Cancer Imaging Archive Research Articles

Geometric and dosimetric impact of 3D generative adversarial network-based metal artifact reduction algorithm on VMAT and IMPT for the head and neck region

BackgroundWe investigated the geometric and dosimetric impact of three-dimensional (3D) generative adversarial network (GAN)-based metal artifact reduction (MAR) algorithms on volumetric-modulated arc therapy (VMAT) and intensity-modulated proton therapy (IMPT) for the head and neck region, based on artifact-free computed tomography (CT) volumes with dental fillings.MethodsThirteen metal-free CT volumes of the head and neck regions were obtained from The Cancer Imaging Archive. To simulate metal artifacts on CT volumes, we defined 3D regions of the teeth for pseudo-dental fillings from the metal-free CT volumes. HU values of 4000 HU were assigned to the selected teeth region of interest. Two different CT volumes, one with four (m4) and the other with eight (m8) pseudo-dental fillings, were generated for each case. These CT volumes were used as the Reference. CT volumes with metal artifacts were then generated from the Reference CT volumes (Artifacts). On the Artifacts CT volumes, metal artifacts were manually corrected for using the water density override method with a value of 1.0 g/cm3 (Water). By contrast, the CT volumes with reduced metal artifacts using 3D GAN model extension of CycleGAN were also generated (GAN-MAR). The structural similarity (SSIM) index within the planning target volume was calculated as quantitative error metric between the Reference CT volumes and the other volumes. After creating VMAT and IMPT plans on the Reference CT volumes, the reference plans were recalculated for the remaining CT volumes.ResultsThe time required to generate a single GAN-MAR CT volume was approximately 30 s. The median SSIMs were lower in the m8 group than those in the m4 group, and ANOVA showed a significant difference in the SSIM for the m8 group (p < 0.05). Although the median differences in D98%, D50% and D2% were larger in the m8 group than the m4 group, those from the reference plans were within 3% for VMAT and 1% for IMPT.ConclusionsThe GAN-MAR CT volumes generated in a short time were closer to the Reference CT volumes than the Water and Artifacts CT volumes. The observed dosimetric differences compared to the reference plan were clinically acceptable.

Multi-channel multi-task deep learning for predicting EGFR and KRAS mutations of non-small cell lung cancer on CT images.

Predicting the mutation statuses of 2 essential pathogenic genes [epidermal growth factor receptor (EGFR) and Kirsten rat sarcoma (KRAS)] in non-small cell lung cancer (NSCLC) based on CT is valuable for targeted therapy because it is a non-invasive and less costly method. Although deep learning technology has realized substantial computer vision achievements, CT imaging being used to predict gene mutations remains challenging due to small dataset limitations. We propose a multi-channel and multi-task deep learning (MMDL) model for the simultaneous prediction of EGFR and KRAS mutation statuses based on CT images. First, we decomposed each 3D lung nodule into 9 views. Then, we used the pre-trained inception-attention-resnet model for each view to learn the features of the nodules. By combining 9 inception-attention-resnet models to predict the types of gene mutations in lung nodules, the models were adaptively weighted, and the proposed MMDL model could be trained end-to-end. The MMDL model utilized multiple channels to characterize the nodule more comprehensively and integrate patient personal information into our learning process. We trained the proposed MMDL model using a dataset of 363 patients collected by our partner hospital and conducted a multi-center validation on 162 patients in The Cancer Imaging Archive (TCIA) public dataset. The accuracies for the prediction of EGFR and KRAS mutations were, respectively, 79.43% and 72.25% in the training dataset and 75.06% and 69.64% in the validation dataset. The experimental results demonstrated that the proposed MMDL model outperformed the latest methods in predicting EGFR and KRAS mutations in NSCLC.

Uncontrolled Confounders May Lead to False or Overvalued Radiomics Signature: A Proof of Concept Using Survival Analysis in a Multicenter Cohort of Kidney Cancer.

