Interest in frozen section examination of margins and lymph nodes in laryngeal surgery.

One hundred and one patients presenting with a squamous cell carcinoma of the larynx underwent surgery in our department between January 1980 and May 1985. In most of these patients, nodes were removed from the main lymphatic drainage pathways and subjected to immediate frozen section examination. The results from frozen section examination of the nodes were then compared with those from the surgical specimens of cervical neck dissections performed on the patients according to the classic rules. In addition, margin resections were made and examined by frozen section after removal of the tumour. In the event of a positive finding, these resections were continued until healthy tissue was reached, the specimens being examined in addition by classic methods. Immediate frozen sections enable the margins of the resection to be verified correctly. In our series we were brought to extend the limits of resection in 10 cases out of 68 (15 per cent). It can also be seen that the accuracy of the pathologist's reading of the frozen sections is satisfactory. The overall level of error is three out of 68 (4.5 per cent). All the errors correspond to false negatives. The aim of avoiding neck dissections in the presence of N0, thanks to nodal selection with frozen section, is not attained. We find a 6/61 rate of false negatives for N0-N1 (10 per cent) when we compare the frozen sections of the selected nodes and the neck dissections. This is due to the fact that the surgeon may be led astray by a reactive hyperplastic node whilst other less inflammatory neighbouring nodes are in fact the site of metastasis.