[Critical analysis of the new TNM staging (UICC, 1987) of cancer of the cervical esophagus in relation to therapeutic decisions].

66 consecutive patients with a tumor confined to the cervical esophagus underwent surgical resection. The comparison between clinical and pathological TNM stage showed a clinical understaging in 30 patients. 25 of the 56 patients who had undergone curative resection had lymph node metastases: positive mediastinal and abdominal nodes were found in 8 (32%) and 0 cases, respectively. The mean survival after curative resection of the 10 evaluable patients with metastatic periesophageal, recurrent and/or paratracheal nodes was 22.4 months; of the 6 evaluable patients with positive mediastinal nodes it was 10.3 months; and of the 5 patients with positive deep latero-cervical nodes it was 5.8 months. The 2-year actuarial survival after curative resection (in the 53 operative survivors) was as follows (according to pathologic TNM staging): Stage I (n = 3) 100%, Stage IIA (n = 17) 30%, Stage IIB (n = 3) 33%, and Stage III (n = 30) 22%. The exact location of neoplastic recurrence after curative resection was documented in 13 cases; it was in the neck in 8 cases (61%); both neck and at a distance in 3 cases (23%) and only at a distance in 2 (16%). The clinical TNM staging of cervical esophageal cancer was not in agreement with the pathological findings in nearly 50% of the cases and is, therefore, inaccurate and unreliable both for therapeutic decision-making and for prognostic evaluations. Endoscopic ultrasound, which was not used in most of the patients studied here, may improve the accuracy of clinical TNM staging. The N classification, which defines only the cervical nodes as regional nodes, appears to be arbitrary since the pathological staging showed metastatic mediastinal nodes in 32% of the N + cases, with a survival comparable to that of patients with metastatic nodes only in the neck. The prognostic value of pathological TNM staging was not confirmed in the present study since only Stage I patients had a significantly better prognosis than patients in the other stages. This may be due to the small number of patients considered or to lymph node understaging caused by the fact that most patients did not undergo mediastinal lymphadenectomy through a thoracotomy or a sternum splitting.