Diagnosis with confidence: deep learning for reliable classification of laryngeal dysplasia.

BACKGROUND: Diagnosis of head and neck (HN) squamous dysplasias and carcinomas is critical for patient care, cure, and follow-up. It can be challenging, especially for grading intraepithelial lesions. Despite recent simplification in the last WHO grading system, the inter- and intraobserver variability remains substantial, particularly for nonspecialized pathologists, exhibiting the need for new tools to support pathologists. METHODS: In this study we investigated the potential of deep learning to assist the pathologist with automatic and reliable classification of HN lesions following the 2022 WHO classification system. We created, for the first time, a large-scale database of histological samples (>2000 slides) intended for developing an automatic diagnostic tool. We developed and trained a weakly supervised model performing classification from whole-slide images (WSI). We evaluated our model on both internal and external test sets and we defined and validated a new confidence score to assess the predictions that can be used to identify difficult cases. RESULTS: Our model demonstrated high classification accuracy across all lesion types on both internal and external test sets (respectively average area under the curve [AUC]: 0.878 (95% confidence interval [CI]: [0.834-0.918]) and 0.886 (95% CI: [0.813-0.947])) and the confidence score allowed for accurate differentiation between reliable and uncertain predictions. CONCLUSION: Our results demonstrate that the model, associated with confidence measurements, can help in the difficult task of classifying HN squamous lesions by limiting variability and detecting ambiguous cases, taking us one step closer to a wider adoption of AI-based assistive tools.

Subcutaneous Panniculitis-Like T-Cell Lymphoma Revealed By Immunophenotyping of Necrosis.

ABSTRACT: Subcutaneous panniculitis-like T-cell lymphoma (SPTCL) is a rare but well-defined entity, often associated with autoimmune manifestations, generally with a good prognosis unless associated with hemophagocytic syndrome. Typically, the lymphoma cells rim the adipocytes and are characterized by a CD8 + cytotoxic phenotype. We report 2 cases of SPTCL where the first biopsies only showed subcutaneous fat necrosis without any lymphoma cell visible. The diagnoses were allowed by immunophenotypic markers which characterized necrotic neoplastic T cells and confirmed on further biopsies with a typical pattern of SPTCL. These observations should prompt dermatologists to perform as large biopsies as possible, and pathologists to perform immunophenotyping in all suspected cases even if only lobular necrosis is seen morphologically.

Inactivation of kindlin-3 increases human melanoma aggressiveness through the collagen-activated tyrosine kinase receptor DDR1.

The role of the focal adhesion protein kindlin-3 as a tumor suppressor and its interaction mechanisms with extracellular matrix constitute a major field of investigation to better decipher tumor progression. Besides the well-described role of kindlin-3 in integrin activation, evidence regarding modulatory functions between melanoma cells and tumor microenvironment are lacking and data are needed to understand mechanisms driven by kindlin-3 inactivation. Here, we show that kindlin-3 inactivation through knockdown or somatic mutations increases BRAFV600mut melanoma cells oncogenic properties via collagen-related signaling by decreasing cell adhesion and enhancing proliferation and migration in vitro, and by promoting tumor growth in mice. Mechanistic analysis reveals that kindlin-3 interacts with the collagen-activated tyrosine kinase receptor DDR1 (Discoidin domain receptor 1) modulating its expression and its interaction with ß1-integrin. Kindlin-3 knockdown or mutational inactivation disrupt DDR1/ß1-integrin complex in vitro and in vivo and its loss improves the anti-proliferative effect of DDR1 inhibition. In agreement, kindlin-3 downregulation is associated with DDR1 over-expression in situ and linked to worse melanoma prognosis. Our study reveals a unique mechanism of action of kindlin-3 in the regulation of tumorigenesis mediated by the collagen-activated tyrosine kinase receptor DDR1 thus paving the way for innovative therapeutic targeting approaches in melanoma.

Idylla MSI test as a new tool for microsatellite instability detection in sebaceous tumours and keratoacanthomas.

AIM: Sebaceous tumours and keratoacanthomas can be associated with mismatch repair (MMR) deficiency and thus microsatellite instability (MSI). In such tumours, MSI phenotype could be an argument to search for an underlying Muir-Torre syndrome (MTS). MTS has been recognised as a variant of Lynch syndrome, characterised by a deficiency of the MMR proteins. In Lynch syndrome, the sensitivity and specificity of the techniques used to detect MSI is well described, which is not the case for skin tumours. In our hands, immunohistochemistry is a sensitive and specific method to detect MMR deficiency in those tumours. Contrasting with tumours of Lynch spectrum, sensitivity and specificity of molecular methods has not been extensively studied. This study aimed at evaluating two molecular methods to detect MSI phenotype in MTS associated tumours: a commonly used pentaplex PCR using Bethesda markers and the fully automated method using the Idylla MSI assay. METHODS: A comparison between PCR, and Idylla was performed on 39 DNA extracted from cutaneous tumours. Immunohistochemistry was used as the gold standard to calculate sensitivity and specificity of both molecular techniques. RESULTS: Concordant results were found in 32 cases (82%) with pentaplex PCR and in 36 cases (92%) with Idylla. The sensitivity of pentaplex PCR to detect MSI phenotype was 76% whereas Idylla sensitivity was 90%. CONCLUSION: Idylla is more performant than PCR, for the detection of MSI in MTS-associated tumours and is a reliable additional technique to help detecting MTS in these tumours.