Wilms´ tumor (nephroblastoma) represents 8-10% of the malignances that occur in the pediatric age group and is the most frequent of common solid tumors in children (except for CNS tumors) [1,2]. The classic nephroblastoma is made up of varying proportions of three cell types - blastemal, stromal and epithelial, but they are not all present in every case. The blastemal cells usually show distinctive patterns of rosettes, tubules and pseudonodules. The stromal structure may contain mesenchymal component with metaplasia to striated muscles, cartilage and fat tissue [1,3].
The prognosis of patients with nephroblastoma is related to the tumor clinical stage, histologic features of the tumor and treatment protocols. The overall survival rate in children with Wilms´ tumor is approximately 90% regardless the clinical stage [1,2,4-7].
The aim of this study was to evaluate the course of disease and results of therapy in patients with Wilms´ tumor in relation to histologic tumor features according to SIOP (International Society of Pediatric Oncology) classification of 1993 and verified by means of classification SIOP-2001 [2,3,7,8].
Material and methods
Records of 44 children with Wilms´ tumor treated at the Department of Pediatric Surgery and Oncology of the University School of Medicine in Lodz in the years 1993-2001 were reviewed. There were 23 girls (52%) and 21 boys (48%), aged one month to 11 years (mean age 5 years). All patients underwent multimodal therapy according to the SIOP protocols. Treatment was initiated with neoadjuvant chemotherapy (ACTD - dactinomycin, VCR - vincristine) without histologic verification, diagnosis being based on clinical presentation and results of imaging studies. Surgical removal of the tumor with involved kidney was performed after 4-week-long (35 patients) or 6-week-long (5 children) initial chemotherapy. Restaging was performed and adjuvant chemotherapy was administered after histologic diagnosis of the tumor was established (ACTD - dactinomycin, VCR - vincristine, EPI - epirubicin, IFO - ifosfamide, VP-16 - etoposide, CBDCA - carboplatin). Radiation therapy was applied additionally in 10 children with locoregional progression of the neoplasm. In 4 infants under 6 months of age the treatment began with primary surgical removal of the tumor without neoadjuvant chemotherapy.
The assessment of the clinical course of the disease included: presence of distant metastases (M) detected at the diagnosis or during therapy, local recurrence of tumor confirmed by means of imaging studies, progression of the disease and deaths. The overall survival rate was calculated in the entire group of children with Wilms´ tumor as well as in consecutive subgroups related to the clinical stage and tumor histologic features.
The removed tumors were assessed histologically at the Department of Pathomorphology of the Institute of Pediatrics in Łódź according to the SIOP-1993 working classification of renal tumors of childhood. Results were verified at the Department of Pathology of the Institute of Mother and Child in Warsaw. Patients were grouped in four pathological categories:
n Group 1 - tumors with "favourable" histology (low grade malignancy),
n Group 2 - tumors with "standard" histology (medium grade malignancy),
n Group 3 - tumors with "unfavourable" histology (high grade malignancy),
n Group 4 - tumors of unclassified histology due to extensive necrosis.
For the purpose of this work the histopathologic diagnoses were reclassified according to the revised SIOP-2001 working classification of renal tumors of childhood to create a simulated analysis of a clinical course. Patients were divided in three groups: low, intermediate and high risk tumors.
The clinical staging of tumors was done following the National Wilms´ Tumor Study Group (NWTSG) in the SIOP modification [1,3,4].
Follow up period ranged from 4 to 12 yrs. (mean 8.7).
The obtained results were subjected to a statistical analysis using chi2 test and Yule ratio (Q).
Of 44 patients with Wilms´ tumor, favourable histology was found in 9 children (20.5%), standard histology in 30 instances (68.2%), unfavourable histology in 4 cases (9%) and in one patient (2.3%) histological type of tumor could not be determined due to its extensive necrosis, the effect of neoadjuvant chemotherapy (SIOP-93). There was no predominance of any of sexes observed in any of four histological groups. The average age was comparable in patients with standard and unfavourable tumor types. Children with favourable histological feature were significantly younger (p<0,001) (Table I). 36 out of 44 children were in clinical stage I and II (81.8%) and the majority were classified as having standard histological type of tumor (30/44 - 68.1%). Nine patients with favourable tumor histology were in clinical stage I (66.7%) and II (33.3%). Patients with standard histology were distributed in four clinical stage groups, although most of them were in stage I and II (23/30 - 76.7%). All 4 children with unfavourable histological features were classified as stage II only, and the patient with uncategorized histological type of Wilms´ tumor was in stage IV (Table II). Survival analysis was performed in consecutive clinical stages in order to examine possible correlation between the course of disease and clinical stage of tumor determined at the beginning of disease following NWTS (Table II). But no such correlation was found in the examined group. Complications of neoplastic disease were observed in ten children (22.7%). Metastatic disease was present in 4 cases (9.1%); in two patients at the diagnosis and in two during therapy with lung (3/4 - 75%) and liver (1/4 - 25%) involvement. There were 6 relapses (13.6%) and four deaths (9.1%). The highest diversity of the course of disease was observed in children with a standard histologic type of tumor. In this group of patients there were 3 cases of metastatic disease, five relapses and three deaths. Out of 4 patients with unfavourable histological type one child died after recurrence and progression of disease (Table III). The remaining 38 patients (86.4%) completed the therapy and are free of disease with follow up period ranging from 3 to 11 years (mean 7.7 years). Event-free survival (EFS) was 86.4% as compared to overall survival (OS) of 90.9%. Survival rate of 100% was observed in patients with favourable and unclassified histological type of tumor. In the group of standard histology, 90% of children were cured compared to 75% survivors in the group of unfavourable histology (Table II). Comparison of OS between SIOP-93 and simulated SIOP-01 analysis revealed increase in strategic groups of intermediate and high risk tumors (Tables II and IV).
