Dataset Information

Deep Convolutional Neural Network-Assisted Feature Extraction for Diagnostic Discrimination and Feature Visualization in Pancreatic Ductal Adenocarcinoma (PDAC) versus Autoimmune Pancreatitis (AIP).

ABSTRACT: The differentiation of autoimmune pancreatitis (AIP) and pancreatic ductal adenocarcinoma (PDAC) poses a relevant diagnostic challenge and can lead to misdiagnosis and consequently poor patient outcome. Recent studies have shown that radiomics-based models can achieve high sensitivity and specificity in predicting both entities. However, radiomic features can only capture low level representations of the input image. In contrast, convolutional neural networks (CNNs) can learn and extract more complex representations which have been used for image classification to great success. In our retrospective observational study, we performed a deep learning-based feature extraction using CT-scans of both entities and compared the predictive value against traditional radiomic features. In total, 86 patients, 44 with AIP and 42 with PDACs, were analyzed. Whole pancreas segmentation was automatically performed on CT-scans during the portal venous phase. The segmentation masks were manually checked and corrected if necessary. In total, 1411 radiomic features were extracted using PyRadiomics and 256 features (deep features) were extracted using an intermediate layer of a convolutional neural network (CNN). After feature selection and normalization, an extremely randomized trees algorithm was trained and tested using a two-fold shuffle-split cross-validation with a test sample of 20% (n = 18) to discriminate between AIP or PDAC. Feature maps were plotted and visual difference was noted. The machine learning (ML) model achieved a sensitivity, specificity, and ROC-AUC of 0.89 ± 0.11, 0.83 ± 0.06, and 0.90 ± 0.02 for the deep features and 0.72 ± 0.11, 0.78 ± 0.06, and 0.80 ± 0.01 for the radiomic features. Visualization of feature maps indicated different activation patterns for AIP and PDAC. We successfully trained a machine learning model using deep feature extraction from CT-images to differentiate between AIP and PDAC. In comparison to traditional radiomic features, deep features achieved a higher sensitivity, specificity, and ROC-AUC. Visualization of deep features could further improve the diagnostic accuracy of non-invasive differentiation of AIP and PDAC.

SUBMITTER: Ziegelmayer S

PROVIDER: S-EPMC7764649 | biostudies-literature | 2020 Dec

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Deep Convolutional Neural Network-Assisted Feature Extraction for Diagnostic Discrimination and Feature Visualization in Pancreatic Ductal Adenocarcinoma (PDAC) versus Autoimmune Pancreatitis (AIP).

Ziegelmayer Sebastian S Kaissis Georgios G Harder Felix F Jungmann Friederike F Müller Tamara T Makowski Marcus M Braren Rickmer R

Journal of clinical medicine 20201211 12

The differentiation of autoimmune pancreatitis (AIP) and pancreatic ductal adenocarcinoma (PDAC) poses a relevant diagnostic challenge and can lead to misdiagnosis and consequently poor patient outcome. Recent studies have shown that radiomics-based models can achieve high sensitivity and specificity in predicting both entities. However, radiomic features can only capture low level representations of the input image. In contrast, convolutional neural networks (CNNs) can learn and extract more co ...[more]

PMID: 33322559

Similar Datasets

Project description:Background and Objective: The emergence of the nonnutritive suck (NNS) pattern in preterm infants reflects the integrity of the brain and is used by clinicians in the neonatal intensive care unit (NICU) to assess feeding readiness and oromotor development. A critical need exists for an integrated software platform that provides NNS signal preprocessing, adaptive waveform discrimination, feature detection, and batch processing of big data sets across multiple NICU sites. Thus, the goal was to develop and describe a cross-platform graphical user interface (GUI) and terminal application known as NeoNNS for single and batch file time series and frequency-domain analyses of NNS compression pressure waveforms using analysis parameters derived from previous research on NNS dynamics. Methods. NeoNNS was implemented with Python and the Tkinter GUI package. The NNS signal-processing pipeline included a low-pass filter, asymmetric regression baseline correction, NNS peak detection, and NNS burst classification. Data visualizations and parametric analyses included time- and frequency-domain view, NNS spatiotemporal index view, and feature cluster analysis to model oral feeding readiness. Results. 568 suck assessment files sampled from 30 extremely preterm infants were processed in the batch mode (<50 minutes) to generate time- and frequency-domain analyses of infant NNS pressure waveform data. NNS cycle discrimination and NNS burst classification yield quantification of NNS waveform features as a function of postmenstrual age. Hierarchical cluster analysis (based on the Tsfresh python package and NeoNNS) revealed the capability to label NNS records for feeding readiness. Conclusions. NeoNNS provides a versatile software platform to rapidly quantify the dynamics of NNS development in time and frequency domains at cribside over repeated sessions for an individual baby or among large numbers of preterm infants at multiple hospital sites to support big data analytics. The hierarchical cluster feature analysis facilitates modeling of feeding readiness based on quantitative features of the NNS compression pressure waveform.

