Development of deep learning model for diagnosing muscle-invasive bladder cancer on MRI with vision transformer.

Rationale and objectives: To develop and validate a deep learning (DL) model to automatically diagnose muscle-invasive bladder cancer (MIBC) on MRI with Vision Transformer (ViT). Materials and methods: This multicenter retrospective study included patients with BC who reported to two institutions between January 2016 and June 2020 (training dataset) and a third institution between May 2017 and May 2022 (test dataset). The diagnostic model for MIBC and the segmentation model for BC on MRI were developed using the training dataset with 5-fold cross-validation. ViT- and convolutional neural network (CNN)-based diagnostic models were developed and compared for diagnostic performance using the area under the curve (AUC). The performance of the diagnostic model with manual and auto-generated regions of interest (ROImanual and ROIauto, respectively) was validated on the test dataset and compared to that of radiologists (three senior and three junior radiologists) using Vesical Imaging Reporting and Data System scoring. Results: The training and test datasets included 170 and 53 patients, respectively. Mean AUC of the top 10 ViT-based models with 5-fold cross-validation outperformed those of the CNN-based models (0.831 ± 0.003 vs. 0.713 ± 0.007-0.812 ± 0.006, p < .001). The diagnostic model with ROImanual achieved AUC of 0.872 (95 % CI: 0.777, 0.968), which was comparable to that of junior radiologists (AUC = 0.862, 0.873, and 0.930). Semi-automated diagnosis with the diagnostic model with ROIauto achieved AUC of 0.815 (95 % CI: 0.696, 0.935). Conclusion: The DL model effectively diagnosed MIBC. The ViT-based model outperformed CNN-based models, highlighting its utility in medical image analysis.

Spatial Pattern of Intraprostatic Recurrence after Definitive External-Beam Radiation Therapy for Prostate Cancer: Implications for Focal Boost to Intraprostatic Dominant Lesion.

Purpose: We retrospectively investigated spatial pattern associations between primary and recurrent tumor sites after definitive external-beam radiation therapy (EBRT) for prostate cancer, using positron emission tomography/computed tomography (PET/CT) with a prostate-specific membrane antigen (PSMA)-targeted probe, 18F-FSU-880. Methods and Materials: We used data from our prior phase 2 trial involving patients who received PET/CT with 18F-FSU-880, which was designed to evaluate the tumor detection efficacy of PSMA-PET/CT for recurrent prostate cancer. Data from patients with local intraprostatic recurrence detected by PSMA-PET/CT after definitive EBRT were retrospectively analyzed. The prostate and seminal vesicles were divided into 14 sections. Two diagnostic radiologists separately re-evaluated the intraprostatic location of the primary tumor on magnetic-resonance imaging and that of the recurrent tumor on PSMA-PET/CT, respectively, and the rate of overlap between primary and recurrent tumors was calculated. The overlap rate was defined as "the number of sections that overlapped between the primary tumor and recurrent tumor" divided by "the total number of sections of recurrent tumor". A recurrent tumor was considered to be at the same location as the primary tumor when the overlap rate was equal to or greater than 75%, and a partial overlap was defined as an overlap rate between 25 and 74%. Results: Twelve patients had local recurrence detected by PSMA-PET/CT. The median time to diagnosis of local recurrence was 9.1 (range, 2.2-12.3) years after definitive EBRT. The recurrent tumor was detected at the same location in 25.0%, and a partial overlap was noted in 41.7%. Conclusions: Local intraprostatic recurrence after definitive EBRT often occurs at the same site or at a partially overlapping site adjacent to the primary intraprostatic dominant lesion. Our results support the merit of focal dose-escalation for intraprostatic dominant lesions in definitive EBRT.

Diagnosis of a Dural Arteriovenous Fistula Involving the Superior Sagittal Sinus: The MMA Sign on Routine 3D-TOF MR Angiography.

