In this study, we developed an automated workflow using a deep learning model (DL) to measure the lateral ventricle linearly in fetal brain MRI, which are subsequently classified into normal or ventriculomegaly, defined as a diameter wider than 10 mm at the level of the thalamus and choroid plexus. To accomplish this, we first trained a UNet-based deep learning model to segment the brain of a fetus into seven different tissue categories using a public dataset (FeTA 2022) consisting of fetal T2-weighted images. Then, an automatic workflow was developed to perform lateral ventricle measurement at the level of the thalamus and choroid plexus. The test dataset included 22 cases of normal and abnormal T2-weighted fetal brain MRIs. Measurements performed by our AI model were compared with manual measurements performed by a general radiologist and a neuroradiologist. The AI model correctly classified 95% of fetal brain MRI cases into normal or ventriculomegaly. It could measure the lateral ventricle diameter in 95% of cases with less than a 1.7 mm error. The average difference between measurements was 0.90 mm in AI vs. general radiologists and 0.82 mm in AI vs. neuroradiologists, which are comparable to the difference between the two radiologists, 0.51 mm. In addition, the AI model also enabled the researchers to create 3D-reconstructed images, which better represent real anatomy than 2D images. When a manual measurement is performed, it could also provide both the right and left ventricles in just one cut, instead of two. The measurement difference between the general radiologist and the algorithm (p = 0.9827), and between the neuroradiologist and the algorithm (p = 0.2378), was not statistically significant. In contrast, the difference between general radiologists vs. neuroradiologists was statistically significant (p = 0.0043). To the best of our knowledge, this is the first study that performs 2D linear measurement of ventriculomegaly with a 3D model based on an artificial intelligence approach. The paper presents a step-by-step approach for designing an AI model based on several radiological criteria. Overall, this study showed that AI can automatically calculate the lateral ventricle in fetal brain MRIs and accurately classify them as abnormal or normal.
Background. Corpus callosal abnormalities (CCA) are midline developmental brain malformations and are usually associated with a wide spectrum of other neurological and non-neurological abnormalities. The study aims to highlight the diagnostic role of fetal MRI to characterize heterogeneous corpus callosal abnormalities using the latest classification system. It also helps to identify associated anomalies, which have prognostic implications for the postnatal outcome. Methods. In this study, retrospective data from antenatal women who underwent fetal MRI between January 2014 and July 2023 at Rush University Medical Center were evaluated for CCA and classified based on structural morphology. Patients were further assessed for associated neurological and non-neurological anomalies. Results. The most frequent class of CCA was complete agenesis (79.1%), followed by hypoplasia (12.5%), dysplasia (4.2%), and hypoplasia with dysplasia (4.2%). Among them, 17% had isolated CCA, while the majority (83%) had complex forms of CCA associated with other CNS and non-CNS anomalies. Out of the complex CCA cases, 58% were associated with other CNS anomalies, while 8% were associated with non-CNS anomalies. 17% of cases had both. Conclusion. The use of fetal MRI is valuable in the classification of abnormalities of the corpus callosum after the confirmation of a suspected diagnosis on prenatal ultrasound. This technique is an invaluable method for distinguishing between isolated and complex forms of CCA, especially in cases of apparent isolated CCA. The use of diffusion-weighted imaging or diffusion tensor imaging in fetal neuroimaging is expected to provide further insights into white matter abnormalities in fetuses diagnosed with CCA in the future.
31 August 2023 in American Journal of Neuroradiology
BACKGROUND AND PURPOSE: An MRI of the fetus can enhance the identification of perinatal developmental disorders, which improves the accuracy of ultrasound. Manual MRI measurements require training, time, and intra-variability concerns. Pediatric neuroradiologists are also in short supply. Our purpose was developing a deep learning model and pipeline for automatically identifying anatomic landmarks on the pons and vermis in fetal brain MR imaging and suggesting suitable images for measuring the pons and vermis. MATERIALS AND METHODS: We retrospectively used 55 pregnant patients who underwent fetal brain MR imaging with a HASTE protocol. Pediatric neuroradiologists selected them for landmark annotation on sagittal single-shot T2-weighted images, and the clinically reliable method was used as the criterion standard for the measurement of the pons and vermis. A U-Net-based deep learning model was developed to automatically identify fetal brain anatomic landmarks, including the 2 anterior-posterior landmarks of the pons and 2 anterior-posterior and 2 superior-inferior landmarks of the vermis. Four-fold cross-validation was performed to test the accuracy of the model using randomly divided and sorted gestational age–divided data sets. A confidence score of model prediction was generated for each testing case. RESULTS: Overall, 85% of the testing results showed a ≥90% confidence, with a mean error of <2.22 mm, providing overall better estimation results with fewer errors and higher confidence scores. The anterior and posterior pons and anterior vermis showed better estimation (which means fewer errors in landmark localization) and accuracy and a higher confidence level than other landmarks. We also developed a graphic user interface for clinical use. CONCLUSIONS: This deep learning–facilitated pipeline practically shortens the time spent on selecting good-quality fetal brain images and performing anatomic measurements for radiologists.
