenting clinical and labora- tory features or in constitutional p53 mutations [6], a find- ing that suggests that ACT in these two groups is the same
disease but manifests itself earlier in girls than in boys. However, the clinical presentation seen in the older group differed from that observed in the younger groups. Adoles- cents and young adults tended to have Cushing’s syndrome or nonfunctional tumors at presentation, whereas more than 90% of young children had virilizing features. More- over, constitutional mutations of the p53 gene are fre- quently reported in young children with a diagnosis of ACTs [10,11], but are relatively rare in adults [12,13]. These observations suggest that pediatric ACTs comprise at least two subtypes. In the young child, the early manifestations of the disease, the predominance of virilizing tumors, and the tumor’s histopathologic features [14] suggest that ACTs arise from the fetal zone of the fetal adrenal cortex. The fetal zone represents 85% of the adrenal cortex during fetal de- velopment and is oriented toward dehydroepiandrosterone production, as indicated by the fact that the fetal zone cells primarily express the enzyme 17a-hydroxylase/17,20 lyase (P450C17) [15]. In contrast, ACTs of adolescence and adulthood may originate from the definitive adrenal cortex.
It is plausible that the presence of a constitutional p53 mutation increases the penetrance of ACTs in the fetal adrenal cortex but not in the definitive adrenal cortex. This premise is consistent with the fact that in individuals with constitutional p53 mutations (Li-Fraumeni syndrome), ACTs usually occur only during the first decade of life [16]. Because of the strong association between constitutional p53 mutations and a diagnosis of ACTs at a young age, genetic testing of these patients for p53 status should be considered. To date, no comparison of gene expression in pediatric and adult adrenocortical tumors has been re- ported. The reason(s) for the predominance of female sex in children with ACTs is not known.
Approximately 12% of the patient cases with ACTs were diagnosed during the first year of life. One patient was diagnosed during the neonatal period. The clinical charac- teristics and outcome of the infants were similar to those of children aged 1 to 3 years at the time of diagnosis. There is a paucity of data on ACTs in the first few months of life. Satge et al [17] described one patient and reviewed the literature from 1900 to 1987. He found reports of 13 children with ACTs diagnosed during the neonatal period among 56 re- viewed cases of solid malignancies. Eight patients reported before 1980 died as a result of postsurgical complications. Three neonates received no treatment and died during the first few days of life. Three patients, including the one reported by Satge et al [17], survived surgery but experi- enced relapse. Two died and one was free of disease 14 years after diagnosis. The latter patient was treated with mitotane at the time of relapse. Other sporadic cases of neonatal ACTs have been reported [18-23]. Interestingly, Saracco et al [18] reported the case of a newborn with ACT who had complete spontaneous resolution of tumor, including all
metastatic lesions, 4 months after initial resection of the primary tumor.
Most patients enrolled onto the registry were from southern Brazil. This fact reflects the increased incidence of the disease in this area (15 times greater than that in other geographic regions) and the heightened interest of investi- gators in this region. A unique constitutional p53 mutation was recently identified in the Brazilian patients [6,24], al- though the reason for the frequency of this abnormality in southern Brazil remains unclear. Patients reported from other countries rarely have been tested for constitutional or somatic p53 abnormalities. At St Jude, among 11 patients with ACTs who were tested, five had a constitu- tional p53 mutation (unpublished observation), al- though none had the specific mutation that was identi- fied in the Brazilian patients.
These registry data confirmed reports [2,4,7] that the diagnosis of ACTs is usually delayed. The median interval between the first signs or symptoms of ACTs and the diag- nosis was 5 months (range, a few days to 8 years). The reasons for the delay are probably multifactorial. First, be- cause small tumors can produce striking clinical manifesta- tions, an extensive endocrine work-up may be required before the tumor is detected. Second, these children do not seem to be ill, except for their endocrine manifestations, and in fact, many experience accelerated growth and devel- opment. Finally, tumor growth can be slow, as shown by the time lag between parents’ reported initial awareness of the clinical endocrine changes and detection of the tumor. This latter observation is consistent with a multistep process of carcinogenesis resulting from accumulation of genetic changes [25]. It can be argued that retrospective parental accounts are not reliable, but in some cases, these accounts were confirmed by photographs of the patients showing signs of overproduction of adrenal cortical hormones sev- eral years before diagnosis. Moreover, in a few cases, pa- tients whose clinical and laboratory evaluation was consis- tent with increased production of adrenocortical hormones were lost to follow-up and then returned several years later, when the tumor was finally detected.