PurposeWe aimed to explore potential confounders of prognostic radiomics signature predicting survival outcomes in clear cell renal cell carcinoma (ccRCC) patients and demonstrate how to control for them.Materials and MethodsPreoperative contrast enhanced abdominal CT scan of ccRCC patients along with pathological grade/stage, gene mutation status, and survival outcomes were retrieved from The Cancer Imaging Archive (TCIA)/The Cancer Genome Atlas—Kidney Renal Clear Cell Carcinoma (TCGA-KIRC) database, a publicly available dataset. A semi-automatic segmentation method was applied to segment ccRCC tumors, and 1,160 radiomics features were extracted from each segmented tumor on the CT images. Non-parametric principal component decomposition (PCD) and unsupervised hierarchical clustering were applied to build the radiomics signature models. The factors confounding the radiomics signature were investigated and controlled sequentially. Kaplan–Meier curves and Cox regression analyses were performed to test the association between radiomics signatures and survival outcomes.Results183 patients of TCGA-KIRC cohort with available imaging, pathological, and clinical outcomes were included in this study. All 1,160 radiomics features were included in the first radiomics signature. Three additional radiomics signatures were then modelled in successive steps removing redundant radiomics features first, removing radiomics features biased by CT slice thickness second, and removing radiomics features dependent on tumor size third. The final radiomics signature model was the most parsimonious, unbiased by CT slice thickness, and independent of tumor size. This final radiomics signature stratified the cohort into radiomics phenotypes that are different by cancer-specific and recurrence-free survival; HR (95% CI) = 3.0 (1.5–5.7), p <0.05 and HR (95% CI) = 6.6 (3.1–14.1), p <0.05, respectively.ConclusionRadiomics signature can be confounded by multiple factors, including feature redundancy, image acquisition parameters like slice thickness, and tumor size. Attention to and proper control for these potential confounders are necessary for a reliable and clinically valuable radiomics signature.

Automated imaging-based stratification of early-stage lung cancer patients prior to receiving surgical resection using deep learning applied to CTs.

1552 Background: Computed tomography (CT) imaging is an important tool to guide further investigation and treatment in patients with lung cancer. For patients with early stage lung cancer, surgery remains an optimal treatment option. Artificial intelligence applied to pretreatment CTs may have the ability to quantify mortality risk and stratify patients for more individualized diagnostic, treatment and monitoring decisions. Methods: A fully automated, end-to-end model was designed to localize the 36cm x 36cm x 36cm space centered on the lungs and learn deep prognostic features using a 3-dimensional convolutional neural network (3DCNN) to predict 5-year mortality risk. The 3DCNN was trained and validated in a 5-fold cross-validation using 2,924 CTs of 1,689 lung cancer patients from 6 public datasets made available in The Cancer Imaging Archive. We evaluated 3DCNN’s ability to stratify stage I &amp; II patients who received surgery into mortality risk quintiles using the Cox proportional hazards model. Results: 260 of the 1,689 lung cancer patients in the withheld validation dataset were diagnosed as stage I or II, received a surgical resection within 6 months of their pretreatment CT and had known 5-year disease and survival outcomes. Based on the 3DCNN’s predicted mortality risk, patients in the highest risk quintile had a 14.2-fold (95% CI 4.3-46.8, p &lt; 0.001) increase in 5-year mortality hazard compared to patients in the lowest risk quintile. Conclusions: Deep learning applied to pretreatment CTs provides personalised prognostic insights for early stage lung cancer patients who received surgery and has the potential to inform treatment and monitoring decisions.[Table: see text]

TCGA-TCIA-Based CT Radiomics Study for Noninvasively Predicting Epstein-Barr Virus Status in Gastric Cancer.

OBJECTIVE. The purpose of this study was to investigate the value of TCGA-TCIA (The Cancer Genome Atlas and The Cancer Imaging Archive)-based CT radiomics for noninvasive prediction of Epstein-Barr virus (EBV) status in gastric cancer (GC). MATERIALS AND METHODS. A total of 133 patients with pathologically confirmed GC (94 in the training cohort and 39 in the validation cohort) who were identified from the TCGA-TCIA public data repository and two hospitals were retrospectively enrolled in the study. Two-dimensional and 3D radiomics features were extracted to construct corresponding radiomics signatures. Then, 2D and 3D nomograms were built by combining radiomics signatures and clinical information on the basis of multivariable analysis. Their performance and clinical practicability were determined, validated, and compared with respect to discrimination, calibration, reclassification, and time spent on tumor segmentation. RESULTS. Both 2D and 3D nomograms were robust and showed good calibration. The AUCs of the 2D and 3D nomograms showed no significant difference in the training cohort (0.919 vs 0.945, respectively; p = .41) or validation cohort (0.939 vs 0.955, respectively; p = .71). The net reclassification index showed that the 3D nomogram revealed no significant improvement in risk reclassification when compared with the 2D nomogram in the training cohort (net reclassification index, 0.68%; p = .14) and the validation cohort (net reclassification index, 6.06%; p = .08). Of note, the time spent on 3D segmentation (median, 907 seconds) was higher than that spent on 2D segmentation (median, 129 seconds). CONCLUSION. The 2D and 3D radiomics nomograms might have the potential to be used as effective tools for prediction of EBV in GC. When time spent on segmentation is considered, the 2D nomogram is more highly recommended for clinical application.

