van Overhagen H, Laméris J S, Berger M Y, Klooswijk A I, Tilanus H W, van Pel R, Schütte H E
Department of Radiology, University Hospital Rotterdam-Dijkzigt, Erasmus University, The Netherlands.
J Comput Assist Tomogr. 1993 May-Jun;17(3):367-73. doi: 10.1097/00004728-199305000-00005.
The ability of preoperative CT to assess resectability and to stage carcinoma of the esophagus and gastroesophageal junction was studied in 71 patients who underwent transhiatal esophagectomy. Patients with preoperatively proven distant metastases who did not have surgery were not included in the present study. At surgery the tumor invaded adjacent mediastinal or abdominal structures in 18 patients (prevalence 25%), but was nonresectable in only 7 of these 18 patients (39%). Invasions of the tracheobronchial tree, the aorta, and the diaphragm were correctly detected on CT in 5 of 6, 1 of 2, and 2 of 10 patients. There were four false-positive results on CT; tracheobronchial invasion and pericardial invasion were incorrectly predicted in one and three patients, respectively. Invasion of adjacent structures was correctly assessed on CT in 58 (82%) patients and the depth of tumor invasion was correctly determined in 49 (69%) patients. Computed tomography correctly staged 57% of patients according to the classification of the American Joint Committee on Cancer. Understaging (31%) occurred more often than overstaging (11%). In the present study, computed tomography was not effective in assessing non-resectability by diagnosing invasion because of the relatively low prevalence of invasion of adjacent structures and the fact that invasion was often not associated with nonresectability. In assessing invasion itself, CT was accurate in diagnosing tracheobronchial involvement, but was limited in diagnosing invasion of other adjacent structures. In assessing stage grouping, CT was limited in detecting either diaphragmatic invasion or lymph node involvement.
对71例行经裂孔食管切除术的患者进行研究，以评估术前CT对食管癌及食管胃交界癌可切除性及分期的能力。术前已证实有远处转移且未接受手术的患者未纳入本研究。手术中，18例患者（发生率25%）的肿瘤侵犯了相邻的纵隔或腹部结构，但在这18例患者中只有7例（39%）不可切除。6例气管支气管树受侵患者中，CT正确检测出5例；2例主动脉受侵患者中，CT正确检测出1例；10例膈肌受侵患者中，CT正确检测出2例。CT有4例假阳性结果；分别有1例和3例患者被错误预测为气管支气管受侵和心包受侵。58例（82%）患者的相邻结构受侵情况在CT上得到正确评估，49例（69%）患者的肿瘤侵犯深度得到正确判定。根据美国癌症联合委员会的分类，CT正确分期的患者占57%。分期过低（31%）的情况比分期过高（11%）更常见。在本研究中，由于相邻结构受侵的发生率相对较低，且受侵往往与不可切除性无关，因此CT在通过诊断受侵来评估不可切除性方面并不有效。在评估受侵本身时，CT在诊断气管支气管受累方面准确，但在诊断其他相邻结构受侵方面有限。在评估分期分组时，CT在检测膈肌受侵或淋巴结受累方面有限。
