CORRESPONDENCE
To the Editor:
We read with interest the article by Sabbaga et al1 entitled “Adrenocortical Carcinoma in Children: Clinical Aspects and Prognosis,” in which they reported the results of analyzing the data regarding 55 children from two institutions in Curitiba, Brazil (Hospital de Clinicas and Hospital Infantil Pequeno Principe). The authors found that age > 2 years and duration of symptoms 26 months before the diagnosis were significantly associated with poor prognosis. Surprisingly, tumor size was not a prognostic indicator.
An analysis of prognostic factors, the clinical features, and outcomes of 40 children from the Hospital de Clinicas. Curitiba, has been previously reported.2-4 Important discrepancies between the conclusions drawn by Sabbaga et al and those from previously published studies were noted. With the aim of clarifying the differences in conclusions drawn from the results of these studies that included some of the same patients from the same institution, we offer the following comments on some of the methodological aspects of these studies, which may explain these differences.
In our first study2 from the Hospital de Clinicas, which included 40 children, univariate analysis disclosed that age ≥3.5 years at diagnosis, an interval of ≥6 months between first symptoms and diagnosis, tumor weight ≥ 100 g, tumor volume ≥ 200 cm3, and higher-than-normal levels of urinary 17-ketosteroid and 17- hydroxycorticosteroids were associated with an unfavorable out- come. Multivariate analysis disclosed that only a tumor volume ≥200 cm3 independently identified those patients with disease having an unfavorable prognosis. Histology, designated as either malignant, benign or indeterminate, was not associated with outcome. Subsequently, in another study,4 we investigated the prognostic importance of histology (examined by two pathologists without prior knowledge of the patients’ outcome, and classified according to the modified criteria of Weiss et al5), DNA ploidy, proliferative index, and tumor size in a larger series of patients from southern Brazil, including the cases from Curitiba. Tumor size and histological classification were predictive of therapeutic
failure. There were no failures in 11 adenomas, 5 failures in 27 low-grade carcinomas, and 9 failures in 16 high-grade carcinomas.
In the study by Sabbaga et al, no reasons were given for excluding approximately one third of the patients from their analyses. Moreover, tumor weight estimates for each patient are not pro- vided, and there was no mention as to how tumor weights of patients with metastatic disease at diagnosis, partially excised tumors or tumor biopsy alone, were factored for statistical analysis. Finally, no histological classification of the tumors was provided.
Results of their analysis of prognostic variables is potentially misleading if, for example, patients with metastatic disease at diagnosis and/or with partially excised tumor were excluded when analyzing the prognostic importance of tumor size. In addition, unequal distribution of histological subtypes by age group may have accounted for the better outcome observed in the younger age group.
We think that prognostic indicators of childhood adrenocortical tumors have not been thoroughly defined. Complete resection of a localized tumor will cure the disease in approximately 60% of patients. Based on our studies,2-4 regardless of age or interval between disease onset and diagnosis, patients with large tumors or high-grade carcinoma histiotype, and those with gross local re- sidual or metastatic disease, have an increased risk of tumor recurrence or disease progression, respectively.
Raul C. Ribeiro, MD Department of Hematology-Oncology St Jude Children’s Research Hospital Memphis, TN
Romolo Sandrini, MD Department of Pediatrics Hospital de Clinicas Parana State University Curitiba, Brazil
REFERENCES
1. Sabbaga CC. Avilla SG, Schulz C, et al: Adrenocortical carcinoma in children: Clinical aspects and prognosis. J Pediatr Surg 28:841-843, 1993
2. Ribeiro RC, Sandrini RN, Schell MJ, et al: Adrenocortical carcinoma in children: A study of 40 cases. J Clin Oncol 8:67-74, 1990
3. Ribeiro RC, Sandrini RN, Bugg MF, et al: Pediatric adreno- cortical tumors: Relation of tumor DNA content to clinical and histologic features. Proc Am Soc Clin Oncol 11:375, 1992
4. Bugg MF, Ribeiro RC, Roberson PK, et al: Correlation of pathologic features with clinical outcome in pediatric adrenocorti- cal neoplasia: A study of a Brazilian population. Am J Pathol (in press)
5. Weiss LM, Medeiros LJ, Vickery AL Jr: Pathologic features of prognostic significance in adrenocortical carcinoma. Am J Surg Pathol 13:202-206, 1989