Multi-Task Learning for Motion Analysis and Segmentation in 3D Echocardiography

Characterizing left ventricular deformation and strain using 3D+time echocardiography provides useful insights into cardiac function and can be used to detect and localize myocardial injury. To achieve this, it is imperative to obtain accurate motion estimates of the left ventricle. In many strain analysis pipelines, this step is often accompanied by a separate segmentation step; however, recent works have shown both tasks to be highly related and can be complementary when optimized jointly. In this work, we present a multi-task learning network that can simultaneously segment the left ventricle and track its motion between multiple time frames. Two task-specific networks are trained using a composite loss function. Cross-stitch units combine the activations of these networks by learning shared representations between the tasks at different levels. We also propose a novel shape-consistency unit that encourages motion propagated segmentations to match directly predicted segmentations. Using a combined synthetic and in-vivo 3D echocardiography dataset, we demonstrate that our proposed model can achieve excellent estimates of left ventricular motion displacement and myocardial segmentation. Additionally, we observe strong correlation of our image-based strain measurements with crystal-based strain measurements as well as good correspondence with SPECT perfusion mappings. Finally, we demonstrate the clinical utility of the segmentation masks in estimating ejection fraction and sphericity indices that correspond well with benchmark measurements.

RAPID resistance to BET inhibitors is mediated by FGFR1 in glioblastoma

Bromodomain and extra-terminal domain (BET) proteins are therapeutic targets in several cancers including the most common malignant adult brain tumor glioblastoma (GBM). Multiple small molecule inhibitors of BET proteins have been utilized in preclinical and clinical studies. Unfortunately, BET inhibitors have not shown efficacy in clinical trials enrolling GBM patients. One possible reason for this may stem from resistance mechanisms that arise after prolonged treatment within a clinical setting. However, the mechanisms and timeframe of resistance to BET inhibitors in GBM is not known. To identify the temporal order of resistance mechanisms in GBM we performed quantitative proteomics using multiplex-inhibitor bead mass spectrometry and demonstrated that intrinsic resistance to BET inhibitors in GBM treatment occurs rapidly within hours and involves the fibroblast growth factor receptor 1 (FGFR1) protein. Additionally, small molecule inhibition of BET proteins and FGFR1 simultaneously induces synergy in reducing GBM tumor growth in vitro and in vivo. Further, FGFR1 knockdown synergizes with BET inhibitor mediated reduction of GBM cell proliferation. Collectively, our studies suggest that co-targeting BET and FGFR1 may dampen resistance mechanisms to yield a clinical response in GBM.

DuDoCFNet: Dual-Domain Coarse-to-Fine Progressive Network for Simultaneous Denoising, Limited-View Reconstruction, and Attenuation Correction of Cardiac SPECT

Single-Photon Emission Computed Tomography (SPECT) is widely applied for the diagnosis of coronary artery diseases. Low-dose (LD) SPECT aims to minimize radiation exposure but leads to increased image noise. Limited-view (LV) SPECT, such as the latest GE MyoSPECT ES system, enables accelerated scanning and reduces hardware expenses but degrades reconstruction accuracy. Additionally, Computed Tomography (CT) is commonly used to derive attenuation maps (μ-maps) for attenuation correction (AC) of cardiac SPECT, but it will introduce additional radiation exposure and SPECT-CT misalignments. Although various methods have been developed to solely focus on LD denoising, LV reconstruction, or CT-free AC in SPECT, the solution for simultaneously addressing these tasks remains challenging and under-explored. Furthermore, it is essential to explore the potential of fusing cross-domain and cross-modality information across these interrelated tasks to further enhance the accuracy of each task. Thus, we propose a Dual-Domain Coarse-to-Fine Progressive Network (DuDoCFNet), a multi-task learning method for simultaneous LD denoising, LV reconstruction, and CT-free μ-map generation of cardiac SPECT. Paired dual-domain networks in DuDoCFNet are cascaded using a multi-layer fusion mechanism for cross-domain and cross-modality feature fusion. Two-stage progressive learning strategies are applied in both projection and image domains to achieve coarse-to-fine estimations of SPECT projections and CT-derived μ-maps. Our experiments demonstrate DuDoCFNet's superior accuracy in estimating projections, generating μ-maps, and AC reconstructions compared to existing single- or multi-task learning methods, under various iterations and LD levels. The source code of this work is available at https://github.com/XiongchaoChen/DuDoCFNet-MultiTask.

Unique vulnerability of RAC1-mutant melanoma to combined inhibition of CDK9 and immune checkpoints

RAC1P29S is the third most prevalent hotspot mutation in sun-exposed melanoma. RAC1 alterations in cancer are correlated with poor prognosis, resistance to standard chemotherapy, and insensitivity to targeted inhibitors. Although RAC1P29S mutations in melanoma and RAC1 alterations in several other cancers are increasingly evident, the RAC1-driven biological mechanisms contributing to tumorigenesis remain unclear. Lack of rigorous signaling analysis has prevented identification of alternative therapeutic targets for RAC1P29S-harboring melanomas. To investigate the RAC1P29S-driven effect on downstream molecular signaling pathways, we generated an inducible RAC1P29S expression melanocytic cell line and performed RNA-sequencing (RNA-seq) coupled with multiplexed kinase inhibitor beads and mass spectrometry (MIBs/MS) to establish enriched pathways from the genomic to proteomic level. Our proteogenomic analysis identified CDK9 as a potential new and specific target in RAC1P29S-mutant melanoma cells. In vitro, CDK9 inhibition impeded the proliferation of in RAC1P29S-mutant melanoma cells and increased surface expression of PD-L1 and MHC Class I proteins. In vivo, combining CDK9 inhibition with anti-PD-1 immune checkpoint blockade significantly inhibited tumor growth only in melanomas that expressed the RAC1P29S mutation. Collectively, these results establish CDK9 as a novel target in RAC1-driven melanoma that can further sensitize the tumor to anti-PD-1 immunotherapy.

KRAS mutation-selective requirement for ACSS2 in colorectal adenoma formation

Oncogenic KRAS mutations are prevalent in colorectal cancer (CRC) and are associated with poor prognosis and resistance to therapy. There is a substantial diversity of KRAS mutant alleles observed in CRC. Emerging clinical and experimental analysis of common KRAS mutations suggest that each mutation differently influences the clinical properties of a disease and response to therapy. Although there is some evidence to suggest biological differences between mutant KRAS alleles, these are yet to be fully elucidated. One approach to study allelic variation involves the use of isogenic cell lines that express different endogenous Kras mutants. Here, we generated Kras isogenic Apc-/- mouse colon epithelial cell lines using CRISPR-driven genome editing by altering the original G12D Kras allele to G12V, G12R, or G13D. We utilized these cell lines to perform transcriptomic and proteomic analysis to compare different signaling properties between these mutants. Both screens indicate significant differences in pathways relating to cholesterol and lipid regulation that we validated with targeted metabolomic measurements and isotope tracing. We found that these processes are upregulated in G12V lines through increased expression of nuclear SREBP1 and higher activation of mTORC1. G12V cells showed higher expression of ACSS2 and ACSS2 inhibition sensitized G12V cells to MEK inhibition. Finally, we found that ACSS2 plays a crucial role early in the development of G12V mutant tumors, in contrast to G12D mutant tumors. These observations highlight differences between KRAS mutant cell lines in their signaling properties. Further exploration of these pathways may prove to be valuable for understanding how specific KRAS mutants function, and identification of novel therapeutic opportunities in CRC.

Grand Challenges at the Interface of Engineering and Medicine

Over the past two decades Biomedical Engineering has emerged as a major discipline that bridges societal needs of human health care with the development of novel technologies. Every medical institution is now equipped at varying degrees of sophistication with the ability to monitor human health in both non-invasive and invasive modes. The multiple scales at which human physiology can be interrogated provide a profound perspective on health and disease. We are at the nexus of creating "avatars" (herein defined as an extension of "digital twins") of human patho/physiology to serve as paradigms for interrogation and potential intervention. Motivated by the emergence of these new capabilities, the IEEE Engineering in Medicine and Biology Society, the Departments of Biomedical Engineering at Johns Hopkins University and Bioengineering at University of California at San Diego sponsored an interdisciplinary workshop to define the grand challenges that face biomedical engineering and the mechanisms to address these challenges. The Workshop identified five grand challenges with cross-cutting themes and provided a roadmap for new technologies, identified new training needs, and defined the types of interdisciplinary teams needed for addressing these challenges. The themes presented in this paper include: 1) accumedicine through creation of avatars of cells, tissues, organs and whole human; 2) development of smart and responsive devices for human function augmentation; 3) exocortical technologies to understand brain function and treat neuropathologies; 4) the development of approaches to harness the human immune system for health and wellness; and 5) new strategies to engineer genomes and cells.

