T-cell Receptor Profiling in Early Stage Breast Cancer

Background Immunotherapies have transformed cancer treatment and how we understand and interpret the tumour microenvironment.

Furthermore, it is becoming increasingly apparent that the immune system plays a key role in the effectiveness of chemotherapy.

The Caldas laboratory has performed extensive immunogenomic analysis of patients (n=168) recruited to the TransNEO observational study.

This study takes three serial biopsies from the same tumour site of patients with early breast cancer during standard of care neoadjuvant chemotherapy treatment. The Caldas group has noted that the adaptive immune infiltrate predicts pathological complete response.

Aim To characterise key aspects of the T-cell infiltrate and assess their utility in predicting response and clinical outcomes.

Primary End Points • Characterise the T-cell receptor (TCR) repertoire of breast tumours treated with neoadjuvant therapy (NAT) at diagnosis, midway through chemotherapy and on completion of chemotherapy. • Examine whether TCR repertoire diversity at baseline correlates with pathological complete response (pathCR) and the degree of residual disease as quantified using the Residual Cancer Burden (RCB) score.

Secondary End Point • Incorporate TCR data into a novel integrated machine learning (ML) predictor of response developed by the Caldas Lab. This ML model is able to predict response to chemotherapy using DNA, RNA and digital pathology data. Methods We will sequence TCR-a and TCR-ß cDNA of each TransNEO patient's samples using digital sequencing.

The reads will be filtered, forward and reverse reads will be paired and the sequence edited to leave TCR and complementarity-determining region 3 (CDR3) sequences.

Network properties of the TCRs will be calculated to understand the specific sequences present and the proportionate number of identical reads. After enriching for patient-specific antigens, we will assign Gini scores of T cell clonality.

Analysis of variance (ANOVA) will be performed whether different RCB scores at resection are associated with different distributions of TCR clonality scores. Then the data will be integrated into a multi-variable logistic regression model for response outcomes using R.

How the Results of the Research will be Used I will accrue data and skills which will support my PhD applications at the end of this year.

This data should be published either by myself as a complete manuscript or be part of a larger, higher impact publication.

Either publication should expand the knowledge of how to interpret early breast cancer, help others to better plan treatment or trials in breast cancer, and could inform research strategy in further chemoresponsive cancers.