|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
© 2002 American Society for Clinical Oncology Role of Chest Computed Tomography at Diagnosis in the Management of Wilms’ Tumor: A Study by the United Kingdom Children’s Cancer Study GroupByFrom the Departments of Radiology and Haematology/Oncology, Great Ormond Street Hospital for Children National Health Service Trust, London; Department of Surgery, Institute of Child Health, London; and United Kingdom Children’s Cancer Study Group, Leicester, United Kingdom. Address reprint requests to C.M. Owens, MD, Department of Diagnostic Radiology, Great Ormond Street Hospital for Children National Health Service Trust, London WC1N 3JH, United Kingdom.
PURPOSE: This study sought to determine whether the identification of minimal pulmonary metastatic disease by chest computed tomography (CT) performed at diagnosis in patients with Wilms’ tumor and normal chest x-rays (CXR) could predict a subgroup of children at increased risk of pulmonary relapse. PATIENTS AND METHODS: A retrospective analysis was carried out of the records of 449 children entered onto the United Kingdom Childrens’ Cancer Study Group Second Wilms’ Tumor Study between July 1986 and September 1991. The imaging protocol did not stipulate chest CT at diagnosis, but 141 children who had normal frontal and lateral CXRs and a chest CT scan performed at diagnosis were eligible for analysis. After surgery, children with stage I Wilms’ tumor received single-agent chemotherapy (vincristine), whereas children with stages II, III, and bilateral Wilms’ tumor received combination chemotherapy. Most children with stage III tumors were also treated with abdominal radiotherapy (20 Gy). RESULTS: In 31 patients (22%), pulmonary nodules were visible on chest CT; eight experienced relapse, four (15%) in the lungs. When only stage I patients were analyzed, there was a significant difference between the pulmonary relapse rate of 43% (three of seven) in the CT-positive group and 10% (five of 48) in the CT-negative group (P = .02). Four of eight patients with stage I disease with pulmonary relapse died. CONCLUSION: CT seemed to identify a subgroup of stage I patients who were at increased risk of pulmonary relapse. These children had received only single-agent chemotherapy. A prospective randomized trial is needed to clarify whether these children would benefit from combination chemotherapy.
THE STAGING AND therapy for Wilms’ tumor patients are based on data collected before computed tomography (CT) was available. CT of the chest is now recommended by some investigators as an essential component of the pretreatment evaluation of children with Wilms’ tumor,1,2 but although chest CT is unquestionably more sensitive than chest x-ray (CXR) for the detection of pulmonary parenchymal nodules,3-5 these nodules may represent either metastatic cancer or various types of benign lung pathology. Three studies from the United States have attempted to assess the clinical significance of a positive chest CT in children with Wilms’ tumor and a negative CXR. Green et al6 were unable to demonstrate a difference in prognosis between similarly treated children with pulmonary nodules identified on CT and those with a normal scan. Wilimas et al,7 on the other hand, demonstrated an increased risk of pulmonary relapse among their cohort of comparable patients. In a separate study, however, these authors demonstrated that there was considerable variability in interpretation of chest CT scans,8 which limited the predictive value of their analysis. In view of this controversy, we evaluated the outcome of children entered into the United Kingdom Children’s Cancer Study Group’s (UKCCSG) Second Wilms’ Tumor Study (UKW2) Trial, with a negative plain CXR and a positive chest CT scan at diagnosis.