Patient-Specific Heart Geometry Modeling for Solid Biomechanics Using Deep Learning

Automated volumetric meshing of patient-specific heart geometry can help expedite various biomechanics studies, such as post-intervention stress estimation. Prior meshing techniques often neglect important modeling characteristics for successful downstream analyses, especially for thin structures like the valve leaflets. In this work, we present DeepCarve (Deep Cardiac Volumetric Mesh): a novel deformation-based deep learning method that automatically generates patient-specific volumetric meshes with high spatial accuracy and element quality. The main novelty in our method is the use of minimally sufficient surface mesh labels for precise spatial accuracy and the simultaneous optimization of isotropic and anisotropic deformation energies for volumetric mesh quality. Mesh generation takes only 0.13 seconds/scan during inference, and each mesh can be directly used for finite element analyses without any manual post-processing. Calcification meshes can also be subsequently incorporated for increased simulation accuracy. Numerous stent deployment simulations validate the viability of our approach for large-batch analyses. Our code is available at https://github.com/danpak94/Deep-Cardiac-Volumetric-Mesh.

Kinome profiling identifies MARK3 and STK10 as potential therapeutic targets in uveal melanoma

Most uveal melanoma cases harbor activating mutations in either GNAQ or GNA11. Despite activation of the mitogen-activated protein kinase (MAPK) signaling pathway downstream of Gαq/11, there are no effective targeted kinase therapies for metastatic uveal melanoma. The human genome encodes numerous understudied kinases, also called the "dark kinome". Identifying additional kinases regulated by Gαq/11 may uncover novel therapeutic targets for uveal melanoma. In this study, we treated GNAQ-mutant uveal melanoma cell lines with a Gαq/11 inhibitor, YM-254890, and conducted a kinase signaling proteomic screen using multiplexed-kinase inhibitors followed by mass spectrometry. We observed downregulated expression and/or activity of 22 kinases. A custom siRNA screen targeting these kinases demonstrated that knockdown of microtubule affinity regulating kinase 3 (MARK3) and serine/threonine kinase 10 (STK10) significantly reduced uveal melanoma cell growth and decreased expression of cell cycle proteins. Additionally, knockdown of MARK3 but not STK10 decreased ERK1/2 phosphorylation. Analysis of RNA-sequencing and proteomic data showed that Gαq signaling regulates STK10 expression and MARK3 activity. Our findings suggest an involvement of STK10 and MARK3 in the Gαq/11 oncogenic pathway and prompt further investigation into the specific roles and targeting potential of these kinases in uveal melanoma.

FedFTN: Personalized federated learning with deep feature transformation network for multi-institutional low-count PET denoising

Low-count PET is an efficient way to reduce radiation exposure and acquisition time, but the reconstructed images often suffer from low signal-to-noise ratio (SNR), thus affecting diagnosis and other downstream tasks. Recent advances in deep learning have shown great potential in improving low-count PET image quality, but acquiring a large, centralized, and diverse dataset from multiple institutions for training a robust model is difficult due to privacy and security concerns of patient data. Moreover, low-count PET data at different institutions may have different data distribution, thus requiring personalized models. While previous federated learning (FL) algorithms enable multi-institution collaborative training without the need of aggregating local data, addressing the large domain shift in the application of multi-institutional low-count PET denoising remains a challenge and is still highly under-explored. In this work, we propose FedFTN, a personalized federated learning strategy that addresses these challenges. FedFTN uses a local deep feature transformation network (FTN) to modulate the feature outputs of a globally shared denoising network, enabling personalized low-count PET denoising for each institution. During the federated learning process, only the denoising network's weights are communicated and aggregated, while the FTN remains at the local institutions for feature transformation. We evaluated our method using a large-scale dataset of multi-institutional low-count PET imaging data from three medical centers located across three continents, and showed that FedFTN provides high-quality low-count PET images, outperforming previous baseline FL reconstruction methods across all low-count levels at all three institutions.

A Multiclass Radiomics Method-Based WHO Severity Scale for Improving COVID-19 Patient Assessment and Disease Characterization From CT Scans

OBJECTIVES

The aim of this study was to evaluate the severity of COVID-19 patients' disease by comparing a multiclass lung lesion model to a single-class lung lesion model and radiologists' assessments in chest computed tomography scans.

MATERIALS AND METHODS

The proposed method, AssessNet-19, was developed in 2 stages in this retrospective study. Four COVID-19-induced tissue lesions were manually segmented to train a 2D-U-Net network for a multiclass segmentation task followed by extensive extraction of radiomic features from the lung lesions. LASSO regression was used to reduce the feature set, and the XGBoost algorithm was trained to classify disease severity based on the World Health Organization Clinical Progression Scale. The model was evaluated using 2 multicenter cohorts: a development cohort of 145 COVID-19-positive patients from 3 centers to train and test the severity prediction model using manually segmented lung lesions. In addition, an evaluation set of 90 COVID-19-positive patients was collected from 2 centers to evaluate AssessNet-19 in a fully automated fashion.

RESULTS

AssessNet-19 achieved an F1-score of 0.76 ± 0.02 for severity classification in the evaluation set, which was superior to the 3 expert thoracic radiologists (F1 = 0.63 ± 0.02) and the single-class lesion segmentation model (F1 = 0.64 ± 0.02). In addition, AssessNet-19 automated multiclass lesion segmentation obtained a mean Dice score of 0.70 for ground-glass opacity, 0.68 for consolidation, 0.65 for pleural effusion, and 0.30 for band-like structures compared with ground truth. Moreover, it achieved a high agreement with radiologists for quantifying disease extent with Cohen κ of 0.94, 0.92, and 0.95.

CONCLUSIONS

A novel artificial intelligence multiclass radiomics model including 4 lung lesions to assess disease severity based on the World Health Organization Clinical Progression Scale more accurately determines the severity of COVID-19 patients than a single-class model and radiologists' assessment.

Copy Number Variation Informs fMRI-based Prediction of Autism Spectrum Disorder

The multifactorial etiology of autism spectrum disorder (ASD) suggests that its study would benefit greatly from multimodal approaches that combine data from widely varying platforms, e.g., neuroimaging, genetics, and clinical characterization. Prior neuroimaging-genetic analyses often apply naive feature concatenation approaches in data-driven work or use the findings from one modality to guide posthoc analysis of another, missing the opportunity to analyze the paired multimodal data in a truly unified approach. In this paper, we develop a more integrative model for combining genetic, demographic, and neuroimaging data. Inspired by the influence of genotype on phenotype, we propose using an attention-based approach where the genetic data guides attention to neuroimaging features of importance for model prediction. The genetic data is derived from copy number variation parameters, while the neuroimaging data is from functional magnetic resonance imaging. We evaluate the proposed approach on ASD classification and severity prediction tasks, using a sex-balanced dataset of 228 ASD and typically developing subjects in a 10-fold cross-validation framework. We demonstrate that our attention-based model combining genetic information, demographic data, and functional magnetic resonance imaging results in superior prediction performance compared to other multimodal approaches.

Cross-Attention for Improved Motion Correction in Brain PET

Head movement during long scan sessions degrades the quality of reconstruction in positron emission tomography (PET) and introduces artifacts, which limits clinical diagnosis and treatment. Recent deep learning-based motion correction work utilized raw PET list-mode data and hardware motion tracking (HMT) to learn head motion in a supervised manner. However, motion prediction results were not robust to testing subjects outside the training data domain. In this paper, we integrate a cross-attention mechanism into the supervised deep learning network to improve motion correction across test subjects. Specifically, cross-attention learns the spatial correspondence between the reference images and moving images to explicitly focus the model on the most correlative inherent information - the head region the motion correction. We validate our approach on brain PET data from two different scanners: HRRT without time of flight (ToF) and mCT with ToF. Compared with traditional and deep learning benchmarks, our network improved the performance of motion correction by 58% and 26% in translation and rotation, respectively, in multi-subject testing in HRRT studies. In mCT studies, our approach improved performance by 66% and 64% for translation and rotation, respectively. Our results demonstrate that cross-attention has the potential to improve the quality of brain PET image reconstruction without the dependence on HMT. All code will be released on GitHub: https://github.com/OnofreyLab/dl_hmc_attention_mlcn2023.

Preserved Edge Convolutional Neural Network for Sensitivity Enhancement of Deuterium Metabolic Imaging (DMI)

PURPOSE

Common to most MRSI techniques, the spatial resolution and the minimal scan duration of Deuterium Metabolic Imaging (DMI) are limited by the achievable SNR. This work presents a deep learning method for sensitivity enhancement of DMI.

