Staging of Cervical Cancer: A Practical Approach Using MRI and FDG PET.

This review provides a practical approach to the imaging evaluation of patients with cervical cancer (CC), from initial diagnosis to restaging of recurrence, focusing on MRI and FDG PET. The primary updates to the International Federation of Gynecology and Obstetrics (FIGO) CC staging system, as well as these updates' relevance to clinical management, are discussed. The recent literature investigating the role of MRI and FDG PET in CC staging and image-guided brachytherapy is summarized. The utility of MRI and FDG PET in response assessment and posttreatment surveillance is described. Important findings on MRI and FDG PET that interpreting radiologists should recognize and report are illustrated. The essential elements of structured reports during various phases of CC management are outlined. Special considerations, including the role of imaging in patients desiring fertility-sparing management, differentiation of CC and endometrial cancer, and unusual CC histologies, are also described. Finally, future research directions including PET/MRI, novel PET tracers, and artificial intelligence applications are highlighted.

Multimodal data integration using machine learning improves risk stratification of high-grade serous ovarian cancer.

Patients with high-grade serous ovarian cancer suffer poor prognosis and variable response to treatment. Known prognostic factors for this disease include homologous recombination deficiency status, age, pathological stage and residual disease status after debulking surgery. Recent work has highlighted important prognostic information captured in computed tomography and histopathological specimens, which can be exploited through machine learning. However, little is known about the capacity of combining features from these disparate sources to improve prediction of treatment response. Here, we assembled a multimodal dataset of 444 patients with primarily late-stage high-grade serous ovarian cancer and discovered quantitative features, such as tumor nuclear size on staining with hematoxylin and eosin and omental texture on contrast-enhanced computed tomography, associated with prognosis. We found that these features contributed complementary prognostic information relative to one another and clinicogenomic features. By fusing histopathological, radiologic and clinicogenomic machine-learning models, we demonstrate a promising path toward improved risk stratification of patients with cancer through multimodal data integration.

Accelerating Prostate Diffusion-weighted MRI Using a Guided Denoising Convolutional Neural Network: Retrospective Feasibility Study.

PURPOSE: To investigate the feasibility of accelerating prostate diffusion-weighted imaging (DWI) by reducing the number of acquired averages and denoising the resulting image using a proposed guided denoising convolutional neural network (DnCNN). MATERIALS AND METHODS: Raw data from the prostate DWI scans were retrospectively gathered between July 2018 and July 2019 from six single-vendor MRI scanners. There were 103 datasets used for training (median age, 64 years; interquartile range [IQR], 11), 15 for validation (median age, 68 years; IQR, 12), and 37 for testing (median age, 64 years; IQR, 12). High b-value diffusion-weighted (hb DW) data were reconstructed into noisy images using two averages and reference images using all 16 averages. A conventional DnCNN was modified into a guided DnCNN, which uses the low b-value DW image as a guidance input. Quantitative and qualitative reader evaluations were performed on the denoised hb DW images. A cumulative link mixed regression model was used to compare the readers' scores. The agreement between the apparent diffusion coefficient (ADC) maps (denoised vs reference) was analyzed using Bland-Altman analysis. RESULTS: Compared with the original DnCNN, the guided DnCNN produced denoised hb DW images with higher peak signal-to-noise ratio (32.79 ± 3.64 [standard deviation] vs 33.74 ± 3.64), higher structural similarity index (0.92 ± 0.05 vs 0.93 ± 0.04), and lower normalized mean square error (3.9% ± 10 vs 1.6% ± 1.5) (P < .001 for all). Compared with the reference images, the denoised images received higher image quality scores from the readers (P < .0001). The ADC values based on the denoised hb DW images were in good agreement with the reference ADC values (mean ADC difference ranged from -0.04 to 0.02 × 10-3 mm2/sec). CONCLUSION: Accelerating prostate DWI by reducing the number of acquired averages and denoising the resulting image using the proposed guided DnCNN is technically feasible. Supplemental material is available for this article. © RSNA, 2020.

Feasibility of a sub-3-minute imaging strategy for ungated quiescent interval slice-selective MRA of the extracranial carotid arteries using radial k-space sampling and deep learning-based image processing.

