%0 Journal Article
%T Integrating Clinical and Diffusion-Weighted MRI Radiomics via Machine Learning for Predicting Lymph Node Metastasis in Endometrial Cancer
%A G. Papadopoulos
%A N. Vlachos
%A E. Georgiou
%J Asian Journal of Current Research in Clinical Cancer
%@ 3062-4444
%D 2022
%V 2
%N 2
%R 10.51847/ZiZPj56i2s
%P 88-99
%X Effective evaluation of lymphadenectomy risk is crucial in managing endometrial cancer (EC), as it requires balancing therapeutic outcomes with the potential for surgical complications and mortality. This study aimed to determine the additional diagnostic benefit of employing computer-assisted image segmentation combined with machine learning techniques that merge clinical indicators and diffusion-weighted magnetic resonance imaging (DWI-MRI) radiomic data for forecasting lymph node (LN) metastasis in EC. A total of 236 women diagnosed with EC (average age 51.2 ± 11.6 years) who underwent preoperative MRI between July 2010 and July 2018 were prospectively enrolled. Participants were randomly allocated into a training group (n = 165) and a validation group (n = 71). A decision tree classifier was established using several features: the tumor’s mean apparent diffusion coefficient (ADC; cutoff 1.1 × 10⁻³ mm²/s), relative ADC skewness (cutoff 1.2), LN short-axis diameter (cutoff 1.7 mm), LN ADC skewness (cutoff 7.2 × 10⁻²), tumor grade (grade 1 versus grades 2–3), and clinical tumor diameter (cutoff 20 mm). The model demonstrated sensitivities of 94% (training) and 86% (validation) and specificities of 80% and 78%, respectively. Its diagnostic performance, indicated by an area under the ROC curve (AUC) of 0.85, significantly outperformed both the mean ADC-based model (AUC = 0.54) and LN short-axis diameter criteria (AUC = 0.62) (p < 0.0001). In summary, integrating clinical factors with MRI-derived radiomic attributes using a machine learning approach provides a superior framework for predicting LN metastasis in EC compared to conventional ADC and size assessments.
%U https://galaxypub.co/article/integrating-clinical-and-diffusion-weighted-mri-radiomics-via-machine-learning-for-predicting-lymph-rxsfgt5twopeuv2