Deep Learning in Prostate Cancer Diagnosis Using Multiparametric Magnetic Resonance Imaging With Whole-Mount Histopathology Referenced Delineations

Background: Multiparametric magnetic resonance imaging (mpMRI) plays an important role in the diagnosis of prostate cancer (PCa) in the current clinical setting. However, the performance of mpMRI usually varies based on the experience of the radiologists at different levels; thus, the demand for MRI interpretation warrants further analysis. In this study, we developed a deep learning (DL) model to improve PCa diagnostic ability using mpMRI and whole-mount histopathology data.Methods: A total of 739 patients, including 466 with PCa and 273 without PCa, were enrolled from January 2017 to December 2019. The mpMRI (T2 weighted imaging, diffusion weighted imaging, and apparent diffusion coefficient sequences) data were randomly divided into training (n = 659) and validation datasets (n = 80). According to the whole-mount histopathology, a DL model, including independent segmentation and classification networks, was developed to extract the gland and PCa area for PCa diagnosis. The area under the curve (AUC) were used to evaluate the performance of the prostate classification networks. The proposed DL model was subsequently used in clinical practice (independent test dataset; n = 200), and the PCa detective/diagnostic performance between the DL model and different level radiologists was evaluated based on the sensitivity, specificity, precision, and accuracy.Results: The AUC of the prostate classification network was 0.871 in the validation dataset, and it reached 0.797 using the DL model in the test dataset. Furthermore, the sensitivity, specificity, precision, and accuracy of the DL model for diagnosing PCa in the test dataset were 0.710, 0.690, 0.696, and 0.700, respectively. For the junior radiologist without and with DL model assistance, these values were 0.590, 0.700, 0.663, and 0.645 versus 0.790, 0.720, 0.738, and 0.755, respectively. For the senior radiologist, the values were 0.690, 0.770, 0.750, and 0.730 vs. 0.810, 0.840, 0.835, and 0.825, respectively. The diagnosis made with DL model assistance for radiologists were significantly higher than those without assistance (P < 0.05).Conclusion: The diagnostic performance of DL model is higher than that of junior radiologists and can improve PCa diagnostic accuracy in both junior and senior radiologists.

Explainable AI for CNN-based Prostate Tumor Segmentation in Multi-parametric MRI Correlated to Whole Mount Histopathology

Automatic prostate tumor segmentation is often unable to identify the lesion even if in multi-parametric MRI data is used as input, and the segmentation output is difficult to verify due to the lack of clinically established ground truth images. In this work we use an explainable deep learning model to interpret the predictions of a convolutional neural network (CNN) for prostate tumor segmentation. The CNN uses a U-Net architecture which was trained on multi-parametric MRI data from 122 patients to automatically segment the prostate gland and prostate tumor lesions. In addition, co-registered ground truth data from whole mount histopathology images were available in 15 patients that were used as a test set during CNN testing. To be able to interpret the segmentation results of the CNN, heat maps were generated using the Gradient Weighted Class Activation Map (Grad-CAM) method. With the CNN a mean Dice Sorensen Coefficient for the prostate gland and the tumor lesions of 0.62 and 0.31 with the radiologist drawn ground truth and 0.32 with wholemount histology ground truth for tumor lesions could be achieved. Dice Sorensen Coefficient between CNN predictions and manual segmentations from MRI and histology data were not significantly different. In the prostate the Grad-CAM heat maps could differentiate between tumor and healthy prostate tissue, which indicates that the image information in the tumor was essential for the CNN segmentation.

Techniques for visualizing fibroblast-vessel interactions in the developing and adult CNS

Fibroblasts are found associated with blood vessels in various locations across the CNS: in the meninges, the choroid plexus, and in the parenchyma within perivascular spaces. CNS fibroblasts have been characterized using transcriptional profiling and a Col1a1-GFP mouse line used to identify CNS fibroblasts in vivo. However, current methods for visualizing CNS fibroblasts are lacking and, in particular, prevent adequate assessment of fibroblast-vessel interactions. Here, we describe methods for whole mount visualization of meningeal and choroid plexus fibroblasts, and optical tissue clearing methods for visualization of parenchymal vessel-associated fibroblasts. Importantly, these techniques can be combined with immunohistochemistry methods for labeling different cell types in the meninges and blood vasculature as well as EdU-based cell proliferation assays. These methods are ideal for visualization of vessel-fibroblast interactions in these CNS structures and provide significant improvement over traditional sectioning and staining methods. We expect these methods will advance studies of CNS fibroblast development and functions in homeostasis, injury, and disease.

