Development of Gestational Age-Based Fetal Brain and Intracranial Volume Reference Norms Using Deep Learning.

BACKGROUND AND PURPOSE: Fetal brain MR imaging interpretations are subjective and require subspecialty expertise. We aimed to develop a deep learning algorithm for automatically measuring intracranial and brain volumes of fetal brain MRIs across gestational ages. MATERIALS AND METHODS: This retrospective study included 246 patients with singleton pregnancies at 19-38 weeks gestation. A 3D U-Net was trained to segment the intracranial contents of 2D fetal brain MRIs in the axial, coronal, and sagittal planes. An additional 3D U-Net was trained to segment the brain from the output of the first model. Models were tested on MRIs of 10 patients (28 planes) via Dice coefficients and volume comparison with manual reference segmentations. Trained U-Nets were applied to 200 additional MRIs to develop normative reference intracranial and brain volumes across gestational ages and then to 9 pathologic fetal brains. RESULTS: Fetal intracranial and brain compartments were automatically segmented in a mean of 6.8 (SD, 1.2) seconds with median Dices score of 0.95 and 0.90, respectively (interquartile ranges, 0.91-0.96/0.89-0.91) on the test set. Correlation with manual volume measurements was high (Pearson r = 0.996, P < .001). Normative samples of intracranial and brain volumes across gestational ages were developed. Eight of 9 pathologic fetal intracranial volumes were automatically predicted to be >2 SDs from this age-specific reference mean. There were no effects of fetal sex, maternal diabetes, or maternal age on intracranial or brain volumes across gestational ages. CONCLUSIONS: Deep learning techniques can quickly and accurately quantify intracranial and brain volumes on clinical fetal brain MRIs and identify abnormal volumes on the basis of a normative reference standard.

Predictive Value of Noncontrast Head CT with Negative Findings in the Emergency Department Setting.

BACKGROUND AND PURPOSE: Noncontrast head CTs are routinely acquired for patients with neurologic symptoms in the emergency department setting. Anecdotally, noncontrast head CTs performed in patients with prior negative findings with the same clinical indication are of low diagnostic yield. We hypothesized that the rate of acute findings in noncontrast head CTs performed in patients with a preceding study with negative findings would be lower compared with patients being imaged for the first time. MATERIALS AND METHODS: We retrospectively evaluated patients in the emergency department setting who underwent noncontrast head CTs at our institution during a 4-year period, recording whether the patient had undergone a prior noncontrast head CT, the clinical indication for the examination, and the examination outcome. Positive findings on examinations were defined as those that showed any intracranial abnormality that would necessitate a change in acute management, such as acute hemorrhage, hydrocephalus, herniation, or interval worsening of a prior finding. RESULTS: During the study period, 8160 patients in the emergency department setting underwent a total of 9593 noncontrast head CTs; 88.2% (7198/8160) had a single examination, and 11.8% (962/8160) had at least 1 repeat examination. The baseline positive rate of the "nonrepeat" group was 4.3% (308/7198). The 911 patients in the "repeat" group with negative findings on a baseline/first CT had a total of 1359 repeat noncontrast head CTs during the study period. The rate of positive findings for these repeat examinations was 1.8% (25/1359), significantly lower than the 4.3% baseline rate (P < .001). Of the repeat examinations that had positive findings, 80% (20/25) had a study indication that was discordant with that of the prior examination, compared with only 44% (593/1334) of the repeat examinations that had negative findings (P < .001). CONCLUSIONS: In a retrospective observational study based on approximately 10,000 examinations, we found that serial noncontrast head CT examinations in patients with prior negative findings with the same study indication are less likely to have positive findings compared with first-time examinations or examinations with a new indication. This finding suggests a negative predictive value of a prior noncontrast head CT examination with negative findings with the same clinical indication.

Hybrid 3D/2D Convolutional Neural Network for Hemorrhage Evaluation on Head CT.

BACKGROUND AND PURPOSE: Convolutional neural networks are a powerful technology for image recognition. This study evaluates a convolutional neural network optimized for the detection and quantification of intraparenchymal, epidural/subdural, and subarachnoid hemorrhages on noncontrast CT. MATERIALS AND METHODS: This study was performed in 2 phases. First, a training cohort of all NCCTs acquired at a single institution between January 1, 2017, and July 31, 2017, was used to develop and cross-validate a custom hybrid 3D/2D mask ROI-based convolutional neural network architecture for hemorrhage evaluation. Second, the trained network was applied prospectively to all NCCTs ordered from the emergency department between February 1, 2018, and February 28, 2018, in an automated inference pipeline. Hemorrhage-detection accuracy, area under the curve, sensitivity, specificity, positive predictive value, and negative predictive value were assessed for full and balanced datasets and were further stratified by hemorrhage type and size. Quantification was assessed by the Dice score coefficient and the Pearson correlation. RESULTS: A 10,159-examination training cohort (512,598 images; 901/8.1% hemorrhages) and an 862-examination test cohort (23,668 images; 82/12% hemorrhages) were used in this study. Accuracy, area under the curve, sensitivity, specificity, positive predictive value, and negative-predictive value for hemorrhage detection were 0.975, 0.983, 0.971, 0.975, 0.793, and 0.997 on training cohort cross-validation and 0.970, 0.981, 0.951, 0.973, 0.829, and 0.993 for the prospective test set. Dice scores for intraparenchymal hemorrhage, epidural/subdural hemorrhage, and SAH were 0.931, 0.863, and 0.772, respectively. CONCLUSIONS: A customized deep learning tool is accurate in the detection and quantification of hemorrhage on NCCT. Demonstrated high performance on prospective NCCTs ordered from the emergency department suggests the clinical viability of the proposed deep learning tool.