With reference to introduction of the revised SIOP-2001 working classification of renal tumors of childhood, the authors performed "simulated" analysis of the clinical course of disease and distribution of patients in consecutive clinical stages. Simulation revealed that 5 patients (11.4%) with favourable tumor histology moved to the intermediate risk type, while 18 children (40.9%) with standard histology moved to the high risk type, increasing the total number of children in this group to 22 (50%) (Table IV). All but two patients with complications of neoplastic disease were classified as having high risk tumors in the simulated analysis based on the revised SIOP-01 classification. The only death in the group of intermediate risk tumors was the patient with poorly differentiated epithelial feature of Wilms´ tumor (Table III).
Analyzing the possible correlation between the stage of disease and histological features of tumor according to SIOP-93 classification, patients were divided into groups depending on histological diagnosis. The majority of cases with favourable histology were highly differentiated epithelial and foetal-cystic subtypes. Most of the patients with standard tumor histology had a blastemal subtype (18/30 - 60%), while within the unfavourable histology group an anaplastic subtype of tumor dominated (3/4 - 75%) (Table V). Simulated SIOP-01 analysis showed that blastemal subtype became the most common feature of high risk histology (18/22 - 81.8%) and together with the anaplastic subtype was responsible for mortality in this group (3/22 - 13.6%). A poorly differentiated epithelial subtype became the most frequent (7/17 - 41.2%) and the only feature with decreased OS in the group of intermediate risk tumors (Table 5.). Survival rate of children with Wilms´ tumor ranged in the present study from 66.7% to 100% depending on histological subtype of the tumor. Mortality was associated with three histological subtypes: diffuse anaplasia, low differentiated epithelial and blastemal (Tables V and VI).
More than thirty years ago Currie at al. noticed the correlation between histological type of the tumor and clinical course of the disease . Beckwith at al. were next, whose studies on Wilms´ tumor histologic structure allowed for the first time to separate two histologic groups of favourable and unfavourable correlation with the clinical course . Subsequent histological classifications supervised by SIOP and NWTSG created a third group: tumors with standard histology or tumors of intermediate malignancy [3,5,7,10].
The retrospective analysis performed by the authors in the group of 44 patients with Wilms´ tumor treated in one centre between 1993 and 2001 revealed the 9% incidence of tumors with unfavourable histological features (Table II) which was comparable with the SIOP data . However, in the light of the revised SIOP-01 classification these data appeared to be considerably lower and misleading. The simulation (SIOP-01) performed by authors demonstrated clearly that half of all treated patients (22/44 - 50%) were classified as high risk tumors which was confirmed by the clinical course of disease (Table IV). Such a spectacular and significant (p<0.0001) increase in number of patients of this group can be explained by reclassification and moving the children with blastemal tumor feature from the standard (SIOP-93) to the high risk group (SIOP-01) (Table V). The observation by Reinhard et al. of a 90% incidence of intermediate risk and 7% only of high risk tumors differed markedly from our findings .
The statistical dependence between patient´s age and sex and histological type of tumor was not demonstrated. However, the mean age of children with tumors of favourable histology (2.2 years) was lower than mean age of patients with standard and unfavourable types of tumor (respectively: 3.4 and 3.5 years). The low clinical stages (I and II) were observed in tumors with favourable histology (low risk tumors according to SIOP-01) while the high ones (III and IV) in the standard type tumors (high risk tumors according to SIOP-01). However, the clinical course of disease was not statistically dependent on the tumor stage in the analyzed group of patients.
In performing the study the authors examined the incidence of complications of neoplastic disease (metastases, recurrences, disease progression and mortality rate) as well as the survival rate in relation to histological diagnosis. It appeared that the highest incidence of metastases (3/30 - 10%), relapses (5/30 - 16,7%) and fatal course (3/30 - 10%) that was found in the group of standard histological features of tumor according to SIOP-93 moved to the group of high risk tumors after the SIOP-01 reclassification (Table III) with corresponding data of metastases - 13.6% (3/22), relapses - 22.7% (5/22) and deaths - 13.6% (3/22). The OS after reclassification increased with statistical significance from 90% to 94% in the standard histology group (Q=0.23) and from 75% to 86.4% in the high risk group (Q=0.36) (Tables II and IV) and the last one was higher than given in literature [5,11].
The overall survival rate of patients with Wilms´ tumor presented in this study differed in consecutive histologic subtypes. It was significantly lower (Q=1.0) in patients with the following subtypes of tumor independent of classification: anaplastic (66.7%), poorly differentiated epithelial (85.7%) and blastemal (88.9%).This clearly shows that not only anaplastic and blastemal but also poorly differentiated epithelial subtype of Wilms´ tumor is associated with high malignancy and poor prognosis in comparison with all other subtypes in which a fatal course of disease was not observed. The authors have paid particular attention to the poorly differentiated epithelial feature of Wilms´ tumor, which as the only one out of three discussed histologic subtypes was not included in the high risk tumor group of the new revised SIOP-01 classification (the common epithelial subtype of intermediate risk tumors).
This finding needs further investigation. If it proves to be true, an adjustment of the classification should be considered. In the authors´ opinion, because the survival rate of patients with poorly differentiated epithelial subtype was even lower than in children with blastemal subtype of tumor, the former subtype should not be considered a standard histological feature of Wilms´ tumor.
The authors thank Dr. John N. Schullinger of New York Babies Hospital for reviewing the manuscript.