D0639 - Algorithmic Approach to Differentiate Between Non-Specific and Specific Etiologies of Chronic Terminal Ileitis

Award: Presidential Poster Award

Karan Sachdeva, MBBS¹, Samagra Agarwal, MBBS, MD, DM¹, Peeyush Kumar, MBBS, MD, DM¹, David Mathew, MBBS, MD, DM¹, Sudheer Kumar Vuyyuru, MBBS, MD, DM¹, Bhaskar Kante, MBBS, MD, DM¹, Pabitra Sahu, MBBS, MD, DM¹, Sandeep K Mundhra, MBBS, MD, DM¹, Shubi Virmani, MD¹, Venigalla Pratap Mouli, MBBS, MD, DM¹, Rajan Dhingra, MBBS, MD, DM¹, Govind Makharia, MBBS, MD, DM², Saurabh Kedia, MBBS, MD, DM¹, Vineet Ahuja, MBBS, MD, DM^2
1All India Institute of Medical Sciences, Delhi, Delhi, India; ²All India Institute of Medical Sciences, New Delhi, Delhi, India

Introduction: Chronic isolated terminal ileal (TI) involvement (terminal ileitis) may be seen in numerous etiologies including Crohn’s disease (CD) and intestinal tuberculosis (ITB) in addition to other etiologies that may improve with symptomatic management alone. We aimed to improve our previously suggested algorithm to distinguish patients who would merit specific treatment from those who do not.

Methods: Patients(n=153) with isolated TI involvement (ulcers or nodularity) following-up at All India Institute of Medical Sciences (2007-2022) were retrospectively reviewed. A specific (ITB/CD) diagnosis was made based on standardized criteria and other relevant data was collected. The cohort was utilized for validation of previously suggested algorithm while multivariate analysis with bootstrap validation was used to develop a revised algorithm.

Results: We included 153 patients (mean age 36.9 years, 70% males, median duration of symptoms 1.5 years) with isolated terminal ileal involvement of whom 109 (71.2%) received a specific diagnosis (69 CD, 40 ITB). Multivariate regression and validation statistics suggested that based on a combination of clinical (blood in stools, weight loss, hemoglobin), radiological (necrotic lymph nodes, long segment ileal involvement) and colonoscopic findings (presence or absence of deep ulcers), an optimism corrected c-statistic of 0.975 and 0.958 could be reached with and without histopathological findings respectively. Based on these, a revised algorithm was developed which showed a sensitivity of 99.08% (95%CI:94.99 - 99.98) and a specificity of 75.0%(95%CI:59.66 - 86.81). The PPV was 90.76%(95%CI:85.47 - 94.25) and NPV was 97.06%(95%CI:82.32 – 99.57) with an overall accuracy of 92.16% (95%CI:86.70 - 95.88). This was a more sensitive and specific than the previous algorithm (accuracy 83.6%, sensitivity 95.4%, and specificity 54.4%).

Discussion: We suggested a revised algorithm to stratify patients of terminal ileitis into those who have a specific etiology and those who need only symptomatic treatment. Our algorithm has the potential to avoid missed diagnosis as well as unnecessary side effects of treatment.



Disclosures:


Karan Sachdeva, MBBS¹, Samagra Agarwal, MBBS, MD, DM¹, Peeyush Kumar, MBBS, MD, DM¹, David Mathew, MBBS, MD, DM¹, Sudheer Kumar Vuyyuru, MBBS, MD, DM¹, Bhaskar Kante, MBBS, MD, DM¹, Pabitra Sahu, MBBS, MD, DM¹, Sandeep K Mundhra, MBBS, MD, DM¹, Shubi Virmani, MD¹, Venigalla Pratap Mouli, MBBS, MD, DM¹, Rajan Dhingra, MBBS, MD, DM¹, Govind Makharia, MBBS, MD, DM², Saurabh Kedia, MBBS, MD, DM¹, Vineet Ahuja, MBBS, MD, DM². D0639 - Algorithmic Approach to Differentiate Between Non-Specific and Specific Etiologies of Chronic Terminal Ileitis, ACG 2022 Annual Scientific Meeting Abstracts. Charlotte, NC: American College of Gastroenterology.