How many segments are necessary to characterize delayed colonic transit time?

Area of Science:

• Gastroenterology
• Medical Diagnostics
• Computational Biology

Background:

• Colonic transit time measurement using radiopaque markers is a valuable tool for diagnosing constipation.
• Current methods for segmenting the colon are subjective, lacking experimental validation.
• A more objective approach is needed to define colonic partitions for accurate transit time analysis.

Purpose of the Study:

• To establish a rational, data-driven method for determining colonic partitions in transit time measurements.
• To objectively segment patients based on their colonic transit patterns.
• To validate the sufficiency of classical colonic divisions for characterizing transit delays.

Main Methods:

• Colonic transit time was assessed in 852 patients across seven defined colonic segments.
• An unsupervised Gaussian mixture modeling algorithm was employed for initial segmentation.
• K-means clustering was subsequently used to group patients into homogeneous transit time profiles.

Main Results:

• Four distinct patient populations were identified: normal/fast transit (n=548) and three delayed transit groups (right, left, outlet).
• Delayed transit was characterized by specific patterns: right colon delay (n=82), left colon delay (n=87), and outlet constipation (n=135).
• The clustering approach demonstrated high accuracy, with only 3.7% of patients misclassified.

Conclusions:

• Unsupervised analysis confirms that a three-segment division of the colon and rectum is adequate for characterizing delayed colonic transit.
• This objective method enhances the diagnostic utility of colonic transit time measurements in constipated patients.
• The findings support a more standardized and reliable approach to diagnosing and managing constipation.