Technical University of Munich’s Post

Reimagining Cancer Detection with DIM12 We’re entering a new era in healthcare innovation. DIM12 our AI + Quantum-powered Multi-Cancer Detection Platform is setting new standards for early and accurate diagnosis. What Makes DIM12 Different: • Combines multi-omics data, radiology, and AI imaging for a holistic diagnostic approach • Achieves high sensitivity & specificity, minimizing false negatives and positives • Designed to be scalable across hospitals, diagnostic labs, and national screening programs * DIM12 Impact: - 95%+ Detection Accuracy - 12+ Cancer Types - Early-Stage Detection that saves lives By integrating the power of AI and Quantum Computing, DIM12 represents a leap forward in preventive healthcare making early detection faster, smarter, and more accessible. #DIM12 #QuantumAI #HealthcareInnovation #CancerDetection #ArtificialIntelligence #MedicalTechnology #FutureOfHealth #Innovation #Dimechain #HealthTech #DigitalTransformation #dimechain Dimechain

✨ Excited to share that our work, “PS3: A Multimodal Transformer Integrating Pathology Reports with Histology Images and Biological Pathways for Cancer Survival Prediction” will be presented at ICCV 2025! In this work, we introduce PS3, a Multimodal Transformer that predicts cancer survival outcomes by integrating three complementary data modalities: ·     Pathology reports ·     Histology whole slide images ·     Genomic data PS3 uses prototype-based representations to distill key information from each modality, enabling the model to learn meaningful relationships across textual, visual, and molecular data. This work highlights how multimodal AI can bridge language, vision, and biology within a unified framework. 📍 Poster presentation: Exhibit Hall I #222 🗓️ Thursday, Oct 23 10:45 a.m. —12:45 p.m. HST presented by Prof. Nasir Rajpoot 📄 Read the preprint: https://lnkd.in/eiNHz7CX Grateful to my co-authors Ayesha Azam, Talha Q. and Nasir Rajpoot for their support and collaboration! #ICCV2025 #ICCV #DeepLearning #Multimodal #ComputationalPathology #AI #ComputerVision

🧬 mCODEGPT: Turning Cancer Notes into Structured Research Data — No Labels Required A new AI system called mCODEGPT is changing how cancer data is captured. Using GPT-4o and a zero-shot hierarchical prompting method, it transforms messy clinical notes into standardized cancer records (mCODE™) — with 94% accuracy and fewer than 2.5% hallucinations. Instead of relying on human labeling, mCODEGPT guides the model step-by-step to pull out key details like tumor type, stage, treatment, and genomic info. This innovation could streamline EHR data sharing, accelerate research, and power precision oncology — all while protecting patient privacy. 💡 From scattered text to structured insight, mCODEGPT shows how AI can make cancer data smarter, faster, and more usable than ever. Source: https://lnkd.in/ebdSj96w Published Date: October 13, 2025 👉 Comment BIOHACK if you want more science and health news like this! #BiohackYourself #HealthNews #ScienceNews #ResearchUpdates #Biohack Disclaimer: This content is for educational and entertainment purposes only and is not a substitute for medical advice. Always consult a healthcare professional. Full disclaimer: https://lnkd.in/eJE9Rsty 🧠 We explore all angles — ancient wisdom, modern science, and everything in between. No allegiance to Big Pharma or Big Natural. 🔍 We cite studies, but encourage you to read them, question funding, and review the methods. Stay curious. 📚 Not all journals are equal. Peer-reviewed ≠ perfect. Check the source, think critically, and decide for yourself. ⚠️ One study isn’t the full story. Science evolves. We’re here to inform, not to tell you what to believe.

