European Journal of Radiology

Magnetic resonance imaging features of mesonephric-like adenocarcinoma of the uterine corpus in comparison with endometrial endometrioid carcinoma: Multi-institutional study

To examine magnetic resonance imaging (MRI) features of mesonephric-like adenocarcinoma (MLA) of the uterine corpus. MRI features of 19 patients with pathologically proven MLA of the uterine corpus were retrospectively compared with those of 95 patients with endometrial endometrioid carcinoma (EEC). Most patients with MLA were postmenopausal. Advanced FIGO stages were more common in the MLA than in the EEC group (63.2 % vs. 18.9 %, p < 0.001). On MRI, endophytic growth into the myometrium were more frequent in the MLA than in the EEC group (68.4 % vs. 14.7 %, p = 0.005). The median maximum tumor diameter in the MLA group (52.4 mm) tended to be larger than that in the EEC group (38.9 mm), although the difference was not statistically significant (p = 0.374). The tumor-to-muscle signal intensity ratio (SIR) on fat-suppressed gadolinium-enhanced T1-weighted gradient-echo imaging was higher in the MLA group than in the EEC group. (1.67 vs. 1.36, p = 0.002). The SIR on diffusion-weighted imaging (DWI) was comparable between the two groups (8.35 vs. 6.72, p = 0.330). The apparent diffusion coefficient value was lower in the MLA than in the EEC group (0.69 10 MRI of MLA of the uterine corpus typically demonstrates large, diffuse, and endophytic growth into the myometrium, with strong contrast enhancement and more restricted diffusion compared to EEC, with coexisting adenomyosis present in over 50% of patients. Therefore, postmenopausal women with adenomyosis should be carefully evaluated for MLA on MRI, particularly using DWI.

Multiple mathematical models of diffusion-weighted imaging for endometrial cancer characterization: Correlation with prognosis-related risk factors

To investigate mono-exponential, bi-exponential, and stretched-exponential models of diffusion-weighted imaging (DWI) for evaluation of prognosis-related risk factors of endometrial cancer (EC). Sixty-one consecutive patients with EC who preoperatively underwent pelvic MRI with multiple b value DWI between September 2016 and May 2018 were enrolled. The apparent-diffusion-coefficient (ADC), bi-exponential model parameters (D, D* and f) and stretched-exponential model parameters (DDC and α) were measured and compared to analyze the following prognosis-related risk factors confirmed by pathology: histological grade, depth of myometrial invasion, cervical stromal infiltration (CSI) and lymphovascular invasion (LVSI). A stepwise multilvariate logistic regression and the receiver operating characteristic (ROC) curves were performed for further statistical analysis. Lower ADC, D, f, and DDC were observed in tumor with high grade compared with a low-grade group, and the largest area under curve (AUC) was obtained when combining f and DDC values. ADC, D, f, DDC, and α were significantly different in patients with deep myometrial invasion (DMI) compared to those without DMI; the combination of f, DDC and α showed the highest AUC. Significantly different ADC and f were found between patients' presence and absence CSI; the f values showed the highest diagnostic performance with an AUC of 0.825. Regarding the LVSI, ADC, D*, f, and DDC were significantly lower in tumors with LVSI compared to those without LVSI; the combination of f and DDC showed the largest AUC. Multiple mathematical DWI models are a useful approach for the prediction of prognosis-related risk factors in EC.

The application of Machine learning in predicting the outcomes of minimally invasive treatments for uterine Fibroids: A systematic review and meta-analysis

Uterine fibroids (UFs) are common benign tumors that impact women's health, particularly through symptoms such as abnormal bleeding or reproductive dysfunction. Interventional radiology (IR) techniques like uterine artery embolization (UAE) and high-intensity focused ultrasound (HIFU) are minimally invasive alternatives to surgery. Machine learning (ML) has shown promise in predicting treatment outcomes, though the optimal model remains uncertain. This systematic review and meta-analysis evaluate models predicting outcomes of minimally invasive treatments for uterine fibroids. A comprehensive search was conducted across five databases (PubMed, Embase, Scopus, Web of Science, and Cochrane) through November 2024, following PRISMA guidelines and registered in PROSPERO. Studies using ML to predict different outcomes of UFs treatment via minimally invasive treatments were included. PROBAST + AI was used to assess study quality. Pooled sensitivity, specificity, and AUC values were calculated using a bivariate random effect model. Out of 1,114 records, fourteen studies met the inclusion criteria, with 12 focusing on HIFU and two on UAE. Logistic regression was the most commonly used approach, while gradient‑boosting models reported high discrimination in some individual studies; however, external validation was uncommon and risk of bias was frequently high. AUCs for radiomics-based models ranged from 0.668 to 0.887, and combined models ranged from 0.773 to 0.93. Meta-analysis of five HIFU-based radiomics studies demonstrate pooled sensitivity of 75% and specificity of 76% respectively, with an AUC of 0.82. ML models, particularly those integrating radiomics and clinical data, show strong performance in predicting image-guided treatment outcomes in UFs. These approaches support a promising path toward individualized treatment planning and may improve patient selection in clinical workflow.

[18F]-FAPI-04 PET/CT improves metastatic lesions detection for patients with uterine malignancy after radical surgery: prospectively compared with [18F]-FDG PET/CT

This prospective study aimed to investigate the potential of [ From August 2022 to August 2024, patients with suspected recurrent/metastatic uterine malignancy following radical surgery received both [ 36 patients with 482 local relapses and distant metastases were included. Local relapses demonstrated comparable diagnostic performance between [ [

MRI-based assessment of residual disease after neoadjuvant chemotherapy in pregnant women with cervical cancer

This study explores the performance of MRI in detecting residual disease after platinum-based neoadjuvant chemotherapy (NACT) in pregnant cervical cancer (CC) patients, which would serve as a surrogate of treatment response. In this retrospective single-centre study, consecutive pregnant cervical cancer patients treated with NACT and undergoing MRI examination before and at the end of the therapeutic protocol between 2010 and 2021 were included. Tumour maximum diameter and nodal status were evaluated in MRI at staging and after NACT. Patients exhibiting increased tumor burden post-NACT were excluded. On post-NACT the presence or absence of residual disease was recorded and the MRI diagnostic performance for assessing residual disease was calculated using histopathology at radical hysterectomy as the reference standard. The study included 12 pregnant patients (median age 36 years, 27-42). At post-NACT MRI, residual disease was absent in 2/12 patients (16.7 %) while present in 10/12 (83.3 %). Histopathology was concordant in all patients without MRI residual disease and in 9/10 patients with MRI residual disease, while discordant in 1/10. MRI sensitivity, specificity, positive predictive value, negative predictive and accuracy for detecting residual disease were 100.0 % (95 %CI: 1.00, 1.00), 66.7 % (95 %CI: 0.13, 1.00), 90 % (95 %CI: 0.71, 1.00), 100 % (95 %CI: 1.00, 1.00), and 91.7 % (95 %CI: 0.76, 1.00) respectively (p = 0.045). This study suggests that MRI has good diagnostic performance to detect residual disease after NACT in pregnant CC patients, and potentially assess response to treatment in this setting.

Effect of “T2-rim sign” related parameters on high-intensity focused ultrasound ablation of uterine fibroids

To investigate the effect of "high-signal-intensity peripheral rim on T2-weighted MR images (T2-rim sign)" related parameters on non-perfused volume ratio (NPVR) after high-intensity focused ultrasound (HIFU) ablation of uterine fibroids. Data from 616 patients with uterine fibroids treated with HIFU were retrospectively analyzed. Univariate and multivariate logistic regression was used to analyze the factors influencing the ablation effect. The effect of T2-rim sign on ablation parameters and results was also analyzed. Spearman correlation analysis was used to compare the correlation between coverage ratio, average thickness of T2-rim sign and NPVR in 207 cases of fibroids with T2-rim sign. The presence of T2-rim sign was an independent risk factor affecting the ablation effect. The coverage ratio of T2-rim sign was negatively correlated with treatment efficiency (r = -0.174, p = 0.012) and NPVR (r = -0.186, p = 0.007), and positively correlated with energy efficiency factor (EEF) (r = 0.156, p = 0.024). The average thickness of T2-rim sign was positively correlated with treatment intensity (r = 0.203, p = 0.003) and negatively correlated with NPVR (r = -0.363, p < 0.001). There was a negative correlation between the average thickness of the T2-rim sign and NPVR in isointense fibroids (r = -0.484, p < 0.001). The presence of T2-rim sign increases the difficulty of ablation and reduces the ablation effect. In clinical practice, the presence and related parameters of T2-rim sign should be fully considered when screening for HIFU indications and formulating treatment plans.

Development and validation of a deep learning-based method for automatic measurement of uterus, fibroid, and ablated volume in MRI after MR-HIFU treatment of uterine fibroids

The non-perfused volume divided by total fibroid load (NPV/TFL) is a predictive outcome parameter for MRI-guided high-intensity focused ultrasound (MR-HIFU) treatments of uterine fibroids, which is related to long-term symptom relief. In current clinical practice, the MR-HIFU outcome parameters are typically determined by visual inspection, so an automated computer-aided method could facilitate objective outcome quantification. The objective of this study was to develop and evaluate a deep learning-based segmentation algorithm for volume measurements of the uterus, uterine fibroids, and NPVs in MRI in order to automatically quantify the NPV/TFL. A segmentation pipeline was developed and evaluated using expert manual segmentations of MRI scans of 115 uterine fibroid patients, screened for and/or undergoing MR-HIFU treatment. The pipeline contained three separate neural networks, one per target structure. The first step in the pipeline was uterus segmentation from contrast-enhanced (CE)-T1w scans. This segmentation was subsequently used to remove non-uterus background tissue for NPV and fibroid segmentation. In the following step, NPVs were segmented from uterus-only CE-T1w scans. Finally, fibroids were segmented from uterus-only T2w scans. The segmentations were used to calculate the volume for each structure. Reliability and agreement between manual and automatic segmentations, volumes, and NPV/TFLs were assessed. For treatment scans, the Dice similarity coefficients (DSC) between the manually and automatically obtained segmentations were 0.90 (uterus), 0.84 (NPV) and 0.74 (fibroid). Intraclass correlation coefficients (ICC) were 1.00 [0.99, 1.00] (uterus), 0.99 [0.98, 1.00] (NPV) and 0.98 [0.95, 0.99] (fibroid) between manually and automatically derived volumes. For manually and automatically derived NPV/TFLs, the mean difference was 5% [-41%, 51%] (ICC: 0.66 [0.32, 0.85]). The algorithm presented in this study automatically calculates uterus volume, fibroid load, and NPVs, which could lead to more objective outcome quantification after MR-HIFU treatments of uterine fibroids in comparison to visual inspection. When robustness has been ascertained in a future study, this tool may eventually be employed in clinical practice to automatically measure the NPV/TFL after MR-HIFU procedures of uterine fibroids.