Accurately predicting a fetus’s gestational age (GA) is crucial in prenatal care. This study aimed to develop an artificial intelligence (AI) model to predict GA using biometric measurements from fetal brain magnetic resonance imaging (MRI). We assessed the significance of using different reference standards for interpreting GA predictions. Measurements of biparietal diameter (BPD), fronto-occipital diameter (FOD), and head circumference (HC) were obtained from 52 normal fetal MRI cases from Rush University. Both manual and AI-based methods were utilized, and comparisons were made using three reference standards (Garel, Freq, and Bio). The AI model showed a strong correlation with manual measurements, particularly for HC, which exhibited the highest correlation with actual values. Differences between GA predictions and picture archiving and communication system (PACS) records varied by reference, ranging from 0.47 to 2.17 weeks for BPD, 0.46 to 2.26 weeks for FOD, and 0.75 to 1.74 weeks for HC. Pearson correlation coefficients between PACS records and GA predictions exceeded 0.97 across all references. In conclusion, the AI model demonstrated high accuracy in predicting GA from fetal brain MRI measurements. This approach offers improved accuracy and convenience over manual methods, highlighting the potential of AI in enhancing prenatal care through precise GA estimation.
Background/Objective: The accurate diagnosis of congenital central nervous system abnormalities is critical to pre- and postnatal prognostication and management. When an abnormality is found in the posterior fossa of the fetal brain, parental counseling is challenging because of the wide spectrum of clinical and neurodevelopmental outcomes in patients with Dandy–Walker (DW) spectrum posterior malformations. The objective of this study was to evaluate the utility of biometric measurements obtained from fetal magnetic resonance imaging (MRI) to facilitate the prenatal differentiation of Dandy–Walker (DW) spectrum malformations, including vermian hypoplasia (VH), Blake’s pouch cyst (BPC), and classic Dandy–Walker malformation (DWM). Methods: This retrospective single-center study evaluated 34 maternal–infant dyads referred for fetal MRI evaluation of suspected DW spectrum malformations identified on antenatal ultrasound. Radiologists took posterior fossa measurements, including the vermis anteroposterior (AP) diameter, vermis height (VH), and tegmento–vermian angle (TVA). The posterior fossa, fourth ventricle, and cisterna magna were classified as normal, large, or dilated. The postnatal imaging findings were evaluated for concordance. The acquired values were compared between the groups and with normative data. The genetic testing results are reported when available. Results: A total of 27 DW spectrum fetal MRI cases were identified, including 7 classic DWMs, 14 VHs, and 6 BPCs. The TVA was significantly higher in the DWM group compared with the VH and BPC groups (p < 0.001). All three groups had reduced AP vermis measurements for gestational age compared with normal fetal brains, as well as differences in the means across the groups (p = 0.002). Conclusions: Biometric measurements derived from fetal MRI can effectively facilitate the prenatal differentiation of VH, BPC, and classic DWM when assessing DW spectrum posterior fossa lesions. Standardizing biometric measurements may increase the diagnostic utility of fetal MRI and facilitate improved antenatal counseling and clinical decision-making.
Accurately predicting a fetus's gestational age (GA) is of utmost importance in prenatal care. This study aimed to develop an artificial intelligence (AI) model that can automatically predict GA using biometric measurements derived from fetal brain mag-netic resonance imaging (MRI). Additionally, we aimed to assess the significance of con-sidering different references when interpreting GA predictions. To achieve this, we obtained measurements such as Biparietal Diameter (BPD), Fron-to-occipital Diameter (FOD), and Head Circumference (HC) from a dataset comprising 52 normal fetal MRI cases with T2 Haste sequences from Rush University. Both manual and AI-based methods were utilized to acquire these measurements. We also employed three reference papers (Garel, Freq, and Bio) for comparison purposes. The results demonstrated a strong correlation between manual and AI measure-ments, indicating consistency between the two methods. The AI-based measurement of HC exhibited a higher correlation with actual values compared to BPD, FOD, and correct-ed BPD (BPDC). When comparing these measurements with GA in the Picture Archiving and Communication System (PACS), the differences varied depending on the reference used. Specifically, the differences ranged from 0.47 to 2.17 weeks for BPD, 0.46 to 2.26 weeks for FOD, and 0.75 to 1.74 weeks for HC. Furthermore, the Pearson correlation coeffi-cient analysis revealed that all correlation coefficients between PACS records and GA pre-dictions using different references were greater than 0.97. In conclusion, the AI model based on fetal brain MRI accurately predicts GA by uti-lizing BPD, FOD, and HC measurements. The AI approach, which involves combining line segments to calculate fetal head circumference, offers improved accuracy and con-venience compared to manual estimation. This study underscores the potential of AI models in accurately estimating gestational age and highlights their utility in prenatal care. By integrating AI as a valuable tool in prenatal care, we can enhance the accuracy, ef-ficiency, and decision-making involved in assessing fetal development and monitoring pregnancies using MRI measurements.