Endometrial cancer (EC) is one of the most common malignancies of the female reproductive system, and its incidence continues to rise globally, particularly in high-income countries [1]. According to statistics from the American Cancer Society, approximately 69,120 new cases and 13,860 related deaths are expected in the United States in 2025 [2]. EC has become a significant threat to women’s health worldwide, reducing quality of life and placing a considerable burden on healthcare systems.

In recent years, advances in imaging technology have made imaging evaluation pivotal in determining treatment strategies for EC. Multiparametric magnetic resonance imaging (mpMRI) enables comprehensive assessment of tumor characteristics, depth, and extent, including evaluation of deep myometrial invasion (DMI) and cervical stromal invasion (CSI) [3, 4]. mpMRI is increasingly recognized by gynecologists and has become a routine pre-treatment examination.

Previous study has demonstrated that different sequences, such as T2-weighted imaging (T2WI), diffusion-weighted imaging (DWI), and dynamic contrast-enhanced MRI (DCE-MRI), vary in their diagnostic efficacy for detecting tumors, as well as assessing DMI and CSI [5]. The current consensus on routine scanning protocols for EC includes T2WI, T1WI, DWI, and DCE-MRI sequences. Although this comprehensive protocol provides detailed information, the use of contrast agents prolongs scan time, increases examination costs, and adds to the burden on patients and healthcare systems. Moreover, gadolinium-based contrast agents may lead to adverse reactions in certain patients, such as nephrogenic systemic fibrosis, deposition of free gadolinium ions in tissues, and allergic responses (including anaphylactic shock), and in sporadic cases, can even result in a persistent vegetative state or fatality. The long-term clinical implications of gadolinium deposition remain uncertain [6]. Nevertheless, DCE-MRI is still widely performed in daily clinical practice.

The 2019 European Society of Urogenital Radiology (ESUR) guidelines note that whether the combination of T2WI and DWI is superior to DCE-MRI remains uncertain [7]. Therefore, identifying a method that relies solely on non-contrast sequences (e.g., T2WI and DWI) while maintaining diagnostic efficacy is clinically important.

In this study, we performed a retrospective analysis to maximize efficiency and reduce costs by using existing data, while enabling comparison across a relatively large sample size [8]. By evaluating different mpMRI sequence combinations and comparing their efficacy in assessing DMI and CSI, we aimed to determine whether diagnostic performance equivalent to conventional protocols can be achieved without contrasting agents, thereby potentially reducing the healthcare burden. Few studies have explored this topic, and our findings may provide new insights and form the basis for future clinical practice.

The choice of treatment strategy for EC depends on the local tumor stage, which involves evaluating the depth of myometrial invasion (50% or $\ge$50%), the presence or absence of CSI, extra-uterine spread, and invasion of adjacent organs [10]. The International Federation of Gynecology and Obstetrics (FIGO) revised the EC staging system in 2023, combining molecular and histological information to more accurately reflect tumor biology and prognosis. Nevertheless, imaging methods, including ultrasound, computed tomography (CT), MRI, and positron emission tomography (PET), remain fundamental for preoperative assessment of tumor spread. The roles and assessment criteria of CT and MRI have not changed; therefore, evaluation of DMI, CSI, lymph node metastasis, and distant metastasis remains necessary before treatment [11].

Pre-treatment MRI staging of EC is critical. The depth of myometrial invasion serves as a surrogate marker for lymph node metastasis and is associated with poorer prognosis. Therefore, accurate identification of DMI by MRI is essential, not only for local staging but also as a recognized adverse prognostic factor [12]. Axial MRI of the uterine corpus provides optimal visualization of the myometrium, allowing layer-by-layer assessment of tumor depth and extent.

Neves et al. [13] reported that T2WI combined with DWI achieved 92.1% accuracy in assessing myometrial invasion, comparable to the 90.52% observed in the T2&DWI group in our study. Stanzione et al. [14] analyzed 92 EC cases and found no significant difference in diagnostic accuracy among DWI, DCE, and conventional comprehensive MRI protocols for identifying DMI, consistent with our findings. Bonatti et al. [15] previously reported that DWI outperformed DCE in predicting DMI, although not statistically significant.