BACKGROUND AND PURPOSE: The main feeding artery of a dural arteriovenous fistula (DAVF) involving the superior sagittal sinus (SSS) is the middle meningeal artery (MMA). MMA extends continuously from the proximal part to SSS in DAVF involving SSS. It is possible to diagnose DAVF involving SSS by evaluating the proximal part of MMA on routine three-dimensional time-of-flight MR angiography (MRA). We termed the finding in an anteroposterior maximum intensity projection (MIP) image of routine MRA in which MMA was continuously visualized at the top of the imaging slab (the upper corpus callosum level) without tapering as the MMA sign. The purpose of the present study was to examine the frequency of the MMA sign on routine MRA in patients with DAVF involving SSS and control patients. MATERIALS AND METHODS: Subjects comprised 18 patients with DAVF involving SSS confirmed by angiography and 871 control subjects who underwent routine MRA. The presence of the MMA sign was retrospectively evaluated using an anteroposterior MIP image of routine MRA in patients with DAVF involving SSS and control patients. RESULTS: The MMA sign was observed in 17 of the 18 (94%) patients with DAVF involving SSS. In one patient with DAVF involving the posterior part of SSS without the MMA sign, the main feeding artery was the occipital artery, not MMA. The MMA sign was observed in 13 of the 871 (1.5%) control patients. CONCLUSION: The MMA sign on routine MRA is useful, suggesting DAVF involving SSS.

Detection of Hyperdense Arterial Sign in Acute Ischemic Stroke with Dual-Energy Computed Tomography: Optimal Combination with X-ray Energy and Slice Thickness.

BACKGROUND: The hyperdense artery sign (HAS) in acute ischemic stroke (AIS) is considered an important marker of a thrombus on computed tomography (CT). An advantage of scanning with dual-energy CT (DECT) is its ability to reconstruct CT images with various energies using the virtual monochromatic imaging (VMI) technique. The aim of this study was to investigate the optimal combination of X-ray energy and slice thickness to detect HASs on DECT. METHODS: A total of 32 patients with confirmed occlusion of the horizontal (M1) portion of the middle cerebral artery were included in this study. Modified contrast-to-noise ratio (modified CNR) analysis was used as a method for evaluating HASs in AIS. A region of interest (ROI) was set as an HAS, the M1 portion, and an approximately 2 cm diameter ROI was set as the background including the HAS and measured. CT images with X-ray energies from 40 to 190 keV, with increments of 10 keV, were reconstructed based on VMI with 1, 2, and 3 mm slice thicknesses. RESULTS: The top five combinations of X-ray energy and slice thickness in descending order of the mean HAS-modified CNR were as follows: Rank 1, 60 keV-1 mm; Rank 2, 70 keV-1 mm; Rank 3, 60 keV-2 mm; Rank 4, 80 keV-2 mm; Rank 5, 60 keV-3 mm. CONCLUSIONS: Our study showed that the optimal combination to detect an HAS was 60 keV and a 1 mm slice thickness on DECT.

Preoperative Imaging Evaluation of Endometrial Cancer in FIGO 2023.

The staging of endometrial cancer is based on the International Federation of Gynecology and Obstetrics (FIGO) staging system according to the examination of surgical specimens, and has revised in 2023, 14 years after its last revision in 2009. Molecular and histological classification has incorporated to new FIGO system reflecting the biological behavior and prognosis of endometrial cancer. Nonetheless, the basic role of imaging modalities including ultrasound, computed tomography (CT), magnetic resonance imaging (MRI), and positron emission tomography, as a preoperative assessment of the tumor extension and also the evaluation points in CT and MRI imaging are not changed, other than several point of local tumor extension. In the field of radiology, it has also undergone remarkable advancement through the rapid progress of computational technology. The application of deep learning reconstruction techniques contributes the benefits of shorter acquisition time or higher quality. Radiomics, which extract various quantitative features from the images, is also expected to have the potential for the quantitative prediction of risk factors such as histological types and lymphovascular space invasion, which is newly included in the new FIGO system. This article reviews the preoperative imaging diagnosis in new FIGO system and recent advances in imaging analysis and their clinical contributions in endometrial cancer. EVIDENCE LEVEL: 4 TECHNICAL EFFICACY: Stage 3.

Nodal infiltration in endometrial cancer: a prediction model using best subset regression.

