Screening Colonoscopy Withdrawal Time Threshold for Adequate Proximal Serrated Polyp Detection Rate

Development and validation of a novel colonoscopy withdrawal time indicator based on YOLOv5

BACKGROUND AND AIM

The study aims to introduce a novel indicator, effective withdrawal time (WTS), which measures the time spent actively searching for suspicious lesions during colonoscopy and to compare WTS and the conventional withdrawal time (WT).

METHODS

Colonoscopy video data from 472 patients across two hospitals were retrospectively analyzed. WTS was computed through a combination of artificial intelligence (AI) and manual verification. The results obtained through WTS were compared with those generated by the AI system. Patients were categorized into four groups based on the presence of polyps and whether resections or biopsies were performed. Bland Altman plots were utilized to compare AI-computed WTS with manually verified WTS. Scatterplots were used to illustrate WTS within the four groups, among different hospitals, and across various physicians. A parallel box plot was employed to depict the proportions of WTS relative to WT within each of the four groups.

RESULTS

The study included 472 patients, with a median age of 55 years, and 57.8% were male. A significant correlation with manually verified WTS (r = 0.918) was observed in AI-computed WTS. Significant differences in WTS/WT among the four groups were revealed by the parallel box plot (P < 0.001). The group with no detected polyps had the highest WTS/WT, with a median of 0.69 (interquartile range: 0.40, 0.97). WTS patterns were found to be varied between the two hospitals and among senior and junior physicians.

CONCLUSIONS

A promising alternative to traditional WT for quality control and training assessment in colonoscopy is offered by AI-assisted computation of WTS.

Artificial intelligence-assisted real-time monitoring of effective withdrawal time during colonoscopy: a novel quality marker of colonoscopy

BACKGROUND AND AIMS

The importance of withdrawal time during colonoscopy cannot be overstated in mitigating the risk of missed lesions and postcolonoscopy colorectal cancer. We evaluated a novel colonoscopy quality metric called the effective withdrawal time (EWT), which is an artificial intelligence (AI)-derived quantitative measure of quality withdrawal time, and its association with various colonic lesion detection rates as compared with standard withdrawal time (SWT).

METHODS

Three hundred fifty video recordings of colonoscopy withdrawal (from the cecum to the anus) were assessed by the new AI model. The primary outcome was adenoma detection rate (ADR) according to different quintiles of EWT. Multivariate logistic regression, adjusting for baseline covariates, was used to determine the adjusted odd ratios (ORs) for EWT on lesion detection rates, with the lowest quintile as reference. The area under the receiver-operating characteristic curve of EWT was compared with SWT.

RESULTS

The crude ADR in different quintiles of EWT, from lowest to highest, was 10.0%, 31.4%, 33.3%, 53.5%, and 85.7%. The ORs of detecting adenomas and polyps were significantly higher in all top 4 quintiles when compared with the lowest quintile. Each minute increase in EWT was associated with a 49% increase in ADR (aOR, 1.49; 95% confidence interval [CI], 1.36-1.65). The area under the receiver-operating characteristic curve of EWT was also significantly higher than SWT on adenoma detection (.80 [95% CI, .75-.84] vs .70 [95% CI, .64-.74], P < .01).

CONCLUSIONS

AI-derived monitoring of EWT is a promising novel quality indicator for colonoscopy, which is more associated with ADR than SWT.

Using Natural Language Processing to Automatically Identify Dysplasia in Pathology Reports for Patients With Barrett's Esophagus

BACKGROUND & AIMS

Identifying dysplasia of Barrett's esophagus (BE) in the electronic medical record (EMR) requires manual abstraction of unstructured data. Natural language processing (NLP) creates structure to unstructured free text. We aimed to develop and validate an NLP algorithm to identify dysplasia in BE patients on histopathology reports with varying report formats in a large integrated EMR system.

METHODS

We randomly selected 600 pathology reports for NLP development and 400 reports for validation from patients with suspected BE in the national Veterans Affairs databases. BE and dysplasia were verified by manual review of the pathology reports. We used NLP software (Clinical Language Annotation, Modeling, and Processing Toolkit; Melax Tech, Houston, TX) to develop an algorithm to identify dysplasia using findings. The algorithm performance characteristics were calculated as recall, precision, accuracy, and F-measure.

