Towards automated extraction of 2D standard fetal head planes from 3D ultrasound acquisitions: A clinical evaluation and quality assessment comparison.

INTRODUCTION: Clinical evaluation of deep learning (DL) tools is essential to compliment technical accuracy metrics. This study assessed the image quality of standard fetal head planes automatically-extracted from three-dimensional (3D) ultrasound fetal head volumes using a customised DL-algorithm. METHODS: Two observers retrospectively reviewed standard fetal head planes against pre-defined image quality criteria. Forty-eight images (29 transventricular, 19 transcerebellar) were selected from 91 transabdominal fetal scans (mean gestational age = 26 completed weeks, range = 20+5-32+3 weeks). Each had two-dimensional (2D) manually-acquired (2D-MA), 3D operator-selected (3D-OS) and 3D-DL automatically-acquired (3D-DL) images. The proportion of adequate images from each plane and modality, and the number of inadequate images per plane was compared for each method. Inter and intra-observer agreement of overall image quality was calculated. RESULTS: Sixty-seven percent of 3D-OS and 3D-DL transventricular planes were adequate quality. Forty-five percent of 3D-OS and 55% of 3D-DL transcerebellar planes were adequate. Seventy-one percent of 3D-OS and 86% of 3D-DL transventricular planes failed with poor visualisation of intra-cranial structures. Eighty-six percent of 3D-OS and 80% of 3D-DL transcerebellar planes failed due to inadequate visualisation of cerebellar hemispheres. Image quality was significantly different between 2D and 3D, however, no significant difference between 3D-modalities was demonstrated (p < 0.005). Inter-observer agreement of transventricular plane adequacy was moderate for both 3D-modalities, and weak for transcerebellar planes. CONCLUSION: The 3D-DL algorithm can automatically extract standard fetal head planes from 3D-head volumes of comparable quality to operator-selected planes. Image quality in 3D is inferior to corresponding 2D planes, likely due to limitations with 3D-technology and acquisition technique. IMPLICATIONS FOR PRACTICE: Automated image extraction of standard planes from US-volumes could facilitate use of 3DUS in clinical practice, however image quality is dependent on the volume acquisition technique.

Optimization and initial experience of a multisection balanced steady-state free precession cine sequence for the assessment of fetal behavior in utero.

BACKGROUND AND PURPOSE: The assessment of motor function is an essential component of neurologic examinations, which imaging studies have extended to the fetus. US assessment is hampered by a limited FOV, whereas MR imaging has the potential to be an alternative. Our objectives were to optimize a cine MR imaging sequence for capturing fetal movements and to perform a pilot analysis of the relationship between the frequency of movements and uterine spatial constrictions in healthy fetuses. MATERIALS AND METHODS: Initially, a bSSFP cine sequence was selected for optimization, and various compromises were explored in all acquisition parameters to achieve an effective balance between anatomic coverage of the fetus and the temporal resolution of cine data, with the aim of maximizing both. Subsequently, cross-sectional qualitative and quantitative analyses of fetal movements were performed prospectively by using a cohort of 37 healthy fetuses (median GA, 29 weeks; range, 20-37 weeks) with the optimized cine protocol. Two smaller subgroups were selected for representative sampling of overall behavior patterns by using cine data of longer duration and for volumetric quantification of free intrauterine space. RESULTS: The optimized cine sequence, with TR/TE of 3.21/1.59 ms, coupled with parallel imaging and partial-Fourier imaging, resulted in a section-acquisition time of 0.303 seconds. Anatomic coverage was enhanced by using a combination of thick sagittal sections (30-40 mm) and multisection acquisitions to display movements in all fetal limbs, head, and trunk simultaneously. All expected motor patterns were observed throughout this gestational period, and a significant decreasing trend in overall movement frequency with age was demonstrated (r = -0.514, P = .0011). Also a significant negative correlation was found between overall movement frequency and the total intrauterine free space (r = -0.703, P = .0001). Furthermore, a significant decrease in the frequency of leg movements was shown in fetuses older then 30 weeks' GA compared with those younger than that (P = .015). CONCLUSIONS: Cine MR imaging is effective for observing fetal movements from midgestation with near full-body coverage. Also, reductions in free space with increasing GA appear to be a factor in the gradual reductions in overall levels of fetal activity as well as in restrictions in movement within specific regions of the fetal anatomy.