PURPOSE: To develop and test the feasibility of a sub-3-minute imaging strategy for non-contrast evaluation of the extracranial carotid arteries using ungated quiescent interval slice-selective (QISS) MRA, combining single-shot radial sampling with deep neural network-based image processing to optimize image quality. METHODS: The extracranial carotid arteries of 12 human subjects were imaged at 3 T using ungated QISS MRA. In 7 healthy volunteers, the effects of radial and Cartesian k-space sampling, single-shot and multishot image acquisition (1.1-3.3 seconds/slice, 141-423 seconds/volume), and deep learning-based image processing were evaluated using segmental image quality scoring, arterial temporal SNR, arterial-to-background contrast and apparent contrast-to-noise ratio, and structural similarity index. Comparison of deep learning-based image processing was made with block matching and 3D filtering denoising. RESULTS: Compared with Cartesian sampling, radial k-space sampling increased arterial temporal SNR 107% (P < .001) and improved image quality during 1-shot imaging (P < .05). The carotid arteries were depicted with similar image quality on the rapid 1-shot and much lengthier 3-shot radial QISS protocols (P = not significant), which was corroborated in patient studies. Deep learning-based image processing outperformed block matching and 3D filtering denoising in terms of structural similarity index (P < .001). Compared with original QISS source images, deep learning image processing provided 24% and 195% increases in arterial-to-background contrast (P < .001) and apparent contrast-to-noise ratio (P < .001), and provided source images that were preferred by radiologists (P < .001). CONCLUSION: Rapid, sub-3-minute evaluation of the extracranial carotid arteries is feasible with ungated single-shot radial QISS, and benefits from the use of deep learning-based image processing to enhance source image quality.

Dynamic quantitative nonenhanced magnetic resonance angiography of the abdominal aorta and lower extremities using cine fast interrupted steady-state in combination with arterial spin labeling: a feasibility study.

BACKGROUND: Cine fast interrupted steady-state in combination with arterial spin labeling is a recently described nonenhanced magnetic resonance angiography (MRA) technique that relies on bolus tracking for time-resolved digital subtraction angiography-like displays of blood flow patterns. We evaluated the feasibility of applying this technique to display in-plane flow patterns in two regions: the abdominal aorta and lower extremity peripheral arteries. METHODS: We performed an institutional review board-approved study in healthy subjects and patients. In 7 healthy subjects, in-plane flow was imaged at 4 stations ranging from the lower legs to the aorto-iliac bifurcation (junction of the distal thigh and upper calf, mid-thigh, junction of the upper thigh and pelvis, upper pelvis). In 5 healthy subjects and 6 patients without abdominal aortopathy, images were acquired through the suprarenal abdominal aorta. Ten patients with known peripheral arterial disease and two patients with stable disease of the abdominal aorta were also evaluated. Peak velocity was compared at each of the 4 stations for cine fast interrupted steady-state in combination with arterial spin labeling and two-dimensional cine phase contrast in patients with normal vessels. RESULTS: In-plane flow patterns were well visualized in all peripheral arterial stations and in the abdominal aorta, providing a high quality display of hemodynamic patterns along extensive lengths of the vessels. There was very strong positive correlation (r = 0.952, P < 0.05) and excellent agreement (intraclass correlation coefficient, 0.935; 95% confidence interval, 0.812-0.972) between peak flow velocities measured by cine fast interrupted steady-state in combination with arterial spin labeling and two-dimensional cine phase contrast. In 10 patients with peripheral artery disease and 2 patients with aortic pathology, cine fast interrupted steady-state in combination with arterial spin labeling provided a visual demonstration of abnormal hemodynamics. CONCLUSION: This feasibility study suggests that cine fast interrupted steady-state in combination with arterial spin labeling provides an efficient, high quality and physiologically accurate display of in-plane flow patterns over extensive lengths of the lower extremity peripheral arteries, which can be difficult to achieve using other MRA techniques.

Free-Breathing Fast Low-Angle Shot Quiescent-Interval Slice-Selective Magnetic Resonance Angiography for Improved Detection of Vascular Stenoses in the Pelvis and Abdomen: Technical Development.

