In this IRB approved retrospective study 41 women with biopsy-proven invasive breast cancers (IBC) were imaged using contrast-enhanced mammography (CEM), prior to any treatment. Size-matched regions of interest (ROIs) were manually contoured by an experienced breast radiologist on the CEM capturing the breast lesion and breast parenchymal enhancement (BPE), respectively. Radiomics analysis was performed using LifEx software and 109 radiomics metrics spanning 6 different texture families were extracted from each ROI. Predictive models of lesion malignancy were developed using multiple classifiers and used to subclassify breast cancers based on their hormone receptor status. The 10- fold cross validation was used to construct the decision classifier and performance was assessed. CEM radiomics models based on Random Forest, Real Adaboost, and ElasticNet classifiers achieved an AUC of 0.83, 0.82 and 0.74, respectively in discriminating malignant breast lesions from varying amounts of BPE. Accounting for the varying levels of BPE, revealed a reduction in AUC-based prediction of lesion vs. BPE as the qualitative assessment of BPE increased from minimal to moderate (AUCs of 0.89 vs 0.74). Further analyses of the IBC based on their hormone receptor status showed that triple negative breast lesions showed statistically significant differences in multiple radiomics metrics compared to ER+ PR+ HER2- and HER2+. The predicted probability of the radiomics model was significantly different across three receptor-based subtypes and between high and low nuclear grade breast cancers. CEM Radiomics demonstrated good discrimination (AUC>0.8) of malignant breast lesions despite varying BPE levels and supports breast lesion subtyping.
In this Institutional Review Board (IRB) approved, Health Insurance Portability and Accountability Act (HIPAA) compliant, prospective study, uncompressed envelope data (RF data) were collected from 100 patients with focal renal masses using an RS80A ultrasound scanner with B-mode and CEUS. By summing and averaging the Nakagami images formed using sliding windows, we use the average ‘m’ to stratify manually segmented masses, using data from both the B-mode and CEUS scans. Wilcoxon rank sum test using an alpha value of 0.05 was used detect differences between the groups. Logistic regression was used for classification and the area under the receiver operator curve (AUC) was used to assess performance. Among the 100 masses, 40 were benign, 37 were malignant based on histopathology, and 23 were radiologically and clinically presumed malignant but with no pathological proof at the time of data analysis. Univariate analyses showed significant (p<0.01) differences between the benign and non-benign masses on both B-mode and CEUS, with non-benign masses having smaller ‘m’. Predictive models constructed using Nakagami parameters extracted from Bmode and CEUS-based RF scans showed an AUC of 0.67 95% CI: (0.56, 0.78) and 0.61 95% CI: (0.5, 0.73), respectively for discriminating benign from non-benign renal masses. The concordance between the two assessments was 95%. We present a framework for characterizing images using speckle textural properties, for example Nakagami analysis, to aid in objective tissue characterization using ultrasound.
t-Distributed Stochastic Neighbor Embedding (t-SNE) and k-means have been increasingly utilized for dimension reduction and graphical illustration in medical imaging (e.g., CT) informatics. Mapping a grid network onto a slide is a prerequisite for implementing cluster analysis. Traditionally, the performance of cluster analysis is driven by hyperparameters, however, grid size which also affects performance is often set arbitrarily. In this study, we evaluated the effect of varying grid sizes, perplexity and learning rate hyperparameters for unsupervised clustering using CT images of renal masses. We investigated the impact of grid size to cluster analysis. The number of clusters was determined by Gap-statistics. The grid size selections were 2x2, 4x4, 5x5, and 8x8. The results showed that the number of output clusters increased with decreasing grid sizes from 8x8 to 4x4. However, when grid size reached 2x2, the model yielded the same cluster number as 8x8. This finding was consistent across different hyperparameter settings. Additional analyses were conducted to understand the nesting structure between the cluster membership (the mutually exclusive cluster number assigned to each grid in a cluster analysis) from large (8x8) grid and small (2x2) grid, although both grid size selections yielded the same number of clusters. We report that the cluster membership between large grid and small grid is only partially overlaid. This suggests that additional pattern/information is detected by using the small grid. In conclusion, the grid size should be treated as another hyperparameter when using unsupervised clustering methods for pattern recognition in medical imaging analysis.
