A Topological Loss Function for Deep-Learning Based Image Segmentation Using Persistent Homology.

We introduce a method for training neural networks to perform image or volume segmentation in which prior knowledge about the topology of the segmented object can be explicitly provided and then incorporated into the training process. By using the differentiable properties of persistent homology, a concept used in topological data analysis, we can specify the desired topology of segmented objects in terms of their Betti numbers and then drive the proposed segmentations to contain the specified topological features. Importantly this process does not require any ground-truth labels, just prior knowledge of the topology of the structure being segmented. We demonstrate our approach in four experiments: one on MNIST image denoising and digit recognition, one on left ventricular myocardium segmentation from magnetic resonance imaging data from the UK Biobank, one on the ACDC public challenge dataset and one on placenta segmentation from 3-D ultrasound. We find that embedding explicit prior knowledge in neural network segmentation tasks is most beneficial when the segmentation task is especially challenging and that it can be used in either a semi-supervised or post-processing context to extract a useful training gradient from images without pixelwise labels.

Improved co-registration of ex-vivo and in-vivo cardiovascular magnetic resonance images using heart-specific flexible 3D printed acrylic scaffold combined with non-rigid registration.

BACKGROUND: Ex-vivo cardiovascular magnetic resonance (CMR) imaging has played an important role in the validation of in-vivo CMR characterization of pathological processes. However, comparison between in-vivo and ex-vivo imaging remains challenging due to shape changes occurring between the two states, which may be non-uniform across the diseased heart. A novel two-step process to facilitate registration between ex-vivo and in-vivo CMR was developed and evaluated in a porcine model of chronic myocardial infarction (MI). METHODS: Seven weeks after ischemia-reperfusion MI, 12 swine underwent in-vivo CMR imaging with late gadolinium enhancement followed by ex-vivo CMR 1 week later. Five animals comprised the control group, in which ex-vivo imaging was undertaken without any support in the LV cavity, 7 animals comprised the experimental group, in which a two-step registration optimization process was undertaken. The first step involved a heart specific flexible 3D printed scaffold generated from in-vivo CMR, which was used to maintain left ventricular (LV) shape during ex-vivo imaging. In the second step, a non-rigid co-registration algorithm was applied to align in-vivo and ex-vivo data. Tissue dimension changes between in-vivo and ex-vivo imaging were compared between the experimental and control group. In the experimental group, tissue compartment volumes and thickness were compared between in-vivo and ex-vivo data before and after non-rigid registration. The effectiveness of the alignment was assessed quantitatively using the DICE similarity coefficient. RESULTS: LV cavity volume changed more in the control group (ratio of cavity volume between ex-vivo and in-vivo imaging in control and experimental group 0.14 vs 0.56, p < 0.0001) and there was a significantly greater change in the short axis dimensions in the control group (ratio of short axis dimensions in control and experimental group 0.38 vs 0.79, p < 0.001). In the experimental group, prior to non-rigid co-registration the LV cavity contracted isotropically in the ex-vivo condition by less than 20% in each dimension. There was a significant proportional change in tissue thickness in the healthy myocardium (change = 29 ± 21%), but not in dense scar (change = - 2 ± 2%, p = 0.034). Following the non-rigid co-registration step of the process, the DICE similarity coefficients for the myocardium, LV cavity and scar were 0.93 (±0.02), 0.89 (±0.01) and 0.77 (±0.07) respectively and the myocardial tissue and LV cavity volumes had a ratio of 1.03 and 1.00 respectively. CONCLUSIONS: The pattern of the morphological changes seen between the in-vivo and the ex-vivo LV differs between scar and healthy myocardium. A 3D printed flexible scaffold based on the in-vivo shape of the LV cavity is an effective strategy to minimize morphological changes in the ex-vivo LV. The subsequent non-rigid registration step further improved the co-registration and local comparison between in-vivo and ex-vivo data.

Creating three dimensional models of the right ventricular outflow tract: influence of contrast, sequence, operator, and threshold.