Enhanced microvasculature formation and patterning in iPSC-derived kidney organoids cultured in physiological hypoxia

Functional kidney organoids have the potential to be used in implantable kidney grafts for patients with end-stage kidney disease, because they have been shown to self-organize from induced pluripotent stem cells into most important renal structures. To date, however, long-term kidney organoid culture has not succeeded, as nephrons lose their phenotype after approximately 25 days. Furthermore, the renal structures remain immature with diminishing endothelial networks with low connectivity and limited organoid invasion. We hypothesized that introducing long-term culture at physiological hypoxia, rather than the normally applied non-physiological, hyperoxic 21% O2, could initiate angiogenesis, lead to enhanced growth factor expression and improve the endothelial patterning. We therefore cultured the kidney organoids at 7% O2 instead of 21% O2 for up to 25 days and evaluated nephrogenesis, VEGF-A expression and vascularization. Whole mount imaging revealed a homogenous morphology of the endothelial network with enhanced sprouting and interconnectivity when the kidney organoids were cultured in hypoxia. Three-dimensional quantification confirmed that the hypoxic culture led to an increased average vessel length, likely due to the observed upregulation of proangiogenic VEGF-A189 mRNA and downregulation of the antiangiogenic protein VEGF-A165b measured in hypoxia. This research indicates the importance of optimization of oxygen availability in organoid systems and the potential of hypoxic culture conditions in improving the vascularization of organoids.

Quantitative Imaging in Whole-mount Zebrafish Embryos Traces Morphogen Gradient Maintenance and Noise Propagation in BMP Signaling

Dorsoventral (DV) embryonic patterning relies on precisely controlled interpretation of morphogen signaling. In all vertebrates, DV axis specification is informed by gradients of bone morphogenetic proteins (BMPs). We developed a 3D single-molecule mRNA quantification method in whole-mount zebrafish to quantify the inputs and outputs in this pathway. In combination with 3D computational modeling of zebrafish embryo development, data from this method revealed that sizzled (Szl), shaped by BMP and Nodal signaling, maintained a consistent inhibition level with chordin (Chd) to maintain the BMP morphogen gradient. Intriguingly, intrinsic BMP morphogen expression is highly noisy at the ventral marginal layer in the early zebrafish gastrula, where the gradient for DV patterning is established, which implies an unexpected role for noise in gradient shaping.

Using IVIM Parameters to Differentiate Prostate Cancer and Contralateral Normal Tissue through Fusion of MRI Images with Whole-Mount Pathology Specimen Images by Control Point Registration Method

The intravoxel incoherent motion (IVIM) model may enhance the clinical value of multiparametric magnetic resonance imaging (mpMRI) in the detection of prostate cancer (PCa). However, while past IVIM modeling studies have shown promise, they have also reported inconsistent results and limitations, underscoring the need to further enhance the accuracy of IVIM modeling for PCa detection. Therefore, this study utilized the control point registration toolbox function in MATLAB to fuse T2-weighted imaging (T2WI) and diffusion-weighted imaging (DWI) MRI images with whole-mount pathology specimen images in order to eliminate potential bias in IVIM calculations. Sixteen PCa patients underwent prostate MRI scans before undergoing radical prostatectomies. The image fusion method was then applied in calculating the patients’ IVIM parameters. Furthermore, MRI scans were also performed on 22 healthy young volunteers in order to evaluate the changes in IVIM parameters with aging. Among the full study cohort, the f parameter was significantly increased with age, while the D* parameter was significantly decreased. Among the PCa patients, the D and ADC parameters could differentiate PCa tissue from contralateral normal tissue, while the f and D* parameters could not. The presented image fusion method also provided improved precision when comparing regions of interest side by side. However, further studies with more standardized methods are needed to further clarify the benefits of the presented approach and the different IVIM parameters in PCa characterization.