RESULTS

In the development set of 600 patients, 457 patients had confirmed BE (60 with dysplasia). The NLP identified dysplasia with 98.0% accuracy, 91.7% recall, and 93.2% precision, with an F-measure of 92.4%. All 7 patients with confirmed high-grade dysplasia were classified by the algorithm as having dysplasia. Among the 400 patients in the validation cohort, 230 had confirmed BE (39 with dysplasia). Compared with manual review, the NLP algorithm identified dysplasia with 98.7% accuracy, 92.3% recall, and 100.0% precision, with an F-measure of 96.0%.

CONCLUSIONS

NLP yielded a high degree of sensitivity and accuracy for identifying dysplasia from diverse types of pathology reports for patients with BE. The application of this algorithm would facilitate research and clinical care in an EMR system with text reports in large data repositories.

Leveraging Natural Language Processing to Extract Features of Colorectal Polyps From Pathology Reports for Epidemiologic Study

PURPOSE

Histopathologic features are critical for studying risk factors of colorectal polyps, but remain deeply embedded within unstructured pathology reports, requiring costly and time-consuming manual abstraction for research. In this study, we developed and evaluated a natural language processing (NLP) pipeline to automatically extract histopathologic features of colorectal polyps from pathology reports, with an emphasis on individual polyp size. These data were then linked with structured electronic health record (EHR) data, creating an analysis-ready epidemiologic data set.

METHODS

We obtained 24,584 pathology reports from colonoscopies performed at the University of Utah's Gastroenterology Clinic. Two investigators annotated 350 reports to determine inter-rater agreement, develop an annotation scheme, and create a reference standard for performance evaluation. The pipeline was then developed, and performance was compared against the reference for extracting polyp location, histology, size, shape, dysplasia, and the number of polyps. Finally, the pipeline was applied to 24,225 unseen reports and NLP-extracted data were linked with structured EHR data.

RESULTS

Across all features, our pipeline achieved a precision of 98.9%, a recall of 98.0%, and an F1-score of 98.4%. In patients with polyps, the pipeline correctly extracted 95.6% of sizes, 97.2% of polyp locations, 97.8% of histology, 98.3% of shapes, and 98.3% of dysplasia levels. When applied to unseen data, the pipeline classified 12,889 patients as having polyps, 4,907 patients without polyps, and extracted the features of 28,387 polyps. Tubular adenomas were the most common subtype (55.9%), 8.1% of polyps were advanced adenomas, and the mean polyp size was 0.57 (±0.4) cm.

CONCLUSION

Our pipeline extracted histopathologic features of colorectal polyps from colonoscopy pathology reports, most notably individual polyp sizes, with considerable accuracy. This study demonstrates the utility of NLP for extracting polyp features and linking these data with EHR data to create an epidemiologic data set to study colorectal polyp risk factors and outcomes.

Factors affecting colorectal adenoma detection rate

Colorectal cancer is a kind of malignant tumors that seriously threatens the health of Chinese people, and its morbidity and mortality rank third and fifth among malignant tumors in China, respectively. High-quality colonoscopy is an effective means of preventing colorectal cancer. Adenoma detection rate (ADR) is defined as the proportion of those who have adenomas in the total number of patients undergoing colonoscopy. With the increase of ADR, the incidence rate and mortality rate of colorectal cancer gradually decrease. Previous studies have found that ADR is affected by many factors. In this paper, we describe the factors that affect ADR from the aspects of colonoscopy, doctors, and equipment, with a purpose to strengthen the understanding of endoscopists on ADR and to increase ADR during colonoscopy.

