Abstract
Objective
To evaluate the efficiency of diffusion weighted imaging (DWI), the mean apparent diffusion coefficient (ADCmean) and the minimum apparent diffusion coefficient (ADCmin) values with different b-values (800 s/mm² and 1000 s/mm²) in the diagnosis and staging of endometrial carcinoma (EC).
Methods
Preoperative DWI images of 412 patients with EC and 134 patients with benign endometrial lesions were analyzed retrospectively. The performance of DWI images, ADCmean and ADCmin values with different b-values (800 s/mm² and 1000 s/mm²) for the diagnosis and staging (deep myometrial invasion, cervical stromal invasion and lymph node metastasis) of EC was assessed by using receiver operating characteristic curve (ROC). The comparison between AUCs was performed using the DeLong test, and a P value < 0.05 was considered statistically significant.
Results
The area under the curves (AUCs) of DWI protocol with b = 1000 s/mm2 for qualitative assessment of EC diagnosis and staging (0.850, 0.837, 0.906, and 0.820 for diagnosis, deep myometrial invasion, cervical stromal invasion and lymph node metastasis, respectively) were higher than those of b = 800 s/mm2 (0.821, 0.795, 0.860 and 0.814, respectively) (all p < 0.05). The AUCs for the quantitative assessment of EC diagnosis and staging with ADCmean and ADCmin values with b = 1000 s/mm2 were higher than those of b = 800 s/mm2 (all p < 0.05).
Conclusions
DWI images, ADCmean and ADCmin values with b = 1000 s/mm2 had higher performance than those of b = 800 s/mm2 in the assessment of EC diagnosis and staging. This study highlights the potential of using b = 1000 s/mm² as an optimized protocol for EC assessment in clinical practice.
Keywords: Diffusion weighted imaging, Apparent diffusion coefficient, Endometrial carcinoma, B-value, Diagnosis
Introduction
Endometrial carcinoma (EC) is the sixth most common cancer in women worldwide [1], and its incidence is increasing significantly [2, 3]. Endometrial hyperplasia and polyps are the most common benign endometrial diseases which have resembled clinical symptoms with EC. Therefore, it is crucial to differentiate the benign and malignant endometrial diseases before treatment [4]. Clinically, the primary treatment for EC is surgery for all stages, but depending on the stage, adjuvant therapies such as chemotherapy, radiotherapy, or hormone therapy may also be provided. The International Federation of Gynecology and Obstetrics (FIGO) surgical staging system [5], revised in 2023, is currently used for EC staging. In the new classification, the major determinants for staging are histopathology and molecular features. While the morphological and imaging factors, such as the depth of myometrial invasion, the presence of cervical invasion, and the status of lymphatic system, still play a role, their impact on staging is less than that in the previous system [5, 6]. It is of great clinical significance to explore the best testing method for EC based on the above factors.
Magnetic resonance imaging (MRI) pelvic examination is considered to be the most accurate modality for preoperative assessment of endometrial lesions. MRI pelvic examination was recommended by the American College of Radiology (ACR) as the best method for treatment planning [7]. Diffusion weighted imaging (DWI) can improve the sensitivity and accuracy of the diagnosis of malignant endometrial lesions and the Apparent diffusion coefficient (ADC) values can quantificationally identify benign and malignant endometrial lesions. Moreover, the monoexponential diffusion model DWI is still the most prevalent diffusion sequence utilized in clinical practice [7, 8]. But, there is a lack of standardization of the ADC values in the clinical use due to the range of b-value affects the calculation of ADC values [9]. Li et al. [10] The ability of different mathematical models of multiple b-value DWI in predicting lymphovascular invasion in endometrial cancer was not explored for different stages. The European Society of Urogenital Radiology (ESUR) panel recommends including DWI sequence to stage EC, the added value can be provide by DWI for EC staging, particularly for assessing the depth of myometrial invasion (a minimum of two b values of 0 and 800–1000 s/mm2 are recommended) [11–13].
It has reported that b-value has an effect on ability to diagnose [12]. However, the b-value of DWI has not been standardized due to the use of different MRI devices and scan parameters at different medical centers. There are currently no studies regarding which b-value has the best diagnostic and staging performance for EC. Therefore, the aim of this study was to compare the efficiency of DWI, the mean apparent diffusion coefficient (ADCmean) and the minimum apparent diffusion coefficient (ADCmin) at different b-values for the diagnosis and staging of EC.
Materials and methods
Study populations
Ethical approval for this study was granted by the ethics committee of the First People’s Hospital of Yunnan Province. The need for written informed consent to participate was waived by the ethics committee of the hospital due to retrospective nature of the study. Preoperative DWI images of 412 patients with pathologically confirmed EC (395 endometrioid adenocarcinoma, 12 serous carcinoma, 3 clear cell carcinoma, 1 undifferentiated endometrioid carcinoma, and 1 carcinosarcoma) and 134 patients with benign endometrial lesions (60 endometrial hyperplasia without atypia, 43 atypical hyperplasia and 31 endometrial polyps) between January 2017 and March 2022 in our hospital were collected. The patient inclusion standards were as follows: (1) underwent preoperative pelvic MRI examination containing DWI sequences with a b-value of 800 s/mm2 or 1000 s/mm2 for uterine axial images; (2) no more than 2 weeks between MRI examination and surgery; and (3) complete clinical data. The exclusion criteria were as follows: (1) history of other pelvic malignant tumors (n = 5); (2) MR images with poor quality (patient motion, physiological factors, post-processing errors) (n = 22); and (3) patient who had received endocrine therapy, chemotherapy or radiotherapy before MRI examination (n = 3). Finally, using postoperative histopathological findings as the gold standard, all patients were divided into benign and malignant groups, and EC patients were then divided into a deep/superficial myometrial invasion group, with/without cervical stromal invasion group and with/without lymph node metastasis group based on FIGO staging.