The use of 3D printed models of the right ventricular outflow tract (RVOT) for surgical and interventional planning is growing and often requires image segmentation of cardiac magnetic resonance (CMR) images. Segmentation results may vary based on contrast, image sequence, signal threshold chosen by the operator, and manual post-processing. The purpose of this study was to determine potential biases and post-processing errors in image segmentation to enable informed decisions. Models of the RVOT and pulmonary arteries from twelve patients who had contrast enhanced CMR angiography with gadopentetate dimeglumine (GPD), gadofosveset trisodium (GFT), and a post-GFT inversion-recovery (IR) whole heart sequence were segmented, trimmed, and aligned by three operators. Geometric agreement and minimal RVOT diameters were compared between sequences and operators. To determine the contribution of threshold, interoperator variability was compared between models created by the same two operators using the same versus different thresholds. Geometric agreement by Dice between objects was high (intraoperator: 0.89-0.95; interoperator: 0.95-0.97), without differences between sequences. Minimal RVOT diameters differed on average by - 1.9 to - 1.3 mm (intraoperator) and by 0.4 to 1.4 mm (interoperator). The contribution of threshold to interoperator geometric agreement was not significant (same threshold: 0.96 ± 0.06, different threshold: 0.93 ± 0.05; p = 0.181), but minimal RVOT diameters were more variable with different versus constant thresholds (- 9.12% vs. 2.42%; p < 0.05). Thresholding does not significantly change interoperator variability for geometric agreement, but does for minimal RVOT diameter. Minimal RVOT diameters showed clinically relevant variation within and between operators.

Patient-specific 3D Printing: A Novel Technique for Complex Pediatric Renal Transplantation.

OBJECTIVE: The authors investigated a novel application of patient-specific three-dimensional (3D) printing, to enhance preoperative, multidisciplinary planning in complex, living-donor pediatric renal transplantation. SUMMARY BACKGROUND DATA: For children with end-stage kidney disease, the transplantation of adult-sized, living-donor kidneys into small recipients (<20 kg) with increasingly complex structural anomalies can be difficult. Establishing the operative feasibility in such cases demands a surgical understanding of anatomy to be derived from medical imaging. However, this is hampered by the representation of complex structures in 2D, the inherent interpretive expertise this demands, and the challenge of conveying this appreciation to others. METHODS: We report the novel use of patient-specific 3D printed models to achieve personalized management for 3 children who underwent living-donor renal transplantation. Each presented a unique surgical challenge that would otherwise prevent preoperative determination of transplantation feasibility. Patient-specific geometries were segmented from imaging data and fabricated using polyjet, 3D printing technology. Models were verified by an expert radiologist and presented during multidisciplinary discussion and surgical simulation. RESULTS: 3D printed models enhanced preoperative deliberation and surgical simulation and allowed on-table exploration of a small child to be avoided. We have critically determined specific clinical indications, technical insights, limitations, and outcomes of this approach. At latest follow-up (>16 mo) all patients remain well with functioning renal allografts. CONCLUSIONS: We report the new and safe integration of patient-specific 3D printing into complex pediatric renal transplantation. This technique enhances surgical planning and can inform operative feasibility in those cases which would otherwise be uncertain.

Evaluation of a modified Cheatham-Platinum stent for the treatment of aortic coarctation by finite element modelling.

OBJECTIVES: Stent implantation for the treatment of aortic coarctation has become a standard approach for the management of older children and adults. Criteria for optimal stent design and construction remain undefined. This study used computational modelling to compare the performance of two generations of the Cheatham-Platinum stent (NuMED, Hopkinton, NY, USA) deployed in aortic coarctation using finite element analysis. DESIGN: Three-dimensional models of both stents, reverse engineered from microCT scans, were implanted in the aortic model of one representative patient. They were virtually expanded in the vessel with a 16 mm balloon and a pressure of 2 atm. RESULTS: The conventional stent foreshortened to 96.5% of its initial length, whereas the new stent to 99.2% of its initial length. Diameters in 15 slices across the conventional stent were 11.6-15 mm (median 14.2 mm) and slightly higher across the new stent: 10.7-15.3 mm (median 14.5 mm) (p= 0.021). Apposition to the vessel wall was similar: conventional stent 31.1% and new stent 28.6% of total stent area. CONCLUSIONS: The new design Cheatham-Platinum stent showed similar deployment results compared to the conventional design. The new stent design showed slightly higher expansion, using the same delivery balloon. Patient-specific computational models can be used for virtual implantation of new aortic stents and promise to inform subsequent in vivo trials.

Three-dimensional printed models for surgical planning of complex congenital heart defects: an international multicentre study.