Tumor classification based solely on the predominant clinical signs and symptoms (virilization and/or Cushing’s syndrome) does not accurately reflect the tumor’s hor- monal secretory profile. For example, patients were rarely tested for aldosterone, renin, and estrogen levels. Therefore, many patients who were classified as having virilizing tu- mors might have been classified as having a mixed clinical and hormonal presentation if an extensive panel of adrenal hormones had been assayed.
Hypertension was observed at the time of diagnosis in 43% of patients and was distributed among all clinical syn- dromes. In a few patients, it was severe enough to cause encephalopathy, which resulted in the death of one patient. Hypertension in these patients is attributed in part to tumor
production of either glucocorticoids (in most patients) and/or mineralocorticoids. In fact, hypertension was signif- icantly (P < . 04) more common in patients presenting with Cushing’s syndrome features than in other patients. How- ever, 27% of patients with virilizing or nonfunctional tumors also had hypertension. In the latter patients, tumor compression of the renal artery or increased pro- duction of aldosterone may have contributed to hyper- tension. Regardless of the mechanism, hypertension re- solved after tumor resection.
Tumor size was useful in identifying patients at high risk of relapse after complete tumor resection. These data are consistent with the results of several studies [4,14,26-30] that found tumor size to be associated with prognosis. We estimated tumor weight, where it was not available, from tumor dimensions by using an equation derived from pa- tients in this cohort for which both tumor weight and tumor dimensions were available. This equation was appli- cable specifically to our patient group, and it should not be used in other settings. As in many other pediatric malignan- cies, tumor burden is only a surrogate marker of more fundamental cellular events. Genetic studies of pediatric ACTs have disclosed several abnormalities, but their rela- tion to prognosis is still unknown [31,32].
Interestingly, the interval between the first signs and symptoms and diagnosis was not associated with prognosis or tumor weight (data not shown); therefore, tumor size may reflect the quality of cumulative acquired genetic ab- normalities. Thus, it seems that there is an initial phase of slow tumor progression characterized mainly by the endo- crine manifestations in the absence of a palpable mass, and subsequent faster tumor progression. The types of still- unknown acquired genetic abnormalities are likely to deter- mine the tumor’s proliferative potential. This premise is consistent with observations that regardless of the interval between the first signs and symptoms and initial diagnosis, tumor growth and metastatic spread tend to occur rapidly in recurrent pediatric ACTs.
Age younger than 4 years was independently associated with a better prognosis in a multivariate analysis. Other investigators also have reported better survival rates in young children than in older age groups [27]. Similarly, virilization as the sole clinical manifestation was associated with better prognosis. Patients with Cushing’s syndrome as either the sole clinical manifestation or combined with other endocrine clinical syndromes (mixed), as well as those
with nonfunctional tumors, had worse prognosis. How- ever, these observations remain to be confirmed, given that many patients did not have an extensive laboratory evalua- tion for elevated levels of adrenal cortical hormones.
Disease staging criteria for childhood ACTs generally have not been agreed on. Most investigators use a classifi- cation scheme proposed by MacFarlane [33] and modified by Sullivan et al [34]. This classification, which is based on tumor size, lymph node involvement, and metastatic dis- ease, was developed primarily for adults. In our study, we used a modification of the disease staging system proposed by Sandrini et al [5,8] These modifications were based on the premise that complete tumor resection and tumor size are the most important factors for disease control. These criteria allowed the identification of three distinctive prog- nostic groups. Patients with completely resected small tu- mors (≤ 200 g) had an excellent prognosis, whereas patients with either microscopic or gross residual disease had a poor prognosis. Patients with completely resected large tumors had an intermediate prognosis. The prognostic accuracy of this classification probably can be improved by adding in- formation on histology, lymph node involvement, capsular rupture during surgery, and tumor biologic markers.
