Project Summary Pancreatic cancer is a deadly disease with a dismal prognosis and a high frequency of distant metastasis. One likely reason for the frequent metastases in pancreatic cancer is the ability of malignant cells in the primary tumor to invade veins, providing direct access to the liver. While this intravasation is a key first step in metastasis, little is known about the molecular and cellular alterations that underlie this process. Foci of cancer cell intravasation are small and complex, requiring high-resolution three-dimensional (3D) approaches to study them. In this Research Test Bed of our Center for 3D Multiscale Cancer Imaging, we will comprehensively characterize the morphological, cellular, and molecular alterations of foci of intravasation in 3D in human pancreatic cancer samples. In collaboration with Technology Development Unit 1, we will employ our novel deep learning 3D reconstruction approach CODA to serially sectioned pancreatic cancer samples with venous invasion to quantify morphological and cellular features of intravasation. We will also perform multi-dimensional immune profiling of these intravasation foci in 3D using imaging mass cytometry, which can label with 40 antibodies on each tissue section. These analyses will allow us to compare the tumor microenvironment in foci of intravasation to that in other areas of cancer and in uninvolved veins. In collaboration with Technology Development Unit 2, we will analyze the gene expression changes in these intravasation foci by applying a newly developed spatial transcriptomics/proteomics approach (DBit-Seq) to pancreatic cancer samples previously reconstructed by CODA. This approach will provide the first 3D multi-scale models of venous invasion in human pancreatic cancer, providing unprecedented morphological, cellular, and molecular detail of the process of intravasation and initiation of metastasis. In addition, our Research Test Bed will provide a critical opportunity to implement, validate, and iteratively improve the novel approaches from the two Technology Development Centers. We will work closely with these Technology Development Centers in order to evaluate the performance of CODA and DBit-Seq in our human pancreatic tissue samples, allowing validation of the cell type identification and molecular alterations by complementary techniques. The studies in our Research Test Bed will promote significant improvement of the CODA and DBit-Seq imaging platforms while also providing important biological insights into the initiation of metastasis in pancreatic cancer.