Abstract Background: Epithelial-mesenchymal transition (EMT) is associated with metastasis in human prostate cancer (PCa). As metastatic tumors are the major cause of death in most cancer patients, identifying instances of EMT in patients may serve as a good prognostic indicator of progressive disease. Abnormal morphologic deformation of the cell nucleus is strongly associated with an aggressive cancer phenotype and computer-aided quantitative nuclear morphometry (QNM) has been used to quantify this deformation. In the present study, we go further to investigate the association of nuclear morphologic features between an epithelial PCa phenotype and an EMT-induced mesenchymal PCa phenotype across a panel of extracellular matrix (ECM) substrates on which cells are grown in vitro. Methods: PC3 epithelial cells (PC3-Epi) are established human PCa cultures from bone metastatic lesions. M2 macrophages induced PC3-Epi cells to undergo EMT and become mesenchymal PC3 cells (PC3- EMT). PC3-Epi and PC3-EMT cells were cultured independently on glass tissue-culture slides and also on a panel of ECM component substrates, including collagen 1, poly-D-Lysine, and various molecular weight hyaluronic acid polymers. Slides were stained with Feulgen or hematoxylin and eosin (H&E). The images were digitized with a slide-scanning microscope. Computer aided image processing software measured a total of 41 non-redundant features. This included nuclear size, shape, and texture features. Features were statistically analyzed using both supervised and unsupervised machine learning algorithms. Results: Statistical classification between PC3-Epi and PC3-EMT was performed on a panel of ECM substrates. The PC3-Epi and PC3-EMT morphological features were characterized by the type of ECM substrate on which cells were cultured. Chromatin texture and shape features, such as Feret diameter, were among the most significant in contributing to separation of the PC3-Epi and PC3-EMT. In contrast, the morphological features could not be separated when cells were grown on glass surfaces. Conclusions: The PC3-Epi and PC3-EMT cells were statistically separated more effectively with extracellular matrix substrates than with glass. Furthermore, our findings suggest that ECM components affect chromatin texture, a function of chromatin organization, and nuclear morphology. Our results also indicate quantitative nuclear morphometric analysis may be a viable tool for demonstrating instances of EMT in clinical samples, potentially adding quantitative prognostic value to this technique. Nuclear morphometric factors, such as those used here, have also found applications to distinguish aggressive human prostate cancers in fixed biopsy samples (See review: Veltri, R.W., et al. “Nuclear morphometry, nucleomics and prostate cancer progression”. Asian Journal of Andrology (2012) 14, 375-384.) Citation Format: James E. Verdone, Robert W. Veltri, Steven M. Mooney, James R. Hernandez, Calvin A. Harberg, Donald S. Coffey, Kenneth J. Pienta. Extracellular matrix components direct chromatin texture and nuclear morphological changes in epithelial-mesenchymal transition. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 1139. doi:10.1158/1538-7445.AM2014-1139