Improving risk stratification of indeterminate adnexal masses on MRI: What imaging features help predict malignancy in O-RADS MRI 4 lesions?

Ovarian-Adnexal Reporting and Data System (O-RADS) MRI uses a 5-point scale to establish malignancy risk in sonographically-indeterminate adnexal masses. The management of O-RADS MRI score 4 lesions is challenging, as the prevalence of malignancy is widely variable (5-90%). We assessed imaging features that may sub-stratify O-RADS MRI 4 lesions into malignant and benign subgroups. Retrospective single-institution study of women with O-RADS MRI score of 4 adnexal masses between April 2021-August 2022. Imaging findings were assessed independently by 2 radiologists according to the O-RADS lexicon white paper. MRI and clinical findingswere compared between malignant and benign adnexal masses, and inter-reader agreement was calculated. Seventy-four women (median age 52 years, IQR 36-61) were included. On pathology, 41 (55.4%) adnexal masses were malignant. Patients with malignant masses were younger (p = 0.02) with higher CA-125 levels (p = 0.03). Size of solid tissue was greater in malignant masses (p = 0.01-0.04). Papillary projections and larger solid portion were more common in malignant lesions; irregular septations and predominantly solid composition were more frequent in benign lesions (p < 0.01). Solid tissue of malignant lesions was more often hyperintense on T2-weighted and diffusion-weighted imaging (p ≤ 0.03). Other imaging findings were not significantly different (p = 0.09-0.77). Inter-reader agreement was excellent-good for most features (ICC = 0. 662-0.950; k = 0. 650-0.860). Various MRI and clinical features differed between malignant and benign O-RADS MRI score 4 adnexal masses. O-RADS MRI 4 lesions may be sub-stratified (high vs low risk) based on solid tissue characteristics and CA-125 levels.

Training radiology residents to evaluate deep myometrial invasion in endometrial cancer patients on MRI: A learning curve study

To evaluate the impact of a four-month training program on radiology residents' diagnostic accuracy in assessing deep myometrial invasion (DMI) in endometrial cancer (EC) using MRI. Three radiology residents with limited EC MRI experience participated in the training program, which included conventional didactic sessions, case-centric workshops, and interactive classes. Utilizing a training dataset of 120 EC MRI scans, trainees independently assessed subsets of cases over five reading sessions. Each subset consisted of 30 scans, the first and the last with the same cases, for a total of 150 reads. Diagnostic accuracy metrics, assessment time (rounded to the nearest minute), and confidence levels (using a 5-point Likert scale) were recorded. The learning curve was obtained plotting the diagnostic accuracy of the three trainees and the average over the subsets. Anatomopathological results served as the reference standard for DMI presence. The three trainees exhibited heterogeneous starting point, with a learning curve and a trend to more homogeneous performance with training. The diagnostic accuracy of the average trainee raised from 64 % (56 %-76 %) to 88 % (80 %-94 %) across the five subsets (p < 0.001). Reductions in assessment time (5.92 to 4.63 min, p < 0.018) and enhanced confidence levels (3.58 to 3.97, p = 0.12) were observed. Improvements in sensitivity, specificity, positive predictive value, and negative predictive value were noted, particularly for specificity which raised from 56 % (41 %-68 %) in the first to 86 % (74 %-94 %) in the fifth subset (p = 0.16). Although not reaching statistical significance, these advancements aligned the trainees with literature performance benchmarks. The structured training program significantly enhanced radiology residents' diagnostic accuracy in assessing DMI for EC on MRI, emphasizing the effectiveness of active case-based training in refining oncologic imaging skills within radiology residency curricula.

MRI characteristics for predicting histological subtypes in patients with uterine cervical adenocarcinoma

To evaluate the magnetic resonance imaging (MRI) findings of uterine cervical adenocarcinoma for predicting different histological subtypes. We retrospectively analyzed MRI findings of 76 consecutive patients with histopathologically-confirmed uterine cervical adenocarcinoma undergoing preoperative MRI examination. An experienced pathologist classified the histological subtypes based on World Health Organization's 2020 classification and into human papillomavirus (HPV)-associated adenocarcinomas (HPVAs, n = 54) (usual type and variants) and HPV-independent adenocarcinomas (HPVIs, n = 22) (gastric type adenocarcinoma (GAS), clear cell type, and other types). Different MRI variables were compared quantitatively and qualitatively between HPVA and HPVI and between GAS and non-GAS tumor types. The maximum tumor diameter was significantly greater in HPVIs than HPVAs (41.9 ± 18.6 vs 32.7 ± 15.6 mm; p < 0.05). Heterogeneous enhancement on fat-suppressed gadolinium-enhanced T1-weighted images was more frequently seen in HPVIs than HPVAs (62 % vs 15 %; p < 0.01) and in GASs than non-GASs (78 % vs 16 %; p < 0.01). Also, infiltrative growth pattern (58 % vs 20 %; p < 0.05) and intratumoral cyst formation (83 % vs 47 %) (p < 0.05) were more frequent in GASs than non-GASs. Compared with HPVAs, HPVIs tend to have a larger tumor size with heterogeneous enhancement, of which GASs frequently show infiltrative growth patterns with intratumoral cyst formation and heterogeneous enhancement.

Estimating pathological prognostic factors in epithelial ovarian cancers using apparent diffusion coefficients of functional tumor volume

To assess the utility of apparent diffusion coefficients (ADCs) of whole tumor volume (WTV) and functional tumor volume (FTV) in determining the pathologicalprognostic factors in epithelial ovarian cancers (EOCs). A total of 155 consecutive patients who were diagnosed with EOC between January 2017 and August 2022 and underwent both conventional magnetic resonance imaging and diffusion-weighted imaging were assessed in this study. The maximum, minimum, and mean ADC values of the whole tumor (ADCwmax, ADCwmin, and ADCwmean, respectively) and functional tumor (ADCfmax, ADCfmin, and ADCfmean, respectively) as well as the WTV and FTV were derived from the ADC maps. The univariate and multivariate logistic regression analyses and receiver operating characteristic curve (ROC) analysis were used to assess the correlation between these ADC values and the pathological prognostic factors, namely subtypes, lymph node metastasis (LNM), Ki-67 index, and p53 expression. The ADCfmean value was significantly lower in type II EOC, LNM-positive, and high-Ki-67 index groups compared to the type I EOC, LNM-negative, and low-Ki-67 index groups (p ≤ 0.001). Similarly, the ADCwmean and ADCfmean values were lower in the mutant-p53 group compared to the wild-type-p53 group (p ≤ 0.001). Additionally, the ADCfmean showed the highest area under the ROC curve (AUC) for evaluating type II EOC (0.725), LNM-positive (0.782), and high-Ki-67 index (0.688) samples among the given ROC curves, while both ADCwmean and ADCfmean showed high AUCs for assessing p53 expression (0.694 and 0.678, respectively). The FTV-derived ADC values, especially ADCfmean, can be used to assess preoperative prognostic factors in EOCs.

Elevated 18F-FDG uptake in non-metastatic lymph nodes of POLE-mutated endometrial cancer on PET/CT

DNA polymerase epsilon (POLE) exonuclease domain-mutated endometrial cancer (EC) is associated with a favorable prognosis and significant immune cell infiltration. Cancers with significant immune cell infiltration often exhibit higher maximum standardized uptake value (SUVmax) on This retrospective study enrolled 102 patients (mean age, 56.2 ± 11.8 years) with EC who underwent preoperative FDG-PET/CT between July 2018 and August 2023. Patients were categorized into molecular subtypes based on molecular profiling. SUVmax of the primary tumor and lymph nodes was measured, and correlations with molecular subtypes were examined. The SUVmax in the primary tumor of POLE-mutated ECs was higher than that in non-POLE-mutated ECs (mean ± SD, 13.21 ± 4.78 vs 9.82 ± 3.78; p = 0.001). Multiple regression analysis showed that POLE status was associated with SUVmax in primary tumors (B = 2.30, p = 0.031). The SUVmax in non-metastatic lymph nodes of POLE-mutated ECs was higher than that in non-POLE-mutated ECs (median SUVmax, 1.31 [IQR, 1.05-2.21] vs. 1.06 [IQR, 0.91-1.20]; p = 0.003). Multiple regression analysis revealed that POLE status was the only factor associated with SUVmax in non-metastatic lymph nodes (B = 0.339, p = 0.046), while age, BMI, blood glucose level, histological type, and primary tumor size showed no significant association. Patients with POLE-mutated EC exhibit high SUVmax in primary tumors and also tend to show elevated uptake in non-metastatic lymph nodes.

The value of multiparametric MRI combined with clinical prognostic parameters in predicting the 5-year survival of stage IIIC1 cervical squamous cell carcinoma

To explore the value of multiparametric magnetic resonance imaging（MRI）in predicting the 5-year progression-free survival (PFS) and overall survival (OS) of cervical squamous cell carcinoma (CSCC) in 2018 FIGO stage IIIC1. This retrospective study collected156 patients with CSCC from Dec. 2014 to Jul. 2018. Sixty-one patients underwent radical hysterectomy (RH), and 95 patients underwent concurrent chemoradiotherapy (CCRT). Clinical and MR parameters of primary tumours were analysed. A 1:1 ratio propensity score matching (PSM) was performed for the RH group and CCRT group according to T stage. The Cox proportional hazard model was used to evaluate the associations between imaging or clinical variables and PFS and OS. The 5-year PFS and OS rates were 72.6% and 78.3%, respectively. The analysis results show that the treatment method, ADC Functional MRI parameters and SCC-Ag have potential predictive value for the 5-year survival of 2018 FIGOIIIC1 CSCC. There were no significant differences in survival between CCRT and RH + adjuvant therapy for IIIC1 stage CSCC if the T stage was earlier.

Clues to the diagnosis of borderline ovarian tumours: An imaging guide

Borderline Ovarian Tumours (BOTs) are an interesting subset of epithelial neoplasms defined histologically by atypical epithelial proliferation without stromal invasion. These tumours typically affect young women in the reproductive age group and have a good prognosis. Although ultrasonography is the primary screening imaging technique in the evaluation of any suspected adnexal mass, grey-scale and colour Doppler have limited value in characterizing BOTs. Thus, a pelvic magnetic resonance imaging (MRI) is recommended for further characterization on account of its multiplanar capabilities, excellent soft-tissue contrast and high spatial resolution. BOTs histological subtypes display specific features on MRI that are useful in differential diagnosis. However, the final diagnosis and staging of BOTs require pathologic evaluation after surgical excision. Therefore, the purpose of this review is to describe, illustrate and compare the imaging characteristics of the different subtypes of BOTs - serous, mucinous and seromucinous - focusing on MRI, as well as to correlate with pathology findings considering the recent 2020 World Health Organization (WHO) classification, in order to improve the accuracy of preoperative diagnosis and facilitate optimal patient management.

