Introduction
Endometrial carcinoma (EC) remains a leading gynecologic cancer globally, with incidence and mortality trending upward in some regions. The 2023 FIGO update shifted the focus from purely anatomical staging to prognostic risk and tumor biology. In this context, functional MRI techniques—especially intravoxel incoherent motion (IVIM) diffusion imaging—are gaining attention as non-invasive biomarkers for preoperative risk stratification. This article summarizes a prospective study that compares IVIM-derived metrics with conventional diffusion-weighted imaging (DWI) metrics, notably the apparent diffusion coefficient (ADC), in predicting aggressive EC subtypes according to FIGO 2023.
Rationale: Why IVIM?
Conventional DWI using ADC maps reflects both true molecular diffusion and microvascular perfusion. In hypervascular tumors like aggressive EC, perfusion effects can confound diffusion measurements. IVIM separates true diffusion (D) from perfusion-related parameters: the perfusion fraction (f) and the pseudo-diffusion coefficient (D*). This separation may yield a more accurate assessment of tumor cellularity and vascularity, potentially improving discrimination between aggressive and non-aggressive subtypes under the 2023 FIGO framework.
Methods at a Glance
The study enrolled 34 patients with pathologically confirmed EC. MR imaging included non-contrast DWI and IVIM sequences on a 1.5T scanner. Regions of interest were placed to capture solid tumor portions, avoiding necrosis and cystic change. Histopathology served as the reference standard, with aggressive EC defined per FIGO 2023 as including high-grade endometrioid, serous, clear cell, carcinosarcoma, and undifferentiated carcinoma. A biexponential IVIM model yielded D, D*, and f values, which were compared to ADC in differentiating aggressive from non-aggressive subtypes.
Key Findings: D Value Takes the Lead
Across the cohort, aggressive EC demonstrated lower D, ADC, and f values compared with non-aggressive tumors. Notably, the D value achieved the highest diagnostic performance with an area under the ROC curve (AUC) of 0.87 and a cut-off of 0.56 × 10−3 mm2/s, yielding an overall accuracy of 79.4% in distinguishing aggressive from non-aggressive subtypes. By comparison, ADC achieved an AUC compatible with strong discrimination but slightly below D, while f provided a comparable but slightly inferior performance to ADC. D* did not reach statistical significance, likely due to SNR and intrinsic variability in pseudo-diffusion measurements.
Interpretation: Why D Outperforms ADC
The superior performance of D can be attributed to its independence from perfusion effects, offering a purer measure of tissue cellularity and microstructural restriction. In EC, where tumor biology and microvascular patterns vary with histology, D appears to reflect biologic differences with greater fidelity than ADC alone. The perfusion fraction (f) also differentiated subtypes, supporting the notion that vascularity contributes to tumor aggressiveness but with slightly less diagnostic accuracy than D. D* remained unreliable in this cohort due to technical limitations and noise.
Clinical Implications
If validated in larger, multicenter studies, IVIM-derived D could become a valuable non-invasive biomarker for preoperative risk stratification in EC. Integrating IVIM with conventional MRI could enhance multiparametric models, aiding decisions about surgical planning, adjuvant therapy, and surveillance in line with FIGO 2023 categories. This approach aligns with the trend toward precision imaging in gynecologic oncology, where imaging biomarkers help stratify patients by biologic risk rather than relying solely on anatomical staging.
Limitations and Future Directions
The study’s modest sample size (n=34) and single-center design limit generalizability. D maps sometimes showed suboptimal quality, and the addition of 10 b-values extended scan time, increasing susceptibility to motion. The DCE-MRI component was not incorporated, reflecting the study’s non-contrast focus. Future work should include larger, multicenter cohorts, standardized IVIM acquisition/post-processing, and exploration of integrating IVIM with molecular markers and other imaging biomarkers to construct robust multiparametric risk models for FIGO 2023 classification.
Conclusion
IVIM MRI, particularly the diffusion parameter D, demonstrates superior diagnostic accuracy over ADC in predicting aggressive EC subtypes in the context of FIGO 2023. While further validation is needed, IVIM holds promise as a non-invasive biomarker to refine preoperative risk assessment and guide personalized management for endometrial cancer patients.