A retrospective study was performed of the imaging of 449 children entered onto the UKCCSG UKW2, which succeeded the first trial, the UKW1 trial,9 and ran from July 1986 until September 1991. All patients from the 22 major centers within the British Isles involved in the treatment of children with Wilms’ tumor were registered centrally, and follow-up data were submitted once every 6 months. As in the UKW1 study,9 histopathologic subtyping (favorable histology [FH] or unfavorable histology) and staging (I, II, III, IV, or bilateral) were assessed according to the National Wilms’ Tumor Study (NWTS) criteria. After surgery, stage I FH patients were treated with single agent chemotherapy (vincristine); stage II FH patients with a combination of vincristine and dactinomycin; stage III FH patients with vincristine, dactinomycin, and doxorubicin, and in most cases, 20-Gy abdominal irradiation; stage IV FH disease the same way as stage III disease, with or without 12-Gy whole-lung irradiation. Children with unfavorable histology tumors were treated with more intensive three-drug therapy (vincristine, dactinomycin, doxorubicin); stage III and IV patients also received 30 Gy hemiabdominal irradiation, and in the case of stage IV disease, 12-Gy whole-lung irradiation. The UKW2 protocol did not stipulate pulmonary CT scanning at diagnosis, and only centers with easy access to pediatric CT routinely performed lung scans. Both frontal and lateral CXRs were, however, mandatory. Patients with positive peripheral lymph nodes or known hematogenous metastases to liver or bone, and those with nodular pulmonary lesions identified as metastases on frontal or lateral CXRs (ie, all stage IV patients, n = 56), were excluded from the study. Four children with no allocated stage were also excluded. A total of 141 children (36% of all UKW2 patients) with normal frontal and lateral CXRs who had lung CT performed at diagnosis were eligible for analysis. Non–contrast-enhanced 8-mm (Siemens Somaton) or 10-mm (GE 9800 or Toshiba TCT 60 A) contiguous slices were obtained from lung apices to bases and imaged at settings appropriate for visualizing the mediastinum and lung parenchyma. Scans were performed before nephrectomy or biopsy and chemotherapy while the patients were under sedation or general anesthesia. Patients with negative CXR but positive CT scans were reviewed centrally. The frontal and lateral chest radiographs were reviewed in a blinded fashion (ie, without knowledge of the CT assessment). Abnormal pulmonary CT scans were reviewed independently by two radiologists, and a consensus view was reached if there were any discrepancies in opinions. Because of the wide range of possible differential diagnoses, patients with normal CXRs but positive pulmonary CT scans (the CT-positive group) were not submitted to biopsy of the lung nodules, and the patients were treated according to their abdominal stage locoregional disease; hence, histologic information is not available regarding the nature of the pulmonary lesions.
At the time of analysis (April 1996), the median follow-up period from the time of diagnosis was 75 months (range, 3 to 115 months). Analysis was performed on all 141 patients together and independently on stage I patients alone and on patients with stages II, III, and bilateral Wilms’ tumor combined to determine whether chemotherapy had any effect on subsequent relapse. Survival curves were calculated by the method of Kaplan and Meier.10 Log-rank tests11 were used to evaluate the statistical significance of individual prognostic factors, particularly histologic subtype, stage, and positive pulmonary CT scan. Multivariate analyses were carried out by Cox’s proportional hazard regression12 to assess the effect of, for example, CT positivity after allowance for other variables. Event-free survival (EFS) was defined as time from diagnosis to relapse or death or date of last follow-up. Pulmonary relapse-free survival was defined as the time from diagnosis to pulmonary relapse or last follow-up, with censoring of other events. Relapse rate was taken as the complement of pulmonary relapse-free survival.
The 141 patients who underwent pulmonary CT scanning seem to be representative of the 389 children with nonmetastatic disease entered into UKW2 with respect to age, stage, and histology (Table 1). In addition, EFS at 5 years from diagnosis was similar: 83% for the 141 patients with CT scans and 78% for the other 249 patients. The average proportion of patients scanned was 36%, but some centers carried out proportionally more CT scans than others. For centers entering more than 10 patients, for example, the proportion scanned ranged from 0% to 69%.
