Evaluation of Natural Language Processing for the Identification of Crohn Disease-Related Variables in Spanish Electronic Health Records: A Validation Study for the PREMONITION-CD Project.

BACKGROUND: The exploration of clinically relevant information in the free text of electronic health records (EHRs) holds the potential to positively impact clinical practice as well as knowledge regarding Crohn disease (CD), an inflammatory bowel disease that may affect any segment of the gastrointestinal tract. The EHRead technology, a clinical natural language processing (cNLP) system, was designed to detect and extract clinical information from narratives in the clinical notes contained in EHRs. OBJECTIVE: The aim of this study is to validate the performance of the EHRead technology in identifying information of patients with CD. METHODS: We used the EHRead technology to explore and extract CD-related clinical information from EHRs. To validate this tool, we compared the output of the EHRead technology with a manually curated gold standard to assess the quality of our cNLP system in detecting records containing any reference to CD and its related variables. RESULTS: The validation metrics for the main variable (CD) were a precision of 0.88, a recall of 0.98, and an F1 score of 0.93. Regarding the secondary variables, we obtained a precision of 0.91, a recall of 0.71, and an F1 score of 0.80 for CD flare, while for the variable vedolizumab (treatment), a precision, recall, and F1 score of 0.86, 0.94, and 0.90 were obtained, respectively. CONCLUSIONS: This evaluation demonstrates the ability of the EHRead technology to identify patients with CD and their related variables from the free text of EHRs. To the best of our knowledge, this study is the first to use a cNLP system for the identification of CD in EHRs written in Spanish.

Clinical characteristics and prognostic factors for Crohn's disease relapses using natural language processing and machine learning: a pilot study.

BACKGROUND: The impact of relapses on disease burden in Crohn's disease (CD) warrants searching for predictive factors to anticipate relapses. This requires analysis of large datasets, including elusive free-text annotations from electronic health records. This study aims to describe clinical characteristics and treatment with biologics of CD patients and generate a data-driven predictive model for relapse using natural language processing (NLP) and machine learning (ML). METHODS: We performed a multicenter, retrospective study using a previously validated corpus of CD patient data from eight hospitals of the Spanish National Healthcare Network from 1 January 2014 to 31 December 2018 using NLP. Predictive models were created with ML algorithms, namely, logistic regression, decision trees, and random forests. RESULTS: CD phenotype, analyzed in 5938 CD patients, was predominantly inflammatory, and tobacco smoking appeared as a risk factor, confirming previous clinical studies. We also documented treatments, treatment switches, and time to discontinuation in biologics-treated CD patients. We found correlations between CD and patient family history of gastrointestinal neoplasms. Our predictive model ranked 25 000 variables for their potential as risk factors for CD relapse. Of highest relative importance were past relapses and patients' age, as well as leukocyte, hemoglobin, and fibrinogen levels. CONCLUSION: Through NLP, we identified variables such as smoking as a risk factor and described treatment patterns with biologics in CD patients. CD relapse prediction highlighted the importance of patients' age and some biochemistry values, though it proved highly challenging and merits the assessment of risk factors for relapse in a clinical setting.

The impact of submucosal fatty tissue during colon endoscopic submucosal dissection in a western center.

OBJECTIVES: Obesity is associated with submucosal fatty tissue. The main aim of this study was to assess the impact of submucosal fatty tissue on the success of colonic endoscopic submucosal dissection (C-ESD) in a western population. METHODS: This was a retrospective analysis of 125 consecutive C-ESDs performed between October 2015 and July 2017. Fatty tissue sign was defined as positive when the submucosal layer was covered with fatty tissue. The complexity of performing an ESD was assessed by the performing endoscopist, defined by the occurrence of intraprocedural perforation, inability to complete an en-bloc resection or a procedure time exceeding 180 min. RESULTS: Fatty tissue sign positive was present in 44.8% of the procedures. There were 28 (22.4%) c-ESD defined as complex. Factors associated with complex ESD included; fatty tissue sign [odds ratio (OR) 12.5; 95% confidence interval (CI), 1.9-81.9; P = 0.008], severe fibrosis (OR 148.6; 95% CI, 6.6-3358.0; P = 0.002), poor maneuverability (OR 267.4; 95% CI, 11.5-6212.5; P < 0.001) and polyp size ≥35 mm (OR 17.2; 95% CI, 2.6-113.8; P = 0.003). In patients demonstrating the fatty tissue sign, BMI and waist-to-height ratio (WHtR) were higher (27.8 vs. 24.7; P < 0.001 and 0.56 vs. 0.49; P < 0.001, respectively) and en-bloc resection was achieved less frequently (76.8 vs. 97.1%, P = 0.001). Multivariate analysis revealed higher risk of fatty tissue sign positive associated with WHtR ≥0.52 (OR 26.10, 95% CI, 7.63-89.35, P < 0.001). CONCLUSION: This study demonstrates that the fatty tissue sign contributes to procedural complexity during C-ESD. Central obesity correlates with the likelihood of submucosal fatty tissue and as such should be taken into account when planning procedures.

Feasibility and learning curve of unsupervised colorectal endoscopic submucosal hydrodissection at a Western Center.

OBJECTIVES: Colorectal endoscopic submucosal dissection (CR-ESD) is an evolving technique in Western countries. We aimed to determine the results of the untutored implementation of endoscopic submucosal hydrodissection for the treatment of complex colorectal polyps and establish the learning curve for this technique. METHODS: This study included data from 80 consecutive CR-ESDs performed by a single unsupervised western therapeutic endoscopist. To assess the learning curve, procedures were divided into four groups of 20 each. RESULTS: En bloc resection was achieved in 55, 75, 75 and 95% cases in the consecutive time periods (period 1 vs. 4, P = 0.003). Curative resection was achieved in 55, 75, 70 and 95%, respectively (P = 0.037). Overall, series results demonstrated R0 resection in 75% of cases, with 23.7% requiring conversion to endoscopic piecemeal mucosal resection, and 1.25% incomplete resections. Complications included perforations (7.5%) and bleeding (3.7%). Multivariate analysis revealed factors more likely to result in association with non en bloc vs. En bloc resection, where polyp size ≥35 mm [70 vs. 23.4%; odds ratio (OR) 13.2 (1.7-100.9); P = 0. 013], severe fibrosis [40 vs. 11.7%; OR 10.2 (1.2-86.3); P = 0.033] and where carbon dioxide for insufflation was not used [65 vs. 30%; OR 0.09 (0.01-0.53); P = 0.008]. CONCLUSION: CR-ESD by hydrodissection has good safety and efficacy profile and offers well tolerated and effective treatment for complex polyps. As such, this technique may be useful in the West, in centers, where previous gastric ESD is not frequent or Japanese mentoring is not possible.