MRI protocols
This study was performed on a GE 3.0T MRI scanner (GE Signa HDXt, America) equipped with an 8-channel abdominal phased-array coil. Each patient underwent a preoperative pelvic MRI examination with a standard protocol. Some parameters were adjusted according to individual differences of patients. In the present study, we analyzed uterine axial DWI (b = 50, 800/1000 s/mm2), and ADC maps. ADC maps required manual reconstruction using ADW 4.7 FuncTool software. Although T2-weighted imaging (T2WI) is commonly used in clinical practice for its high spatial resolution and morphological information, this study specifically evaluated the diagnostic performance of DWI and ADC values alone to provide a standardized assessment of diffusion characteristics in endometrial lesions. In this study, the specific parameters of the DWI sequence are as follows: oblique axial imaging was performed with b-values of 50 and 800/1000 s/mm². The repetition time (TR) was 4250 ms, the echo time (TE) was 70 ms, the field of view (FOV) was 200 × 200 mm², the matrix was 240 × 240, the slice thickness was 3 mm, the slice gap was 1 mm, and the number of excitations was 1–4. The DWI imaging used a single-shot echo-planar imaging (SE-EPI) sequence and was analyzed using the intravoxel incoherent motion (IVIM) model. ADC maps were manually reconstructed using workstation software (ADW 4.7 FuncTool).
Image analysis
Image analysis was performed independently by two radiologists with 3 and 8 years of experience, respectively, in pelvic MRI interpretation. Both radiologists were unaware of the patient’s histopathological findings and FIGO staging. First, they distinguish EC from benign endometrial lesions and detected deep myometrial invasion, cervical stromal invasion and lymph node metastases of EC only by observing the DWI images. Then, ADCmean and ADCmin values were measured on the ADC maps for quantitative assessment of the above indicators by the above two radiologists. According to a previous report [14], EC with unclear outline, presented significant high signal on DWI with high b-values and low signal on the corresponding ADC map. According to the FIGO staging system, the criteria for deep myometrial invasion was greater than 50% involvement of the myometrial. Normally, lymph node metastasis was routinely evaluated by lymph node with a short diameter greater than 1 cm, but the lymph node with a short diameter greater than 0.8 cm were included when they presented rounded or irregular shapes, hairy margins, or low signal on ADC maps. To measure the ADC values, the two radiologists placed the region of interests (ROIs) in the largest level of the solid components of the lesions. The round ROIs were as large as possible and avoided bleeding and fluid in the uterine cavity and the normal myometrial [15].
Statistical analysis
All data analyses were performed by R software (v4.0.3, Foundation for Statistical Computing, Vienna, Austria) and SPSS 27.0 (IBM, New York, USA). We calculated accuracy, sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) under two protocols (b = 800 s/mm2 and b = 1000 s/mm2). The performance of the two protocols was analyzed using the receiver operating characteristic curve (ROC) and the area under the curve (AUC). The hypothesis testing comparison between AUCs was performed using the DeLong test, and a P value < 0.05 was considered statistically significant. Interobserver agreement for qualitative diagnosis under the two protocols was analyzed using the Cohen’s Kappa statistic, which was interpreted as 0.1 ≤ k < 0.2, poor; 0.2 ≤ k < 0.4, fair; 0.4 ≤ k < 0.6, moderate; 0.6 ≤ k < 0.8, substantial; and 0.8 ≤ k < 1.0, excellent [16].
Results
Consistency test
The basic histopathological feature information of patients has been presented in Table 1. Interobserver agreement between DWI and ADC values for qualitative diagnosis with both protocols was evaluated using Cohen’s Kappa analysis, and the results were substantial and above, with b = 800 s/mm2 and b = 1000 s/mm2, respectively: DWI: k = 0.789 vs. 0.800; ADCmean: k = 0.798 vs. 0.802; ADCmin: k = 0.737 vs. 0.732.
Table 1.
Patients histopathological features
| Characteristics | Number |
|---|---|
| DMI | |
| < 50% | 300 |
| ≥ 50% | 84 |
| FIGO stage | |
| Stage IA | 234 |
| Stage IB | 84 |
| Stage II | 43 |
| Stage III–IV | 23 |
Abbreviations: DMI Depth of myometrial invasion, FIGO International Federation of Gynecology and Obstetrics (2009)
Qualitative assessment of DWI
The results of the qualitative evaluation of DWI for EC diagnosis and staging are displayed in Table 2; Figs. 1 and 2. The AUCs of DWI images with b = 1000 s/mm2 (0.850, 0.837, 0.906, and 0.820 for EC diagnosis, deep myometrial invasion, cervical stromal invasion and lymph node metastasis, respectively) were higher than those of b = 800 s/mm2 (0.821, 0.795, 0.860, and 0.814, respectively) (all P > 0.05). Meanwhile, the accuracies of qualitative evaluation of DWI images with b = 1000 s/mm2 (0.859, 0.864, 0.930, and 0.974 for EC diagnosis, deep myometrial invasion, cervical stromal invasion and lymph node metastasis, respectively) were also higher than those of b = 800 s/mm2 (0.795, 0.804, 0.890, and 0.939, respectively). The specificities of both protocols (b = 1000 s/mm2 and b = 800 s/mm2) were high for EC staging (all specificities > 0.900), the sensitivities were comparatively low (all sensitivities < 0.650 except for detecting lymph node metastasis at b = 1000 s/mm2).
