DeepEn’s Post

Excited to share that our recent work has been published in Physics in Medicine and Biology! The paper is titled "A robust auto-contouring and data augmentation pipeline for adaptive MRI-guided radiotherapy of pancreatic cancer with a limited dataset." In this study, we explored multiple auto-contouring algorithms and enhanced their performance through a novel augmentation approach. By simulating new patients using structure-guided deformations, we significantly improved the AI model's accuracy. Notably, we compared our approach to a customized PatchGAN discriminator CycleGAN and demonstrated superior performance with our method. Our findings also highlight that this improvement could lead to dose distributions that more closely align with ground-truth contours, enabling more accurate and effective treatments. 🔗 Read more about our work here: https://lnkd.in/e5uTPpmr

Glad to share that our paper that explored the potentials of extreme learning algorithms and Deep Feed -Forward Networks for Diagnosis of Breast cancer'' has been accepted in Biomedical Engineering and Computational Biology published by Sage indexed in Clarivate Analytics

MVision AI very important and interesting study here from Alexander Valdman MD, PhD showing that #AI auto-segmentation, based on deep learning, can help reduce variability in organs at risk (OARS) definitions https://lnkd.in/eV8iHYPq

A recent study in Nature Physics on the early stages of bladder cancer shows the importance of mechanical changes in tumour tissue during growth. The study was led by the Computational Biology group of Dagmar Iber, ETH Zürich in collaboration with Lukas Bubendorf’s pathology lab at the University Hospital Basel and the D-BSSE group of Biophysics. 🎥 In this video, the team presents the findings of their research on the onset of epithelial bladder cancer. The researchers’ motivation was to understand what governs the direction in which bladder tumour grows. This growth direction can play a key role in whether it proves malignant or benign. In turn, this also determines the course of treatment and the patient’s chances of survival. Combining 3D-microscopy with computational modelling, the researchers found that the tumour type, its structure and the biomechanical properties of the bladder tissue influence the growth direction of bladder tumours, a critical factor in determining the aggressiveness of bladder cancer. The team calls for a stronger focus on tissue mechanics in future cancer research, suggesting that this approach could lead to more targeted and effective therapies for patients. Thanks to all collaborators: Roman Vetter, Kevin Yamauchi, Yifan Wang, Steve Runser, Nico Strohmeyer, Florian Meer, Marie Hussherr, Gieri Camenisch, Helge Seifert, Cyrill Rentsch, Clémentine Le Magnen, Daniel J. Müller, Lukas Bubendorf, Dagmar Iber Find full news article > https://u.ethz.ch/D5Bw3 View full-length video > https://lnkd.in/ei4-qK-t #Biomechanics #CancerResearch #BladderCancer #PersonalizedMedicine #MedicalInnovation #NaturePhysics #3DMicroscopy #ComputationalBiology #TumorResearch #HealthcareInnovation

A recent study published in Nature Physics highlights how 𝐦𝐞𝐜𝐡𝐚𝐧𝐢𝐜𝐚𝐥 𝐜𝐡𝐚𝐧𝐠𝐞𝐬 𝐢𝐧 𝐭𝐮𝐦𝐨𝐫 𝐭𝐢𝐬𝐬𝐮𝐞, such as alterations in the stiffness of cancer cells and their surroundings, play a crucial role in the early stages of bladder cancer development. This groundbreaking research, led by Franziska Lisa Lampart from Dagmar Iber's Computational Biology group at ETH Zürich, underscores the importance of early detection. At ARTIDIS, we are committed to leveraging these insights to adapt our diagnostic technology for all solid tumors, aiming to improve treatment outcomes for patients. 💡 𝐂𝐡𝐞𝐜𝐤 𝐨𝐮𝐭 𝐭𝐡𝐞 𝐫𝐞𝐬𝐞𝐚𝐫𝐜𝐡 𝐩𝐚𝐩𝐞𝐫 𝐡𝐞𝐫𝐞: https://lnkd.in/dKgUwA73

🙌👏 Pleased to share with you that we published a paper at MDPI Journal of Imaging entitled "State-of-the-Art Deep Learning Methods for Microscopic Image Segmentation: Applications to Cells, Nuclei, and Tissues" ✍️🤝 An article coordinated by Fatma Krikid and written with Hugo Rositi within a project funded by Institut national du cancer 📝📍 https://lnkd.in/eJ3esk2D #research #stateoftheart #review #openaccess #deeplearning #imagesegmentation #microscopicimaging #biomedicalengineering #biomedicalimaging

