An imaging-based model to predict the malignant potential of intraductal papillary mucinous neoplasm of the pancreas.

OBJECTIVES: To develop and validate imaging-based models for predicting the malignancy risk of intraductal papillary mucinous neoplasm (IPMN). MATERIALS AND METHODS: We retrospectively analyzed data from 241 IPMN patients who underwent preoperative CT and MRI for model development. Cyst size, presence and size of the enhancing mural nodule (EMN), main pancreatic duct (MPD) diameter, thickened/enhancing cyst wall, abrupt MPD caliber change with distal atrophy, and lymphadenopathy were assessed. Multiple logistic regression models predicting malignancy risk were created using either continuous (Model C) or dichotomized variables (Model D) using these imaging features. Validation included internal (n = 55) and external (n = 43) datasets. Model performance was assessed using the area under the receiver operating characteristic curve (AUC) and compared with that of the international guideline-based model (Model F). RESULTS: Model C identified age, EMN size, MPD diameter, and lymphadenopathy as independent predictors on CT, and age and presence and size of EMN on MRI. Model D identified age ≥ 68, cyst size ≥ 31 mm, EMN ≥ 6 mm, MPD ≥ 7 mm, and lymphadenopathy as independent predictors on CT, and age ≥ 68, EMN ≥ 4.5 mm, and lymphadenopathy on MRI. Model C (AUC, 0.763-0.899) performed slightly better than Model D (AUC, 0.753-0.912) without statistical significance. No significant difference was observed between Models C and F (AUC, 0.729-0.952). Combining Model C with obstructive jaundice improved performance (AUC, 0.802-0.941) without statistical significance. CONCLUSION: Our imaging-based models effectively predicted the malignancy risk of IPMNs, comparable to international consensus guidelines. CLINICAL RELEVANCE STATEMENT: Imaging features are important for predicting the malignant potential of IPMNs. Our imaging-based model may help determine surgical candidacy for patients with IPMNs. KEY POINTS: Non-invasively determining the malignant potential of intraductal papillary mucinous neoplasms (IPMNs) allows for appropriate treatment decision-making We identified multiple imaging features that are associated with malignant transformation and developed models for this prediction. Our model performs comparably with international consensus guidelines in predicting the malignant potential of IPMNs.

Application of a deep learning algorithm for three-dimensional T1-weighted gradient-echo imaging of gadoxetic acid-enhanced MRI in patients at a high risk of hepatocellular carcinoma.

PURPOSE: To evaluate the efficacy of a vendor-specific deep learning reconstruction algorithm (DLRA) in enhancing image quality and focal lesion detection using three-dimensional T1-weighted gradient-echo images in gadoxetic acid-enhanced liver magnetic resonance imaging (MRI) in patients at a high risk of hepatocellular carcinoma. MATERIALS AND METHODS: In this retrospective analysis, 83 high-risk patients with hepatocellular carcinoma underwent gadoxetic acid-enhanced liver MRI using a 3-T scanner. Triple arterial phase, high-resolution portal venous phase, and high-resolution hepatobiliary phase images were reconstructed using conventional reconstruction techniques and DLRA (AIRTM Recon DL; GE Healthcare) for subsequent comparison. Image quality and solid focal lesion detection were assessed by three abdominal radiologists and compared between conventional and DL methods. Focal liver lesion detection was evaluated using figures of merit (FOMs) from a jackknife alternative free-response receiver operating characteristic analysis on a per-lesion basis. RESULTS: DLRA-reconstructed images exhibited significantly improved overall image quality, image contrast, lesion conspicuity, vessel conspicuity, and liver edge sharpness and reduced subjective image noise, ringing artifacts, and motion artifacts compared to conventionally reconstructed images (all P < 0.05). Although there was no significant difference in the FOMs of non-cystic focal liver lesions between the conventional and DL methods, DLRA-reconstructed images showed notably higher pooled sensitivity than conventionally reconstructed images (P < 0.05) in all phases and higher detection rates for viable post-treatment HCCs in the arterial and hepatobiliary phases (all P < 0.05). CONCLUSIONS: Implementing DLRA can enhance the image quality in 3D T1-weighted gradient-echo sequences of gadoxetic acid-enhanced liver MRI examinations, leading to improved detection of viable post-treatment HCCs.