OBJECTIVES: To build preoperative prediction models with and without MRI for regional lymph node metastasis (r-LNM, pelvic and/or para-aortic LNM (PENM/PANM)) and for PANM in endometrial cancer using established risk factors. METHODS: In this retrospective two-center study, 364 patients with endometrial cancer were included: 253 in the model development and 111 in the external validation. For r-LNM and PANM, respectively, best subset regression with ten-time fivefold cross validation was conducted using ten established risk factors (4 clinical and 6 imaging factors). Models with the top 10 percentile of area under the curve (AUC) and with the fewest variables in the model development were subjected to the external validation (11 and 4 candidates, respectively, for r-LNM and PANM). Then, the models with the highest AUC were selected as the final models. Models without MRI findings were developed similarly, assuming the cases where MRI was not available. RESULTS: The final r-LNM model consisted of pelvic lymph node (PEN) ≥ 6 mm, deep myometrial invasion (DMI) on MRI, CA125, para-aortic lymph node (PAN) ≥ 6 mm, and biopsy; PANM model consisted of DMI, PAN, PEN, and CA125 (in order of correlation coefficient ß values). The AUCs were 0.85 (95%CI: 0.77-0.92) and 0.86 (0.75-0.94) for the external validation, respectively. The model without MRI for r-LNM and PANM showed AUC of 0.79 (0.68-0.89) and 0.87 (0.76-0.96), respectively. CONCLUSIONS: The prediction models created by best subset regression with cross validation showed high diagnostic performance for predicting LNM in endometrial cancer, which may avoid unnecessary lymphadenectomies. CLINICAL RELEVANCE STATEMENT: The prediction risks of lymph node metastasis (LNM) and para-aortic LNM can be easily obtained for all patients with endometrial cancer by inputting the conventional clinical information into our models. They help in the decision-making for optimal lymphadenectomy and personalized treatment. KEY POINTS: •Diagnostic performance of lymph node metastases (LNM) in endometrial cancer is low based on size criteria and can be improved by combining with other clinical information. •The optimized logistic regression model for regional LNM consists of lymph node ≥ 6 mm, deep myometrial invasion, cancer antigen-125, and biopsy, showing high diagnostic performance. •Our model predicts the preoperative risk of LNM, which may avoid unnecessary lymphadenectomies.

Prognostic value of pretreatment FDG PET/CT in uterine cervical cancer according to two major histologic types: squamous cell carcinoma and adenocarcinoma.

Objectives: The aim of this study was to assess the prognostic value of pretreatment Positron emission tomography / computed tomography using 18F-fluorodeoxyglucose (FDG-PET/CT) in cervical cancer according to two major histologic types. Methods: Eighty-three squamous cell carcinoma (SCC) patients and 35 adenocarcinoma (AC) patients who underwent pretreatment FDG-PET/CT were retrospectively analyzed. Maximum standardized uptake value (SUVmax), mean standardized uptake value (SUVmean), metabolic tumor volume (MTV), and total lesion glycolysis (TLG) of the primary tumor were calculated. Kaplan-Meier analyses were used to compare correlations between each PET parameter and overall survival (OS). The prognostic values of imaging and clinical parameters were assessed using uni- and multivariable Cox proportional hazard models. Results: SUVmax, SUVmean, and TLG were significantly higher in SCC than in AC (p<0.01 each). No significant difference in MTV was seen between the two groups (p=0.10). As for Kaplan-Meier analyses, in SCC, patients with SUVmax, SUVmean, MTV, and TLG exceeding cutoff values tended to show worse OS than patients with lower values (p=0.07, p=0.27, p<0.01, and p=0.01, respectively, for OS). On the other hand, in AC, patients with MTV and TLG exceeding cutoff values showed significantly worse PFS and OS (p<0.01 each for OS), while SUVmax and SUVmean were unrelated to OS (p=0.91 and p=0.83, respectively for OS). As for multivariable analyses, in SCC, TLG was identified as an independent prognostic factor for OS (p=0.01). In AC, MTV was identified as an independent prognostic factor for OS (p=0.02). Conclusion: Our preliminary data suggest that FDG-PET/CT would be useful for predicting prognosis in cervical cancer, although the clinical significance of quantitative values may differ according to histopathological type.

Development of an Algorithm to Differentiate Uterine Sarcoma from Fibroids Using MRI and LDH Levels.