METHODS

A convolutional neural network (CNN) was designed to estimate the 2H-labeled metabolite concentrations from low SNR and distorted DMI FIDs. The CNN was trained with synthetic data that represent a range of SNR levels typically encountered in vivo. The estimation precision was further improved by fine-tuning the CNN with MRI-based edge-preserving regularization for each DMI dataset. The proposed processing method, PReserved Edge ConvolutIonal neural network for Sensitivity Enhanced DMI (PRECISE-DMI), was applied to simulation studies and in vivo experiments to evaluate the anticipated improvements in SNR and investigate the potential for inaccuracies.

RESULTS

PRECISE-DMI visually improved the metabolic maps of low SNR datasets, and quantitatively provided higher precision than the standard Fourier reconstruction. Processing of DMI data acquired in rat brain tumor models resulted in more precise determination of 2H-labeled lactate and glutamate + glutamine levels, at increased spatial resolution (from >8 to 2 $\mu$L) or shortened scan time (from 32 to 4 min) compared to standard acquisitions. However, rigorous SD-bias analyses showed that overuse of the edge-preserving regularization can compromise the accuracy of the results.

CONCLUSION

PRECISE-DMI allows a flexible trade-off between enhancing the sensitivity of DMI and minimizing the inaccuracies. With typical settings, the DMI sensitivity can be improved by 3-fold while retaining the capability to detect local signal variations.

Specific pathway abundances in the neonatal calf faecal microbiome are associated with susceptibility to Cryptosporidium parvum infection: a metagenomic analysis

BACKGROUND

Cryptosporidium parvum is the main cause of calf scour worldwide. With limited therapeutic options and research compared to other Apicomplexa, it is important to understand the parasites' biology and interactions with the host and microbiome in order to develop novel strategies against this infection. The age-dependent nature of symptomatic cryptosporidiosis suggests a link to the undeveloped immune response, the immature intestinal epithelium, and its associated microbiota. This led us to hypothesise that specific features of the early life microbiome could predict calf susceptibility to C. parvum infection.

RESULTS

In this study, a single faecal swab sample was collected from each calf within the first week of life in a cohort of 346 animals. All 346 calves were subsequently monitored for clinical signs of cryptosporidiosis, and calves that developed diarrhoea were tested for Rotavirus, Coronavirus, E. coli F5 (K99) and C. parvum by lateral flow test (LFT). A retrospective case-control approach was taken whereby a subset of healthy calves (Control group; n = 33) and calves that went on to develop clinical signs of infectious diarrhoea and test positive for C. parvum infection via LFT (Cryptosporidium-positive group; n = 32) were selected from this cohort, five of which were excluded due to low DNA quality. A metagenomic analysis was conducted on the faecal microbiomes of the control group (n = 30) and the Cryptosporidium-positive group (n = 30) prior to infection, to determine features predictive of cryptosporidiosis. Taxonomic analysis showed no significant differences in alpha diversity, beta diversity, and taxa relative abundance between controls and Cryptosporidium-positive groups. Analysis of functional potential showed pathways related to isoprenoid precursor, haem and purine biosynthesis were significantly higher in abundance in calves that later tested positive for C. parvum (q ≤ 0.25). These pathways are either absent or streamlined in the C. parvum parasites. Though the de novo production of isoprenoid precursors, haem and purines are absent, C. parvum has been shown to encode enzymes that catalyse the downstream reactions of these pathway metabolites, indicating that C. parvum may scavenge those products from an external source.

CONCLUSIONS

The host has previously been put forward as the source of essential metabolites, but our study suggests that C. parvum may also be able to harness specific metabolic pathways of the microbiota in order to survive and replicate. This finding is important as components of these microbial pathways could be exploited as potential therapeutic targets for the prevention or mitigation of cryptosporidiosis in bovine neonates.

Multiaxial pressure-strain analysis of regional myocardial work in the setting of graded coronary stenoses and dobutamine stress

We present a detailed analysis of regional myocardial blood flow and work to better understand the effects of coronary stenoses and low-dose dobutamine stress. Our analysis is based on a unique open-chest model in anesthetized canines that features invasive hemodynamic monitoring, microsphere-based blood flow analysis, and an extensive three-dimensional (3-D) sonomicrometer array that provides multiaxial deformational assessments in the ischemic, border, and remote vascular territories. We use this model to construct regional pressure-strain loops for each territory and quantify the loop subcomponent areas that reflect myocardial work contributing to the ejection of blood and wasted work that does not. We demonstrate that reductions in coronary blood flow markedly alter the shapes and temporal relationships of pressure-strain loops, as well as the magnitudes of their total and subcomponent areas. Specifically, we show that moderate stenoses in the mid-left anterior descending coronary artery decrease regional midventricle myocardial work indices and substantially increase indices of wasted work. In the midventricle, these effects are most pronounced along the radial and longitudinal axes, with more modest effects along the circumferential axis. We further demonstrate that low-dose dobutamine can help to restore or even improve function, but often at the cost of increased wasted work. This detailed, multiaxial analysis provides unique insight into the physiology and mechanics of the heart in the presence of ischemia and low-dose dobutamine, with potential implications in many areas, including the detection and characterization of ischemic heart disease and the use of inotropic support for low cardiac output.NEW & NOTEWORTHY Our unique experimental model assesses cardiac pressure-strain relationships along multiple axes in multiple regions. We demonstrate that moderate coronary stenoses decrease regional myocardial work and increase wasted work and that low-dose dobutamine can help to restore myocardial function, but often with further increases in wasted work. Our findings highlight the significant directional variation of cardiac mechanics and demonstrate potential advantages of pressure-strain analyses over traditional, purely deformational measures, especially in characterizing physiological changes related to dobutamine.

Improvement in cardiac function and regional LV strain following intramyocardial injection of a theranostic hydrogel early postmyocardial infarction in a porcine model

Myocardial infarction (MI) is often complicated by left ventricular (LV) remodeling and heart failure. We evaluated the feasibility of a multimodality imaging approach to guide delivery of an imageable hydrogel and assessed LV functional changes with therapy. Yorkshire pigs underwent surgical occlusions of branches of the left anterior descending and/or circumflex artery to create an anterolateral MI. We evaluated the hemodynamic and mechanical effects of intramyocardial delivery of an imageable hydrogel in the central infarct area (Hydrogel group, n = 8) and a Control group (n = 5) early post-MI. LV and aortic pressure and ECG were measured and contrast cineCT angiography was performed at baseline, 60 min post-MI, and 90 min post-hydrogel delivery. LV hemodynamic indices, pressure-volume measures, and normalized regional and global strains were measured and compared. Both Control and Hydrogel groups demonstrated a decline in heart rate, LV pressure, stroke volume, ejection fraction, and pressure-volume loop area, and an increase in myocardial performance (Tei) index and supply/demand (S/D) ratio. After hydrogel delivery, Tei index and S/D ratio were reduced to baseline levels, diastolic and systolic functional indices either stabilized or improved, and radial strain and circumferential strain increased significantly in the MI regions (ENrr: +52.7%, ENcc: +44.1%). However, the Control group demonstrated a progressive decline in all functional indices to levels significantly below those of Hydrogel group. Thus, acute intramyocardial delivery of a novel imageable hydrogel to MI region resulted in rapid stabilization or improvement in LV hemodynamics and function.NEW & NOTEWORTHY Our study demonstrates that contrast cineCT imaging can be used to evaluate the acute effects of intramyocardial delivery of a therapeutic hydrogel to the central MI region early post MI, which resulted in a rapid stabilization of LV hemodynamics and improvement in regional and global LV function.

