An effective algorithm to detect the possibility of being MSI phenotype in endometrial cancer given the BMI status and histological subtype: a statistical study.
Isabel González VillaEnrique Francisco González DávilaIdaira Jael Expósito AfonsoLeynis Isabel Martínez BlancoJuan Francisco Loro FerrerJuan José Cabrera GalvánPublished in: Clinical & translational oncology : official publication of the Federation of Spanish Oncology Societies and of the National Cancer Institute of Mexico (2022)
22.4% of our patients were MSI phenotype. We obtained statistically significant differences by multivariate analysis between endometrioid subtype and higher FIGO classification grade with MSI phenotype and obesity with MSS phenotype. Given these statistical results, we propose a function for predicting the probability of being MSI phenotype taking into account the histological subtype (endometrioid/non-endometrioid carcinoma) and FIGO grade as well as obesity. This prediction may be useful prior to hysterectomy, for genetic study of the MLH1 promoter and subsequent genetic counseling.
Keyphrases
- endometrial cancer
- metabolic syndrome
- machine learning
- insulin resistance
- type diabetes
- end stage renal disease
- weight loss
- weight gain
- chronic kidney disease
- deep learning
- genome wide
- newly diagnosed
- body mass index
- dna methylation
- gene expression
- ejection fraction
- human immunodeficiency virus
- hepatitis c virus
- data analysis
- neural network