Marilyn Bui is an Anatomic and Clinical Pathologist and Scientific Director of the Analytic Microscopy Core at Moffitt Cancer Center & Research Institute, in Tampa, FL USA. She is also professor at the University of South Florida Health Morsani College of Medicine. Throughout her career, Bui has promoted the image of pathologists at different speaking engagements – the latest of which being the DP&AI Congress in December 2024. Here, Bui joins us to share her views on digital pathology and where the field is heading.
In your talk at the 2024 DP&AI Congress, you said that “we are the inherent leaders in precision medicine.” Could you expand on this?
Pathologists play a crucial role in making diagnosis, identifying disease drivers, and discovering biomarkers by analyzing patient samples. By studying the diseases at the levels of tissue, cell, protein, and gene, we provide critical insight into how to effectively treat them. Pathology is the magical sword that arms the dragon slayers, who are our patient-facing clinical colleagues, fighting to conquer the beasts. Without these efforts, we wouldn’t have targeted therapy and advancements in immunotherapy. Of course, this is only possible through collaboration with researchers, clinicians, and oncologists to translate discoveries from the lab into patient care. Pathology is the foundational work that drives precision medicine forward.
Wanting to directly impact patient care, I pursued pathology because of its vital role in research, discovery, and innovation. Pathology leads the way in precision medicine, shaping the future of patient care, and making meaningful contributions to medical advancements. This specialty makes the best use of my medical doctor degree and the PhD training in Molecular Pathology & Immunology.
At the conference, you mentioned that we are in the age of the third revolution in pathology. Could you tell us more about that?
Of course – pathology has undergone three major revolutions since its inception. The first began with the invention of the microscope, allowing us to study cell morphology for the first time. This foundational tool, along with H&E staining, has helped us understand the mechanisms of diseases for decades.
The second revolution is immunohistochemistry (IHC), which allows us to study diseases at the protein level. By identifying protein markers, we can determine the origin of cancers, such as whether a tumor in the lung originated there or if it had metastasized from the breast. IHC also helps classify cancers, such as carcinomas, lymphomas, and sarcomas, and study biomarkers to guide treatment. For example, in breast cancer, markers like ER, PR, and HER2 not only indicate prognosis but also suggest targeted therapies, significantly improving patient outcomes.
The third revolution is molecular pathology, which examines DNA and RNA using techniques like PCR and next-generation sequencing (NGS). These tools have unlocked enormous potential in precision medicine, but they come with challenges. Molecular pathology generates massive amounts of data, often analyzed as graphs or lines, which can seem disconnected from traditional pathology, where we view cells under a microscope.
To bridge this gap, digital pathology and AI have emerged as transformative tools. Digital pathology digitizes microscope slides, allowing images to be shared instantly across healthcare systems, enabling remote collaboration, tumor boards, and consultations without physical slides. This improves workflow, standardizes care, and ensures that even patients in remote areas can access expert diagnoses through telehealth.
AI adds another layer by analyzing vast amounts of data from both traditional and molecular pathology. It helps us measure and interpret complex data more accurately and efficiently. For example, AI can integrate molecular findings with what we see in H&E or IHC slides, offering deeper insights and improving decision-making in precision medicine.
Without embracing these advancements, the field risks becoming outdated, potentially compromising the entire healthcare system, as pathology is foundational to patient care. With digital pathology and AI, we can remain relevant, drive innovation, and lead the way in precision medicine.
Why should pathologists use image analysis or AI – can the human eye no longer be trusted?
In most cases, the human eye is reliable for analyzing histology and identifying morphological patterns, which are the foundation of pathology. However, when it comes to biomarker testing, especially for companion diagnostics, human observation alone may not be sufficient.
For example, when determining if a biomarker is positive or negative in basic IHC testing, the human eye can usually make an accurate judgment. But in advanced biomarker testing, where accurate quantification is crucial, relying solely on human interpretation can lead to errors. These tests often guide expensive, targeted therapies that can be life-saving but also come with toxicity and high costs. A wrong call not only wastes resources but, more importantly, can harm the patient.
Quantification in biomarker testing often involves specific thresholds, such as 1, 5, 10, 20, 50, or 75 percent, which are challenging for the human eye to assess consistently. Manual quantification is not able to reliably access non-round figures. This is where AI can assist by providing more accurate and reproducible measurements – offering new hope to patients who previously had limited options.