At central review of the CT scans in children with normal CXR and with no other evidence of hematogenous metastasis (eg, liver, bone, distant lymph nodes), one or more pulmonary nodules were identified in 31 (22%) of the 141 patients with chest CT scans performed at diagnosis (CT-positive group). Agreement was assessed between the reading radiologist by each making separate blind assessments of all CT scans for the presence or absence of pulmonary nodules. In the event that there was disagreement, a consensus (after discussion) was reached. This occurred in two of 141 scans—that is, overall, there was 99% agreement (  , 0.958; SE, 0.084). The CT scans of three of 34 patients reported as having bilateral basal nodules at their local centers were thought to demonstrate atelectasis by both reviewers and were included in the CT-negative, CXR-negative group. The CT-positive, CXR-negative group therefore included a total of 31 patients. Most of the nodules identified on CT at diagnosis were less than 8 mm in size (range, 2 to 8 mm; median, 5 mm), and all disappeared with therapy. Pulmonary relapse was diagnosed when CXR unequivocally demonstrated new nodular opacities compatible with secondary tumors. In these circumstances, histologic confirmation was deemed unnecessary. Repeat chest CT scans were performed in only two of these children, one of whom had disease recurrence at the site of each of the three original subpleural lesions, as well as multiple new parenchymal lesions. The other patient had multiple parenchymal lesions, one of which was at the site of a previous nodule. The numbers of patients relapsing or dying in relation to the CT scan result and tumor stage are listed in Table 2. Overall EFS at 5 years was lower in the CT-positive group 67% (95% confidence interval, 49% to 81%) than in the CT-negative group (95% confidence interval, 80% to 92%; P = .01 by log-rank test).
Eight of the 31 patients in the CT-positive group subsequently experienced relapse, with four episodes of relapse in the lungs (relapse rate, 13%). All patients who experienced relapse had FH disease. This figure was not significantly different from the pulmonary relapse rate in the CT-negative group of 7.3% (eight of 110). All the stage I patients who experienced relapse had FH tumors. For patients with stage II, III, and bilateral disease combined, the overall risk of pulmonary relapse was four (5%) of 86 patients. This was similar in the CT-negative group (three [5%] of 62) and the CT-positive group (one [4%] of 24). However, the risk of pulmonary relapse in the children with stage I tumors was significantly (P = .02) greater in the CT-positive group (three [43%] of seven) than in the CT-negative group (five [10%] of 48; Fig 1 and Table 2). There was also a suggestion of more pulmonary relapses (eight [15%] of 55) among stage I patients than in patients with stage II, III, and bilateral disease (four [5%] of 86, P = .05 by log-rank test), although the small sample size and multiple significance tests mean that these results must be interpreted with caution.
This study suggests that children in the UKW2 study with stage I Wilms’ tumor and pulmonary nodules detected only on chest CT (the CT-positive group) were at increased risk of pulmonary relapse compared with other patients. In contrast, the prognosis of CT-positive patients with other stage II, III, and bilateral disease was the same as the CT-negative group. However, the small number of patients in the two groups also limits the power of the study to confirm minimal differences in the outcome, and the power to detect a moderate difference in relapse rate is therefore low. The trend toward a higher pulmonary relapse rate in stage I patients overall, regardless of the CT scan result, implies that chemotherapy with vincristine alone might not be sufficient treatment for this group of children. In fact, the EFS for stage IV patients in the UKW2 study group was 71% at 3 years and at 6 years.13 This is higher than the EFS for the stage I CT-positive patients, which is 57%. Thus, stage IV patients treated with combination chemotherapy seem to have a better prognosis. Three other published studies address the significance of pulmonary nodules visible on chest CT, but not on CXR, in children with Wilms’ tumor. Wilimas et al7 reported 33 such children who had initial nephrectomy and