Role of MRI in characterizing serous borderline ovarian tumor and its subtypes: Correlation of MRI features with clinicopathological characteristics

This study aimed to investigate the diagnostic value of MRI in serous borderline ovarian tumor (SBOT), and to determine the MRI features of SBOT and their correlations with clinicopathological characteristics. A total of 121 patients suspected of SBOT by preoperative MRI and then underwent surgery at our hospital were retrospectively reviewed. The accuracy of MRI in diagnosing SBOT was assessed. MRI features of the SBOT subtypes were compared and their correlations with clinicopathological characteristics were evaluated. SBOT was confirmed by postoperative pathology in 95 patients, including 77 patients with conventional SBOT (SBOT-C) and 18 patients with micropapillary SBOT (SBOT-MP). The accuracy of MRI in diagnosing SBOT was 87.6%. Three MRI morphological patterns of SBOT were identified: (i) mainly solid, (ii) mainly cystic, and (iii) mixed. Branching papillary architecture and internal branching (PA&IB) structures corresponding to multiple branching papillary projections and internal fibrous stalks in tumors were observed in 69.7% of SBOTs on T2-weighted images. MRI findings were consistent with postoperative pathology. Compared with SBOT-C, patients with SBOT-MP were more likely to display elevated cancer antigen 125, bilateral tumors, peritoneal implantation, lymph node metastasis, and advanced tumor staging. No significant differences were observed in MRI features between SBOT-C and SBOT-MP groups. MRI has good performance in diagnosing SBOT. MRI findings of SBOT are consistent with clinicopathological characteristics. The PA&IB structure is the characteristic MRI finding of SBOT. Compared to SBOT-C, SBOT-MP tends to display more aggressive clinical behavior, but their MRI features are similar.

A CT-based radiomics nomogram for predicting early recurrence in patients with high-grade serous ovarian cancer

To develop and validate a radiomics nomogram for predicting early recurrence in high-grade serous ovarian cancer (HGSOC) patients. From May 2008 to December 2019, 256 eligible HGSOC patients were enrolled and divided into training (n = 179) and test cohorts (n = 77) in a 7:3 ratio. A radiomics signature (Radscore) was selected by using recursive feature elimination based on a support vector machine (SVM-RFE) and building a radiomics model for recurrence prediction. Independent clinical risk factors were generated by univariable and multivariable Cox regression analyses. A combined model was developed based on the Radscore and independent clinical risk factors and presented as a radiomics nomogram. Its performance was assessed by AUC, Kaplan-Meier survival analysis and decision curve analysis. Seven radiomics features were selected. The radiomics model yielded AUCs of 0.715 (95% CI: 0.640, 0.790) and 0.717 (95% CI: 0.600, 0.834) in the training and test cohorts, respectively. The clinical model (FIGO stage and residual disease) yielded AUCs of 0.632 and 0.691 in the training and test cohorts, respectively. The combined model demonstrated AUCs of 0.749 (95% CI: 0.678, 0.821) and 0.769 (95% CI: 0.662, 0.877) in the training and test cohorts, respectively. In the combined model, PFS was significantly shorter in the high-risk group than in the low-risk group (P < 0.0001). The radiomics nomogram performed well for early individualized recurrence prediction in patients with HGSOC and can also be used to differentiate high-risk patients from low-risk patients.

Diagnosing uterine cervical cancer on a single T2-weighted image: Comparison between deep learning versus radiologists

To compare deep learning with radiologists when diagnosing uterine cervical cancer on a single T2-weighted image. This study included 418 patients (age range, 21-91 years; mean, 50.2 years) who underwent magnetic resonance imaging (MRI) between June 2013 and May 2020. We included 177 patients with pathologically confirmed cervical cancer and 241 non-cancer patients. Sagittal T2-weighted images were used for analysis. A deep learning model using convolutional neural networks (DCNN), called Xception architecture, was trained with 50 epochs using 488 images from 117 cancer patients and 509 images from 181 non-cancer patients. It was tested with 60 images for 60 cancer and 60 non-cancer patients. Three blinded experienced radiologists also interpreted these 120 images independently. Sensitivity, specificity, accuracy, and area under the receiver operating characteristic curve (AUC) were compared between the DCNN model and radiologists. The DCNN model and the radiologists had a sensitivity of 0.883 and 0.783-0.867, a specificity of 0.933 and 0.917-0.950, and an accuracy of 0.908 and 0.867-0.892, respectively. The DCNN model had an equal to, or better, diagnostic performance than the radiologists (AUC = 0.932, and p for accuracy = 0.272-0.62). Deep learning provided diagnostic performance equivalent to experienced radiologists when diagnosing cervical cancer on a single T2-weighted image.

Feasibility of MRI-based radiomics features for predicting lymph node metastases and VEGF expression in cervical cancer

To investigate the predictive value of MRI-based radiomics features for lymph node metastasis (LNM) and vascular endothelial growth factor (VEGF) expression in patients with cervical cancer. A total of 163 patients with cervical cancer were enrolled in this study. A total of 134 patients were included for LNM differentiation, and 118 were included for VEGF expression discrimination. The patients were randomly assigned to the training group or test group at a ratio of 2:1. Radiomics features were extracted from T1WI enhanced and T2WI MRI scans of each patient, and tumor stage was also documented according to the International Federation of Gynecology and Obstetrics (FIGO) guidelines. The least absolute shrinkage and selection operator algorithm was used for feature selection. The results of 5-fold cross validation were applied to select the best classification models. The performances of the constructed models were further evaluated with the test group. Sixteen radiomics features and the FIGO stage were selected to construct the LNM discrimination model. The LNM prediction model achieved the best diagnostic performance, with areas under the receiver operating curve (AUCs) of 0.95 and 0.88 in the training group and test group, respectively. Nine radiomics characteristics were screened to build the VEGF prediction model, with AUCs of 0.82 and 0.70 in the training group and test group, respectively. Decision curve analysis confirmed their clinical usefulness. The presented radiomics prediction models demonstrated potential to noninvasively differentiate LNM and VEGF expression in cervical cancer.

Combination of clinical and MRI features in diagnosing ovarian granulosa cell tumor: A comparison with other ovarian sex cord-gonadal stromal tumors

To evaluate the combination of magnetic resonance imaging (MRI) findings and clinical features in diagnosing ovarian granulosa cell tumor (OGCT) and comparing OGCTs with other ovarian sex cord-gonadal stromal tumors (OSGTs). Women who underwent MRI and were surgically confirmed with OSGTs between January 2015 and January 2022 were included in the study. Histology was used as a primary method of diagnosis. T1WI, T2WI, and DWI MR scans were performed for all patients. All MR images were reviewed by two radiologists. The clinic baseline characteristics of all patients were recorded. A total of 58 patients were enrolled, with 21 OGCTs found in 20 patients and 39 other OSGTs found in 38 patients. In terms of clinical, the proportion of vaginal discharge/bleeding and menstrual abnormalities were significantly higher in OGCTs than in the control group. A multivariate analysis of the combined clinical MRI revealed that symptomatic, T2 signals of the solid component, Honeycomb-sign, Swiss cheese-sign, and ADC values were independent features for discriminating between OGCTs and other OSGTs. Clinical features, MRI features, and a combined model were established; the areas under the curve of the three models in predicting OGCTs and other OSGTs were 0.694, 0.852, and 0.927, respectively. The DeLong test showed that the combined model had the highest efficiency in predicting OGCTs (p < 0.05), which was significantly different from the AUC of the other two models (p < 0.05). Combining clinic and MRI findings helps differentiate OGCTs from other OSGTs. These results help optimize clinical management and indicate that radiologists should focus on clinical information to help improve diagnostic accuracy.

Leiomyoma or sarcoma? MRI performance in the differential diagnosis of sonographically suspicious uterine masses

To assess the diagnostic performance of MRI in distinguishing between leiomyomas and malignant/potentially malignant mesenchymal neoplasms in patients with rapidly enlarging/sonographically suspicious uterine masses. IRB-approved retrospective study including 88 patients (51 ± 11 years) who underwent MRI for rapidly enlarging/sonographically suspicious uterine mass at our Institution between January 2016 and December 2021, followed by surgery or >12 months follow-up. Qualitative image analysis was independently performed by 2 radiologists and included lesion's margins (sharp/irregular), architecture (homogeneous/inhomogeneous), presence of endometrial infiltration (yes/no), necrotic areas (yes/no), hemorrhagic areas (yes/no), predominant signal intensity on T1-WI, T2-WI, CE T1-WI, DWI, and ADC map. The same radiologists performed quantitative image analysis in consensus, which included lesion's maximum diameter, lesion/myometrium signal intensity ratio on T2-WI and CE T1-weighted images, lesion/endometrium signal intensity ratio on DWI and ADC map and necrosis percentage. Lesions were classified as benign or malignant. Imaging findings were compared with pathology and/or follow-up. After surgery (52/88 patients) or follow-up (36/88 patients, 33 ± 20 months), 83/88 (94.3%) lesions were classified as benign and 5/88 (5.7%) as malignant/potentially malignant. Presence of necrotic areas, high necrosis percentage, hyperintensity on DWI and high lesion/endometrium DWI signal intensity ratio were significantly associated with malignant/potentially malignant lesions (p = 0.027, 0.002, 0.008 and 0.015, respectively). The two readers identified malignant/potentially malignant lesions with 95.5% accuracy, 80.0% sensitivity, 96.4% specificity, 57.1 % PPV, 93.3% NPV. MRI has high accuracy in identifying malignant/potentially malignant myometrial masses. In everyday practice, however, MRI positive predictive value is relatively low given the low pre-test malignancy probability.

