Use of decision analysis in planning a management strategy for the stage N0 neck.

OBJECTIVE: There are three major strategies in the treatment of patients with a stage N0 neck with squamous cell cancer of the head and neck: elective neck dissection, elective neck irradiation, and observation. Each has appropriate salvage strategies in the event of local recurrence. We used decision analysis to determine the optimal strategy for neck treatment as a function of the probability of occult cervical metastasis. DATA SOURCES: We used the bibliographies of current articles and books to access clinical studies of patients with stage N0 neck cancer. STUDY SELECTION: Studies that included large numbers of patients and contained a minimum 2-year follow-up, with results analyzed in terms of outcome as a function of stage of neck disease, were included. DATA EXTRACTION: Specific data points were extracted from the studies independently by multiple observers, and mean values were used in the decision analysis. DATA SYNTHESIS: A decision tree was constructed with use of a computer model to compare the three management strategies. Probabilities of each of the possible events depicted in the trees were inserted into the tree structure. These probabilities were gleaned from the literature as described above. A sensitivity analysis was performed to determine the optimal threshold for treatment of the neck.
CONCLUSION: A patient with primary squamous cell carcinoma of the head and neck and stage N0 neck status should be observed if the probability of occult cervical metastasis is less than 20%. If the probability is greater than 20%, treatment of the neck is warranted. The treatment plan should ideally involve a single modality of therapy; both neck dissection and radiation therapy are quite efficacious in the clinical context, and the decision of which one to employ should be driven by the treatment of the primary lesion. Decision analysis is useful in complex clinical situations.