Introduction: IVIM MRI and the FIGO 2023 framework
Endometrial carcinoma (EC) remains a leading gynecologic cancer worldwide, with rising incidence and diverse histological subtypes. The FIGO 2023 classification emphasizes prognosis based on biological behavior, separating tumors into aggressive and non-aggressive groups. Magnetic resonance imaging (MRI) has long been the backbone of preoperative assessment, balancing anatomical staging with functional information. In recent years, diffusion imaging has evolved from conventional diffusion-weighted imaging (DWI) and apparent diffusion coefficient (ADC) maps to more advanced techniques like intravoxel incoherent motion (IVIM) MRI. This prospective study investigates whether IVIM metrics (D, D*, f) can better predict aggressive EC subtypes than standard DWI-derived ADC values.
What IVIM adds to diffusion imaging
Conventional ADC maps conflate true water diffusion and microvascular perfusion, which can blur differences in tumors with hypervascularity or variable perfusion. IVIM offers a biexponential model that separates D (true diffusion), D* (pseudo-diffusion related to microcirculation), and f (perfusion fraction). By disentangling diffusion from perfusion, IVIM aims to provide a more accurate read on tumor cellularity and vascularity — important factors in tumor aggressiveness and histology under FIGO 2023.
Key findings: D stands out as the strongest discriminator
The study enrolled 34 pathologically confirmed EC cases, categorized into aggressive (high grade endometrioid, serous, clear cell, carcinosarcoma, undifferentiated) and non-aggressive (low grade endometrioid) groups. The main results show:
- Aggressive tumors had lower diffusion metrics overall, with D and ADC values reduced compared to non-aggressive lesions.
- D demonstrated the highest diagnostic accuracy for distinguishing aggressive from non-aggressive EC, with a cut-off around 0.56 × 10−3 mm²/s and an AUC of 0.87, yielding about 79.4% accuracy.
- ADC also differentiated groups but with lower performance (cut-off ~0.53 × 10−3 mm²/s; accuracy near 73.5%).
- The perfusion fraction f differed between groups, but its diagnostic accuracy was slightly inferior to D and close to ADC; D* showed no statistically meaningful separation due to variability.
These findings indicate that, among IVIM parameters, the diffusion component D most reliably flags aggressive EC subtypes, aligning with the notion that higher cellular density reduces true diffusion. The perfusion component f also contributes meaningful information but may be more variable in routine practice.
Clinical implications: A non-invasive risk stratification tool
With FIGO 2023 placing greater emphasis on tumor biology for prognosis, preoperative risk stratification becomes crucial for surgical planning and adjuvant strategies. The study supports incorporating IVIM into MRI protocols, particularly to identify aggressive EC subtypes non-invasively. When integrated with conventional DWI and other multiparametric MRI features, IVIM could enhance preoperative counseling, help tailor surgical approaches, and refine risk-adapted therapy.
Limitations and future directions
Key limitations include the single-center setting and a relatively small cohort (n=34), which may affect generalizability. D maps sometimes exhibited suboptimal image quality, and the use of 10 b-values lengthened scan time, increasing motion risk. DCE-MRI was not included, as the focus was a non-contrast IVIM approach. The D* parameter showed high variability and limited reliability in this study. Future multicenter trials with standardized IVIM acquisition and post-processing protocols are needed to validate these findings. Larger datasets could enable robust multiparametric models that combine IVIM with other imaging biomarkers and molecular markers for precise risk stratification under FIGO 2023.
Conclusion: IVIM–D as a promising biomarker for aggressive EC
Overall, the study demonstrates that IVIM diffusion parameter D has superior diagnostic accuracy over ADC in predicting aggressive endometrial carcinoma subtypes. While IVIM adds complexity to imaging protocols, its potential to improve preoperative risk assessment and guide treatment decisions supports continued research and integration into routine MRI workflows.
