Taylor & Francis Taylor & Francis Group
Pediatric Hematology and Oncology CAtrois Foodit Comes o-to 0 CATHERINE COLLARD IDRICO O TOTO
informa
ISSN: 0888-0018 (Print) 1521-0669 (Online) Journal homepage: http://www.tandfonline.com/loi/ipho20
Recurrence and mortality prognostic factors in childhood adrenocortical tumors: Analysis from the Brazilian National Institute of Cancer experience
Daniel Bulzico, Paulo Antônio Silvestre de Faria, Marcela Pessoa de Paula, Maria Alice Neves Bordallo, Cencita H. C. N. Pessoa, Rossana Corbo, Sima Ferman, Mario Vaisman & Leonardo Vieira Neto
To cite this article: Daniel Bulzico, Paulo Antônio Silvestre de Faria, Marcela Pessoa de Paula, Maria Alice Neves Bordallo, Cencita H. C. N. Pessoa, Rossana Corbo, Sima Ferman, Mario Vaisman & Leonardo Vieira Neto (2016): Recurrence and mortality prognostic factors in childhood adrenocortical tumors: Analysis from the Brazilian National Institute of Cancer experience, Pediatric Hematology and Oncology, DOI: 10.3109/08880018.2016.1173148
To link to this article: http://dx.doi.org/10.3109/08880018.2016.1173148
Published online: 31 May 2016.
Submit your article to this journal
Article views: 14
View related articles
View Crossmark data
CrossMark
Taylor & Francis Taylor & Francis Group
Recurrence and mortality prognostic factors in childhood adrenocortical tumors: Analysis from the Brazilian National Institute of Cancer experience
Daniel Bulzicoa,c, Paulo Antônio Silvestre de Fariab, Marcela Pessoa de Paulac, Maria Alice Neves Bordalloª, Cencita H. C. N. Pessoaª, Rossana Corboa,d, Sima Fermane, Mario Vaismanf, and Leonardo Vieira Netoc,f
aEndocrine Oncology Unit, Brazilian National Institute of Cancer-INCA, Rio de Janeiro, Rio de Janeiro, Brazil; bDivision of Pathology, Brazilian National Institute of Cancer-INCA, Rio de Janeiro, Rio de Janeiro, Brazil; “Endocrinology Section, Federal Hospital of Lagoa, Rio de Janeiro, Rio de Janeiro, Brazil; dNuclear Medicine Section, Medical School and Clementino Fraga Filho University Hospital, Rio de Janeiro Federal University, Rio de Janeiro, Rio de Janeiro, Brazil; eDepartment of Pediatric Oncology, Brazilian National Institute of Cancer-INCA, Rio de Janeiro, Rio de Janeiro, Brazil; ‘Department of Internal Medicine and Endocrinology Section, Medical School and Clementino Fraga Filho University Hospital, Rio de Janeiro Federal University, Rio de Janeiro, Rio de Janeiro, Brazil
ABSTRACT
Prognostic markers that can help identifying precocious risk of unfa- vorable outcomes in patients with childhood adrenocortical tumors (ACTs) are still unclear. This observational and retrospective study aimed to identify clinical and pathology prognostic factors of recurrence and death in a tertiary cancer center population. Clinical, pathology, demo- graphic, staging, and therapy data from patients with childhood ACT (median age: 3.6 years) treated at the Brazilian National Institute of Cancer between 1997 and 2015 were assessed. Univariate and bivariate analyses were used to study the association of clinical and pathology characteristics with recurrence and mortality. Recurrence and disease- related mortality were the main outcomes. Twenty-seven patients were included. Complete tumor resection was performed in 21 cases. The median tumor size was 8.2 cm. Mitotane was the most common adju- vant/palliative therapy (n = 13). Recurrence occurred in 6 patients, after a median time of 7.2 months, and was more common among those with larger tumors (P = . 008), higher Weiss score (P = . 001), and microscopic tumoral necrosis (P =. 002). Ten patients died from the disease. Older age (P = . 04), larger tumor size (P = . 002), metastatic disease (P = . 003), pre- vious recurrence (P = . 003), incomplete resection (P = . 002), intraopera- tive tumor spillage (P = . 005), higher Weiss score (P = . 03), microscopic necrosis (P = . 005), and capsular invasion (P = . 02) were all associated with increased death risk. Even though complete tumor resection was performed in most cases, a considerable number of cases of childhood ACT resulted in recurrence and death. Early identification of unfavor- able outcomes is essential to determine ideal therapy and appropriate surveillance.
ARTICLE HISTORY
Received 4 March 2016 Revised 27 March 2016 Accepted 29 March 2016
KEYWORDS
Adrenocortical carcinoma; cancer staging; pediatric adrenal tumors; prognosis
Endocrine Oncology Unit, Brazilian National Institute of Cancer-INCA, Praça da Cruz Vermelha, 23, 8°Floor, Rio de Janeiro/RJ, Brazil, CEP 20.230-130. @ 2016 Taylor & Francis Group, LLC
Introduction
Adrenocortical tumors (ACTs) are common in the general population, with the vast major- ity of them represented by benign adrenocortical adenomas [1]. In contrast, adrenocortical carcinomas (ACCs) are extremely rare, with incidence peaks occurring during first and fifth decades [2, 3]. Interestingly, the incidence of childhood ACTs in southern Brazil is almost 15-fold higher than the global incidence. The reason of this discrepancy is not fully under- stood, but a TP53 germline mutation is present in more than 90% of these cases [4].