Deep Learning for Fully Automated Prediction of Overall Survival in Patients with Oropharyngeal Cancer Using FDG-PET Imaging.

Accurate prognostic stratification of patients with oropharyngeal squamous cell carcinoma (OPSCC) is crucial. We developed an objective and robust deep learning-based fully-automated tool called the DeepPET-OPSCC biomarker for predicting overall survival (OS) in OPSCC using [18F]fluorodeoxyglucose (FDG)-PET imaging. The DeepPET-OPSCC prediction model was built and tested internally on a discovery cohort (n = 268) by integrating five convolutional neural network models for volumetric segmentation and ten models for OS prognostication. Two external test cohorts were enrolled-the first based on the Cancer Imaging Archive (TCIA) database (n = 353) and the second being a clinical deployment cohort (n = 31)-to assess the DeepPET-OPSCC performance and goodness of fit. After adjustment for potential confounders, DeepPET-OPSCC was found to be an independent predictor of OS in both discovery and TCIA test cohorts [HR = 2.07; 95% confidence interval (CI), 1.31-3.28 and HR = 2.39; 95% CI, 1.38-4.16; both P = 0.002]. The tool also revealed good predictive performance, with a c-index of 0.707 (95% CI, 0.658-0.757) in the discovery cohort, 0.689 (95% CI, 0.621-0.757) in the TCIA test cohort, and 0.787 (95% CI, 0.675-0.899) in the clinical deployment test cohort; the average time taken was 2 minutes for calculation per exam. The integrated nomogram of DeepPET-OPSCC and clinical risk factors significantly outperformed the clinical model [AUC at 5 years: 0.801 (95% CI, 0.727-0.874) vs. 0.749 (95% CI, 0.649-0.842); P = 0.031] in the TCIA test cohort. DeepPET-OPSCC achieved an accurate OS prediction in patients with OPSCC and enabled an objective, unbiased, and rapid assessment for OPSCC prognostication.

PUM1 is upregulated by DNA methylation to suppress antitumor immunity and results in poor prognosis in pancreatic cancer.

BackgroundPancreatic carcinoma (PAAD) is a highly malignant cancer with a poor prognosis and high mortality rate. Pumilio homologous protein 1 (PUM1) promotes cell growth, invasion, and metastasis and suppresses apoptosis in many different kinds of cancers, such as non-small-cell lung carcinoma (NSCLC), ovarian cancer and lymphocyte leukemia. However, the underlying mechanism and potential role of PUM1 in PAAD have not been investigated.MethodsBioinformatics analysis was performed using multiple databases [The Cancer Genome Atlas (TCGA), Gene Expression Profiling Interactive Analysis (GEPIA), BBCancer, Human Protein Atlas (HPA), MethSurv, cBioPortal, The Cancer Imaging Archive (TCIA), xCell, Gene Expression Omnibus (GEO)] to explore the diagnostic and prognostic role of PUM1, and the relationship between expression of PUM1 and prognosis of patients with PAAD. The analysis was further validated using the Kaplan-Meier plotter.ResultsPUM1 plays a role in both diagnostic and prognostic prediction. The PUM1 mRNA expression level correlates with both the prognosis and incidence of pancreatic cancer. PUM1 can serve as a potential diagnostic indicator for pancreatic cancer. Furthermore, the DNA methylation levels of PUM1 affects its oncogene function in pancreatic cancer. PUM1 can also inhibit the immune microenvironment by altering immune cell infiltration, which affects immunotherapy response in pancreatic cancer.ConclusionsPUM1 takes a crucial part in the immune microenvironment and immunotherapy response of PAAD and is potentially useful for the development of novel diagnostic and treatment strategies.

Reproducibility of radiomic features using network analysis and its application in Wasserstein k-means clustering.