Table 2.
Diagnostic performance of DWI with different b-values
| N | TP | FP | FN | TN | Accuracy (%) | Sensitivity (%) | Specificity (%) |
PPV(%) | NPV(%) | AUC | P | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| EC diagnosis(yes/no) | ||||||||||||
| b = 800 s/mm2 | 112 | 70 | 3 | 20 | 19 | 0.795 | 0.959 | 0.487 | 0.778 | 0.864 | 0.821 | 0.542 |
| b = 1000 s/mm2 | 434 | 280 | 19 | 42 | 93 | 0.859 | 0.936 | 0.689 | 0.870 | 0.830 | 0.850 | |
| Deep myometrial invasion(with/without) | ||||||||||||
| b = 800 s/mm2 | 82 | 17 | 11 | 5 | 49 | 0.804 | 0.607 | 0.907 | 0.773 | 0.817 | 0.795 | 0.478 |
| b = 1000 s/mm2 | 302 | 49 | 28 | 13 | 212 | 0.864 | 0.636 | 0.942 | 0.790 | 0.883 | 0.837 | |
| Cervical stromal invasion(with/without) | ||||||||||||
| b = 800 s/mm2 | 82 | 9 | 7 | 2 | 64 | 0.890 | 0.563 | 0.970 | 0.818 | 0.901 | 0.860 | 0.513 |
| b = 1000 s/mm2 | 302 | 28 | 17 | 4 | 253 | 0.930 | 0.622 | 0.984 | 0.875 | 0.937 | 0.906 | |
| Lymph node metastasis(with/without) | ||||||||||||
| b = 800 s/mm2 | 82 | 4 | 3 | 2 | 73 | 0.939 | 0.571 | 0.973 | 0.667 | 0.961 | 0.814 | 0.958 |
| b = 1000 s/mm2 | 302 | 11 | 2 | 6 | 283 | 0.974 | 0.846 | 0.979 | 0.647 | 0.993 | 0.820 | |
EC Endometrial carcinoma, DWI Diffusion weighted imaging, N Number, TP True positive, FP False positive, FN False negative, TN True negative, AUC Area under the curve, PPV Positive predict value, NPV Negative predict value
Fig. 1.
DWI and ADC maps of patients with benign endometrial lesions and early stage EC. The endometrial hyperplasia (a and b), EC with superficial myometrial invasion (c and d), and EC with deep myometrial invasion (e and f) were correctly diagnosed and staging on uterus-axial DWI with b = 1000 s/mm2. The endometrial hyperplasia (g and h) was also correctly diagnosed on uterus-axial DWI with b = 800 s/mm2. EC with superficial myometrial invasion (i and j) and deep myometrial invasion (k and l) were misjudged on DWI with b = 800 s/mm2uterus-axial
Fig. 2.
DWI and ADC maps of patients with advanced stage EC. EC with cervical stromal invasion (a and b) and EC with lymph node metastasis (c and d) were correctly diagnosed on DWI with b = 1000 s/mm2. EC with cervical stromal invasion (e and f) and with lymph node metastasis (g and h) were misjudged on DWI with b = 800 s/mm2
Quantitative assessment of ADCmean values
The results of the quantitative evaluation of ADC values for the diagnosis and staging of EC are shown in Table 3; Fig. 3. The ADCmean values with both protocols (b = 1000 s/mm2 and b = 800 s/mm2) showed higher AUCs (0.927 vs. 0.924, respectively) for the assessment of EC diagnosis. Meanwhile, they were superior to the performance of visual assessment. The AUCs of ADCmean values with b = 1000 s/mm2 for EC staging (0.863, 0.810 and 0.842 for deep myometrial invasion, cervical stromal invasion and lymph node metastasis, respectively) were higher than those of b = 800 s/mm2 (0.846, 0.713 and 0.769, respectively) (all P > 0.05). The effect sizes for EC diagnosis, deep myometrial invasion, and cervical stromal invasion were all below the conventional threshold for a small effect (Cohen’s d < 0.2). The accuracies of ADCmean values at b = 1000 s/mm2 (0.857, 0.795, 0.692, and 0.662 for EC diagnosis, deep myometrial invasion, cervical stromal invasion and lymph node metastasis, respectively) were higher than those of b = 800 s/mm2 (0.830, 0.765, 0.622, and 0.573, respectively). The sensitivities of ADCmean of both protocols were high for EC staging (all sensitivities > 0.800). The cutoff values for ADCmean to assess the different stages of EC were very stable at both protocols (ranged from 0.760 to 0.797 × 10−3 mm2/s). Details of cutoff values of ADC values for the diagnosis and staging of EC are shown in Fig. 4.
Table 3.