Except for the presence of distant metastasis, no consensus on the histological diagnosis of ACC has been defined for the pediatric population. Although tumor size and volume are related to poorer prognosis, no isolated histological characteristic or cluster of microscopic features is able to define the malignant potential in the pediatric population, differently from what is observed in adulthood [5, 6]. Therefore, terms such as “adrenocortical tumors” or “adrenocortical neoplasia” are currently preferable instead of adenoma or carcinoma in pedi- atric patients.
Complete tumor resection is the only intervention able to cure ACC [7, 8]. Mitotane (o,p- dichlorodiphenyldichloroethane) alone or in combination with chemotherapy is formally indicated for stage III and IV cases, but its true efficacy is undetermined because random- ized controlled trials are lacking [3, 8, 9].
Most ACC cases are very aggressive, and since a significant proportion of patients will present with advanced disease, recurrence and mortality are very common [7]. Therefore, prognostic markers are pivotal in order to identify precocious risk of recurrence and death.
This study aimed to describe clinical and pathology prognostic factors for recurrence and mortality in a population of childhood ACT cases treated at a single oncology tertiary insti- tution between 1997 and 2015.
Materials and methods
Study design and population
This observational and retrospective study was conducted by the Endocrine Oncology Unit in collaboration with the Division of Pathology (DIPAT) and the Department of Pediatric Oncology from the Brazilian National Institute of Cancer-INCA, Rio de Janeiro, Brazil. Cases were identified from the DIPAT laboratory information system, which has been coding all cases since 1997. For this study, after filtering for adrenal cortex localization and age below 18 years at diagnosis, all cases registered from January 1, 1997, and March 31, 2015, were considered eligible. Exclusion criterion was evidence of histological diagnosis other than ACT.
Ethical aspects
This study was approved by INCA’s independent institutional advisory committee in September 24, 2014 (protocol 33847514.4.0000.5274), and conducted according to the prin- ciples expressed in the Declaration of Helsinki. All involved subjects or their legal guardians provided informed consent before entering the study.
Procedures
Pathology diagnosis review
In order to confirm the ACT diagnosis, original hematoxylin-eosin-stained slides from resected or biopsied cases were initially reviewed by the same 2 investigators (D.B., P.A.S.F.).
After confirmation, the presence of each of the following 9 histological features were evalu- ated and the Weiss score [10] calculated: nuclear atypia (grade III/IV according to Fuhrman criteria [11]); increased mitotic rate (more than 5 mitotic figures in 50 high-power fields [40x objective]); atypical mitotic figure; eosinophilic (“dark”) cytoplasm in at least 75% of tumor cells; diffuse pattern in tumor architecture; microscopic tumor necrosis; unequivocal venous invasion; sinusoidal invasion; and tumor capsule invasion. Cases only submitted to biopsy were excluded from Weiss score analysis.
Clinical data review
Medical records from all patients included were reviewed, and a specific questionnaire was completed with demographic information, disease presentation characteristics, and data related to therapy and follow-up. The following variables were registered: (a) sex; (b) age at diagnosis; (c) obesity (defined as a body mass index above the sex- and age-matched 95th percentile); (d) first-degree relative cancer history; (e) initial clinical symptoms; (f) time until diagnosis (defined as the interval between initial symptoms and first medical evaluation); (g) evidence of endocrine syndrome defined by an unequivocal description of signs or symp- toms compatible with autonomous hormone secretion, and/or by evidence of age-specific elevated levels of serum total testosterone, androstenedione, or dehydroepiandrosteronesul- fate (DHEA-S) or laboratorial evidence of hypercortisolism (impaired dexamethasone low- dose suppression test or elevated late-night salivary cortisol); (h) tumor’s largest diameter and fibroid volume (a x b x cx 0.523); (i) localization (laterality); (j) computed tomography or magnetic resonance evidence of para-aortic lymphatic and/or distant metastatic involvement; (k) disease’s stage according to Sandrini et al. [6]; (1) initial treatment; (m) adjuvant or pal- liative therapies; (n) performed procedure (complete or incomplete tumor resection, or only biopsy); (o) report of tumor spillage during surgery; and (p) radiotherapy.
Outcomes
Recurrence was defined as evidence of a new local or distant expansive mass in a previously considered free-of-disease patient. The time until recurrence (defined as the interval between complete tumor resection and image confirmed relapse), disease-specific mortality, and over- all survival were assessed. Only cases with at least 12 months of follow-up were included for recurrence analysis.
