No Prep CT Colonography
| Purpose | Find colon cancers and polyps on CT colonography without the need for bowel prep. Fecal tagging if necessary. |
| Tag(s) | |
| Panel | Abdominal Panel |
| Define-AI ID | 20020013 |
| Originator | William Ladd |
| Panel Chair | Luther B. Adair |
| Panel Reviewers | Luther B. Adair and Andrew Smith |
| License | |
| Status | Public Comment |
| RadElement Set | RDES177 |
Clinical Implementation
Value Proposition
It would be a game changer for making CT colonography more acceptable to patients than fiberoptic colonoscopy, barium enema, or CT colonography post bowel prep. Most patients feel that the bowel prep is the worst part of colon cancer screening procedures. CT colonography is a current standard but not widely used screening procedure to detect colon cancer, one of the most common lethal cancers. A CT scan is performed with colon inflation by inserting a rectal tube with balloon tip and inflating with room air by hand with the rectal tube inflation bulb, but without pre-procedure colon cleansing, followed by CT scanning of the abdomen in supine and prone positions, would increase acceptability and hence lead to more widespread use. It could be with or without stool tagging.
Narrative(s)
A 65-year-old patient is due for colon cancer screening. The patient had a screening colonoscopy 10 years ago; however, they are anxious about a repeat exam because of the uncomfortable bowel preparation. Artificial intelligence-enabled CT colonography does not require bowel preparation and is a more convenient option for the patient. Upon receiving this option, a colon polyp with a small focus of early in-situ cancer is discovered and subsequently removed, preventing development of a potentially lethal colon cancer.
Workflow Description
A CT scan is performed with colon inflation by inserting a rectal tube with balloon tip and inflating with room air by hand with the rectal tube inflation bulb, but without pre-procedure colon cleansing, followed by CT scanning of the abdomen in supine and prone positions. The data is sent directly from modality to the AI engine, which would identify colon inner wall abnormalities, primarily pedunculated and sessile polyps, flat (low plateau-like) lesions, and adherent pieces of stool (for exclusion). This information is then sent to PACS for incorporation into a radiology report. It should not be necessary to identify diverticular outpouchings.
Considerations for Dataset Development
| Procedures | CTC |
| View(s) | Axial CT slices, obtained both supine and prone. |
| Age | Adults older than the age of 50 |
| Sex at birth | Male or Female |
| Training set | Training on bowel-prepped CTC exams, since no supply of non-prep exams are available initially. |
Technical Specifications
Inputs
DICOM Study
| Procedure | Colon inflation |
| Views | Axial CT slices with multiplanar and 3D surface reconstructions, obtained supine and prone |
| Data Type | DICOM |
| Modality | CTC |
| Body Region | Colon |
Primary Outputs
Detect presence of colon lesion
| RadElement ID | |
| Definition | Detect if a colon lesion is present |
| Data Type | Categorical |
| Value Set |
|
| Units | N/A |
Secondary Outputs
Location of lesion
| RadElement ID | |
| Definition | Identify the location of all lesions |
| Data Type | Coordinates |
| Value Set | |
| Units | N/A |
Probability of marked finding representing a true colon wall tumor
| RadElement ID | |
| Definition | Probability of true lesion |
| Data Type | Numeric |
| Value Set | [0,1] |
| Units | N/A |
Probability of lesion malignancy
| RadElement ID | |
| Definition | Probability of malignant lesion |
| Data Type | Numeric |
| Value Set | [0,1] |
| Units | N/A |
Future Development Ideas
Try to see if it can work acceptably well without colon inflation.
Related Datasets
The Cancer Imaging Archive (TCIA)
Know of more related datasets? Please let us know