BACKGROUND: This study aimed to establish an evaluation method for detecting uterine sarcoma with 100% sensitivity using MRI and serum LDH levels. METHODS: One evaluator reviewed the MRI images and LDH values of a total of 1801 cases, including 36 cases of uterine sarcoma and 1765 cases of uterine fibroids. The reproducibility of the algorithm was also examined by four evaluators with different imaging experience and abilities, using a test set of 61 cases, including 14 cases of uterine sarcoma. RESULTS: From the MRI images and LDH values of 1801 cases of uterine sarcoma and uterine fibroids, we found that all sarcomas were included in the group with a high T2WI and either a high T1WI, an unclear margin, or high LDH values. In addition, when cases with DWI were examined, all sarcomas had high DWI. Among the 36 sarcoma cases, the group with positive findings for T2WI, T1WI, margins, and serum LDH levels all had a poor prognosis (p = 0.015). The reproducibility of the algorithm was examined by four evaluators and the sensitivity of sarcoma detection ranged from 71% to 93%. CONCLUSION: We established an algorithm to distinguish uterine sarcoma if tumors in the myometrium with low T2WI and DWI are present.

A case of paraovarian tumor of borderline malignancy with decrease of apparent diffusion coefficient value and marked 18F-fluorodeoxyglucose accumulation.

Para-ovarian cysts are occasionally encountered in clinical practice; however, malignant tumors derived from them are rare. Due to its rarity, the characteristic imaging findings of para-ovarian tumors with borderline malignancy (PTBM) are largely unknown. Herein, we report a case of PTBM, along with imaging findings. A 37-year-old woman came to our department with a suspected malignant adnexal tumor. Pelvic contrast-enhanced magnetic resonance imaging (MRI) revealed a solid part within the cystic tumor with a decrease in the apparent diffusion coefficient (ADC) value (1.16 × 10-3 mm2/s). We also performed Positron Emission Tomography-MRI and showed a strong accumulation of 18F-fluorodeoxyglucose (FDG) in the solid part (SUVmax = 14.8). In addition, the tumor appeared to develop independently of the ovary. Because tumor was derived from para-ovarian cyst, we suspected PTBM preoperatively and planned fertility sparing treatment. Pathological examination revealed a serous borderline tumor and PTBM was confirmed. PTBM can have unique imaging characteristics, including a low ADC value and high FDG accumulation. When a tumor appears to develop from para-ovarian cysts, borderline malignancy can be suspected, even if imaging findings suggest malignant potential.

Obstetric pelvimetry by three-dimensional computed tomography in non-pregnant Japanese women: a retrospective single-center study.

OBJECTIVE: While a basic understanding of pelvic size and typology is still important for obstetricians, pelvic measurement data for Japanese women are very scarce. To our best knowledge, no large-scale pelvimetry studies of Japanese women have been made for the past 50 years. This study aimed to investigate the accurate size, particularly the obstetric conjugate (OC) and transverse diameter of the pelvic inlet (TD), of modern Japanese women, using three-dimensional (3D) computed tomography (CT), and to obtain their reference values. METHODS: This retrospective, single-center observational study enrolled Japanese non-pregnant women aged between 20 and 40 years, who underwent pelvic CT examination from 2016 to 2021. CT was performed for various reasons, including acute abdomen, search for cancer metastases, and follow-up of existing disease. However, no cases were taken for pelvic measurements. Pelvimetry was performed retrospectively using a 3D workstation. The OC was measured on a strict lateral view and the TD was measured on an axial-oblique view. Other clinical data, such as age, height, and weight, were also extracted from the medical charts and analyzed. RESULTS: A total of 1,263 patients were enrolled, with the mean age of 32.7 years (standard deviation [SD] 6.2). The mean height, weight, and body mass index were 158.8 cm (SD 5.8), 54.8 kg (SD 11.7), and 21.7 kg/m2 (SD 4.4), respectively. The mean OC length was 127.0 mm (SD 9.5, 95% confidence interval [CI] 126.5-127.5), while the mean TD length was 126.8 mm (SD 7.5, 95% CI 126.4-127.2). Both values were normally distributed. Height was significantly associated with OC (regression coefficient = 0.75 [95% CI 0.66-0.84], p < .001) and TD (regression coefficient = 0.63 [95% CI 0.56-0.70], p < .001). Age showed a weak but statistically significant positive association with TD (regression coefficient = 0.14 [95% CI 0.07-0.20], p < .001) and OC (regression coefficient = -0.10 [95% CI -0.18 to -0.01], p = .026). CONCLUSION: The 3D CT pelvimetry in 1,263 non-pregnant Japanese women of childbearing age revealed the mean OC and TD of 127.0 mm, and 126.8 mm, which were 11.8 mm and 4.3 mm larger, respectively, than those in the survey in 1972. Our data will be referred to in clinical practice as the standard pelvic measurement values for the Japanese population.

