Main Content
Stener Lesion
| |
Clinical Implementation
Value Proposition
A Stener lesion is a type of injury to the thumb. It is seen with a traumatic injury, called the "gamekeeper's thumb," in which there is a tear of the ulnar collateral ligament (UCL) at the level of the metacarpophalangeal joint. The Stener lesion occurs when the aponeurosis of the adductor pollicis muscle is interposed between the torn UCL and its original site of insertion at the base of the proximal phalanx of the thumb. Normally, the UCL is located deep to the adductor pollicis (not superficial to, as in a Stener lesion). Because there is interposition of the aponeurosis between the UCL and its insertion site, the torn UCL fails to heal as a result. The UCL is an important stabilizer of the thumb; therefore, failure of the UCL to heal will cause impairment of the overall hand function. As this requires surgical repair, missing this finding can be harmful to the patient. Radiologists would benefit from an algorithm segmenting the UCL and detecting its abnormality. An AI algorithm meeting this use case would help to reduce the false negative rate, patient risk, and medical-legal risk for the radiologists.Narrative(s)
A 17-year-old patient with a hyperabduction injury to the thumb following a fall on an outstretched hand presents with a painful thumb. The initial radiographs of the thumb demonstrate a small avulsion fracture at the ulnar base of the proximal phalanx. Subsequently, an MRI of the thumb is obtained. The AI algorithm evaluates the MR images, identifies the UCL, and determines it as normal, abnormal, or indeterminate. The abnormal UCL can be further subdivided into partial tear and full-thickness tear. If there is a tear of the UCL, then its location with respect to the aponeurosis of the adductor pollicis muscle is determined. If the UCL is located deep to the aponeurosis, then no Stener lesion is identified. If the UCL is located superficially, then a Stener lesion is identified. The algorithm should also be designed to look for other signs, such as the “yo-yo on a string” sign, which is formed by the retracted, torn UCL looking like a small mass lying superficial to the aponeurosis. The radiologist is informed of these findings at the time of interpretation.Workflow Description
Images obtained from MRI are sent to the PACS/viewer and the AI model. The images are analyzed by the AI model. The model categorizes the UCL as either normal, abnormal, or indeterminate. A message is sent to both the PACS/viewer and the reporting solution from the model with the classification information. If the type of injury (ie, partial tear or full-thickness tear) is also identified, such information can also be sent to the PACS/viewer and reporting solution to highlight the region the engine identified. Once the tear is identified, then its location with respect to the aponeurosis of the adductor pollicis muscle is determined. If the UCL is superficial to the aponeurosis, then the diagnosis of a Stener lesion is made and sent to the PACS/viewer and reporting solution.Considerations for Dataset Development
Procedures(s): {MRI, thumb}
Planes of Imaging: {AP, PA, Lat}
Sex at Birth: {Male, Female}
Technical Specifications
Inputs
DICOM Study
| Procedure | MRI, thumb |
| Views | AP, PA, Lat |
| Data Type | DICOM |
| Modality | MRI |
| Body Region | Upper Extremity |
| Anatomic Focus | Thumb |
Primary Outputs
UCL Status
| RadElement ID | |
| Definition | Determine the status of the UCL |
| Data Type | Categorical |
| Value Set | 0-Unknown 1-Normal 2-Abnormal |
| Units | N/A |
Stener Lesion Detection
| RadElement ID | |
| Definition | If UCL is abnormal, detect the presence of a Stener lesion |
| Data Type | Categorical |
| Value Set | 0-Unknown 1-Stener lesion present 2-Stener lesion absent |
| Units | N/A |
Secondary Outputs
UCL Tear Type
| RadElement ID | RDE256 |
| Definition | If UCL is abnormal, determine the UCL tear type |
| Data Type | Categorical |
| Value Set | 0-Unknown 1-Partial tear 2-Full-thickness tear |
| Units | N/A |
Future Development Ideas
Consider running an algorithm on the scanner to detect when imaging is not adequate. Issues with thumb-imaging orientation are common: Techs may confuse thumb and hand protocols; the patient may not be properly secured to the nailbed; and the coronal plane is more difficult to acquire.Related Datasets
No known related public datasets at this time, please us alert us if you know of any.