Development of a deep-learning model for classification of LI-RADS major features by using subtraction images of MRI: a preliminary study.

PURPOSE: Liver Imaging Reporting and Data System (LI-RADS) is limited by interreader variability. Thus, our study aimed to develop a deep-learning model for classifying LI-RADS major features using subtraction images using magnetic resonance imaging (MRI). METHODS: This single-center retrospective study included 222 consecutive patients who underwent resection for hepatocellular carcinoma (HCC) between January, 2015 and December, 2017. Subtraction arterial, portal venous, and transitional phase images of preoperative gadoxetic acid-enhanced MRI were used to train and test the deep-learning models. Initially, a three-dimensional (3D) nnU-Net-based deep-learning model was developed for HCC segmentation. Subsequently, a 3D U-Net-based deep-learning model was developed to assess three LI-RADS major features (nonrim arterial phase hyperenhancement [APHE], nonperipheral washout, and enhancing capsule [EC]), utilizing the results determined by board-certified radiologists as reference standards. The HCC segmentation performance was assessed using the Dice similarity coefficient (DSC), sensitivity, and precision. The sensitivity, specificity, and accuracy of the deep-learning model for classifying LI-RADS major features were calculated. RESULTS: The average DSC, sensitivity, and precision of our model for HCC segmentation were 0.884, 0.891, and 0.887, respectively, across all the phases. Our model demonstrated a sensitivity, specificity, and accuracy of 96.6% (28/29), 66.7% (4/6), and 91.4% (32/35), respectively, for nonrim APHE; 95.0% (19/20), 50.0% (4/8), and 82.1% (23/28), respectively, for nonperipheral washout; and 86.7% (26/30), 54.2% (13/24), and 72.2% (39/54) for EC, respectively. CONCLUSION: We developed an end-to-end deep-learning model that classifies the LI-RADS major features using subtraction MRI images. Our model exhibited satisfactory performance in classifying LI-RADS major features.

Abbreviated MRI for Secondary Surveillance of Recurrent Hepatocellular Carcinoma After Presumed Curative Treatment.

BACKGROUND: Little is known about the performance of abbreviated MRI (AMRI) for secondary surveillance of recurrent hepatocellular carcinoma (HCC) after curative treatment. PURPOSE: To evaluate the detection performance of AMRI for secondary surveillance of HCC after curative treatment. STUDY TYPE: Retrospective. POPULATION: A total of 243 patients (183 men and 60 women; median age, 65 years) who underwent secondary surveillance for HCC using gadoxetic acid-enhanced MRI after more than 2 year of disease-free period following curative treatment, including surgical resection or radiofrequency ablation (RFA). FIELD STRENGTH/SEQUENCE: A 3.0 T/noncontrast AMRI (NC-AMRI) (T2-weighted fast spin-echo, T1-weighted gradient echo, and diffusion-weighted images), hepatobiliary phase AMRI (HBP-AMRI) (T2-weighted fast spin-echo, diffusion-weighted, and HBP images), and full-sequence MRI ASSESSMENT: Four board-certified radiologists independently reviewed NC-AMRI, HBP-AMRI, and full-sequence MRI sets of each patient for detecting recurrent HCC. STATISTICAL TESTS: Per-lesion sensitivity, per-patient sensitivity and specificity for HCC detection at each set were compared using generalized estimating equation. RESULTS: A total of 42 recurred HCCs were confirmed in the 39 patients. The per-lesion and per-patient sensitivities did not show significant differences among the three image sets for either reviewer (P ≥ 0.358): per-lesion sensitivity: 59.5%-83.3%, 59.5%-85.7%, and 59.5%-83.3%, and per-patient sensitivity: 53.9%-83.3%, 56.4%-85.7%, and 53.9%-83.3% for NC-AMRI, HBP-AMRI, and full-sequence MRI, respectively. Per-lesion pooled sensitivities of NC-AMRI, HBP-AMRI, and full-sequence MRI were 72.6%, 73.2%, and 73.2%, with difference of -0.6% (95% confidence interval: -6.7, 5.5) between NC-AMRI and full-sequence MRI and 0.0% (-6.1, 6.1) between HBP-AMRI and full-sequence MRI. Per-patient specificity was not significantly different among the three image sets for both reviewers (95.6%-97.1%, 95.6%-97.1%, and 97.6%-98.5% for NC-AMRI and HBP-AMRI, respectively; P ≥ 0.117). DATA CONCLUSION: NC-AMRI and HBP-AMRI showed no significant difference in detection performance to that of full-sequence gadoxetic acid-enhanced MRI during secondary surveillance for HCC after more than 2-year disease free interval following curative treatment. Based on its good detection performance, short scan time, and lack of contrast agent-associated risks, NC-AMRI is a promising option for the secondary surveillance of HCC. EVIDENCE LEVEL: 3. TECHNICAL EFFICACY: Stage 2.

