MedAI

Описание: Title: From Data Integration to Clinical Decision Support - An End-to-End Multimodal AI Framework for Precision Oncology

Speaker: Asim Waqas, Ph.D. ; Aakash Tripathi, Ph.D.

Abstract:
We present an end-to-end, open-source pipeline that links three complementary efforts into a single workflow for precision oncology: (1) data curation and harmonization with MINDS, which fuses heterogeneous oncology sources into a patient-centric, machine learning - ready graph/metadata layer; (2) representation learning with HONeYBEE, which produces modality-specific and fused patient embeddings across structured EHR, clinical text, whole-slide pathology images, radiology, and molecular profiles; and (3) multi-omics modeling via SeNMo, a self-normalizing foundation model integrated into HONeYBEE to generate robust omics embeddings that tolerate missing modalities. Together, these components enable downstream tasks such as survival risk stratification, cancer-type classification, patient-similarity retrieval, cohort clustering, and other decision-support functions. In public-cohort evaluations (e.g., TCGA), clinical embeddings consistently emerge as a strong single modality, while multimodal fusion adds complementary gains for selected cancers; for text, general-purpose LLMs perform competitively and task-specific fine-tuning improves performance on heterogeneous report styles. The result is a scalable chain, from interoperable data to unified embeddings to actionable analytics, designed to accelerate reproducible research and translation to clinical decision support.

Speaker Bio:
Asim Waqas, Ph.D., is an Applied Research Scientist in the Machine Learning/ Cancer Epidemiology Departments at Moffitt Cancer Center. Aakash Tripathi, Ph.D. is a Machine Learning Engineer in the Machine Learning Department at Moffitt. They both specialize in multimodal medical AI. With a background in Electrical and Computer Engineering from the University of South Florida, they work at the intersection of machine learning and oncology to improve cancer research and treatment outcomes. Their research focuses on developing novel AI architectures for integrating complex medical data types, creating scalable frameworks for clinical applications, and advancing survival prediction models in oncology. They have multiple publications and are working on several projects with high clinical translation impact.

Related Papers:
HONeYBEE: https://doi.org/10.1038/s41746-025-02...
SeNMO: https://doi.org/10.3390/ijms26157358
MINDS: https://doi.org/10.3390/s24051634
AACR 2026: Multi-agent AI orchestration for temporal-aware extraction of social determinants of health from unstructured clinical records in cancer populations.
AACR 2026: Real-world evaluation of multimodal AI: Foundation model-driven multimodal AI for GBM, NSCLC, and PDAC

------

The MedAI Group Exchange Sessions are a platform where we can critically examine key topics in AI and medicine, generate fresh ideas and discussion around their intersection and most importantly, learn from each other.

We will be having weekly sessions where invited speakers will give a talk presenting their work followed by an interactive discussion and Q&A.

Our sessions are held every Monday from 1pm-2pm PST.

To get notifications about upcoming sessions, please join our mailing list: https://mailman.stanford.edu/mailman/...

For more details about MedAI, check out our website: https://medai.stanford.edu. You can follow us on Twitter @MedaiStanford

Organized by members of the Rubin Lab (http://rubinlab.stanford.edu) and Machine Intelligence in Medicine and Imaging (MI-2) Lab:
Nandita Bhaskhar (https://www.stanford.edu/~nanbhas)
Amara Tariq (  / amara-tariq-475815158  )
Avisha Das (https://dasavisha.github.io/)
Weijie Chen (  / winstonchen2017  )