Ability of known colorectal cancer susceptibility SNPs to predict colorectal cancer risk: A cohort study within the UK Biobank

Colorectal cancer risk stratification is crucial to improve screening and risk-reducing recommendations, and consequently do better than a one-size-fits-all screening regimen. Current screening guidelines in the UK, USA and Australia focus solely on family history and age for risk prediction, even though the vast majority of the population do not have any family history. We investigated adding a polygenic risk score based on 45 single-nucleotide polymorphisms to a family history model (combined model) to quantify how it improves the stratification and discriminatory performance of 10-year risk and full lifetime risk using a prospective population-based cohort within the UK Biobank. For both 10-year and full lifetime risk, the combined model had a wider risk distribution compared with family history alone, resulting in improved risk stratification of nearly 2-fold between the top and bottom risk quintiles of the full lifetime risk model. Importantly, the combined model can identify people (n = 72,019) who do not have family history of colorectal cancer but have a predicted risk that is equivalent to having at least one affected first-degree relative (n = 44,950). We also confirmed previous findings by showing that the combined full lifetime risk model significantly improves discriminatory accuracy compared with a simple family history model 0.673 (95% CI 0.664–0.682) versus 0.666 (95% CI 0.657–0.675), p = 0.0065. Therefore, a combined polygenic risk score and first-degree family history model could be used to improve risk stratified population screening programs.

Impact of the National Endoscopy Database (NED) on colonoscopy withdrawal time: a tertiary centre experience

OBJECTIVE

Colonoscopy withdrawal time (CWT) is a key performance indicator affecting polyp detection rate (PDR) and adenoma detection rate (ADR). However, studies have shown wide variation in CWT and ADR between different endoscopists. The National Endoscopy Database (NED) was implemented to enable quality assurance in all endoscopy units across the UK and also to reduce variation in practice. We aimed to assess whether CWT changed since the introduction of NED and whether CWT affected PDR.

METHODS

We used NED to retrospectively collect data regarding CWT and PDR of 25 endoscopists who performed (n=4459 colonoscopies) in the four quarters of 2019. We then compared this data to their performance in 2016, before using NED (n=4324 colonoscopies).

RESULTS

Mean CWT increased from 7.66 min in 2016 to 9.25 min in 2019 (p=0.0001). Mean PDR in the two periods was 29.9% and 28.3% (p=0.64). 72% of endoscopists (18/25) had CWT>6 min in 2016 versus 100% (25/25) in 2019, the longer CWT in 2019 positively correlated with the PDR (r=0.50, p=0.01). Gastroenterology consultants and trainee endoscopists had longer CWT compared with colorectal surgeons both before and after using NED.

CONCLUSION

NED usage increased withdrawal times in colonoscopy. Longer withdrawal times were associated with higher PDR. A national colonoscopy audit using data from NED is required to evaluate whether wide variations in practice across endoscopy units in the UK still exist and to ensure minimum colonoscopy quality standards are achieved.

Guidelines for accreditation of endoscopy units: quality measures from the Korean Society of Coloproctology

Purpose

Colonoscopy is an effective method of screening for colorectal cancer (CRC), and it can prevent CRC by detection and removal of precancerous lesions. The most important considerations when performing colonoscopy screening are the safety and satisfaction of the patient and the diagnostic accuracy. Accordingly, the Korean Society of Coloproctology (KSCP) herein proposes an optimal level of standard performance to be used in endoscopy units and by individual colonoscopists for screening colonoscopy. These guidelines establish specific criteria for assessment of safety and quality in screening colonoscopy.

Methods

The Colonoscopy Committee of the KSCP commissioned this Position Statement. Expert gastrointestinal surgeons representing the KSCP reviewed the published evidence to identify acceptable quality indicators and indicators that lacked sufficient evidence.

Results

The KSCP recommends an optimal standard list for quality control of screening colonoscopy in the following 6 categories: training and competency of the colonoscopist, procedural quality, facilities and equipment, performance indicators and auditable outcomes, disinfection of equipment, and sedation and recovery of the patient.

Conclusion

The KSCP recommends that endoscopy units performing CRC screening evaluate 6 key performance measures during daily practice.

Evolving Role and Future Directions of Natural Language Processing in Gastroenterology

In line with the current trajectory of healthcare reform, significant emphasis has been placed on improving the utilization of data collected during a clinical encounter. Although the structured fields of electronic health records have provided a convenient foundation on which to begin such efforts, it was well understood that a substantial portion of relevant information is confined in the free-text narratives documenting care. Unfortunately, extracting meaningful information from such narratives is a non-trivial task, traditionally requiring significant manual effort. Today, computational approaches from a field known as Natural Language Processing (NLP) are poised to make a transformational impact in the analysis and utilization of these documents across healthcare practice and research, particularly in procedure-heavy sub-disciplines such as gastroenterology (GI). As such, this manuscript provides a clinically focused review of NLP systems in GI practice. It begins with a detailed synopsis around the state of NLP techniques, presenting state-of-the-art methods and typical use cases in both clinical settings and across other domains. Next, it will present a robust literature review around current applications of NLP within four prominent areas of gastroenterology including endoscopy, inflammatory bowel disease, pancreaticobiliary, and liver diseases. Finally, it concludes with a discussion of open problems and future opportunities of this technology in the field of gastroenterology and health care as a whole.

