"Brain Cancer Detection with Explainable CNNs: A New Method"

🤖 Artificial Intelligence in Breast Cancer Detection — From Promise to Clinical Reality A comprehensive review by Ioannis Sechopoulos and colleagues examine the evolution of AI-assisted breast cancer screening, tracing its journey from early computer-aided detection (CADe/CADx) systems to modern deep learning–based algorithms that now rival expert radiologists. Unlike early systems that struggled with high false-positive rates, deep learning and convolutional neural networks have revolutionized image interpretation in digital mammography and tomosynthesis, achieving human-level accuracy in retrospective datasets. While prospective real-world studies remain essential, evidence strongly supports AI’s potential to enhance screening efficiency, reduce workload, and improve early cancer detection — signaling a paradigm shift in breast imaging. https://lnkd.in/eyXNyDen #BreastCancer #AI #Mammography #DigitalHealth #Tomosynthesis #Radiology #CancerScreening #OncologyFrontier #MediaMedic

Thrilled to share that my book chapter “Breast Cancer Diagnosis Based on Deep Convolutional Neural Networks Using Image Clustering” has been published in Computing and Emerging Technologies: First International Conference, ICCET 2023 (Springer CCIS, Part I). This research leverages Deep Learning and CNN-based clustering for enhanced breast cancer diagnosis, contributing to the growing field of AI in healthcare. A big thank you to my mentors, colleagues, and collaborators at The University of Faisalabad (TUF) for their continuous support. 🔗 Read here: Springer Link #ArtificialIntelligence #MachineLearning #DeepLearning #MedicalImaging #BreastCancerDiagnosis #HealthcareAI #Research #Springer #AcademicPublishing #ComputerVision #AIForGood

New method combines multiphoton microscopy with machine learning and neural networks to quickly identify pancreatic cancer https://lnkd.in/epucAM-w Image: Label-free multiphoton microscopy and deep learning can be used in combination to classify pancreatic neuroendocrine neoplasms with high accuracy—a significant step toward automated digital pathology for such tumors. (Image Credit: N. Daigle|University of Arizona) #microscopy #multiphotonmicroscopy #machinelearning #cancerdiagnostics #digitalpathology #biophotonics #wileyanalyticalscience

🚀 𝐓𝐰𝐨 𝐬𝐮𝐜𝐜𝐞𝐬𝐬𝐟𝐮𝐥 𝐰𝐞𝐞𝐤𝐬 — 𝐭𝐰𝐨 𝐧𝐞𝐰 𝐚𝐫𝐭𝐢𝐜𝐥𝐞𝐬 𝐟𝐫𝐨𝐦 𝐨𝐮𝐫 𝐥𝐚𝐛 𝐨𝐧 𝐡𝐲𝐩𝐞𝐫𝐬𝐩𝐞𝐜𝐭𝐫𝐚𝐥 𝐢𝐦𝐚𝐠𝐢𝐧𝐠 𝐚𝐧𝐝 𝐀𝐈 𝐢𝐧 𝐞𝐬𝐨𝐩𝐡𝐚𝐠𝐞𝐚𝐥 𝐚𝐝𝐞𝐧𝐨𝐜𝐚𝐫𝐜𝐢𝐧𝐨𝐦𝐚! 🧬In two recent studies, we explored how hyperspectral imaging (HSI) combined with artificial intelligence (AI) can support diagnosis and treatment planning in esophageal adenocarcinoma. 💡 First: Using advanced neural networks, we demonstrated that: HSI data can accurately distinguish healthy and cancerous tissue in EAC resection specimens. Second: AI models can potentially predict chemotherapy response from pre-treatment samples in EAC. 🎯 These findings highlight the potential of AI-assisted hyperspectral imaging as a non-invasive, data-driven tool for personalized oncology and intraoperative guidance. 📖 Read more: 👇 - https://lnkd.in/eTa2Z7X6 - https://lnkd.in/eyMTW7Kd 🙏 A heartfelt thank you to all co-authors and collaborators, especially Marianne Maktabi from Hochschule Anhalt, Christel Trifone, Claudia Hain, and the entire ICCAS - Innovation Center Computer Assisted Surgery team for their expertise, teamwork, and dedication. #research #publication #AI #cancerresearch #hyperspectralimaging #machinelearning #oncology #teamwork #innovation

🌟 Proud to Announce Another Research Publication! I am pleased to share that our research article titled: “A Convolutional Neural Network and Vision Transformer Based Framework for Effective Detection of Liver Cancer” has been published in the Journal of Computing & Biomedical Informatics (JCBI), Volume 09, Issue 02, 2025. 🩺 Why this research matters: Liver cancer, particularly Hepatocellular Carcinoma (HCC), remains one of the most aggressive malignancies worldwide. Early detection is critical but often challenging due to limitations of conventional diagnostic methods such as CT imaging. ⚡ Our Contribution: We propose a hybrid Deep Learning framework leveraging Convolutional Neural Networks (CNNs) and Vision Transformers (ViTs) for effective liver cancer detection. Key highlights: • Utilized EfficientNet-B0, TinyViT, and MobileViT v2 for robust classification. • Applied augmentation techniques to enhance model generalization. • Achieved high accuracy and interpretability with Grad-CAM for visual explanations of tumor regions. • Designed for resource-constrained clinical environments, ensuring efficiency and transparency. 🙌 This collaborative effort with Asma Zahoor, Erssa Arif, Naila Nawaz, Muhammad Amjad, Arslan Baig showcases how AI-driven medical imaging can significantly enhance early cancer detection and support better patient outcomes. 🔗 Read the full article here: https://lnkd.in/d3EMfFT5 #AI #MedicalImaging #CancerDetection #DeepLearning #HealthcareInnovation #VisionTransformers #CNN #Research