OBJECTIVES: To evaluate the impact of 3D printed models (3D models) on surgical planning in complex congenital heart disease (CHD). METHODS: A prospective case-crossover study involving 10 international centres and 40 patients with complex CHD (median age 3 years, range 1 month-34 years) was conducted. Magnetic resonance imaging and computed tomography were used to acquire and segment the 3D cardiovascular anatomy. Models were fabricated by fused deposition modelling of polyurethane filament, and dimensions were compared with medical images. Decisions after the evaluation of routine clinical images were compared with those after inspection of the 3D model and intraoperative findings. Subjective satisfaction questionnaire was provided. RESULTS: 3D models accurately replicate anatomy with a mean bias of -0.27 ± 0.73 mm. Ninety-six percent of the surgeons agree or strongly agree that 3D models provided better understanding of CHD morphology and improved surgical planning. 3D models changed the surgical decision in 19 of the 40 cases. Consideration of a 3D model refined the planned biventricular repair, achieving an improved surgical correction in 8 cases. In 4 cases initially considered for conservative management or univentricular palliation, inspection of the 3D model enabled successful biventricular repair. CONCLUSIONS: 3D models are accurate replicas of the cardiovascular anatomy and improve the understanding of complex CHD. 3D models did not change the surgical decision in most of the cases (21 of 40 cases, 52.5% cases). However, in 19 of the 40 selected complex cases, 3D model helped redefining the surgical approach.

Improved passive catheter tracking with positive contrast for CMR-guided cardiac catheterization using partial saturation (pSAT).

BACKGROUND: Cardiac catheterization is a common procedure in patients with congenital heart disease (CHD). Although cardiovascular magnetic resonance imaging (CMR) represents a promising alternative approach to fluoroscopy guidance, simultaneous high contrast visualization of catheter, soft tissue and the blood pool remains challenging. In this study, a novel passive tracking technique is proposed for enhanced positive contrast visualization of gadolinium-filled balloon catheters using partial saturation (pSAT) magnetization preparation. METHODS: The proposed pSAT sequence uses a single shot acquisition with balanced steady-state free precession (bSSFP) readout preceded by a partial saturation pre-pulse. This technique was initially evaluated in five healthy subjects. The pSAT sequence was compared to conventional bSSFP images acquired with (SAT) and without (Non-SAT) saturation pre-pulse. Signal-to-noise ratio (SNR) of the catheter balloon, blood and myocardium and the corresponding contrast-to-noise ratio (CNR) are reported. Subjective assessment of image suitability for CMR-guidance and ideal pSAT angle was performed by three cardiologists. The feasibility of the pSAT sequence is demonstrated in two adult patients undergoing CMR-guided cardiac catheterization. RESULTS: The proposed pSAT approach provided better catheter balloon/blood contrast and catheter balloon/myocardium contrast than conventional Non-SAT sequences. It also resulted in better blood and myocardium SNR than SAT sequences. When averaged over all volunteers, images acquired with a pSAT angle of 20° to 40° enabled simultaneous visualization of the catheter balloon and the cardiovascular anatomy (blood and myocardium) and were found suitable for CMR-guidance in >93% of cases. The pSAT sequence was successfully used in two patients undergoing CMR-guided diagnostic cardiac catheterization. CONCLUSIONS: The proposed pSAT sequence offers real-time, simultaneous, enhanced contrast visualization of the catheter balloon, soft tissues and blood. This technique provides improved passive tracking capabilities during CMR-guided catheterization in patients.

Lighten the Load: Scaffolding Visual Literacy in Biochemistry and Molecular Biology.

The development of scientific visual literacy has been identified as critical to the training of tomorrow's scientists and citizens alike. Within the context of the molecular life sciences in particular, visual representations frequently incorporate various components, such as discipline-specific graphical and diagrammatic features, varied levels of abstraction, and spatial arrangements of visual elements to convey information. Visual literacy is achieved when an individual understands the various ways in which a discipline uses these components to represent a particular way of knowing. Owing to the complex nature of visual representations, the activities through which visual literacy is developed have high cognitive load. Cognitive load can be reduced by first helping students to become fluent with the discrete components of visual representations before asking them to simultaneously integrate these components to extract the intended meaning of a representation. We present a taxonomy for characterizing one component of visual representations-the level of abstraction-as a first step in understanding the opportunities afforded students to develop fluency. Further, we demonstrate how our taxonomy can be used to analyze course assessments and spur discussions regarding the extent to which the development of visual literacy skills is supported by instruction within an undergraduate biochemistry curriculum.

Annular modes and apparent eddy feedbacks in the Southern Hemisphere.