DuSFE: Dual-Channel Squeeze-Fusion-Excitation co-attention for cross-modality registration of cardiac SPECT and CT

Myocardial perfusion imaging (MPI) using single-photon emission computed tomography (SPECT) is widely applied for the diagnosis of cardiovascular diseases. Attenuation maps (μ-maps) derived from computed tomography (CT) are utilized for attenuation correction (AC) to improve the diagnostic accuracy of cardiac SPECT. However, in clinical practice, SPECT and CT scans are acquired sequentially, potentially inducing misregistration between the two images and further producing AC artifacts. Conventional intensity-based registration methods show poor performance in the cross-modality registration of SPECT and CT-derived μ-maps since the two imaging modalities might present totally different intensity patterns. Deep learning has shown great potential in medical imaging registration. However, existing deep learning strategies for medical image registration encoded the input images by simply concatenating the feature maps of different convolutional layers, which might not fully extract or fuse the input information. In addition, deep-learning-based cross-modality registration of cardiac SPECT and CT-derived μ-maps has not been investigated before. In this paper, we propose a novel Dual-Channel Squeeze-Fusion-Excitation (DuSFE) co-attention module for the cross-modality rigid registration of cardiac SPECT and CT-derived μ-maps. DuSFE is designed based on the co-attention mechanism of two cross-connected input data streams. The channel-wise or spatial features of SPECT and μ-maps are jointly encoded, fused, and recalibrated in the DuSFE module. DuSFE can be flexibly embedded at multiple convolutional layers to enable gradual feature fusion in different spatial dimensions. Our studies using clinical patient MPI studies demonstrated that the DuSFE-embedded neural network generated significantly lower registration errors and more accurate AC SPECT images than existing methods. We also showed that the DuSFE-embedded network did not over-correct or degrade the registration performance of motion-free cases. The source code of this work is available at https://github.com/XiongchaoChen/DuSFE_CrossRegistration.

Unique vulnerability of RAC1-mutant melanoma to combined inhibition of CDK9 and immune checkpoints

RAC1P29S is the third most prevalent hotspot mutation in sun-exposed melanoma. RAC1 alterations in cancer are correlated with poor prognosis, resistance to standard chemotherapy, and insensitivity to targeted inhibitors. Although RAC1P29S mutations in melanoma and RAC1 alterations in several other cancers are increasingly evident, the RAC1-driven biological mechanisms contributing to tumorigenesis remain unclear. Lack of rigorous signaling analysis has prevented identification of alternative therapeutic targets for RAC1P29S-harboring melanomas. To investigate the RAC1P29S-driven effect on downstream molecular signaling pathways, we generated an inducible RAC1P29S expression melanocytic cell line and performed RNA-sequencing (RNA-seq) coupled with multiplexed kinase inhibitor beads and mass spectrometry (MIBs/MS) to establish enriched pathways from the genomic to proteomic level. Our proteogenomic analysis identified CDK9 as a potential new and specific target in RAC1P29S-mutant melanoma cells. In vitro, CDK9 inhibition impeded the proliferation of in RAC1P29S-mutant melanoma cells and increased surface expression of PD-L1 and MHC Class I proteins. In vivo, combining CDK9 inhibition with anti-PD-1 immune checkpoint blockade significantly inhibited tumor growth only in melanomas that expressed the RAC1P29S mutation. Collectively, these results establish CDK9 as a novel target in RAC1-driven melanoma that can further sensitize the tumor to anti-PD-1 immunotherapy.

Bootstrapping Semi-supervised Medical Image Segmentation with Anatomical-Aware Contrastive Distillation

Contrastive learning has shown great promise over annotation scarcity problems in the context of medical image segmentation. Existing approaches typically assume a balanced class distribution for both labeled and unlabeled medical images. However, medical image data in reality is commonly imbalanced (i.e., multi-class label imbalance), which naturally yields blurry contours and usually incorrectly labels rare objects. Moreover, it remains unclear whether all negative samples are equally negative. In this work, we present ACTION, an Anatomical-aware ConTrastive dIstillatiON framework, for semi-supervised medical image segmentation. Specifically, we first develop an iterative contrastive distillation algorithm by softly labeling the negatives rather than binary supervision between positive and negative pairs. We also capture more semantically similar features from the randomly chosen negative set compared to the positives to enforce the diversity of the sampled data. Second, we raise a more important question: Can we really handle imbalanced samples to yield better performance? Hence, the key innovation in ACTION is to learn global semantic relationship across the entire dataset and local anatomical features among the neighbouring pixels with minimal additional memory footprint. During the training, we introduce anatomical contrast by actively sampling a sparse set of hard negative pixels, which can generate smoother segmentation boundaries and more accurate predictions. Extensive experiments across two benchmark datasets and different unlabeled settings show that ACTION significantly outperforms the current state-of-the-art semi-supervised methods.

Abstract 2621: Characterizing MEK1/2 degradation in K-ras mutant cells

Activating KRAS mutations frequently occur in several cancers including the majority of pancreatic ductal adenocarcinoma (96%), ~50% of colorectal cancers, and ~30% of lung adenocarcinoma. It has been shown that mutant K-ras preferentially signals through RAF-MEK-ERK signaling and that blockade of this pathway is essential for tumor regression. However, targeted MEK inhibitors (MEKi) have shown limited clinical benefit in the treatment of K-ras mutant cancers due to resistance caused by the feedback activation of the upstream kinase CRAF. Recently, PROTACs (Proteolysis-Targeting Chimeras) targeting MEK1 and MEK2 were developed, providing a new therapeutic strategy to degrade MEK1/2 in K-ras mutant cancer cells. Due to the fact that feedback-induced CRAF causes MEKi resistance in K-ras mutant cells, we postulated that degrading MEK1/2 could bypass CRAF-mediated resistance. Indeed, treatment of MEKi-sensitive or resistant K-ras mutant cell lines with MEK1/2 degraders rapidly reduced MEK1/2 protein levels and downstream ERK signaling causing growth inhibition and apoptosis. Proteomics characterization of K-ras mutant cell lines following MEK1/2 degrader or inhibitor treatment revealed MEK1/2 degradation exhibited distinct signaling pathway responses compared to inhibition. Together, our studies showed MEK1/2 degraders represent a promising avenue for the treatment K-ras mutant cancers that can by-pass resistance mechanisms observed with traditional MEK1/2 inhibitors.

Citation Format: Carlos Herrera-Montavez, Alison M. Kurimchak, Jian Jin, James S. Duncan. Characterizing MEK1/2 degradation in K-ras mutant cells [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 2621.

RAPID BRAIN MENINGES SURFACE RECONSTRUCTION WITH LAYER TOPOLOGY GUARANTEE

The meninges, located between the skull and brain, are composed of three membrane layers: the pia, the arachnoid, and the dura. Reconstruction of these layers can aid in studying volume differences between patients with neurodegenerative diseases and normal aging subjects. In this work, we use convolutional neural networks (CNNs) to reconstruct surfaces representing meningeal layer boundaries from magnetic resonance (MR) images. We first use the CNNs to predict the signed distance functions (SDFs) representing these surfaces while preserving their anatomical ordering. The marching cubes algorithm is then used to generate continuous surface representations; both the subarachnoid space (SAS) and the intracranial volume (ICV) are computed from these surfaces. The proposed method is compared to a state-of-the-art deformable model-based reconstruction method, and we show that our method can reconstruct smoother and more accurate surfaces using less computation time. Finally, we conduct experiments with volumetric analysis on both subjects with multiple sclerosis and healthy controls. For healthy and MS subjects, ICVs and SAS volumes are found to be significantly correlated to sex (p<0.01) and age (p ≤ 0.03) changes, respectively.

FAM122A ensures cell cycle interphase progression and checkpoint control as a SLiM-dependent substrate-competitive inhibitor to the B55⍺/PP2A phosphatase

The Ser/Thr protein phosphatase 2A (PP2A) is a highly conserved collection of heterotrimeric holoenzymes responsible for the dephosphorylation of many regulated phosphoproteins. Substrate recognition and the integration of regulatory cues are mediated by B regulatory subunits that are complexed to the catalytic subunit (C) by a scaffold protein (A). PP2A/B55 substrate recruitment was thought to be mediated by charge-charge interactions between the surface of B55α and its substrates. Challenging this view, we recently discovered a conserved SLiM [ RK ]- V -x-x-[ VI ]- R in a range of proteins, including substrates such as the retinoblastoma-related protein p107 and TAU (Fowle et al. eLife 2021;10:e63181). Here we report the identification of this SLiM in FAM122A, an inhibitor of B55α/PP2A. This conserved SLiM is necessary for FAM122A binding to B55α in vitro and in cells. Computational structure prediction with AlphaFold2 predicts an interaction consistent with the mutational and biochemical data and supports a mechanism whereby FAM122A uses the 'SLiM' in the form of a short α-helix to dock to the B55α top groove. In this model, FAM122A spatially constrains substrate access by occluding the catalytic subunit with a second α-helix immediately adjacent to helix 1. Consistently, FAM122A functions as a competitive inhibitor as it prevents binding of substrates in in vitro competition assays and the dephosphorylation of CDK substrates by B55α/PP2A in cell lysates. Ablation of FAM122A in human cell lines reduces the rate of proliferation, progression through cell cycle transitions and abrogates G1/S and intra-S phase cell cycle checkpoints. FAM122A-KO in HEK293 cells results in attenuation of CHK1 and CHK2 activation in response to replication stress. Overall, these data strongly suggest that FAM122A is a 'SLiM'-dependent, substrate-competitive inhibitor of B55α/PP2A that suppresses multiple functions of B55α in the DNA damage response and in timely progression through the cell cycle interphase.