Mr. Arunabha Tarafdar (Assistant Professor, #scsetbennett), Dr. Susmita Das (Assistant Professor, #scsetbennett), Dr. Sounak Sadhukhan, Ph.D (Assistant Professor, #scsetbennett), Atrija Haldar (B.Tech, #scsetbennett, Batch 2021-2025) for acceptance of the #conference paper, “Evaluating Deep Learning Models for Histopathologic Oral Cancer Detection” for #publication in 5th Asian Conference on Innovation and Technology (ASIANCON). This paper explores early detection of oral cancer using deep learning to automate histopathological analysis. We compared popular CNN architectures—including MobileNetV2, DenseNet, Xception, Inception, VGG19, and ResNet—on accuracy and computational efficiency. MobileNetV2 achieved the best performance with 87% accuracy, showing that lightweight, optimized models can significantly enhance clinical diagnostics where both speed and accuracy are critical. #bennettuniversity #ai #research #multicloud #socialmedia# #AIInHealthcare #MachineLearning #CancerResearch #ClinicalAI #DigitalPathology

"More broadly, the breakthrough lays the foundation for achieving the long-term goal of building digital models of human cells. DOLPHIN generates richer single-cell profiles than conventional methods, enabling virtual simulations of how cells behave and respond to drugs before moving to lab or clinical trials, saving time and money. The researchers’ next step will be to expand the tool’s reach from a few datasets to millions of cells, paving the way for more accurate virtual cell models in the future." https://lnkd.in/g28_WyWe

At #Biomarkers & #PrecisionMedicine UK, Dr. Evans will be discussing critical biomarker program considerations, including: • Avoiding Pitfalls • Building for Impact • Selecting Biospecimens with Purpose • Tracking What Matters • Looking Beyond Single Omics • Being Regulatory-Ready from Day 1 Learn more: https://hubs.ly/Q03Jjvcv0 #BIOMUK2025 #OGPREMED

Thrilled to share that our team successfully presented our research paper titled “Automated Skin Cancer Detection and Risk Prediction using ResNet50��� at the ICIDCA 2025 International Conference! ✨ This project was a collaborative effort by our amazing team — PEER SHAHEDA, Rithika Gadewar and SABBANI SAMHITHA 🧠 About the Work: Our research introduces a hybrid deep learning model that detects skin cancer and predicts its risk level (low/high). We used ResNet50 for deep feature extraction and combined it with SVM classification to enhance diagnostic accuracy. By fusing both handcrafted and deep features, the model achieved outstanding performance with: ✅ 94% accuracy ✅ 95% precision ✅ 95% AUC-ROC 💡 Key Contributions: 🔹 Two-phase model for detection and risk assessment 🔹 Feature fusion for higher accuracy and interpretability 🔹 Clinically relevant design inspired by the ABCDE rule 🔹 Supports early and automated diagnosis to assist dermatologists Presenting at ICIDCA 2025 was an incredible experience — learning from experts, exchanging ideas, and showcasing how AI can revolutionize healthcare. We’re deeply grateful to our mentors and the organizing committee for this opportunity! 🙏 #ICIDCA2025 #ResearchPresentation #TeamWork #SkinCancerDetection #ResNet50 #DeepLearning #MedicalAI #HealthcareInnovation #ComputerVision #MachineLearning #ProudMoment #AcademicResearch #AIForGood

Proud to have spoken at #TECHDay DS4H on behalf of Median Technologies , sharing how data and AI can enable early lung cancer detection. From data acquisition and engineering to annotation and modeling, I presented how we built our eyonis® LCS AI solution for assisted diagnosis (CADe/x) and how innovative data workflows drive cutting-edge performance in medical image analysis. Inspiring discussions and great energy around the future of precision diagnostics and AI in healthcare! #AI #DataScience #HealthcareInnovation #MedicalImaging #PrecisionDiagnostics #LungCancerScreening #MedianTechnologies

If AI sat next to you in clinic tomorrow, what job would you hand it first? A few days ago at the MedFuel Cardio-Oncology AI Conference (Panel 1), speakers across genomics, cardiology, and oncology agreed—AI isn’t replacing clinicians, it’s augmenting them. Highlights that hit home: 1. Clinician-in-the-loop or bust: if it adds friction, it fails. 2. Trials & rare cancers: let AI find pattern-matched look-alikes when N is tiny. 3. Cardio-oncology: imaging + NLP are reshaping risk detection, but validation still lags. 4. Data governance: integration is the chokepoint; privacy and consent aren’t optional. 5. Patients as partners: telehealth success stories prove connection scales with trust. Moonshots mentioned: digital-twin modeling, proteomics at scale, affordable risk prediction tools, and true system-to-system interoperability. Thank you, Frank Aziz and #MedFuel, for the great event! #MedFuel #PrecisionOncology #CardioOncology #Multiomics #ClinicalTrials #HealthcareInnovation