RESUMEN

Histopathology is a critical tool in the diagnosis and stratification of cancer. Digital Pathology involves the scanning of stained and fixed tissue samples to produce highresolution images that can be used for computer-aided diagnosis and research. A common challenge in digital pathology related to the quality and characteristics of staining, which can vary widely from center to center and also within the same institution depending on the age of the stain and other human factors. In this paper we examine the use of deep learning models for colorizing H&E stained tissue images and compare the results with traditional image processing/statistical approaches that have been developed for standardizing or normalizing histopathology images. We adapt existing deep learning models that have been developed for colorizing natural images and compare the results with models developed specifically for digital pathology. Our results show that deep learning approaches can standardize the colorization of H&E images. The performance as measured by the chi-square statistic shows that the deep learning approach can be nearly as good as current state-of-the art normalization methods.

Asunto(s)

RESUMEN

Recent cyber attacks provide evidence of increased threats to our critical systems and infrastructure. A common reaction to a new threat is to harden the system by adding new rules and regulations. As federal and state governments request new procedures to follow, each of their organizations implements their own cyber defense strategies. This unintentionally increases time and effort that employees spend on training and policy implementation and decreases the time and latitude to perform critical job functions, thus raising overall levels of stress. People's performance under stress, coupled with an overabundance of information, results in even more vulnerabilities for adversaries to exploit. In this article, we embed a simple regulatory model that accounts for cybersecurity human factors and an organization's regulatory environment in a model of a corporate cyber network under attack. The resulting model demonstrates the effect of under- and overregulation on an organization's resilience with respect to insider threats. Currently, there is a tendency to use ad-hoc approaches to account for human factors rather than to incorporate them into cyber resilience modeling. It is clear that using a systematic approach utilizing behavioral science, which already exists in cyber resilience assessment, would provide a more holistic view for decisionmakers.

RESUMEN

The current protocol for detecting and ruling out prostate cancer involves serum PSA testing followed by sampling of the prostate using a transrectal ultrasonography (TRUS)-guided biopsy. Many specialists have discussed how PSA screening has contributed to underdetection of clinically significant prostate cancer, overdiagnosis of clinically insignificant disease and poor risk stratification; however, little consideration has been given to the role of TRUS-guided biopsy in these errors. The performance of TRUS-guided biopsy is constrained by the biomechanical attributes of the sampling strategy, resulting in suboptimal detection efficiency of each core. By using a biomedical engineering approach, a uniform grid sampling strategy could be used to improve the detection efficiency of prostate biopsy. Moreover, the calibration of the sampling can be adjusted by altering the distance between needle deployments. Our model shows that for any given number of needle trajectories, a uniform grid approach will be superior to a divergent, nonuniform strategy for the detection of clinically important disease. This is an important message that should result in a move away from divergent sampling to a uniform grid approach for prostate biopsy.

Asunto(s)

RESUMEN

BACKGROUND: Analyze an approach to distributing transperineal prostate biopsy cores that yields data on the volume of a tumor that might be present when the biopsy is negative, and also increases detection efficiency. METHODS: Basic principles of sampling and probability theory are employed to analyze a transperineal biopsy pattern that uses evenly-spaced parallel cores in order to extract quantitative data on the volume of a small spherical tumor that could potentially be present, even though the biopsy did not detect it, i.e., negative biopsy. RESULTS: This approach to distributing biopsy cores provides data for the upper limit on the volume of a small, spherical tumor that might be present, and the probability of smaller volumes, when biopsies are negative and provides a quantitative basis for evaluating the effectiveness of different core spacing distances. CONCLUSIONS: Distributing transperineal biopsy cores so they are evenly spaced provides a means to calculate the probability that a tumor of given volume could be present when the biopsy is negative, and can improve detection efficiency.

Asunto(s)