T staging esophageal tumors with x rays.

With histopathology results typically taking several days, the ability to stage tumors during interventions could provide a step change in various cancer interventions. X-ray technology has advanced significantly in recent years with the introduction of phase-based imaging methods. These have been adapted for use in standard labs rather than specialized facilities such as synchrotrons, and approaches that enable fast 3D scans with conventional x-ray sources have been developed. This opens the possibility to produce 3D images with enhanced soft tissue contrast at a level of detail comparable to histopathology, in times sufficiently short to be compatible with use during surgical interventions. In this paper we discuss the application of one such approach to human esophagi obtained from esophagectomy interventions. We demonstrate that the image quality is sufficiently high to enable tumor T staging based on the x-ray datasets alone. Alongside detection of involved margins with potentially life-saving implications, staging tumors intra-operatively has the potential to change patient pathways, facilitating optimization of therapeutic interventions during the procedure itself. Besides a prospective intra-operative use, the availability of high-quality 3D images of entire esophageal tumors can support histopathological characterization, from enabling "right slice first time" approaches to understanding the histopathology in the full 3D context of the surrounding tumor environment.

Enhanced detection of threat materials by dark-field x-ray imaging combined with deep neural networks.

X-ray imaging has been boosted by the introduction of phase-based methods. Detail visibility is enhanced in phase contrast images, and dark-field images are sensitive to inhomogeneities on a length scale below the system's spatial resolution. Here we show that dark-field creates a texture which is characteristic of the imaged material, and that its combination with conventional attenuation leads to an improved discrimination of threat materials. We show that remaining ambiguities can be resolved by exploiting the different energy dependence of the dark-field and attenuation signals. Furthermore, we demonstrate that the dark-field texture is well-suited for identification through machine learning approaches through two proof-of-concept studies. In both cases, application of the same approaches to datasets from which the dark-field images were removed led to a clear degradation in performance. While the small scale of these studies means further research is required, results indicate potential for a combined use of dark-field and deep neural networks in security applications and beyond.