Federated Transfer Learning for Low-dose PET Denoising: A Pilot Study with Simulated Heterogeneous Data

Positron emission tomography (PET) with a reduced injection dose, i.e., low-dose PET, is an efficient way to reduce radiation dose. However, low-dose PET reconstruction suffers from a low signal-to-noise ratio (SNR), affecting diagnosis and other PET-related applications. Recently, deep learning-based PET denoising methods have demonstrated superior performance in generating high-quality reconstruction. However, these methods require a large amount of representative data for training, which can be difficult to collect and share due to medical data privacy regulations. Moreover, low-dose PET data at different institutions may use different low-dose protocols, leading to non-identical data distribution. While previous federated learning (FL) algorithms enable multi-institution collaborative training without the need of aggregating local data, it is challenging for previous methods to address the large domain shift caused by different low-dose PET settings, and the application of FL to PET is still under-explored. In this work, we propose a federated transfer learning (FTL) framework for low-dose PET denoising using heterogeneous low-dose data. Our experimental results on simulated multi-institutional data demonstrate that our method can efficiently utilize heterogeneous low-dose data without compromising data privacy for achieving superior low-dose PET denoising performance for different institutions with different low-dose settings, as compared to previous FL methods.

Co-attention spatial transformer network for unsupervised motion tracking and cardiac strain analysis in 3D echocardiography

Myocardial ischemia/infarction causes wall-motion abnormalities in the left ventricle. Therefore, reliable motion estimation and strain analysis using 3D+time echocardiography for localization and characterization of myocardial injury is valuable for early detection and targeted interventions. Previous unsupervised cardiac motion tracking methods rely on heavily-weighted regularization functions to smooth out the noisy displacement fields in echocardiography. In this work, we present a Co-Attention Spatial Transformer Network (STN) for improved motion tracking and strain analysis in 3D echocardiography. Co-Attention STN aims to extract inter-frame dependent features between frames to improve the motion tracking in otherwise noisy 3D echocardiography images. We also propose a novel temporal constraint to further regularize the motion field to produce smooth and realistic cardiac displacement paths over time without prior assumptions on cardiac motion. Our experimental results on both synthetic and in vivo 3D echocardiography datasets demonstrate that our Co-Attention STN provides superior performance compared to existing methods. Strain analysis from Co-Attention STNs also correspond well with the matched SPECT perfusion maps, demonstrating the clinical utility for using 3D echocardiography for infarct localization.

Class-Aware Adversarial Transformers for Medical Image Segmentation

Transformers have made remarkable progress towards modeling long-range dependencies within the medical image analysis domain. However, current transformer-based models suffer from several disadvantages: (1) existing methods fail to capture the important features of the images due to the naive tokenization scheme; (2) the models suffer from information loss because they only consider single-scale feature representations; and (3) the segmentation label maps generated by the models are not accurate enough without considering rich semantic contexts and anatomical textures. In this work, we present CASTformer, a novel type of adversarial transformers, for 2D medical image segmentation. First, we take advantage of the pyramid structure to construct multi-scale representations and handle multi-scale variations. We then design a novel class-aware transformer module to better learn the discriminative regions of objects with semantic structures. Lastly, we utilize an adversarial training strategy that boosts segmentation accuracy and correspondingly allows a transformer-based discriminator to capture high-level semantically correlated contents and low-level anatomical features. Our experiments demonstrate that CASTformer dramatically outperforms previous state-of-the-art transformer-based approaches on three benchmarks, obtaining 2.54%-5.88% absolute improvements in Dice over previous models. Further qualitative experiments provide a more detailed picture of the model's inner workings, shed light on the challenges in improved transparency, and demonstrate that transfer learning can greatly improve performance and reduce the size of medical image datasets in training, making CASTformer a strong starting point for downstream medical image analysis tasks.

DuDoUFNet: Dual-Domain Under-to-Fully-Complete Progressive Restoration Network for Simultaneous Metal Artifact Reduction and Low-Dose CT Reconstruction

To reduce the potential risk of radiation to the patient, low-dose computed tomography (LDCT) has been widely adopted in clinical practice for reconstructing cross-sectional images using sinograms with reduced x-ray flux. The LDCT image quality is often degraded by different levels of noise depending on the low-dose protocols. The image quality will be further degraded when the patient has metallic implants, where the image suffers from additional streak artifacts along with further amplified noise levels, thus affecting the medical diagnosis and other CT-related applications. Previous studies mainly focused either on denoising LDCT without considering metallic implants or full-dose CT metal artifact reduction (MAR). Directly applying previous LDCT or MAR approaches to the issue of simultaneous metal artifact reduction and low-dose CT (MARLD) may yield sub-optimal reconstruction results. In this work, we develop a dual-domain under-to-fully-complete progressive restoration network, called DuDoUFNet, for MARLD. Our DuDoUFNet aims to reconstruct images with substantially reduced noise and artifact by progressive sinogram to image domain restoration with a two-stage progressive restoration network design. Our experimental results demonstrate that our method can provide high-quality reconstruction, superior to previous LDCT and MAR methods under various low-dose and metal settings.

NetrinG1+ cancer-associated fibroblasts generate unique extracellular vesicles that support the survival of pancreatic cancer cells under nutritional stress

It is projected that in 5 years, pancreatic cancer will become the second deadliest cancer in the United States. A unique aspect of pancreatic ductal adenocarcinoma (PDAC) is its stroma; rich in cancer-associated fibroblasts (CAFs) and a dense CAF-generated extracellular matrix (ECM). These pathogenic stroma CAF/ECM units cause the collapse of local blood vessels rendering the tumor microenvironment nutrient-poor. PDAC cells are able to survive this state of nutrient stress via support from CAF-secreted material, which includes small extracellular vesicles (sEVs). The tumor-supportive CAFs possess a distinct phenotypic profile, compared to normal-like fibroblasts, expressing NetrinG1 (NetG1) at the plasma membrane, and active Integrin α₅β₁ localized to the multivesicular bodies; traits indicative of poor patient survival. We herein report that NetG1⁺ CAFs secrete sEVs that stimulate Akt-mediated survival in nutrient-deprived PDAC cells, protecting them from undergoing apoptosis. Further, we show that NetG1 expression in CAFs is required for the pro-survival properties of sEVs. Additionally, we report that the above-mentioned CAF markers are secreted in distinct subpopulations of EVs; with NetG1 being enriched in exomeres, and Integrin α₅β₁ being enriched in exosomes. Finally, we found that NetG1 and Integrin α₅β₁ were detected in sEVs collected from plasma of PDAC patients, while their levels were significantly lower in plasma-derived sEVs of sex/age-matched healthy donors. The discovery of these tumor-supporting CAF-EVs elucidates novel avenues in tumor-stroma interactions and pathogenic stroma detection.

SimCVD: Simple Contrastive Voxel-Wise Representation Distillation for Semi-Supervised Medical Image Segmentation

Automated segmentation in medical image analysis is a challenging task that requires a large amount of manually labeled data. However, most existing learning-based approaches usually suffer from limited manually annotated medical data, which poses a major practical problem for accurate and robust medical image segmentation. In addition, most existing semi-supervised approaches are usually not robust compared with the supervised counterparts, and also lack explicit modeling of geometric structure and semantic information, both of which limit the segmentation accuracy. In this work, we present SimCVD, a simple contrastive distillation framework that significantly advances state-of-the-art voxel-wise representation learning. We first describe an unsupervised training strategy, which takes two views of an input volume and predicts their signed distance maps of object boundaries in a contrastive objective, with only two independent dropout as mask. This simple approach works surprisingly well, performing on the same level as previous fully supervised methods with much less labeled data. We hypothesize that dropout can be viewed as a minimal form of data augmentation and makes the network robust to representation collapse. Then, we propose to perform structural distillation by distilling pair-wise similarities. We evaluate SimCVD on two popular datasets: the Left Atrial Segmentation Challenge (LA) and the NIH pancreas CT dataset. The results on the LA dataset demonstrate that, in two types of labeled ratios (i.e., 20% and 10%), SimCVD achieves an average Dice score of 90.85% and 89.03% respectively, a 0.91% and 2.22% improvement compared to previous best results. Our method can be trained in an end-to-end fashion, showing the promise of utilizing SimCVD as a general framework for downstream tasks, such as medical image synthesis, enhancement, and registration.

