The aim of this study was to evaluate the use of texture analysis for differentiation between benign from malignant adrenal lesions on contrast-enhanced abdominal computed tomography (CT).
After ...institutional review board approval, a retrospective analysis was performed, including an electronic search of pathology records for all biopsied adrenal lesions. Patients were included if they also had a contrast-enhanced abdominal CT in the portal venous phase. Computed tomographic images were manually segmented, and texture analysis of the segmented tumors was performed. Texture analysis results of benign and malignant tumors were compared, and areas under the curve (AUCs) were calculated.
One hundred twenty-five patients were included in the analysis. Excellent discriminators of benign from malignant lesions were identified, including entropy and standard deviation. These texture features demonstrated lower values for benign lesions compared with malignant lesions. Entropy values of benign lesions averaged 3.95 using a spatial scaling factor of 4 compared with an average of 5.08 for malignant lesions (P < .0001). Standard deviation values of benign lesions averaged 19.94 on the unfiltered image compared with an average of 34.32 for malignant lesions (P < .0001). Entropy demonstrated AUCs ranging from 0.95 to 0.97 for discriminating tumors, with sensitivities and specificities ranging from 81% to 95% and 88% to 100%, respectively. Standard deviation demonstrated AUCs ranging from 0.91 to 0.94 for discriminating tumors, with sensitivities and specificities ranging from 73% to 93% and 86% to 95%, respectively.
Texture analysis offers a noninvasive tool for differentiating benign from malignant adrenal tumors on contrast-enhanced CT images. These results support the further development of texture analysis as a quantitative biomarker for characterizing adrenal tumors.
Purpose
To evaluate the utility of texture analysis for the differentiation of renal tumors, including the various renal cell carcinoma subtypes and oncocytoma.
Materials and methods
Following IRB ...approval, a retrospective analysis was performed, including all patients with pathology-proven renal tumors and an abdominal computed tomography (CT) examination. CT images of the tumors were manually segmented, and texture analysis of the segmented tumors was performed. A support vector machine (SVM) method was also applied to classify tumor types. Texture analysis results were compared to the various tumors and areas under the curve (AUC) were calculated. Similar calculations were performed with the SVM data.
Results
One hundred nineteen patients were included. Excellent discriminators of tumors were identified among the histogram-based features noting features skewness and kurtosis, which demonstrated AUCs of 0.91 and 0.93 (
p
< 0.0001), respectively, for differentiating clear cell subtype from oncocytoma. Histogram feature median demonstrated an AUC of 0.99 (
p
< 0.0001) for differentiating papillary subtype from oncocytoma and an AUC of 0.92 for differentiating oncocytoma from other tumors. Machine learning further improved the results achieving very good to excellent discrimination of tumor subtypes. The ability of machine learning to distinguish clear cell subtype from other tumors and papillary subtype from other tumors was excellent with AUCs of 0.91 and 0.92, respectively.
Conclusion
Texture analysis is a promising non-invasive tool for distinguishing renal tumors on CT images. These results were further improved upon application of machine learning, and support the further development of texture analysis as a quantitative biomarker for distinguishing various renal tumors.
Purpose
The purpose of this study was to assess the diagnostic yield of abdomen magnetic resonance imaging (MRI) in the inpatient setting following a computed tomography (CT).
Methods
All inpatient ...abdominopelvic MRIs performed on patients for a 1-year period were identified and medical records were retrospectively reviewed for the following information. Only MRIs with a preceding CT were included in the study.