Wang et al. [16] conducted a meta-analysis of 19 studies involving 961 patients, which showed that in a subgroup of 7 studies, the AUCs of DWI combined with T2WI and DCE combined with T2WI for diagnosing DMI in EC were 0.959 (95% confidence interval [CI]: 0.932–0.986) and 0.929 (95% CI: 0.847–1.000), respectively, with no statistically significant difference between the two. The study further indicated that the use of DWI alone, DCE alone, or the combination of DWI and DCE were all reasonable approaches for diagnosing DMI. In addition, sensitivity analysis suggested that DWI may be more sensitive than DCE in detecting DMI. This conclusion is consistent with our findings. To corroborate this, we present a representative case from our study (Fig. 3), illustrating the potential advantage of DWI in identifying DMI. Based on our results and previous studies, individualized MRI protocols without DCE are feasible, offering advantages such as avoiding contrast agents, reduced invasiveness, lower cost, shorter acquisition time, and maintenance of diagnostic reliability for DMI.

Fig. 3.

In a 54-year-old patient with EC, postoperative pathology confirmed the presence of DMI. (A) Sagittal T2WI of the uterine body. (B–F) Axial images of the uterine body: (B) DWI (b = 800), (C) ADC_map, (D) T2WI, (E) early-phase, and (F) late-phase DCE images. Fig. 3B showed DMI of the anterior wall of the uterine corpus with a high signal area (as indicated by the white arrows). The corresponding ADC_map showed a slightly low signal shadow. However, the dynamic enhancement scans of early and late images did not show DMI. EC, endometrial cancer.

In the analysis of imaging characteristics of misdiagnosed cases in this study, we found that when tumor involvement reached approximately half of the uterine wall thickness, it was often difficult to measure the depth of infiltration accurately, leading to misinterpretation. We also observed that assessment inaccuracies were likely to occur under the following conditions, causing significant thinning or deformation of the myometrium: (1) large tumor volume; (2) presence of substantial intrauterine fluid and/or hematometra; and (3) marked atrophy of the uterine body. In such scenarios, a comprehensive MRI protocol should be employed for improved assessment accuracy. Through analysis of accurately assessed case imaging features, we found that when the T2&DWI group shows small tumor volume, no definite tumor signal, or no apparent signs of superficial myometrial invasion, the T2&DCE group usually yields similar assessment conclusions. Similarly, when the T2&DWI group clearly indicates DMI, the evaluation results of the two sequence groups are largely consistent. However, when the assessment of DMI is ambiguous in the T2&DWI group, the supplementary information provided by the T2&DCE group may offer additional value for the final evaluation. Based on these findings, we recommend that when the T2&DWI group can clearly determine either the absence of superficial myometrial invasion or the presence of DMI, the DCE sequence may not be strictly necessary. However, when image quality is suboptimal for assessing the superficial myometrium or when tumor involvement approximates half of the myometrial thickness, performing a DCE sequence may provide valuable assistance in evaluating DMI.

EC invades the cervical stroma in two main patterns [17, 18]: (1) simultaneous invasion of both the cervical mucosa and stroma, and (2) direct penetration of the cervical stroma through the uterine myometrium, bypassing the mucosa. EC with CSI is associated with lymph node metastasis and poor survival [19]. Assessment of CSI can help tailor surgical planning and provide fertility-preserving treatment options for young women without CSI [20].

In a 1.5 T MRI study for localized EC staging, including CSI, the accuracy, sensitivity, and specificity of DWI were 95%, 98%, and 80%, respectively. Based on these findings, Bonatti et al. [15] concluded that DWI could reliably replace contrast-enhanced MRI. Bi et al. [17] conducted a meta-analysis of 42 studies involving CSI and found that 3.0 T MRI offered higher diagnostic sensitivity, and DWI alone demonstrated higher sensitivity than DCE or T2WI alone. Lin et al. [21] analyzed 83 EC patients using 3.0 T MRI and reported that DWI, including ADC_maps, had significantly higher diagnostic efficacy for CSI compared with DCE (p = 0.009) or T2WI (p = 0.013). Hori et al. [22] studied 71 EC patients, including 6 cases of CSI and showed that DWI and DCE had almost identical performance in accuracy, sensitivity, and specificity for CSI at 3.0 T MRI. Our study enrolled a total of 138 EC patients, of whom 12 were pathologically confirmed to have CSI. Statistical analysis indicated that there was no statistically significant difference between the T2&DWI and T2&DCE groups in assessing CSI, a finding consistent with previous studies.

