Using Deep Learning to Predict Final HER2 Status in Invasive Breast Cancers That are Equivocal (2+) by Immunohistochemistry.

Invasive breast carcinomas are routinely tested for HER2 using immunohistochemistry (IHC), with reflex in situ hybridization (ISH) for those scored as equivocal (2+). ISH testing is expensive, time-consuming, and not universally available. In this study, we trained a deep learning algorithm to directly predict HER2 gene amplification status from HER2 2+ IHC slides. Data included 115 consecutive cases of invasive breast carcinoma scored as 2+ by IHC that had follow-up HER2 ISH testing. An external validation data set was created from 36 HER2 IHC slides prepared at an outside institution. All internal IHC slides were digitized and divided into training (80%), and test (20%) sets with 5-fold cross-validation. Small patches (256×256 pixels) were randomly extracted and used to train convolutional neural networks with EfficientNet B0 architecture using a transfer learning approach. Predictions for slides in the test set were made on individual patches, and these predictions were aggregated to generate an overall prediction for each slide. This resulted in a receiver operating characteristic area under the curve of 0.83 with an overall accuracy of 79% (sensitivity=0.70, specificity=0.82). Analysis of external validation slides resulted in a receiver operating characteristic area under the curve of 0.79 with an overall accuracy of 81% (sensitivity=0.50, specificity=0.82). Although the sensitivity and specificity are not high enough to negate the need for reflexive ISH testing entirely, this approach may be useful for triaging cases more likely to be HER2 positive and initiating treatment planning in centers where HER2 ISH testing is not readily available.

Assessment of HER2 using the 2018 ASCO/CAP guideline update for invasive breast cancer: a critical look at cases classified as HER2 2+ by immunohistochemistry.

In 2018, the American Society of Clinical Oncology/College of American Pathologists revised the criteria for HER2 immunohistochemistry (IHC) equivocal (2+) classification in their updated guideline. We reviewed invasive breast cancer specimens originally classified as equivocal (2+) under the 2018 guideline that underwent HER2 fluorescence in situ hybridization (FISH) testing from August 2018 to August 2019 at our Canadian reference hospital to investigate cases with ambiguous staining patterns between the 1+ and 2+ definitions. Demographics, pathologic features, and pre-analytic conditions were recorded. The H&E and corresponding HER2 IHC slides were reviewed to confirm tumor type and grade, and classify as HER2 indeterminate, 0, 1+, 2+, or "Intermediate" (staining features between the 1+ and 2+ classifications). FISH testing was performed on 289 cases and 273 met inclusion criteria. The FISH-amplified rate was 12.1%. Upon IHC review, 44.7% (122/273) of cases were reclassified as Intermediate. These cases had incomplete staining with moderate intensity (43/122, 35.3%) and/or <10% complete weak or moderate staining (102/122, 83.6%). Intermediate cases had a significantly lower frequency of amplified FISH results than 2+ cases (p < 0.0001), with only four (3.3%) FISH positive and two (1.6%) FISH heterogeneous. Our study highlights the ambiguity in the current guideline for classifying some HER2 IHC patterns. As the rate of gene amplification in these cases was low (4.9%), we recommend adhering to the 2018 HER2 2+ criteria for reflex FISH testing. However, cases with <10% moderate complete staining and certain heterogeneous patterns warrant special consideration. Further descriptive clarification of 1+ criteria is needed.

Sea urchin neural development and the metazoan paradigm of neurogenesis.

Summary:Urchin embryos continue to prove useful as a means of studying embryonic signaling and gene regulatory networks, which together control early development. Recent progress in understanding the molecular mechanisms underlying the patterning of ectoderm has renewed interest in urchin neurogenesis. We have employed an emerging model of neurogenesis that appears to be broadly shared by metazoans as a framework for this review. We use the model to provide context and summarize what is known about neurogenesis in urchin embryos. We review morphological features of the differentiation phase of neurogenesis and summarize current understanding of neural specification and regulation of proneural networks. Delta-Notch signaling is a common feature of metazoan neurogenesis that produces committed progenitors and it appears to be a critical phase of neurogenesis in urchin embryos. Descriptions of the differentiation phase of neurogenesis indicate a stereotypic sequence of neural differentiation and patterns of axonal growth. Features of neural differentiation are consistent with localized signals guiding growth cones with trophic, adhesive, and tropic cues. Urchins are a facile, postgenomic model with the potential of revealing many shared and derived features of deuterostome neurogenesis.

Neural development in Eucidaris tribuloides and the evolutionary history of the echinoid larval nervous system.

The structure and development of the larval nervous systems of all classes of echinoderms have been described and details of embryonic signaling mechanisms patterning neurogenesis have been revealed experimentally in sea urchins. Several features of neuroanatomy and neural development indicate that echinoids are the most derived group. Here we describe the development and organization of the nervous system of a cidaroid, Eucidaris tribuloides. The cidaroids are one of two major clades of echinoids, and are considered to have features of anatomy and development that represent the common ancestor to all echinoids. The embryos of E. tribuloides lack a thickened animal plate and serotonergic neurons arise laterally, associated with the ciliary band. Although lacking a discrete apical organ, plutei have serotonergic neurons associated with the pre-oral ciliary band joined by a few diffusely arranged connecting axons. Chordin and Hnf6, early markers for oral ectoderm and ciliary band, are expressed in similar patterns to euechinoids. However, an animal pole domain marker, Nk2.1, is expressed in a broader region of anterior ectoderm than in euechinoids. Six3, a proneural marker that is restricted to the animal plate of euechinoids, is expressed laterally in the preoral ciliary band at the same location as the serotonergic neurons. We conclude that the organization and development of the larval nervous system of E. tribuloides retains features shared with other echinoderm larvae, but not with euechinoids. These data support a model in which several distinctive features of euechinoid neural organization are derived, having arisen after the divergence of the two clades of echinoids about 265 million years ago. We hypothesize that differences in the developmental mechanisms that restrict neurogenesis to the animal pole forms the basis for the distinctive neuroanatomy of euechinoids.