In summary, this analysis of the registry data con- firmed that pediatric ACTs are more common in girls and are almost always associated with increased production of adrenal androgens. It also revealed that pediatric ACTs are heterogeneous and comprise at least two subtypes with different pathogenetic mechanisms. Regardless of the type, complete resection is mandatory for definitive dis- ease control. Among patients who have had complete resection, tumor size is a strong prognostic indicator. In addition, this study demonstrates the feasibility of orga- nizing a disease-specific database as a means of obtaining meaningful clinical and outcome information about rare pediatric malignancies.
Appendix
The appendix is included in the full-text version of this article, available on-line at www.jco.org. It is not included in the PDF (via Adobe® Acrobat Reader®) version.
Authors’ Disclosures of Potential Conflicts of Interest
The authors indicated no potential conflicts of interest.
REFERENCES
1. Bernstein L, Gurney JG: Carcinomas and other malignant epithelial neoplasms, in Ries LAG, Smith MA, Gurney JG, et al (eds): Cancer Incidence and Survival Among Chil- dren and Adolescents: United States SEER Program 1975-1995. Bethesda, MD, National
Cancer Institute, SEER Program, 1999, pp 139- 147
2. Ribeiro RC, Michalkiewicz EL: Adrenocor- tical tumors in children, in Raghavan D, Brecher ML, Johnson DH, et al (eds): Textbook of Un- common Cancer (ed 2). New York, NY, John Wiley & Sons, 1999, pp 611-620
3. Lefevre M, Gerard-Marchant R, Chaussain JL, et al: Adrenal cortical carcinoma in children:
42 patients treated from 1958 to 1980 at Villejuif, in Humphrey GB, Grindey GB, Dehner LP, et al (eds): Adrenal and Endocrine Tumors in Children. Boston, MA, Martinus Nijhoff, 1983, pp 265-276
4. Ribeiro RC, Sandrini Neto RS, Schell MJ, et al: Adrenocortical carcinoma in chil- dren: A study of 40 cases. J Clin Oncol 8:67- 74, 1990
Childhood Adrenocortical Tumors
5. Sandrini R, Ribeiro RC, DeLacerda L: Childhood adrenocortical tumors. J Clin Endocri- nol Metab 82:2027-2031, 1997
6. Ribeiro RC, Sandrini F, Figueiredo B, et al: An inherited p53 mutation that contributes in a tissue-specific manner to pediatric adrenal corti- cal carcinoma. Proc Natl Acad Sci U S A 98:9330- 9335, 2001
7. Liou LS, Kay R: Adrenocortical carcinoma in children: Review and recent innovations. Urol Clin North Am 27:403-421, 2000
8. Sandrini R, De Lacerda L, Sampaio G, et al: Adrenocortical tumors in children: Relationship between disease stage and outcome. Pediatr Res 33:S22, 1993 (suppl 5)
9. Kaplan EL, Meier P: Nonparametric esti- mation from incomplete observations. J Am Stat Assoc 53:457-481, 1958
10. Wagner J, Portwine C, Rabin K, et al: High frequency of germline p53 mutations in child- hood adrenocortical cancer. J Natl Cancer Inst 86:1707-1710, 1994
11. Varley JM, McGown G, Thorncroft M, et al: Are there low-penetrance TP53 alleles? Evi- dence from childhood adrenocortical tumors. Am J Hum Genet 65:995-1006, 1999
12. Reincke M, Karl M, Travis WH, et al: p53 mutations in human adrenocortical neoplasms: Immunohistochemical and molecular studies. J Clin Endocrinol Metab 78:790-794, 1994
13. Stojadinovic A, Ghossein RA, Hoos A, et al: Adrenocortical carcinoma: Clinical, morpho- logic, and molecular characterization. J Clin On- col 20:941-950, 2002
