Blog

All eyes are on emerging molecules, new indications, and new biomarkers as we approach the upcoming World Antibody-Drug Conjugate (ADC) 2024 conference this week.

ADCs have been shown to effectively deliver potent cancer-killing agents directly to tumor cells while sparing healthy tissue. However, despite their promise as a “magic bullet” for cancer, ADCs have been known to have a relatively narrow therapeutic index in clinical trials¹. This is where patient selection strategies based on expression of the target in tumors have the potential to maximize benefit and provide the best chance of clinical success.

The recent failures of AstraZeneca and Daiichi Sankyo’s datopotamab deruxtecan (Dato-DXd) to significantly improve overall survival in breast cancer and lung cancer highlight the critical need for biomarkers to guide therapy decisions^2,3.

Fortunately, there are promising biomarker technologies to watch for at World ADC 2024. AstraZeneca recently reported an AI-enabled TROP2 biomarker strategy, demonstrating that using advanced biomarkers enhances the efficacy of ADCs by targeting the right patients ⁴.

In this post, we outline how the TROP2 story shows that AI and advanced digital pathology techniques can make a game-changing difference in precision oncology.

The ADC Landscape: Potential and Challenges

Just this year, we have seen a 53% year-over-year rise in new ADC trials, with over 200 ADCs in active clinical development and 11 already approved⁵. This surge of activity reflects the exciting potential of ADCs to transform cancer care. By precisely targeting tumor cells and delivering a toxic payload, ADCs offer the possibility of potent anti-cancer effects with reduced systemic toxicity compared to traditional chemotherapy.

However, several key challenges have emerged as the ADC field has matured:

• A narrow therapeutic index, with a fine line between efficacy and toxicity
• Difficulty in selecting optimal patients – a good target alone does not guarantee clinical success
• Complex pharmacology involving multiple steps: antibody binding, internalization, payload release, and spatially-dependent bystander effects, where therapeutically-treated cancer cells can suppress tumor progression in neighboring, untreated cells.

These factors underscore the critical importance of developing robust biomarkers to guide patient selection and predict efficacy. Unfortunately, biomarker development using conventional immunohistochemistry (IHC) approaches have proven inadequate for this task.

The Limitations of Traditional IHC

Standard IHC scoring by pathologists, while valuable, has significant drawbacks when it comes to ADCs:

• Semi-quantitative and subjective
• Limited dynamic range, especially for low-level staining
• Unable to capture spatial heterogeneity and cellular relationships
• Prone to inter-observer variability

For ADCs with complex, spatially dependent mechanisms like bystander effects, these limitations can result in suboptimal patient selection and missed opportunities to deliver effective therapy.

Enter AI-Powered Digital Pathology

Recently, the combination of AI and advanced image analysis has shown to be a fruitful approach for developing next-generation digital biomarkers for ADCs. Recent data demonstrates the power of AI-based approaches like Quantitative Continuous Scoring (QCS) to overcome the limitations of manual IHC and deliver superior predictive performance. This approach to scoring combines the magnitude of a cell’s positive signal intensity with a spatial radius score measuring how close a cell is to positive cells.

Key advantages of AI-powered digital pathology for ADC biomarkers include:

• Greatly improved precision and reproducibility
• Ability to capture the full spectrum of marker expression, including low-level staining
• Quantification of spatial relationships between cells
• Potential to model complex biological phenomena like bystander effects

Impacts of AI Digital Biomarkers on ADC Efficacy

The power of this approach has been demonstrated in recent clinical studies of ADCs targeting HER2 and TROP2. For example, a QCS-based spatial proximity score for HER2 showed superior enrichment for response and progression-free survival (PFS) benefit compared to conventional IHC scoring in gastric cancer patients treated with trastuzumab deruxtecan⁶ (Figure 1).

Figure 1. Using an AI-based spatial biomarker (right) yielded a significant PFS benefit (survival probability of 50% at 9.7 years) compared to using manual scoring of IHC expression (left; survival probability of 50% at 7.1 years).

Similarly, for Dato-DXd in an exploratory analysis of a Phase III non-small cell lung cancer trial, a QCS-based TROP2 biomarker identified a larger proportion of likely responders compared to standard IHC⁷. Importantly, this translated to meaningful improvements in PFS for biomarker-positive patients.