were then treated according to the abdominal stage of their disease. Four (36%) of 11 patients developed recurrent disease in the lungs, half of them at the site or sites of the original nodules, compared with a 20% relapse rate in the study as a whole. In a more recent study of a larger group of 202 children, the same authors8 confirmed a significantly higher pulmonary relapse rate in patients with nodules identified on CT. However, they noted considerable observer variation in the interpretation of chest CT scans, which clouded the clinical significance of their findings. In contrast to our findings, Wilimas et al8 found no increased risk of pulmonary relapse in CT-positive children with stage I disease: overall, two (3%) of 61 of their patients experienced pulmonary relapse, compared with eight (15%) of 55 in our study. However, their patients were all treated with two drugs (vincristine and dactinomycin) rather than with vincristine alone. Green et al6 were unable to demonstrate a difference in outcome for children with pulmonary nodules on chest CT, whether they were treated intensively (ie, with vincristine, dactinomycin, doxorubicin, and whole-lung radiation therapy) or for stage I to III disease with either vincristine plus dactinomycin for stage I or II tumors or vincristine, dactinomycin, and doxorubicin for stage III disease. In a later publication from the third and fourth NWTS groups,14 the authors report the 4-year EFS for 53 FH CT-positive, CXR-negative patients treated as stage IV (with three or four drugs and whole-lung irradiation) as 88.6% and 90.6%, respectively, whereas the 4-year EFS and overall survival for 38 FH patients treated according to the extent of their locoregional disease with two of three drugs without total lung irradiation were 80% and 85%, respectively. These results are not significantly different, and there was no difference between the outcome of the stage IV CT-positive, CXR-negative group and the group treated as stage IV on the basis of their positive CXR (relapse-free survival and overall survival 78.7% and 82.7%, respectively). The NWTS investigation concluded that children with positive chest CT scans and normal CXR should be randomized and treated according to the stage of their locoregional disease. Before providing intensive three-drug treatment and pulmonary radiation, the NWTS recommends biopsy of CT-detected lesions to provide a precise diagnosis. CT scanning is widely used in the staging of patients with malignant solid tumors, both to assess the extent of the primary tumor and to detect and quantify metastatic disease. Many studies, particularly in adults, have demonstrated that CT scans identify more pulmonary nodules than CXR.1,3-5,15-17 A study by Wooton-Gorges et al18 included 83 patients with Wilms’ tumor and compared the sensitivity of four view CXRs—anteroposterior, lateral, and right and left obliques—with pulmonary CT in detecting pulmonary nodules and found discrepancy in only two patients, one of which was a false-negative finding and one a false-positive finding. However, because this study used only one observer and did not indicate whether or not the observer was blinded to the clinical details, its findings are not supported generally in the literature.
In cancer patients, not all pulmonary nodules detected by CT scanning represent metastatic disease, and the differential diagnosis includes Histoplasma capsulatum infection,19 intrapulmonary lymph nodes,20 "round pneumonia," round atelectasis, inflammatory pseudotumor, and hamartoma.20,21 In a study by Chang et al16 of adults with various cancers, for instance, biopsy was performed on 69 CT-detected pulmonary lesions, and only 31 (45%) were malignant. A similar study in children was reported, in abstract form, by Crisp et al.22 These workers carried out a retrospective review of pathology reports from 52 children with solid malignant tumors who underwent 74 thoracotomies for removal of lung lesions detected by CT but missed by CXR. Overall, 80% of the lung nodules were In a smaller study, Green et al6 reported that of five patients with Wilms’ tumor, negative CXR, and positive CT scans who underwent thoracotomy, four had histopathologically confirmed tumor and one had a nonmalignant lesion. Because the relapse rate for all patients with pulmonary nodules, rather than that for those with