Search

The development of next-generation sequencing (NGS) has enabled the discovery of cancer-specific driver gene alternations, making precision medicine possible. However, accurate genetic testing requires a sufficient amount of tumor cells in the specimen. The evaluation of tumor content ratio (TCR) from hematoxylin and eosin (H&E)-stained images has been found to vary between pathologists, making it an important challenge to obtain an accurate TCR. In this study, three pathologists exhaustively labeled all cells in 41 regions from 41 lung cancer cases as either tumor, non-tumor or indistinguishable, thus establishing a "gold standard" TCR. We then compared the accuracy of the TCR estimated by 13 pathologists based on visual assessment and the TCR calculated by an AI model that we have developed. It is a compact and fast model that follows a fully convolutional neural network architecture and produces cell detection maps which can be efficiently post-processed to obtain tumor and non-tumor cell counts from which TCR is calculated. Its raw cell detection accuracy is 92% while its classification accuracy is 84%. The results show that the error between the gold standard TCR and the AI calculation was significantly smaller than that between the gold standard TCR and the pathologist's visual assessment ( p < 0.05 ). Additionally, the robustness of AI models across institutions is a key issue and we demonstrate that the variation in AI was smaller than that in the average of pathologists when evaluated by institution. These findings suggest that the accuracy of tumor cellularity assessments in clinical workflows is significantly improved by the introduction of robust AI models, leading to more efficient genetic testing and ultimately to better patient outcomes. [ABSTRACT FROM AUTHOR]

Abstract: Five bacterial strains capable of utilizing 17β-estradiol (E2) and estrone (E1) were isolated from soil samples. Using their morphological and physiological features and 16S rDNA sequences, we classified these isolates into two groups: Group A (Rhodococcus sp. strains ED6, ED7, and ED10) and Group B (Sphingomonas sp. strains ED8 and ED9). All isolates used E2 and E1 as the sole carbon sources and showed high E1 and E2 degradation activities. In all strains, more than 50% of 0.8 mg of E1 or E2 was degraded in 4 mL of inorganic medium over 24 h, and 90% was degraded over 120 h. By incubating the resting ED8 cells with E2 and the meta-cleavage inhibitor 3-chlorocatechol, we identified two metabolites, 4-hydroxyestrone (4-OH-E1) and 4-hydroxyestradiol (4-OH-E2), and confirmed their identity using authentic chemicals. The 4-OH-E1 and 4-OH-E2 compounds were assumed to be intermediate metabolites formed before meta-cleavage, as they were not identified in culture without 3-chlorocatechol. Degradation of E2 by strain ED8 can be initiated by hydroxylation of the C-4 position, followed by meta-cleavage of the benzene ring. When strains ED8 degraded E2, we further identified hydroxy-E2, keto-E1 and -E2, and an additional degradation product via mass spectrometry. The presence of these compounds implied degradation through a second pathway initiated through an attack of the saturated ring. [Copyright &y& Elsevier]

Abstract: Each chromosome occupies its own-specific space called a ‘territory’ within the interphase nucleus, and the arrangement of chromosome territories (CTs) is important in epigenetic mechanisms. The molecular mechanism to determine the positioning of CTs, however, remains unknown. On the other hand, dioxin is known to be the typical environmental pollutant that affects a wide variety of biological events in many species. Here, we show that dioxin enlarges the minimum distance between chromosome 12 and chromosome 16 territories in human preadipocyte cells, and the alteration of chromosome positioning is canceled by an aryl hydrocarbon receptor (AhR) antagonist α-naphthoflavone. Thus, AhR may be a key molecule to regulate chromosome positioning. Our results suggest a novel effect of dioxin toxicity, and demonstrate a clue to reveal the novel molecular mechanism for the arrangement of CTs. [Copyright &y& Elsevier]

Here, we propose an algorithm to automatically obtain extraction filters for the affected regions from cancer images. The proposed algorithm consists of two steps: extraction of affected region candidates and elimination of false positives. Useful features of cancer images, such as the area and degree of circularity of cancer nests, etc., are extracted using the derived filters. These features are useful for supporting pathological diagnosis. Automatic Construction of Tree-structural Image Transformation (ACTIT) was used to construct these filters to extract the affected regions from the image. The proposed algorithm was applied to a mouth cancer pathological image. The results confirmed that the proposed algorithm can obtain good filters capable of extracting cancer nests. The derived filters were also applied to other images from the same specimen. The results also indicated that the generated filters show general versatility in extracting cancer nest candidates. The area and degree of circularity of the cancer nets were also derived automatically.