Combined Radiomic and Visual Assessment for Improved Detection of Lung Adenocarcinoma Invasiveness on Computed Tomography Scans: A Multi-Institutional Study.

Objective: The timing and nature of surgical intervention for semisolid abnormalities are dependent upon distinguishing between adenocarcinoma-in-situ (AIS), minimally invasive adenocarcinoma (MIA), and invasive adenocarcinoma (INV). We sought to develop and evaluate a quantitative imaging method to determine invasiveness of small, ground-glass lesions on computed tomography (CT) chest scans. Methods: The study comprised 268 patients from 4 institutions with resected (<=3 cm) semisolid lesions with confirmed histopathological diagnosis of MIA/AIS or INV. A total of 248 radiomic texture features from within the tumor nodule (intratumoral) and adjacent to the nodule (peritumoral) were extracted from manually annotated lung nodules of chest CT scans. The datasets were randomly divided, with 40% of patients used for training and 60% used for testing the machine classifier (Training DTrain, N=106; Testing, DTest, N=162). Results: The top five radiomic stable features included four intratumoral (Laws and Haralick feature families) and one peritumoral feature within 3 to 6 mm of the nodule (CoLlAGe feature family), which successfully differentiated INV from MIA/AIS nodules with an AUC of 0.917 [0.867-0.967] on DTrain and 0.863 [0.79-0.931] on DTest. The radiomics model successfully differentiated INV from MIA cases (<1 cm AUC: 0.76 [0.53-0.98], 1-2 cm AUC: 0.92 [0.85-0.98], 2-3 cm AUC: 0.95 [0.88-1]). The final integrated model combining the classifier with the radiologists' score gave the best AUC on DTest (AUC=0.909, p<0.001). Conclusions: Addition of advanced image analysis via radiomics to the routine visual assessment of CT scans help better differentiate adenocarcinoma subtypes and can aid in clinical decision making. Further prospective validation in this direction is warranted.

The Role of Imaging in Health Screening: Overview, Rationale of Screening, and Screening Economics.

Imaging screening examinations are growing in their indications and volume to identify conditions at an early, treatable stage. The Radiology Research Alliance's 'Role of Imaging in Health Screening' Task Force provides a review of imaging-based screening rationale, economics, and describes established guidelines by various organizations. Various imaging modalities can be employed in screening, and are often chosen based on the specific pathology and patient characteristics. Prevalent disease processes with identifiable progression patterns that benefit from early potentially curative interventions are ideal for screening. Two such examples include colonic precancerous polyp progression to adenocarcinoma in colon cancer formation and atypical ductal hyperplasia progression to ductal carcinoma in situ and invasive ductal carcinoma in breast cancer. Economic factors in imaging-based screening are reviewed, including in the context of value-based reimbursements. Global differences in screening are outlined, along with the role of various organizational guidelines, including the American Cancer Society, the US Preventive Services Task Force, and the American College of Radiology.

Mediastinal hemangiomas: Spectrum of CT and MRI findings - retrospective case series study and systematic review of the literature.

PURPOSE: To analyze the imaging manifestations of mediastinal hemangioma (MH) by CT and MRI to aid in its successful diagnosis and preoperative evaluation. METHODS: Seventeen cases of MH diagnosed by histopathology combined with CT and MRI were retrospectively collected; and their CT and MRI features, including the lesions' site and range, shape, size, margin, density or signal, enhancement pattern, mass-cardiovascular interface, mass-pulmonary interface, and other characteristics were evaluated. RESULTS: The anterior, middle, and posterior mediastinum were involved in 13, 13, and 8 cases, respectively. The masses size varied from 20 to 233 mm. Irregular, dumbbell-like, and oval masses were found in 13, 2, and 2 cases, respectively, while with pampiniform growth in 16 cases and expansive growth in 1 case. Mixed density, homogeneous density solid masses, and heterogeneous density masses with dominant fat were found in 9, 5, and 3 cases, respectively, showing mild or significant enhancement in aortic phase while no or mild enhancement in pulmonary artery phase. Draining veins were found in 16 cases and feeding arteries in 10 cases. Phleboliths were detected in 10 cases, splenic hemangiomas in 6 cases, and left lateral-chest-wall hemangioma in 1 case. In MRI sequences, mixed signal was found on T1WI and heterogeneous hypersignal with nodular or linear hyposignal on T2WI in 5 cases, mild or significant enhancement in 4 cases, draining veins in 2 cases, and no feeding arteries or phleboliths were seen. CONCLUSION: Presence of phleboliths, pampiniform growth pattern, and aberrant draining veins are relatively specific characteristics in diagnosing MH.

Utility of virtual monoenergetic images derived from a dual-layer detector-based spectral CT in the assessment of aortic anatomy and pathology: A retrospective case control study.

