Is reseeding from the primary a plausible cause of node failure?

In a previous analysis of node failures in 1251 consecutive patients with node positive oropharyngeal and pharyngolaryngeal squamous cell carcinomas treated by external radiotherapy alone at the Institut Curie, the main reasons for patient exclusion were node recurrence associated with primary failure (N+T failures) and doses less than 55 Gy. These exclusions reduced the number of node failures from 399/1251 (32%) to 77/798 (10%). Multivariate analysis of node recurrence indicated that node size and fixity, treatment duration, and T stage of primary were significant (higher probability of isolated node failure for the T1-T2 primaries). In the present analysis, it is noted that 60% of the N+T failures were observed less than 1 month after the completion of the irradiation and, therefore, were not likely the result of reseeding from the primary tumor. When all 1251 patients were included in the analysis, the probability of nodal failure increased for larger nodes, T4 primaries, lower nodal doses, presence of contralateral node metastases, and nodal fixation to the surrounding structures. No influence of the primary site was found. Treatment duration was closely associated with total dose to the nodes. The best description of the data was obtained with a model including total dose and not treatment time. However, as in the previous analysis, the exclusion of low-dose (less than 55 Gy) treatments resulted in the loss of a significant dose-control relationship. We conclude that the majority of node failures is unlikely to result from reseeding from the primary tumor, and therefore should not be excluded from local-control analyses. From a more radiobiological point of view, the exclusion of palliative treatments is questionable when studying the effect of dose on local control.