.Purpose: The goal of this study is to develop innovative methods for identifying radiomic features that are reproducible over varying image acquisition settings.Approach: We propose a regularized partial correlation network to identify reliable and reproducible radiomic features. This approach was tested on two radiomic feature sets generated using two different reconstruction methods on computed tomography (CT) scans from a cohort of 47 lung cancer patients. The largest common network component between the two networks was tested on phantom data consisting of five cancer samples. To further investigate whether radiomic features found can identify phenotypes, we propose a -means clustering algorithm coupled with the optimal mass transport theory. This approach following the regularized partial correlation network analysis was tested on CT scans from 77 head and neck squamous cell carcinoma (HNSCC) patients in the Cancer Imaging Archive (TCIA) and validated using an independent dataset.Results: A set of common radiomic features was found in relatively large network components between the resultant two partial correlation networks resulting from a cohort of lung cancer patients. The reliability and reproducibility of those radiomic features were further validated on phantom data using the Wasserstein distance. Further analysis using the network-based Wasserstein -means algorithm on the TCIA HNSCC data showed that the resulting clusters separate tumor subsites as well as HPV status, and this was validated on an independent dataset.Conclusion: We showed that a network-based analysis enables identifying reproducible radiomic features and use of the selected set of features can enhance clustering results.

Optimized convolutional neural network by firefly algorithm for magnetic resonance image classification of glioma brain tumor grade

The most frequent brain tumor types are gliomas. The magnetic resonance imaging technique helps to make the diagnosis of brain tumors. It is hard to get the diagnosis in the early stages of the glioma brain tumor, although the specialist has a lot of experience. Therefore, for the magnetic resonance imaging interpretation, a reliable and efficient system is required which helps the doctor to make the diagnosis in early stages. To make classification of the images, to which class the glioma belongs, convolutional neural networks, which proved that they can obtain an excellent performance in the image classification tasks, can be used. Convolutional network hyperparameters’ tuning is a very important issue in this domain for achieving high accuracy on the image classification; however, this task takes a lot of computational time. Approaching this issue, in this manuscript, we propose a metaheuristics method to automatically find the near-optimal values of convolutional neural network hyperparameters based on a modified firefly algorithm and develop a system for automatic image classification of glioma brain tumor grades from magnetic resonance imaging. First, we have tested the proposed modified algorithm on the set of standard unconstrained benchmark functions and the performance is compared to the original algorithm and other modified variants. Upon verifying the efficiency of the proposed approach in general, it is applied for hyperparameters’ optimization of the convolutional neural network. The IXI dataset and the cancer imaging archive with more collections of data are used for evaluation purposes, and additionally, the method is evaluated on the axial brain tumor images. The obtained experimental results and comparative analysis with other state-of-the-art algorithms tested under the same conditions show the robustness and efficiency of the proposed method.

A secure medical image transmission scheme aided by quantum representation

Quantum information and quantum image processing have made immense progress in the most recent years. The proposed paper investigates the proficiency of quantum image encryption in the field of Telehealth which can be built by incorporating quantum block-based spatial transformations. The proposed quantum cryptosystem introduced two levels of security for medical image and its sensitive content. The secrecy of medical image is reinforced by choosing a plain image based initial seed values for encryption. It is found that the suggested quantum block-based scrambling can serve as a novel method for intensifying the privacy-preserving process of medical image. It additionally ensures the medical image integrity by introducing dedicated quantum encryption for ROI based regional data. With a bilateral image transmission perspective, the proposed paper introduced InterPlanetary Features based file system for proficient retrieval of medical images. The quantitative simulation and performance evaluation done on the Cancer Imaging Archive dataset manifests that the suggested quantum-based two-level security for medical images is suitable for reinforcing the confidentiality and privacy-preserving process. It also enlarges the effectiveness of quantum image processing applications in Telemedicine.

A survey on lung CT datasets and research trends

Lung cancer is the most dangerous of all forms of cancer and it has the highest occurrence rate, world over. Early detection of lung cancer is a difficult task. Medical images generated by computer tomography (CT) are being used extensively for lung cancer analysis and research. However, it is essential to have a well-organized image database in order to design a reliable computer-aided diagnosis (CAD) tool. Identifying the most appropriate dataset for the research is another big challenge. The objective of this paper is to present a review of literature related to lung CT datasets. The Cancer Imaging Archive (TCIA) consortium collates different types of cancer datasets and permits public access through an integrated search engine. This survey summarizes the research work done using lung CT datasets maintained by TCIA. The motivation to present this survey was to help the research community in selecting the right lung dataset and to provide a comprehensive summary of the research developments in the field.