histologically proven metastases, was not determined in our study, the true rate of pulmonary recurrence may have been underestimated. After pulmonary relapse or recurrence, the stage I patients we assessed were not all curable with second-line chemotherapy. This raises the following questions: Should stage I patients presenting with a negative CXR but with a positive chest CT scan be upstaged? That is, should they be categorized as having stage IV disease? Should they receive more intensive chemotherapy and possibly whole-lung irradiation? The possible benefit of additional treatment has to be gauged against the risks. Dactinomycin causes myelosuppression, and there is a 1% to 2% risk of life-threatening hepatopathy associated with thrombocytopenia,23,24 probably caused by hepatic veno-occlusive disease. Doxorubicin is also myelosuppressive and causes cardiomyopathy. Cardiac function abnormalities are detectable in up to 25% of young children receiving this drug, and congestive heart failure develops in around 2% of patients.25,26 Pulmonary irradiation can cause interstitial pneumonitis,27 which may be fatal, and lung function abnormalities28 and may exacerbate doxorubicin-induced cardiotoxicity. There was no increase in the rate of pulmonary relapse among patients with stage II or III tumors and those with bilateral disease within UKW2, and only two of 61 pulmonary relapses in stage I patients treated with vincristine and dactinomycin reported by Wilimas et al.8 Therefore, eradication of minimal pulmonary disease with acceptable toxicity might be achieved by the addition of dactinomycin to vincristine in CT-positive stage I patients. Chest CT in children should only be performed if the results will lead to a change in therapy and thereby outcome. Our study suggests that compared with CXRs, pulmonary CT scanning at diagnosis has no additional predictive value, except in stage I patients. Assuming that children with stage II, III, and bilateral tumors always receive multiagent chemotherapy, the number of pulmonary CT scans could therefore be reduced by performing postoperative lung CT only in children confirmed as having stage I disease after surgery. Allowing a 7-day interval between operation and the CT scan should reduce the incidence of false-positive scans caused by areas of atelectasis that could be mistaken for metastatic tumor.29 CT scans should be performed by optimal, standardized techniques and central review of positive scans considered. Pulmonary nodules identified on CT could then undergo biopsy and only those children with confirmed metastatic disease considered for combination chemotherapy.
We acknowledge the cooperation of the many clinicians who registered their patients and participated in this UKCCSG study. We dedicate this article to the memory of our colleague and coauthor, Claire Dicks-Mireaux, MD, who died, suddenly and tragically, as the study reached its conclusion.
1. Cohen MD, Siddiqui A, Weetman R, et al: A rational approach to the radiological evaluation of children with Wilms’ tumor. Cancer 50: 887-892, 1982[CrossRef][Medline] 2. Cohen MD: Commentary: Imaging and staging of Wilms’ tumors—Problems and controversies. Pediatr Radiol 26: 307-311, 1996[CrossRef][Medline] 3. Muhm JR, Pritchard DJ: Computed tomography for the detection of pulmonary metastasis in patients with osteogenic sarcoma. Proc Am Assoc Cancer Res 21: 148, 1980 (abstr 593) 4. Muhm JR, Brown LR, Crowe JK: Detection of pulmonary nodules by computed tomography. Am J Roentgenol 128: 267-270, 1977[Abstract] 5. Schaner EG, Chang AE, Doppman JL: Comparison of computed and conventional whole lung tomography in detecting pulmonary nodules: A prospective radiologic-pathologic study. Am J Roentgenol 131: 51-54, 1978[Abstract] 6. Green DM, Fernbach DJ, Norkool P, et al: The treatment of Wilms’ tumor patients with pulmonary metastases detected only with computed tomography: A report from the National Wilms’ Tumor Study. J Clin Oncol 9: 1776-1781, 1991[Abstract]
7.
Wilimas J, Douglass EC, Magill HL, et al: Significance of pulmonary computed tomography at diagnosis in Wilms’ tumor. J Clin Oncol 6: 1144-1146, 1988
8.
Wilimas J, Kaste SC, Kauffman WM, et al: Use of chest computed tomography in the staging of pediatric Wilms’ tumor: Interobserver variability and prognostic significance. J Clin Oncol 15: 2631-2635, 1997
9.
