Abstract Cell morphologies represent a phenotypic integration of intrinsic cellular processes, changes in cell state, and interactions with neighboring cells. While analyses of large cell line collections such as the Cancer Dependency Map (DepMap) have greatly enhanced discovery of novel molecular biomarkers and therapeutic targets, morphological analysis of these models has not yet been performed. Hence, there is a rich opportunity to discover novel regulators and functional properties of cell morphologies by linking image-based profiling to -omics based assays. Here, we studied the transcriptional and functional properties of 16 human pancreatic cancer cell lines, associations with distinct cell morphologies, and treatment-induced remodeling. At baseline, we observed extensive morphological heterogeneity across and within cell lines, as well as differences in multicellular organization. We categorized cell line morphologies into epithelial, mesenchymal, and spheroid, and organizational patterns into tightly aggregated, multilayered, and dispersed. We performed differential expression analysis across morphological subtypes utilizing DepMap transcriptomics data and queried functional correlates including CRISPR dependency, drug sensitivity, and proclivity for metastasis. For example, we discovered that the mesenchymal morphology and multilayered organizational pattern [WH3] were associated with metastatic potential, while expressing reduced levels of tight junction and cell adhesion genes. We then explored treatment associated morphological changes and their potential as cost-effective biomarkers for cell-intrinsic resistance to chemotherapy (e.g., 5-FU and gemcitabine) and KRAS inhibitors (e.g., MRTX1133 and RMC- 6236). Cells were treated continuously and monitored with live cell imaging (Incucyte SX5) followed by Cell Painting at the three-day endpoint. We observed conserved morphological responses to therapy across cell lines including a shift towards spindle-like cell shapes with neurite-like projections in chemotherapy-treated conditions, and a cell bloating phenomena characterized by an increase in cell area in KRAS inhibitor treated conditions. To directly link transcriptional state to morphology, we performed 1000-plex spatial molecular imaging (Nanostring CosMx) at subcellular resolution to pools of cell lines, demonstrating that morphological diversity within cell lines corresponded to distinct transcriptional states. In conclusion, this study highlights the potential of harnessing cell morphological information in a rapid, cost-effective phenotyping assay to aid precision oncology efforts leveraging patient- derived in vitro models. Citation Format: Dennis Gong, William L Hwang. Morphological diversity predicts functional traits in pancreatic cancer [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Advances in Pancreatic Cancer Research; 2024 Sep 15-18; Boston, MA. Philadelphia (PA): AACR; Cancer Res 2024;84(17 Suppl_2):Abstract nr B076.