RESUMEN
OBJECTIVES: To investigate whether in patients undergoing surgery for oral squamous cell carcinoma, stimulated Raman histology (SRH), in comparison with H&E-stained frozen sections, can provide accurate diagnoses regarding neoplastic tissue and sub-classification of non-neoplastic tissues. MATERIALS AND METHODS: SRH, a technology based on Raman scattering, was applied to generate digital histopathologic images of 80 tissue samples obtained from 8 oral squamous cell carcinoma (OSCC) patients. Conventional H&E-stained frozen sections were then obtained from all 80 samples. All images/sections (SRH and H&E) were analyzed for squamous cell carcinoma, normal mucosa, connective tissue, muscle tissue, adipose tissue, salivary gland tissue, lymphatic tissue, and inflammatory cells. Agreement between SRH and H&E was evaluated by calculating Cohen's kappa. Accuracy of SRH compared to H&E was quantified by calculating sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) as well as area under the receiver operating characteristic curve (AUC). RESULTS: Thirty-six of 80 samples were classified as OSCC by H&E-based diagnosis. Regarding the differentiation between neoplastic and non-neoplastic tissue, high agreement between H&E and SRH (kappa: 0.880) and high accuracy of SRH (sensitivity: 100%; specificity: 90.91%; PPV: 90.00%, NPV: 100%; AUC: 0.954) were demonstrated. For sub-classification of non-neoplastic tissues, SRH performance was dependent on the type of tissue, with high agreement and accuracy for normal mucosa, muscle tissue, and salivary glands. CONCLUSION: SRH provides high accuracy in discriminating neoplastic and non-neoplastic tissues. Regarding sub-classification of non-neoplastic tissues in OSCC patients, accuracy varies depending on the type of tissue examined. CLINICAL RELEVANCE: This study demonstrates the potential of SRH for intraoperative imaging of fresh, unprocessed tissue specimens from OSCC patients without the need for sectioning or staining.
Asunto(s)
Carcinoma de Células Escamosas , Neoplasias de Cabeza y Cuello , Neoplasias de la Boca , Humanos , Carcinoma de Células Escamosas de Cabeza y Cuello , Carcinoma de Células Escamosas/patología , Neoplasias de la Boca/patología , Valor Predictivo de las PruebasRESUMEN
BACKGROUND: Mass spectrometry imaging (MSI) derives spatial molecular distribution maps directly from clinical tissue specimens and thus bears great potential for assisting pathologists with diagnostic decisions or personalized treatments. Unfortunately, progress in translational MSI is often hindered by insufficient quality control and lack of reproducible data analysis. Raw data and analysis scripts are rarely publicly shared. Here, we demonstrate the application of the Galaxy MSI tool set for the reproducible analysis of a urothelial carcinoma dataset. METHODS: Tryptic peptides were imaged in a cohort of 39 formalin-fixed, paraffin-embedded human urothelial cancer tissue cores with a MALDI-TOF/TOF device. The complete data analysis was performed in a fully transparent and reproducible manner on the European Galaxy Server. Annotations of tumor and stroma were performed by a pathologist and transferred to the MSI data to allow for supervised classifications of tumor vs. stroma tissue areas as well as for muscle-infiltrating and non-muscle infiltrating urothelial carcinomas. For putative peptide identifications, m/z features were matched to the MSiMass list. RESULTS: Rigorous quality control in combination with careful pre-processing enabled reduction of m/z shifts and intensity batch effects. High classification accuracy was found for both, tumor vs. stroma and muscle-infiltrating vs. non-muscle infiltrating urothelial tumors. Some of the most discriminative m/z features for each condition could be assigned a putative identity: stromal tissue was characterized by collagen peptides and tumor tissue by histone peptides. Immunohistochemistry confirmed an increased histone H2A abundance in the tumor compared to the stroma tissues. The muscle-infiltration status was distinguished via MSI by peptides from intermediate filaments such as cytokeratin 7 in non-muscle infiltrating carcinomas and vimentin in muscle-infiltrating urothelial carcinomas, which was confirmed by immunohistochemistry. To make the study fully reproducible and to advocate the criteria of FAIR (findability, accessibility, interoperability, and reusability) research data, we share the raw data, spectra annotations as well as all Galaxy histories and workflows. Data are available via ProteomeXchange with identifier PXD026459 and Galaxy results via https://github.com/foellmelanie/Bladder_MSI_Manuscript_Galaxy_links . CONCLUSION: Here, we show that translational MSI data analysis in a fully transparent and reproducible manner is possible and we would like to encourage the community to join our efforts.
