A recent study led by researchers at the Mass General Brigham Cancer Institute has identified a marker on circulating tumor cells that can help predict which lung cancer patients are likely to benefit from tarlatamab, a new immunotherapy drug. The research, published in Cancer Discovery, suggests this method could offer clinicians a noninvasive way to determine patient suitability for the treatment.
Tarlatamab received full approval in late 2025 as a therapy for small cell lung cancer (SCLC) following chemotherapy. It is an antibody designed to direct T cells toward cancer cells that display DLL3, a neuro-endocrine marker.
The study examined whether characteristics of circulating tumor cells were linked to patient responses to tarlatamab. Although clinical trials showed potential for the drug, nearly half of SCLC patients saw their disease progress within six months after starting treatment. The prevailing assumption was that all SCLC cases expressed DLL3; however, the research team found that only about half of the 20 patients studied had high levels of DLL3-positive cancer cells in their blood. Those with abundant DLL3 responded better to tarlatamab.
Testing for DLL3 on circulating tumor cells accurately identified 85% of patients who experienced clinical benefits from tarlatamab and correctly determined all patients who did not benefit (85% sensitivity and 100% specificity).
The technology used to enrich blood samples and analyze rare cancer cells has been licensed to TellBio, Inc., reflecting collaboration between bioengineering specialists and lung cancer clinicians.
"Our work may help predict which patients with SCLC are likely to respond to tarlatamab and potentially other antibodies targeting DLL3, many of which are in development," said Justin Gainor, MD, program director of the Center for Thoracic Cancers at Mass General Brigham Cancer Institute. "It also has potential implications for other cancers that express DLL3 as they become more aggressive and for the field of antibody-directed cancer therapies."
