Introduction
Endoscopic resection (ER) is curative for Barrett’s oesophagus-related early-stage oesophageal adenocarcinoma (EAC) without high-risk features, such as lymphovascular invasion (LVI), positive margins (R1), and deep submucosal invasion (T1bsm2-3). Piecemeal endoscopic mucosal resection (pEMR), often used for lesions larger than 15mm, prevents assessment of lateral margins and therefore complicates precise risk estimation for residual EAC at the first post-ER endoscopy. Thus, we investigated alternative clinical factors that could potentially be used to assess the risk for residual and recurrent EAC post-ER.
Aims & Methods
We performed a longitudinal cohort study of patients who underwent ER between 2006-2023. Inclusion criteria were: T1a or well/moderately differentiated T1bsm1 EAC; LVI-, and R0 resection. ER was categorised as pEMR (n=105) or en-bloc ER (n=22 en-bloc EMR, n=9 endoscopic submucosal dissection). The primary outcome was residual EAC at first post-ER endoscopy. Secondary outcomes were residual HGD at first post-ER endoscopy, recurrence of EAC and HGD at any post-ER endoscopy, and remission of EAC, HGD and intestinal metaplasia (IM) at latest endoscopy. Multivariate logistic regression, receiver operating characteristics, chi-squared, and Kaplan Meier analyses were performed to investigate possible associated risk factors.
Results
The cumulative incidence of recurrence of HGD was higher post-pEMR compared to en-bloc ER (HR = 2.43, 95% CI: 1.1-5.36; p=0.021), with residual EAC at first post-ER endoscopy accounting for 35% of such post-pEMR recurrence. The percentage of pEMR specimens with EAC was an independent risk factor for residual EAC at first post-pEMR endoscopy (p=0.013, Table 1). A 53.5% cut-off of involved pEMR specimens was identified as the optimal threshold for residual EAC prediction (specificity=0.68, sensitivity=0.67, AUC=0.66). Indeed, rate of residual (p=0.013) and recurrent (p=0.00041) EAC were higher when EAC involvement was >50%.However, rates of remission of EAC (p=0.13), HGD (p=0.20) and IM (p=0.81) were unaffected by high EAC burden.
| Univariate | Multivariate |
| OR (95% CI) | P-Value | OR (95% CI) | P-Value |
| Age * | 0.96 (0.90-1.01) | 0.11 | 0.92 (0.87-0.99) | 0.025 |
| Sex, male * | 0.96 (0.20-3.38) | 0.95 | 0.40 (0.057-1.92) | 0.30 |
| High alcohol consumption (≥14 units weekly) * | 0.65 (0.17-3.15) | 0.55 | 0.44 (0.088-2.58) | 0.32 |
| Ever smoker * | 1.93 (0.70-5.70) | 0.21 | 2.09 (0.62-7.43) | 0.24 |
| Percentage of cancerous samples *† | 1.25 (1.03-1.52) | 0.026 | 1.32 (1.07-1.66) | 0.013 |
| Number of samples taken | 1.11 (0.92-1.33) | 0.25 |
|
|
| Differentiation * | 1.19 (0.48-3.02) | 0.71 | 1.05 (0.37-3.04) | 0.93 |
| Stage, T1b sm1 * | 3.50 (0.44-22.79) | 0.19 | 11.01 (1.01-106.05) | 0.035 |
Table 1: logistic regression model for residual EAC at the first post-pEMR endoscopy
* : inclusion in multivariate analysis
† : odds ratio for a 10% increase in the percentage of cancerous samples
Bold indicates statistical significance
Conclusion
High burden of cancer in pEMR specimens correlates with increased risk of residual and recurrent EAC at post-pEMR endoscopy. Post-pEMR site check is recommended prior to endoscopic ablation, particularly in patients where more than 50% of pEMR specimens show EAC, where repeat ER should be performed in cases of flat subtle lesions. This will ensure eradication of any focus of residual disease and high success rates for treatment of early EAC using endotherapy.