Results
A total of 221 MRIs were included. Forty exams were deemed technically inadequate due to motion, while 9 more patients did not tolerate a full examination. The most common indications were focal liver lesion (
n
= 101), pancreaticobiliary ductal dilatation (
n
= 39), abnormal liver function tests (
n
= 26), acute pancreatitis (
n
= 14), abdominal pain (
n
= 10), and fever/sepsis (
n
= 10). 83 MRIs were recommended on CT and 138 were requests from the care team. In 63 cases, MRI offered new information over CT. Thirty-two MRIs recommended by radiologists affected patient management, while only 31 MRIs recommended by the care team affected management. Of these 63 MRIs, 29 cases changed immediate inpatient management, requiring further intervention. In these cases, MRI identified abscesses, choledocholithiasis, or made other diagnoses such as cholecystitis, which were not diagnosed on CT. Patient LOS increased in 24 patients in order to receive an MRI. Average costs of outpatient CTs and MRIs are typically 20% less than inpatient costs.
Conclusion
Inpatient abdomen MRIs have limited impact on patient care following a CECT and entail higher cost, utilize more resources, scanner time, and increase patient LOS. Therefore, it should be reserved for select clinical indications.
Purpose
To evaluate the potential utility of texture analysis of proton density maps for quantifying hepatic fibrosis in a murine model of hepatic fibrosis.
Materials and Methods
Following ...Institutional Animal Care and Use Committee (IACUC) approval, a dietary model of hepatic fibrosis was used and 15 ex vivo murine liver tissues were examined. All images were acquired using a 30 mm bore 11.7T magnetic resonance imaging (MRI) scanner with a multiecho spin‐echo sequence. A texture analysis was employed extracting multiple texture features including histogram‐based, gray‐level co‐occurrence matrix‐based (GLCM), gray‐level run‐length‐based features (GLRL), gray level gradient matrix (GLGM), and Laws' features. Texture features were correlated with histopathologic and digital image analysis of hepatic fibrosis.
Results
Histogram features demonstrated very weak to moderate correlations (r = –0.29 to 0.51) with hepatic fibrosis. GLCM features correlation and contrast demonstrated moderate‐to‐strong correlations (r = –0.71 and 0.59, respectively) with hepatic fibrosis. Moderate correlations were seen between hepatic fibrosis and the GLRL feature short run low gray‐level emphasis (SRLGE) (r = –0. 51). GLGM features demonstrate very weak to weak correlations with hepatic fibrosis (r = –0.27 to 0.09). Moderate correlations were seen between hepatic fibrosis and Laws' features L6 and L7 (r = 0.58).
Conclusion
This study demonstrates the utility of texture analysis applied to proton density MRI in a murine liver fibrosis model and validates the potential utility of texture‐based features for the noninvasive, quantitative assessment of hepatic fibrosis. J. Magn. Reson. Imaging 2015;42:1259–1265.
Purpose
To test the hypothesis that an automated post-processing workflow reduces trauma panscan exam completion times and variability.
Methods
One-hundred-fifty consecutive trauma panscans performed ...between June 2018 and December 2019 were included, half before and half after implementation of an automated software-driven post-processing workflow. Acquisition and reconstruction timestamps were used to calculate total examination time (first acquisition to last reformation), setup time (between the non-contrast and contrast-enhanced acquisitions), and reconstruction time (for the contrast-enhanced reconstructions and reformations). The performing technologist was recorded and accounted for in analyses using linear mixed models to assess differences between the pre- and post-intervention groups.
Results
Exam, setup, and recon times were (mean ± standard deviation) 33.5 ± 4.6, 9.2 ± 2.4, and 23.6 ± 4.7 min before and 27.8 ± 1.5, 8.9 ± 1.4, and 18.9 ± 1.7 min after intervention. These reductions of 5.7 and 4.7 min in the mean exam and recon times were statistically significant (
p
< 0.001) while the setup time was not (
p
= 0.49). The reductions in standard deviation were statistically significant for exam and recon times
(p
< 0.0001) but not for setup time (
p
= 0.13). All automated panscans were completed within 36 min, versus 65% with the traditional workflow.
Conclusion
Automation of image reconstruction workflow significantly decreased mean exam and reconstruction times as well as variability between exams, thus facilitating a consistently rapid imaging assessment, and potentially reducing delays in critical management decisions.