Learnable image histograms-based deep radiomics for renal cell carcinoma grading and staging

Fuhrman cancer grading and tumor-node-metastasis (TNM) cancer staging systems are typically used by clinicians in the treatment planning of renal cell carcinoma (RCC), a common cancer in men and women worldwide. Pathologists typically use percutaneous renal biopsy for RCC grading, while staging is performed by volumetric medical image analysis before renal surgery. Recent studies suggest that clinicians can effectively perform these classification tasks non-invasively by analyzing image texture features of RCC from computed tomography (CT) data. However, image feature identification for RCC grading and staging often relies on laborious manual processes, which is error prone and time-intensive. To address this challenge, this paper proposes a learnable image histogram in the deep neural network framework that can learn task-specific image histograms with variable bin centers and widths. The proposed approach enables learning statistical context features from raw medical data, which cannot be performed by a conventional convolutional neural network (CNN). The linear basis function of our learnable image histogram is piece-wise differentiable, enabling back-propagating errors to update the variable bin centers and widths during training. This novel approach can segregate the CT textures of an RCC in different intensity spectra, which enables efficient Fuhrman low (I/II) and high (III/IV) grading as well as RCC low (I/II) and high (III/IV) staging. The proposed method is validated on a clinical CT dataset of 159 patients from The Cancer Imaging Archive (TCIA) database, and it demonstrates 80% and 83% accuracy in RCC grading and staging, respectively.

Automated detection and segmentation of thoracic lymph nodes from CT using 3D foveal fully convolutional neural networks

BackgroundIn oncology, the correct determination of nodal metastatic disease is essential for patient management, as patient treatment and prognosis are closely linked to the stage of the disease. The aim of the study was to develop a tool for automatic 3D detection and segmentation of lymph nodes (LNs) in computed tomography (CT) scans of the thorax using a fully convolutional neural network based on 3D foveal patches.MethodsThe training dataset was collected from the Computed Tomography Lymph Nodes Collection of the Cancer Imaging Archive, containing 89 contrast-enhanced CT scans of the thorax. A total number of 4275 LNs was segmented semi-automatically by a radiologist, assessing the entire 3D volume of the LNs. Using this data, a fully convolutional neuronal network based on 3D foveal patches was trained with fourfold cross-validation. Testing was performed on an unseen dataset containing 15 contrast-enhanced CT scans of patients who were referred upon suspicion or for staging of bronchial carcinoma.ResultsThe algorithm achieved a good overall performance with a total detection rate of 76.9% for enlarged LNs during fourfold cross-validation in the training dataset with 10.3 false-positives per volume and of 69.9% in the unseen testing dataset. In the training dataset a better detection rate was observed for enlarged LNs compared to smaller LNs, the detection rate for LNs with a short-axis diameter (SAD) ≥ 20 mm and SAD 5–10 mm being 91.6% and 62.2% (p < 0.001), respectively. Best detection rates were obtained for LNs located in Level 4R (83.6%) and Level 7 (80.4%).ConclusionsThe proposed 3D deep learning approach achieves an overall good performance in the automatic detection and segmentation of thoracic LNs and shows reasonable generalizability, yielding the potential to facilitate detection during routine clinical work and to enable radiomics research without observer-bias.

Predicting cell invasion in breast tumor microenvironment from radiological imaging phenotypes

BackgroundThe abundance of immune and stromal cells in the tumor microenvironment (TME) is informative of levels of inflammation, angiogenesis, and desmoplasia. Radiomics, an approach of extracting quantitative features from radiological imaging to characterize diseases, have been shown to predict molecular classification, cancer recurrence risk, and many other disease outcomes. However, the ability of radiomics methods to predict the abundance of various cell types in the TME remains unclear. In this study, we employed a radio-genomics approach and machine learning models to predict the infiltration of 10 cell types in breast cancer lesions utilizing radiomic features extracted from breast Dynamic Contrast Enhanced Magnetic Resonance Imaging.MethodsWe performed a retrospective study utilizing 73 patients from two independent institutions with imaging and gene expression data provided by The Cancer Imaging Archive (TCIA) and The Cancer Genome Atlas (TCGA), respectively. A set of 199 radiomic features including shape-based, morphological, texture, and kinetic characteristics were extracted from the lesion volumes. To capture one-to-one relationships between radiomic features and cell type abundance, we performed linear regression on each radiomic feature/cell type abundance combination. Each regression model was tested for statistical significance. In addition, multivariate models were built for the cell type infiltration status (i.e. “high” vs “low”) prediction. A feature selection process via Recursive Feature Elimination was applied to the radiomic features on the training set. The classification models took the form of a binary logistic extreme gradient boosting framework. Two evaluation methods including leave-one-out cross validation and external independent test, were used for radiomic model learning and testing. The models’ performance was measured via area under the receiver operating characteristic curve (AUC).ResultsUnivariate relationships were identified between a set of radiomic features and the abundance of fibroblasts. Multivariate models yielded leave-one-out cross validation AUCs ranging from 0.5 to 0.83, and independent test AUCs ranging from 0.5 to 0.68 for the multiple cell type invasion predictions.ConclusionsOn two independent breast cancer cohorts, breast MRI-derived radiomics are associated with the tumor’s microenvironment in terms of the abundance of several cell types. Further evaluation with larger cohorts is needed.