Diagnostic performance of ADCmean values with different b-values
| ADCmean | Accuracy (%) | Sensitivity (%) | Specificity (%) | PPV (%) |
NPV (%) |
AUC | P | Cutoff values | ||
|---|---|---|---|---|---|---|---|---|---|---|
|
EC Diagnosis |
b = 800 s/mm2 | |||||||||
| EC (n=90) | 0.849±0.198 | 0.830 | 0.811 | 0.909 | 0.973 | 0.541 | 0.924 | 0.904 | 1.062 | |
| Benign lesions (n=22) | 1.257±0.180 | |||||||||
| b = 1000 s/mm2 | ||||||||||
| EC (n=322) | 0.847±0.161 | 0.857 | 0.835 | 0.919 | 0.968 | 0.660 | 0.927 | 1.000 | ||
| Benign lesions (n=112) | 1.199±0.177 | |||||||||
| Deep myometrial invasion | b = 800 s/mm2 | |||||||||
| Yes (n=22) | 0.685±0.063 | 0.765 | 1.000 | 0.667 | 0.523 | 1.000 | 0.846 | 0.705 | 0.773 | |
| No (n=60) | 0.907±0.204 | |||||||||
| b = 1000 s/mm2 | ||||||||||
| Yes (n=62) | 0.699±0.079 | 0.795 | 0.942 | 0.736 | 0.587 | 0.970 | 0.863 | 0.797 | ||
| No (n=240) | 0.886±0.157 | |||||||||
| Cervical stromal invasion | b = 800 s/mm2 | |||||||||
| Yes (n=11) | 0.730±0.125 | 0.622 | 0.818 | 0.592 | 0.237 | 0.955 | 0.713 | 0.244 | 0.760 | |
| No (n=71) | 0.866±0.207 | |||||||||
| b = 1000 s/mm2 | ||||||||||
| Yes (n=32) | 0.707±0.081 | 0.692 | 0.875 | 0.670 | 0.239 | 0.978 | 0.810 | 0.777 | ||
| Lymph node metastasis | No (n=270) | 0.864±0.162 | ||||||||
| b = 800 s/mm2 | ||||||||||
| Yes (n=6) | 0.693±0.049 | 0.573 | 1.000 | 0.539 | 0.146 | 1.000 | 0.769 | 0.346 | 0.767 | |
| No (n=76) | 0.860±0.206 | |||||||||
| b = 1000 s/mm2 | ||||||||||
| Yes (n=17) | 0.680±0.078 | 0.662 | 0.941 | 0.646 | 0.137 | 0.995 | 0.842 | 0.777 | ||
| No (n=285) | 0.857±0.161 | |||||||||
EC Endometrial carcinoma, DWI Diffusion weighted imaging, AUC Area under the curve, PPV Positive predict value, NPV Negative predict value
Fig. 3.
ROC plots for ADCmean and ADCmin values with different b-values. The AUCs of ADCmean (a-d) and ADCmin values (e-h) at b = 1000 s/mm2 were higher than those of b = 800 s/mm2. There was a significantly difference for AUC of detecting lymph node metastasis between the ADCmin values at b = 1000 s/mm2 and that of b = 800 s/mm2
Fig. 4.
Box plots of ADCmean (a) and ADCmin (b) values for different b-values. The red triangles in the figure represented the cutoff values. The cutoff values of ADCmean for different stages of EC were very similar, and they were close to the median. The data distribution was more dispersed at b = 800 s/mm2 protocol and more concentrated at b = 1000 s/mm2
Quantitative assessment of ADCmin values
The AUCs of ADCmin values with b = 1000 s/mm2 (0.918, 0.849, 0.789, and 0.829 for EC diagnosis, deep myometrial invasion, cervical stromal invasion and lymph node metastasis, respectively) were higher than those of b = 800 s/mm2 (0.903, 0.848, 0.691, and 0.655, respectively) (all P > 0.05 excepted lymph node metastasis). The accuracies of ADCmin values at b = 1000 s/mm2 were 0.855, 0.755, 0.692, and 0.682 for EC diagnosis, deep myometrial invasion, cervical stromal invasion and lymph node metastasis, respectively. And the accuracies of ADCmin values at b = 800 s/mm2 were 0.795, 0.854, 0.561, and 0.646, respectively. The sensitivities of ADCmin with b = 1000 s/mm2 were high for EC diagnosis and staging (all sensitivities > 0.800), but the ADCmin with b = 800 s/mm2 were comparatively low in evaluation of deep myometrial invasion (sensitivities = 0.638) (Table 4). The cutoff values of ADCmin to assess the different stages of EC with both protocols ranged from 0.574 to 0.705 × 10−3 mm2/s.
Table 4.
