Pre-resection biopsy is usually performed for gastric lesions and non-ampullary duodenal adenomas. However, the final histological diagnosis often diverges, revealing the presence of high-grade dysplasia or carcinoma not previously detected. We aimed to assess the agreement between biopsy and definitive histopathological diagnosis in this setting.

We performed a retrospective analysis of gastric and non-ampullary duodenal  lesions submitted to endoscopic resection. Patients’ and lesions’ characteristics were prospectively collected. A descriptive analysis was performed.

We collected 135 lesions (54%  F, mean age 75.6 years) treated by endoscopic resection and previous biopsy with forceps, performed between January 2021 and October 2025 at Forlì-Cesena Gastroenterology Unit. Gastric (71%) and duodenal (29%) lesions were included, with mean dimensions of 20 mm, respectively submitted to Endoscopic Submucosal Dissection (ESD) and Endoscopic Mucosal Resection (EMR) according to European Guidelines.

In 35% of cases we observed disagreement between biopsies and definitive histology ; in details, for gastric lesions understaging rate was 41% and among them carcinoma was diagnosed in one third of cases not previously  detected. Notably 5% emerged as mixed type gastric cancer including signet ring cells, leading to a significant change in the oncological approach. 

In our experience and in accordance with previous literature, pre-resection sampling showed a low prediction rate of definitive histology despite its economic and ecological burden, as well potential delay in therapeutic decisions. In this setting, a future change in the diagnostic algorithm for upper GI lesions, especially gastric ones, should be considered. The adoption of artificial systems and chromoendoscopy could strengthen this approach.