𝗔𝗻𝘁𝗶𝗯𝗼𝗱𝗶𝗲𝘀 𝗰𝗶𝘁𝗲𝗱 𝗶𝗻 𝗼𝘃𝗲𝗿 𝟯,𝟬𝟬𝟬 𝗽𝗮𝗽𝗲𝗿𝘀—𝗮𝗻𝗱 𝘁𝗵𝗲𝘆 𝗱𝗶𝗱𝗻’𝘁 𝗲𝘃𝗲𝗻 𝗯𝗶𝗻𝗱 𝘁𝗼 𝘁𝗵𝗲 𝘁𝗮𝗿𝗴𝗲𝘁 𝗽𝗿𝗼𝘁𝗲𝗶𝗻…. This shocking statistic underscores the profound impact unreliable reagents have on research reproducibility. A recent 𝗡𝗮𝘁𝘂𝗿𝗲 𝗮𝗿𝘁𝗶𝗰𝗹𝗲, "The antibodies don’t work! The race to rid labs of molecules that ruin experiments" (Nov 2024), shines a spotlight on this long-standing issue—and it’s eye-opening. The problem isn’t just frustrating; it’s slowing down science across fields like neuroscience, cancer biology, and more. Ref: https://lnkd.in/gt5-jteC

Full article here: https://lnkd.in/dDSaQz9C First principles......

As developers of minimally invasive deep tissue imaging equipment, the Day of Intravital Microscopy has always been a great source of inspiration for us. This November was the second time DeepEn'ers participated in the event, where we demonstrated the potential of #ultrathin holographic endoscopy live in front of the attendants. At the workshops, we learned a lot about recent developments in intravital microscopy for #cancer and #brain research. We also presented new imaging results from our 2024 collaborations, showing the high #spatial and #temporal resolutions achievable through a 100µm-thin fibre probe in-vivo with DeepEn's technology. For researchers, this makes examining subcellular details hidden deeply inside sensitive tissues possible, as well as observing fast processes at several kilohertz in the same imaging session. This year, the event occurred at VIB-KU Leuven Center for Cancer Biology in Belgium, which allowed us to discuss new applications with experts from across #Europe working in different research fields. Thanks for having us, and see you next year! German BioImaging - Gesellschaft für Mikroskopie und Bildanalyse e.V. Nerf - empowered by imec, KU Leuven and VIB German Center for Neurodegenerative Diseases (DZNE) VIB-KU Leuven Center For Brain & Disease Research

𝗖𝗵𝗲𝗰𝗸 𝗼𝘂𝘁 𝘁𝗵𝗲 𝗮𝗿𝘁𝗶𝗰𝗹𝗲 “𝗔 𝗖𝗲𝗹𝗹𝘂𝗹𝗮𝗿 𝗚𝗿𝗼𝘂𝗻𝗱 𝗧𝗿𝘂𝘁𝗵 𝘁𝗼 𝗗𝗲𝘃𝗲𝗹𝗼𝗽 𝗠𝗥𝗜 𝗦𝗶𝗴𝗻𝗮𝘁𝘂𝗿𝗲𝘀 𝗶𝗻 𝗚𝗹𝗶𝗼𝗺𝗮 𝗠𝗼𝗱𝗲𝗹𝘀 𝗯𝘆 𝗖𝗼𝗿𝗿𝗲𝗹𝗮𝘁𝗶𝘃𝗲 𝗟𝗶𝗴𝗵𝘁 𝗦𝗵𝗲𝗲𝘁 𝗠𝗶𝗰𝗿𝗼𝘀𝗰𝗼𝗽𝘆 𝗮𝗻𝗱 𝗔𝘁𝗹𝗮𝘀-𝗕𝗮𝘀𝗲𝗱 𝗖𝗼𝗿𝗲𝗴𝗶𝘀𝘁𝗿𝗮𝘁𝗶𝗼𝗻” 𝗯𝘆 𝗞𝗮𝘁𝗵𝗮𝗿𝗶𝗻𝗮 𝗦𝗰𝗵𝗿𝗲𝗴𝗲𝗹 𝗲𝘁 𝗮𝗹.📖 https://lnkd.in/dSkV7dh5 This group created an image analysis workflow by integrating light-sheet fluorescence microscopy and magnetic resonance imaging to examine mouse glioma models. Their approach enables a better understanding of glioma biology and provides improved preclinical imaging analysis tools. This article was featured in our January Fluorescence Microscopy Journal Club. If you would like to be notified of other featured publications once a month, you can subscribe here: https://lnkd.in/dwG8Dsr3 🔽 Workflow overview, including multiparametric MRI, tissue clearing, and light sheet microscopy brain imaging. Brain atlas regions and tumor details can be analyzed using coregistration and image segmentation. Reproduced Figure 1 from Schregel et al., 2023 (reproduced under CC-BY 4.0 license). #biology #oncology #microscopy