The transition from benign to tumor tissue is poorly understood. Determining the molecular events that underpin tumor transformation will be fundamental to improving the early diagnosis of cancer. In our paper presented at AUA2022 and recently published in Nature, we report the first step to generate a view of the genome integrity in situ using spatial transcriptomics (video available at https://youtube/YdzF0-PFXhc).¹ This approach provides a high-resolution map of the entire transcriptome over thousands of areas (spots) at near–single cell resolution. We applied this approach to 10 patients’ tissues, including 3 whole axial sections of prostates, to study clonal dynamics in heterogeneous multifocal prostate cancer.

Using the spatial transcriptomic information as input, we developed a computational framework to detect gains and losses of genomic information at near–single cell resolution. Surprisingly, not only were we able to identify distinct clonal patterns within tumors, but also in nearby benign epithelia. This allowed us to construct evolutionary clone trees from healthy benign tissue, through altered benign to multiclonal tumors (see Figure). We were able to corroborate these findings in a separate patient’s cutaneous squamous cell carcinoma which shared copy-number amplifications with nearby benign squamous epithelial cells.

This unsupervised approach captures molecular and spatial continuums in a tissue context. Future studies could use this approach to determine primary clonal populations with metastatic potential, thereby informing prognosis and treatment selection. Further work will also investigate events leading to prostatic transformation, providing further insights for early detection and future screening trials.