Variation Over Time and Factors Associated With Detection Rates of Sessile Serrated Lesion Across the United States: Results Form a National Sample Using the GIQuIC Registry

INTRODUCTION

Sessile serrated lesions (SSLs) are important precursor lesions for the CpG island-methylated pathway to colorectal cancer. The reported detection rates of SSL are highly variable, and national or population-based estimates are not available. Patient-, provider-, and procedure-level factors associated with the detection rates of SSL have not been well described. The aim of our study was to study the detection rates of SSL, variability of rates over time, and factors associated with detection rates of SSL in a national sample of patients undergoing colonoscopy using the GIQuIC registry.

METHODS

We used colonoscopies submitted to the GIQuIC registry from 2014 to 2017 on adults, aged 18-89 years. Only the first colonoscopy record per patient was included. Indications for colonoscopy were categorized as screening, diagnostic, and surveillance. We used the hierarchical logistic models to study the factors associated with the detection rates of SSL. The Cochrane-Armitage test was used to study the significance of trend over time.

RESULTS

There were a total of 5,173,211 colonoscopies performed by 3,934 endoscopists during the study period. Among the 2,101,082 screening colonoscopies over the study period in adults older than or equal to 50 years that were complete to the cecum, the average detection rate per endoscopist for SSL was 6.43% (SD 5.18) and 6.25% standardized for the 2010 US population. There was a significant increase in the detection rates of SSLs from screening colonoscopies over the study period from 4.99% in 2014 to 7.09% in 2017 (P trend <0.001). Clinically significant factors associated with higher detection rates of SSL were longer withdrawal times (>11 minutes vs ≤6 minutes) (odds ratio [OR] 9.61; 9.03-10.24), adequate preparation (OR 1.25; 1.22-1.28), female sex (OR 1.17; 1.16-1.18), and use of a specialized gastrointestinal pathology group (OR 1.12; 95% confidence interval 1.04, 1.19).

DISCUSSION

Population-based estimates of the detection rates of SSL are 6% and have increased over time.

Impact of Moderate versus Deep Sedation and Trainee Participation on Adenoma Detection Rate-Analysis of a Veteran Population

BACKGROUND/AIMS

The adenoma detection rate (ADR) is used as a quality indicator for screening and surveillance colonoscopy. The study aimed to determine if moderate versus deep sedation affects the outcomes of the ADR and other quality metrics in the veteran population.

METHODS

A retrospective review of colonoscopies performed at Memphis Veterans Affairs Medical Center over a one-year period was conducted. A total of 900 colonoscopy reports were reviewed. After exclusion criteria, a total of 229 index, average-risk screening colonoscopies were identified. Data were collected to determine the impact of moderate (benzodiazepine plus opioids) versus deep (propofol) sedation on the ADR, polyp detection rate (PDR), and withdrawal time.

RESULTS

Among 229 screening colonoscopies, 103 (44.9%) used moderate sedation while 126 (55%) were done under deep sedation. The ADR and PDR were not significantly different between moderate versus deep sedation at 35.9% vs. 37.3% (p=0.82) and 58.2% vs. 48.4% (p=0.13), respectively. Similarly, there was no significant difference in withdrawal time between moderate and deep sedation (13.4 min vs. 14 min, p=0.56) during screening colonoscopies.

CONCLUSION

In veterans undergoing index, average-risk screening colonoscopies, the quality metrics of the ADR, PDR, and withdrawal time are not influenced by deep sedation compared with moderate sedation.

