Prajwal Ghimire

“🚨 Podcast 🚨” Dive deep into the world of neuro-oncology with ‘Neuro-oncology Talks | The Deep Dive.’ Explore groundbreaking research, AI in medicine, and the future of brain tumor treatment. 🎙️ Listen now: podcasts.apple.com/gb/podcast/neu… #NeuroOncology #CancerResearch #AIInMedicine

Grateful to co-authors and my supervisors for support and guidance @ThomasCBooth and Dr Marc Modat

3/3 This is a major step toward "digital biopsies" for personalized brain cancer care. 🚀 📖 Full Paper: doi.org/10.1093/noajnl… 💻 Code: github.com/prazg/PRECISE_… #NeuroOnc #AI #Radiomics #Immunotherapy #MedTwitter #PrecisionMedicine #BrainTumor

New Research Alert! 🧠💻 1/3 Can AI map the #glioblastoma immune environment without a biopsy? Our new PRECISE-GBM study in @EditorNeuro says YES. We’ve developed a radiogenomic pipeline to non-invasively predict macrophage M0 signatures using standard pre-operative MRI:

Nature research paper: Non-invasive profiling of the tumour microenvironment with spatial ecotypes go.nature.com/4wf1BFl

We've known how important the tumor microenvironment is for cancer progression and treatment, but we never had a non-invasive blood test to assess it. Today, as reported @nature, one has been discovered nature.com/articles/s4158…

#JNSCaseLessons. Researchers highlight variant case of spontaneous CSF rhinorrhea arising from combined Sternberg canal defect and intracranial hypertension. thejns.org/doi/10.3171/CA…

Building strong medical AI means large datasets. But in healthcare, data is hard to get. A new study shows you may not need it. By focusing training on key clinical concepts, a vision-language model achieved state-of-the-art performance using just 1% of the typical data while preserving its reasoning ability. What do we think of this shift? Does progress in medical AI come less from more data and more from teaching models what actually matters? nature.com/articles/s4174…

Multi-Embed is an interpretable framework that enables integrated analyses of histological images and multilayer molecular profiles. nature.com/articles/s4159…

1/Do radiologists sound like they are speaking a different language when they talk about MRI? T1 shortening what? T2 prolongation who? Here’s a translation w/an introductory thread to MRI.

The operating room is one of the hardest environments for AI: constant motion, multiple devices, and high stakes. There is really zero room for error. But general-purpose models are not built for that reality. A new study introduced ORQA, a specialized multimodal foundation model that understands surgical scenes using video, audio, and structured data. It significantly outperformed generalist AI systems and was designed in multiple sizes for real-world deployment. How do we continue building systems specifically trained for the complexity of care? nature.com/articles/s4174…

@HimalayaAILab Excellent initiative. I am happy to contribute and collaborate.

We are proposing the Nepal AI Sangraha—a sovereign national data repository for public access to datasets and models Inspired global initiatives like India’s AIKosh and the open-source power-house of Hugging Face, this platform will host anonymized & non-sensitive public dataset

Can't remember the types of craniotomies off the top of your head?   Do you need another thing to memorize like you need a hole in your head?   Here's a quick figure with a review of some common craniotomies to help you out!   Let’s dig deeper into these holes!   Frontal: For large midline anterior and sometimes middle skull base lesions   Temporal: For middle cranial fossa lesions, including intra-axial lesions in the mesial temporal lobe 
Parietal: For mid to posterior cerebral hemisphere lesions while sparing the motor and sensory cortices.   Parietal interhemispheric approach for parafalcine, medial parietal, and splenial lesions. Parietal transcortical route for intra-axial lesions through the functionally “silent” superior parietal lobule   Pterional (frontotemporal): Access to the frontal lobe, temporal lobe, and Sylvian fissure and allow for direct visualization of the microvasculature surrounding lesions in the suprasellar cistern

Mapping nerves in a whole embryos. We find that across species and development stages, embryonic nerves display (beautiful) fractal geometry. More here: biorxiv.org/content/10.648…

Tumor cells often metastasize in clusters with other cells, offering novel targets for treatments. Learn more: scim.ag/4dF9dui @NewsfromScience

🎓 Completed "Digital Clinical Safety Training" with @NHSEngland & @NHSDigital Learning how to build safer, smarter digital health systems is so important as the NHS continues to grow digitally. 💙 🔗digital.nhs.uk/services/clini… #NHSDigital #HealthTech #DigitalHealth #NHS

Our paper is out! This has been years in the making. We found that mental imagery and perception do share a neural substrate, but we see it in in higher-order transmodal networks rather than earlier sensory systems. Check it out! doi.org/10.1016/j.neur…

Aging clocks may be shaped by neurosyndemics, multiple interacting physical and social real-world environments jointly influencing brain health. Out in Nature Medicine (nature.com/articles/s4159…), we assessed 18,701 participants from 34 countries, showing that the combined aggregate-level exposome (73 physical, social, and political factors measured at country-level) predicts multimodal brain aging far better than isolated exposures (up to 15-fold more variance). Moving beyond single risks, we provide evidence that synergistic, nonlinear exposome burden accelerates brain clocks across health and disease, with physical exposures linking more strongly to structural brain aging and social exposures to functional brain aging. Exposome burden increased the risk of accelerated brain aging by 3.3–9.1-fold, in some cases exceeding the effects associated with dementia, and these findings held in out-of-sample, longitudinal, individual-level variation, and sensitivity analyses. Thus, the pace at which the brain ages may be shaped by syndemic environmental and societal conditions, calling for much more intersectoral policies. Congrats @AgustinaLegaz Sebastian Moguilner @HernHdezL & all coauthors. 1/5👇 @GBHI_Fellows

Bridging the Gap: How TRIBE v2 is Set to Transform Neuro-Oncology 🧠💻: Infographic Summary

Very happy to finally see this work published, where we investigate the role of anti-epileptic drugs (AEDs) in the survival of a cohort of GBM patients. Thank you @lavrador_jose for the support and invaluable guidance. @KingsNeuro @EditorNeuro academic.oup.com/noa/article/8/…