C0214 - Predictive Model for Refractory Benign Esophageal Strictures

Matthew Bell, MD¹, Karthik Ravi, MD², Ross Dierkhising, MS^2
1Mayo Clinic Graduate Medical Education, Rochester, MN; ²Mayo Clinic, Rochester, MN

Introduction: Refractory benign esophageal strictures (RBES) are defined by inadequate response to endoscopic dilation. While adjunctive modalities such as corticosteroid injection improve outcomes in RBES, lack of reliable predictors results in therapeutic delays with associated cost and morbidity. We sought to establish a predictive model for RBES.

Methods: Patients were identified through search of CPT codes for esophagogastroduodenoscopy (EGD) with esophageal stricture dilation, with identified cases performed after October 1, 2012 reviewed sequentially. In addition, a cohort of RBES patients from a prospectively maintained clinical database of self-dilation patients was identified. Demographic information, endoscopic findings, and dilation characteristics were collected. Malignant strictures, Schatzki rings, and previously treated strictures were excluded. RBES was defined by inability to achieve or maintain a diameter ≥14mm over 5 consecutive dilation sessions. Univariate and multivariable regression models were performed. Multivariable models were chosen by minimizing the AIC statistic, with model intercepts adjusted due to the increased prevalence RBES.

Results: 128 patients with index EGD and esophageal dilation were identified, with 25 (19.5%) meeting criteria for RBES. An additional 63 RBES patients were identified from the self-dilation patient cohort for a total of 88 RBES and 103 non RBES patients included in the analysis. Male gender, longer length, smaller diameter, upper/middle esophageal location, and radiation induced strictures were associated with RBES (p< 0.05). Given inconsistent reporting of stricture length and diameter, multivariable analysis both with and without these variables was performed with both yielding strong predictive models with c-statistic of 0.87 and 0.85, respectively. (Table 1) Subsequently, a predicted probability formula (Score=Intercept +coefficient1 *variable1 + coefficient2 *variable2 +…. Risk= 1/1+exp(-Score)) was created for RBES risk prediction at the time of index EGD for individual patients. (Figure 1)

Discussion: RBES can be predicted at index EGD based on patient characteristics and stricture features. Further, we demonstrate that a strongly predictive formula can calculate RBES risk on a case-by-case basis, potentially allowing for individualized patient care to guide therapeutic approach and reduce associated morbidity and cost in esophageal stricture management.




Disclosures:


Matthew Bell, MD¹, Karthik Ravi, MD², Ross Dierkhising, MS². C0214 - Predictive Model for Refractory Benign Esophageal Strictures, ACG 2022 Annual Scientific Meeting Abstracts. Charlotte, NC: American College of Gastroenterology.