Predicting Brain Tumor Response to Therapy using a Hybrid Deep Learning and Radiomics Approach https://lnkd.in/efpm7qrr Accurate evaluation of the response of glioblastoma to therapy is crucial for clinical decision-making and patient management. The Response Assessment in Neuro-Oncology (RANO) criteria provide a standardized framework to assess patients' clinical response, but their application can be complex and subject to observer variability. This paper presents an automated method for classifying the intervention response from longitudinal MRI scans, developed to predict tumor response during therapy as part of the BraTS 2025 challenge. We propose a novel hybrid framework that combines deep learning derived feature extraction and an extensive set of radiomics and clinically chosen features. Our approach utilizes a fine-tuned ResNet-18 model to extract features from 2D regions of interest across four MRI modalities. These deep features are then fused with a rich set of more than 4800 radiomic and clinically driven features, including 3D radiomics of tumor growth and shrinkage masks, volumetric changes relative to the nadir, and tumor centroid shift. Using the fused feature set, a CatBoost classifier achieves a mean ROC AUC of 0.81 and a Macro F1 score of 0.50 in the 4-class response prediction task (Complete Response, Partial Response, Stable Disease, Progressive Disease). Our results highlight that synergizing learned image representations with domain-targeted radiomic features provides a robust and effective solution for automated treatment response assessment in neuro-oncology. --- Newsletter https://lnkd.in/emCkRuA More story https://lnkd.in/enY7VpM LinkedIn https://lnkd.in/ehrfPYQ6 #AINewsClips #AI #ML #ArtificialIntelligence #MachineLearning #ComputerVision

My first post in this space! I’m excited to share our new preprint introducing CenSegNet—a deep learning framework for high-throughput, spatially resolved, single-cell centrosome profiling in heterogeneous tissues. This project brilliantly co-led by PhD students Jiaoqi Cheng & Keqiang Fan, is a great interdisciplinary collaboration across University of Southampton and University of Cambridge We applied CenSegNet in human breast cancer tissues, and uncovered that numerical and structural centrosome amplification (CA) are decoupled—not only in cellular architecture but also in their associations with tumour grade, hormone receptor status, germline mutation, and patient age. Our findings challenge the longstanding view of CA as a monolithic driver of malignancy. Instead, they highlight CA subtypes as distinct functional modules in tumour evolution, opening new directions for understanding tumour heterogeneity and progression. CenSegNet is: Fast Generalisable Open-source Designed for biologists and pathologists Check it out: https://lnkd.in/ecUfiy77 This work started back in 2021 as a “wild idea” that AI could help us better understand the cell biology of breast tumour heterogeneity. I’m grateful to the University of Southampton Institute for Life Sciences for believing in this vision, and delighted to see it come to life. https://lnkd.in/eFp-c58K

Exciting news! Our research paper “An integrated deep learning approach for enhancing brain tumor diagnosis” has been published in Healthcare Analytics (Elsevier, Q1 journal). This work integrates multiple deep learning techniques to enhance brain tumor detection, contributing toward more accurate and reliable diagnosis in healthcare. Grateful to my co-authors and collaborators for their contributions throughout this journey. 🔗 Read the article here: https://lnkd.in/grV5QYJ3

Distinguishing cancerous extracellular vesicles via nanoaperture optical tweezers and deep learning: in their recent work, Hao Zhang at UCLA, Reuven Gordon at University of Victoria and collaborators demonstrated near-perfect classification of cancerous (two types) and non-cancerous cell-derived extracellular vesicles using nanoaperture optical tweezers (NOTs) combined with deep learning. EVs, which are abundant in biological fluids and carry molecular “fingerprints” of their parent cells, offer great potential for minimally invasive cancer diagnostics. To enhance clinical applicability, improved sensitivity and specificity were needed 🔗 https://lnkd.in/eSF3ZDyB The NOT approach provided label-free, single-EV sensitivity by detecting only the scattered light from the laser tweezers. Exceptional specificity was obtained through a custom deep learning architecture consisting of a four-layer convolutional network and a Kolmogorov–Arnold linear layer, outperforming several alternative models. Beyond diagnostics, their platform enabled real-time EV profiling, offering new insights into EV biology and supporting the advancement of personalized, minimally invasive medicine. An article co-authored by Tianyu(Chris) Zhao, Wenwen Zhang, Sina Halvaei, Matthew Peters, Tsz Shing Cheung and Karla Williams. #extracellularvesicles #exosomes #cancer #diagnosis #Vesiculab