Statistical analysis
Analyses were performed using SPSS version 20.0 for Macintosh (IBM, Armonk, NY, USA). In descriptive analysis, categorical variables were expressed as percentages, whereas numerical variables were expressed as mean ± standard deviation (SD) and/or median (minimum and maximum). The Kolmogorov-Smirnov test was used to evaluate whether numeric variables were normally distributed. The Kruskal-Wallis test was used to compare numerical variables among 3 or 4 groups and the Mann-Whitney test was performed for comparison between 2 groups. Means of normally distributed variables were compared using Student t test. Chi-square or Fisher exact tests were used to evaluate categorical variables. A P value <. 05 was considered statistically significant, except for comparisons among more than 2 groups, when P values <. 017 (3 groups) or <. 013 (4 groups) were considered significant (Bonferroni post hoc analysis). Sensitivity, specificity, positive predictive value, negative predictive value, accuracy, and area under the curve from receiver operating characteristic (ROC) curves of
| Clinical characteristic | Value |
|---|---|
| Age, years | |
| Mean (SD) | 5.1 (4.1) |
| Median (min-max) | 3.6 (1.1-15.6) |
| Female, n (%) | 19 (70.4%) |
| Obesity, n (%) (N = 25) | 13 (52%) |
| Initial clinical presentation, n (%) | |
| Precocious puberty/virilization | 19 (70.4%) |
| Hypercortisolism | 7 (25.9%) |
| Abdominal pain | 1 (3.7%) |
| Malignancy family history*, n (%) (N = 24) | 15 (62.5%) |
| Time between symptoms and diagnosis, months | |
| Mean (SD) | 7.5 (8.1) |
| Median (min-max) | 5 (2-38) |
| Tumor localization, n (%) | |
| Right adrenal | 15 (55.6%) |
| Left adrenal | 12 (44.4%) |
| Tumor largest diameter, cm (N = 24) | |
| Median (min-max) | 8.2 (3-17) |
| Tumor volume, cm3 (N = 23) | |
| Median (min-max) | 105 (8.8-935) |
| Lymph node involvement, n (%) | 5 (18.5%) |
| Distant metastasis, n (%) | 5 (18.5%) |
| Surgery performed, n (%) | |
| Complete resection | 21 (77.8%) |
| Macroscopic residual tumor | 3 (11.1%) |
| Biopsy | 3 (11.1%) |
| Adjuvant and palliative therapy, n (%) | |
| Mitotane alone | 5 (18.5%) |
| Mitotane + cytotoxic chemotherapy | 8 (30.8%) |
| Cytotoxic chemotherapy alone | 0 |
| Radiotherapy | 0 |
* Among first-degree relatives.
variables associated with recurrence and mortality were calculated. Kaplan-Meier survival analysis from the entire population according to pediatric adrenocortical staging system pub- lished by Sandrini et al. [6] was carried out. Data were analyzed up to December 1, 2015.
Results
Population characteristics
Twenty-seven children with adrenocortical tumors were included. No case of misdiagnosed tumor (adrenal tumor other than adrenocortical) was identified. The median age at diagno- sis was 3.6 years (range: 1.1-15.6 years), with a median follow-up of 27 months (range: 1- 154 months). Twenty-four patients were submitted to complete or incomplete tumor removal without prior biopsy. In the remaining 3 cases, surgery was not possible because of extensive local invasion, and only biopsy was performed. Data on clinical and tumor characteristics are summarized in Table 1.
The staging system proposed by Sandrini et al. [6] was applied to all 27 cases and Weiss score of pathology features could be reviewed in all surgically treated cases (n = 22). A signif- icantly lower Weiss score was observed in stage I in comparison with stage II (P = . 004) and stage III P = . 008) patients (Table 2). Analyses on tumor size and time between initial symp- toms and diagnosis revealed no difference between groups. Stage III and stage IV patients
| Stage | n (%) | Weiss score Median (min-max) | Tumor size, cm Mean (SD) | Time until diagnosis, months Median (min-max) Mean (SD) |
|---|---|---|---|---|
| Stage I | 11 (40.7%) | 5 (2-5) | 4.5 (1.7) | 5 (2-11)5.6 (3.0) |
| Stage II | 5 (18.5%) | 7 (5-8) | 9.5 (4.4) | 4 (2-13)6.0 (4.7) |
| Stage III | 6 (22.2%) | 7.5 (4-9) | 8.9 (4.0) | 7 (2-28)9.7 (9.5) |
| Stage IV | 5 (18.5%) | 6.5 (5,8)# | 14.5 (4.4) | 3 (2-38)10.6 (15.6) |
* Tumor staging according to Sandrini et al. [6].
# Weiss score was not assessed in 3 cases only submitted to biopsy.
were more likely to be referred either to mitotane therapy alone (2 cases each), or in combina- tion with cytotoxic chemotherapy (4 and 3 cases, respectively). Two stage II patients received mitotane adjuvant therapy because of inconclusive vascular invasion at diagnosis. No stage I patient received adjuvant therapy. No patient received external beam radiotherapy.