Timing of MRI for early treatment response prediction of chemoradiotherapy in uterine cervical cancer

To explore the optimal use of MRI including time point to predict early treatment response during definitive chemoradiotherapy in cervical carcinoma. Pilot study including 15 patients with cervical carcinoma stage IIB-IIIB (FIGO 2009) scheduled for chemoradiotherapy. All patients underwent four MRI examinations (at baseline, 3 weeks, 5 weeks, and 12 weeks after treatment start). Maximum tumor size, size change (Δsize), visibility on diffusion-weighted imaging (DWI), apparent diffusion coefficient (ADC) mean and change in ADC (ΔADC) at the different time points were recorded. 7/15 patients relapsed during the study period, referred to as "poor prognosis" (PP) and the remaining 8/15 are referred to as "good prognosis" (GP). At all four time points, the tumors in the PP patients were larger than in the GP patients. Tumor size did change between the time point but this was not significant between the groups. Visible tumor on high b-value DWI reached a sensitivity and specificity for predicting prognostic group being highest at 5 weeks, 83 % (5/6) and 63 % (5/8), respectively. The combination of tumor size at baseline and visible tumor on DWI at 5 weeks, resulted in an area under the curve (AUC) = 0.83. At 12 weeks, no GP patients, but 2/7 PP patients showed visible tumor on DWI. Addition of ADC-values in the analysis did not improve the predictive value at any time point. This pilot study indicates that the tumor size on baseline MRI, combined with tumor visibility on DWI at 5 weeks, may predict outcome of chemoradiotherapy.

Clinical efficacy of the novel heat-balance technique in ultrasound-guided percutaneous microwave ablation for uterine fibroids: A retrospective study

To evaluate the clinical efficacy of the novel Heat-Balance Technique (HB type) in Ultrasound-Guided Percutaneous Microwave Ablation for Uterine Fibroids (PMWA). According to the inclusion and exclusion criteria, 40 patients who underwent PMWA between June 2022 and June 2024 were selected. These patients were divided into two groups based on the specific surgical technique: one using the conventional technique and the other using a novel PMWA (HB-type). Relevant clinical data of the patients were retrospectively analyzed, and all patients were followed up postoperatively to assess the clinical efficacy of the two groups. All cases achieved effective ablation. Among them, 15 cases (93.75 %) in the HB-type group and 14 cases (87.5 %) in the conventional ablation group achieved marked efficacy, with no significant statistical difference between the two groups. Both modalities significantly improved patients' clinical symptoms (p < 0.001). The HB-type group showed significantly shorter surgical duration (8 [6-12] min vs. 25 [20-30] min; p < 0.001) and fewer antenna placements (1 [1,2] vs. 5 [3,5]; p < 0.001) than the conventional group. The HB-type technique in PMWA achieves comparable ablation effects and clinical efficacy to conventional ablation methods for uterine fibroids, while offering distinct advantages: shortened surgical duration, reduced number of antenna placements, and lowered the complication rates.

Diagnostic performance of ADC values and MRI-based radiomics analysis for detecting lymph node metastasis in patients with cervical cancer: A systematic review and meta-analysis

To evaluate and compare the diagnostic performance of apparent diffusion coefficient (ADC) values and MRI-based radiomics analysis for lymph node metastasis (LNM) detection in patients with cervical cancer (CC). We searched relevant databases for studies on ADC values and MRI-based radiomics analysis for LNM detection in CC between January 2001 and December 2021. Methodological quality assessment of risk of bias using Quality Assessment of Diagnostic Accuracy Studies 2 and radiomics quality score (RQS) of the studies was conducted. The pooled sensitivity, specificity, positive likelihood ratio (LR+), negative likelihood ratio (LR-), diagnostic odds ratio (DOR), and area under the curve (AUC) were calculated. Diagnostic performance was compared between the two quantitative analyses using a two-sample Z-test. In total, 22 studies including 2314 patients were included. Unclear risk of bias was observed in 4.5-36.4% of the studies. The 8 radiomics studies exhibited a median (interquartile range) RQS of 13.5 (5.5-15.75). The pooled sensitivity, specificity, LR+, LR-, DOR, and AUC of the ADC values vs radiomics analysis were 0.86 vs 0.84, 0.85 vs 0.73, 5.7 vs 3.1, 0.17 vs 0.22, 34 vs 14, and 0.91 vs 0.86, respectively. There was no threshold effect or publication bias, but significant heterogeneity existed among the studies. No significant difference was detected in the diagnostic performance of the two quantitative analyses using the Z-test. ADC values are more clinically promising because they are more easily accessible and widely applied, and exhibit a non-statistically significant trend to outperform radiomics analysis.

ADC and kinetic parameter of primary tumor: Surrogate imaging markers for fertility-sparing vaginal radical trachelectomy in patients with stage IB cervical cancer

To investigate the role of ADC and kinetic parameters derived from DCE-MRI in selecting eligible candidates for fertility-sparing vaginal radical trachelectomy (VRT). Female patients with FIGO stage IB cervical cancers between March 2019 and January 2022 were retrospectively included. All patients underwent hysterectomy and bilateral lymphadenectomy. According to the surgical pathology, the study population was divided into VRT-eligible group and VRT-ineligible group. ADC, semi-quantitative and quantitative kinetic parameters of the primary tumor were compared between the two groups. Logistic regression analysis was used to determine the independent predictors for VRT eligibility and ROC curve was used to evaluate the predictive performance. 19 patients were deemed eligible for VRT and 50 were ineligible. Compared with VRT-eligible group, time to peak and ADC were significantly lower in VRT-ineligible group (P = 0.004 and 0.001, respectively) while volume fraction of plasma (V ADC and the kinetic DCE-MRI parameter V

Added-value of texture analysis of ADC in predicting the survival of patients with 2018 FIGO stage IIICr cervical cancer treated by concurrent chemoradiotherapy

To investigate the value of texture analysis of ADC in predicting the survival of patients with 2018 International Federation of Gynecology and Obstetrics (FIGO) stage IIICr cervical squamous cell cancer (CSCC) treated with concurrent chemoradiotherapy (CCRT). A total of 91 patients with stage IIICr CSCC treated by CCRT between January 2014 and December 2018 were retrospectivelyenrolled in this study. Clinical variables and 21 first-order texture features extracted from ADC maps were collected. Univariate and multivariate Cox hazard regression analyses were performed to evaluate these parameters in predicting progression-free survival (PFS) and overall survival (OS). The independent variables were combined to build a prediction model and compared with the 2018 FIGO staging system. Survival curves were generated using the Kaplan-Meier method, and the log-rank test was used for comparison. Mean Absolute Deviation (MAD), T stage, and the number of lymph node metastasis (LNM) were independently associated with PFS, while MAD, energy, T stage, number of LNM, and tumor grade were independently associated with OS. The C-index values of the combined models for PFS and OS, which were respectively 0.750 and 0.832, were significantly higher compared to 2018 FIGO staging system values of 0.629 and 0.630, respectively (P < 0.05). The texture analysis of the ADC maps could be used along with clinical prognostic biomarkers to predict PFS and OS in patients with stage IIICr CSCC treated by CCRT.

Assessment of lymph node metastases in patients with ovarian high-grade serous carcinoma: Incremental diagnostic value of dual-energy CT combined with morphologic parameters

To explore the feasibility of Dual-Energy Computed Tomography (DECT) in distinguishing metastatic from non-metastatic lymph nodes (LNs) in ovarian High-Grade Serous Carcinoma (HGSC), and to assess the incremental diagnostic value of combining DECT with morphologic parameters in differentiating metastatic and non-metastatic LNs. From October 2021 to May 2024, 141 LNs from 39 patients with HGSC who underwent DECT were retrospectively enrolled. LNs were matched with the pathological report. Five morphologic parameters and nine DECT parameters were assessed. DECT parameters were obtained from both the arterial and venous phases, including the attenuation at 40 and 70 keV, slope of the spectral Hounsfield unit curve (λ 86 metastatic LNs and 55 non-metastatic LNs were finally enrolled in our study. The short diameter (S), long diameter (L), and S/L ratio were significantly larger in metastatic LNs compared to non-metastatic LNs (9.69 ± 4.06 vs. 6.37 ± 1.24 mm, P < 0.001; 13.99 ± 5.36 vs.9.61 ± 2.30 mm, P < 0.001; 0.70 ± 0.15 vs. 0.67 ± 0.12, P = 0.023). In the venous phase, λ DECT parameters provide incremental diagnostic value in assessing metastatic LNs in patients with HGSC. The combination of the morphology and DECT models significantly improves diagnostic performance compared to the standalone morphology model.

Mr-based radiomics analysis of intra-tumor heterogeneity for early treatment response prediction in locally advanced cervical cancer treated with concurrent chemoradiotherapy

Accurately predicting the sensitivity to concurrent chemoradiotherapy (CCRT) is crucial for patients with locally advanced cervical cancer (LACC). The objective of this study was to develop an imaging intra-tumoral heterogeneity (IITH) model based on IITH features to predict the early treatment response to CCRT in patients with LACC. The data was collected retrospectively from three medical centers, and a total of 211 patients with LACC who underwent CCRT were included in the study. Radiomics feature characterization was performed on pre-treatment contrast-enhanced T1-weighted imaging (CE-T1WI) at the voxel level to generate IITH feature vectors. These feature vectors were then clustered into one class using the K-means method, resulting in the formation of multiple tumor subregions. Radiomics features were extracted from multiple tumor subregions, filtered, and the best features were selected to build the IITH model. Center A (n = 160) was designated as the training cohort, while Center B (n = 36) and CC-Tumor-Heterogeneity (CCTH) (n = 15) datasets were utilized as external test cohorts. At the patient cohort level, three distinct radiomics feature vectors of information were incorporated at each voxel. When the tumor was segmented into seven sub-regions, clustering was optimal. Each subregional radiomic feature was extracted, and after filtering, the IITH prediction model was constructed. The AUC of the IITH model was 0.858 for the training cohort and 0.829 for the external test cohort. An IITH model was developed from tumor subregions, which demonstrated promising performance in predicting early treatment response to CCRT in LACC.

MRI characteristics of FIGO stage IA epithelial ovarian cancer (EOC)

To investigate the MRI characteristics of FIGO stage IA EOC with different pathologic subtypes in order to improve the radiologists' understanding of these diseases. In this retrospective study, we recruited patients who underwent surgery due to EOC at our hospital and were staged as FIGO IA from January 2013 to May 2024. The MR imaging and clinical features were evaluated by radiologists specialized in gynecology. Kruskal-Wallis test and Chi-squared test were performed to assess the difference between groups. A total of 34 patients with a mean age of 56 years included serous carcinoma 9 cases (26 %), endometrioid carcinoma 9 cases (26 %), clear cell carcinoma 10 cases (29 %), mucinous carcinoma 6 cases (18 %). 2 patients synchronously developed ovarian cancer and uterus endometrial cancer. The median CA125 level was 65.6 U/ml (95 % CI, 21.6 to 92.8) and laterality ratio was 17:18 (left: right). Median diameter of tumor was 10.9 cm (95 % CI, 8.0 to 11.8). There were 6 cases of pure solid tumor, 5 cases of unilocular cystic-solid tumor and 23 multilocular cystic-solid tumor. 17 cases appeared hemorrhage signal and 6 of which had mixed signals in the loculi. As for enhancement pattern, total 10 cases of type I, 11 cases of type II and 13 cases of type III. Solid tissue in 33 (97 %) cases demonstrated restricted diffusion in DWI sequence. 9 cases presented ascites. There were 5 factors showed significant differences among different pathologic subtypes (p < 0.05), including maximum diameter, general morphology, growth pattern of solid tissue, grouped septa and presence of mixed signals in the loculi. The imaging performance of FIGO stage IA EOC were variable, and there were differences in imaging characteristics among different pathological subtypes, which can improve current understanding of FIGO stage IA EOC and reduce clinical diagnostic omissions.

