User login
TOPLINE:
The adenoma missed rate (AMR), including lesions > 5 mm, was lower in patients whose precolonoscopy bowel preparation was deemed suitable using an artificial intelligence (AI)–guided assessment.
METHODOLOGY:
- Individuals with inappropriate bowel preparation before a colonoscopy face a higher lesion miss rate and may need to repeat the procedure.
- This prospective single-center study analyzed the screening colonoscopies of 393 individuals (mean age, 55 years; 50% men) using assessments made by endoscopists and AI.
- The AI-based method was the e-Boston Bowel Preparation Scale (e-BBPS), in which a score of 3 is the threshold to guarantee an adenoma detection rate of > 25%. Patients with an e-BBPS score of ≤ 3 or > 3 were considered by the AI as being qualified or unqualified, respectively.
- If the bowel preparation was considered adequate by the endoscopists and qualified by AI, individuals immediately underwent a repeat colonoscopy to assess for any missed lesions; otherwise, they underwent bowel recleansing before a repeat colonoscopy.
- The primary outcome was a > 5-mm AMR.
TAKEAWAY:
- The > 5-mm AMR was higher in individuals whose bowel preparation was deemed unqualified vs qualified by AI (35.71% vs 13.19%), particularly in the cecum (50.00% vs 25.00%), ascending colon (25.00% vs 9.09%), transverse colon (58.82% vs 14.71%), and descending colon (40.00% vs 21.43%).
- Similarly, any AMR (50.89% vs 20.79%), > 5-mm polyp miss rate (35.82% vs 19.48%), and any polyp miss rate (43.05% vs 25.51%) were higher in the unqualified AI vs qualified AI individuals.
- The rate of detection of adenomas > 5 mm (2.88% vs 11.25%) or any adenoma (15.97% vs 46.25%) was lower among the qualified AI vs unqualified AI individuals during the repeat colonoscopy.
- The e-BBPS also showed a high pairwise agreement with the analysis of expert endoscopists and moderate pairwise agreement with that of general endoscopists.
IN PRACTICE:
“The use of AI in bowel preparation assessment can relieve endoscopists’ workload, enabling them to concentrate more on detecting lesions during colonoscopy without being distracted by preparation evaluation, thus enhancing both efficiency and overall medical quality,” the authors wrote.
SOURCE:
The study, led by Liwen Yao, PhD, Department of Gastroenterology, Renmin Hospital of Wuhan University, Wuhan, China, was published online in Gastrointestinal Endoscopy.
LIMITATIONS:
Limitations included the study’s lack of external validity, including Western populations. Different bowel preparation regimens were not compared; therefore, conclusions about their efficacy cannot be deduced. The use of AI in the assessment of bowel preparation may lead to ethical issues, such as increased colonoscopy costs due to the technology and whether patients are fully informed.
DISCLOSURES:
This study was funded by the Science and Technology Achievement Transformation Platform Construction Project of Ministry of Education and Public Health Research Project of Futian District, Shenzhen. The authors declared no conflicts of interest.
A version of this article first appeared on Medscape.com.
TOPLINE:
The adenoma missed rate (AMR), including lesions > 5 mm, was lower in patients whose precolonoscopy bowel preparation was deemed suitable using an artificial intelligence (AI)–guided assessment.
METHODOLOGY:
- Individuals with inappropriate bowel preparation before a colonoscopy face a higher lesion miss rate and may need to repeat the procedure.
- This prospective single-center study analyzed the screening colonoscopies of 393 individuals (mean age, 55 years; 50% men) using assessments made by endoscopists and AI.
- The AI-based method was the e-Boston Bowel Preparation Scale (e-BBPS), in which a score of 3 is the threshold to guarantee an adenoma detection rate of > 25%. Patients with an e-BBPS score of ≤ 3 or > 3 were considered by the AI as being qualified or unqualified, respectively.
- If the bowel preparation was considered adequate by the endoscopists and qualified by AI, individuals immediately underwent a repeat colonoscopy to assess for any missed lesions; otherwise, they underwent bowel recleansing before a repeat colonoscopy.
- The primary outcome was a > 5-mm AMR.
TAKEAWAY:
- The > 5-mm AMR was higher in individuals whose bowel preparation was deemed unqualified vs qualified by AI (35.71% vs 13.19%), particularly in the cecum (50.00% vs 25.00%), ascending colon (25.00% vs 9.09%), transverse colon (58.82% vs 14.71%), and descending colon (40.00% vs 21.43%).
