Five ways AI can unlock the full potential of ADCs

Antibody-drug conjugates (ADCs) represent a groundbreaking advancement in oncology. These innovative therapies combine the targeting precision of monoclonal antibodies with the potent cytotoxicity of chemotherapy, enabling highly specific and effective cancer treatment. Over the past few years, ADCs have captured the spotlight in the oncology world, dominating discussions at global conferences such as the American Society of Clinical Oncology (ASCO) and its European version, the European Society for Medical Oncology (ESMO). Their success has sparked significant multi-million-dollar deals in the pharmaceutical and biotech industry, driving a rapidly expanding market presence.

The numbers speak for themselves: With 12 FDA-approved antibody-drug conjugates across more than 20 indications and over 200 in active clinical development, the ADC market is expected to be worth $28 billion by 2028¹. As ADCs continue demonstrating remarkable success in treating solid and hematologic malignancies, their future looks brighter than ever. But what makes ADCs so compelling, and what challenges remain in unlocking their full potential?

What are ADCs, and why are pharma investing so heavily in them?

ADCs are a sophisticated drug class made up of three key components:

• A monoclonal antibody that binds specifically to a tumor-associated antigen
• A highly potent cytotoxic payload that can kill cancer cells
• A chemical linker that connects the antibody to the payload

This innovative design combines the specificity and long circulation half-life of antibodies with the potency of chemotherapies. By specifically targeting tumor cells with the antibody component, ADCs minimize exposure of healthy tissues to the cytotoxic payload. This precision reduces off-target toxicity and expands the therapeutic index—the range between an effective dose and a toxic dose—allowing the safe and effective use of highly potent agents for treatment.

Since the first FDA-approved ADC in 2000², the field has advanced significantly, with payloads and targets now meticulously fine-tuned to maximize effectiveness. This progress reached a new milestone in April 2024 with the first tumor-agnostic approval for a HER2-targeted therapy in metastatic HER2-positive solid tumors, highlighting the transformative potential of ADCs in reshaping cancer treatment.³

Pharmaceutical companies are investing heavily in ADCs for good reason. They represent a fundamental advancement in oncology, with the potential to:

• Provide targeted treatment for specific patient populations
• Address tumors resistant to standard therapies
• Reduce systemic toxicity to improve patient quality of life

Recent clinical successes have further validated ADCs’ promise. For instance, in the phase 3 DESTINY-Breast03 trial for HER2-positive metastatic breast cancer, the ADC ENHERTU achieved a remarkable 72% reduction in the risk of tumor progression or death compared to the standard of care⁴—further consolidating ADCs as a transformative force in oncology and driving increased innovation and investment in the field.

The challenges ahead

Despite their potential, ADCs are not without challenges. Developing and manufacturing ADCs is a complex and costly process, and their safety profiles must be carefully managed. ^5, 6

ADC resistance‍

Intratumoral and intertumoral heterogeneity can lead to acquired resistance, limiting therapeutic outcomes. The mechanisms behind resistance remain poorly characterized, likely due to the complex mechanism of action of ADCs. Tumors can evade ADC activity through various pathways, including reduced antigen expression, altered intracellular transport, and payload resistance. Intracellular transport is critical for ADC efficacy, enabling the ADC to reach the lysosome, where the payload is released. Disruptions in this process can hinder payload release and reduce therapeutic effectiveness. While pre-clinical studies have validated these mechanisms, clinical evidence remains limited.

Toxicity risks

Despite being designed to be tumor-selective, ADCs face toxicity challenges both similar and distinct to those of chemotherapies. Many ADCs have been withdrawn from clinical trials or from the market due to unacceptable toxicities or narrow therapeutic windows. Even with recent successes, clinical development remains fraught with significant failure rates, as off-target toxicities frequently limit ADC doses below levels needed for optimal anti-cancer efficacy. How and when the ADCs are processed in the body is closely linked to their level of toxicity—when the payload is released in normal tissues rather than the tumor, it can damage healthy cells and contribute to adverse effects.