Momentum Contrastive Voxel-wise Representation Learning for Semi-supervised Volumetric Medical Image Segmentation

Contrastive learning (CL) aims to learn useful representation without relying on expert annotations in the context of medical image segmentation. Existing approaches mainly contrast a single positive vector (i.e., an augmentation of the same image) against a set of negatives within the entire remainder of the batch by simply mapping all input features into the same constant vector. Despite the impressive empirical performance, those methods have the following shortcomings: (1) it remains a formidable challenge to prevent the collapsing problems to trivial solutions; and (2) we argue that not all voxels within the same image are equally positive since there exist the dissimilar anatomical structures with the same image. In this work, we present a novel Contrastive Voxel-wise Representation Learning (CVRL) method to effectively learn low-level and high-level features by capturing 3D spatial context and rich anatomical information along both the feature and the batch dimensions. Specifically, we first introduce a novel CL strategy to ensure feature diversity promotion among the 3D representation dimensions. We train the framework through bi-level contrastive optimization (i.e., low-level and high-level) on 3D images. Experiments on two benchmark datasets and different labeled settings demonstrate the superiority of our proposed framework. More importantly, we also prove that our method inherits the benefit of hardness-aware property from the standard CL approaches. Codes will be available soon.

The drug efflux pump MDR1 promotes intrinsic and acquired resistance to PROTACs in cancer cells

Proteolysis-targeting chimeras (PROTACs) are a promising new class of drugs that selectively degrade cellular proteins of interest. PROTACs that target oncogene products are avidly being explored for cancer therapies, and several are currently in clinical trials. Drug resistance is a substantial challenge in clinical oncology, and resistance to PROTACs has been reported in several cancer cell models. Here, using proteomic analysis, we found intrinsic and acquired resistance mechanisms to PROTACs in cancer cell lines mediated by greater abundance or production of the drug efflux pump MDR1. PROTAC-resistant cells were resensitized to PROTACs by genetic ablation of ABCB1 (which encodes MDR1) or by coadministration of MDR1 inhibitors. In MDR1-overexpressing colorectal cancer cells, degraders targeting either the kinases MEK1/2 or the oncogenic mutant GTPase KRAS^G12C synergized with the dual epidermal growth factor receptor (EGFR/ErbB)/MDR1 inhibitor lapatinib. Moreover, compared with single-agent therapies, combining MEK1/2 degraders with lapatinib improved growth inhibition of MDR1-overexpressing KRAS-mutant colorectal cancer xenografts in mice. Together, our findings suggest that concurrent blockade of MDR1 will likely be required with PROTACs to achieve durable protein degradation and therapeutic response in cancer.

Inactivation of p21-Activated Kinase 2 (Pak2) Inhibits the Development of Nf2-Deficient Tumors by Restricting Downstream Hedgehog and Wnt Signaling

Because loss of the NF2 tumor suppressor gene results in p21-activated kinase (Pak) activation, PAK inhibitors hold promise for the treatment of NF2-deficient tumors. To test this possibility, we asked if loss of Pak2, a highly expressed group I PAK member, affects the development of malignant mesothelioma in Nf2;Cdkn2a-deficient (NC) mice and the growth properties of NC mesothelioma cells in culture. In vivo, deletion of Pak2 resulted in a markedly decreased incidence and delayed onset of both pleural and peritoneal malignant mesotheliomas in NC mice. In vitro, Pak2 deletion decreased malignant mesothelioma cell viability, migration, clonogenicity, and spheroid formation. RNA-sequencing analysis demonstrated downregulated expression of Hedgehog and Wnt pathway genes in NC;Pak2-/- mesothelioma cells versus NC;Pak2+/+ mesothelioma cells. Targeting of the Hedgehog signaling component Gli1 or its target gene Myc inhibited cell viability and spheroid formation in NC;P+/+ mesothelioma cells. Kinome profiling uncovered kinase changes indicative of EMT in NC;Pak2-/- mesothelioma cells, suggesting that Pak2-deficient malignant mesotheliomas can adapt by reprogramming their kinome in the absence of Pak activity. The identification of such compensatory pathways offers opportunities for rational combination therapies to circumvent resistance to anti-PAK drugs.

IMPLICATIONS

We provide evidence supporting a role for PAK inhibitors in treating NF2-deficient tumors. NF2-deficient tumors lacking Pak2 eventually adapt by kinome reprogramming, presenting opportunities for combination therapies to bypass anti-PAK drug resistance.

Using Adversarial Images to Assess the Robustness of Deep Learning Models Trained on Diagnostic Images in Oncology

Deep learning (DL) models have rapidly become a popular and cost-effective tool for image classification within oncology. A major limitation of DL models is their vulnerability to adversarial images, manipulated input images designed to cause misclassifications by DL models. The purpose of the study is to investigate the robustness of DL models trained on diagnostic images using adversarial images and explore the utility of an iterative adversarial training approach to improve the robustness of DL models against adversarial images.

Exploring vulnerabilities of deep learning algorithms to adversarial images across oncologic imaging modalities.

Simultaneous Segmentation and Motion Estimation of Left Ventricular Myocardium in 3D Echocardiography Using Multi-task Learning

Motion estimation and segmentation are both critical steps in identifying and assessing myocardial dysfunction, but are traditionally treated as unique tasks and solved as separate steps. However, many motion estimation techniques rely on accurate segmentations. It has been demonstrated in the computer vision and medical image analysis literature that both these tasks may be mutually beneficial when solved simultaneously. In this work, we propose a multi-task learning network that can concurrently predict volumetric segmentations of the left ventricle and estimate motion between 3D echocardiographic image pairs. The model exploits complementary latent features between the two tasks using a shared feature encoder with task-specific decoding branches. Anatomically inspired constraints are incorporated to enforce realistic motion patterns. We evaluate our proposed model on an in vivo 3D echocardiographic canine dataset. Results suggest that coupling these two tasks in a learning framework performs favorably when compared against single task learning and other alternative methods.

DuDoDR-Net: Dual-domain data consistent recurrent network for simultaneous sparse view and metal artifact reduction in computed tomography

Sparse-view computed tomography (SVCT) aims to reconstruct a cross-sectional image using a reduced number of x-ray projections. While SVCT can efficiently reduce the radiation dose, the reconstruction suffers from severe streak artifacts, and the artifacts are further amplified with the presence of metallic implants, which could adversely impact the medical diagnosis and other downstream applications. Previous methods have extensively explored either SVCT reconstruction without metallic implants, or full-view CT metal artifact reduction (MAR). The issue of simultaneous sparse-view and metal artifact reduction (SVMAR) remains under-explored, and it is infeasible to directly apply previous SVCT and MAR methods to SVMAR which may yield non-ideal reconstruction quality. In this work, we propose a dual-domain data consistent recurrent network, called DuDoDR-Net, for SVMAR. Our DuDoDR-Net aims to reconstruct an artifact-free image by recurrent image domain and sinogram domain restorations. To ensure the metal-free part of acquired projection data is preserved, we also develop the image data consistent layer (iDCL) and sinogram data consistent layer (sDCL) that are interleaved in our recurrent framework. Our experimental results demonstrate that our DuDoDR-Net is able to produce superior artifact-reduced results while preserving the anatomical structures, that outperforming previous SVCT and SVMAR methods, under different sparse-view acquisition settings.

Late-life terminal seizure freedom in drug-resistant epilepsy: "Burned-out epilepsy"

The course of established epilepsy in late life is not fully known. One key question is whether the resolution of an epileptic diathesis is a natural outcome in some people with long-standing epilepsy. We investigated this with a view to generating a hypothesis. We retrospectively explored whether terminal seizure-freedom occurs in older people with previous drug-resistant epilepsy at the Chalfont Centre for Epilepsy over twenty years. Of the 226 people followed for a median period of 52 years, 39 (17%) achieved late-life terminal seizure-freedom of at least two years before death, which occurred at a median age of 68 years with a median duration of 7 years. Multivariate analysis suggests that a high initial seizure frequency was a negative predictor (p < 0.0005). Our findings indicate that the 'natural' course of long-standing epilepsy in some people is one of terminal seizure freedom. We also consider the concept of "remission" in epilepsy, its definition challenges, and the evolving terminology used to describe the state of seizure freedom. The intersection of ageing and seizure freedom is an essential avenue of future investigation, especially in light of current demographic trends. Gaining mechanistic insights into this phenomenon may help broaden our understanding of the neurobiology of epilepsy and potentially provide targets for therapeutic intervention.