Morphological metrics such as fractal dimension (FD) have shown value as diagnostic and prognostic markers in diverse cancers. A lack of procedural consensus on fractal techniques may lead to a non-generalization of results across different studies. This study reports variations of Computed Tomography (CT) derived FD renal masses across different fractal analysis implementations. The Fraclac grayscale pixel size 512x512 pixel setting Area Under Curve (AUC) showed the highest AUC value (0.59) among all pixel settings in classifying clear cell renal cell carcinoma (ccRCC) vs. Oncocytoma and liquid poor angiomyolipoma (AML). Similarly, for the multiphase analysis, we also explored MATLAB grayscale pixel sizes from 7x7 to 256x256 pixels. Results showed that the 64x64 pixel setting had the highest AUC of 0.60-0.72 for ccRCC vs. Oncocytoma and AML and AUC of 0.58-0.69 for chromophobe renal cell carcinoma (RCC) vs Oncocytoma.
In this prospective study, forty patients with solid renal masses who underwent contrast-enhanced ultrasound (CEUS) examinations were selected. Using the ImageJ software, renal masses and adjacent normal tissue were manually segmented from CEUS cine exams obtained using the built-in RS85 Samsung scanner software. For the radiomics analysis, one frame representing precontrast, early, peak, and delay enhancement phase were selected post segmentation from each CEUS clip. From each region of interest (ROI) within a tumor tissue normalized renal mass, 112 radiomic metrics were extracted using custom Matlab® code. For the time-intensity curve (TIC) analysis, the segmented ROIs were plotted as a function of time, and the data were fit to a washout curve. From these time-signal intensity curves, perfusion quantitative parameters, were generated. Wilcoxon rank sum test or univariate independent t-test depending on data normality were used for descriptive analyses. Agreement was analyzed using Kappa statistic. Of the 40 solid masses, 31 (77.5%) were malignant, 9 (22.5%) were benign based on histopathology. Excellent agreement was found between histopathological confirmation and visual assessment based on CEUS in discriminating solid renal masses into benign vs. malignant categories (κ=0.89 95% confidence interval (CI): (0.77,1)). The total agreement between the two was 92.5%. The sensitivity and specificity of CEUS-based visual assessment was found to be 100% and 66.7%, respectively. Quantitative analysis revealed TIC metrics revealed statistically significant differences between the malignant and benign groups and between clear cell renal cell carcinoma (ccRCC) and papillary renal cell carcinoma (pRCC) subtypes. The study shows excellent agreement between visual assessment and histopathology, but with the room to improve in specificity.
Clear cell renal cell carcinoma (ccRCC) is a common cancer and could result in poor prognosis. Understanding individual tumor immune microenvironment (TIME) in ccRCC patients may predict prognosis and response to therapy. In this work, we explore the concept of using radiomic features extracted from computer tomography (CT) imaging to correlate the TIME measurements from multiplex immunohistochemistry (mIHC) analysis. Since CT imaging has long been the standard for evaluation of RCCs, it has the potential to provide noninvasive approximations of the tissue-based mIHC biomarkers. We selected two biomarkers that were grounded by clinical research: PD-L1 expression and CD8+PD-1+ T cell to CD8+ T cell ratio of the tumor epithelium. Then we extracted these two markers from a preliminary set of 52 patients using automated mIHC analysis. We used Random Forest, AdaBoost and ElasticNet to classify each sample as either expressing high or low levels of these markers. We found the radiomic features can correlate tumor epithelium PD-L1 >5%, PD-L1 >10%, and CD8+PD1+/CD8+ >37% with AUROC 0.75, 0.85 and 0.71, respectively.