Diagnostic performance of ADCmin values with different b-values
| ADCmin | Accuracy (%) |
Sensitivity (%) |
Specificity (%) |
PPV (%) |
NPV (%) |
AUC | P | Cutoff values | ||
|---|---|---|---|---|---|---|---|---|---|---|
|
EC Diagnosis |
b = 800 s/mm2 | |||||||||
| EC (n = 90) | 0.716 ± 0.175 | 0.795 | 0.744 | 1.000 | 1.000 | 0.489 | 0.903 | 0.642 | 0.824 | |
| Benign lesions (n = 22) | 1.056 ± 0.195 | |||||||||
| b = 1000 s/mm2 | ||||||||||
| EC (n = 322) | 0.755 ± 0.141 | 0.855 | 0.829 | 0.929 | 0.971 | 0.654 | 0.918 | 0.897 | ||
| Benign lesions (n = 112) | 1.044 ± 0.164 | |||||||||
| Deep myometrial invasion | b = 800 s/mm2 | |||||||||
| Yes (n = 22) | 0.570 ± 0.078 | 0.854 | 0.636 | 0.933 | 0.778 | 0.875 | 0.848 | 0.989 | 0.574 | |
| No (n = 60) | 0.771 ± 0.176 | |||||||||
| b = 1000 s/mm2 | ||||||||||
| Yes (n = 62) | 0.632 ± 0.080 | 0.755 | 0.855 | 0.729 | 0.449 | 0.951 | 0.849 | 0.705 | ||
| No (n = 240) | 0.791 ± 0.133 | |||||||||
| Cervical stromal invasion | b = 800 s/mm2 | |||||||||
| Yes (n = 11) | 0.620 ± 0.102 | 0.561 | 0.909 | 0.507 | 0.222 | 0.973 | 0.691 | 0.221 | 0.692 | |
| No (n = 71) | 0.732 ± 0.185 | |||||||||
| b = 1000 s/mm2 | ||||||||||
| Yes (n = 32) | 0.642 ± 0.082 | 0.692 | 0.844 | 0.674 | 0.235 | 0.973 | 0.789 | 0.702 | ||
| No (n = 270) | 0.773 ± 0.139 | |||||||||
| Lymph node metastasis | b = 800 s/mm2 | |||||||||
| Yes (n = 6) | 0.615 ± 0.051 | 0.646 | 0.833 | 0.632 | 0.152 | 0.980 | 0.655 | 0.029* | 0.620 | |
| No (n = 76) | 0.725 ± 0.184 | |||||||||
| b = 1000 s/mm2 | ||||||||||
| Yes (n = 17) | 0.619 ± 0.075 | 0.682 | 0.882 | 0.674 | 0.139 | 0.990 | 0.829 | 0.692 | ||
| No (n = 285) | 0.767 ± 0.138 | |||||||||
EC Endometrial carcinoma, DWI Diffusion weighted imaging, AUC Area under the curve, PPV Positive predict value, NPV Negative predict value
*significant statistical difference
Discussion
MRI is currently the most reliable modality for preoperative staging of EC, and DWI is regarded as the most convenient, rapid, and non-invasive MRI technique [8, 12]. Meanwhile, DWI is also considered as an imaging biomarker of tumors and has been extensively used in the diagnosis of gynecologic tumors [2, 17, 18]. However, the option of b-value has a considerable impact on the visual assessment of DWI image and the calculation of ADC value [19]. There have been many studies exploring the optimal b-value for pelvic tumors such as prostate cancer, rectal cancer, and cervical cancer. and some researchers have found that the DWI with b = 1000 s/mm2 has better performance in pelvic scans [20, 21]. The use of an optimized b-value may be an important step in the optimization of DWI. This retrospective study compared the diagnostic performance of DWI, the ADCmean and ADCmin values with b = 1000 s/mm2 and b = 800 s/mm2 for the diagnosis and staging of EC. The results demonstrated that the protocol of DWI and ADC values with b = 1000 s/mm2 had all slightly higher performance than those of b = 800 s/mm2 for the assessment of EC diagnosis and staging. However, the improvement was only slightly higher, and b = 800 s/mm2 could still be used as an alternative, especially in clinical settings where b = 1000 s/mm2 is not available or feasible. It is worth noting that other studies have used multiple b-values, such as 50, 500, and 1000 s/mm2, for DWI in similar contexts [8]. The use of multiple b-values can potentially provide more comprehensive information about the diffusion characteristics of tissues. However, our study focused on comparing the two commonly used b-values (800 and 1000 s/mm2) to determine their relative diagnostic performance. The results suggest that while b = 1000 s/mm2 provides slightly better diagnostic accuracy and stability, b = 800 s/mm2 remains a viable option, especially considering that the differences in performance were not statistically significant in all parameters. And the ADC values with b = 1000 s/mm2 had higher accuracies and better stabilities. The better stabilities of the b = 1000 s/mm2 protocol can be attributed to several factors. First, higher b-values (such as b = 1000 s/mm2) provide greater sensitivity to diffusion, which enhances the contrast between tissues with different diffusion properties. This increased contrast allows for more accurate differentiation between benign and malignant endometrial lesions, as well as more precise assessment of myometrial invasion, cervical stromal invasion, and lymph node metastasis. Second, the ADC values derived from higher b-values tend to be more stable and less influenced by T2 shine-through effects, which can interfere with the accuracy of ADC measurements at lower b-values [2, 18]. In our study, the cutoff values for ADCmean and ADCmin at b = 1000 s/mm2 were more consistent and had a narrower range compared to those at b = 800 s/mm2 (Fig. 4). This suggests that the b = 1000 s/mm2 protocol provides more reliable and reproducible ADC values, which are crucial for accurate diagnosis and staging.
Regarding the cutoff value for ADC in the diagnosis of lymph node metastases, our study found that the ADCmean values with b = 1000 s/mm2 had a cutoff value of 0.777 × 10 − 3 mm2/s, while the ADCmin values had a cutoff value of 0.692 × 10 − 3 mm2/s. These cutoff values were determined based on the ROC analysis and were found to be highly effective in differentiating between metastatic and non-metastatic lymph nodes. The use of these cutoff values in conjunction with established morphological, criteria such as lymph node size and shape, can significantly enhance the diagnostic accuracy for lymph node metastases. Specifically, lymph nodes with a short diameter greater than 1 cm are typically considered suspicious for metastasis, but the inclusion of ADC values can provide additional quantitative evidence to support the diagnosis, especially for smaller lymph nodes with a short diameter greater than 0.8 cm that exhibit rounded or irregular shapes, hairy margins, or low signal on ADC maps [7]. Our findings suggest that incorporating ADC values into the diagnostic workflow can complement morphological criteria and improve the sensitivity and specificity of lymph node metastasis detection in patients with endometrial carcinoma.
