Biological correlates associated with high-risk breast cancer patients identified using a computational method.

Using a novel unsupervised method to integrate multi-omic data, we previously identified a breast cancer group with a poor prognosis. In the current study, we characterize the biological features of this subgroup, defined as the high-risk group, using various data sources. Assessment of three published hypoxia signatures showed that the high-risk group exhibited higher hypoxia scores (p < 0.

Tumor aware recurrent inter-patient deformable image registration of computed tomography scans with lung cancer.

Background: Voxel-based analysis (VBA) for population level radiotherapy (RT) outcomes modeling requires topology preserving inter-patient deformable image registration (DIR) that preserves tumors on moving images while avoiding unrealistic deformations due to tumors occurring on fixed images.

Purpose: We developed a tumor-aware recurrent registration (TRACER) deep learning (DL) method and evaluated its suitability for VBA.

Methods: TRACER consists of encoder layers implemented with stacked 3D convolutional long short term memory network (3D-CLSTM) followed by decoder and spatial transform layers to compute dense deformation vector field (DVF).

Normal tissue transcriptional signatures for tumor-type-agnostic phenotype prediction.

Cancer transcriptional patterns reflect both unique features and shared hallmarks across diverse cancer types, but whether differences in these patterns are sufficient to characterize the full breadth of tumor phenotype heterogeneity remains an open question. We hypothesized that these shared transcriptomic signatures reflect repurposed versions of functional tasks performed by normal tissues. Starting with normal tissue transcriptomic profiles, we use non-negative matrix factorization to derive six distinct transcriptomic phenotypes, called archetypes, which combine to describe both normal tissue patterns and variations across a broad spectrum of malignancies.