These early clinical successes highlight the potential for AI-powered spatial biomarkers to expand the population of patients who can benefit from ADCs while also improving predictive accuracy.

Looking Ahead: A New Paradigm for Precision Oncology

Translational research teams are facing an unmissable opportunity to drive a paradigm shift in how tissue biomarkers are developed and applied for cancer therapy. By harnessing the power of artificial intelligence and advanced image analysis, we can:

• Achieve greater analytical precision in quantifying target expression
• Better understand tumor heterogeneity and spatial biology at the cellular level
• Develop objective, reproducible, and scalable biomarker solutions
• Ultimately deliver greater clinical accuracy in matching the right patients to the right therapies

As the field of ADCs continues to expand and evolve, AI-powered digital pathology will be essential to unlocking the full potential of these promising therapies. The future of precision oncology is here, and at Nucleai, we’re proud to be at the forefront of AI digital biomarker development.

Interested in implementing quantitative scoring and other AI-enabled approaches in your clinical programs? Contact us to chat with our team of biomarker scientists and digital pathology experts. >> Contact Us

References

1. Clin Cancer Res(2019) 25 (18): 5441–5448.

2. Daiichi Sankyo, Inc. September 23, 2024, Datopotamab Deruxtecan Final Overall Survival Results Reported in Patients with Metastatic HR Positive, HER2 Low or Negative Breast Cancer in TROPION-Breast01 Phase 3 Trial [Press Release] https://www.businesswire.com/news/home/20240921214470/en/Datopotamab-Deruxtecan-Final-Overall-Survival-Results-Reported-in-Patients-with-Metastatic-HR-Positive-HER2-Low-or-Negative-Breast-Cancer-in-TROPION-Breast01-Phase-3-Trial/?feedref=JjAwJuNHiystnCoBq_hl-WBlLQO14j4QZPTyX5CYOnwC4VjN53jumZRlnzhBoBxGrCOi9QzgjCezTS3Nw_X6kJUrpSBm-Hav1w-UkdSlG3ltkRSnqzf6ourQGu_UA28CzZCGORvG0LE20YOvo49uqw==

3. AstraZeneca, Inc. 9 September 2024, TROPION-Lung01, evaluating AstraZeneca and Daiichi Sankyo’s datopotamab deruxtecan versus chemotherapy, previously met the dual primary endpoint of progression-free survival in the overall trial population [Press Release] https://www.astrazeneca.com/media-centre/press-releases/2024/dato-dxd-showed-median-overall-survival-of-146-months-in-patients-with-advanced-nsclc-in-tropion-lung01-phase-iii-trial.html

4. AstraZeneca, 8 September 2024, Novel computational pathology-based TROP2 biomarker for datopotamab deruxtecan was predictive of clinical outcomes in patients with non-small cell lung cancer in TROPION-Lung01 Phase III trial, [Press Release] https://www.astrazeneca.com/media-centre/press-releases/2024/novel-computational-pathology-based-trop2-biomarker-for-dato-dxd-was-predictive-of-clinical-outcomes-in-patients-with-nsclc-in-tropion-lung01-phase-iii-trial.html

5. Beacon by Hanson Wade, H1 2024 ADC Landscape Review, https://beacon-intelligence.com/infographic/h1-2024-adc-landscape-review/ [Downloaded Mon, Nov 4, 2024]

6. Kapil A, Spitzmüller A, Brieu N, Haneder S, Shumilov A, Meier A, Cecchi F, Barkell A, Harder N, Mittermaier K, Hidalgo-Sastre A, Alleze R, Schick M, Schmidt G, Sade H, Tsuchihashi Z, Suto F, Gustavson M, Barrett JC, Carroll D. HER2 quantitative continuous scoring for accurate patient selection in HER2 negative trastuzumab deruxtecan treated breast cancer. Sci Rep. 2024 May 27;14(1):12129. doi: 10.1038/s41598-024-61957-9. PMID: 38802399; PMCID: PMC11130140.

7. Garassino MC, Sands J, Paz-Ares L, et al. Normalized membrane ratio of TROP2 by quantitative continuous scoring is predictive of clinical outcomes in TROPION-Lung01. Presented at: 2024 IASLC World Conference on Lung Cancer; September 7-11, 2024; San Diego, CA. Abstract 4037