Preoperative unveiling of mucinous ovarian cancer subtypes: Clinical and CT features

To compare the frequency, clinical, imaging, and survival characteristics of mucinous ovarian carcinoma (MOC) growth pattern subtypes. According to ESGO 2024 guidelines, fertility-sparing surgery can be considered for stages IA and IC1 expansile MOC due to their excellent prognoses and low recurrence risk. The ability to diagnose MOC subtypes preoperatively would aid treatment planning. However, variables for preoperatively distinguishing MOC subtypes have not been established. This IRB-approved, retrospective, single-center study included consecutive patients (≥18 years) with pre-treatment CT from 01/2000-12/2020 and pathological diagnosis of MOC. One experienced radiologist interpreted all CTs, reviewing challenging cases with a senior radiologist. Clinical (age, CA-125, FIGO stage), outcome (local recurrence, distant metastasis, death from cancer/unrelated cause), and CT (laterality, size, mass type, calcifications, adenopathy, peritoneal/serosal implants, ascites) were compared (p  0.10 for all differences but age, potentially due to the small sample. Although CT features of MOC subtypes were similar, expansile MOC presented in younger patients with lower CA-125 levels, was more often diagnosed at stage I, and was uniformly unilateral.

Time-dependent diffusion MRI-based microstructural mapping for differentiating high-grade serous ovarian cancer from serous borderline ovarian tumor

To investigate the value of microstructural characteristics derived from time-dependent diffusion MRI in distinguishing high-grade serous ovarian cancer (HGSOC) from serous borderline ovarian tumor (SBOT) and the associations of immunohistochemical markers with microstructural features. Totally 34 HGSOC and 12 SBOT cases who received preoperative pelvic MRI were retrospectively included in this study. Two radiologists delineated the tumors to obtain the regions of interest (ROIs). Time-dependent diffusion MRI signals were fitted by the IMPULSED (imaging microstructural parameters using limited spectrally edited diffusion) model, to extract microstructural parameters, including fraction of the intracellular component (f In this study, f Time-dependent diffusion MRI had value in evaluating the microstructures of HGSOC and SBOT and could discriminate between these tumors.

Combined dynamic contrast enhanced MRI parameter with clinical factors predict the survival of concurrent chemo-radiotherapy in patients with 2018 FIGO IIICr stage cervical cancer

Combined clinical prognostic factors and magnetic resonance imaging (MRI) parameters on predicting the prognosis after concurrent chemo-radiotherapy (CCRT)in patients with 2018 International Federation of Gynecology and Obstetrics (FIGO) IIICr stage patients. A total of 117 patients with cervical cancer (2018 FIGO stage IIICr) who underwent CCRT were enrolled from Dec.2014 to Jul.2017. 47 patients developed outcome events, including 32 recurrences and 15 deaths. Clinical and MR parameters of primary tumors were analyzed, including apparent diffusion coefficient (ADC) values (ADC The median PFS was 35 months (6-68 month). The 1-year and 3-year PFS rates were was 90.4 %, 74.4 %, respectively. Multivariate analysis showed that 2018 FIGOIIIC2r stage (HR 2.701,95 %CI1.259to. 5.797; p = 0.011), K In conclusion, we have identified IIIC2r stage, K

The role of magnetic resonance imaging in the rare pathologies of the vulva

Pathologies of the vulva encompass a wide range of mesenchymal and epithelial benign and malignant lesions. Suspicion is raised by non-specific symptoms or clinical findings detected during routine gynecological examinations, and histopathology is essential for the diagnosis. The role of imaging has often been limited, but it can be essential in guiding treatment and, in some cases, in helping differential diagnosis. In particular, magnetic resonance imaging (MRI) can play a central role in identifying the extent of disease and planning surgical treatment. To this aim, rigorous image acquisition, correct disease evaluation in the context of vulvar anatomy and understanding of the possible differential diagnosis are essential. The aim of this article is to review the role of MRI in the evaluation of rare vulvar pathologies, focusing on different sites of origin, imaging characteristics, and local extent.

Comparison of reduced field-of-view diffusion-weighted imaging (DWI) and conventional DWI techniques in the assessment of Cervical carcinoma at 3.0T: Image quality and FIGO staging

To evaluate imaging quality (IQ) and International Federation of Gynecology and Obstetrics (FIGO) staging of reduced field-of-view (r-FOV) diffusion-weighted imaging (DWI) in cervical carcinoma (CC). Sixty patients with pathologically proven CC who underwent both pre-treatment r-FOV DWI and full field-of-view (f-FOV) DWI on a 3.0T MRI scanner were retrospectively reviewed. The subjective qualitative image scores were compared using the Wilcoxon signed-rank test. Objective quality values and apparent diffusion coefficient (ADC) were estimated by paired t-test or Wilcoxon signed-rank test for the two DWI sequences according to Normality test. Spearman rank correlation analysis was used to evaluate the relationship between pathological results and mean ADC value. The subjective IQ scores for r-FOV DWI were significantly higher than those for f-FOV DWI (P < 0.001). Similarly, the contrast-to-noise (CNR) value of r-FOV DWI was superior to that of f-FOV DWI (10.30 ± 3.676, 8.91 ± 3.008, P = 0.021). However, the signal-to-noise ratio (SNR) value of r-FOV DWI was considerably lower than that of f-FOV DWI (27.80 ± 6.056, 33.67 ± 7.833, P＜0.001). No significant difference was found between mean ADC values of f-FOV DWI and r-FOV DWI. There was a significant tendency for a negative correlation between the ADC values and FIGO stages of CC for both two sequences (r=-0. 436, P＜0.01; r=-0.470, P＜0.01, respectively). The rFOV DWI sequence provided significantly better IQ and lesion conspicuity than the fFOV DWI sequence. In addition, rFOV sequences can be used in evaluation of FIGO staging of cervical cancer.

O-RADS scoring system for adnexal lesions: Diagnostic performance on TVUS performed by an expert sonographer and MRI

To determine the diagnostic performance of transvaginal ultrasound (TVUS) performed by an US specialist and MRI based on the O-RADS scoring system. Between March 5th 2013 and December 31st 2021, 227 patients, referred to our center, underwent TVUS and pelvic MRI for characterization of an adnexal lesion proven by surgery or two years of negative follow-up. All lesions were classified according to O-RADS US and O-RADS MRI risk scoring systems. Imaging data were then correlated with histopathological diagnosis or negative follow-up for 2 years. The prevalence of malignancy was 11.1%. Sensitivity of O-RADS US / O-RADS MRI were respectively of 83.3%/83.3% and specificity was 73.2%/92.9% (p < 0.001). O-RADS MRI was more accurate than O-RADS US even when performed by an US specialist (p < 0.001). When MRI was used after US, 51 lesions were reclassified correctly by MRI and only 4 lesions incorrectly reclassified. Most of the lesions (49/51) rated O-RADS US 4 or 5 and reclassified correctly by MRI were benign, mainly including cystadenomas or cystadenofibromas. Only 4 lesions were misclassified by MRI but correctly classified by ultrasound. Our study suggests that MR imaging has equally high sensitivity but higher specificity than TVUS for the characterization of adnexal lesions based on O-RADS scoring system. MRI should be the recommended second-line technique when a mass is discovered during TVUS and is rated O-RADS 4 and 5 over than TVUS by an US specialist.

MRI-based deep learning and radiomics for preoperative prediction of P53abn endometrial cancer: A multicenter study

To develop and validate a non-invasive magnetic resonance imaging (MRI)-based deep learning and radiomics approach for the preoperative differentiation of p53 abnormal (P53abn) endometrial cancer, facilitating refined risk stratification for personalized treatment planning. In this retrospective multi-institutional analysis, we examined data from 920 patients with histologically confirmed endometrial cancer who underwent preoperative MRI. A two-stage deep learning architecture (V-Net followed by VB-Net) was developed to automate tumor delineation across three participating centers. Extracted radiomic features from these segmented regions were leveraged to build machine learning classifiers-support vector machines (SVM), random forests (RF), logistic regression (LR), and decision trees (DT)-aimed at distinguishing p53-abnormal tumors from other molecular subtypes. Model efficacy was assessed using the Dice similarity coefficient (DSC) for segmentation accuracy and the area under the receiver operating characteristic curve (AUC) for classification performance. The automated segmentation achieved Dice similarity coefficients (DSC) of 77.4%, 84.9%, and 80.1% on T2-weighted imaging (T2WI), diffusion-weighted imaging (DWI), and contrast-enhanced T1-weighted imaging (CE-T1WI) sequences, respectively. Among the four classification models developed, the RF classifier demonstrated the highest AUC values in both internal CV cohort (0.924) and external test cohort (0.863). No statistically significant differences were observed between automated and manual segmentation results across all models (P = 0.109-0.454). The integrated deep learning and radiomics pipeline developed in this study provides a promising non-invasive approach for preoperative risk stratification of endometrial cancer. The model has demonstrated high performance in identifying the P53abn subtype, offering a valuable tool to support personalized treatment planning.

The role of volumetric ADC histogram analysis in preoperatively evaluating the tumour subtype and grade of endometrial cancer

To assess the value of volumetric ADC histogram metrics in evaluating the histological subtype and grade of endometrial cancer. Preoperative MRI datasets of 317 patients with endometrial cancer were used to obtain volumetric ADC histogram metrics (tumour volume; minADC, maxADC and meanADC; 10th, 25th, 50th, 75th and 90th percentiles of ADC; skewness; and kurtosis). The Mann-Whitney test or Student's t-test was used to compare the difference in ADC histogram metrics between endometrioid adenocarcinomas (EACs) and serous endometrial cancers (SECs) and between different tumour grades (G1, G2, G3). The area under the curve (AUC) of the receiver operating characteristic (ROC) curve was used to evaluate the performance of ADC histogram metrics or combined models in predicting the tumour subtype and grade. SECs showed a significantly larger tumour volume (P  0.05). A tumour volume ≥ 7.752 cm Volumetric ADC histogram analysis is helpful for non-invasive preoperatively predicting the subtype of endometrial cancer and differentiating G1 from G2 and G3 EACs.