Automatic segmentation of bladder cancer on MRI using a convolutional neural network and reproducibility of radiomics features: a two-center study.

This study aimed to develop a versatile automatic segmentation model of bladder cancer (BC) on MRI using a convolutional neural network and investigate the robustness of radiomics features automatically extracted from apparent diffusion coefficient (ADC) maps. This two-center retrospective study used multi-vendor MR units and included 170 patients with BC, of whom 140 were assigned to training datasets for the modified U-net model with five-fold cross-validation and 30 to test datasets for assessment of segmentation performance and reproducibility of automatically extracted radiomics features. For model input data, diffusion-weighted images with b = 0 and 1000 s/mm2, ADC maps, and multi-sequence images (b0-b1000-ADC maps) were used. Segmentation accuracy was compared between ours and existing models. The reproducibility of radiomics features on ADC maps was evaluated using intraclass correlation coefficient. The model with multi-sequence images achieved the highest Dice similarity coefficient (DSC) with five-fold cross-validation (mean DSC = 0.83 and 0.79 for the training and validation datasets, respectively). The median (interquartile range) DSC of the test dataset model was 0.81 (0.70-0.88). Radiomics features extracted from manually and automatically segmented BC exhibited good reproducibility. Thus, our U-net model performed highly accurate segmentation of BC, and radiomics features extracted from the automatic segmentation results exhibited high reproducibility.

Placental functional assessment and its relationship to adverse pregnancy outcome: comparison of intravoxel incoherent motion (IVIM) MRI, T2-relaxation time, and umbilical artery Doppler ultrasound.

BACKGROUND: Early identification of placental insufficiency can lead to appropriate treatment selections and can improve neonates' outcomes. Possible contributions of magnetic resonance imaging (MRI) have been suggested. PURPOSE: To evaluate the prognostic capabilities of placental intravoxel incoherent motion (IVIM) parameters and T2-relaxation time, and their correlation with fetal growth and adverse outcomes, comparing umbilical artery (UmA) pulsatility index (PI). MATERIAL AND METHODS: A total of 68 singleton pregnancies at 24-40 weeks of gestation underwent placental MRI and were reviewed retrospectively. UmA-PI was measured using Doppler ultrasound by obstetricians. IVIM parameters (Dfast, Dslow, and f) were calculated with a Bayesian model fitting. First, the associations between gestational age (GA) with placental IVIM parameters, T2-relaxation time, and placental thickness (PT) were evaluated. Second, IVIM parameters, T2 value (Z-score), PT (Z-score), and UmA-PI (Z-score) were compared between ( 1) those delivering small for gestational age (SGA) and appropriate for gestational age (AGA) neonates, ( 2) emergency cesarean section (ECS), and non-ECS, and ( 3) preterm birth and full-term birth. RESULTS: Low birth weight was observed in 15/68 cases (22%). GA was significantly associated only with T2-relaxation time and PT. SGA was significantly associated with T2 value (Z-score), f, and UmA-PI (Z-score). In the ECS groups, T2 value (Z-score), f, and Dfast were significantly lower than those in non-ECS groups. All IVIM parameters and T2 values (Z-score) showed significantly lower scores in the preterm birth group. CONCLUSION: Placental f and T2 value (Z-score) had significant associations with low birth weight and clinical adverse outcomes and could be potential imaging biomarkers of placental insufficiency.

Risk Stratification for Pregnancies Diagnosed With Fetal Growth Restriction Based on Placental MRI.