14. Bugg MF, Ribeiro RC, Roberson PK, et al: Correlation of pathologic features with clinical outcome in pediatric adrenocortical neoplasia: A study of a Brazilian population-Brazilian Group for Treatment of Childhood Adrenocortical Tu- mors. Am J Clin Pathol 101:625-629, 1994
15. Keegan CE, Hammer GD: Recent insights into organogenesis of the adrenal cortex. Trends Endocrinol Metab 13:200-208, 2002
16. Kleihues P, Schauble B, Hausen A, et al: Tumors associated with p53 germline mutations: A synopsis of 91 families. Am J Pathol 150:1-13, 1997
17. Satge D, Philippe E, Ruppe M, et al: Les carcinomes neonatals: Revue de la litterature a propos d’un cas. Bull Cancer 75:373-384, 1988
18. Saracco S, Abramowsky C, Taylor S, et al: Spontaneously regressing adrenocortical carci- noma in a newborn: A case report with DNA ploidy analysis. Cancer 62:507-511, 1988
19. Godil MA, Atlas MP, Parker RI, et al: Metastatic congenital adrenocortical carcinoma: A case report with tumor remission at 3 1/2 years. J Clin Endocrinol Metab 85:3964-3967, 2000
20. Kakkar N, Vasishta RK, Lamba A, et al: Special feature: Pathological case of the month-Denouement and discussion: Congeni- tal adrenocortical carcinoma. Arch Pediatr Ado- lesc Med 154:1267-1268, 2000
21. Butler H, Bick R, Morrison S: Unsuspected adrenal masses in the neonate: Adrenal cortical carcinoma and neuroblastoma-A report of two cases. Pediatr Radiol 18:237-239, 1988
22. Artigas JLR, Niclewicz ED, Silva APG, et al: Congenital adrenal cortical carcinoma. J Pedi- atr Surg 11:247-252, 1976
23. Goldstein AE, Rubin SW, Askin JA: Carci- noma of adrenal cortex with adrenogenital syn- drome in children: Complete review of the liter- ature and report of a case with recovery in a child 8 months of age. Am J Dis Child 72:563-603, 1946
24. DiGiammarino EL, Lee AS, Cadwell C, et al: A novel mechanism of tumorigenesis involv-
ing pH-dependent destabilization of a mutant p53 tetramer. Nat Struct Biol 9:12-16, 2002
25. Shih IM, Wang TL, Traverso G, et al: Top-down morphogenesis of colorectal tumors. Proc Natl Acad Sci U S A 98:2640-2645, 2001
26. Michalkiewicz EL, Sandrini R, Bugg MF, et al: Clinical characteristics of small functional ad- renocortical tumors in children. Med Pediatr On- col 28:175-178, 1997
27. Humphrey GB, Pysher T, Holcombe J, et al: Overview on the management of adrenocor- tical carcinoma (ACC), in Humphrey GB, Grindey GB, Dehner LP, et al (eds): Adrenal and Endo- crine Tumors in Children. Boston, MA, Martinus Nijhoff, 1983, pp 349-358
28. Weiss LM, Medeiros LJ, Vickery AL Jr: Pathologic features of prognostic significance in adrenocortical carcinoma. Am J Surg Pathol 13: 202-206, 1989
29. Cagle PT, Hough AJ, Pysher J: Compari- son of adrenal cortical tumors in children and adults. Cancer 57:2235-2237, 1986
30. Wieneke JA, Thompson LD, Heffess CS: Adrenal cortical neoplasms in the pediatric pop- ulation: A clinicopathologic and immunopheno- typic analysis of 83 patients. Am J Surg Pathol 27:867-881, 2003
31. Figueiredo BC, Stratakis CA, Sandrini R, et al: Comparative genomic hybridization analysis of adrenocortical tumors of childhood. J Clin Endocrinol Metab 84:1116-1121, 1999
32. James LA, Kelsey AM, Birch JM, et al: Highly consistent genetic alterations in childhood adrenocortical tumours detected by comparative genomic hybridization. Br J Cancer 81:300-304, 1999
33. McFarlane DA: Cancer of adrenal cortex. Ann R Coll Surg Engl 23:155-186, 1958
34. Sullivan M, Boileau M, Hodges CV: Adre- nal cortical carcinoma. J Urol 120:660-665, 1978