Assessment of brain growth in early childhood using deformation-based morphometry.

We present methods for the quantitative analysis of brain growth based on the registration of longitudinal MR image data with the use of Jacobian determinant maps to characterise neuroanatomical changes. The individual anatomies, growth maps and tissue classes are also spatially normalised in an 'average space' and aggregated to provide atlases for the population at each timepoint. The average space representation is obtained using the average intersubject transformation within each timepoint. In an exemplar study, this approach is used to assess brain development in 25 infants between 1 and 2 years, and we show consistency in growth estimates between registration and segmentation approaches.

Quantitative assessment of magnetic resonance imaging lesion load in multiple sclerosis.

Changes of lesion load on yearly conventional spin echo (CSE) T2-weighted scans of the brain from patients with multiple sclerosis, measured using computer assisted techniques, are used to monitor long term disease evolution, either natural or modified by treatment. Although lesion load measurements have several advantages over clinical measures of outcome (they provide a more objective and sensitive measure of disease evolution, which has a linear distribution and a more strict relation with the underlying pathology), the poor correlation between changes of lesion load and changes of disability is of concern when using such an approach for monitoring multiple sclerosis trials. In this review, the main sources of variation in T2 lesion load from brain MRI of patients with multiple sclerosis will be considered, along with possible strategies to, at least partially, overcome them. Also, some of the newer fully automated techniques to segment multiple sclerosis lesions, which have been validated against manual outlining, and a recently developed coregistration technique are presented. It is hoped that a more reliable and standardised approach to lesion load measurements in multiple sclerosis will lead to better correlation with clinical disease course, to a higher confidence in the results of trials, and to reduced numbers of scans needed to conduct the trials, thus improving cost efficiency and reducing discomfort of the patients.

Assessment of brain changes with registered MR before and after bone marrow transplantation for chronic myeloid leukemia.

PURPOSE: To determine the frequency and nature of changes to the brain resulting from chemotherapy, radiation therapy, and bone marrow transplantation for chronic myeloid leukemia and to compare the sensitivity of conventional and registered MR scans for detecting these changes. METHODS: In 15 patients, conventional T1-weighted, T2-weighted, and fluid-attenuated inversion recovery MR sequences, as well as T1-weighted radio frequency spoiled 3-D volume MR scans were performed before, 4 to 6 days after, and up to 339 days after transplantation (13 allografts, two autografts). A subvoxel registration program was used to match the volume images precisely so that small changes could be detected after subtraction of scans. Five healthy adult control subjects were also studied on two occasions 1 month apart. RESULTS: Studies performed 4 to 339 days after transplantation showed ventricular enlargement and cortical atrophy in all 13 patients who had allografts. The changes were evident at 4 to 6 days after transplantation and became more obvious during later follow-up examinations. Similar changes were seen in one patient with an autograft but no significant change was seen in the other patient with an autograft or in the five control subjects. Accurately registered volume scans were more sensitive than unregistered conventional scans in detecting early (9/10 versus 0/10), intermediate (12/13 versus 3/12), and late (10/10 versus 4/9) ventricular enlargement on follow-up examinations. The same applied to cortical atrophy (9/10 versus 0/10, 12/13 versus 0/12, and 10/10 versus 0/9). CONCLUSION: The specific cause and clinical significance of these changes are uncertain. Subvoxel registration of serial MR images may reveal changes that are poorly seen or not apparent on conventional scans.