BACKGROUND: Diagnosis of fetal growth restriction (FGR) entails difficulties with differentiating fetuses not fulfilling their growth potential because of pathologic conditions, such as placental insufficiency, from constitutionally small fetuses. The feasibility of placental MRI for risk stratification among pregnancies diagnosed with FGR remains unexplored. PURPOSE: To explore quantitative MRI features useful to identify pregnancies with unfavorable outcomes and to assess the diagnostic performance of visual analysis of MRI to detect pregnancies with unfavorable outcomes, among pregnancies diagnosed with FGR. STUDY TYPE: Retrospective. POPULATION: Thirteen pregnancies with unfavorable outcomes (preterm emergency cesarean section or intrauterine fetal death) and 11 pregnancies with favorable outcomes performed MRI at gestational weeks 21-36. FIELD STRENGTH/SEQUENCE: A 5-T, half-Fourier-acquired single-shot turbo spin echo (HASTE), spin-echo echo-planar imaging (SE-EPI) and T2 map derived from SE-EPI. ASSESSMENT: Placental size on HASTE sequences and T2 mapping-based histogram features were extracted. Three radiologists qualitatively evaluated the visibility of maternal cotyledon on HASTE and SE-EPI sequences with echo times (TEs) = 60, 90, and 120 msec using 3-point Likert scales: 0, absent; 1, equivocal; and 2, present. STATISTICAL TESTS: Welch's t-test or Mann-Whitney U test for quantitative features between the favorable and unfavorable outcome groups. Areas under the receiver operating curves (AUCs) of the three readers' visual analyses to detect pregnancies with unfavorable outcomes. A P value of <0.05 was inferred as statistically significant. RESULTS: Placental size (major and minor axis, estimated area of placental bed, and volume of placenta) and T2 mapping-based histogram features (mean, skewness, and kurtosis) were statistically significantly different between the two groups. Visual analysis of HASTE and SE-EPI with TE = 60 msec showed AUCs of 0.80-0.86 to detect pregnancies with unfavorable outcomes. DATA CONCLUSION: Placental size, histogram features, and visual analysis of placental MRI may allow for risk stratification regarding outcomes among pregnancies diagnosed with FGR. EVIDENCE LEVEL: 3 TECHNICAL EFFICACY: Stage 5.

MR findings of polypoid endometriosis of female genital organs: report of three cases.

Polypoid endometriosis is a benign, rare variant of endometriosis that forms polypoid nodules mimicking malignant tumors. For three cases of polypoid endometriosis of female genital organs, this report presents characteristic MR imaging features reflecting the histopathological findings. The solid and microcystic pattern or the multilocular pattern both reflecting dilated endometrial glands, and characteristic morphology of the nodules, multilobulated or polypoid-shaped, were helpful diagnostic clues present in these three cases. Earlier reported MR findings were also recognized, including signal intensity similar to that of the endometrium on T2-weighted image and contrast enhanced T1-weighted image, hypointense rim on T2-weighted image, lack of diffusion restriction, and hyperintense foci on T1-weighted image. Two cases were diagnosed preoperatively based on MR imaging findings as polypoid endometriosis. Fertility-preserving treatment was administered for one patient. Preoperative inference of polypoid endometriosis from MR imaging can avoid overtreatment and lead to fertility preservation.

MRI in the Diagnosis of Endometriosis and Related Diseases.

Endometriosis, a common chronic inflammatory disease in female of reproductive age, is closely related to patient symptoms and fertility. Because of its high contrast resolution and objectivity, MRI can contribute to the early and accurate diagnosis of ovarian endometriotic cysts and deeply infiltrating endometriosis without the need for any invasive procedure or radiation exposure. The ovaries, which are the most frequent site of endometriosis, can be afflicted by multiple related conditions and diseases. For the diagnosis of deeply infiltrating endometriosis and secondary adhesions among pelvic organs, fibrosis around the ectopic endometrial gland is usually found as a T2 hypointense lesion. This review summarizes the MRI findings obtained for ovarian endometriotic cysts and their physiologically and pathologically related conditions. This article also includes the key imaging findings of deeply infiltrating endometriosis.

Chronic abruption-oligohydramnios sequence (CAOS) revisited: possible implication of premature rupture of membranes.

AIM: The pathogenic mechanism of chronic abruption-oligohydramnios sequence (CAOS) remains unknown, and there are no objective standards for diagnosis on imaging or using pathological evidence. We aimed to reconsider and clarify the true pathology of CAOS by integrating clinical, magnetic resonance imaging (MRI) and histopathological findings of the placenta. MATERIAL AND METHODS: This is a case series of patients with CAOS managed at our hospital between 2010 and 2020. The clinical data of the patients, including MRI findings and placental pathology, were reviewed retrospectively. RESULTS: A total of 18 patients were eligible. Preterm birth occurred in 17 (94%) cases; the median gestational age at delivery was 25. Three neonates (17%) died within two years, and 10 neonates (56%) developed chronic lung disease. MRI was performed in 13 cases and clearly showed intrauterine hematoma and hemorrhagic amniotic fluid. Pathologically, in all cases, retroplacental hematoma was not detected, and fetal membranes were extremely fragile and ragged. Shedding and necrosis of the amniotic epithelium was a characteristic finding, which was confirmed in 17 cases (94%). Diffuse chorionic hemosiderosis (DCH) was detected in all cases. CONCLUSIONS: The fundamental cause of CAOS is repeated intrauterine hemorrhage and subsequent subchorionic hematoma, which induces hemorrhagic amniotic fluid and DCH. Consequently, these factors result in the necrosis and weakening of the amnion. Therefore, the true pathology of CAOS is believed to be premature rupture of membranes rather than chronic abruption.