Recurrence prognostic factors
Cases only submitted to biopsy (n = 3) or with macroscopic residual tumor after initial surgery (n = 3) were excluded from recurrence analysis. Recurrence occurred in 6 patients, after a mean time of 8.3 ± 6.7 months (median: 7.2 months [range: 1.6-16.5 months]). Five cases presented with distant recurrence (3 with lung and 2 with liver metastasis) and 1 with locoregional disease. Two patients with lung metastasis recurrence were submitted to metasta- sectomy plus mitotane as rescue therapy. Both were currently considered free from disease 5.5 and 9.2 years after therapy, respectively. The remaining 4 cases died after a mean survival time of 6.8 ± 6.7 months (median: 5.2 months [range: 0.6-16.5 months]). Recurrence occurred in 2 stage II and 4 stage III cases. No recurrence was observed in stage I patients. Those with larger and palpable tumors, microscopic tumor necrosis, and higher Weiss score were sig- nificantly more likely to suffer recurrence (Tables 3 and 4). Cases with microscopic vascular invasion or in which tumor spillage occurred during surgery had a trend to higher recurrence rate. Recurrence was significantly higher among mitotane users with more advanced disease. No other demographic, tumor, or therapy-related features were related to increased recur- rence risk. No analysis of the prognostic factor of tumor functional status was performed, since 26 out of 27 cases had endocrine syndrome.
Mortality prognostic factors
Ten patients (37%) died from disease during follow-up, with a mean survival time of 9.8 ± 7.8 months (median: 8.3 months [range: 1-21.4 months]). Among clinical and pathol- ogy features, those significantly associated with mortality risk were older age, palpable abdominal disease, tumor size, preoperative evidence of lymphatic involvement, metastatic disease, incomplete tumor resection, tumor spillage during surgery, previous recurrence, microscopic tumor necrosis and capsular invasion, and Weiss score (Tables 3 and 4). Deaths were significantly more common among mitotane-treated subjects, who also harbored more advanced disease stage. Figure 1 describes Kaplan-Meier survival rate according to staging status.
| Factor | N | Recurrence | No recurrence | P | N | Nonsurvivors | Survivors | P |
|---|---|---|---|---|---|---|---|---|
| n | 21* | 6 | 15 | - | 27 | 10 | 17 | - |
| Age, years | 21 | 27 | ||||||
| Mean (SD) | 6.1 (3.4) | 3.8 (4.0) | .24 | 7.1 (3.9) | 3.8 (3.8) | .04 | ||
| Median (min-max) | 5.2 (2.5-12) | 2.4 (1.1-15.6) | 6.7 (2.1-12.7) | 2.6 (1.1-15) | ||||
| Gender, n (%) | 21 | 27 | ||||||
| Female | 4 (66.7%) | 11 (73.3%) | 1.00 | 7 (70%) | 12 (70.6%) | 1.00 | ||
| Male | 2 (33.3%) | 4 (26.7%) | 3 (30%) | 5 (29.4%) | ||||
| Tumor localization, n (%) | 21 | .36 | 27 | |||||
| Left adrenal | 4 (66.7%) | 6 (40%) | 5 (50%) | 7 (41.2%) | .7 | |||
| Right adrenal | 2 (33.3%) | 9 (60%) | 5 (50%) | 10 (58.8%) | ||||
| Obesity, n (%) | 20 | 2 (40%) | 8 (53.3%) | 1.0 | 25 | 5 (62.5%) | 3 (37.5%) | .67 |
| Palpable tumor, n (%) | 14 | 4 (100%) | 2 (20%) | .01 | 18 | 7 (100%) | 3 (27.3%) | .004 |
| Positive family history for cancer, n (%) | 19 | 4 (100%) | 8 (53.3%) | .24 | 24 | 6 (75%) | 9 (56.2%) | .65 |
| Time between symptoms and diagnosis, months | 21 | 27 | ||||||
| Mean (SD) | 7.6 (10.0) | 6.5 (3.7) | .72 | 9.4 (12.8) | 6.4 (3.5) | .48 | ||
| Para-aortic lymphatic | 21 | 0 | 1 (6.7%) | 1.00 | 27 | 4 (40%) | 1 (5.9%) | .04 |
| involvement, n (%) | ||||||||
| Distant metastasis, n (%) | - | - | - | - | 27 | 5 (50%) | 0 (0%) | .003 |
| Liver metastasis, n (%) | - | - | - | - | 23 | 5 (50%) | 0 (0%) | .007 |
| Lung metastasis, n (%) | - | - | - | - | 23 | 3 (30%) | 0 (0%) | .06 |
| Total testosterone x URR | 19 | 23 | ||||||
| Mean (SD) | 14.6 (10.7) | 13.9 (15.7) | .93 | 15.3 (13.3) | 14.1 (14.6) | .84 | ||
| Incomplete tumor | - | - | - | - | 26 | 5 (55.6%) | 0 (0%) | .002 |
| resection, n (%) | ||||||||
| Tumor spillage during | 13 | 2 (50%) | 0 | .07 | 16 | 5 (71.4%) | 0 (0%) | .005 |
| surgery, n (%) | ||||||||
| Recurrence, n (%) | - | - | - | - | 21 | 4 (100%) | 2 (11.8%) | .003 |
| Mitotane use, n (%) | 21 | 6 (100%) | 1 (6.7%) | .001 | 27 | 10 (100%) | 3 (17.6%) | .001 |
| Cytotoxic chemotherapy, | 20 | 4 (66.7%) | 1 (7.1%) | .01 | 26 | 5 (50%) | 3 (18.8%) | .18 |
| n (%) |
Note. URR = upper reference range.