Most guidelines for assessing the female pelvis, such as those by the ESUR, advise a high b-value of >800–1000 s/mm2 for female pelvic imaging [13]. This recommendation is based on the improved sensitivity and specificity of DWI at higher b-values, which can better delineate the diffusion characteristics of lesions and reduce the T2 shine-through effect. Our study supports this guideline by demonstrating that DWI with b = 1000 s/mm2 outperforms b = 800 s/mm2in both qualitative and quantitative assessments for EC diagnosis and staging.
The study of Basaran et al. [17] with 88 patients (36 malignant cases and 52 benign cases) reported the accuracy, sensitivity and specificity of EC diagnosis with visual evaluation on DWI with b = 1000 s/mm2 were 74%, 81% and 69% and 69% respectively. In our study, the accuracy and sensitivity were relatively higher compared to the study. Which might be related to the increase in the number of malignant tumors.Rechichi et al. found the accuracy, sensitivity and specificity of assessment myometrial invasion on DWI were 75%, 85% and 71% respectively. With the increase of b-value, DWI images could provide higher contrast, minimizing T2-weighted and perfusion effects [22, 23]. The sharp contrast between the hyperintense EC and the hypointense myometrium on DWI significantly facilitates assessment of myometrial invasion depth. Meanwhile, we found that the protocol of DWI with b = 1000 s/mm2 had higher performance than those of b = 800 s/mm2 for detecting deep myometrial invasion. In addition, the findings of this study were similar with the previous study [24]. They demonstrated that DWI had superior diagnostic efficiency in the evaluation of cervical stromal invasion of EC (the accuracy, sensitivity, and specificity respectively for reader 1/reader 2 were 95.2%, 91.6%, 91.7%/100%, and 95.8%/90.1%). In our results, the sensitivity was lower for the detection of cervical stromal invasion. We only evaluated axial DWI. It may be beneficial to increase sagittal DWI for improving the sensitivity [7]. Our results showed that DWI with b = 1000 s/mm2 had a high accuracy and specificity for the assessment of lymph node metastasis. The results of this study support previously published studies that DWI showed low sensitivity and high specificity for detecting lymph node metastases [25].
This study showed that the AUCs and accuracies of ADCmean values with b = 1000 s/mm2 were higher than those of b = 800 s/mm2 for EC diagnosis and staging. Since the ADC values are a quantitative analysis and not easily affected by T2 penetration effects, it has a higher accuracy than the visual assessment. Our findings supported the results of previous studies [], in which the AUC of the ADCmean for distinguishing malignant from benign lesions was 0.985 and the cutoff values of ADCmean was 1.007 × 10−3 mm2/s. Our results were also in this range. According to the study of Koh et al. [23], the estimation of ADC values became more accurate as b-values increases. Although high b-values can highlight the area of lesions, signal-to-noise ratios and an increased susceptibility artifact of the images also need to be considered [2].
Reyes et al. [] demonstrated the AUCs of ADCmean and ADCmin for assessment myometrial invasion were 0.90 and 0.88 respectively, which was similar to our results. They also found that ADCmean and ADCmin values of EC with deep myometrial invasion were lower than those with superficial or no myometrial invasion, which suggested ADC values might be a useful indicator to evaluate the local invasiveness of EC . Cao et al. [9] found that the ADCmin values were correlated with cervical stromal invasion, and the AUC was 0.83. Our results supported the findings and showed that using ADCmean and ADCmin values to detect cervical stromal invasion had a higher sensitivity than visual evaluation. Many studies found that ADCmean and ADCmin values could be used to estimate the lymph node metastasis [27, 27–29]. Quan et al. [29] demonstrated the AUCs of ADCmean and ADCmin to assess lymph node metastasis were 0.737 and 0.643 respectively. They believe that pretreatment ADC values may be a potential biomarker to predict prognosis factors of patients with EC. Our results also showed that the ADC values have the ability to assess lymph node metastasis, and can provide better diagnostic efficacy with b = 1000 s/mm2. As a consequence, the accuracy of staging may be improved by the addition of ADC values in clinical practice.
There are some limitations in our study. First, this was a single-center retrospective study. The sample size was not very large and only two b-values were studied. And this retrospective design is that each DWI acquisition was performed with a single, clinically dictated b-value; therefore, paired images of the same lesion at two different b-values from the same patient are not available. Next, we will include a larger sample with multiple centers and more b-values to further validate our findings. Second, we only studied the staging and identification of EC, the pathological subtype and grade of EC need to be further explored. Third, in order to further validate the efficiency of DWI in staging EC, we should enroll more patients with advanced EC. Finally, it is worth noting that we focused solely on DWI for the assessment of EC. While DWI has shown promising results in our study, it is important to recognize that combining DWI with other MRI sequences, such as T2-weighted imaging (T2WI), might provide more comprehensive diagnostic information. T2WI is known for its excellent soft tissue contrast and is often used in conjunction with DWI in clinical practice to improve the accuracy of lesion characterization and staging. Future studies should consider incorporating T2WI to better understand the complementary roles of these imaging modalities in the diagnosis and staging of endometrial carcinoma.
Conclusion
In conclusion, this study demonstrated that the b = 1000 s/mm² protocol delivered more consistent diagnostic performance—whether assessed visually or quantitatively—while yielding higher AUCs than the b = 800 s/mm² protocol. ADC values can be used as a noninvasive imaging biomarker to accurately identify benign and malignant endometrial lesions and as a reference to evaluate the staging of EC. Therefore, this study suggested that DWI with b = 1000 s/mm2 might be a better protocol for EC assessment than b = 800 s/mm2. Which would provide more accurate diagnostic basis for the clinical practice.