High-Resolution Magnetic Resonance Spectroscopic Imaging using a Multi-Encoder Attention U-Net with Structural and Adversarial Loss

Common to most medical imaging techniques, the spatial resolution of Magnetic Resonance Spectroscopic Imaging (MRSI) is ultimately limited by the achievable SNR. This work presents a deep learning method for 1H-MRSI spatial resolution enhancement, based on the observation that multi-parametric MRI images provide relevant spatial priors for MRSI enhancement. A Multi-encoder Attention U-Net (MAU-Net) architecture was constructed to process a MRSI metabolic map and three different MRI modalities through separate encoding paths. Spatial attention modules were incorporated to automatically learn spatial weights that highlight salient features for each MRI modality. MAU-Net was trained based on in vivo brain imaging data from patients with high-grade gliomas, using a combined loss function consisting of pixel, structural and adversarial loss. Experimental results showed that the proposed method is able to reconstruct high-quality metabolic maps with a high-resolution of 64×64 from a low-resolution of 16 × 16, with better performance compared to several baseline methods.

MDPET: A Unified Motion Correction and Denoising Adversarial Network for Low-Dose Gated PET

In positron emission tomography (PET), gating is commonly utilized to reduce respiratory motion blurring and to facilitate motion correction methods. In application where low-dose gated PET is useful, reducing injection dose causes increased noise levels in gated images that could corrupt motion estimation and subsequent corrections, leading to inferior image quality. To address these issues, we propose MDPET, a unified motion correction and denoising adversarial network for generating motion-compensated low-noise images from low-dose gated PET data. Specifically, we proposed a Temporal Siamese Pyramid Network (TSP-Net) with basic units made up of 1.) Siamese Pyramid Network (SP-Net), and 2.) a recurrent layer for motion estimation among the gates. The denoising network is unified with our motion estimation network to simultaneously correct the motion and predict a motion-compensated denoised PET reconstruction. The experimental results on human data demonstrated that our MDPET can generate accurate motion estimation directly from low-dose gated images and produce high-quality motion-compensated low-noise reconstructions. Comparative studies with previous methods also show that our MDPET is able to generate superior motion estimation and denoising performance. Our code is available at https://github.com/bbbbbbzhou/MDPET.

Learning-Based Regularization for Cardiac Strain Analysis via Domain Adaptation

Reliable motion estimation and strain analysis using 3D+ time echocardiography (4DE) for localization and characterization of myocardial injury is valuable for early detection and targeted interventions. However, motion estimation is difficult due to the low-SNR that stems from the inherent image properties of 4DE, and intelligent regularization is critical for producing reliable motion estimates. In this work, we incorporated the notion of domain adaptation into a supervised neural network regularization framework. We first propose a semi-supervised Multi-Layered Perceptron (MLP) network with biomechanical constraints for learning a latent representation that is shown to have more physiologically plausible displacements. We extended this framework to include a supervised loss term on synthetic data and showed the effects of biomechanical constraints on the network's ability for domain adaptation. We validated the semi-supervised regularization method on in vivo data with implanted sonomicrometers. Finally, we showed the ability of our semi-supervised learning regularization approach to identify infarct regions using estimated regional strain maps with good agreement to manually traced infarct regions from postmortem excised hearts.

Limited View Tomographic Reconstruction Using a Cascaded Residual Dense Spatial-Channel Attention Network With Projection Data Fidelity Layer

Limited view tomographic reconstruction aims to reconstruct a tomographic image from a limited number of projection views arising from sparse view or limited angle acquisitions that reduce radiation dose or shorten scanning time. However, such a reconstruction suffers from severe artifacts due to the incompleteness of sinogram. To derive quality reconstruction, previous methods use UNet-like neural architectures to directly predict the full view reconstruction from limited view data; but these methods leave the deep network architecture issue largely intact and cannot guarantee the consistency between the sinogram of the reconstructed image and the acquired sinogram, leading to a non-ideal reconstruction. In this work, we propose a cascaded residual dense spatial-channel attention network consisting of residual dense spatial-channel attention networks and projection data fidelity layers. We evaluate our methods on two datasets. Our experimental results on AAPM Low Dose CT Grand Challenge datasets demonstrate that our algorithm achieves a consistent and substantial improvement over the existing neural network methods on both limited angle reconstruction and sparse view reconstruction. In addition, our experimental results on Deep Lesion datasets demonstrate that our method is able to generate high-quality reconstruction for 8 major lesion types.

Anatomy-guided multimodal registration by learning segmentation without ground truth: Application to intraprocedural CBCT/MR liver segmentation and registration

Multimodal image registration has many applications in diagnostic medical imaging and image-guided interventions, such as Transcatheter Arterial Chemoembolization (TACE) of liver cancer guided by intraprocedural CBCT and pre-operative MR. The ability to register peri-procedurally acquired diagnostic images into the intraprocedural environment can potentially improve the intra-procedural tumor targeting, which will significantly improve therapeutic outcomes. However, the intra-procedural CBCT often suffers from suboptimal image quality due to lack of signal calibration for Hounsfield unit, limited FOV, and motion/metal artifacts. These non-ideal conditions make standard intensity-based multimodal registration methods infeasible to generate correct transformation across modalities. While registration based on anatomic structures, such as segmentation or landmarks, provides an efficient alternative, such anatomic structure information is not always available. One can train a deep learning-based anatomy extractor, but it requires large-scale manual annotations on specific modalities, which are often extremely time-consuming to obtain and require expert radiological readers. To tackle these issues, we leverage annotated datasets already existing in a source modality and propose an anatomy-preserving domain adaptation to segmentation network (APA2Seg-Net) for learning segmentation without target modality ground truth. The segmenters are then integrated into our anatomy-guided multimodal registration based on the robust point matching machine. Our experimental results on in-house TACE patient data demonstrated that our APA2Seg-Net can generate robust CBCT and MR liver segmentation, and the anatomy-guided registration framework with these segmenters can provide high-quality multimodal registrations.

SHAPE-REGULARIZED UNSUPERVISED LEFT VENTRICULAR MOTION NETWORK WITH SEGMENTATION CAPABILITY IN 3D+TIME ECHOCARDIOGRAPHY

Accurate motion estimation and segmentation of the left ventricle from medical images are important tasks for quantitative evaluation of cardiovascular health. Echocardiography offers a cost-efficient and non-invasive modality for examining the heart, but provides additional challenges for automated analyses due to the low signal-to-noise ratio inherent in ultrasound imaging. In this work, we propose a shape regularized convolutional neural network for estimating dense displacement fields between sequential 3D B-mode echocardiography images with the capability of also predicting left ventricular segmentation masks. Manually traced segmentations are used as a guide to assist in the unsupervised estimation of displacement between a source and a target image while also serving as labels to train the network to additionally predict segmentations. To enforce realistic cardiac motion patterns, a flow incompressibility term is also incorporated to penalize divergence. Our proposed network is evaluated on an in vivo canine 3D+t B-mode echocardiographic dataset. It is shown that the shape regularizer improves the motion estimation performance of the network and our overall model performs favorably against competing methods.

A review of deep learning in medical imaging: Imaging traits, technology trends, case studies with progress highlights, and future promises

Since its renaissance, deep learning has been widely used in various medical imaging tasks and has achieved remarkable success in many medical imaging applications, thereby propelling us into the so-called artificial intelligence (AI) era. It is known that the success of AI is mostly attributed to the availability of big data with annotations for a single task and the advances in high performance computing. However, medical imaging presents unique challenges that confront deep learning approaches. In this survey paper, we first present traits of medical imaging, highlight both clinical needs and technical challenges in medical imaging, and describe how emerging trends in deep learning are addressing these issues. We cover the topics of network architecture, sparse and noisy labels, federating learning, interpretability, uncertainty quantification, etc. Then, we present several case studies that are commonly found in clinical practice, including digital pathology and chest, brain, cardiovascular, and abdominal imaging. Rather than presenting an exhaustive literature survey, we instead describe some prominent research highlights related to these case study applications. We conclude with a discussion and presentation of promising future directions.