Development and external validation of a non-invasive molecular status predictor of chromosome 1p/19q co-deletion based on MRI radiomics analysis of Low Grade Glioma patients

PurposeThe 1p/19q co-deletion status has been demonstrated to be a prognostic biomarker in lower grade glioma (LGG). The objective of this study was to build a magnetic resonance (MRI)-derived radiomics model to predict the 1p/19q co-deletion status. Method209 pathology-confirmed LGG patients from 2 different datasets from The Cancer Imaging Archive were retrospectively reviewed; one dataset with 159 patients as the training and discovery dataset and the other one with 50 patients as validation dataset.Radiomics features were extracted from T2- and T1-weighted post-contrast MRI resampled data using linear and cubic interpolation methods.For each of the voxel resampling methods a three-step approach was used for feature selection and a random forest (RF) classifier was trained on the training dataset. Model performance was evaluated on training and validation datasets and clinical utility indexes (CUIs) were computed. The distributions and intercorrelation for selected features were analyzed. ResultsSeven radiomics features were selected from the cubic interpolated features and five from the linear interpolated features on the training dataset. The RF classifier showed similar performance for cubic and linear interpolation methods in the training dataset with accuracies of 0.81 (0.75−0.86) and 0.76 (0.71−0.82) respectively; in the validation dataset the accuracy dropped to 0.72 (0.6−0.82) using cubic interpolation and 0.72 (0.6−0.84) using linear resampling. CUIs showed the model achieved satisfactory negative values (0.605 using cubic interpolation and 0.569 for linear interpolation). ConclusionsMRI has the potential for predicting the 1p/19q status in LGGs. Both cubic and linear interpolation methods showed similar performance in external validation.

RADIOMICS BASED SINGLE AND MULTI-CLASS GLIOMA CLASSIFICATION USING SUPPORT VECTOR MACHINE VARIANTS

The common type of primary brain tumor is glioma. The mortality rate of glioma patients is high due to delayed diagnosis, incorrect grading and treatment planning. Traditionally, gliomas were classified into Low Grade (grade-I and grade-II) and High Grade (grade-III and grade-IV). However, World Health Organization has insisted to classify the grades into grade-I(G-I), grade II(G-II), grade III(G-III) and grade IV(G-IV) individually to aid the physicians in clinical decision-making. Although there are limited number of studies reported to differentiate individual grades, the classification accuracy was low. Consequently, in this work single-class (G-II vs. G-III, G-II vs. G-IV and G-III vs. G-IV) and multi-class (G-II vs. G-III+IV, G-III vs. G-II+IV and G-IV vs. G-II+III) analysis was performed using specific region of tumor and whole brain as Regions of Interest(ROI) by extracting radiomic features. The images for this study (N=75) were obtained from The Cancer Imaging Archive. Further, the statistically significant features were used in the classification of individual grades by implementing variants of Support Vector Machine (SVM) algorithm: SVM, Linear-SVM and Least-Squared SVM. Among these, Linear-SVM resulted in the highest classification accuracy (&gt;80%) with average sensitivity, specificity and AUC values of &gt;70%. The comparative analysis of whole brain versus tumor ROI showed that the latter yielded better classification accuracy.