An 84-year-old man presented with a 6-week history of painless right testicular swelling. He had no fever or systemic symptoms. Ultrasonography revealed marked asymmetric enlargement and ...hypervascularity of the right testicle. A diagnostic procedure was performed.
Abstract Purpose To compare enhanced Laws textures derived from parametric proton density (PD) maps to other MRI surrogate markers (T2 , PD, apparent diffusion coefficient (ADC)) in assessing degrees ...of liver fibrosis in an ex vivo murine model of hepatic fibrosis imaged using 11.7T MRI. Methods This animal study was IACUC approved. Fourteen male, C57BL/6 mice were divided into control and experimental groups. The latter were fed a 3,5-dicarbethoxy-1,4-dihydrocollidine (DDC) supplemented diet to induce hepatic fibrosis. Ex vivo liver specimens were imaged using an 11.7T scanner, from which the parametric PD, T2 , and ADC maps were generated from spin-echo pulsed field gradient and multi-echo spin-echo acquisitions. A sequential enhanced Laws texture analysis was applied to the PD maps: automated dual-clustering algorithm, optimal thresholding algorithm, global grayscale correction, and Laws texture features extraction. Degrees of fibrosis were independently assessed by digital image analysis (a.k.a. %Area Fibrosis). Scatterplot graphs comparing enhanced Laws texture features, T2 , PD, and ADC values to degrees of fibrosis were generated and correlation coefficients were calculated. Results Hepatic fibrosis and the enhanced Laws texture features were strongly correlated with higher %Area Fibrosis associated with higher Laws textures ( r = 0.89). Without the proposed enhancements, only a moderate correlation was detected between %Area Fibrosis and unenhanced Laws texture features ( r = 0.70). Correlation also existed between %Area Fibrosis and ADC ( r = 0.86), PD ( r = 0.65), and T2 ( r = 0.66). Conclusions Higher degrees of hepatic fibrosis are associated with increased Laws textures. The proposed enhancements could improve the accuracy of Laws texture features significantly.
Abstract Purpose To determine the ability of texture analyses of contrast-enhanced CT images for distinguishing between varying degrees of hepatic fibrosis in patients with chronic liver disease ...using histopathology as the reference standard. Materials and methods Following IRB approval, 83 patients who underwent contrast enhanced 64-MDCT of the abdomen and pelvis in the portal venous phase between 12/2005 and 01/2013 and who had a liver biopsy within 6 months of the CT were included. An in-house developed, MATLAB-based texture analysis program was employed to extract 41 texture features from each of 5 axial segmented volumes of liver. Using the Ishak fibrosis staging scale, histopathologic grades of hepatic fibrosis were correlated with texture parameters after stratifying patients into three analysis groups, comparing Ishak scales 0–2 with 3–6, 0–3 with 4–6, and 0–4 with 5–6. To assess the utility of texture features, receiver operating characteristic (ROC) curves were constructed and the area under the curve (AUC) was used to determine the performance of each feature in distinguishing between normal/low and higher grades of hepatic fibrosis. Results A total of 19 different texture features with 7 histogram features, one grey level co-occurrence matrix, 6 gray level run length, 1 Laws feature, and 4 gray level gradient matrix demonstrated statistically significant differences for discriminating between fibrosis groupings. The highest AUC values fell in the range of fair performance for distinguishing between different fibrosis groupings. Conclusion These findings suggest that texture-based analyses of contrast-enhanced CT images offer a potential avenue toward the non-invasive assessment of liver fibrosis.
Adrenal masses are often indeterminate on single-phase postcontrast CT. Dual-energy CT (DECT) with three-material decomposition algorithms may aid characterization.
The purpose of this study was to ...compare the diagnostic performance of metrics derived from portal venous phase DECT, including virtual noncontrast (VNC) attenuation, fat fraction, iodine density, and relative enhancement ratio, for characterizing adrenal masses.
