Introduction and clinical context
Endometrial carcinoma (EC) is the second most common gynecologic cancer worldwide, with rising incidence and a spectrum of histological subtypes. The 2023 FIGO revision reshapes staging by emphasizing prognostic risk and tumor biology rather than purely anatomical spread. In this setting, magnetic resonance imaging (MRI) remains pivotal for preoperative evaluation. Beyond conventional diffusion-weighted imaging (DWI) and the apparent diffusion coefficient (ADC), intravoxel incoherent motion (IVIM) MRI offers a more nuanced view by separating true molecular diffusion from microvascular perfusion.
What IVIM adds to diffusion imaging
IVIM modeling yields three parameters: D (true diffusion), f (perfusion fraction), and D* (pseudo-diffusion). While ADC blends diffusion and perfusion effects, D isolates cellularity-driven diffusion, and f reflects tissue microvascularity. In aggressive EC subtypes (high-grade endometrioid, serous, clear cell, carcinosarcoma, undifferentiated), higher cellular density and altered vasculature can confound standard ADC interpretation. IVIM’s decoupling of diffusion and perfusion allows a potentially more accurate assessment of tumor biology relevant to FIGO 2023 risk stratification.
Study snapshot: IVIM vs ADC for predicting aggressive EC subtypes
In a prospective cohort of 34 pathologically confirmed EC cases, researchers evaluated IVIM-derived D, f, and D* alongside conventional ADC in distinguishing aggressive from non-aggressive subtypes as defined by FIGO 2023. The aggressive group comprised high-grade and non-endometrioid histologies, while the non-aggressive group included low-grade endometrioid carcinomas.
Key findings showed that aggressive ECs exhibited lower D, ADC, and f values compared with non-aggressive tumors. Among the parameters, D achieved the highest diagnostic performance with a cut-off of 0.56 × 10–3 mm²/s and an area under the curve (AUC) of 0.87, translating to 79.4% diagnostic accuracy. ADC followed as a strong discriminator (cut-off ≈ 0.53 × 10–3 mm²/s) with an accuracy around 73.5%, while f provided useful but slightly lesser discrimination. D* did not reach statistical significance, likely due to variability and SNR limitations inherent to perfusion-sensitive metrics.
These results reinforce a critical point: the D parameter from IVIM may serve as a more robust surrogate for tumor cellularity independent of perfusion, improving preoperative risk stratification in the FIGO 2023 era. The study also notes that while perfusion fraction (f) mirrors some vascular aspects, its diagnostic accuracy aligns more closely with ADC than with D.
Clinical implications for preoperative planning
Accurate prediction of aggressive EC subtypes before surgery informs decisions on lymphadenectomy, adjuvant therapy, and overall treatment intensity. If IVIM-D proves consistently superior across larger cohorts, radiologists may rely more on D as a non-invasive biomarker to stratify patients into risk categories aligned with FIGO 2023 biology-driven staging. This multiparametric approach, combining DWI, IVIM, and conventional MRI sequences, holds promise for refining prognosis and guiding individualized management plans without additional contrast-enhanced sequences.
Limitations and future directions
The cited study is prospective but single-centered with a relatively small sample size (n=34). IVIM acquisition used 10 b-values and faced longer scan times and potential motion artifacts, with D maps showing suboptimal reproducibility in some cases. D* exhibited high variability, limiting its current clinical reliability. Future work should prioritize multicenter validation, standardized acquisition and post-processing, and integration of IVIM with other imaging biomarkers and molecular data to build comprehensive predictive models for FIGO 2023 risk assessment.
Conclusion
In the evolving landscape of FIGO 2023 staging for endometrial cancer, IVIM-derived diffusion metrics—especially the D parameter—demonstrate superior diagnostic accuracy over ADC in predicting aggressive tumor subtypes. While promising, these findings require validation in larger cohorts before IVIM can be routinely deployed as a non-invasive, preoperative risk stratification tool.