Quantitative Evaluation of Hepatic Steatosis Using Advanced Imaging Techniques: Focusing on New Quantitative Ultrasound Techniques.

Nonalcoholic fatty liver disease, characterized by excessive accumulation of fat in the liver, is the most common chronic liver disease worldwide. The current standard for the detection of hepatic steatosis is liver biopsy; however, it is limited by invasiveness and sampling errors. Accordingly, MR spectroscopy and proton density fat fraction obtained with MRI have been accepted as non-invasive modalities for quantifying hepatic steatosis. Recently, various quantitative ultrasonography techniques have been developed and validated for the quantification of hepatic steatosis. These techniques measure various acoustic parameters, including attenuation coefficient, backscatter coefficient and speckle statistics, speed of sound, and shear wave elastography metrics. In this article, we introduce several representative quantitative ultrasonography techniques and their diagnostic value for the detection of hepatic steatosis.

Imaging diagnosis of hepatocellular carcinoma: Future directions with special emphasis on hepatobiliary magnetic resonance imaging and contrast-enhanced ultrasound.

Hepatocellular carcinoma (HCC) is a unique cancer entity that can be noninvasively diagnosed using imaging modalities without pathologic confirmation. In 2018, several major guidelines for HCC were updated to include hepatobiliary contrast agent magnetic resonance imaging (HBA-MRI) and contrast-enhanced ultrasound (CEUS) as major imaging modalities for HCC diagnosis. HBA-MRI enables the achievement of high sensitivity in HCC detection using the hepatobiliary phase (HBP). CEUS is another imaging modality with real-time imaging capability, and it is reported to be useful as a second-line modality to increase sensitivity without losing specificity for HCC diagnosis. However, until now, there is an unsolved discrepancy among guidelines on whether to accept "HBP hypointensity" as a definite diagnostic criterion for HCC or include CEUS in the diagnostic algorithm for HCC diagnosis. Furthermore, there is variability in terminology and inconsistencies in the definition of imaging findings among guidelines; therefore, there is an unmet need for the development of a standardized lexicon. In this article, we review the performance and limitations of HBA-MRI and CEUS after guideline updates in 2018 and briefly introduce some future aspects of imaging-based HCC diagnosis.

Image quality in liver CT: low-dose deep learning vs standard-dose model-based iterative reconstructions.

