Targeting the Blast Radius: Why Spatial AI Diagnostics Are Crucial for Maximizing the Impact of Antibody-Drug Conjugates

Antibody-drug conjugates (ADCs) are designed to deliver precise blows to cancer cells, much like a well-aimed strike. However, the real power of these therapies lies not only in their direct targeted impact but in the broader “blast radius” and “bystander killing” effect they generate within the tumor microenvironment. 

To fully understand and control this effect, we need to go beyond traditional diagnostic pathology methods. Spatial artificial intelligence (AI), which maps, analyzes, and interprets the complex cellular interactions in a patient’s tumor biopsy, and spatial biomarkers are crucial for guiding these powerful therapies and ensuring they achieve their intended results.

The “blast radius” and “bystander killing” in ADCs

The “blast radius” refers to the zone of influence around the target cells where an ADC also affects nearby cells, not just the intended cancer cells. This effect can be beneficial by eliminating neighboring malignant cells, but it can also risk harming healthy tissue if not properly managed. 

“Bystander killing” occurs when a therapy like an ADC targets a specific protein in cancer cells but also affects adjacent cells that do not express the target protein. This secondary effect can contribute to the overall effectiveness of the treatment but requires careful consideration to avoid unintended damage.

Spatial AI offers researchers and pathologists a powerful tool to analyze these complex interactions across hundreds and thousands of patient samples; it’s like how we went from using flat, physical maps to utilizing GPS on our mobile phones for getting directions–scaling a process easily that couldn’t be done manually before. Scientists are calling this new approach AI-powered spatial proximity scoring.

The shortcomings of traditional pathology

Traditional pathology methods often classify cancer cells in binary terms, such as HER2 positive or negative. This oversimplification fails to capture the nuanced interactions within the tumor microenvironment that are critical to understanding how treatments like ADCs will perform. 

For instance, patients with HER2 “ultralow” expression might benefit from ADCs but are often overlooked by traditional methods. While the human eye can easily conduct binary scoring of cancer cells in traditional pathology, it struggles with the current need for continuous linear scoring (e.g., HER2 high, low, ultralow, and negative) or scoring cells within a specific blast radius.

By relying on binary classifications, traditional pathology may overlook patients who could benefit from therapies if their tumor’s complexity were better understood. For example, patients with HER2 “ultralow” expression might not be considered for ADCs even though they could respond to treatment. AI can help scale both the detection of ultralow and low expressors and complex spatial proximity scoring.

Spatial AI: Mapping the tumor microenvironment

Spatial biomarkers are not just individual data points, but a map of interactions and patterns within the tumor. They reveal how different cell types are positioned relative to each other, how they interact, and how these interactions influence the tumor’s behavior and a patient’s response to treatment. This approach moves beyond static snapshots of individual proteins and provides a dynamic view of the tumor environment.

Spatial AI combines advanced imaging techniques with machine learning to analyze tissue samples across multiple parameters, such as cells, proteins, location coordinates, and cellular interactions, to name a few. By mapping the spatial distribution of proteins, cells, and other critical factors, it uncovers patterns that the human eye cannot detect, such as subtle differences in protein expression or the optimal radius at which bystander killing leads to effective treatment outcomes.

Innovative companies are drawing from geospatial analysis methodologies and applying those techniques to the biological realm. This approach, termed “biospatial,” uses AI to create detailed maps of the tumor microenvironment, allowing for a more accurate prediction of how therapies like ADCs will interact with both their intended targets and surrounding cells.

Why spatial AI is essential for improving patient outcomes with ADCs

ADCs are a powerful targeted and personalized treatment option for cancer patients. However, as the oncology field advances, the need for better companion diagnostics to predict treatment response is essential. 

Researchers are moving towards developing a deeper understanding of the comprehensive map of cell locations and interactions within each patient’s tissue sample and using AI algorithms to determine signature patterns that predict ADC response. By understanding the surrounding tumor biology and creating spatial biomarkers, clinicians can better determine which ADC therapies will lead to the best outcomes for specific patient subpopulations.

For example, AstraZeneca and Daiichi Sankyo have shown how AI spatial scoring can help identify breast cancer patients who positively responded to HER2-directed ADC trastuzumab deruxtecan treatment even though traditional diagnostic methods scored them as a false HER2 negative. Per the guidelines, these patients would have been deemed ineligible for ADC treatment. Since these patients responded positively to therapy, possibly due to additional bystander effects, it demonstrates the potential of spatial AI to expand treatment options.

Additionally, the partners developed a novel AI-powered biomarker to assess expression of the protein TROP2 and to conduct an exploratory analysis of their Phase III TROPION-Lung01 trial, evaluating TROP2-directed ADC datopotamab deruxtecan (Dato-DXd) in non-small cell lung cancer. In TROP2-QCS-positive patients, Dato-DXd lowered the risk of disease progression or death by 43%. This approach demonstrates possibilities for spatial AI to further derisk clinical trials and another pathway for pharma companies to improve patient selection.

AstraZeneca is also deploying this novel AI-powered TROP2 biomarker for prospective enrollment in the AVANZAR trial, a combo study with Dato-DXd plus Imfinzi and chemotherapy, as first-line treatment of advanced NSCLC without actionable genomic alterations.

Scientists are working closely with leading precision medicine and biopharmaceutical companies to leverage spatial AI and solve challenges in ADC development, such as improving patient selection and optimizing trial outcomes. As an example, groundbreaking technologies are currently being deployed to analyze failed trial samples using AI spatial proximity scoring. The goal is to determine whether spatial scoring could better predict treatment outcomes and accurately classify responders and non-responders. These types of studies could lead to new strategies informing patient selection for future trials where traditional methods failed to show efficacy.

The future of cancer treatment

As the field of oncology evolves, we need to shift away from traditional binary “yes-no” biomarkers to a more nuanced, spatial understanding of tumor biology. This approach will allow for more personalized treatment plans that are tailored to the unique characteristics of each patient’s tumor. Spatial AI is essential for fully realizing the potential of advanced cancer therapies, providing the detailed insights necessary to guide these treatments effectively. This ensures that each therapy hits its targets precisely while managing the broader impact on the tumor environment.

Integrating spatial AI into clinical practice could become the standard for evaluating and administering complex therapies like ADCs. This integration would ensure that every aspect of the tumor’s biology is considered in treatment decisions, leading to better patient outcomes.

We need to go beyond finding the right drug, and rather, amplify our understanding of how that drug interacts with the entire tumor. With spatial AI, we can bring a new level of precision to cancer therapy, offering patients safer, more effective and personalized treatments.

Editor’s Note: The author has no financial relationship with any of the companies / products mentioned.

Photo: FatCamera, Getty Images

This post appears through the MedCity Influencers program. Anyone can publish their perspective on business and innovation in healthcare on MedCity News through MedCity Influencers. Click here to find out how.