Computer Assisted Planning for Curved Laser Interstitial Thermal Therapy

Laser interstitial thermal therapy (LiTT) is a minimally invasive alternative to conventional open surgery for drug-resistant focal mesial temporal lobe epilepsy (MTLE). Recent studies suggest that higher seizure freedom rates are correlated with maximal ablation of the mesial hippocampal head, whilst sparing of the parahippocampal gyrus (PHG) may reduce neuropsychological sequelae. Current commercially available laser catheters are inserted following manually planned straight-line trajectories, which cannot conform to curved brain structures, such as the hippocampus, without causing collateral damage or requiring multiple insertions. The clinical feasibility and potential of curved LiTT trajectories through steerable needles has yet to be investigated. This is the focus of our work. We propose a GPU-accelerated computer-assisted planning (CAP) algorithm for steerable needle insertions that generates optimized curved 3D trajectories with maximal ablation of the amygdalohippocampal complex and minimal collateral damage to nearby structures, while accounting for a variable ablation diameter ( 5-15mm). Simulated trajectories and ablations were performed on 5 patients with mesial temporal sclerosis (MTS), which were identified from a prospectively managed database. The algorithm generated obstacle-free paths with significantly greater target area ablation coverage and lower PHG ablation variance compared to straight line trajectories. The presented CAP algorithm returns increased ablation of the amygdalohippocampal complex, with lower patient risk scores compared to straight-line trajectories. This is the first clinical application of preoperative planning for steerable needle based LiTT. This study suggests that steerable needles have the potential to improve LiTT procedure efficacy whilst improving the safety and should thus be investigated further.

Identification of Wee1 as a target in combination with avapritinib for gastrointestinal stromal tumor treatment

Management of gastrointestinal stromal tumors (GISTs) has been revolutionized by the identification of activating mutations in KIT and PDGFRA and clinical application of RTK inhibitors in advanced disease. Stratification of GISTs into molecularly defined subsets provides insight into clinical behavior and response to approved targeted therapies. Although these RTK inhibitors are effective in most GISTs, resistance remains a significant clinical problem. Development of effective treatment strategies for refractory GISTs requires identification of novel targets to provide additional therapeutic options. Global kinome profiling has the potential to identify critical signaling networks and reveal protein kinases essential in GISTs. Using multiplexed inhibitor beads and mass spectrometry, we explored the majority of the kinome in GIST specimens from the 3 most common molecular subtypes (KIT mutant, PDGFRA mutant, and succinate dehydrogenase deficient) to identify kinase targets. Kinome profiling with loss-of-function assays identified an important role for G₂/M tyrosine kinase, Wee1, in GIST cell survival. In vitro and in vivo studies revealed significant efficacy of MK-1775 (Wee1 inhibitor) in combination with avapritinib in KIT mutant and PDGFRA mutant GIST cell lines as well as notable efficacy of MK-1775 as a monotherapy in the engineered PDGFRA mutant line. These studies provide strong preclinical justification for the use of MK-1775 in GIST.

Deep learning-assisted differentiation of pathologically proven atypical and typical hepatocellular carcinoma (HCC) versus non-HCC on contrast-enhanced MRI of the liver

OBJECTIVES

To train a deep learning model to differentiate between pathologically proven hepatocellular carcinoma (HCC) and non-HCC lesions including lesions with atypical imaging features on MRI.

METHODS

This IRB-approved retrospective study included 118 patients with 150 lesions (93 (62%) HCC and 57 (38%) non-HCC) pathologically confirmed through biopsies (n = 72), resections (n = 29), liver transplants (n = 46), and autopsies (n = 3). Forty-seven percent of HCC lesions showed atypical imaging features (not meeting Liver Imaging Reporting and Data System [LI-RADS] criteria for definitive HCC/LR5). A 3D convolutional neural network (CNN) was trained on 140 lesions and tested for its ability to classify the 10 remaining lesions (5 HCC/5 non-HCC). Performance of the model was averaged over 150 runs with random sub-sampling to provide class-balanced test sets. A lesion grading system was developed to demonstrate the similarity between atypical HCC and non-HCC lesions prone to misclassification by the CNN.

RESULTS

The CNN demonstrated an overall accuracy of 87.3%. Sensitivities/specificities for HCC and non-HCC lesions were 92.7%/82.0% and 82.0%/92.7%, respectively. The area under the receiver operating curve was 0.912. CNN's performance was correlated with the lesion grading system, becoming less accurate the more atypical imaging features the lesions showed.

CONCLUSION

This study provides proof-of-concept for CNN-based classification of both typical- and atypical-appearing HCC lesions on multi-phasic MRI, utilizing pathologically confirmed lesions as "ground truth."

KEY POINTS

• A CNN trained on atypical appearing pathologically proven HCC lesions not meeting LI-RADS criteria for definitive HCC (LR5) can correctly differentiate HCC lesions from other liver malignancies, potentially expanding the role of image-based diagnosis in primary liver cancer with atypical features. • The trained CNN demonstrated an overall accuracy of 87.3% and a computational time of < 3 ms which paves the way for clinical application as a decision support instrument.

Multi-frame Attention Network for Left Ventricle Segmentation in 3D Echocardiography

Echocardiography is one of the main imaging modalities used to assess the cardiovascular health of patients. Among the many analyses performed on echocardiography, segmentation of left ventricle is crucial to quantify the clinical measurements like ejection fraction. However, segmentation of left ventricle in 3D echocardiography remains a challenging and tedious task. In this paper, we propose a multi-frame attention network to improve the performance of segmentation of left ventricle in 3D echocardiography. The multi-frame attention mechanism allows highly correlated spatiotemporal features in a sequence of images that come after a target image to be used to augment the performance of segmentation. Experimental results shown on 51 in vivo porcine 3D+time echocardiography images show that utilizing correlated spatiotemporal features significantly improves the performance of left ventricle segmentation when compared to other standard deep learning-based medical image segmentation models.

Phosphorylation of RIAM by src promotes integrin activation by unmasking the PH domain of RIAM

Integrin activation controls cell adhesion, migration, invasion, and extracellular matrix remodeling. RIAM (RAP1-GTP-interacting adaptor molecule) is recruited by activated RAP1 to the plasma membrane (PM) to mediate integrin activation via an inside-out signaling pathway. This process requires the association of the pleckstrin homology (PH) domain of RIAM with the membrane PIP2. We identify a conserved intermolecular interface that masks the PIP2-binding site in the PH domains of RIAM. Our data indicate that phosphorylation of RIAM by Src family kinases disrupts this PH-mediated interface, unmasks the membrane PIP2-binding site, and promotes integrin activation. We further demonstrate that this process requires phosphorylation of Tyr267 and Tyr427 in the RIAM PH domain by Src. Our data reveal an unorthodox regulatory mechanism of small GTPase effector proteins by phosphorylation-dependent PM association of the PH domain and provide new insights into the link between Src kinases and integrin signaling.

Kinome Profiling of Primary Endometrial Tumors Using Multiplexed Inhibitor Beads and Mass Spectrometry Identifies SRPK1 as Candidate Therapeutic Target

Endometrial carcinoma (EC) is the most common gynecologic malignancy in the United States, with limited effective targeted therapies. Endometrial tumors exhibit frequent alterations in protein kinases, yet only a small fraction of the kinome has been therapeutically explored. To identify kinase therapeutic avenues for EC, we profiled the kinome of endometrial tumors and normal endometrial tissues using Multiplexed Inhibitor Beads and Mass Spectrometry (MIB-MS). Our proteomics analysis identified a network of kinases overexpressed in tumors, including Serine/Arginine-Rich Splicing Factor Kinase 1 (SRPK1). Immunohistochemical (IHC) analysis of endometrial tumors confirmed MIB-MS findings and showed SRPK1 protein levels were highly expressed in endometrioid and uterine serous cancer (USC) histological subtypes. Moreover, querying large-scale genomics studies of EC tumors revealed high expression of SRPK1 correlated with poor survival. Loss-of-function studies targeting SRPK1 in an established USC cell line demonstrated SRPK1 was integral for RNA splicing, as well as cell cycle progression and survival under nutrient deficient conditions. Profiling of USC cells identified a compensatory response to SRPK1 inhibition that involved EGFR and the up-regulation of IGF1R and downstream AKT signaling. Co-targeting SRPK1 and EGFR or IGF1R synergistically enhanced growth inhibition in serous and endometrioid cell lines, representing a promising combination therapy for EC.

A Semi-supervised Joint Network for Simultaneous Left Ventricular Motion Tracking and Segmentation in 4D Echocardiography

This work presents a novel deep learning method to combine segmentation and motion tracking in 4D echocardiography. The network iteratively trains a motion branch and a segmentation branch. The motion branch is initially trained entirely unsupervised and learns to roughly map the displacements between a source and a target frame. The estimated displacement maps are then used to generate pseudo-ground truth labels to train the segmentation branch. The labels predicted by the trained segmentation branch are fed back into the motion branch and act as landmarks to help retrain the branch to produce smoother displacement estimations. These smoothed out displacements are then used to obtain smoother pseudo-labels to retrain the segmentation branch. Additionally, a biomechanically-inspired incompressibility constraint is implemented in order to encourage more realistic cardiac motion. The proposed method is evaluated against other approaches using synthetic and in-vivo canine studies. Both the segmentation and motion tracking results of our model perform favorably against competing methods.