Integrative radiogenomics analysis for predicting molecular features and survival in clear cell renal cell carcinoma

Objectives: To assess the feasibility of predicting molecular characteristics by computed tomography (CT) radiomics features, and predicting overall survival (OS) using combination of omics data in clear cell renal cell carcinoma (ccRCC).Methods: Genetic data of 207 ccRCC patients was retrieved from The Cancer Genome Atlas (TCGA) and matched contrast-enhanced CT images were obtained from The Cancer Imaging Archive (TCIA). Another cohort of 175 ccRCC patients from West China Hospital was used as external validation. We first applied radiomics features and machine learning algorithms to predict genetic mutations and mRNA-based molecular subtypes. Next, we established predictive models for OS based on single omics, combined omics (radiomics+genomics, radiomics+transcriptomics, radiomics+proteomics) and all features (multi-omics).Results: Using radiomics features, random forest algorithm showed good capacity to identify the mutations VHL (AUC=0.971), BAP1 (AUC=0.955), PBRM1 (AUC=0.972), SETD2 (AUC=0.949), and molecular subtypes m1 (AUC=0.973), m2 (AUC=0.968), m3 (AUC=0.961), m4 (AUC=0.953). The TCGA cohort was divided into training (n=104) and validation (n=103) sets. The radiomics model had promising prognostic value for OS in validation set (5-year AUC=0.775) and external validation set (5-year AUC=0.755). In the validation set, the radiomics+omics models enhanced predictive accuracy than single-omics models, and the multi-omics model made further improvement (5-year AUC=0.846). High-risk group of validation set predicted by multi-omics model showed significantly poorer OS (HR=6.20, 95%CI: 3.19-8.44, p<0.0001).Conclusions: CT radiomics might be a feasible approach to predict genetic mutations, molecular subtypes and OS in ccRCC patients. Integrative analysis of radiogenomics may improve the survival prediction of ccRCC patients.

Classification of Prostate Transitional Zone Cancer and Hyperplasia Using Deep Transfer Learning From Disease-Related Images.

PurposeThe diagnosis of prostate transition zone cancer (PTZC) remains a clinical challenge due to their similarity to benign prostatic hyperplasia (BPH) on MRI. The Deep Convolutional Neural Networks (DCNNs) showed high efficacy in diagnosing PTZC on medical imaging but was limited by the small data size. A transfer learning (TL) method was combined with deep learning to overcome this challenge.Materials and methodsA retrospective investigation was conducted on 217 patients enrolled from our hospital database (208 patients) and The Cancer Imaging Archive (nine patients). Using T2-weighted images (T2WIs) and apparent diffusion coefficient (ADC) maps, DCNN models were trained and compared between different TL databases (ImageNet vs. disease-related images) and protocols (from scratch, fine-tuning, or transductive transferring).ResultsPTZC and BPH can be classified through traditional DCNN. The efficacy of TL from natural images was limited but improved by transferring knowledge from the disease-related images. Furthermore, transductive TL from disease-related images had comparable efficacy to the fine-tuning method. Limitations include retrospective design and a relatively small sample size.ConclusionDeep TL from disease-related images is a powerful tool for an automated PTZC diagnostic system. In developing regions where only conventional MR scans are available, the accurate diagnosis of PTZC can be achieved via transductive deep TL from disease-related images.

A Voxel-Based Radiographic Analysis Reveals the Biological Character of Proneural-Mesenchymal Transition in Glioblastoma.

Introduction: Proneural and mesenchymal subtypes are the most distinct demarcated categories in classification scheme, and there is often a shift from proneural type to mesenchymal subtype in the progression of glioblastoma (GBM). The molecular characters are determined by specific genomic methods, however, the application of radiography in clinical practice remains to be further studied. Here, we studied the topography features of GBM in proneural subtype, and further demonstrated the survival characteristics and proneural-mesenchymal transition (PMT) progression of samples by combining with the imaging variables.Methods: Data were acquired from The Cancer Imaging Archive (TCIA, http://cancerimagingarchive.net). The radiography image, clinical variables and transcriptome subtype from 223 samples were used in this study. Proneural and mesenchymal subtype on GBM topography based on overlay and Voxel-based lesion-symptom mapping (VLSM) analysis were revealed. Besides, we carried out the comparison of survival analysis and PMT progression in and outside the VLSM-determined area.Results: The overlay of total GBM and separated image of proneural and mesenchymal subtype revealed a correlation of the two subtypes. By VLSM analysis, proneural subtype was confirmed to be related to left inferior temporal medulla, and no significant voxel was found for mesenchymal subtype. The subsequent comparison between samples in and outside the VLSM-determined area showed difference in overall survival (OS) time, tumor purity, epithelial-mesenchymal transition (EMT) score and clinical variables.Conclusions: PMT progression was determined by radiography approach. GBM samples in the VLSM-determined area tended to harbor the signature of proneural subtype. This study provides a valuable VLSM-determined area related to the predilection site, prognosis and PMT progression by the association between GBM topography and molecular characters.