Taking the HER2 testing in breast cancer as an example, this paradigm shift has expanded the number of patients eligible for treatment, but identifying HER2-low and HER2-ultralow cases is complex and requires precise evaluation. To meet this challenge, pathologists must undergo additional training, supported by educational resources like HER2know.com. At the same time, the scientific community is working to develop AI algorithms to assist in these difficult cases, ensuring accuracy and reproducibility in testing, interpretation, and reporting. By combining the strengths of human expertise with machine precision, pathologists can better serve patients and help drive advances in precision medicine.
There are only six FDA approved AI tools accessible to the pathology field – why do you think this is?
The adoption of digital pathology and AI has been a slow journey, even for someone like me, who has worked in the field for 20 years. Over the past decade, as a member and former president of the Digital Pathology Association (DPA), I’ve seen the challenges firsthand. Early on, it felt like climbing a steep hill, breaking down barriers to gain acceptance.
However, the COVID-19 pandemic became a turning point. During the crisis, pathologists were thrust into the spotlight to develop and report COVID-19 test results quickly. With in-person work disrupted, there was a sudden need to work remotely. Many critical decisions in patient care – about 90 percent – depend on pathology lab results, so delays weren’t an option. To address this, the US government temporarily loosened regulations, allowing validated remote sign-outs. This flexibility gave digital pathology and AI a significant boost, driving faster adoption.
Since then, investments in digital pathology and AI have surged, and their use has grown rapidly. We’ve now passed the initial hurdles, and over the next decade, we can expect to see even more advanced digital tools widely available, transforming the field and improving patient care.
Why is collaboration so important in digital pathology integration?
Digital pathology involves many stakeholders, and while pathologists are often at the center, collaboration is essential. For example, machine learning scientists bring their technical expertise to solve problems in clinical use cases; IT teams ensure digital pathology systems function smoothly; industry partners develop tools and need input from pathologists to understand what is clinically relevant; administrators manage the infrastructure and support crucial elements in scaling and sustaining innovations, and so forth.
Collaboration is broad, and I’ve leveraged my professional network extensively. Organizations like the DPA, the Society for Imaging Informatics in Medicine, the College of American Pathologists, the Association for Pathology Informatics, the United States and Canadian Academy of Pathology, the American Society for Clinical Pathology, and the American Society of Cytopathology, are working together to unify efforts. Instead of isolated initiatives, we’re creating a larger, collaborative movement to advance digital pathology and AI.
Tell us about the training resources that you’ve helped to develop for pathologists.
Digital pathology and AI can attract the next generation of pathologists by showcasing the exciting opportunities in the field. To support this, I spearheaded to establish the Digital Anatomic Pathology Academy (DAPA) during my presidency at the DPA, alongside Raj Singh.
The idea is simple: future pathologists will likely work in a digital environment. Medical students already learn histology through digital images, and board exams now include whole-slide imaging. However, training during residency varies widely, so we wanted to standardize and democratize access to digital pathology education.
DAPA offers free membership to all trainees, providing access to valuable resources, like case of the month, grand rounds, board reviews, and fellowship opportunities. Trainees can use these tools on any device – computer, phone, or laptop – without needing advanced equipment. They can also attend meetings for free, making this a gift of education for the next generation of pathologists. Whether you’re a department chair, senior professor, or medical student, everyone’s voice is valued equally at DPA. This creates an energizing and welcoming environment that helps students discover how engaging and inspiring pathology can be.
Another amazing gift of education to all pathologists and trainees on precision medicine is the Precision Medicine Academy (PMC) at the Florida Society of Pathologists (FSP). As the President of FSP (2023-2025), I championed this project to fill the gap on education of pathologists in molecular pathology, biomarker testing, digital pathology, and AI. The goal is to empower pathologists, who are the match makers of our patients to benefit from the advancement of targeted therapy and immunotherapy.
For those interested in precision medicine, AI, and making an impact, pathology offers an incredibly rewarding career. Personally, I have achieved success and happiness by being a practicing pathologist and have helped others (patients, colleagues, and trainees) – I encourage others to explore this fulfilling path to pathology.
Newsletters
Receive the latest pathology news, personalities, education, and career development – weekly to your inbox.
