Development of a deep-learning model tailored for HER2 detection in breast cancer to aid pathologists in interpreting HER2-low cases.

AIMS: Over 50% of breast cancer cases are "Human epidermal growth factor receptor 2 (HER2) low breast cancer (BC)", characterized by HER2 immunohistochemistry (IHC) scores of 1+ or 2+ alongside no amplification on fluorescence in situ hybridization (FISH) testing. The development of new anti-HER2 antibody-drug conjugates (ADCs) for treating HER2-low breast cancers illustrates the importance of accurately assessing HER2 status, particularly HER2-low breast cancer. In this study we evaluated the performance of a deep-learning (DL) model for the assessment of HER2, including an assessment of the causes of discordances of HER2-Null between a pathologist and the DL model. We specifically focussed on aligning the DL model rules with the ASCO/CAP guidelines, including stained cells' staining intensity and completeness of membrane staining. METHODS AND RESULTS: We trained a DL model on a multicentric cohort of breast cancer cases with HER2-IHC scores (n = 299). The model was validated on two independent multicentric validation cohorts (n = 369 and n = 92), with all cases reviewed by three senior breast pathologists. All cases underwent a thorough review by three senior breast pathologists, with the ground truth determined by a majority consensus on the final HER2 score among the pathologists. In total, 760 breast cancer cases were utilized throughout the training and validation phases of the study. The model's concordance with the ground truth (ICC = 0.77 [0.68-0.83]; Fisher P = 1.32e-10) is higher than the average agreement among the three senior pathologists (ICC = 0.45 [0.17-0.65]; Fisher P = 2e-3). In the two validation cohorts, the DL model identifies 95% [93% - 98%] and 97% [91% - 100%] of HER2-low and HER2-positive tumours, respectively. Discordant results were characterized by morphological features such as extended fibrosis, a high number of tumour-infiltrating lymphocytes, and necrosis, whilst some artefacts such as nonspecific background cytoplasmic stain in the cytoplasm of tumour cells also cause discrepancy. CONCLUSION: Deep learning can support pathologists' interpretation of difficult HER2-low cases. Morphological variables and some specific artefacts can cause discrepant HER2-scores between the pathologist and the DL model.

Rare-Earth-Mediated Optomechanical System in the Reversed Dissipation Regime.

Strain-mediated interaction between phonons and telecom photons is demonstrated using excited states of erbium ions embedded in a mechanical resonator. Owing to the extremely long-lived nature of rare-earth ions, the dissipation rate of the optical resonance falls below that of the mechanical one. Thus, a "reversed dissipation regime" is achieved in the optical frequency region. We experimentally demonstrate an optomechanical coupling rate g_{0}=2π×21.7 Hz, and numerically reveal that the interaction causes stimulated excitation of erbium ions. Numerical analyses further indicate the possibility of g_{0} exceeding the dissipation rates of erbium and mechanical systems, thereby leading to single-photon strong coupling. This strain-mediated interaction, moreover, involves the spin degree of freedom, and has a potential to be extended to highly coherent opto-electro-mechanical hybrid systems in the reversed dissipation regime.

When LC-HRMS metabolomics gets ISO17025 accredited and ready for official controls - application to the screening of forbidden compounds in livestock.

Within the particular context of controlling chemical residues in food, an alternative to targeted approaches has emerged; it consists in the characterisation of physiological perturbations induced upon exposure of animals to a given chemical substance/class of substances to highlight suitable biomarkers addressing safety and/or regulatory issues. Metabolomics in particular has been investigated in the hope of identifying such biomarkers, and a range of studies have demonstrated the efficiency of the strategy. Until very recently, steps remained to be taken towards official or commercial implementation of corresponding tools. In particular, the lack of guidelines and criteria to validate such methods that do not target specific chemical species per se, constituted a bottleneck. In the present work, a metabolomics model dedicated to the detection of ß-agonist administration in bovines has been developed and fully validated; criteria (selectivity, robustness, stability, suspicion threshold definition, false positive and false negative rates) have been proposed in agreement with EU expectations (Dec 2002/657), enabling demonstration that performances comply with screening requirements. Although some of the biomarkers involved in the prediction model remain un-elucidated, the corresponding LC-HRMS method has recently been ISO17025 accredited, allowing for the very first official implementation of a metabolomics based strategy within French National Monitoring Plans.