Shu Boles, how does AI play into Qiagen’s corporate vision?
Shu Boles, PhD, MBA, director of strategic marketing, clinical oncology, Qiagen: AI is being leveraged in our variant interpretation package, as well as database and knowledge base. We are serving customers with AI-augmented variant interpretation to match patients with clinical trials or identify therapies. We’re also serving the biopharma segment—taking all the incoming data sets, applying a biomedical knowledge base to make sense of the variants, and extrapolating that data to feed into cancer drug development pipelines. In this data-rich environment, there’s a space for AI to assist us.
I agree with José that while we integrate AI tools, it is important to have a human touch. For example, for a patient workup, we need someone to sign off with this assist of AI, as a final quality control.
Ul Balis, what is your impression of what you’ve heard so far?
Ulysses G. J. Balis, MD, A. James French professor of pathology informatics and director of the Division of Pathology Informatics and of the Computational Pathology Laboratory Section, Michigan Medicine: In one word, opportunity. We have unprecedented variability in the classes of data now available for interrogation and analysis, not just the omics data, but also now the availability of all classes of federated clinical data from the health system’s overall electronic health record. This is a reversal of the standard data reporting model, where typically it is the LIS feeding the central EHR. With reverse federation, the lab system, pathologists, and laboratorians can benefit not just from the lab’s own multiomics data but also from multiplexed data of all classes across the enterprise, and this includes all-important time series data.
When you layer in AI—whether it be agentic or machine learning, targeted pipelines—to transform what we generate from diagnostic information to prognostic, the lab becomes the first generator of understanding the biologic trajectory of disease, as opposed to clinicians being compelled to figure it out. The lab becomes the center point for identifying the starting point of the treatment plan. For example, if a patient has aggressive chronic kidney disease and is expected to quickly evolve to complete renal failure, you’re going to be more aggressive than the typical watchful waiting approach that happens now. We’re on the precipice of a completely new way of recognizing the encoded data we’ve been sitting on for decades, but not using, because we didn’t know it was there. And when you further enrich it by having additional classes of multiplexed data, including all the omics the lab already generates, the predictive power skyrockets.
Gilbert Hakim, you understand the need to structure data and databases properly to solve clinical questions. What comes to mind as you listen to the conversation?
Gilbert Hakim, founder and CEO, SCC Soft Computer: The role we play is in total automation, from sample collection to histology, to wet lab automation and integrating instruments and pipelines, to massaging the data and passing it to scientists. The reality is, the same next-gen sequencing works on the molecular side for oncology as well as for reproductive and hereditary diseases. With our application, you can pull up a patient’s family history to see if a family member had a mutated gene and whether it resulted in disease, or use AI to search a database of patients with the same characteristics to see what the outcomes were or what therapies were used. You have an integration of anatomic pathology and a genetics information system, which is HLA for transplant, flow cytometry, biochemistry, molecular, cytogenetics—all must be accessible to the pathologist or scientist for sign-out. It dramatically reduces the time to look for data in multiple systems. The integration provides the ability to define a search across multiple disciplines in the lab, and the connection to multiple pipelines provides different types of information. In areas like histology, we can now extract DNA from the slide itself. You don’t need another biopsy.