Validation of a Multitask Deep Learning System at Spine Metastasis CT
Study Details
Study Description
Brief Summary
The multitask deep learning system (DLS) with five algorithms detecting five quantitative factors of Spinal Instability Neoplastic Score (SINS) was developed. Radiologists and oncologists from multicenter will be recruited to read the CT scans in picture archiving and communication system (PACS) independently, comparing with the DLS. One month after reading the CT scans in PACS, the participants will also asked to perform a web-based test in the DLS website using the same CT scans. Area under the receiver operating characteristic curve (AUC), sensitivity, and specificity of the DLS were calculated with professional graders as the reference standard.
Condition or Disease | Intervention/Treatment | Phase |
---|---|---|
|
Study Design
Arms and Interventions
Arm | Intervention/Treatment |
---|---|
routine physicians
|
Diagnostic Test: Deep Learning System
The multitask DLS with five algorithms detecting five quantitative factors of SINS
|
DLS
|
Diagnostic Test: Deep Learning System
The multitask DLS with five algorithms detecting five quantitative factors of SINS
|
Outcome Measures
Primary Outcome Measures
- AUC [1 months]
Area under the receiver operating characteristic curve (AUC) of spinal instability detection
Secondary Outcome Measures
- sensitivity [1 months]
sensitivity of spinal instability detection
- specificity [1 months]
specificity of spinal instability detection
Eligibility Criteria
Criteria
Inclusion Criteria:
-
pathology-proven diagnosis of solid tumor;
-
spinal CT scan indicating spinal metastasis with at least one lesion;
-
no previous surgery for spinal metastasis
Exclusion Criteria:
-
spinal CT scans with no sagittal reconstruction;
-
the radiologist considered that the quality of CT image was unqualified.
Contacts and Locations
Locations
No locations specified.Sponsors and Collaborators
- Shanghai 6th People's Hospital
Investigators
None specified.Study Documents (Full-Text)
None provided.More Information
Publications
None provided.- DLS01