Analysis of misdiagnosed cases based on imaging features revealed that underestimation by both sequence combinations was attributed to the absence of definitive signs of CSI on MRI. Overestimation in two cases by the T2&DWI group was due to imaging features suggesting suspected minimal tumor involvement of the cervical stroma. Furthermore, underestimation by the T2&DWI group and overestimation by the T2&DCE group were primarily caused by the larger tumor tissue extending into the cervical canal, resulting in marked thinning of the cervical wall and altered signal intensity, which made it difficult to distinguish the tumor from the surrounding cervical stroma. Based on these findings, we recommend that the DCE sequence may not be strictly necessary when the T2&DWI group shows no tumor involvement of the uterine isthmus. In contrast, DCE should be considered when the tumor extends deeply into the isthmus or endocervical canal to aid in diagnosis.

In this study, we compared the T2&DWI group with the T2&DCE group, as well as high-sensitivity and high-specificity combinations of MRI sequences. Preliminary results indicated that the T2&DWI group had a diagnostic performance similar to the other groups for assessing DMI and CSI, with no statistically significant differences. Based on the above findings, we propose that in clinical practice, the combination of T2WI and DWI may replace contrast-enhanced imaging under specific conditions. When the T2&DWI group clearly demonstrates either the absence of superficial myometrial invasion or the presence of DMI, and when the tumor does not involve the uterine isthmus, DCE may not be strictly necessary. This approach is particularly suitable for patients with contraindications, such as renal impairment or allergies to contrast agents [23]. Adopting a contrast-free MRI protocol consisting solely of T2WI and DWI can help reduce the medical burden on patients while maintaining diagnostic accuracy. However, when the T2&DWI combination provides unclear visualization of the superficial myometrium, suggests tumor involvement of approximately half of the myometrial thickness, or indicates tumor extension into the isthmus or the endocervical canal, the addition of DCE is recommended to improve diagnostic reliability. In actual clinical decision-making, patients’ overall health status and economic factors should also be comprehensively considered. This study may offer new perspectives for optimizing the MRI evaluation pathway in EC.

In summary, the T2WI & DWI combination demonstrated comparable efficacy to the T2WI & DCE combination for preoperative assessment of DMI and CSI in EC, with no statistically significant differences. These findings provide new evidence supporting the use of unenhanced MRI protocols in specific patient populations, particularly those with contrast allergies or renal insufficiency. Future research should continue to explore more efficient, contrast-free MRI protocols to optimize assessment of EC extent.

The data that support the findings of this study are available from the corresponding author upon reasonable request.

JJL, BLR, and SYW designed the research study. FJW and SYW provided assistance and recommendations for the MRI scanning protocol and image interpretation standards. JJX was responsible for collecting and organizing the cases. KHL and JJX each evaluated the patients’ imaging studies, and FJW served as the arbitrator for disputed cases. JJL and JJX performed statistical analysis on the data and interpreted the results. JJL and KHL drafted this manuscript, with revisions provided by BLR and SYW. All authors contributed to the manuscript revision, and all approved the final version. All authors have participated sufficiently in the work and agreed to be accountable for all aspects of the work.

The study was carried out in accordance with the guidelines of the Declaration of Helsinki. The protocol was approved by the Ethics Committee of the First Hospital of Putian City (Approval Number: 2022-020), and no special informed consent was required because anonymous patient records were used.

We want to express our gratitude to all those who helped us in writing this manuscript. We also thank the peer reviewers for their valuable comments and suggestions. We wish to acknowledge the Editor for their valuable input and assistance with this manuscript.

The study was supported by the Putian University Scientific Research Fund Project (Grant No.: 2022098).

The authors declare no conflict of interest.

Supplementary material associated with this article can be found, in the online version, at https://doi.org/10.31083/CEOG45908.