Did the tumour-specific T cells make it into the product? DETECT as a quality check for adoptive cell therapy

When you expand a patient's own T cells into a cell therapy product, how do you know the tumour-reactive clones survived manufacturing? A case study presented at SITC 2026 and published in the Journal for ImmunoTherapy of Cancer by Vera Donnenberg, Albert Donnenberg and colleagues put DETECT to work on exactly that question.

The team is developing an adoptive cellular therapeutic made from pleural infiltrating T cells, for patients with malignant pleural effusions. For one breast cancer patient, they sequenced the T cell receptors in the uncultured effusion and in the product generated from it, then used DETECT to predict the target epitope and MHC for 65,535 unique CDR3 sequences.

Keeping only the top 1% of predictions by confidence score and clone frequency left ten epitopes: five from mutant K-RAS and p65, and five from Epstein-Barr virus and SARS-CoV-2. In every case, the MHC predicted by DETECT matched the patient's own molecularly typed MHC. And the predicted tumour-specific clones were significantly amplified in the final product.

That is the kind of readout a cell therapy quality check needs. From TCR sequencing data alone, the team could see whether tumour-directed clones were present in the starting material and whether they were enriched after expansion. The authors describe it as a potentially economical way to infer the presence of tumour-specific T cells in cellular therapeutics.

The authors note that the predicted K-RAS and p65 mutations still need to be confirmed by whole genome sequencing of the tumour.