RESUMEN
MOTIVATION: Mass spectrometry imaging (MSI) characterizes the molecular composition of tissues at spatial resolution, and has a strong potential for distinguishing tissue types, or disease states. This can be achieved by supervised classification, which takes as input MSI spectra, and assigns class labels to subtissue locations. Unfortunately, developing such classifiers is hindered by the limited availability of training sets with subtissue labels as the ground truth. Subtissue labeling is prohibitively expensive, and only rough annotations of the entire tissues are typically available. Classifiers trained on data with approximate labels have sub-optimal performance. RESULTS: To alleviate this challenge, we contribute a semi-supervised approach mi-CNN. mi-CNN implements multiple instance learning with a convolutional neural network (CNN). The multiple instance aspect enables weak supervision from tissue-level annotations when classifying subtissue locations. The convolutional architecture of the CNN captures contextual dependencies between the spectral features. Evaluations on simulated and experimental datasets demonstrated that mi-CNN improved the subtissue classification as compared to traditional classifiers. We propose mi-CNN as an important step toward accurate subtissue classification in MSI, enabling rapid distinction between tissue types and disease states. AVAILABILITY AND IMPLEMENTATION: The data and code are available at https://github.com/Vitek-Lab/mi-CNN_MSI.
Asunto(s)
Redes Neurales de la Computación , Espectrometría de MasasRESUMEN
In cancer biology, the architectural concept "form follows function" is reflected by cell morphology, migration, and epithelial-mesenchymal transition protein pattern. In vivo, features of epithelial-mesenchymal transition have been associated with tumor budding, which correlates significantly with patient outcome. Hereby, the majority of tumor buds are not truly detached but still connected to a major tumor mass. For detailed insights into the different tumor bud types and the process of tumor budding, we quantified tumor cells according to histomorphological and immunohistological epithelial-mesenchymal transition characteristics. Three-dimensional reconstruction from adenocarcinomas (pancreatic, colorectal, lung, and ductal breast cancers) was performed as published. Tumor cell morphology and epithelial-mesenchymal transition characteristics (represented by zinc finger E-box-binding homeobox 1 and E-Cadherin) were analyzed qualitatively and quantitatively in a three-dimensional context. Tumor buds were classified into main tumor mass, connected tumor bud, and isolated tumor bud. Cell morphology and epithelial-mesenchymal transition marker expression were assessed for each tumor cell. Epithelial-mesenchymal transition characteristics between isolated tumor bud and connected tumor bud demonstrated no significant differences or trends. Tumor cell count correlated significantly with epithelial-mesenchymal transition and histomorphological characteristics. Regression curve analysis revealed initially a loss of membranous E-Cadherin, followed by expression of cytoplasmic E-Cadherin and subsequent expression of nuclear zinc finger E-box-binding homeobox 1. Morphologic changes followed later in this sequence. Our data demonstrate that connected and isolated tumor buds are equal concerning immunohistochemical epithelial-mesenchymal transition characteristics and histomorphology. Our data also give an insight in the process of tumor budding. While there is a notion that the epithelial-mesenchymal transition zinc finger E-box-binding homeobox 1-E-Cadherin cascade is initiated by zinc finger E-box-binding homeobox 1, our results are contrary and outline other possible pathways influencing the regulation of E-Cadherin.
Asunto(s)
Adenocarcinoma/genética , Cadherinas/biosíntesis , Transición Epitelial-Mesenquimal/genética , Homeobox 1 de Unión a la E-Box con Dedos de Zinc/biosíntesis , Adenocarcinoma/patología , Cadherinas/genética , Línea Celular Tumoral , Movimiento Celular/genética , Proliferación Celular/genética , Femenino , Regulación Neoplásica de la Expresión Génica , Humanos , Masculino , Análisis de Regresión , Transducción de Señal/genética , Homeobox 1 de Unión a la E-Box con Dedos de Zinc/genéticaRESUMEN
Stimulated Raman Histology (SRH) employs the stimulated Raman scattering (SRS) of photons at biomolecules in tissue samples to generate histological images. Subsequent pathological analysis allows for an intraoperative evaluation without the need for sectioning and staining. The objective of this study was to investigate a deep learning-based classification of oral squamous cell carcinoma (OSCC) and the sub-classification of non-malignant tissue types, as well as to compare the performances of the classifier between SRS and SRH images. Raman shifts were measured at wavenumbers k1 = 2845 cm-1 and k2 = 2930 cm-1. SRS images were transformed into SRH images resembling traditional H&E-stained frozen sections. The annotation of 6 tissue types was performed on images obtained from 80 tissue samples from eight OSCC patients. A VGG19-based convolutional neural network was then trained on 64 SRS images (and corresponding SRH images) and tested on 16. A balanced accuracy of 0.90 (0.87 for SRH images) and F1-scores of 0.91 (0.91 for SRH) for stroma, 0.98 (0.96 for SRH) for adipose tissue, 0.90 (0.87 for SRH) for squamous epithelium, 0.92 (0.76 for SRH) for muscle, 0.87 (0.90 for SRH) for glandular tissue, and 0.88 (0.87 for SRH) for tumor were achieved. The results of this study demonstrate the suitability of deep learning for the intraoperative identification of tissue types directly on SRS and SRH images.