The variation in quantitative measures extracted from computed tomography (CT) perfusion parametric maps due to changes in dose was evaluated. A CT perfusion phantom was scanned on a Philips CT scanner using AAPM recommendations at 2 different speeds and varying x-ray exposure. The acquired images were post-processed using the TeraRecon software. The software outputted Cerebral Blood Flow (CBF), Cerebral Blood Volume (CBV), Mean Transit Time (MTT) maps and Time Attenuation Curves (TAC) of the artery and the vein rods of the phantom. Measurements were made in regions of interest (ROIs) in the two tissue rods (foreground) and 5 regions in the background, across the different parametric maps, respectively. Mixed effect model with AR (1) covariance structure was used to compare measurements across different dose levels as repeated measured random effect. Dunnet adjustment was used for posthoc pairwise comparisons. For the foreground ROI, no significant changes in the measured mean CBF, CBV, and MTT values were observed with changes in dose. As expected, the standard deviation (SD) of CBF and CBV decreased as dose increased. At each dose, higher speed settings were consistently associated with higher SD of CBF and lower MTT. For the background ROI, the measured mean CBF and CBV were significantly higher at lower dose levels, and the SD of CBF decreased as the dose increased. The MTT of the background did not vary with dose. We conclude that radiation dose affects perfusion metrics especially for low or no flow conditions.
Clinical imaging techniques have low accuracy in differentiating malignant tumors such as clear cell Renal Cell Carcinoma (ccRCC) and benign tumors such as oncocytoma. Texture metrics i.e., metrics assessing the variations in grey-levels of intensity making up a region of interest extracted from routine clinical images have shown promising results in achieving this objective. To explore the relationship between tumor behavior and texture metrics from images, we test the effectiveness of 2D Curvelet Transform-based texture analysis in differentiating between ccRCC and Oncocytoma using contrast-enhanced computed tomography (CECT) images. Whole lesions were manually segmented on the nephrographic phase using Synapse 3D (Fujifilm, CT) and co-registered to other phases of multiphase CT acquisitions for each tumor. A first-generation curvelet transform code was used to apply forward, inverse transform to segmented images, and texture metrics were extracted from each CT phase. Histopathological diagnosis was obtained following surgical resection. A Wilcoxon rank-sum test showed that curvelet-based metric: energy on corticomedullary phase was significantly (p <0.005) higher in oncocytoma (0.06±0.04) than ccRCC (0.04±0.05). Higher values of energy are associated with homogenous textures. A supportive receiver operator characteristics analysis based on energy metric revealed reasonable discrimination (AUC>0.7, p <0.05) between ccRCC and oncocytoma. We conclude based on these preliminary results that curvelet- based energy metric can differentiate between ccRCC and oncocytoma based on their CECT data. In combination with other metrics, curvelet metrics may advance radiomic analysis in evaluating clinical imaging data.
Purpose: Evaluate the feasibility of using a Nakagami model to create an accurate parametric image from ultrasound imaging data for the differentiation of homogenous and heterogeneous texture phantoms. Analysis was done on the raw data i.e., radiofrequency (RF) data collected before any post processing that can affect the images. Materials and methods: The Nakagami parametric image was constructed on demodulated RF data with the sliding window technique to create a map of local parameters. The Nakagami parameter (m) for the entire image was found by averaging all values. By design, when m is greater than 1, the distribution is post-Rayleigh. When m is equal to 1, the distribution is Rayleigh. To test the technique, two agar phantoms were constructed, using varying amounts of flour as the scatterer. The higher amount of flour scatterer was meant to mimic heterogeneous texture and the lesser amount meant to mimic homogeneous texture. Scans were done on each phantom and analyzed for differences in the Nakagami parameter. Results: Phantom 1 displayed a post-Rayleigh distribution (m = 36.1±7.0), while phantom 2 did so, to a lesser extent (m = 1.64±0.12). As the distribution transitions from Rayleigh to post Rayleigh, the scatterers in the sample go from being periodically located/randomly distributed to large numbers of randomly distributed scatterers. Conclusion: Our study suggests that Nakagami parametric based metrics may be used to increase robustness of texture analysis, considering the analysis is done on the raw data before any post processing that can affect the images is introduced.
