Purpose Neck dissection has traditionally played an important role in the treatment of patients with squamous cell carcinoma of the head and neck who present with regionally advanced neck disease (N2-N3). Radiotherapy and concurrent chemotherapy improves overall survival in advanced head-and-neck cancer compared with radiotherapy alone. The necessity for postchemoradiation neck dissection is controversial. The intent of this report was to define the value of neck dissection in this patient population better. Methods and materials Patients with locally advanced squamous carcinoma of the head and neck who also presented with nodal disease and underwent hyperfractionated radiotherapy and concurrent cisplatin/5-fluorouracil chemotherapy constituted the study population. Adjuvant modified neck dissection (MND) was planned 6 to 8 weeks after completion of chemoradiation in those patients who had a biopsy-proven pathologically complete response at the primary tumor site, irrespective of the clinical/radiographic neck response. A cohort of patients underwent electrode assessment of tumor oxygenation. Pathologic findings from the MND were used to compute the negative and positive predictive values and overall accuracy of the clinical/radiographic response (cCR). Regional control, failure-free survival, and survival were compared according to whether patients actually underwent MND. Results A total of 154 patients received concurrent chemoradiation. Of these, 108 presented with nodal disease: N1, n = 30; and N2-N3, n = 78. MND was performed in 65 (60%) of 108 patients, including 13 (43%) of 30 with Stage N1 and 52 (66%) of 78 with Stage N2-N3. For N1 patients, the negative predictive value of a cCR, positive predictive value of less than a cCR, and the overall accuracy for clinical response was 92%, 100%, and 92%, respectively. For N2-N3 patients, the corresponding values were 74%, 44%, and 60%. Patients with poorly oxygenated tumors were more likely to have residual disease at MND. The median follow-up was 4 years. The 4-year disease-free survival rate was 70% for N1 patients, irrespective of the clinical response or whether MND was performed. The 4-year disease-free survival rate was 75% for N2-N3 patients who had a cCR and underwent MND vs. 53% for patients who had a cCR but did not undergo MND (p = 0.08). The 4-year overall survival rate was 77% vs. 50% for these two groups of patients (p = 0.04). Conclusion The clinical and pathologic responses in the neck correlated poorly with one another for patients with N2-N3 neck disease undergoing concurrent chemoradiation for advanced head-and-neck cancer. MND still appears to confer a disease-free survival and overall survival advantage with acceptably low morbidity. Tumor oxygenation assessment may be useful in selecting patients who are especially prone to have residual disease. Better tools need to be developed to determine prospectively whether this procedure is required for individual patients.