Pelvic insufficiency fractures in locally advanced cervical cancer: the diagnostic yield of post-treatment MRI in a tertiary centre

To assess the incidence of pelvic insufficiency fractures (PIFs) after concurrent chemoradiotherapy (CCRT) in patients with locally advanced cervical cancer (LACC), their time of onset and risk factors. We also analysed the inter-observer agreement between gynaecologic radiologists (GYN readers) and radiologists expert in musculoskeletal imaging (MSK reader) in detecting PIFs in our tertiary care centre. Patients with confirmed LACC who underwent concurrent chemoradiation (CCRT) at our institution from June 2019 to November 2022 were retrospectively included. These patients underwent follow-up pelvic MRI every 3-6 months after CCRT. Cohen's kappa statistics was employed to assess the inter-observer agreement between GYN and MSK readers.Logistic regression analysis was performed calculating odds ratios (OR) to identify risk factors for PIFs, such as age, body mass index (BMI), diabetes, smoking, hypertension, renal function and tumour size. Eighty-seven patients were included. PIFs were diagnosed in 21/87 (24.1 %) patients with a median onset time of 7.4 months from the end of EBRT. Among risk factors, age was statistically associated with PIFs (OR = 1.057, 95 % CI: 1.005-1.118, p = 0.033) with median age in the fracture group of 61.1 years (range: 52.0-71.5) and 53.8 years (range: 43.8-63.3). BMI was a significant predictor of PIFs (OR = 1.134; 95 % CI: 1.013-1.285; p = 0.027), with a higher median BMI among patients with PIFs (26.5; range: 21.5-31.2) compared to non-fractured patients (23.1; range: 20.2-25.1). Also patients with reduced renal function (eGFR < 60 mL/min) had 3.437 times higher odds of experiencing fractures compared to those with normal eGFR. The GYN readers correctly identified PIFs in 2/21 cases and agreed with the MSK reader in 68/87 cases. The interobserver agreement was poor to fair (K = 0.138; 95 % CI: 0-0.311). PIFs are a common complication of CCRT. Their identification on post-CCRT MRI may decrease the need for further targeted investigations and invasive treatments.

Diagnostic performance of IOTA SR and O-RADS combined with CA125, HE4, and risk of malignancy algorithm to distinguish benign and malignant adnexal masses

To compare the diagnostic performance of International Ovarian Tumour Analysis Simple Rules (IOTA SR) and Ovarian-Adnexal Reporting and Data System (O-RADS), and to analyse whether combining IOTA SR and O-RADS with the biomarkers cancer antigen 125 (CA125), human epididymis protein 4 (HE4), and risk of malignancy algorithm (ROMA) further improves diagnostic performance in women with different menopause status. This study retrospectively included patients with ovarian adnexal masses confirmed by surgical pathology between September 2021 and February 2022. The area under the curve (AUC), sensitivity, and specificity were calculated to evaluate the diagnostic efficacy of IOTA SR, O-RADS, and their combination with CA125, HE4, and ROMA. This study included 1,179 ovarian adnexal masses. In all women, the AUC of IOTA SR was comparable to O-RADS (0.879 vs. 0.889, P = 0.361), and O-RADS had a significantly higher sensitivity than IOTA SR (95.77 % vs. 87.32 %, P < 0.001). In premenopausal women, O-RADS had a significantly higher AUC than other diagnostic strategies (all P < 0.05), and the sensitivity, specificity, and accuracy were 93.33 %, 84.74 %, and 85.59 %, respectively. In postmenopausal women, IOTA SR + ROMA had a significantly higher AUC than other diagnostic strategies (all P < 0.05), and the sensitivity, specificity, and accuracy were 85.37 %, 93.88 %, and 90.00 %, respectively. Our study supports the high diagnostic value of IOTA SR or O-RADS alone in all women, and O-RADS was more sensitive than IOTA SR. In premenopausal women, O-RADS had the highest diagnostic value. In postmenopausal women, IOTA SR outperformed O-RADS, and IOTA SR + ROMA had the highest diagnostic value.

Novel approach to MRI based risk stratification of uterine myometrial lesions

Surgery for uterine mesenchymal tumors is common in gynecology. Preoperative diagnosis of malignant tumors can lead to appropriate management for the lesions. This study aims to externally validate a previous MRI-based expert consensus algorithm and evaluate the potential modification of MR-based scoring system's accuracy in diagnosing uterine mesenchymal tumors (UMT). With institutional ethics committee approval and a waiver of informed consent (CRM-2405-410), a bicentric retrospective observational cohort study was conducted from January 2018 to December 2023. The study included women with a pathological diagnosis of uterine mesenchymal tumor following a pelvic MRI within six months. Clinical and MR criteria were blindly recorded by two radiologists (6- and 3-years' experience in gynaecological MR imaging) who assessed several MR features. Continuous variables were analyzed using a Mann-Whitney test, and categorical variables using Fisher's exact test. Odds ratios (OR) for predicting malignancy were calculated with 95% confidence intervals and p-values. The cohort included 455 women (mean age: 43 years, range: 15-82 years) with mesenchymal tumors: 437 leiomyomas, 2 STUMPs (0.4 %), and 16 malignant UMT (3.5 %). Using initial criteria (enlarged pelvic lymph nodes, T2W signal intensity, DW signal intensity compared to endometrium, and ADC cutoff value of 0.9 × 10 Modified MR imaging evaluation algorithms increase true positive diagnosis of malignant UMTs leading to effective differentiation from benign leiomyomas. The new algorithm can allow for appropriate triage of potentially malignant UMTs, alleviating risk associated with morcellation in patients with uterine leiomyosarcoma. Our study demonstrates that combining 5 criteria based on multivariate analysis in a new algorithm (T2W signal, DW signal, ADC cut off value of 1.23 x 10-3 mm2/sec, tumor margins and menopausal status) allows us to distinguish benign from malignant uterine mesenchymal tumors with an accuracy of 98 % (CI95% 97,1%-98,1%), a sensitivity of 83.3 % (CI95% 79-88) and a specificity of 98.6 % (CI95% 98-99). This model allows to build a stratification score that would help in the management of typical and atypical uterine lesions.

Dynamic contrast-enhanced MRI to characterize angiogenesis in primary epithelial ovarian cancer: An exploratory study

Angiogenesis is essential for tumor growth. Currently, there are no established imaging biomarkers to show angiogenesis in tumor tissue. The aim of this prospective study was to evaluate whether semiquantitative and pharmacokinetic DCE-MRI perfusion parameters could be used to assess angiogenesis in epithelial ovarian cancer (EOC). We enrolled 38 patients with primary EOC treated in 2011-2014. DCE-MRI was performed with a 3.0 T imaging system before the surgical treatment. Two different sizes of ROI were used to evaluate semiquantitative and pharmacokinetic DCE perfusion parameters: a large ROI (L-ROI) covering the whole primary lesion on one plane and a small ROI (S-ROI) covering a small solid, highly enhancing focus. Tissue samples from tumors were collected during the surgery. Immunohistochemistry was used to measure the expression of vascular endothelial growth factor (VEGF), its receptors (VEGFRs) and to analyse microvascular density (MVD) and the number of microvessels. VEGF expression correlated inversely with K We observed that several DCE-MRI parameters correlated with VEGF and VEGFR-2 expression and MVD. Thus, both semiquantitative and pharmacokinetic perfusion parameters of DCE-MRI represent promising tools for the assessment of angiogenesis in EOC.

A meta-analysis of MRI-based radiomic features for predicting lymph node metastasis in patients with cervical cancer

To evaluate the ability of preoperative MRI-based radiomic features in predicting lymph node metastasis (LNM) in patients with cervical cancer. PubMed, Embase, Web of Science, Cochrane Library databases, and four Chinese databases were searched to identify relevant studies published up until October 22, 2021. Two reviewers screened all papers independently for eligibility. We included diagnostic accuracy studies that used radiomics-MRI for LNM in patients with cervical cancer, using histopathology as the reference standard.Quality was assessed using the Quality Assessment of Diagnostic Accuracy Studies 2 and Radiomics Quality Score. Overall diagnostic odds ratio (DOR), sensitivity, specificity and area under the curve (AUC) were calculated to assess the prediction efficacy of MRI-based radiomic features in patients with cervical cancer. Spearman's correlation coefficient was calculated and subgroup analysis performed to investigate causes of heterogeneity. Twelve studies comprising 793 female patients were included. The pooled DOR, sensitivity, specificity, and AUC of radiomics in detecting LNM were 12.08 [confidence interval (CI) 8.18, 17.85], 80% (72%, 87%), 76% (72%, 80%), and 0.83 (0.76, 0.89), respectively. The meta-analysis showed significant heterogeneity among the included studies. No threshold effect was detected. Subgroup analysis showed that multiple sequences, and radiomics combined with clinical factors, radiomics approach [DOR:15.49 (6.06, 39.62), 18.93 (8.46, 42.38), and 10.63 (6.23, 18.12), respectively] could slightly improve diagnostic performance compared with apparent diffusion coefficient-based radiomic features, T2 + dynamic contrast-enhanced MRI-based radiomic features, T2 images-based radiomic features, single radiomics, and human reading [DOR: 4.9 (1.91, 12.74), 7.63 (3.78, 15.38), 8.31 (3.05, 22.61), 16.10 (9.10, 28.47), and 6.46 (3.08, 13.56), respectively]. Our meta-analysis showed that preoperative MRI-based radiomic features performs well in predicting LNM in patients with cervical cancer. This noninvasive and convenient tool may be used to facilitate preoperative identification of LNM.