* Number of cases in which recurrence end point could be assessed.
| Factor | N | Recurrence | No recurrence | P | N | Nonsurvivors | Survivors | P |
|---|---|---|---|---|---|---|---|---|
| Weiss score | 21 | 22 | ||||||
| Mean (SD) | 7.5 (1.0) | 4.8 (1.5) | .001 | 7.0 (1.8) | 5.0 (1.5) | .03 | ||
| Pathology feature, n (%) | 21 | 24 | ||||||
| Nuclear grade Fuhrman | 6 (100%) | 13 (86.7%) | 1.00 | 7 (100%) | 15 (88.2%) | 1.00 | ||
| III/IV | ||||||||
| Mitotic rate >5/50 HPF | 6 (100%) | 9 (60%) | .12 | 7 (100%) | 11 (64.7%) | .13 | ||
| Abnormal mitoses | 6 (100%) | 12 (80%) | .52 | 7 (100%) | 14 (82.4%) | .53 | ||
| ≤25% clear cells | 6 (100%) | 10 (66.7%) | .26 | 5 (71.4%) | 12 (70.6%) | 1.00 | ||
| >1/3 diffuse | 5 (83.3%) | 15 (100%) | .28 | 5 (71.4%) | 16 (94.1%) | .19 | ||
| architecture | ||||||||
| Necrosis | 6 (100%) | 3 (20%) | .002 | 7 (100%) | 5 (29.4%) | .005 | ||
| Vascular invasion | 5 (83.3%) | 5 (33.3%) | .06 | 5 (71.4%) | 6 (35.3%) | .18 | ||
| Sinusoid invasion | 2 (33.3%) | 3 (20%) | .59 | 1 (14.3%) | 4 (23.5%) | 1.00 | ||
| Capsular invasion | 3 (50%) | 3 (20%) | .29 | 5 (71.4%) | 3 (17.6%) | .02 | ||
| Tumor size, cm | 21 | 24 | ||||||
| Mean (SD) | 11.4 (2.7) | 6.4 (3.6) | .008 | 12.6 (3.2) | 6.9 (3.7) | .002 | ||
| Tumor volume, cm3 | 20 | 20 | ||||||
| Mean (SD) | 518 (347) | 92 (166) | .02 | 659 (344) | 134 (202) | .001 |
Note. HPF = high-power fields.
Cumulative Survival (%)
100
Stage I (n=11)
80
60
Stage II (n=5)
40
Stage III (n=6)
20
00
Stage IV (n=5)
Log Rank < 0.001
0
24
48
72
96
120
144
Time (months)
| Recurrence/Mortality | AUC | Sensitivity | Specificity | PPV | NPV | Accuracy |
|---|---|---|---|---|---|---|
| Recurrence | ||||||
| Palpable tumor | - | 100% | 80% | 66% | 100% | 85% |
| Tumor spillage during | - | 50% | 100% | 100% | 82% | 84% |
| surgery | ||||||
| Microscopic vascular | - | 83% | 67% | 50% | 91% | 71% |
| invasion | ||||||
| Microscopic necrosis | - | 100% | 80% | 67% | 100% | 85% |
| Weiss score | 0.928 | - | - | - | - | - |
| Tumor size | 0.872 | - | - | - | - | - |
| Tumor volume | 0.967 | - | - | - | - | - |
| Mortality | ||||||
| Palpable tumor | - | 100% | 72.7% | 70% | 100% | 83% |
| Para-aortic lymphatic | - | 40% | 94% | 80% | 72.7% | 74% |
| involvement | ||||||
| Distant metastasis | - | 50% | 100% | 100% | 77.3% | 81% |
| Liver metastasis | - | 50% | 100% | 100% | 72% | 78% |
| Lung metastasis | - | 30% | 100% | 100% | 65% | 69% |
| Incomplete tumor | - | 56% | 100% | 100% | 81% | 84% |
| resection | ||||||
| Recurrence | - | 100% | 88% | 67% | 100% | 90% |
Note. AUC = area under the curve from receiver operating characteristic (ROC) curves; PPV = positive predictive value; NPV = negative predictive value.
Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy related to recurrence and mortality were calculated for those variables associ- ated with these end points (Table 5). The presence of palpable tumor and microscopic tumor necrosis were the most sensitive prognostic factors, with a NPV of 100% for both outcomes. The presence of distant metastasis and incomplete tumor resection were additional factors with high specificity in mortality risk analysis. Table 5 also describes the area under receiver operating characteristic (ROC) curve for variables significantly related to recurrence and mortality.
