Peroperative frozen section and cytology to assess proximal invasion in gastro-oesophageal carcinoma.

Peroperative frozen sections and cytological material have been obtained in 39 consecutive gastro-oesophageal resections between October 1978 and February 1980. Material from the proximal edge of the resected specimen was unreliable because of 3 false positive cytology results and 1 false positive frozen section report. Frozen section (no errors) was shown to be significantly more accurate than cytology (6 errors) when material was obtained from the distal oesophagus. Fatal anastomotic dehiscence occurred in only 1 of 30 resections when the excision was adequate. Where the results of frozen section were acted upon, in 5 of 9 patients with an inadequate excision, only 1 fatal dehiscence occurred compared with fatal dehiscence in all 4 patients where a second excision was not performed. These results indicate that frozen section of the distal oesophagus should be used to determine the extent of oesophageal resection in operations for gastro-oesophageal cancer.