Glycolytic phenotypes in an evaluation of ovarian carcinoma based on carcinogenesis and BRCA mutation

Recently, a dualistic carcinogenesis model of ovarian cancer has emerged. We aimed to investigate differences in the glycolytic phenotypes of type I and type II ovarian carcinoma on the basis of FDG uptake and in the pathological features according to tumour grade and histology. In total, 386 epithelial ovarian carcinoma patients underwent debulking surgery, and the histopathological results of the patients were retrospectively reviewed from 2003 to 2017. Among these patients, 170 patients had histopathological data that were available due to primary cytoreductive surgery and could be analysed regarding FDG avidity in type I and type II ovarian cancer. The FDG uptake of the tumour (SUVmax), metabolic tumor volume (MTV) and total lesion glycolysis (TLG) were analysed according to the tumour grade, histology and type of ovarian carcinogenesis (type I and II) and prognosis. Among the 386 patients, there was a significant difference in SUVmax among ovarian cancer subtypes. There was a significant increase in SUVmax as the tumour grade increased (8.08 ± 0.63, 10.5 ± 0.40, and 12.7 ± 0.38 for grades I, II and III, respectively, Kruskal-Wallis test, p < 0.0001). Among the 90 type I and 80 type II ovarian carcinoma patients, there was a significant difference in SUVmax (type I and II, 9.47 ± 0.54 and 12.97 ± 0.70, respectively, Mann-Whitney test, p = 0.0003). However, no significant change in SUVmax was observed between BRCA-positive and BRCA-negative patients (N = 80, 13.8 ± 5.78 and 12.4 ± 6.30, Student's t-test, p = 0.3075). Among clinicopathologic and metabolic parameters, type of ovarian cancer, MTV and CA125 were significant factors in the prediction of recurrence. The glycolytic phenotype was related to tumour grade and histological subtype, with significant differences between type I and II ovarian cancer. SUVmax of the ovarian cancer would be considered in the differentiation of type I and II ovarian cancer.

Can MRI features differentiate ovarian mucinous carcinoma from mucinous borderline tumor?

Our purpose was to determine if MRI could be used to distinguish ovarian mucinous carcinoma (MC) from mucinous borderline tumor (MBT). This study included 63 consecutive patients with histopathologically proven ovarian mucinous neoplasms (11 MCs and 52 MBTs) who underwent preoperative contrast-enhanced MRI. MRI images were retrospectively reviewed and compared between the 2 pathologies. The maximum tumor diameters (219.7 ± 80.8 mm vs. 177.4 ± 56.5 mm, p <  0.05) and maximum mural nodule (MN) diameters (41.7 ± 33.8 mm vs. 6.6 ± 8.9 mm, p <  0.01) were significantly larger in MCs than in MBTs. MNs larger than 5 mm (82 % vs. 29 %, p <  0.01) and abnormal ascites (45 % vs. 12 %, p <  0.05) were significantly more frequent in MCs than in MBTs. Apparent diffusion coefficient (ADC) values of MN were significantly lower in MCs than in MBTs (1.20 ± 0.25 × 10 MRI findings of these 2 pathologies overlapped considerably. Compared with MBTs, MCs exhibited larger tumor size, larger MN size, and lower ADC values of MN, and MCs more frequently had MNs larger than 5 mm and abnormal ascites. T2 hypointense microcysts might be a specific MRI finding in MBTs.

Differentiation between ovarian metastasis from colorectal carcinoma and primary ovarian carcinoma: Evaluation of tumour markers and “mille-feuille sign” on computed tomography/magnetic resonance imaging

The purpose of this retrospective study was to evaluate the usefulness of serum tumour markers and morphological characteristics in CT/MRI to differentiate between ovarian metastases from colorectal carcinomas (OMCRC) and primary ovarian carcinomas (POC). Preoperative radiological images of 41 OMCRCs from 27 patients (mean age ± SD: 52.2 ± 10.7 years) and 46 POCs from 36 patients (52.1 ± 12.7 years) were included. Three blinded gynecological radiologists classified tumour morphology into 'mille-feuille sign', 'solid and cystic', 'multicystic without nodules', and 'multicystic with nodules' groups and analysed using Fisher's exact test. Serum carcinoembryonic antigen (CEA), cancer antigen 125 (CA125), and carbohydrate antigen 19-9 levels were compared by Wilcoxon rank-sum test. 'Mille-feuille sign' indicated OMCRC (OMCRC: 8/41, POC: 1/46, specificity = 0.98, p = 0.011) and had excellent interobserver agreement (Fleiss's kappa value = 0.96). 'Solid and cystic' indicated POC (18/41 vs 41/45, p < 0.001) and 'multicystic without nodules' indicated OMCRC (8/41 vs 2/46, p = 0.041). There was no significant difference in 'multicystic with nodules'. CA125 levels were higher in POCs (292.5 U/mL vs. 41.0 U/mL, p = 0.003). CEA levels were higher in OMCRCs (24.5 ng/mL vs 2 ng/mL, p < 0.001). CEA (< 6.3 ng/mL) AND (CA125 (≥87.0 U/mL) OR 'solid and cystic') indicated POC with high accuracy (3/41 vs 44/46, accuracy = 0.94, p < 0.001). Our new method with morphological classification and tumour markers were useful for differentiating the two tumours. In particular, the 'mille-feuille sign' frequently indicated OMCRC with high specificity and excellent interobserver agreement.

Abdominopelvic complications of gynecologic malignancy: Essentials for radiologists

Gynecologic cancers are among the leading causes of cancer-related deaths among female patients, with over 80 % of patients experiencing persistent or long-term effects even after curative treatment. Abdominopelvic complications can arise from the disease itself or treatment-related factors. Tumor-related complications include effects from locoregional invasion (malignant bowel obstruction, obstructive uropathy), tumor rupture (and associated hemorrhage), hypercoagulability (leading to deep vein thrombosis), and infections (including tumor fistulization to the bowel or lower urinary tract, abscesses, pyometra, and/or superinfected necrosis). Treatment-related complications can be subdivided into those following surgery, radiotherapy, or systemic therapy, including immunotherapy. Postoperative complications include paralytic ileus, obstructions, fistulas, anastomotic leaks or strictures, vaginal cuff dehiscence, wound infections, lymphocele, and lymphedema. Radiotherapy-related toxicities include acute toxicities of diarrhea, cystitis, and vaginal mucositis, as well as chronic toxic effects, including radiation enteritis, bladder dysfunction, fistulas, pelvic insufficiency fractures, and sexual dysfunction. Complications of cytotoxic chemotherapy and targeted agents include myelosuppression, neuropathy, mucositis, neutropenic enterocolitis, pneumatosis intestinalis, bowel perforation, tumor-to-bowel fistula, pancreatitis, nephrotoxicity, osteoporosis, and bone loss. Immunotherapy-related toxicities include colitis, enteritis, hepatitis, and pancreatitis. The role of the radiologist in the detection and characterization of these complications is paramount, as imaging is integral to timely diagnosis and multidisciplinary management. An awareness of the spectrum of abdominopelvic complications affecting gynecologic oncology patients is essential to maximal diagnostic accuracy and optimal patient care.

Diffusion-weighted imaging of cervical cancer: Feasibility of ultra-high b-value at 3T

We sought to evaluate the image quality and compare the signal intensity (SI) and apparent diffusion coefficient (ADC) maps of ultra-high b-value (2000 s/mm This study was approved by the institutional review board. Sixty patients diagnosed with cervical cancer by pathology were prospectively included. Female pelvic magnetic resonance imaging was performed using a 3 T magnetic resonance scanner; B1 and B2 images were obtained for evaluation. Two radiologists blinded to the scan parameters evaluated the images for signal loss in the background, spatial distortion, image ghosting, confidence in the lesion delineation, and overall image quality using a 5-point scoring system. The scores were compared using a paired Wilcoxon test. SI was measured in the B1 and B2 images for the tumour and normal reference tissues. Additionally, the SI contrast ratios were calculated and compared using the Mann-Whitney U test, the ADC values of tumours and normal tissues were measured, and the maximum tumour diameters were measured from the B1 and B2 images and compared with those from the T2-weighted images, which was the reference standard. The signal loss in the background, confidence of the lesion delineation and overall image quality scores were higher for the B2 images than for the B1 images (all p < 0.001). The contrast ratios of the tumour-to-normal SI were also higher for the B2 images than for the B1 images (p < 0.01). The mean ADC values derived from the B2 images showed better correlations with the tumour differentiation grades than those from the B1 images. The tumour diameters measured from the B2 images experienced less bias than those from the B1 images. B2 images of DWI are technically feasible to acquire and provide more promising additional information for the delineation of cervical cancer tumours than B1 images of the female pelvis.

Time-synchronized 2-deoxy-2-[18F]fluoro-D-glucose PET/MRI with MR-active trigger and Bayesian penalized likelihood reconstruction: Diagnostic utility for locoregional extension of endometrial cancer

Minimal misregistration of fused PET and MRI images can be achieved with simultaneous positron emission tomography/magnetic resonance imaging (PET/MRI). However, the acquisition of multiple MRI sequences during a single PET emission scan may impair fusion precision of each sequence. This study evaluated the diagnostic utility of time-synchronized PET/MRI using an MR active trigger and a Bayesian penalized likelihood reconstruction algorithm (BPL) to assess the locoregional extension of endometrial cancer. Fifty-five patients with endometrial cancer who underwent pelvic 2-deoxy-2-[ The T classification by time-synchronized PET/MRI was the most concordant with the pathological T classification for the two readers. Time-synchronized PET/MRI had a significantly higher diagnostic performance for deep MI and higher confidence level scores than overlapping PET/MRI for the novice reader (p = 0.033 and p = 0.038, respectively). The differences in bladder dimension on sequential PET/MRI were significantly larger than those on overlapping and time-synchronized PET/MRI (p <0.001). Time-synchronized PET/MRI is superior to conventional PET/MRI for assessing the locoregional extension of endometrial cancer.

Validation of MRI short axis analysis for predicting lymphovascular invasion in endometrial cancer patients

In the context of FIGO classification updates in Endometrial Cancer (EC), lymphovascular space invasion (LVSI) is often either missing or wrongly assessed in preoperative histological analysis. This retrospective study aimed to validate the diagnostic efficacy of systematic short-axis measurement on preoperative MRI for predicting lymphovascular space invasion (LVSI) in patients with EC. A total of 116 patients who underwent preoperative pelvic MRI between January 2015 and December 2019 were included. Two expert radiologists specializing in female pelvic MRI measured the tumor's short axis (previously described by Lavaud et al) on all sequences in sagittal axes T2-weighted and post-contrast T1-weighted images fat suppressed. MRI findings were compared with preoperative biopsy results and postoperative histopathology. The analysis revealed the highest discrepancies between preoperative histology combined with MRI images and final pathology in tumor grade (21.6 %), FIGO stage (39.6 %), and myometrial invasion (27.6 %). A 24 mm threshold for the anteroposterior measurement was used as a predictor of LVSI. The model utilizing this cutoff demonstrated good performance (AUC = 0.61, p < 0.001) and correctly reclassified 19.8 % of patients with preoperative FIGO stage I tumors as FIGO stage II or more after surgery. This approach may enhance the preoperative prediction of LVSI and improve the application of the updated FIGO classification in endometrial cancer. The results suggest that MRI-derived short-axis measurement could be a valuable tool for refining the preoperative assessment of LVSI in EC patients.

