Acute colonic flexures: the basis for developing an artificial intelligence-based tool for predicting the course of...

Area of Science:

• Medical Imaging
• Gastroenterology
• Biomedical Engineering

Background:

• Colon tortuosity is a key factor influencing colonoscopy outcomes.
• Understanding colonic flexures can aid in predicting procedural difficulty.

Purpose of the Study:

• To analyze the number of acute colonic flexures in adult subjects.
• To investigate correlations between flexure count and demographic factors (gender, age, height, BMI).
• To develop an artificial intelligence algorithm for predicting colonoscopy course and identifying patients at risk for incomplete procedures.

Main Methods:

• Computed tomography (CT) scans of the abdominal cavity were analyzed for 224 adult subjects.
• The number of acute bends (≤90°) between adjacent colonic segments was quantified.
• Data were correlated with gender, age, height, and weight; an AI algorithm was proposed.

Main Results:

• Females exhibited an average of 9.74 acute flexions, while males had 8.7.
• More acute flexions were observed in women than men, and in older individuals compared to younger ones.
• The sigmoid colon showed the greatest variability in flexures (0-9). Older and female subjects had more flexures in the transverse colon; older subjects had more in the descending colon.

Conclusions:

• The number of acute colonic flexures can be used to classify patients into risk categories for incomplete colonoscopies.
• Specific ranges of acute flexures (4-7, 8-12, >12) can indicate varying levels of procedural risk.
• The study predicted a higher likelihood of troublesome colonoscopies in 14.9% of females and 6.1% of males based on flexure analysis.