Patient and Physician Factors Associated with Adenoma and Sessile Serrated Lesion Detection Rates

BACKGROUND AND AIMS

Sessile serrated lesions (SSLs) have been increasingly recognized as precursors to colorectal cancer. Unlike adenoma detection rate (ADR), there is currently no agreed-upon benchmark for SSL detection rate (SSLDR), and data on factors that impact SSL detection are limited. We aimed to identify patient, endoscopist, and procedural factors associated with SSL and adenoma detection.

METHODS

We used a single-center electronic endoscopy database to identify all patients ages ≥ 50 years who underwent outpatient screening colonoscopy from January 1, 2012, to June 30, 2018. Univariable Chi-square analysis was used to determine patient, endoscopist, and procedure-related factors associated with SSL or adenoma detection. We used logistic regression with generalized estimating equations, accounting for clustering by individual endoscopist, to determine factors independently associated with ADR and SSLDR.

RESULTS

We identified 10,538 unique patients who underwent colonoscopy performed by 28 endoscopists. Overall SSLDR was 2.2%, and overall ADR was 29.1%. On multivariable analysis, patient age, sex, BMI, smoking, endoscopist withdrawal time, and year of colonoscopy were independent predictors of ADR. Smoking and year of colonoscopy were independent predictors of SSLDR. Sub-optimal bowel preparation was inversely associated with SSL detection but not ADR.

CONCLUSIONS

In this large study of patients undergoing average-risk screening colonoscopy, overall SSLDR was low, indicating that methods for increasing SSLDR are needed. Our findings suggest that endoscopists may take into account risk factors for SSLs, such as smoking history, and recognize that the detection of such lesions, even more so than for adenomas, is dependent on optimal bowel preparation.

Difference in Physician- and Patient-Dependent Factors Contributing to Adenoma Detection Rate and Serrated Polyp Detection Rate

BACKGROUND

Adenoma detection rate (ADR) is correlated with the risk of interval colorectal cancer and is considered as a quality benchmark for colonoscopy. Serrated polyp detection rate (SPDR) might be a more stringent indicator of quality in polyp detection.

AIMS

To evaluate in a 2-year monocentric observational study patient-dependent and endoscopist-dependent factors influencing ADR and SPDR in daily practice.

METHODS

We determined ADR and SPDR. We collected patient-dependent factors and endoscopist-dependent factors. Links between these data and detection rates were assessed by uni- and multivariate analysis.

RESULTS

A total of 11682 colonoscopies were performed (female: 54.3%; male: 45.7%; median age 58) by 30 endoscopists (female: 9; male: 21). ADR and SPDR were 29.2% and 8%, respectively. In multivariate analysis, ADR was associated with patient-dependent factors: age (OR 1.044, CI 95% 1.040-1.048), male gender (OR 1.7, CI 95% 1.56-1.85), personal history of polyp/cancer (OR 1.53, CI 95% 1.3-1.9), and positive fecal immunochemical test (OR 2.47, CI 95% 2.0-3.1). In multivariate analysis, SPDR was associated with withdrawal time (OR 1.25, CI 95% 1.17-1.32), low volume activity (OR 1.3, CI 95% 1.1-1.52), and personal history of polyp/cancer (OR 1.61, CI 95% 1.15-2.25).

CONCLUSION

In this large series of routine colonoscopies, we found that ADR was mainly driven by patient-dependent conditions, i.e., age, male gender, colonoscopy indication for positive FIT, and a personal history of polyp or cancer. In contrast, SPDR was mainly related to endoscopist-dependent factor, i.e., withdrawal time and low volume activity.

Ileal intubation is not associated with higher detection rate of right-sided conventional adenomas and serrated polyps compared to cecal intubation after adjustment for overall adenoma detection rate

Background

High cecal intubation rate (CIR) is associated with significant improved adenoma detection rate (ADR), however, self-reported CIR may be overestimated and inadequate documentation of cecal intubation is associated with a lower polyp detection rate compared to clear documentation. We aimed to investigate if ileal intubation may be associated with higher detection rates (DR) for right-sided conventional adenomas (cAD) and serrated polyps (SP) compared to cecal intubation in a large screening colonoscopy cohort.