OBJECTIVES: To evaluate the ability of the retrospectively generated virtual monoenergetic images (VMIs) from a dual-layer detector-based spectral computed tomography (SDCT) to augment aortic enhancement for the evaluation of aortic anatomy and pathology. METHODS: 98 patients with suboptimal aortic enhancement (≤200 HU) were retrospectively identified from SDCT scans. VMI from 40 to 80 keV were generated. Attenuation, noise, SNR, and CNR were measured at seven levels in the aorta. Image quality was graded on a 5-point scale, 5 being the best. From the VMI, an ideal set was chosen with mean vascular attenuation above 200 HU while maintaining diagnostic quality. Image parameters and quality of this ideal-set were compared to the standard 120-kVp images. RESULTS: The mean attenuation of all seven measured anatomical regions was 156.6 ±â€¯61.7 HU in the 120-kVp images. Attenuation of the VMI from 40 to 70 keV were higher than the 120-kVp image, measuring 439.2 ±â€¯215.3 HU, 298.5 ±â€¯140.6 HU, 213.4 ±â€¯94.3 HU, and 164.7 ±â€¯90.2 HU, for 40 keV, 50 keV, 60 keV, and 70 keV, respectively (p value <0.01 for 40, 50, 60 keV; 0.07 for 70 keV). SNR and CNR showed similar trends. The 50 keV VMI had the best image quality (4.48 ±â€¯0.84 vs. 2.24 ±â€¯0.92 on 120-kVp images, p < 0.001). Attenuation, CNR, and SNR increased by 90.6%, 85.0%, and 108.1% at 50 keV compared to 120-kVp. CONCLUSIONS: A contrast-enhanced CT study can be optimized for the assessment of the aorta by using low-energy VMI obtained using SDCT. At the optimal monoenergetic level, attenuation, SNR, CNR and image quality were significantly higher than that of conventional polyenergetic images.

Multimodality imaging assessment of endoleaks post-endovascular aortic repair.

Endoleaks are a common complication of endovascular aortic repair (EVAR). As a result, patients require lifelong imaging surveillance following EVAR. In current clinical practice, evaluation for endoleaks is predominantly performed with CT angiography (CTA). Due to the significant cumulative radiation burden associated with repetitive CTA imaging, as well as the repeated administration of nephrotoxic contrast agent, contrast-enhanced ultrasound (CEUS) and magnetic resonance angiography (MRA) have evolved as potential modalities for lifelong surveillance post-EVAR. In this paper, multimodality imaging, including CTA, CEUS and MRA, for the surveillance of endoleaks is discussed. Further, new CTA techniques for radiation reduction are elaborated. Additionally, imagery for three cases of aortic endoleak detection using CTA and five cases using MRA are presented. Imaging for different types of endoleaks with CTA, MRA and CEUS are presented. For lifelong endoleak surveillance post-EVAR, CTA is still regarded as the imaging modality of choice. However, advancements in CEUS and MRA technique enable partial replacement of CTA in certain patients.

Assessment of 70-keV virtual monoenergetic spectral images in abdominal CT imaging: A comparison study to conventional polychromatic 120-kVp images.

PURPOSE: To evaluate the image quality of 70-keV virtual monoenergetic (monoE) abdominal CT images compared to 120-kVp polychromatic images generated from a spectral detector CT (SDCT) scanner. METHODS: This prospective study included generation of a 120-kVp polychromatic dataset and a 70-keV virtual monoE dataset after a single contrast-enhanced CT acquisition on a SDCT scanner (Philips Healthcare) during portal venous phase. The attenuation values (HU), noise, signal-to-noise ratio (SNR), and contrast-to-noise ratio (CNR) were measured in the liver, spleen, pancreas, kidney, aorta, portal vein, and muscle. The subjective image quality including noise, soft tissue contrast, sharpness, and overall image quality were graded on a 5-point Likert scale by two radiologists independently (1-worst image quality, 5-best image quality). Statistical analysis was performed using paired sample t test and Fleiss's Kappa. RESULTS: Fifty-five patients (54.3 ± 16.8 y/o; 28 M, 27 F) were recruited. The noise of target organs was significantly lower in virtual monoE images in comparison to polychromatic images (p < 0.001). The SNR and CNR were significantly higher in virtual monoE images (p < 0.001 for both). Subjective image quality of 70-keV virtual monoE images was significantly better (p < 0.001) for all evaluated parameters. Median scores for all subjective parameters were 3.0 versus 4.0 for polychromatic vs virtual monoE images, respectively. The inter-reader agreement for overall image quality was good (Kappa were 0.767 and 0.762 for polychromatic and virtual monoE images, respectively). CONCLUSION: In abdominal imaging, 70-keV virtual monoE CT images demonstrated significantly better noise, SNR, CNR, and subjective score compared to conventional 120-kVp polychromatic images.

