Using a computational biomarker, the companies say they can identify which patients would derive significant clinical benefit from their experimental antibody-drug conjugate. AstraZeneca and Roche are co-developing and commercializing a companion diagnostic for the biomarker.
AstraZeneca and Daiichi Sankyo unveiled a novel AI-powered biomarker on Sunday which can predict clinical outcomes in non-small cell lung cancer patients treated with the partners’ investigational antibody-drug conjugate datopotamab deruxtecan.
Using this computational biomarker—called the normalized membrane ratio (NMR) of the TROP2 protein by quantitative continuous scoring (QCS), or TROP2-QCS for short—AstraZeneca and Daiichi Sankyo were able to identify which patients would derive significant clinical benefit from datopotamab deruxtecan (Dato-DXd).
The partners conducted an exploratory analysis of their Phase III TROPION-Lung01 study and found that in patients positive for TROP2-QCS, Dato-DXd led to a 43% drop in the risk of disease progression or death versus docetaxel. This effect was statistically significant, according to the companies. Median progression-free survival (PFS) was 6.9 months in the Dato-DXd arm, versus 4.1 months in docetaxel comparators.
By contrast, in the primary analysis of the overall study population—without taking into account TROP2-QCS—Dato-DXd could only lower the risk of disease progression or death by 25% versus docetaxel.
Ken Takeshita, Daiichi Sankyo’s global head of R&D, in a statement said that the findings from QCS analysis “support the potential of TROP2, as measured by quantitative continuous scoring, as a predictive biomarker” for Dato-DXd. The new AI biomarker can also “begin to answer the question of why certain patients with non-small cell lung cancer (NSCLC) respond better to treatment,” Takeshita said.
The partners unveiled TROP2-QCS at the International Association for the Study of Lung Cancer’s 2024 World Conference on Lung Cancer. AstraZeneca will work with Roche Tissue Diagnostics business to collaboratively develop and commercialize the AI biomarker, the pharma announced.
While most biomarkers typically rely on the simple expression levels of certain proteins in a cancer cell, TROP2-QCS instead looks at the amount of TROP2 expressed on tumor cells’ membrane relative to the protein’s expression in the cytoplasm. Patients are then classified as TROP2-QCS-positive if at least 75% of their tumor cells have an NMR below a prespecified value, which is indicative of a higher TROP2 expression in the cytoplasm.
TROP2-QCS was developed using AstraZeneca’s QCS “fully automated computational pathology” platform, which uses AI to generate biomarker data beyond the simple presence or absence of a protein. According to the pharma’s website, QCS also determines the intensity of the signal and its localization in the cell.
AstraZeneca and Daiichi Sankyo used QCS to explain the failure of TROPION-Lung01, which in May 2024 showed that Dato-DXd failed to significantly improve overall survival in the overall trial population. At the time, the partners zeroed in on patients with non-squamous disease, in whom the antibody-drug conjugate resulted in a “clinically meaningful improvement” in overall survival.
According to the QCS analysis on Sunday, “a greater proportion” of patients with non-squamous NSCLC were also TROP2-QCS-positive—66% versus the 44% biomarker positivity in patients with squamous disease.
With data from TROPION-Lung01, AstraZeneca and Daiichi Sankyo have filed for Dato-DXd’s approval in previously treated patients with non-squamous NSCLC. The FDA’s decision is due in the fourth quarter of this year.
