Search

The detection of semantic and covariate out-of-distribution (OOD) examples is a critical yet overlooked challenge in digital pathology (DP). Recently, substantial insight and methods on OOD detection were presented by the ML community, but how do they fare in DP applications? To this end, we establish a benchmark study, our highlights being: 1) the adoption of proper evaluation protocols, 2) the comparison of diverse detectors in both a single and multi-model setting, and 3) the exploration into advanced ML settings like transfer learning (ImageNet vs. DP pre-training) and choice of architecture (CNNs vs. transformers). Through our comprehensive experiments, we contribute new insights and guidelines, paving the way for future research and discussion.

Nano- and microparticles are popular media to enhance optical signals, including fluorescence from a dye proximal to the particle. Here we show that homogeneous, lossless, all-dielectric spheres with diameters in the mesoscale range, between nano- ($\lesssim 100$~nm) and micro- ($\gtrsim 1$ $\mu$m) scales, can offer surprisingly large fluorescence enhancements, up to $F\sim 10^4$. With the absence of nonradiative Ohmic losses inherent to plasmonic particles, we show that $F$ can increase, decrease or even stay the same with increasing intrinsic quantum yield $q_0$, for suppressed, enhanced or intact radiative decay rates of a fluorophore, respectively. Further, the fluorophore may be located inside or outside the particle, providing additional flexibility and opportunities to design fit for purpose particles. The presented analysis with simple dielectric spheres should spur further interest in this less-explored scale of particles and experimental investigations to realize their potential for applications in imaging, molecular sensing, light coupling, and quantum information processing., Comment: main document 22 pages, 6 figures; supplementary materials 7 pages

An optical microscopic examination of thinly cut stained tissue on glass slides prepared from a FFPE tissue blocks is the gold standard for tissue diagnostics. In addition, the diagnostic abilities and expertise of any pathologist is dependent on their direct experience with common as well as rarer variant morphologies. Recently, deep learning approaches have been used to successfully show a high level of accuracy for such tasks. However, obtaining expert-level annotated images is an expensive and time-consuming task and artificially synthesized histological images can prove greatly beneficial. Here, we present an approach to not only generate histological images that reproduce the diagnostic morphologic features of common disease but also provide a user ability to generate new and rare morphologies. Our approach involves developing a generative adversarial network model that synthesizes pathology images constrained by class labels. We investigated the ability of this framework in synthesizing realistic prostate and colon tissue images and assessed the utility of these images in augmenting diagnostic ability of machine learning methods as well as their usability by a panel of experienced anatomic pathologists. Synthetic data generated by our framework performed similar to real data in training a deep learning model for diagnosis. Pathologists were not able to distinguish between real and synthetic images and showed a similar level of inter-observer agreement for prostate cancer grading. We extended the approach to significantly more complex images from colon biopsies and showed that the complex microenvironment in such tissues can also be reproduced. Finally, we present the ability for a user to generate deepfake histological images via a simple markup of sematic labels.

In clinical diagnostics and research involving histopathology, formalin fixed paraffin embedded (FFPE) tissue is almost universally favored for its superb image quality. However, tissue processing time (more than 24 hours) can slow decision-making. In contrast, fresh frozen (FF) processing (less than 1 hour) can yield rapid information but diagnostic accuracy is suboptimal due to lack of clearing, morphologic deformation and more frequent artifacts. Here, we bridge this gap using artificial intelligence. We synthesize FFPE-like images ,virtual FFPE, from FF images using a generative adversarial network (GAN) from 98 paired kidney samples derived from 40 patients. Five board-certified pathologists evaluated the results in a blinded test. Image quality of the virtual FFPE data was assessed to be high and showed a close resemblance to real FFPE images. Clinical assessments of disease on the virtual FFPE images showed a higher inter-observer agreement compared to FF images. The nearly instantaneously generated virtual FFPE images can not only reduce time to information but can facilitate more precise diagnosis from routine FF images without extraneous costs and effort., Comment: 24 pages, 6 figures, and 3 tables

Label-free imaging approaches seek to simplify and augment histopathologic assessment by replacing the current practice of staining by dyes to visualize tissue morphology with quantitative optical measurements. Quantitative phase imaging (QPI) operates with visible/UV light and thus provides a resolution matched to current practice. Here we introduce and demonstrate confocal QPI for label-free imaging of tissue sections and assess its utility for manual histopathologic inspection. Imaging cancerous and normal adjacent human breast and prostate, we show that tissue structural organization can be resolved with high spatial detail comparable to conventional H&E stains. Our confocal QPI images are found to be free of halo, solving this common problem in QPI. We further describe and apply a virtual imaging system based on Finite-Difference Time-Domain (FDTD) calculations to quantitatively compare confocal with wide-field QPI methods and explore performance limits using numerical tissue phantoms., Comment: 18 pages, 6 figures

Chemical imaging provides information about the distribution of chemicals within a target. When combined with structural information about the target, in situ chemical imaging opens the door to applications ranging from tissue classification to industrial process monitoring. The combination of infrared spectroscopy and optical microscopy is a powerful tool for chemical imaging of thin targets. Unfortunately, extending this technique to targets with appreciable depth is prohibitively slow. We combine confocal microscopy and infrared spectroscopy to provide chemical imaging in three spatial dimensions. Interferometric measurements are acquired at a small number of focal depths, and images are formed by solving a regularized inverse scattering problem. A low-dimensional signal model is key to our approach: we assume the target comprises a finite number of distinct chemical species. We establish conditions on the constituent spectra and the number of measurements needed for unique recovery of the target. Simulations illustrate imaging of cellular phantoms and sub-wavelength targets from noisy measurements.

Infrared (IR) microscopes measure spectral information that quantifies molecular content to assign the identity of biomedical cells but lack the spatial quality of optical microscopy to appreciate morphologic features. Here, we propose a method to utilize the semantic information of cellular identity from IR imaging with the morphologic detail of pathology images in a deep learning-based approach to image super-resolution. Using Generative Adversarial Networks (GANs), we enhance the spatial detail in IR imaging beyond the diffraction limit while retaining their spectral contrast. This technique can be rapidly integrated with modern IR microscopes to provide a framework useful for routine pathology., Comment: corrected typos (the word "lack" was missing in the abstract)

3D printing enables the fabrication of complex architectures by automating long sequences of additive steps. The increasing sophistication of printers, materials, and generative design promises to make geometric complexity a non-issue in manufacturing; however, this complexity can only be realized if a design can be translated into a physically executable sequence of printing operations. We investigate this planning problem for freeform direct-write assembly, in which filaments of material are deposited through a nozzle translating along a 3D path to create sparse, frame-like structures. We enumerate the process constraints for different variants of the freeform assembly process and show that, in the case where material stiffens via a glass transition, determining whether a feasible sequence exists is NP-complete. Nonetheless, for topologies typically encountered in real-world applications, finding a feasible or even optimal sequence is a tractable problem. We develop a sequencing algorithm that maximizes the fidelity of the printed part and minimizes the probability of print failure by modeling the assembly as a linear, elastic frame. We implement the algorithm and validate our approach experimentally, printing objects composed of thousands of sugar alcohol filaments with diameters of 100-200 microns. The assembly planner allows the freeform process to be applied to arbitrarily complex parts, from tissue engineering and microfluidics at the micrometer scale, to vascularized functional materials and soft robots at the millimeter scale, to structural components at the meter scale, thus opening a variety of assembly possibilities., Comment: Submitted to IEEE Transactions on Automation Science and Engineering