Cardiac MR imaging in constrictive pericarditis: multiparametric assessment in patients with surgically proven constriction.

To assess the utility of cardiac magnetic resonance (MR) imaging in the diagnosis of constrictive pericarditis (CP). This study was approved by the institutional review board, with a waiver of informed consent. A total of 42 consecutive patients (mean age, 55 ± 16 years; 3 women, 39 men) with CP treated with pericardiectomy who had undergone cardiac MR before surgery were evaluated retrospectively. An additional 21 patients were evaluated as a control group; of these, 10 consecutive patients received cardiac MR for reasons other than suspected pericardial disease, and 11 consecutive patients had a history of pericarditis but no clinical suspicion of pericardial constriction. MR imaging parameters were analyzed independently and with a decision tree algorithm for usefulness in the prediction of CP. Catheterization data were also reviewed when available. A model combining pericardial thickness and relative interventricular septal (IVS) excursion provided the best overall performance in prediction of CP (C statistic, 0.98, 100% sensitivity, 90% specificity). Several individual parameters also showed strong predictive value in the assessment of constriction, including relative IVS excursion (sensitivity, 93%; specificity, 95%), pericardial thickness (sensitivity, 83%; specificity, 100%), qualitative assessment of pathologic coupling (sensitivity, 88%; specificity, 100%), diastolic IVS bounce (sensitivity, 90%; specificity, 85%), left ventricle area change (sensitivity, 86%; specificity, 100%), and eccentricity index (sensitivity, 86%; specificity, 90%; all P < 0.001). Strong agreement was observed between catheterization and surgical findings of constriction (97%). Cardiac MR provides robust quantitative and qualitative analysis for the diagnosis of CP.

Cardiac MR assessment of aortic regurgitation: holodiastolic flow reversal in the descending aorta helps stratify severity.

PURPOSE: To assess the utility of holodiastolic flow reversal (HDR) in the descending aorta on velocity-encoded cardiac magnetic resonance (MR) images in the stratification of aortic regurgitation (AR) severity. MATERIALS AND METHODS: This study was approved by the institutional review board, with waiver of informed consent. A total of 80 patients (overall mean age, 49 years ± 18 [standard deviation]; 22 women and 58 men) with clinical indication for cardiac MR imaging of the aorta were analyzed retrospectively. Velocity-encoded MR imaging was used to quantify AR and assess for HDR at the level of the middescending aorta. These indexes were compared with a qualitative integrated echocardiographic evaluation of AR severity. Sensitivity and specificity for HDR in the prediction of substantial AR were determined, and logistic regression analysis (with associated odds ratios and C statistics) was performed, with HDR and regurgitant fraction as independent predictors. An additional 42 patients (overall mean age, 48 years ± 21; 12 female and 30 male) were then prospectively evaluated in similar fashion to evaluate a decision model derived from analysis of the first group. RESULTS: HDR predicted severe AR (echo grade, 4) with high sensitivity (100%) and specificity (93%). HDR was highly specific (100%) but had lower sensitivity (61%) for moderate to severe AR (echo grade, 3-4). Integration of HDR and direct AR quantification into a combined stratification model based on analysis of the primary group showed good predictive results in the validation group, with a C statistic of 0.94 for moderate to severe AR and 0.93 for severe AR. CONCLUSION: HDR in the middescending thoracic aorta observed at cardiac MR is indicative of severe AR and can be used in conjunction with quantified regurgitant values obtained from velocity-encoded MR imaging to stratify AR severity.

Utility of free-breathing, whole-heart, three-dimensional magnetic resonance imaging in the assessment of coronary anatomy for congenital heart disease.

The incidence of coronary anomalies is increased in congenital heart disease (CHD). Whole-heart magnetic resonance imaging (MRI) has been proposed as a robust approach to coronary artery imaging without ionizing radiation. The proximal coronary arteries were imaged in 112 CHD patients (63 males) age 17 ± 13 years (range 11 days-68 years) using a navigator-gated, whole-heart, three-dimensional (3D) technique at 1.5 T. Two observers assessed image quality overall and for left anterior descending coronary artery (LAD), left circumflex coronary artery (LCX), and right coronary artery (RCA) using a 5-point scale ranging from 0 (not visible) to 4 (clear margins). Weighted kappa was used to assess interobserver agreement. Coronary artery origins were visible in 99% of the patients. The left main origin was not visualized in one patient, although the LAD, LCX, and RCA were visualized. Eight patients (7%) had anomalies. The overall image quality was 3.3 ± 0.8 for reader 1 and 3.1 ± 1.0 for reader 2. Age had a significant effect on image quality, with younger patients having lower scores. Agreement between readers was moderate (overall kappa, 0.60). Free-breathing, navigator-gated, whole-heart 3D MRI is a useful, robust, and reliable noninvasive technique for assessing coronary artery origins and their proximal course with diagnostic quality in CHD patients.