Clomiphene Citrate and Enclomiphene Hydrochloride Exposure Is Associated With Interfrontal Suture Fusion in Zebrafish

Background: The last several decades have witnessed an increase in metopic craniosynostosis incidence. Population-based studies suggest that pharmacological exposure in utero may be responsible. This study examined effects of the fertility drug clomiphene citrate (CC) on calvarial development in an established model for craniofacial development, the zebrafish Danio rerio. Results: Zebrafish larvae were exposed to clomiphene citrate or its isomer enclomiphene for five days at key points during calvarial development. Larvae were then raised to adulthood in normal rearing water. Zebrafish were analyzed using whole-mount skeletal staining. We observed differential effects on survivability, growth and suture formation depending on the treatment. Treatments with CC or enclomiphene at 5.5 mm SL led to increased fusion of the interfrontal suture (p < .01) compared to controls. Conclusions: Exposure to fertility drugs appears to affect development of the cranial vault, specifically the interfrontal suture, in zebrafish. Further research is required to identify the signaling mechanisms at play. This work suggests that fertility drug treatment may contribute to the increased incidence of metopic craniosynostosis observed globally.

Spatiotemporal Quantification of Metastatic Tumour Cell Growth and Distribution in Lymph Nodes by Whole-Mount Tissue 3D Imaging

Lymph nodes (LNs) are a common site of metastasis in many solid cancers. Tumour cells can migrate to LNs for further metastatic colonization of distant organs, indicating poor prognosis and requiring different clinical interventions. The current histopathological diagnostic methods used for the detection of clinical lymph node metastasis (LNM) still have some limitations, such as incomplete observation. To obtain a complete picture of tumour-metastasized LNs at spatial and temporal scales, we used 3D imaging of solvent-cleared organs (uDISCO) and 3D rapid immunostaining. MC38 cells tagged with EGFP were injected into the left footpad of C57BL/6 mice. Draining lymph nodes (DLNs) obtained from these mice were cleared using uDISCO. Metastatic colorectal cancer (CRC) cells in various regions of DLNs from mice at different time points were quantified using whole-mount tissue 3D imaging. The results revealed several stages of tumour cell growth and distribution in LNs: 1) invasion of lymphatic vessels (LVs) and blood vessels (BVs); 2) dispersion outside LVs and BVs for proliferation and expansion; and 3) re-entry into BVs and efferent lymphatic vessels (ELVs) for further distant metastasis. Moreover, these data demonstrated that mouse fibroblast cells (MFCs) could not only promote the LNM of tumour cells but could also metastasize to LNs together with tumour cells, thus providing a “soil” for tumour cell colonization. In conclusion, whole-mount tissue 3D imaging and spatiotemporal analysis of LNM may together constitute an auxiliary method to improve the accuracy of clinical LNM detection in the future.

Involvement of the zebrafish trrap gene in craniofacial development

Trrap (transformation/transcription domain-associated protein) is a component shared by several histone acetyltransferase (HAT) complexes and participates in transcriptional regulation and DNA repair; however, the developmental functions of Trrap in vertebrates are not fully understood. Recently, it has been reported that human patients with genetic mutations in the TRRAP gene show various symptoms, including facial dysmorphisms, microcephaly and global developmental delay. To investigate the physiological functions of Trrap, we established trrap gene-knockout zebrafish and examined loss-of-function phenotypes in the mutants. The trrap zebrafish mutants exhibited smaller eyes and heads than the wild-type zebrafish. The size of the ventral pharyngeal arches was reduced and the mineralization of teeth was impaired in the trrap mutants. Whole-mount in situ hybridization analysis revealed that dlx3 expression was narrowly restricted in the developing ventral pharyngeal arches, while dlx2b expression was diminished in the trrap mutants. These results suggest that trrap zebrafish mutants are useful model organisms for a human disorder associated with genetic mutations in the human TRRAP gene.

Protocol for CUBIC Clearing and Whole Mount Imaging of Mouse Lung Lobes v1

This protocol is for Clear, Unobstructed Brain/Body Imaging Cocktails and Computational analysis (CUBIC) of mouse lung tissue for whole lobe imaging using Zeiss Lightsheet Imaging. All experimental procedures were performed in the American Association for Accreditation of Laboratory Animal Care (AAALAC)-certified laboratory animal facility at the University of California San Diego, following protocols approved by the institutional animal care and use committee (IACUC). The procedures should incorporate all local requirements for standards of animal experimentation.