Ron Alfa, MD, PhD’s Post

📣 We have a new CMIG journal paper on detecting rare adrenal and extra-adrenal tumors in CT! https://lnkd.in/depPVmDK ❗Pheochromocytomas and Paragangliomas (PPGLs) are tumors that have metastatic potential, and the tumor's genetic type dictates the therapy pathway for a patient. If we can accurately classify the genetic type non-invasively with CT, then it could potentially remove the need for an invasive genetic test. But in CT, it's difficult to identify the tumors and even harder to annotate them. So, we simply drew boxes around the identified tumors, and used the incomplete/noisy labels to directly train a nnUNet segmentation model. Surprisingly, the model worked well enough to identify PPGLs in new CT scans, and can be used for downstream tasks. 🎉 David Oluigbo, Bikash Santra, Pritam Mukherjee, Jianfei Liu, Abhishek Jha, Mayank Patel, Karel Pacak, Ronald Summers

We're thrilled to share the preprint of our article titled, "Medulloblastoma Spatial Transcriptomics Reveals Tumor Microenvironment Heterogeneity with High-Density Progenitor Cell Regions Correlating with High-Risk Disease" at bioRxiv. Thanks to our dedicated team members and collaborators MacDonald Tobey for making this possible. Big thanks to Franklin Chien, MD, and Marina Michaud for leading the effort and performing excellent analysis. This is the first story mapping spatial architecture and cellular landscape of the medulloblastoma (MB) TME and identifying dense regions of quiescent progenitor-like MB cells enriched in patients with high-risk (HR) features. Read Full Article: https://lnkd.in/e82vVt5T

Just published! In this paper, the delineation of treatment effect from true tumor progression in Glioblastoma multiforme (GBM) remains a challenging task. the use of cerebral blood flow (CBF) as a potential imaging marker holds promise in improving the accuracy of diagnosis and treatment decisions for GBM patients. Further research and validation studies are needed to established the utility of CBF as a reliable marker in the management of Glioblastoma multiforme. https://lnkd.in/dDj57cmg

Our pilot trial of the first use of 5-ala for intaoperative tumor fluorescence was chosen as the cover in the latest publication of the Laryngoscope. https://lnkd.in/ePyTxGrZ

In this short video, we tell you about the main features of Altris AI ✅ Severity differentiation for fast detection of pathological scans ✅ Retina layer thickness ✅ Progression tracking ✅ AI Pathologies detection module for analysis of OCT scans and detection of 70+ retina pathologies and biomarkers ✅ Biomarkers quantitative analysis ✅ Early glaucoma risk assessment ✅ Smart reports and more See how the system works in a Demo Account: https://demo.altris.ai/

This afternoon, checking recent publications with our anti- #QuinolinicAcid #antibody. Here, a nice paper from Pravin Kesarwani et al. in Nature Comms, demonstrating the accumulation of quinolinic acid in glioblastoma and its immunosuppressive role. Paper is based on human tumor specimens, and describes a dedicated rodent model, used to demonstrate underlying signaling mechanisms & to flag kynureninase as potential therapeutic target. Link to the paper: https://lnkd.in/dvqErTV7 Immusmol

#IndicaLabs will be at #AACR2024! Visit my colleagues in booth 3352 to learn about the latest features coming to Indica's AI-powered HALO platforms, from AI-based annotation tools to new algorithms for tumor analysis! #digitalpathology

Dr. He and his team have developed new methods to map the targets of the SNORD genes in PWS. FPWR has funded him to use these methods on mouse models and human cell lines of PWS to better understand the function of these critical PWS genes. https://fpwr.us/3QrTMch

Main features of Altris AI in one video. Altris AI automates the selection of pathological OCT scans and the detection of 70+ pathologies and biomarkers. FDA clearance. CE – certified.

Why did it lag behind? I explored Support Vector Machine (SVM) kernels for cancer cell classification: Benign or Malignant. 🧬 I focused on different kernels—linear, polynomial, RBF, and sigmoid—and evaluated their performance using the F1-score metric. Key Takeaways: 🔍 Achieved high F1-scores with linear, polynomial, and RBF kernels. 📈 Encountered challenges with the sigmoid kernel—0.37 score. I'd love to hear your thoughts on why the sigmoid kernel lagged behind! #MachineLearning #CancerClassification #SVM #DataScience #healthtech