OBJECTIVES: Balanced steady-state free precession-based quiescent-interval slice-selective (bSSFP QISS) magnetic resonance angiography (MRA) is accurate for the noncontrast evaluation of peripheral arterial disease (PAD); however, drawbacks include the need for breath-holding when imaging the abdomen and pelvis, and sensitivity to off-resonance artifacts. The purpose of this study was to evaluate the image quality and diagnostic accuracy in the pelvis and abdomen of free-breathing fast low-angle shot-based QISS (FLASH QISS) techniques in comparison to bSSFP QISS in patients with PAD, using computed tomographic angiography as the reference. MATERIALS AND METHODS: Twenty-seven patients (69 ± 10 years, 17 men) with PAD were enrolled in this institutional review board-approved, Health Insurance Portability and Accountability Act-compliant prospective study between April and December 2018. Patients underwent noncontrast MRA using standard bSSFP QISS and prototype free-breathing radial-FLASH and Cartesian-FLASH QISS at 3 T. A subset of patients (n = 22) also underwent computed tomographic angiography as the reference standard. Nine arterial segments per patient were evaluated spanning the abdomen, pelvis, and upper thigh regions. Objective (signal intensity ratio and relative standard deviation) and subjective image quality (4-point scale) and stenosis (>50%) were evaluated by 2 readers and compared using one-way analysis of variance, Wilcoxon, and McNemar tests, respectively. RESULTS: A total of 179 vascular segments were available for analysis by all QISS techniques. No significant difference was observed among bSSFP, radial-FLASH, and Cartesian-FLASH QISS techniques in signal intensity ratio (P = 0.428) and relative standard deviation (P = 0.220). Radial-FLASH QISS demonstrated the best image quality (P < 0.0001) and the highest interreader agreement (κ = 0.721). The sensitivity values of bSSFP, radial-FLASH, and Cartesian-FLASH QISS for the detection of greater than 50% stenosis were 76.0%, 84.0%, and 80.0%, respectively, whereas specificity values were 97.6%, 94.0%, and 92.8%, respectively. Moreover, FLASH QISS consistently reduced off-resonance artifacts compared with bSSFP QISS. CONCLUSIONS: Free-breathing FLASH QISS MRA techniques provide improved image quality and sensitivity, high specificity, and reduced off-resonance artifacts for vascular stenosis detection in the abdomen and pelvis.

Ungated nonenhanced radial quiescent interval slice-selective (QISS) magnetic resonance angiography of the neck: Evaluation of image quality.

BACKGROUND: Standard-of-care time-of-flight (TOF) techniques for nonenhanced magnetic resonance angiography (NEMRA) of the carotid bifurcation and other cervical arteries often provide nondiagnostic image quality due to motion and flow artifacts. PURPOSE: To perform an initial evaluation of an ungated radial quiescent-interval slice-selective (QISS) technique for NEMRA of the neck, in comparison with 2D TOF and contrast-enhanced magnetic resonance angiography (CEMRA). STUDY TYPE: Retrospective. POPULATION: Sixty patients referred for neck MR angiography. FIELD STRENGTH/SEQUENCE: Ungated radial QISS at 3T. ASSESSMENT: Three radiologists scored image quality of 18 arterial segments using a 4-point scale (1, nondiagnostic; 2, fair; 3, good; 4, excellent), and two radiologists graded proximal internal carotid stenosis using five categories (<50%, 50-69%, 70-99%, occlusion, nondiagnostic). STATISTICAL TESTS: Friedman tests with post-hoc Wilcoxon signed-rank tests; unweighted Gwet's AC1 statistic; tests for equality of proportions. RESULTS: Ungated radial QISS provided image quality that significantly exceeded 2D TOF (mean scores of 2.7 vs. 2.0, 2.7 vs. 2.2, and 2.9 vs. 2.3; P < 0.001, all comparisons), while CEMRA provided the best image quality (mean scores of 3.6, 3.7, and 3.5 for the three reviewers). Interrater agreement of image quality scores was substantial for CEMRA (AC1 = 0.70, P < 0.001), and moderate for QISS (AC1 = 0.43, P < 0.001) and TOF (AC1 = 0.41, P < 0.001). Compared with TOF, QISS NEMRA provided a significantly higher percentage of diagnostic segments for all three reviewers (91.0% vs. 71.7%, 93.5% vs. 72.9%, 95.5% vs. 85.2%; P < 0.0001) and demonstrated better agreement with CEMRA for grading of proximal internal carotid stenosis (AC1 = 0.94 vs. 0.73 for reviewer 1, P < 0.05; AC1 = 0.89 vs. 0.68 for reviewer 2, P < 0.05). DATA CONCLUSION: In this initial study, ungated radial QISS significantly outperformed 2D TOF for the evaluation of the neck arteries, with overall better image quality and more diagnostic arterial segments, and improved agreement with CEMRA for grading stenosis of the proximal internal carotid artery. LEVEL OF EVIDENCE: 3 Technical Efficacy: Stage 1 J. Magn. Reson. Imaging 2019;50:1798-1807.

Intrahepatic cholangiocarcinoma: can imaging phenotypes predict survival and tumor genetics?