Material and methods

Retrospective analysis of individuals ≥50 years with average risk for colorectal cancer (CRC) who underwent screening colonoscopy between 01/01/2012 and 14/12/2016 at a tertiary academic hospital and six community-based private practices. Exclusion criteria were conditions with increased risk for CRC (e.g. inflammatory bowel disease, history of CRC, hereditary cancer syndromes), previous colonoscopy at the same institution, and incomplete procedures. Right-sided colon was defined as caecum and ascending colon.

Results

4.138 individuals were analysed (mean age 62 years, 52.1% female). DR for right-sided cADs and SPs were significantly higher after ileal compared to cecal intubation in univariate (12.5% vs. 6.8%, p < 0.001, and 6.3% vs. 3.3%, p < 0.001), but not in multivariate analysis (OR 1.025, 95%-CI 0.639–1.646, p = 0.918, and OR 0.937, 95%-CI 0.671–1.309, p = 0.704). DRs did not differ between ileal and cecal intubation for endoscopists with ADR ≥25 and < 25%, respectively. ADR ≥25% was significantly associated with ileal intubation (OR 21.862, 95%-CI 18.049–26.481, p < 0.001).

Conclusion

Ileal intubation may not provide any benefit over cecal intubation concerning the detection of cADs and SPs in the right-sided colon.

Impact of moderate versus deep sedation on adenoma detection rate in index average-risk screening colonoscopies

BACKGROUND AND AIMS

The debate between moderate sedation versus deep sedation for index average-risk screening colonoscopies is well known to gastroenterologists. Ensuring the best of all metrics to perform quality colonoscopies for colon cancer prevention is paramount for both patients and physicians alike, because colon cancer remains the leading cause of cancer death and is the most-used screening tool in the United States. The aim of this study was to determine if moderate sedation versus deep sedation affects outcomes of adenoma detection rate (ADR) or polyp detection rate (PDR) in index, average-risk colonoscopies for colon cancer screening.

METHODS

A retrospective, single, tertiary care outpatient center study of 585 healthy average-risk patients who underwent index screening colonoscopy between June 1, 2015 to December 31, 2015 (moderate sedation only) and June 1, 2016,to December 31, 2016 (deep sedation only) was performed after Institutional Review Board approval. Demographic data and polyp details were collected to determine ADR and PDR. Patients who were not average risk were excluded from the study.

RESULTS

A total of 585 index average-risk screening colonoscopies were included in this study with 57.7% moderate sedation and 42.2% deep sedation. Neither ADR nor PDR was significantly different between the 2 groups (44.1% vs 38.5% [P = .18] and 71.9% vs 67.6% [P = .27], respectively).

CONCLUSIONS

In index average-risk screening colonoscopies, deep sedation appears to have no benefit compared with moderate sedation for ADR and PDR.

The dark side of the colon: current issues surrounding the significance, prevalence, detection, diagnosis and management of serrated polyps

PURPOSE OF REVIEW

Hyperplastic polyps, once considered to have no malignant potential, are now recognized to be part of a larger group of polyps known as serrated polyps. Serrated polyps can progress to CRC through an epigenetic pathway known as CpG Island Methylator Phenotype (CIMP), characterized by hypermethylation of specific DNA regions such as the promoter regions of the DNA mismatch repair genes like MLH1. The CIMP pathway is tightly linked with mutations of the oncogene BRAF. There are three subtypes of serrated polyps - hyperplastic polyps, sessile serrated polyps (SSPs) and traditional serrated adenomas (TSAs). TSAs harbor cytologic dysplasia whereas hyperplastic polyps and SSPs are nondysplastic lesions. Currently, only SSPs and TSAs are believed to progress to CRC whereas hyperplastic polyps are thought to be benign with no malignant potential. This article will review the current evidence while highlighting some of the issues regarding serrated polyps.

RECENT FINDINGS

One challenge has been pathologically distinguishing hyperplastic polyps from SSPs, which is an important distinction, given the potential for progression of SSPs to CRC. Other challenges regarding serrated polyps include adequate detection and resection. Surveillance guideline recommendations for some serrated polyps have been changed in current guidelines to reflect the malignant potential, recommending closer surveillance intervals than the 10-year follow-up that has been traditionally provided for hyperplastic polyps.

SUMMARY

Given the difficulties in diagnosing as well as resecting, it is important for endoscopists to know how to detect, resect and manage follow-up in patients with serrated polyps.