Pritchard J, Imeson J, Barnes J, et al: Results of the United Kingdom Children’s Cancer Study Group first Wilms’ Tumor Study. J Clin Oncol 13: 124-133, 1995 10. Kaplan EL, Meier P: Nonparametric estimation from incomplete observations. J Am Stat Assoc 53: 457-481, 1958[CrossRef] 11. Peto R, Pike C, Armitage P: Design and analysis of randomized clinical trials requiring prolonged observation of each patient. Br J Cancer 35: 1-37, 1977[Medline] 12. Cox DR: Regression models and life-tables. J R Stat Soc B 34: 187-220, 1972 13. Mitchell C, Morris Jones P, Kelsey A, et al: The treatment of Wilms’ tumour: Results of the United Kingdom Children’s Cancer Study Group (UKCCSG) second Wilms’ tumour study. Br J Cancer 83: 602-608, 2000[CrossRef][Medline] 14. Meisel JA, Guthrie KS, Breslow NE, et al: Significance and management of computed tomography detected pulmonary nodules: A report from the National Wilms Tumour Study Group. Int J Rad Oncol Biol Phys 44: 579-585, 1999[CrossRef][Medline] 15. Silverman FN, Kuhn JP: Caffey’s Essentials of Pediatric X-Ray Diagnosis. Chicago, IL, Year Book Medical Publishers, 1990 16. Chang AE, Schaner EG, Conkle DM, et al: Evaluation of computed tomography in the detection of pulmonary metastases. Cancer 43: 913-919, 1979[CrossRef][Medline] 17. Cohen M, Grosfeld J, Baehner R, et al: Lung CT for detection of metastases: Solid tissue neoplasms of children. Am J Roentgenol 139: 895-898, 1992 18. Wooton-Gorges SL, Albano EA, Riggs JM, et al: Chest radiography versus chest CT in the evaluation for pulmonary metastases in patients with Wilms’ tumor: A retrospective review. Paediatr Radiol 30: 533-537, 2000[CrossRef][Medline]
19.
Cohen M, Smith WL, Provisor A: Pulmonary pseudometastases in children with malignant tumors. Radiology 141: 371-374, 1981 20. Robertson PL, Boldt DW, de Campo JF: Pediatric pulmonary nodules: A comparison of computed tomography, thoracotomy findings, and histology. Clin Radiol 39: 607-610, 1988[CrossRef][Medline]
21.
Hidalgo H, Korobkin M, Kinney TR, et al: The problem of benign pulmonary nodules in children receiving cytotoxic chemotherapy. Am J Roentgenol 140: 21-24, 1983 22. Crisp AJ, Babyn PS, Weitzman S, et al: Significance of lung nodules detected at CT in children with solid malignant tumors. International Pediatric Radiology Conference, Boston, MA, May 1996 (abstr 86) 23. Raine J, Bowman A, Wallendszus , et al: Hepatopathy-thrombocytopenia syndrome: A complication of dactinomycin therapy for Wilms’ tumor—A report from the United Kingdom Childrens Cancer Study Group. J Clin Oncol 9: 268-273, 1991[Abstract] 24. Green DM, Finklestein KZ, Norkool P, et al: Severe hepatic toxicity after treatment with single dactinomycin and vincristine. Cancer 62: 270-273, 1988[CrossRef][Medline] 25. Sorenson K, Levitt G, Sebag-Montefiore D, et al: Card-iac function in Wilms’ tumor survivors. J Clin Oncol 13: 17-22, 1994 26. Green DM, Breslow NE, Moskness JE, et al: Congestive cardiac failure after initial therapy for Wilms’ tumour: A report from the National Wilms’ Tumour study. Pediatr Res 35: 161A, 1994 (abstr) 27. Green DM, Finklestein KZ, Tefft ME, et al: Diffuse interstitial pneumonitis after pulmonary irradiation for metastatic Wilms’ tumor. Cancer 63: 450-453, 1989[CrossRef][Medline] 28. Jenney MEM, Shaw NJ: Late respiratory effects of treatment for childhood malignancy. Pediatr Rev Commun 8: 17-22, 1994 29. Damgaard-Pedersen K, Quist T: Pediatric pulmonary CT scanning. Pediatr Radiol 9: 145-148, 1980[CrossRef][Medline] Submitted February 26, 2001; accepted March 19, 2002. This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2002 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
2 tests were performed to determine whether there was any association between CT scans being performed and factors such as age, stage, and histology. 
10 mm in diameter, and 86 (83%) of 104 biopsies of tissue samples demonstrated metastatic tumor. 
