Optimizing endoscopic biopsy detection of early cancers in Barrett's high-grade dysplasia

The of high-grade dysplasia management (HGD) in Barrett's esophagus remains controversial, in part, because of uncertainty about the ability of endoscopic biopsies to consistently detect early, curable cancers.

Here we report cancers we have diagnosed in 45 patients with Barrett's HGD using a protocol involving serial endoscopies with four-quadrant biopsies taken at 1-cm intervals. We compare these results to a modeled endoscopic biopsy protocol in which four-quadrant biopsies are taken every 2 cm in the Barrett's segment.

Thirteen cancers were detected at the baseline endoscopy and 32 in surveillance. In 82% of patients, cancer was detected at a single 1-cm level of the esophagus, and in 69% the cancer was detected in a single endoscopic biopsy specimen. A 2-cm protocol missed 50% of cancers that were detected by a 1-cm protocol in Barrett's segments 2 cm or more without visible lesions. The maximum depth of cancer invasion was intramucosal in 96% of patients. Only 39% of patients who had endoscopic biopsy cancer diagnoses had cancer detected in the esophagectomy specimen. Adverse outcomes included the development of regional metastatic disease during surveillance (1 of 32), operative mortality (3 of 36), including two patients who had their primary surgeries at other institutions, and death from metastatic disease after endoscopic ablation performed at another institution (1 of 3).

A four-quadrant, 1-cm endoscopic biopsy protocol performed at closely timed intervals consistently detects early cancers arising in HGD in Barrett's esophagus and should be used in patients with HGD who do not undergo surgical resection.