Discussion
The current study describes our experience over the institute’s last 18 years of the treatment of childhood ACTs. Although very rare, our study was able to assess 27 ACT cases. As a national
tertiary cancer center, we assume that virtually most cases of ACT during childhood in our state are referred to us. It is noteworthy that even though the elevated incidence of childhood ACTs in southern Brazil is well described, no such data are available from other regions inside the country [8, 12].
To distinguish benign from malignant ACTs in children is still a challenge because the pathology parameters of malignancy are not consistent. Although useful in segregating car- cinomas from adenomas within adults, the Weiss score [10], a 9-histomorphological feature score, has already been applied to children with ACTs with disappointing results [13, 14]. Among adults, a total score of 3 or more points has been associated with malignant tumor behavior, higher risk of metastasis, and unfavorable prognosis [10]. Faria and Almeida [5] studied 39 cases of childhood ACT and found no difference on survival rate according to adult score threshold. In addition, Mendonça et al. [15] described Weiss score as an unreliable out- come predictor in pediatric ACT. Different score systems applying pathology features to pedi- atric ACTs also failed to determine a single threshold capable of defining malignancy [16, 17]. In our study, all specimens were currently reviewed in order to uniform pathology character- istics assessment and apply the Weiss score. Although no significant difference between stage II, III, and IV patients was observed, those who experienced recurrence and/or nonsurvivors had a significantly higher Weiss score. Interestingly, patients who did not suffer from recur- rence or die from the disease presented relatively elevated mean Weiss scores, confirming that the adulthood 3-point threshold for outcomes should not be used as a prognostic parameter in children and adolescents. Nevertheless, isolated microscopic features might play a role as prognostic factors. Among microscopic findings, necrosis had a significant association with recurrence risk with accuracy of 85%. Similarly, in mortality risk evaluation, necrosis and capsular invasion predicted death with accuracy of 79% and 77%, respectively. We should mention that despite suitable negative predictive values, their positive predicted values were quite disappointing for both outcomes.
With a median age of 3.6 years at diagnosis and a predominance of females, our group presented similar demographic characteristics to those described for a 254-Brazilian patient cohort from Michalkiewicz et al. [8]. The prognostic effect of age in survival is not a consen- sus. Whereas Gulack et al. [18], Michalkiewicz et al. [8], and Sabbaga et al. [19] reported best survival rates in children under 4, 3, and 2 years of age, respectively, Sbragia et al. [20] and Hanna el al. [7] found no impact of age in mortality risk on their series. In our study, no asso- ciation between age and recurrence risk was evident, but survivors were significantly younger at diagnosis than nonsurvivors.
As observed in different ethnic populations worldwide [7, 8, 21, 22], the vast majority of our patients presented with autonomous hormonal hypersecretion, with only 1 case searching for medical assistance for abdominal pain. Therefore, any analysis on the relationship between hormonal status and outcomes would be impaired. Androgen excess was the most common hormonal pattern, in contrast to the usual presentation of adult functional tumors, which is most often represented by cortisol hypersecretion [23]. In our series, 25.9% of patients presented with overt Cushing’s syndrome, whereas more than a half of them were obese. Considering that diagnosis of subtle cortisol autonomous secretion in the pediatric popu- lation is still a challenge, we hypothesize that some tumors originally defined as exclusive androgen secreting might actually be underdiagnosed androgen-cortisol-cosecreting ones. This could have contributed to a high prevalence of obesity in our series.
Although 95% of our patients had precocious puberty and/or virilization, which could con- cur to early diagnosis, 40.7% of them had advanced disease at diagnosis. This fact may be
either attributed to the aggressive and rapidly progressive tumor behavior or the delayed time until diagnosis, which had a median value of 5 months.
The staging system proposed by Sandrini et al. [6] was developed upon the pioneer experi- ence on the management of 73 children and adolescents with ACT at a single center. The sys- tem is based on clinical and pathology features such as tumor size, extension of resection, hor- monal status after surgery, and presence of distant metastasis. When applied to our patients, consistent reproducibility of the original mortality risk cohort was observed. Whereas stage IV patients had a cumulative survival of 40% after a year, for stage I patients this figure was 60%. Both lymphatic and distant metastases at diagnosis were independent strong predictors for mortality in our study. Other staging systems, well established in adults, such as ENSAT (European Network for the Study of Adrenal Tumors) and UICC (Union for International Cancer Control) did not have the same mortality prediction power when applied to our cases and therefore were not used for further analysis (data not shown) [24, 25].
Tumor size is consistently the most evidence-based unfavorable outcome prognosticator in cases of childhood ACT. Tumor size, volume, or weight threshold values for poor outcomes have been published [6, 8, 9, 12, 13, 15-18, 20-22]. In our series, patients with large palpable tumors were significantly more likely to suffer recurrence and die from the disease. Nonsur- vivors had a mean tumor size of 12.6 cm, whereas for survivors the mean size was 6.9 cm. Our data are in accordance to those published by Wieneke et al. [17], who described a worse prognosis in cases with tumors larger than 10.5 cm.
