|
Purpose |
Detection and quantification of osteolytic lesions throughout the skeleton caused by Multiple Myeloma (Kahler's disease) |
|
Tag(s) |
|
|
Panel |
Musculoskeletal |
|
Define-AI ID |
20050036 |
|
Originator |
Erik Ranschaert |
| Lead | Erik Ranschaert |
|
Panel Chair |
Jay Patti |
|
Panel Reviewers |
Musculoskeletal |
| License | Creative Commons 4.0 |
| Status | Public Comment |
| RadElement Set | RDES167 |
Traditionally, evaluation for destructive bone lesions from multiple myeloma has been performed using a skeletal survey employing multiple plain radiographs of a patient. Given that CT technology is considerably more sensitive for destructive osseous lesions, recent literature has suggested that CT could be used in place of an x-ray based skeletal survey in evaluation of multiple Myla patients for the evaluation of bone lesions. Unfortunately, evaluation using CT of the entire body of a patient with multiple myeloma would be a tedious endeavor with greater cost to the patient based on the higher cost of CT compared to radiographs. If artificial intelligence algorithms were able to help identify and count destructive bone lesions without a sclerotic margin, this could decrease radiologist interpretive time, and potentially decrease the overall cost of a whole body CT scan to evaluate for myeloma bone lesions. The automation should increase the efficiency and accuracy of the evaluation of the total body CT scan.
68 yo patient under the care of an oncologist for multiple myeloma presents for a whole body CT scan to evaluate for osseous lesions related to multiple myeloma. Many institutions use plain film assessment for osseous lesions which are less sensitive than CT. The use of CT would increase the sensitivity of an examination, however also result in an increase in radiologist interpretation time/complexity, increased cost, and increased radiation. Given that radiation is not a major consideration in a patient with myeloma, the main disadvantages to using CT for osseous myeloma detection is in the time and monetary costs associated with the examination. The use of an AI algorithm to aid in the detection of myeloma lesions reduces the time and monetary cost of the examination and aids in the adoption of CT for following myeloma lesions.
The algorithm should be able to automatically analyze total body CT-scans made with patients suspected of Kahler’s disease, with the intention to identify osteolytic lesions greater than 5mm without a sclerotic border related to Multiple Myeloma and to count them accurately. The algorithm would be expected to identify, locate and highlight bone lesions on the CT scan and provide an output for identified lesions.
|
Inclusion criteria |
patients with confirmed diagnosis of multiple myeloma based upon EPR data who have a low dose CT of the whole body (multiple myeloma protocol). |
|
Age |
> 18 yo |
DICOM Study
|
Procedure |
low dose CT Skeleton (whole body) |
|
Views |
Axial |
|
Data Type |
DICOM |
|
Modality |
CT |
|
Body Region |
Total Body |
|
Anatomic Focus |
Skull, spine, pelvis, upper and lower limbs |
Location of Osteolytic Lesion
|
RadElement ID |
RDE1137 |
|
Definition |
Detect the presence of osteolytic lesions of 5 mm or larger without Sclerotic border. Represented by coordinates/location of each lesion within DICOM volume. |
|
Data Type |
Numeric |
|
Value Set |
N/A |
|
Units |
Coordinates |
Number of Osteloytic Lesion
|
RadElement ID |
RDE1138 |
|
Definition |
The number of osteolytic lesion(s) greater than 5 mm |
|
Data Type |
Numeric |
|
Value Set |
N/A |
|
Units |
N/A |
After identification of lesions and count of lesions which fit the criteria above, it is likely that a total Volumetric calculation of lytic lesions would be of clinical use. This additional step would require an algorithmic understanding of the margins of the lucent lesions rather than the just the location or coordinates of the center of the lesion for identification purposes. Additionally, further development could include the evaluation of the evolution of the lesions over time during treatment, changes in density (fatty content, changes in density due to treatment). Comparison of size and number with previous scans.