Diagnostic Value of DCE-MRI for Differentiating Malignant Adnexal Masses Compared with Contrast-enhanced-T1WI.

PURPOSE: To compare the diagnostic performance of dynamic contrast-enhanced-MR (DCE-MR) and delayed contrast-enhanced (CE)-MRI added to unenhanced MRI, including diffusion weighted image (DWI) for differentiating malignant adnexal tumors, conducting a retrospective blinded image interpretation study. METHODS: Data of 80 patients suspected of having adnexal tumors by ultrasonography between April 2008 and August 2018 were used for the study. All patients had undergone preoperative MRI and surgical resection at our institution. Four radiologists (two specialized in gynecological radiology and two non-specialized) were enrolled for blinded review of the MR images. A 3-point scale was used: 0 = benign, 1 = indeterminate, and 2 = malignant. Three imaging sets were reviewed: Set A, unenhanced MRI including DWI; Set B, Set A and delayed CE-T1WI; and Set C, Set A and DCE-MRI. Imaging criteria for benign and malignant tumors were given in earlier reports. The diagnostic performance of the three imaging sets of the four readers was calculated. Their areas under the curve (AUCs) were compared using the DeLong method. RESULTS: Accuracies of Set B were 81%-88%. Those of Set C were 81%-85%. The AUCs of Set B were 0.83 and 0.89. Those of Set C were 0.81-0.86. For two readers, Set A showed lower accuracy and AUC than Set B/Set C (less than 0.80), although those were equivalent in other readers. No significant difference in AUCs was found among the three sequence sets. Intrareader agreement was moderate to almost perfect in Sets A and B, and substantial to almost perfect in Set C. CONCLUSION: DCE-MR showed no superiority for differentiating malignant adnexal tumors from benign tumors compared to delayed CE-T1WI with conventional MR and DWI.

Radiomic machine learning for pretreatment assessment of prognostic risk factors for endometrial cancer and its effects on radiologists' decisions of deep myometrial invasion.

PURPOSE: To evaluate radiomic machine learning (ML) classifiers based on multiparametric magnetic resonance images (MRI) in pretreatment assessment of endometrial cancer (EC) risk factors and to examine effects on radiologists' interpretation of deep myometrial invasion (dMI). METHODS: This retrospective study examined 200 consecutive patients with EC during January 2004 -March 2017, divided randomly to Discovery (n = 150) and Test (n = 50) datasets. Radiomic features of tumors were extracted from T2-weighted images, apparent diffusion coefficient map, and contrast enhanced T1-weighed images. Using the Discovery dataset, feature selection and hyperparameter tuning for XGBoost were performed. Ten classifiers were built to predict dMI, histological grade, lymphovascular invasion (LVI), and pelvic/paraaortic lymph node metastasis (PLNM/PALNM), respectively. Using the Test dataset, the diagnostic performances of ten classifiers were assessed by the area under the receiver operator characteristic curve (AUC). Next, four radiologists assessed dMI independently using MRI with a Likert scale before and after referring to inference of the ML classifier for the Test dataset. Then, AUCs obtained before and after reference were compared. RESULTS: In the Test dataset, mean AUC of ML classifiers for dMI, histological grade, LVI, PLNM, and PALNM were 0.83, 0.77, 0.81, 0.72, and 0.82. AUCs of all radiologists for dMI (0.83-0.88) were better than or equal to mean AUC of the ML classifier, which showed no statistically significant difference before and after the reference. CONCLUSION: Radiomic classifiers showed promise for pretreatment assessment of EC risk factors. Radiologists' inferences outperformed the ML classifier for dMI and showed no improvement by review.