Background and aim: An inverse correlation between hospital volume and esophageal resection mortality has been reported. In this study, we compared the quality of preoperative metastasis detection between a high-volume referral center with that of low-volume referring regional centers.
Methods: In 573 patients diagnosed with esophageal cancer (1994-2003), the results of preoperative staging investigations (CT-scan, ultrasound of abdomen and neck, and chest x-ray) performed in 61 regional centers were re-evaluated and/or repeated in one referral center. The gold standards were a radiological result with > or =6 months follow-up, fine-needle aspiration, or the postoperative TNM-stage.
Results: In the same group of patients, the preoperative investigations performed in regional centers detected true-positive malignant lymph nodes in 8% of patients and true-positive distant metastases in 7% of patients, whereas these percentages were 16% and 20%, respectively, in the referral center. In 72/573 (13%) patients, one or more metastases detected in the referral center had been missed in the regional centers. After allowing resectability in the presence of M1a lymph nodes, this would still have resulted in futile esophageal resections in 6% of patients. In contrast to the higher diagnostic sensitivity in the referral center, specificity was comparable between referral and regional centers.
Conclusions: This study found that, in assessing the operability of esophageal cancer, the diagnostic sensitivity of metastasis detection in a high-volume referral center was higher than that in referring regional centers. This resulted from both better CT-scanning equipment and more experienced radiologists in the referral center. Should the decision to perform esophagectomy have only been based on metastasis detection in these regional centers, over 1 in 20 patients would have undergone resection in the presence of metastases.