Work

Whole-tumor histogram analysis of non-Gaussian diffusion MRI models combined with clinicopathological indicators for assessing lymph node metastasis in endometrial cancer

Year: 2026

Type: article

Abstract: Objectives: To investigate whether whole-tumor histogram features derived from five non-Gaussian diffusion models, combined with clinicopathological factors, are associated with lymph node metastasis (LNM) in endometrial cancer (EC). Methods: This prospective study enrolled 73 patients with histologically confirmed EC and known nodal status. All underwent preoperative 3.0T MRI including diffusion-weighted imaging with 16 b-values. Thirteen parametric maps were derived from five non-Gaussian models: intravoxel incoherent motion (IVIM), diffusion kurtosis imaging (DKI), stretched exponential model (SEM), fractional-order calculus (FROC), and continuous time random walk (CTRW). Clinicopathological data were collected, including demographics, comorbidities, serum markers, pathological type, myometrial invasion, cervical involvement, and MRI-based local stage. Clinical, radiomics, and combined logistic regression models were built. Performance was evaluated using AUC, calibration curves, decision curve analysis, and bootstrap analysis of intercept variability. Results: = 0.090-0.140). The combined model had the lowest Brier score (0.092) and provided higher net benefit across threshold probabilities of 0.1-0.3. Conclusion: Non-Gaussian diffusion histogram parameters, particularly DKI_K_energy, demonstrate potential for predicting LNM in EC. A combined model incorporating DKI_K_energy with age and pathological type shows promise as a potential biomarker for risk stratification.

Source: Frontiers in Medicine

DOI: 10.3389/fmed.2026.1959334