Interview

Interview with Carla Leibowitz, Chief BD Officer at Paige AI-derived Cancer Biomarker Testing Company

Paige

Carla Leibowitz, Chief BD Officer at Paige AI-derived cancer biomarker testing company talks about deriving complex data leading to early diagnosis

1. Tell us about your role as Chief Business Development Officer at Paige.

As the Chief Business Development Officer at Paige, I lead Strategic Partnerships, Business Development, and Marketing. My role focuses on growth and market strategy. I spend most of my time forming relationships with innovative partners, and making sure that the customers and the market know who we are and what value we bring. 

2. How did your professional journey happen to move into the AI tech space?

Paige is my third venture into the clinical AI space. I led Corporate Development and Strategy at Arterys, the first company to achieve FDA clearance for several products that combine cloud computing and artificial intelligence. I also spent time at NVIDIA as the Global Head of Clinical and Life Sciences Partnerships. There, I worked on the Global Research and Partnership Strategy in Genomics, Medical Imaging, and Population Health. I’ve seen some exciting innovation in AI in healthcare while at NVIDIA, Arterys, and now Paige. As an industry, we’ve only scratched the surface on AI’s potential in medicine and science.

3. Can you elaborate more on how AI is empowering the healthcare industry?

We’re only beginning to understand and realize the full value of AI as it applies to Medicine and Science. As AI and machine learning technologies mature and become more broadly adopted, applications are evolving and beginning to demonstrate clinical and economic value in healthcare settings. Some examples include image acquisition and processing that reduce scan times and analysis times in radiology, workflow efficiencies, and triaging technologies that can flag potential conditions automatically and diagnostic tests that have, in some cases, replaced invasive procedures.

At Paige, our goal is to enable pathologists to make efficient, effective, and precise diagnoses. This includes a portfolio of AI-based diagnostic products and an enterprise imaging solution that can deliver this portfolio. We already have FDA clearance and CE mark for our FullFocus viewer, and we’re steadily working towards obtaining regulatory clearance for our computational pathology solutions across different disease types for routine clinical use.

4. Can you explain in layman terms how ML algorithms enable pathologists to make diagnoses more precise and fast?

The technology pathologists use to diagnose disease in tissue has not changed significantly in the past 100 years. At the same time, the demands for pathology services are increasing, while the number of qualified pathologists around the world is dropping. As a result, pathologists’ burnout, and the delays in diagnosis for patients and care teams start getting more pronounced. Moreover, results from traditional diagnostic testing in the standard of care and clinical trials can often take up to two weeks to process and analyze.

ML has the potential to transform the pathologist’s workflow and how pathologists assess tissue, increasing speed and consistency to what has traditionally been a mostly manual, qualitative practice. Besides this, ML can add layers of information they did not have access to before.

From a workflow perspective, ML can be applied to help care teams triage and prioritize their workloads while ensuring quality. From the data and information perspective, ML can be applied to increase the speed of providing results potentially dramatically Ultimately, we believe that the improvements in workflow and information will enable pathologists to reach accurate diagnoses faster for patients and care teams, resulting in better healthcare outcomes. 

5. How do you think mining relevant historic datasets enhances the predictive capabilities for the health sector?

In reality, this is difficult, as hospital data can often be disjointed. Different types of data are stored separately in most institutions. It’s common to have laboratory systems separate from the EMR and for entirely separate systems to handle Cardiology, Radiology, and other surgical information. Datasets, where these items can be linked, are highly valuable and hard to come by. When they are available, they can be used to ensure that resources go to the patients that need them the most, and can be used to detect rare diseases early.

Two great examples of this are an algorithm developed to flag which cardiac patients would most need flu shots and automated systems that analyze claims data to flag potential rare diseases based on symptoms and patient pathways. In our case, we’re exploring how much prognostic and predictive information can be found on images of tissue alone. 

6. Paige recently received the FDA Clearance for the FullFocus™ Viewer for digital pathology. What role did AI play in achieving this milestone? 

Paige’s offering includes the Paige Platform and a portfolio of AI-based digital diagnostics and biomarkers. While AI did not play a central role here, FullFocus™ Viewer is part of the Paige Platform, designed from the ground up to broadly distribute this portfolio, providing an interoperability layer across many different technology setups in labs and hospitals. FullFocus™ Viewer has been designed for the deployment of these powerful AI-based solutions throughout the pathologist’s workflow.

7. Goldman Sachs is now a part of Paige’s growth and advancement. Can you tell us more about that?

Yes, absolutely. Goldman Sachs Merchant Banking Division invested $20 mn in the firm. The initial $5 mn came in our Series B in April this year. They added to their investment in July because they recognized a significant market opportunity and Paige’s unique confluence of capabilities in technology, data, partnerships, evidence, regulatory, and an experienced leadership team. They saw an urgent opportunity to accelerate development and provided the needed investment to do so.

