The model for the differential diagnosis of uterine sarcomas is based on a machine learning classifier, trained and tested for binary classification: malignant vs benign, using histopathology results collected from a dataset of patients.
Patent title | METHOD FOR CLASSIFYNG A LESION AS A UTERINE LEIOMYOMA OR AS A UTERINE LEIOMYOSARCOMA |
---|---|
Thematic area | Health |
Ownership | ALMA MATER STUDIORUM - UNIVERSITA' DI BOLOGNA, Azienda Ospedaliera- Universitaria di Bologna Policlino S. Orsola- Malpighi |
Inventors | Miriam Santoro, LIDIA STRIGARI, GIULIA PAOLANI, Pierandrea De Iaco, Anna Myriam Perrone, Camelia Alexandra Coada |
Filed on | 27 June 2023 |
The major challenge we face is the difficulty in providing an accurate and rapid diagnosis for patients with uterine sarcoma (SU) and choosing the right surgical approach, not only for the patient's prognosis but also for preserving the uterus in young women who desire future pregnancies. Currently, the diagnosis of SU is almost always confirmed postoperatively through the definitive histological examination, due to the lack of unequivocal sonographic and radiological characteristics that can differentiate it from uterine leiomyomas (MU).
The model is based on a machine learning classifier, trained and tested for the binary classification task of malignant vs. benign, using histopathology results collected from a patient dataset. The model relies on 1409 quantitative features extracted from computed tomography (CT) images for supervised training and testing. The 'gold standard' used is the histopathological result. The model was trained and tested using a 10-fold cross-validation approach, employing this strategy as a technique for increasing the available data and enhancing predictive value on new data sets.
- Differential diagnosis between uterine sarcoma (SU) and uterine leiomyoma (MU) based on CT images
- Standardization of diagnosis
- Pre-surgical diagnosis
- Personalization of therapeutic approach.