Dermoid Cyst Detection on CT

Purpose	Detection of a dermoid cyst on CT
Tag(s)
Panel	Abdominal
Define-AI ID	22020021
Originator	Martha Menchaca
Lead	Martha Menchaca
Panel Chair	Luther Adair II
Panel Reviewers	Abdominal Panel
License	Creative Commons 4.0
Status	Public Comment
RadElement Set	RDES2 16

Clinical Implementation

Value Proposition

AI for this use case would help in detecting and quantifying a benign pelvic mass which could be associated with pain or torsion in some patients. In clinical settings where radiologists are available, such detection could be used to alert the radiologist.

Narrative(s)

A 35-year-old female presents with right pelvic pain to the emergency room.

Workflow Description

The relevant images are obtained from the modality and sent to PACS and the AI engine based on anatomic landmarks. The images are analyzed by the engine. An algorithm evaluates the CT and, based on anatomic landmarks for the location of the pelvic adnexa, determines whether calcium is present, absent, or indeterminate by assessing the Hounsfield Units. If present, it then assesses whether there is also fat in this area by categorizing it as present, absent, or indeterminate. For cases returned where the algorithm returns to calcium or fat, the mass is measured in mm and that measurement is displayed. The radiologist is then made aware of the need for further attention to this area.

Considerations for Dataset Development

Procedures	{CT Abdomen with contrast, without contrast, CT Abdomen and pelvis}
View(s)	{Axial, Coronal, Sagittal}
Age	[0,50]
Sex at Birth	{Female}
Abdominal	{Abdominal or pelvic pain}
Other	{}
Confounding Variables	{pelvic contraception devices such as IUD, tubal hardware, pelvic fat, ruptured dermoid with fat density fluid throughout the pelvis, calcified uterine fibroids, ovarian calcifications}

Technical Specifications

Inputs

DICOM Study

Procedure	CT Abdomen with contrast, CT Abdomen without contrast
Views	Axial, Coronal, Sagittal
Data Type	DICOM
Modality	CT
Body Region	Abdomen and Pelvis
Anatomic Focus	Pelvic Adnexa

Primary Outputs

Fat Detection

RadElement ID	RDE1394
Definition	Detection of Mass with Fat (between −30 and −70 Hounsfield Units)
Data Type	Categorical
Value Set	Absent Present Intermediate
Units	N/A

Calcium Detection

RadElement ID	RDE1395
Definition	Detection of Mass with Fat ( Greater or equal to 1000 Hounsfield Units)
Data Type	Categorical
Value Set	Absent Present Intermediate
Units	N/A

Secondary Outputs

Mass Measurement

RadElement ID	RDE1396
Definition	Measurement of fat or calcium containing pelvic mass
Data Type	Numeric
Value Set	N/A
Units	millimeters

Future Development Ideas

● Future considerations include detection of cysts in the pelvis, categorization of pelvic cysts, and differentiation of pelvic cyst features.