We are also lucky to have David Castelblanco, Managing Director at Goldman Sachs, as a new member of Paige’s Board of Directors. The additional funding helps us continue our work to bring our computational pathology solutions to the market around the globe. It also enables us to grow our team, enter into additional biopharma partnerships, and continue optimizing our platform to deliver clinical use.

8. Can you explain to us in detail about Paige’s infrastructure? 

At Paige, we use the most advanced infrastructure for digital and computational pathology.

For development, Paige has invested in its own private cluster for training machine learning and deep learning models on a very large scale. Given the unique challenges of digital computational pathology, notably the nature of the very large images file sizes to handle, Paige built AIRI, the first AI-ready compute cluster featuring high-end GPUs from NVIDIA associated with extremely fast flash storage.

For its day-to-day operations, and to sustain its global footprint, Paige is built on Amazon Web Services (AWS) Infrastructure-as-a-Service (IaaS) and Platform-as-a-Service (PaaS), which supports Paige in providing a scalable, reliable, secure, available service to all its customers, worldwide. And, we’re proud that the Paige product architecture is developed with a focus on compliance, reliability, security, and privacy. 

9. What do you think will be the next big breakthrough in the Machine Learning and Artificial Intelligence technology space? 

Overall, the next essential step for ML is transitioning from these systems, mainly doing pattern recognition, into systems that can perform complex algorithmic operations and ensure fundamental safety. In terms of healthcare and pathology, specifically over the next five years, the use of diagnosis in digital pathology will likely reach routine clinical use. AI also has enormous potential, particularly for precision oncology, where computational pathology products could significantly improve the selection and stratification of treatments leading to better patient management.

Traditional biomarkers focus on identifying a single gene or protein. In contrast, AI has the potential to learn morphometric phenotypes of these expressions that can capture complex interactions that were previously unseen. Similarly, machine learning can correlate patterns with the outcome. It could potentially mine features for diagnosis that are superior to those in routine practice today, primarily designed by hand.

10. What advice do you have for a company or an individual starting out in the ML space?

Focus on the application of the technology and the market dynamics, not just the technology. Your goal should be to differentiate through one or more of the following: the breadth of the data you’re training on, the novelty of the algorithms, and, ultimately, its impact on users and its attractiveness to the market.

11.  How do you keep pace with the rapidly developing tech world?  

We’re continually recruiting top technical talent for our Machine Learning and Engineering teams. We continually invest in new technology and ensure our team members have time to review the latest research as soon as it is published. This includes an educational stipend for employees to ensure their development is up to date. Personally, I follow what my network is up to and innovations from the large players, researchers, and small companies in the space. 

12. Can you tell us about your team and how it supports you? 

I’m lucky to have stepped into a phenomenal team when I came in as CBDO and all the members who joined us since meet that same high bar. Everyone is intelligent, passionate, and dedicated. We have a culture of hard work and collaboration. I’ve never been let down by my colleagues when it comes to rallying behind a proposal, an initiative, or a big opportunity. Just last week, we had 6 authors contribute to a 16-page piece in a matter of days. It has been amazing how such a small team has achieved so much: 2 CE marks, multiple partnerships, and an FDA clearance and significant fundraising, even in COVID times, and teamwork is how we have succeeded.

13. What book are you currently reading?

I’m re-reading ‘One Hundred Years of Solitude’ by Gabriel García Márquez – one of my all-time favorites.

14. We’d love to get a peek into Paige’s chic and fabulous work culture! Can you share some pictures of get-togethers and office fun activities?

Caroline Hynes

Carla Leibowitz is the Chief Business Development Officer at Paige. She heads upmarket strategy, business development, marketing, and long-term growth initiatives at the company. She previously served as Global Head, Clinical and Life Sciences Partnerships at NVIDIA, where she was responsible for global research and partnership strategy in the fields of genomics, medical imaging, and population health. Carla also led Corporate Development and Strategy at Arterys, the first company to achieve FDA clearances for several products that combine cloud computing and artificial intelligence. Prior to joining Arterys, she spent three years at Bain & Company, consulting for top biotech, diagnostic, and hospital clients. At the beginning of her career, she also designed medical devices and led device development teams at several companies and has more than 16 patents under her name. Carla earned an MBA from the Stanford Graduate School of Business and engineering degrees from both MIT and Stanford.

Related posts

Interview with Matt Schlicht, Co-founder & CEO at Octane AI

AI TechPark

Interview on AI in Healthcare with Harjinder Sandhu, Founder & CEO of Saykara

AI TechPark

Interview with Scott Lucas, Head of Marketing at Concentric Data Risk Monitoring Tool

AI TechPark