Acknowledgements
Not applicable.
Abbreviations
- DWI
Diffusion weighted imaging
- ADC
Apparent diffusion coefficient
- EC
Endometrial carcinoma
- ROC
Receiver operating characteristic
- AUC
The area under the curves
- ESUR
European Society of Urogenital Radiology
- MRI
Magnetic resonance imaging
- ROI
Regions ofinterest
Authors’ contributions
Wu KH and Wang QQ conceived of the study, and Deng YC, Bi Q, Wang Q, Wang J, Yang HY, Ding F and Li QH participated in its design and data analysis and statistics, Deng YC and Bi Q helped to draft the manuscript. All authors read and approved the final manuscript.
Funding
This study was supported by the Rui Ying Research Fund (YXJL-2024-0350-0309); the Basic Research on Application of Joint Special Funding of Science and Technology Department of Yunnan Province-Kunming Medical University (202301AY070001-084);Kunming University of Science and Technology & the First People’s Hospital of Yunnan Province Joint Special Project on Medical Research (KUST-KH2022027Y). The funding bodies played no role in the design of the study or in the collection, analysis, or interpretation of the data.
Data availability
All data generated or analysed during this study are included in this article. Further enquiries can be directed to the corresponding author.
Declarations
Ethics approval and consent to participate
This study was conducted in accordance with the Declaration of Helsinki and approved by the ethics committee of the First People’s Hospital of Yunnan Province.The need for written informed consent to participate was waived by the ethics committee of the hospital due to retrospective nature of the study.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Yuchen Deng and Qiu Bi contributed equally to this study.
Contributor Information
Qinqing Wang, Email: leyouran@126.com.
Kunhua Wu, Email: khcgz@sina.com.
References
- 1.Siegel RL, Miller KD, Jemal A. Cancer statistics. CA Cancer J Clin. 2020;70(1):7–30. 10.3322/caac.21590. [DOI] [PubMed] [Google Scholar]
- 2.Gu B, Shang X, Yan M, et al. Variations in incidence and mortality rates of endometrial cancer at the global, regional, and national levels, 1990–2019. Gynecol Oncol. 2021;161(2):573–80. 10.1016/j.ygyno.2021.01.036. [DOI] [PubMed] [Google Scholar]
- 3.Sorosky JI. Endometrial cancer. Obstet Gynecol. 2012;120(2 Pt 1):383–97. 10.1097/AOG.0b013e3182605bf1. [DOI] [PubMed] [Google Scholar]
- 4.Koc Z, Erbay G, Ulusan S, et al. Optimization of b value in diffusion-weighted MRI for characterization of benign and malignant gynecological lesions. J Magn Reson Imaging. 2012;35(3):650–9. 10.1002/jmri.22871. [DOI] [PubMed] [Google Scholar]
- 5.Berek JS, Matias-Guiu X, Creutzberg C, et al. FIGO staging of endometrial cancer: 2023. Int J Gynaecol Obstet. 2023;162(2):383–94. 10.1002/ijgo.14923. [DOI] [PubMed] [Google Scholar]
- 6.Colombo N, Creutzberg C, Amant F, et al. Esmo-esgo-estro consensus conference on endometrial cancer: diagnosis, treatment and follow-up. Ann Oncol. 2016;27(1):16–41. 10.1093/annonc/mdv484. [DOI] [PubMed] [Google Scholar]
- 7.Reinhold C, Ueno Y, Akin EA, et al. ACR appropriateness Criteria® pretreatment evaluation and Follow-Up of endometrial cancer. J Am Coll Radiol. 2020;17(11S):S472–86. 10.1016/j.jacr.2020.09.001. [DOI] [PubMed] [Google Scholar]
- 8.Li H, Zhang J, Liu B, et al. Histogram analysis of multiple mathematical diffusion-weighted imaging models for preoperative prediction of Ki-67 expression in hepatocellular carcinoma. Front Oncol. 2025;15:1531236. 10.3389/fonc.2025.1531236. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Cao K, Gao M, Sun YS, et al. Apparent diffusion coefficient of diffusion weighted MRI in endometrial carcinoma-relationship with local invasiveness. Eur J Radiol. 2012;81(8):1926–30. 10.1016/j.ejrad.2011.04.019. [DOI] [PubMed] [Google Scholar]
- 10.Li G, Lixia WPH, et al. The ability of multi b-value DWI with different mathematical models to predict lymphatic vessel invasion in endometrial cancer. J Kunming Med Univ. 2022;43(04):118–24. [Google Scholar]
- 11.Wang LJ, Tseng YJ, Wee NK, et al. Diffusion-weighted imaging versus dynamic contrast-enhanced imaging for pre-operative diagnosis of deep myometrial invasion in endometrial cancer: a meta-analysis. Clin Imaging. 2021;80:36–42. 10.1016/j.clinimag.2021.06.027. [DOI] [PubMed] [Google Scholar]
- 12.Maheshwari E, Nougaret S, Stein EB, et al. Update on MRI in evaluation and treatment of endometrial cancer. Radiographics. 2022;42(7):2112–30. 10.1148/rg.220070. [DOI] [PubMed] [Google Scholar]