Prediction of molecular subtypes of endometrial cancer patients on the basis of intratumoral and peritumoral radiomic features from multiparametric MR images

The purpose of this study was to assess the performance of multiparametric MRI-based radiomic models in predicting the molecular subtypes of endometrial cancer (EC) patients. A total of 310 patients with pathologically confirmed EC who underwent preoperative MRI were enrolled this retrospective study and randomly divided into training (n = 217) and testing (n = 93) cohorts. We extracted 22,640 radiomic features from intratumoral and 3-mm peritumoral regions of interest (ROIs) on MR images. Feature selection was performed using the Mann-Whitney U test, Max-Relevance and Min-Redundancy (mRMR) and the least absolute shrinkage and selection operator (LASSO). Twelve radiomic signatures (RSs) were constructed using logistic regression to predict four molecular subtypes (POLEmut, MMRd, NSMP, and p53abn). The performance of these RSs was assessed using receiving operating characteristic (ROC) curve analysis, and the area under the curve (AUC), sensitivity, specificity, and accuracy were calculated. In the testing cohort, the RSs based on intratumoral features for predicting the POLEmut, MMRd, NSMP and p53abn subtypes yielded AUCs of 0.764, 0.812, 0.893 and 0.731, respectively, whereas those based on peritumoral features yielded AUCs of 0.847, 0.836, 0.871 and 0.804, respectively. The RSs constructed by combining intratumoral and peritumoral features for predicting the POLEmut, MMRd, NSMP and p53abn subtypes had the AUCs of 0.844, 0.880, 0.943 and 0.801, respectively. The combination of intratumoral and peritumoral radiomic features from multiparametric MRI enables effective and noninvasive prediction of EC molecular subtypes.

The Node Reporting and Data System (Node-RADS) for standardized MRI evaluation of lymph nodes in endometrial cancer, integrated with clinicopathological and molecular data

To evaluate the diagnostic performance of Node-RADS score using magnetic resonance imaging (MRI) in predicting lymph node involvement (LNI) in patients with endometrial cancer (EC). Additionally, the applicability of the Node-RADS score was evaluated by three readers with different levels of experience in pelvic imaging. Finally, this study investigated the correlation between the Node-RADS score and the extent of myometrial invasion, histological type, lympho vascular invasion (LVI) and molecular subtype. Out of 108 cases, 82 patients with histologically confirmed locally advanced EC met the inclusion criteria for retrospective analysis. LNI risk was assessed for each pelvic lymph node station using a Node-RADS score (1-5). Diagnostic accuracy was determined by comparing scores to histologic findings, considered as the gold standard. Three independent readers with different experience levels assigned scores. The Node-RADS score strongly correlated with histologically confirmed LNI (AUC: 0.832). A cutoff of Node-RADS ≥ 3 optimally detected metastatic lymph nodes, with 85.71 % sensitivity and 76.47 % specificity. Interobserver agreement was high, with κ values of 0.86 (senior vs. junior reader 1) and 0.70 (senior vs. junior reader 2). A significant positive correlation was found between Node-RADS score and myometrial invasion as well as LVI. Node-RADS score is a reliable, standardized tool for assessing LN stations and enhancing diagnostic accuracy in locoregional staging of EC.

A nomogram for preoperative risk stratification based on MRI morphological parameters in patients with endometrioid endometrial carcinoma

To develop and validate a nomogram based on MRI morphological parameters to preoperatively discriminate between low-risk and non-low-risk patients with endometrioid endometrial carcinoma (EEC). Two hundred eighty-one women with histologically confirmed EEC were divided into training (1.5-T MRI, n = 182) and validation cohorts (3.0-T MRI, n = 99). According to the European Society of Medical Oncology guidelines, the patients were divided into four risk groups: low, intermediate, high-intermediate, and high. Binary classification models were developed (low-risk vs. non-low-risk). Univariate logistic regression (LR) analyses were used to determine which variables to select to build the predictive models. Five classification models were constructed, and the best model was selected. The area under the receiver operating characteristic curve (AUC) was calculated to evaluate the performance of the prediction model and nomogram. P < 0.05 indicated a statistically significant difference. Age and four morphological parameters (tumor size, tumor volume, maximum anteroposterior tumor diameter on sagittal T2-weighted images (APsag), and tumor area ratio (TAR)) were selected, and the LR model was used to construct an MRI morphological nomogram. The AUCs for the nomogram in predicting a non-low-risk of EEC among patients in the training and validation cohorts were 0.856 (sensitivity = 75.0%, specificity = 83.1%) and 0.849 (sensitivity = 74.6%, specificity = 85.0%), respectively. An MRI morphological nomogram was developed and achieved high diagnostic performance for classifying low-risk and non-low-risk EEC preoperatively, which could provide support for therapeutic decision-making. Furthermore, our findings indicate that this nomogram is robust in the clinical application of various field strength data.

MRI in uterine cancers with uncertain origin: Endometrial or cervical? Radiological point of view with review of the literature

Therapeutic options and clinical management of cervical and endometrial cancers differs significantly. When clinical and histological analysis of a uterine mass are unable to differentiate between an endocervical or endometrial origin, magnetic resonance imaging (MRI) plays a pivotal role in discriminating the anatomical origin, supporting the clinician in the treatment planning. Cervical adenocarcinomas are more likely to be centered in the cervical region and involving both cervical canal and stromal ring, with possible parametrial invasion. Endometrial adenocarcinomas usually present an elongated morphology and are centered in the endometrial cavity predominantly involving endometrium and myometrium. On contrast-enhanced sequences, cervical cancers are more frequently hypervascular compared to endometrial cancers. In cases of uncertain findings, diffusion-weighted imaging (DWI) can provide additional helpful information with significantly higher apparent coefficient diffusion (ADC) values in cervical adenocarcinomas compared to endometrial adenocarcinomas. However, even when MRI cannot precisely reveal the origin of the tumor, it provides valuable information on several prognostic factors that can help treatment planning.

Radiomics in endometrial cancer and beyond - a perspective from the editors of the EJR

Preoperative risk stratification in women with endometrial cancer: A comparison of contrast-enhanced MR imaging and diffusion-weighted MR imaging

To compare CE MRI and DWI in the risk stratification of women with endometrial cancer for lymph node metastasis. Two readers independently assessed the degree of myometrial invasion on two separate occasions in a retrospective cohort of 84 women with endometrial cancers: once with CE MRI and standard anatomic sequences and another time with DWI and standard anatomic sequences. Participants were stratified according to their risk of lymph node metastasis following the European Society for Medical Oncology guidelines. The rate of lymph node metastasis was compared between the risk stratification groups obtained using CE MRI or DWI by generalized estimating equations. In the low-risk group, the rate of lymph node metastasis was 1.9% (1/53) when using CE MRI and 1.9% (1/54) when using DWI for reader 1, and 3.8% (2/52) when using CE MRI and 1.9% (1/52) when using DWI for reader 2. The rate of lymph node metastasis in the high-risk group was 25.8% (8/31) when using CE MRI and 26.7% (8/30) when using DWI for reader 1, and 21.9% (7/32) when using CE MRI and 25.0% (8/32) when using DWI for reader 2. There was no significant difference between CE MRI and DWI in the rate of lymph node metastasis according to the risk stratification (p > .05 in both low- and high-risk groups for both readers). DWI might be a comparable alternative to CE MRI in the preoperative risk stratification of women with endometrial cancer for lymph node metastasis.

MRI radiomics: A machine learning approach for the risk stratification of endometrial cancer patients

To investigate radiomics and machine learning (ML) as possible tools to enhance MRI-based risk stratification in patients with endometrial cancer (EC). From two institutions, 133 patients (Institution1 = 104 and Institution2 = 29) with EC and pre-operative MRI were retrospectively enrolled and divided in two a low-risk and a high-risk group according to EC stage and grade. T2-weighted (T2w) images were three-dimensionally annotated to obtain volumes of interest of the entire tumor. A PyRadiomics based and previously validated pipeline was used to extract radiomics features and perform feature selection. In particular, feature stability, variance and pairwise correlation were analyzed. Then, the least absolute shrinkage and selection operator technique and recursive feature elimination were used to obtain the final feature set. The performance of a Support Vector Machine (SVM) algorithm was assessed on the dataset from Institution 1 via 2-fold cross-validation. Then, the model was trained on the entire Institution 1 dataset and tested on the external test set from Institution 2. In total, 1197 radiomics features were extracted. After the exclusion of unstable, low variance and intercorrelated features least absolute shrinkage and selection operator and recursive feature elimination identified 4 features that were used to build the predictive ML model. It obtained an accuracy of 0.71 and 0.72 in the train and test sets respectively. Whole-lesion T2w-derived radiomics showed encouraging results and good generalizability for the identification of low-risk EC patients.

MRI-based radiomics model for distinguishing endometrial carcinoma from benign mimics: A multicenter study

To develop and validate an MRI-based radiomics model for preoperatively distinguishing endometrial carcinoma (EC) with benign mimics in a multicenter setting. Preoperative MRI scans of EC patients were retrospectively reviewed and divided into training set (158 patients from device 1 in center A), test set #1 (78 patients from device 2 in center A) and test set #2 (109 patients from device 3 in center B). Two radiologists performed manual delineation of tumor segmentation on the map of apparent diffusion coefficient (ADC), diffusion-weighted imaging (DWI) and T2-weighted imaging (T2WI). The features were extracted and firstly selected using Chi-square test, followed by refining using binary least absolute shrinkage and selection operator (LASSO) regression. The support vector machine (SVM) was employed to build the radiomics model, which is tuned in the training set using 10-fold cross-validation, and then assessed on the test set. Performance of the model was determined by area under the receiver-operating characteristic curve (AUC), accuracy, sensitivity, specificity and F1-score. Five most informative features are selected from the extracted 3142 features. The SVM with linear kernel was employed to build the radiomics model. The AUCs of the model were 0.989 (95% confidence interval [CI]: 0.968-0.997) for the training set, 0.999 (95% CI: 0.991-1.000) for test set #1, 0.961 (95% CI: 0.902-0.983) for test set #2. Accuracies of the model were 0.937 for the training set (precision: 0.919, recall: 0.963, F1-score: 0.940), 0.974 for test set #1 (precision: 0.949, recall: 1.000, F1-score: 0.974) and 0.908 for test set #2 (precision: 0.899, recall: 0.954, F1-score: 0.925). These results confirmed the efficacy of this model in predicting EC in different centers. The MRI-based radiomics model showed promising potential for distinguishing EC with benign mimics and might be useful for surgical management of EC.

Importance of ADC parameters from histogram analysis corresponding to histological components in endometrial cancer