Decision analysis to compare treatment strategies for Stage I/favorable histology Wilms tumor.

BACKGROUND

Decision analysis was used to clarify differences in survival and complication rates comparing surgery alone versus surgery plus chemotherapy for Stage I, favorable histology Wilms tumor patients.

PROCEDURE

A state transition model was used to simulate treatment with nephrectomy-only, nephrectomy with adjuvant vincristine (VCR) or with vincristine plus dactinomcyin (NWTS Regimen EE4A). Rates of relapse and complications of therapy were obtained from the literature. In sensitivity analysis, the model was probed for the value(s) at which the treatment of choice changes.

RESULTS

The overall survival (OS) is essentially the same for patients treated with any of the three strategies (OS(Nephrectomy) = 98.8%; OS(EE4A) = 98.8%; OS(VCR) = 98.6%). Rates of serious long-term complications in the surviving population are also similar across treatment strategies (nephrectomy = 1.4%; VCR = 1.2%; EE4A = 0.3%). Both the progression and salvage rates after nephrectomy-only would have to be much worse than expected for nephrectomy-only to be an unacceptable strategy.

CONCLUSIONS

The differences in overall survival and rates of long-term complications between the three different initial strategies were negligible in the model. Based on this analysis, it was decided by the Children's Oncology Group that it was acceptable to continue to include nephrectomy without adjuvant chemotherapy as an experimental arm of the low risk Wilms tumor protocol with stringent eligibility criteria and close follow-up. Decision analysis can have a role in clinical trial design by making the tradeoffs between strategies more explicit. The robustness of these conclusions can be tested by widely varying the underlying assumptions.