Carlo Bifulco: How AI Can Address Critical Healthcare Needs While Maintaining Clinical Rigor

Carlo Bifulco, Chief Medical Officer at Providence Genomics, shared a recent article by Gabriele Campanella et al, on LinkedIn, adding:

“What an incredible milestone!

Seeing GigaPath deployed in real-world clinical practice in a prospective study represents exactly the kind of transformative impact we envisioned when we made this pathology foundation model open source. The 89% AUROC for EGFR detection and 43% reduction in molecular testing demonstrates how AI can address critical healthcare needs while maintaining clinical rigor.

There is broader potential of GenAI in pathology in several key areas:

• Biomarker Predictions – Beyond EGFR and other genomic precision medicine markers, we’re seeing promise in IHC predictions for ADC targets and other therapeutic markers.
• HnE-based Signatures – AI can replicate proprietary widely used clinical mRNA signatures that currently require centralized labs, dramatically reducing turnaround times and improving access.
• Pathologist Diagnostic Assistants – Emerging multimodal AI models augment (not replace) pathologists, and will help address workforce shortages while improving diagnostic accuracy.
• Multimodal Patient Embeddings – Further in the future, integrating pathology with genomics, medical imaging, lab results, and clinical notes for truly holistic patient insights that go beyond the current narrow companion diagnostics approach.

All this is complementary to the ongoing advances in digital pathology and IHC quantification that are finally transforming pathology practice. To re-emphasize on one key message: AI augments, does not replace. While we’re giving new powers to the pathologist and augmenting the microscope, we are not replacing genomic tools.

Congratulations to the Amazing Teams:

Huge congrats to Thomas Fuchs, Gabriele Campanella, Chad Vanderbilt and the entire team for this groundbreaking clinical deployment! Also want to recognize the incredible work from other pioneers in the field – Faisal Mahmood and team, Richard J. Chen, Ming Yang (Max) Lu, Jan Witowski, Siqi Liu, Razik Yousfi, and all the researchers pushing the boundaries of GenAI in digital pathology.

Special thanks to my Providence colleagues Brian Piening, Bill Wright, Ari Robicsek, Walter Urba, R. Bryan Bell and our senior leadership including Susan Huang for believing in the power of Generative AI and supporting the open-source release of GigaPath for the global community.

Last but not least, deep gratitude to Hoifung Poon, Naoto Usuyama, Sheng Wang, Hanwen Xu for years of collaborative work that enabled the creation of GigaPath.

Future generations will look back at our current fragmented, resource-limited diagnostic landscape as a dark age. The transformation of healthcare through AI represents one of the most profound opportunities of our time.”

Title: Real-world deployment of a fine-tuned pathology foundation model for lung cancer biomarker detection

Authors: Gabriele Campanella, Neeraj Kumar, Swaraj Nanda, Siddharth Singi, Eugene Fluder, Ricky Kwan, Silke Muehlstedt, Nicole Pfarr, Peter J. Schüffler, Ida Häggström, Noora Neittaanmäki, Levent M. Akyürek, Alina Basnet, Tamara Jamaspishvili, Michel R. Nasr, Matthew M. Croken, Fred R. Hirsch, Arielle Elkrief, Helena Yu, Orly Ardon, Gregory M. Goldgof, Meera Hameed, Jane Houldsworth, Maria Arcila, Thomas J. Fuchs, Chad Vanderbilt.

Read the Full Article on Nature Medicine.

Bernard Fox, Co-Founder, President, CEO at UbiVac, shared this post adding:

“WOW!!!

‘Future generations will look back at our current fragmented, resource-limited diagnostic landscape as a dark age. The transformation of healthcare through AI represents one of the most profound opportunities of our time.’

Very proud of my incredibly talented and visionary colleague Carlo Bifulco , whom I quoted above. In addition to teaching me the smidgen of pathology I might know, his perspective on what can be achieved from evaluating a mn HnE stained section of cancer has been ‘mind blowing’. Instead of waiting days or weeks for a genetic test, a program can look at the slide and predict it has the mutation that is targeted by

As a patient, You might not need a potentially expensive genetic test that takes days to weeks to run, and might not be run or covered by your insurance company, before you received a targeted therapy that might give you high quality of life for an extended period.

And the program is free to download and use… Thanks to Providence !!

This wasn’t just Carlo, This was years of effort by teams of ROCKSTARS with diverse talents. led by Hoifung Poon at Microsoft AI and by others at Earle A. Chiles Research Institute / Providence Cancer Institute”

More posts featuring Bernard Fox.