PURPOSE: On computed tomography (CT), intrahepatic cholangiocarcinomas (ICC) are a visibly heterogeneous group of tumors. The purpose of this study was to investigate the associations between CT imaging phenotypes, patient survival, and known genetic markers. METHODS: A retrospective study was performed with 66 patients with surgically resected ICC. Pre-surgical CT images of ICC were assessed by radiologists blinded to tumor genetics and patient clinical data. Associations between qualitative imaging features and overall survival (OS) and disease-free survival (DFS) were performed with Cox proportional hazards regression and visualized with Kaplan-Meier plots. Associations between radiographic features and genetic pathways (IDH1, Chromatin and RAS-MAPK) were assessed with Fisher's Exact test and the Wilcoxon Rank sum test where appropriate and corrected for multiple comparisons within each pathway using the False Discovery Rate correction. RESULTS: Three imaging features were significantly associated with a higher risk of death: necrosis (hazard ratio (HR) 2.95 95% CI 1.44-6.04, p = 0.029), satellite nodules (HR 3.29, 95% CI:1.35-8.02, p = 0.029), and vascular encasement (HR 2.63, 95% CI 1.28-5.41, p = 0.029). Additionally, with each increase in axial size, the risk of death increased (HR 1.14, 95% CI 1.03-1.26, p = 0.029). Similar to findings for OS, satellite nodules (HR 3.81, 95% CI 1.88-7.71, p = 0.002) and vascular encasement (HR 2.25, 95% CI 1.24-4.06, p = 0.019) were associated with increased risk of recurrence/death. No significant associations were found between radiographic features and genes in the IDH1, Chromatin or RAS-MAPK pathways (p = 0.63-84). CONCLUSION: This preliminary analysis of resected ICC suggests associations between CT imaging features and OS and DFS. No association was identified between imaging features and currently known genetic pathways.

What CT characteristics of lepidic predominant pattern lung adenocarcinomas correlate with invasiveness on pathology?

OBJECTIVES: The International Association for the Study of Lung Cancer, American Thoracic Society and European Respiratory Society lung adenocarcinoma classification in 2011 defined three lepidic predominant patterns including adenocarcinoma in situ, minimally invasive adenocarcinoma and lepidic predominant adenocarcinoma. We sought to correlate the radiology and pathology findings and identify any computed tomography (CT) features which can be associated with invasive growth. MATERIALS AND METHODS: An institutional review board approved, retrospective study was conducted evaluating 63 patients with resected, pathologically confirmed, adenocarcinomas with predominant lepidic patterns. Preoperative CT images of the nodules were assessed using quantitative and qualitative radiographic descriptors while blinded to pathologic sub-classification and size. Maximum diameter was measured after evaluation of the axial, sagittal and coronal planes. Radiologic - pathologic associations were examined using Fisher's exact test, the Kruskal-Wallis test and the Spearman correlation coefficient (ρ). RESULTS AND CONCLUSION: Increasing maximum diameter of the whole lesion (ground glass and solid component) on CT was significantly associated with invasiveness (p = .003), as was the maximum pathologic specimen diameter (p = .008). Larger diameter of the solid component on CT was also found in lepidic predominant adenocarcinoma compared to minimally invasive adenocarcinoma (median 10.5 vs 2 mm, p = .005). More invasive tumors had higher visual estimated percentage solid component compared to whole lesion measurement on CT (p = .014). CT and pathologic measurements were positively correlated, although only moderately (ρ = .66) for the maximum whole lesion size and fair (ρ = .49) for solid/invasive component maximum measurements. Larger whole lesion size and solid component size of lepidic predominant pattern adenocarcinomas are associated with lesion invasiveness, although radiologic and pathologic lesion measurements are only fair-moderately positively correlated.

What the Radiologist Should Know About Treatment of Peritoneal Malignancy.

OBJECTIVE: The purpose of this article is to discuss the role of the radiologist in the treatment of peritoneal cancer, with focus placed on advanced treatment options and selection of patients with resectable disease for whom complete cytoreduction can be achieved. CONCLUSION: Peritoneal cancers traditionally have been associated with significant morbidity and universal mortality; however, the management of such cancers has evolved substantially. Advanced treatment options, including cytoreductive surgery and intraperitoneal chemotherapy, are associated with significantly improved long-term patient survival. To ensure that patients benefit from aggressive multimodality treatments, the radiologist plays a pivotal role in the multidisciplinary team to ensure careful patient selection, identifying individuals with resectable disease for whom complete cytoreduction can be achieved.