Biomarker identification‍

Another key challenge in ADC development is identifying biomarkers that accurately predict therapeutic activity. While the targeted nature of ADCs suggests that higher expression of the antibody target on tumor cells should correlate with increased efficacy, this relationship is often inconsistent. Most ADCs' response rates do not align clearly with target expression levels, leading to their approval without requiring prior testing of the corresponding antigen.

A deeper understanding of disease and patient heterogeneity—and the ability to define homogeneous patient populations—will be critical for optimizing ADC efficacy. Striking the right balance between tolerated and efficacious doses requires innovative drug design and patient stratification, with a crucial need for reliable biomarkers to guide therapy decisions and ensure ADCs reach those most likely to benefit.

Owkin is perfectly positioned to drive ADC development success

At Owkin, we accelerate ADC development by harnessing multimodal patient data, including advanced modalities from our MOSAIC spatial oncology atlas, such as single-cell and spatial transcriptomics. Our AI models transform these modalities into novel, data-driven representations, providing unparalleled insights into the underlying biology. These insights empower biotech and pharmaceutical companies to make informed, critical decisions throughout the ADC development process.

Inform ADC development decisions: 

• Select optimal indications and patients: By deciphering inter- and intra-patient heterogeneity, we can identify homogeneous populations most likely to respond to ADCs, enabling precise selection of indications and patient cohorts for development
• Discover key biomarkers: Our AI models can find actionable biomarkers by uncovering previously unseen patterns in multiscale, multimodal data. This supports clinical trial strategies and regulatory success.‍
• Define combination strategies: By analyzing spatial patterns of target expression, resistance mechanisms, and TME interactions, we gain critical insights into how therapeutic targets are distributed and interact within the tumor microenvironment. We can provide recommendations for combination strategies to enhance therapeutic efficacy and overcome resistance.‍

Strengthen clinical trial design

• Refine patient selection: By uncovering heterogeneity within patient groups, based on their biology, we optimize inclusion/exclusion criteria to enrich trials with the most relevant populations.
• Improve clinical success: By integrating biomarker-driven findings, precise cohort definitions, and predictive insights, we minimize trial variability and optimize design to maximize the likelihood of meaningful clinical outcomes.

Our approach combines high-quality, differentiated data, advanced AI, and insights from an expert network of key opinion leaders to deliver tailored solutions that address the unique challenges of ADC development.

An example of MOSAIC data analysis quantifying levels of expression of Nectin-4 in malignant cells in various cancers

The path forward

ADCs represent a turning point in cancer treatment, with the potential to transform oncology care for millions of patients. However, realizing this potential requires addressing complex challenges with innovative solutions. At Owkin, we are committed to partnering with pharma and biotech companies to harness the full promise of ADCs—bringing life-changing therapies to patients faster and more effectively.

By combining world-class datasets, advanced AI capabilities, and unparalleled expertise, we empower our partners to drive innovation and improve outcomes in the evolving field of antibody-drug conjugates, from early discovery to development and diagnostics.

‍

References:

¹ Source: Evaluate Ltd

² Mylotarg (gemtuzumab ozogamicin) was first approved in 2000 for patients with CD33-positive Acute Myeloid Leukemia, withdrawn in 2010 following safety concerns and gained new approval in 2017

³ AstraZeneca and Daiichi Sankyo’s Enhertu (trastuzumab deruxtecan) has been approved in the US for the treatment of adult patients with unresectable or metastatic HER2-positive (IHC 3+) solid tumours who have received prior systemic treatment

⁴ N Engl J Med (2022) 386: 2346-2347

⁵ Cancer Discov (2024) 14 (11): 2089–2108

⁶ Nat. Rev. Clin. Oncol. (2024) 21: 203–223

⁷ Mylotarg (gemtuzumab ozogamicin) was first approved in 2000 for patients with CD33-positive Acute Myeloid Leukemia, withdrawn in 2010 following safety concerns and gained new approval in 2017