Project description:BackgroundThe fecal microbiota in pancreatic ductal adenocarcinoma (PDAC) and in autoimmune pancreatitis (AIP) patients remains largely unknown. We aimed to characterize the fecal microbiota in patients with PDAC and AIP, and explore the possibility of fecal microbial biomarkers for distinguishing PDAC and AIP.Methods32 patients with PDAC, 32 patients with AIP and 32 age- and sex-matched healthy controls (HC) were recruited and the fecal microbiotas were analyzed through high-throughput metagenomic sequencing. Alterations of fecal short-chain fatty acids were measured using gas chromatographic method.ResultsPrincipal coordinate analysis (PCoA) revealed that microbial compositions differed significantly between PDAC and HC samples; whereas, AIP and HC individuals tended to cluster together. Significant reduction of phylum Firmicutes (especially butyrate-producing bacteria, including Eubacterium rectale, Faecalibacterium prausnitzii and Roseburia intestinalis) and significant increase of phylum Proteobacteria (especially Gammaproteobacteria) were observed only among PDAC samples. At species level, when compared with HC samples, we revealed 24 and 12 differently enriched bacteria in PDAC and AIP, respectively. Functional analysis showed a depletion of short-chain fatty acids synthesis associated KO modules (e.g. Wood-Ljungdahl pathway) and an increase of KO modules associated with bacterial virulence (e.g. type II general secretion pathway). Consistent with the downregulation of butyrate-producing bacteria, gas chromatographic analysis showed fecal butyrate content was significantly decreased in PDAC group. Eubacterium rectale, Eubacterium ventrisum and Odoribacter splanchnicus were among the most important biomarkers in distinguishing PDAC from HC and from AIP individuals. Receiver Operating Characteristic analysis showed areas under the curve of 90.74% (95% confidence interval [CI] 86.47-100%), 88.89% (95% CI 73.49-100%), and 76.54% (95% CI 52.5-100%) for PDAC/HC, PDAC/AIP and AIP/HC, respectively.ConclusionsIn conclusion, alterations in fecal microbiota and butyrate of patients with PDAC suggest an underlying role of gut microbiota for the pathogenesis of PDAC. Fecal microbial and butyrate as potential biomarkers may facilitate to distinguish patients with PDAC from patients with AIP and HCs which worth further validation.

Project description:ObjectivesThe purpose of this study was to develop and validate an CT-based radiomics nomogram for the preoperative differentiation of focal-type autoimmune pancreatitis from pancreatic ductal adenocarcinoma.Methods96 patients with focal-type autoimmune pancreatitis and pancreatic ductal adenocarcinoma have been enrolled in the study (32 and 64 cases respectively). All cases have been confirmed by imaging, clinical follow-up and/or pathology. The imaging data were considered as: 70% training cohort and 30% test cohort. Pancreatic lesions have been manually delineated by two radiologists and image segmentation was performed to extract radiomic features from the CT images. Independent-sample T tests and LASSO regression were used for feature selection. The training cohort was classified using a variety of machine learning-based classifiers, and 5-fold cross-validation has been performed. The classification performance was evaluated using the test cohort. Multivariate logistic regression analysis was then used to develop a radiomics nomogram model, containing the CT findings and Rad-Score. Calibration curves have been plotted showing the agreement between the predicted and actual probabilities of the radiomics nomogram model. Different patients have been selected to test and evaluate the model prediction process. Finally, receiver operating characteristic curves and decision curves were plotted, and the radiomics nomogram model was compared with a single model to visually assess its diagnostic ability.ResultsA total of 158 radiomics features were extracted from each image. 7 features were selected to construct the radiomics model, then a variety of classifiers were used for classification and multinomial logistic regression (MLR) was selected to be the optimal classifier. Combining CT findings with radiomics model, a prediction model based on CT findings and radiomics was finally obtained. The nomogram model showed a good sensitivity and specificity with AUCs of 0.87 and 0.83 in training and test cohorts, respectively. The areas under the curve and decision curve analysis showed that the radiomics nomogram model may provide better diagnostic performance than the single model and achieve greater clinical net benefits than the CT finding model and radiomics signature model individually.ConclusionsThe CT image-based radiomics nomogram model can accurately distinguish between focal-type autoimmune pancreatitis and pancreatic ductal adenocarcinoma patients and provide additional clinical benefits.

Dataset Information

Deep Convolutional Neural Network-Assisted Feature Extraction for Diagnostic Discrimination and Feature Visualization in Pancreatic Ductal Adenocarcinoma (PDAC) versus Autoimmune Pancreatitis (AIP).

Publications

Deep Convolutional Neural Network-Assisted Feature Extraction for Diagnostic Discrimination and Feature Visualization in Pancreatic Ductal Adenocarcinoma (PDAC) versus Autoimmune Pancreatitis (AIP).

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets