About SMART-ROMA
About SMART-ROMA
SMART-ROMA uses continuous age, ln(HE4) and ln(CA125) as model inputs. A random forest classifier generates a probability that is mapped by isotonic regression before fixed-threshold risk classification.
Intended use
The model supports malignancy triage in women with suspected adnexal masses and may inform further imaging and specialist assessment. This research service is not a standalone diagnostic or treatment-decision tool.
Current model version
- Algorithm
- RandomForest
- Calibration
- isotonic
- Threshold
- 0.171429
- Features
- ln_HE4, ln_CA125, age
- Model SHA-256
809db3ca3d1d3f7c…