EUS-guided gastroenterostomy (EUS-GE) with lumen-apposing metal stent (LAMS) has emerged as a mini-invasive, effective and safe alternative to surgical gastrojejunostomy and endoluminal stenting to treat malignant gastric outlet obstruction (MGOO). Currently, intra-abdominal carcinomatosis represents a relative contraindication to this procedure, due to a possible higher risk of clinical failure. The primary aim was to compare clinical success of EUS-GE in patients with MGOO with vs without concomitant intra-abdominal carcinomatosis. Secondary aims were to compare in the same subgroups of patients the clinical and procedural characteristics, the technical success rate and the long-term outcomes.

All patients undergoing EUS-GE for MGOO at our institution from June 2018 to December 2024 were retrospectively enrolled. Technical success was defined as successfully establishing EUS-GE and clinical success as low-residue diet tolerance without re-intervention for 14 days after the procedure. Inclusion criteria were: 1) age>18 years; 2) MGOO; 3) complete clinical records. Exclusion criteria was missing data. Intra-abdominal carcinomatosis was defined according either to CT and MRI imaging. Adverse events (AEs) recorded and classified (AGREE). Obstructive symptoms graded according to the GOOSS score. Data expressed as median [range]. The Student-t-test, Mann-Whitney-u test and X2 test used as appropriate.

EUS-GE for MGOO was performed in 74 patients (36 [48.6%] males, age 71 [30-92]), with ECOG performance status ³2 in 29 (39.2%) and obstructive symptoms reported by 72 (97.3%), with a median GOOSS of 1 [0-3]. Malignancy stage was II in 6 (8%) patients, III in 4 (5%) and IV in 64 (87%). Intra-abdominal carcinomatosis was present in 29 (39.2%) patients. Ascites was significantly more frequent in patients with vs without carcinomatosis (19 [65.5%] vs 12 [26.7%], p=0.02). Other clinical characteristics were comparable between groups.

All EUS-GE were performed in deep sedation. LAMS size 20mm was used in a comparable proportion of patients with vs without carcinomatosis (25 [86.2%] vs 37 [83.2%], p=0.89). Technical success rates and procedural time did not differ between groups (28 [96.6%] vs 43 [95.6%], p=0.69; vs 33 [16-85] vs 40 [20-106], p=0.52). Overall, 9 misdeployments were observed (carcinomatosis: 3 [10.3%] vs no-carcinomatosis: 6 [13.3%], p=0.98). AEs were 4 (19.8%) and 9 (20%) in patients with vs without carcinomatosis (p=0.71), all grade III. Time to refeeding was comparable between groups (1 [1-5] vs 1 [1-5], p=0.96), while post-procedural hospital stay was longer in patients with carcinomatosis (8 [2-18] vs 4 [2-15], p=0.0009). Clinical success rate did not differ in patients with vs without carcinomatosis (25 [86.2%] vs 43 [95.6%], p=0.31). All 6 failures were due to malignancy progression. Median follow-up duration was 51 [21-622] days. During follow-up, 27 (93.1%) patients with and 38 (84.4%) without carcinomatosis died (p=0.45). Median survival (days) and oral feeding toleration until last follow-up/death were comparable between patients with vs without carcinomatosis (42 [21-398] vs 66 [26-622], p=0.07 and 26 [89.6%] vs 42 [93.3%], p=0.89). Overall, oncological treatments were resumed in 11 (37.9%) patients with and 23 (51.1%) without carcinomatosis (p=0.38) Re-hospitalization for any cause were observed in a comparable proportion of patients with vs without carcinosis (14 [48.3%] vs 26 [57.7%], p=0.57), requiring reintervention in 1 case.

In our study population the high technical and clinical success rates of EUS-GE were comparable in patients with and without intra-abdominal carcinomatosis. The major outcomes were comparable between these 2 subgroups of patients, thus suggesting the long-term efficacy of this procedure also in patients with more advanced malignancy.