Purpose: To evaluate potential use of wavelets analysis in discriminating benign and malignant renal masses (RM) Materials and Methods: Regions of interest of the whole lesion were manually segmented and co-registered from multiphase CT acquisitions of 144 patients (98 malignant RM: renal cell carcinoma (RCC) and 46 benign RM: oncocytoma, lipid-poor angiomyolipoma). Here, the Haar wavelet was used to analyze the grayscale images of the largest segmented tumor in the axial direction. Six metrics (energy, entropy, homogeneity, contrast, standard deviation (SD) and variance) derived from 3-levels of image decomposition in 3 directions (horizontal, vertical and diagonal) respectively, were used to quantify tumor texture. Independent t-test or Wilcoxon rank sum test depending on data normality were used as exploratory univariate analysis. Stepwise logistic regression and receiver operator characteristics (ROC) curve analysis were used to select predictors and assess prediction accuracy, respectively. Results: Consistently, 5 out of 6 wavelet-based texture measures (except homogeneity) were higher for malignant tumors compared to benign, when accounting for individual texture direction. Homogeneity was consistently lower in malignant than benign tumors irrespective of direction. SD and variance measured in the diagonal direction on the corticomedullary phase showed significant (p<0.05) difference between benign versus malignant tumors. The multivariate model with variance (3 directions) and SD (vertical direction) extracted from the excretory and pre-contrast phase, respectively showed an area under the ROC curve (AUC) of 0.78 (p < 0.05) in discriminating malignant from benign. Conclusion: Wavelet analysis is a valuable texture evaluation tool to add to a radiomics platforms geared at reliably characterizing and stratifying renal masses.
Clinicians can now objectively quantify tumor necrosis by Hounsfield units and enhancement characteristics from multiphase contrast enhanced CT imaging. NecroQuant has been designed to work as part of a radiomics pipelines. The software is a departure from the conventional qualitative assessment of tumor necrosis, as it provides the user (radiologists and researchers) a simple interface to precisely and interactively define and measure necrosis in contrast-enhanced CT images. Although, the software is tested here on renal masses, it can be re-configured to assess tumor necrosis across variety of tumors from different body sites, providing a generalized, open, portable, and extensible quantitative analysis platform that is widely applicable across cancer types to quantify tumor necrosis.
Purpose: Evaluate the feasibility of spectral analysis, particularly fast fourier transform (FFT), to help clinicians differentiate clear cell renal cell carcinoma (ccRCC) tumor grades using contrast-enhanced computed tomography (CECT) images of renal masses, quantitatively, and compare their performance to the Fuhrman grading system. Materials and Methods: Regions of interest of the whole lesion were manually segmented and co-registered from multiphase CT acquisitions of 95 patients with ccRCC. Here, FFT is employed to objectively quantify the texture of a tumor surface by evaluating tissue gray-level patterns and automatically measure frequency-based texture metrics. An independent t-test or a Wilcoxon rank sum test (depending on the data distribution) was used to determine if the spectral analysis metrics would produce statistically significant differences between the tumor grades. Receiver operating characteristic (ROC) curve analysis was used to evaluate the usefulness of spectral metrics in predicting the ccRCC grade. Results: The Wilcoxon test showed that there was a significant difference in complexity index between the different tumor grades, p < 0.01 at all the four phases of CECT acquisition. In all cases a positive correlation was observed between tumor grade and complexity index. ROC analysis revealed the importance of the entropy of FFT amplitude, FFT phase and complexity index and its ability to identify grade 1 and grade 4 tumors from the rest of the population. Conclusion: Our study suggests that FFT-based spectral metrics can differentiate between ccRCC grades, and in combination with other metrics improve patient management and prognosis of renal masses.
Radiomics workflows are high-throughput disease descriptive or predictive tools that extract mineable quantitative data of pathological phenotypes from standard-of-care grayscale images using advanced image processing algorithms. The success of these workflows rely on establishing large image datasets from which diverse disease descriptors can be extracted, with the expectation that large numbers may be able to overcome some of the inherent heterogeneities inherent in standard-of-care medical imaging workflows. Here, we present such a radiomics platform which relies on a combination of existing standard-of-care imaging clinical and research software as well as custom written code. The key components of the workflow include a file organization schema for centralized data storage, deployment of image registration strategies, and frontend GUI design for ease of use by the clinical researcher, all of which increase the transparency, flexibility, and portability of our radiomics platform. Widespread establishment of such radiomics platform can greatly revolutionize radiomics research and aid in successful translation into clinical decision support systems.
Presented are three preliminary studies completed using our proposed radiomics research workflow to investigate various diseases. The radiomics research workflow is modality and disease independent which allow it to serve as a general platform for medical image post-processing experimentation.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.