Lagged correlation analysis is often used to infer intraseasonal dynamical effects but is known to be affected by nonstationarity. We highlight a pronounced quasi 2 year peak in the anomalous zonal wind and eddy momentum flux convergence power spectra in the Southern Hemisphere, which is prima facie evidence for nonstationarity. We then investigate the consequences of this nonstationarity for the Southern Annular Mode and for eddy momentum flux convergence. We argue that positive lagged correlations previously attributed to the existence of an eddy feedback are more plausibly attributed to nonstationary interannual variability external to any potential feedback process in the midlatitude troposphere. The findings have implications for the diagnosis of feedbacks in both models and reanalysis data as well as for understanding the mechanisms underlying variations in the zonal wind.

Use of a semi-automated cardiac segmentation tool improves reproducibility and speed of segmentation of contaminated right heart magnetic resonance angiography.

Three-dimensional printing has an increasing number of clinical applications in pediatric cardiology. Time required for dataset segmentation and conversion to stereolithography (STL) format remains a significant limitation. We investigated the impact of semi-automated cardiovascular-specific segmentation software on time and reproducibility of segmentation. Magnetic resonance angiograms (MRAs) of 19 patients undergoing intervention for right ventricular outflow lesions were segmented to demonstrate the right heart. STLs were created by two independent clinicians using semi-automated cardiovascular segmentation (SAS) and traditional manual segmentation (MS). Time was recorded and geometric STL disagreement was determined (0 % = no disagreement, 100 % = complete disagreement). MRA datasets were categorized as clean when only right heart structures were present in the MRA, or contaminated when left heart structures were also present and required removal. Eighteen (seven clean and 11 contaminated) cases were successfully segmented with both methods. Time to STL for clean datasets was faster with MS than SAS [median 209 s (IQR 192-252) vs. 296 s (272-317), p = 0.018] while contaminated datasets were faster with SAS [455 s (384-561) vs. 866 s (310-1429), p = 0.033]. Interobserver STL geometric disagreement was significantly lower using SAS than MS overall (0.70 ± 1.15 % vs. 1.31 ± 1.52 %, p = 0.030), and for the contaminated subset (0.81 ± 1.08 % vs. 1.75 ± 1.57 %, p = 0.036). Most geometric disagreement occurred at areas where left heart contamination was removed. Semi-automated segmentation was faster and more reproducible for contaminated datasets, while MS was faster but equally reproducible for clean datasets. Semi-automated segmentation methods are preferable for contaminated datasets and continued refinement of these tools should be supported.

Adult with cerebellar anaplastic pilocytic astrocytoma associated with BRAF V600E mutation and p16 loss.

Pilocytic astrocytoma (PA) is the most common pediatric tumor, with the vast majority being benign (WHO Grade I). Herein, we present a rare sporadic (not radiation- or NF1-associated) anaplastic PA arising from the cerebellum of an adult patient. The diagnosis was based on the coexistence of classic PA and more cellular foci, associated with both tumor necrosis and up to 27 mitoses per 10 high power fields. Based on these features, the tumor was felt to be equivalent in biological behavior to that of a WHO Grade III astrocytoma. Additional genetic studies revealed the presence of a BRAF V600E mutation. In comparison to the foci of classic PA, the malignant component showed increased p53 protein expression, decreased p16 protein expression, and hemizygous p16 gene deletion by FISH analysis. This case provides additional support for the concept of anaplastic transformation in PA and further elucidates the possible molecular pathways associated with malignant progression.

Identifying modulators of hERG channel activity using the PatchXpress planar patch clamp.

The authors used the PatchXpress 7000A system to measure compound activity at the hERG channel using procedures that mimicked the "gold-standard" conventional whole-cell patch clamp. A set of 70 compounds, including hERG antagonists with potencies spanning 3 orders of magnitude, were tested on hERG302-HEK cells using protocols aimed at either identifying compound activity at a single concentration or obtaining compound potency from a cumulative concentration dependence paradigm. After exposure to compounds and subsequent washout of the wells to determine reversibility of the block, blockade by a reference compound served as a quality control. Electrical parameters and voltage dependence were similar to those obtained using a conventional whole-cell patch clamp. Rank order of compound potency was also comparable to that determined by conventional methods. One exception was flunarizine, a particularly lipophilic compound. The PatchXpress accurately identified the activity of 29 moderately potent antagonists, which only weakly displace radiolabeled astemizole and are false negatives in the binding assay. Finally, no false hits were observed from a collection of relatively inactive compounds. High-quality data acquisition by PatchXpress should help accelerate secondary screening for ion channel modulators and the drug discovery process.