- 13.Nougaret S, Horta M, Sala E, et al. Endometrial cancer MRI staging: updated guidelines of the European society of urogenital radiology. Eur Radiol. 2019;29(2):792–805. 10.1007/s00330-018-5515-y. [DOI] [PubMed] [Google Scholar]
- 14.Whittaker CS, Coady A, Culver L, et al. Diffusion-weighted MR imaging of female pelvic tumors: a pictorial review. Radiographics. 2009;29(3):759–78. 10.1148/rg.293085130. [DOI] [PubMed] [Google Scholar]
- 15.Inoue C, Fujii S, Kaneda S, et al. Apparent diffusion coefficient (ADC) measurement in endometrial carcinoma: effect of region of interest methods on ADC values. J Magn Reson Imaging. 2014;40(1):157–61. 10.1002/jmri.2437. [DOI] [PubMed] [Google Scholar]
- 16.Ma X, Qiang J, Zhang G, et al. Evaluation of the depth of myometrial invasion of endometrial carcinoma: comparison of orthogonal pelvis-axial contrast-enhanced and uterus-axial dynamic contrast-enhanced MRI protocols. Acad Radiol. 2022;29(8):e119–27. 10.1016/j.acra.2021.09.011a. [DOI] [PubMed] [Google Scholar]
- 17.Basaran I, Cengel F, Bayrak AH. Diffusion-weighted imaging in the benign-malignant differentiation of endometrial pathologies; effectiveness of visual evaluation. J Coll Physicians Surg Pak. 2023;33(1):73–8. 10.29271/jcpsp.2023.01.73. [DOI] [PubMed] [Google Scholar]
- 18.Scepanovic B, Andjelic N, Mladenovic-Segedi L, et al. Diagnostic value of the apparent diffusion coefficient in differentiating malignant from benign endometrial lesions. Front Oncol. 2023;13:1109495. 10.3389/fonc.2023.1109495. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Koc Z, Erbay G. Optimal b value in diffusion-weighted imaging for differentiation of abdominal lesions. J Magn Reson Imaging: JMRI. 2014;40(3):559–66. 10.1002/jmri.24403. [DOI] [PubMed] [Google Scholar]
- 20.Liu Y, Zhang Y, Cheng R, et al. Radiomics analysis of apparent diffusion coefficient in cervical cancer: a preliminary study on histological grade evaluation. J Magn Reson Imaging. 2019;49(1):280–90. 10.1002/jmri.26192. [DOI] [PubMed] [Google Scholar]
- 21.Chen L, Shen F, Li Z, et al. Diffusion-weighted imaging of rectal cancer on repeatability and cancer characterization: an effect of b-value distribution study. Cancer Imaging. 2018;18(1):43. 10.1186/s40644-018-0177-1. Published 2018 Nov 15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Rechichi G, Galimberti S, Signorelli M, et al. Myometrial invasion in endometrial cancer: diagnostic performance of diffusion-weighted MR imaging at 1.5-T. Eur Radiol. 2010;20(3):754–62. 10.1007/s00330-009-1597-x. [DOI] [PubMed] [Google Scholar]
- 23.Koh D-M, Collins DJ. Diffusion-weighted MRI in the body: applications and challenges in oncology. AJR. 2007;188(6):1622–35. 10.2214/AJR.06.1403. [DOI] [PubMed] [Google Scholar]
- 24.Lin G, Huang YT, Chao A, et al. Endometrial cancer with cervical stromal invasion: diagnostic accuracy of diffusion-weighted and dynamic contrast enhanced MR imaging at 3T. Eur Radiol. 2017;27(5):1867–76. 10.1007/s00330-016-4583-0. [DOI] [PubMed] [Google Scholar]
- 25.Hori M, Kim T, Onishi H, et al. Endometrial cancer: preoperative staging using three-dimensional T2-weighted turbo spin-echo and diffusion-weighted MR imaging at 3.0 T: a prospective comparative study. Eur Radiol. 2013;23(8):2296–305. 10.1007/s00330-013-2815-0. [DOI] [PubMed] [Google Scholar]
- 26.Reyes-Pérez JA, Villaseñor-Navarro Y, Jiménez de Los Santos ME, et al. The apparent diffusion coefficient (ADC) on 3-T MRI differentiates myometrial invasion depth and histological grade in patients with endometrial cancer. Acta Radiol. 2020;61(9):1277–86. 10.1177/0284185119898658. [DOI] [PubMed] [Google Scholar]
- 27.Koplay M, Dogan NU, Erdogan H, et al. Diagnostic efficacy of diffusion-weighted MRI for preoperative assessment of myometrial and cervical invasion and pelvic lymph node metastasis in endometrial carcinoma. J Med Imaging Radiat Oncol. 2014;58(5):538–648. 10.1111/1754-9485.12209. [DOI] [PubMed] [Google Scholar]
- 28.Shih IL, Yen RF, Chen CA, et al. Standardized uptake value and apparent diffusion coefficient of endometrial cancer evaluated with integrated whole-body PET/MR: correlation with pathological prognostic factors. J Magn Reson Imaging. 2015;42(6):1723–32. 10.1002/jmri.24932. [DOI] [PubMed] [Google Scholar]
- 29.Quan Q, Lu Y, Xuan B, et al. The prominent value of apparent diffusion coefficient in assessing high-risk factors and prognosis for patients with endometrial carcinoma before treatment. Acta Radiol. 2021;62(6):830–8. 10.1177/0284185120940271. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
All data generated or analysed during this study are included in this article. Further enquiries can be directed to the corresponding author.