Regarding treatment, total tumor resection was achieved in most of our patients (77.8%). Surgery is the cornerstone of ACT treatment, with the extent of tumor resection clearly being related to outcomes [6, 7, 12, 18, 22, 26]. In 3 cases (11.1%), only biopsy was performed because of extensive local invasion. In the remaining 3 other cases (11.1%), macroscopic resid- ual tumor was evident at the end of the procedure, although complete surgical removal was attempted. Incomplete tumor resection was significantly associated with death risk in our series. Patients who had tumor spillage during surgery, which might increase the risk of malig- nant cells seeding in abdominal cavity, presented with a significant higher mortality risk.
Traditionally, mitotane, a cytotoxic adrenolytic drug with an incompletely understood mechanism of action, is advocated for advanced stages with residual tumor after surgery, dis- tant metastasis, or inoperable local disease [27]. Most series describe mitotane use alone or in combination with different conventional chemotherapy schemes, preferably doxorubicin, etoposide, and cisplatin [22, 28, 29]. The benefit of mitotane is less clear in the pediatric pop- ulation [30, 31]. Even though an apparent improvement in hormonal status is well described, limited impact on overall survival has been observed [7, 8, 22]. Moreover, as randomized trials with mitotane and chemotherapy are lacking, issues such as the benefit of therapy in early- stage tumors, length of therapy, and best combination of drugs are yet to be determined. In our series, recurrence and mortality were significantly higher among mitotane users. This could be explained by the fact that almost all cases treated with adjuvant or palliative thera- pies were stage III or IV. Only 2 low-stage patients in whom there was doubt about venous invasion after surgery received mitotane. No analysis of mitotane impact on time to disease progression within drug users could be performed because of the limited number of events. No patient in our series was submitted to radiation therapy.
Our study has some strengths and limitations. The main strength is that all patients were treated at the same institution, by the same multidisciplinary team, thereby constituting an intact cohort. We consider that as a rare disease, the number of subjects included was suffi- cient to provide good quality information. Furthermore, data from pathology features of all
cases in this study were reassessed by the same adrenal disease expert observers in the light of updated literature regarding ACT diagnostic limitations. Consequently, any bias due to inexperienced pathologists could be ruled out. The most important limitation of this work is its retrospective design. Because patients’ clinical information was retrieved from medical records, missing data might have limited the statistical power in the analysis of some vari- ables. Unfortunately, genetic testing was not assessed. Therefore, conclusions whether our cases share classical germline TP53 mutation previously described in southern Brazil [4] are impossible. Finally, because of relative small number of events within groups in addition to the large number of variables, multivariate analysis would be of limited statistic power and thus could not be performed.
Conclusions
Most of the cases occurred in females and caused endocrine syndrome. Nearly half of patients were diagnosed with advanced-stage ACT, highlighting the disease’s aggressiveness. Almost 30% of those initially considered cured experienced recurrence. Death occurred in 37% of patients, being more common among those with older age, larger tumors, metastatic dis- ease at diagnosis, previous recurrence, incomplete resection, tumor spillage during surgery, higher Weiss score, microscopic capsular invasion, and necrosis. Thus, the identification of these recurrence and death factors is essential to guide ideal therapy and the most appropriate surveillance.
Declaration of interest
The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.
References
[1] Mansmann G, Lau J, Balk E, et al. The clinically inapparent adrenal mass: update in diagnosis and management. Endocr Rev. 2004;25:309-340.
[2] Kebebew E, Reiff E, Duh QY, et al. Extent of disease at presentation and outcome for adrenocor- tical carcinoma: have we made progress? World J Surg. 2006;30:872-878.
[3] Zancanella P, Pianovski MA, Oliveira BH, et al. Mitotane associated with cisplatin, etoposide, and doxorubicin in advanced childhood adrenocortical carcinoma: mitotane monitoring and tumor regression. J Pediatr Hematol Oncol. 2006;28:513-524.
[4] Ribeiro RC, Sandrini F, Figueiredo B, et al. An inherited p53 mutation that contributes in a tissue- specific manner to pediatric adrenal cortical carcinoma. Proc Natl Acad Sci U S A. 2001;98:9330- 9335.
[5] Faria AM, Almeida MQ. Differences in the molecular mechanisms of adrenocortical tumorigen- esis between children and adults. Mol Cell Endocrinol. 2012;351:52-57.
[6] Sandrini R, Ribeiro RC, DeLacerda L. Childhood adrenocortical tumors. J Clin Endocrinol Metab. 1997;82:2027-2031.
[7] Hanna AM, Pham TH, Askegard-Giesmann JR, et al. Outcome of adrenocortical tumors in chil- dren. J Pediatr Surg. 2008;43:843-849.
[8] Michalkiewicz E, Sandrini R, Figueiredo B, et al. Clinical and outcome characteristics of children with adrenocortical tumors: a report from the International Pediatric Adrenocortical Tumor Reg- istry. J Clin Oncol. 2004;22:838-845.