Two‐stage deep learning model for fully automated pancreas segmentation on computed tomography: Comparison with intra‐reader and inter‐reader reliability at full and reduced radiation dose on an external dataset

To develop a two-stage three-dimensional (3D) convolutional neural networks (CNNs) for fully automated volumetric segmentation of pancreas on computed tomography (CT) and to further evaluate its performance in the context of intra-reader and inter-reader reliability at full dose and reduced radiation dose CTs on a public dataset. A dataset of 1994 abdomen CT scans (portal venous phase, slice thickness≤3.75-mm, multiple CT vendors) was curated by two radiologists (R1 and R2) to exclude cases with pancreatic pathology, suboptimal image quality, and image artifacts (n=77). Remaining 1917 CTs were equally allocated between R1 and R2 for volumetric pancreas segmentation [ground truth (GT)]. This internal dataset was randomly divided into training (n=1380), validation (n=248), and test (n=289) sets for the development of a two-stage 3D CNN model based on a modified U-net architecture for automated volumetric pancreas segmentation. Model's performance for pancreas segmentation and the differences in model-predicted pancreatic volumes vs GT volumes were compared on the test set. Subsequently, an external dataset from The Cancer Imaging Archive (TCIA) that had CT scans acquired at standard radiation dose and same scans reconstructed at a simulated 25% radiation dose was curated (n=41). Volumetric pancreas segmentation was done on this TCIA dataset by R1 and R2 independently on the full dose and then at the reduced radiation dose CT images. Intra-reader and inter-reader reliability, model's segmentation performance, and reliability between model-predicted pancreatic volumes at full vs reduced dose were measured. Finally, model's performance was tested on the benchmarking National Institute of Health (NIH)-Pancreas CT (PCT) dataset. Three-dimensional CNN had mean (SD) Dice similarity coefficient (DSC): 0.91 (0.03) and average Hausdorff distance of 0.15 (0.09)mm on the test set. Model's performance was equivalent between males and females (P=0.08) and across different CT slice thicknesses (P>0.05) based on noninferiority statistical testing. There was no difference in model-predicted and GT pancreatic volumes [mean predicted volume 99cc (31cc); GT volume 101cc (33cc), P=0.33]. Mean pancreatic volume difference was -2.7cc (percent difference: -2.4% of GT volume) with excellent correlation between model-predicted and GT volumes [concordance correlation coefficient (CCC)=0.97]. In the external TCIA dataset, the model had higher reliability than R1 and R2 on full vs reduced dose CT scans [model mean (SD) DSC: 0.96 (0.02), CCC=0.995 vs R1 DSC: 0.83 (0.07), CCC=0.89, and R2 DSC:0.87 (0.04), CCC=0.97]. The DSC and volume concordance correlations for R1 vs R2 (inter-reader reliability) were 0.85 (0.07), CCC=0.90 at full dose and 0.83 (0.07), CCC=0.96 at reduced dose datasets. There was good reliability between model and R1 at both full and reduced dose CT [full dose: DSC: 0.81 (0.07), CCC=0.83 and reduced dose DSC:0.81 (0.08), CCC=0.87]. Likewise, there was good reliability between model and R2 at both full and reduced dose CT [full dose: DSC: 0.84 (0.05), CCC=0.89 and reduced dose DSC:0.83(0.06), CCC=0.89]. There was no difference in model-predicted and GT pancreatic volume in TCIA dataset (mean predicted volume 96cc (33); GT pancreatic volume 89cc (30), p=0.31). Model had mean (SD) DSC: 0.89 (0.04) (minimum-maximum DSC: 0.79 -0.96) on the NIH-PCT dataset. A 3D CNN developed on the largest dataset of CTs is accurate for fully automated volumetric pancreas segmentation and is generalizable across a wide range of CT slice thicknesses, radiation dose, and patient gender. This 3D CNN offers a scalable tool to leverage biomarkers from pancreas morphometrics and radiomics for pancreatic diseases including for early pancreatic cancer detection.