- Similarly, any AMR (50.89% vs 20.79%), > 5-mm polyp miss rate (35.82% vs 19.48%), and any polyp miss rate (43.05% vs 25.51%) were higher in the unqualified AI vs qualified AI individuals.
- The rate of detection of adenomas > 5 mm (2.88% vs 11.25%) or any adenoma (15.97% vs 46.25%) was lower among the qualified AI vs unqualified AI individuals during the repeat colonoscopy.
- The e-BBPS also showed a high pairwise agreement with the analysis of expert endoscopists and moderate pairwise agreement with that of general endoscopists.
IN PRACTICE:
“The use of AI in bowel preparation assessment can relieve endoscopists’ workload, enabling them to concentrate more on detecting lesions during colonoscopy without being distracted by preparation evaluation, thus enhancing both efficiency and overall medical quality,” the authors wrote.
SOURCE:
The study, led by Liwen Yao, PhD, Department of Gastroenterology, Renmin Hospital of Wuhan University, Wuhan, China, was published online in Gastrointestinal Endoscopy.
LIMITATIONS:
Limitations included the study’s lack of external validity, including Western populations. Different bowel preparation regimens were not compared; therefore, conclusions about their efficacy cannot be deduced. The use of AI in the assessment of bowel preparation may lead to ethical issues, such as increased colonoscopy costs due to the technology and whether patients are fully informed.
DISCLOSURES:
This study was funded by the Science and Technology Achievement Transformation Platform Construction Project of Ministry of Education and Public Health Research Project of Futian District, Shenzhen. The authors declared no conflicts of interest.
A version of this article first appeared on Medscape.com.
TOPLINE:
The adenoma missed rate (AMR), including lesions > 5 mm, was lower in patients whose precolonoscopy bowel preparation was deemed suitable using an artificial intelligence (AI)–guided assessment.
METHODOLOGY:
- Individuals with inappropriate bowel preparation before a colonoscopy face a higher lesion miss rate and may need to repeat the procedure.
- This prospective single-center study analyzed the screening colonoscopies of 393 individuals (mean age, 55 years; 50% men) using assessments made by endoscopists and AI.
- The AI-based method was the e-Boston Bowel Preparation Scale (e-BBPS), in which a score of 3 is the threshold to guarantee an adenoma detection rate of > 25%. Patients with an e-BBPS score of ≤ 3 or > 3 were considered by the AI as being qualified or unqualified, respectively.
- If the bowel preparation was considered adequate by the endoscopists and qualified by AI, individuals immediately underwent a repeat colonoscopy to assess for any missed lesions; otherwise, they underwent bowel recleansing before a repeat colonoscopy.
- The primary outcome was a > 5-mm AMR.
TAKEAWAY:
- The > 5-mm AMR was higher in individuals whose bowel preparation was deemed unqualified vs qualified by AI (35.71% vs 13.19%), particularly in the cecum (50.00% vs 25.00%), ascending colon (25.00% vs 9.09%), transverse colon (58.82% vs 14.71%), and descending colon (40.00% vs 21.43%).
- Similarly, any AMR (50.89% vs 20.79%), > 5-mm polyp miss rate (35.82% vs 19.48%), and any polyp miss rate (43.05% vs 25.51%) were higher in the unqualified AI vs qualified AI individuals.
- The rate of detection of adenomas > 5 mm (2.88% vs 11.25%) or any adenoma (15.97% vs 46.25%) was lower among the qualified AI vs unqualified AI individuals during the repeat colonoscopy.
- The e-BBPS also showed a high pairwise agreement with the analysis of expert endoscopists and moderate pairwise agreement with that of general endoscopists.
IN PRACTICE:
“The use of AI in bowel preparation assessment can relieve endoscopists’ workload, enabling them to concentrate more on detecting lesions during colonoscopy without being distracted by preparation evaluation, thus enhancing both efficiency and overall medical quality,” the authors wrote.
SOURCE:
The study, led by Liwen Yao, PhD, Department of Gastroenterology, Renmin Hospital of Wuhan University, Wuhan, China, was published online in Gastrointestinal Endoscopy.
LIMITATIONS:
Limitations included the study’s lack of external validity, including Western populations. Different bowel preparation regimens were not compared; therefore, conclusions about their efficacy cannot be deduced. The use of AI in the assessment of bowel preparation may lead to ethical issues, such as increased colonoscopy costs due to the technology and whether patients are fully informed.
DISCLOSURES:
This study was funded by the Science and Technology Achievement Transformation Platform Construction Project of Ministry of Education and Public Health Research Project of Futian District, Shenzhen. The authors declared no conflicts of interest.
A version of this article first appeared on Medscape.com.