[9] Pianovski MA, Maluf EM, de Carvalho DS, et al. Mortality rate of adrenocortical tumors in chil- dren under 15 years of age in Curitiba, Brazil. Pediatr Blood Cancer. 2006;47:56-60.
[10] Weiss LM, Medeiros LJ, Vickery AL Jr. Pathologic features of prognostic significance in adreno- cortical carcinoma. Am J Surg Pathol. 1989;13:202-206.
[11] Fuhrman SA, Lasky LC, Limas C. Prognostic significance of morphologic parameters in renal cell carcinoma. Am J Surg Pathol. 1982;6:655-663.
[12] Ribeiro RC, Sandrini Neto RS, Schell MJ, et al. Adrenocortical carcinoma in children: a study of 40 cases. J Clin Oncol. 1990;8:67-74.
[13] Bugg MF, Ribeiro RC, Roberson PK, et al. Correlation of pathologic features with clinical out- come in pediatric adrenocortical neoplasia. A study of a Brazilian population. Brazilian Group for Treatment of Childhood Adrenocortical Tumors. Am J Clin Pathol. 1994;101:625-629.
[14] Cagle PT, Hough AJ, Pysher TJ, et al. Comparison of adrenal cortical tumors in children and adults. Cancer. 1986;57:2235-2237.
[15] Mendonça BB, Lucon AM, Menezes CA, et al. Clinical, hormonal and pathological find- ings in a comparative study of adrenocortical neoplasms in childhood and adulthood. J Urol. 1995;154:2004-2009.
[16] Macfarlane DA. Cancer of the adrenal cortex; the natural history, prognosis and treatment in a study of fifty-five cases. Ann R Coll Surg Engl. 1958;23:155-186.
[17] Wieneke JA, Thompson LD, Heffess CS. Adrenal cortical neoplasms in the pediatric population: a clinicopathologic and immunophenotypic analysis of 83 patients. Am J Surg Pathol. 2003;27:867- 881.
[18] Gulack BC, Rialon KL, Englum BR, et al. Factors associated with survival in pediatric adrenocorti- cal carcinoma: an analysis of the National Cancer Data Base (NCDB). J Pediatr Surg. 2016;51:172- 177.
[19] Sabbaga CC, Avilla SG, Schulz C, et al. Adrenocortical carcinoma in children: clinical aspects and prognosis. J Pediatr Surg. 1993;28:841-843.
[20] Sbragia L, Oliveira-Filho AG, Vassallo J, et al. Adrenocortical tumors in Brazilian children: immunohistochemical markers and prognostic factors. Arch Pathol Lab Med. 2005;129:1127- 1131.
[21] Dall’Igna P, Virgone C, De Salvo GL, et al. Adrenocortical tumors in Italian children: analy- sis of clinical characteristics and P53 status. Data from the national registries. J Pediatr Surg. 2014;49:1367-1371.
[22] Teinturier C, Pauchard MS, Brugieres L, et al. Clinical and prognostic aspects of adrenocortical neoplasms in childhood. Med Pediatr Oncol. 1999;32:106-111.
[23] Icard P, Goudet P, Charpenay C, et al. Adrenocortical carcinomas: surgical trends and results of a 253-patient series from the French Association of Endocrine Surgeons study group. World J Surg. 2001;25:891-897.
[24] Fassnacht M, Johanssen S, Quinkler M, et al. Limited prognostic value of the 2004 International Union Against Cancer staging classification for adrenocortical carcinoma: proposal for a Revised TNM Classification. Cancer. 2009;115:243-250.
[25] Lughezzani G, Sun M, Perrotte P, et al. The European Network for the Study of Adrenal Tumors staging system is prognostically superior to the international union against cancer-staging system: a North American validation. Eur J Cancer. 2010;46:713-719.
[26] Hubertus J, Boxberger N, Redlich A, et al. Surgical aspects in the treatment of adrenocortical carcinomas in children: data of the GPOH-MET 97 trial. Klin Padiatr. 2012;224:143-147.
[27] Rodriguez-Galindo C, Figueiredo BC, Zambetti GP, et al. Biology, clinical characteristics, and management of adrenocortical tumors in children. Pediatr Blood Cancer. 2005;45:265-273.
[28] Redlich A, Boxberger N, Strugala D, et al. Systemic treatment of adrenocortical carcinoma in children: data from the German GPOH-MET 97 trial. Klin Padiatr. 2012;224:366-371.
[29] Ribeiro RC, Pinto EM, Zambetti GP, et al. The International Pediatric Adrenocortical Tumor Reg- istry initiative: contributions to clinical, biological, and treatment advances in pediatric adreno- cortical tumors. Mol Cell Endocrinol. 2012;351:37-43.
[30] Terzolo M, Angeli A, Fassnacht M, et al. Adjuvant mitotane treatment for adrenocortical carci- noma. N Engl J Med. 2007;356:2372-2380.
[31] Terzolo M, Ardito A, Zaggia B, et al. Management of adjuvant mitotane therapy following resec- tion of adrenal cancer. Endocrine. 2012;42:521-525.