A selective resection algorithm for barrett’s neoplasia optimizes oncological outcomes

Aims Accepted oncological principles advise en-bloc, R0, excision of cancer to achieve a curative resection. In Barrett’s neoplasia, this includes T1a and superficial-T1b disease. While treatment options include endoscopic mucosal resection (EMR) and endoscopic submucosal dissection (ESD), a mechanism for appropriate technique selection has not been described or validated.

Methods We conducted a prospective multi-centre observational study to evaluate the performance of a selective resection algorithm (SRA) for Barrett’s neoplasia. To achieve an en-bloc resection, ESD was selected if there was a suspicion for≥T1a disease and the lesion was>15 mm (January 2017 to April 2024). This was compared with a historical approach (HA), where ESD was only performed in cases suspicious for deep submucosal invasion (February 2013 to December 2016).

Results A total of 581 resections were performed in 542 patients. Median lesion size was 20 mm (IQR 10-30). EMR was performed in 354 (60.9%) and ESD in 227 (39.1%). The SRA cohort included 392 (67.5%) cases, and HA cohort 189 (32.5%). Histology was T1a adenocarcinoma in 177 (30.5%) and T1b in 94 (16.2%). For T1a disease, en-bloc resection (SRA 110 [83.3%] vs HA 22 [48.9%]; P<0.001), R0 excision (SRA 91 [68.9%] vs HA 17 [37.8%]; P<0.001) and curative resection (SRA 77 [58.3%] vs HA 14 [31.1%]; P=0.002) were higher in the SRA cohort. Lesions undergoing ESD were likely to be larger (29.9±17.6 mm vs 16.8±11.7 mm; P<0.001). The rates of en-bloc resection (ESD 94 [97.9%] vs EMR 38 [46.9%]; P<0.001), R0 excision (ESD 85 [88.5%] vs EMR 23 [28.4%], P<0.001) and curative resection (ESD 72 [75.0%] vs EMR 19 [23.5%]; P<0.001) were higher in the ESD group. Recurrence was lower in the ESD group (7 [7.3%] vs. 17 [20.9%], P=0.008). For T1b disease, en-bloc resection (SRA 70 [95.9%] vs. HA 9 [42.9%]; P<0.001), R0 excision (SRA 52 [71.2%] vs HA 2 [9.5%]; P<0.001) and curative resection (SRA 20 [27.4%] vs HA 0 [0%]; P=0.005) were higher in the SRA cohort. Lesions undergoing ESD were likely to be larger (33.5±18.0 mm vs 19.0±14.7 mm; P=0.007). The rates of en-bloc resection (ESD 76 [98.7%] vs EMR 3 [17.6%]; P<0.001), R0 excision (ESD 54 [70.1%] vs EMR 0 [0%], P<0.001) and curative resection (ESD 19 [24.7%] vs EMR 0 [0%]; P<0.001) were higher in the ESD group. Recurrence was lower in the ESD group (4 [5.2%] vs. EMR 4 [23.5%], P=0.014). Among all T1a and T1b-SM1 cases with favorable histology, which underwent ESD, 86/99 (86.9%) were curative resections.

Conclusions A selective resection algorithm optimizes oncologic outcomes for Barrett’s adenocarcinoma and mitigates the risk of piecemeal resection of cancers.