This retrospective study included 128 patients (82 women, 46 men; mean age, 64.6 ± 12.7 SD years) who between January 2016 and December 2019 underwent portal venous phase abdominopelvic DECT that showed a total of 139 adrenal lesions with an available reference standard based on all imaging, clinical, and pathologic records (87 adenomas, 52 nonadenomas 48 metastases, two adrenal cortical carcinomas, one ganglioneuroma, one hematoma). Two radiologists placed ROIs to determine the following characteristics of the masses: VNC attenuation, fat fraction, iodine density normalized to portal vein, and for masses with VNC greater than 10 HU, relative enhancement ratio (ratio of portal venous phase attenuation to VNC attenuation). Readers' mean measurements were used for ROC analyses, and clinically optimal thresholds were derived as thresholds yielding the highest sensitivity at 100% specificity.
Adenomas and nonadenomas were significantly different (all
< .001) in VNC attenuation (mean ± SD, 18.5 ± 12.9 vs 34.1 ± 8.9 HU), fat fraction (mean ± SD, 24.3% ± 8.2% vs 14.2% ± 5.6%), normalized iodine density (mean ± SD, 0.34 ± 0.15 vs 0.17 ± 0.17), and relative enhancement ratio (mean ± SD, 186% ± 96% vs 58% ± 59%). AUCs for all metrics ranged from 0.81 through 0.91. The metric with highest sensitivity for adenoma at the clinically optimal threshold (i.e., 100% specificity) was fat fraction (threshold, ≥ 23.8%; sensitivity, 59% 95% CI, 48-69%) followed by VNC attenuation (≤ 15.2 HU; sensitivity, 39% 95% CI, 29-50%), relative enhancement ratio (≥ 214%; sensitivity, 37% 95% CI, 25-50%), and normalized iodine density (≥ 0.90; sensitivity, 1% (95% CI, 0-60%). VNC attenuation at the traditional true noncontrast attenuation threshold of 10 HU or lower had sensitivity of 28% (95% CI, 19-38%) and 100% specificity. Presence of fat fraction 23.8% or greater or relative enhancement ratio 214% or greater yielded sensitivity of 68% (95% CI, 57-77%) with 100% specificity.
For adrenal lesions evaluated with single-phase DECT, fat fraction had higher sensitivity than VNC attenuation at both the clinically optimal threshold and the traditional threshold of 10 HU or lower.
By helping to definitively diagnose adenomas, DECT-derived metrics can help avoid downstream imaging for incidental adrenal lesions.
To assess the utility of texture analysis of T
and T
maps for the detection of hepatic fibrosis in a murine model of hepatic fibrosis.
Following Institutional Animal Care and Use Committee approval, ...a dietary model of hepatic fibrosis was used and 15 ex vivo murine livers were examined. Images were acquired using a 30 mm bore 11.7T magnetic resonance imaging (MRI) scanner with a rapid acquisition with relaxation enhancement sequence. Texture analysis was then employed, extracting texture features including histogram-based, gray-level co-occurrence matrix-based (GLCM), gray-level run-length-based features (GLRL), gray-level gradient matrix (GLGM), and Laws' features. Areas under the curve (AUCs) were then calculated to determine the ability of texture features to detect hepatic fibrosis.
Texture analysis of T
maps identified very good to excellent discriminators of hepatic fibrosis within the histogram and GLGM categories. Histogram feature interquartile range (IQR) achieved an AUC value of 0.90 (P < 0.0001) and GLGM feature variance gradient achieved an AUC of 0.91 (P < 0.0001). Texture analysis of T
maps identified very good to excellent discriminators of hepatic fibrosis within the histogram, GLCM, GLRL, and GLGM categories. GLGM feature kurtosis was the best discriminator of hepatic fibrosis, achieving an AUC value of 0.90 (P < 0.0001).
This study demonstrates the utility of texture analysis for the detection of hepatic fibrosis when applied to T
and T
maps in a murine model of hepatic fibrosis and validates the potential use of this technique for the noninvasive, quantitative assessment of hepatic fibrosis.
1 J. Magn. Reson. Imaging 2017;45:250-259.
