Only with the adoption of digital imaging and AI-powered analysis will next-generation precision oncology therapies reach their full potential and ensure no eligible patient is overlooked.
Over the last two decades, cancer treatment has rapidly evolved from conventional chemotherapy and radiation to highly targeted therapies that rely on the molecular profile of a patient’s tumor.
Despite these advancements, many diagnostic labs still rely on manual pathology methods, unable to deliver the precision needed to guide the next generation of treatment. Few labs today are equipped to perform the multiplex immunohistochemistry assays and computational image analysis that tomorrow’s oncology therapies will require. That must change.
Enter computational pathology. By leveraging digital imaging and AI-powered analysis, pathologists are able to quantify biomarker expression with pixel-level precision—far surpassing the capabilities of traditional scoring. As executives at Leica Biosystems focused on the use of AI in biopharma, we have seen firsthand how diagnostic innovation is needed to support novel precision medicine. Only with the adoption of computational pathology in the clinical laboratory will next-generation targeted therapies reach their full potential.
From Qualitative to Quantitative Tumor Scoring
A pivotal moment in the evolution of precision oncology occurred when Herceptin (trastuzumab), a targeted therapy for breast cancer patients with high levels of HER2 expression, was approved in 1998. Companion diagnostics enabled clinicians to identify eligible patients—but these diagnostics were largely qualitative or semi-quantitative in nature.
Today’s antibody-drug conjugates (ADCs) require more sensitive and quantitative approaches. Therapies such as Enhertu (trastuzumab deruxtecan) are effective even for patients with low HER2 expression. Identifying these patients via manual scoring is challenging, but with computational pathology, we can perform continuous, quantitative assessment, helping to ensure no eligible patient is overlooked.
With multiple ADCs now approved or in development for HER2 and another protein overexpressed in some cancers, TROP2, we’re entering into an era where multimarker, multiplex assays are becoming essential. Assessing expression patterns, spatial distribution and co-localization of biomarkers will be key to selecting the right therapy—and, as conventional microscopic review by pathologists cannot effectively handle this level of complexity, computational pathology will be a required component of these assays.
Beyond ADCs, the rise of bispecific antibodies and immunotherapies is driving the need for richer biomarker analysis. These therapies depend not just on the presence of individual markers, but on the context in which they appear: their co-expression, spatial relationships and cellular environments. Algorithms trained on whole slide images can analyze thousands of cells, identify cell types and quantify their interactions—all with precision and reproducibility.
Implementing Precision Diagnostics
Despite these recent advances, fewer than 20% of clinical labs have adopted digital pathology. This is mainly due to a lack of proven clinical utility, which is quickly changing, and the absence of digital infrastructure needed to support high-throughput image analysis.
The path forward begins with digitization. Digital slide scanners generate the data that algorithms need to run. From there, labs can build toward fully integrated, automated workflows—combining staining, scanning and analysis into a seamless process. These workflows reduce variability, improve reproducibility and help alleviate workforce shortages by allowing pathologists to focus on high-value tasks. At Leica Biosystems, we’re seeing growing demand from our pharma partners to validate complex biomarkers, scale diagnostics and accelerate clinical development.
Precision therapeutics cannot reach their full potential without precision diagnostics. As personalized therapies proliferate, the demand for computational pathology will become impossible to ignore. This shift will not only improve treatment outcomes, but also reduce time-to-treatment, minimize unnecessary toxicities and support more cost-effective healthcare delivery.