OBJECTIVES: To compare the overall image quality and detectability of significant (malignant and pre-malignant) liver lesions of low-dose liver CT (LDCT, 33.3% dose) using deep learning denoising (DLD) to standard-dose CT (SDCT, 100% dose) using model-based iterative reconstruction (MBIR). METHODS: In this retrospective study, CT images of 80 patients with hepatic focal lesions were included. For noninferiority analysis of overall image quality, a margin of - 0.5 points (scored in a 5-point scale) for the difference between scan protocols was pre-defined. Other quantitative or qualitative image quality assessments were performed. Additionally, detectability of significant liver lesions was compared, with 64 pairs of CT, using the jackknife alternative free-response ROC analysis, with noninferior margin defined by the lower limit of 95% confidence interval (CI) of the difference of figure-of-merit less than - 0.1. RESULTS: The mean overall image quality scores with LDCT and SDCT were 3.77 ± 0.38 and 3.94 ± 0.34, respectively, demonstrating a difference of - 0.17 (95% CI: - 0.21 to - 0.12), which did not cross the predefined noninferiority margin of - 0.5. Furthermore, LDCT showed significantly superior quantitative results of liver lesion contrast to noise ratio (p < 0.05). However, although LDCT scored higher than the average score in qualitative image quality assessments, they were significantly lower than those of SDCT (p < 0.05). Figure-of-merit for lesion detection was 0.859 for LDCT and 0.878 for SDCT, showing noninferiority (difference: - 0.019, 95% CI: - 0.058 to 0.021). CONCLUSION: LDCT using DLD with 67% radiation dose reduction showed non-inferior overall image quality and lesion detectability, compared to SDCT. KEY POINTS: • Low-dose liver CT using deep learning denoising (DLD), at 67% dose reduction, provided non-inferior overall image quality compared to standard-dose CT using model-based iterative reconstruction (MBIR). • Low-dose CT using DLD showed significantly less noise and higher CNR lesion to liver than standard-dose CT using MBIR and demonstrated at least average image quality score among all readers, albeit with lower scores than standard-dose CT using MBIR. • Low-dose liver CT showed noninferior detectability for malignant and pre-malignant liver lesions, compared to standard-dose CT.

Volumetric CT Texture Analysis of Intrahepatic Mass-Forming Cholangiocarcinoma for the Prediction of Postoperative Outcomes: Fully Automatic Tumor Segmentation Versus Semi-Automatic Segmentation.

OBJECTIVE: To determine whether volumetric CT texture analysis (CTTA) using fully automatic tumor segmentation can help predict recurrence-free survival (RFS) in patients with intrahepatic mass-forming cholangiocarcinomas (IMCCs) after surgical resection. MATERIALS AND METHODS: This retrospective study analyzed the preoperative CT scans of 89 patients with IMCCs (64 male; 25 female; mean age, 62.1 years; range, 38-78 years) who underwent surgical resection between January 2005 and December 2016. Volumetric CTTA of IMCCs was performed in late arterial phase images using both fully automatic and semi-automatic liver tumor segmentation techniques. The time spent on segmentation and texture analysis was compared, and the first-order and second-order texture parameters and shape features were extracted. The reliability of CTTA parameters between the techniques was evaluated using intraclass correlation coefficients (ICCs). Intra- and interobserver reproducibility of volumetric CTTAs were also obtained using ICCs. Cox proportional hazard regression were used to predict RFS using CTTA parameters and clinicopathological parameters. RESULTS: The time spent on fully automatic tumor segmentation and CTTA was significantly shorter than that for semi-automatic segmentation: mean ± standard deviation of 1 minutes 37 seconds ± 50 seconds vs. 10 minutes 48 seconds ± 13 minutes 44 seconds (p < 0.001). ICCs of the texture features between the two techniques ranged from 0.215 to 0.980. ICCs for the intraobserver and interobserver reproducibility using fully automatic segmentation were 0.601-0.997 and 0.177-0.984, respectively. Multivariable analysis identified lower first-order mean (hazard ratio [HR], 0.982; p = 0.010), larger pathologic tumor size (HR, 1.171; p < 0.001), and positive lymph node involvement (HR, 2.193; p = 0.014) as significant parameters for shorter RFS using fully automatic segmentation. CONCLUSION: Volumetric CTTA parameters obtained using fully automatic segmentation could be utilized as prognostic markers in patients with IMCC, with comparable reproducibility in significantly less time compared with semi-automatic segmentation.

Effectiveness of radiologist training in improving reader agreement for Lung-RADS 4X categorization.

OBJECTIVES: To identify the agreement on Lung CT Screening Reporting and Data System 4X categorization between radiologists and an expert-adjudicated reference standard and to investigate whether training led to improvement of the agreement measures and diagnostic potential for lung cancer. METHODS: Category 4 nodules in the Korean Lung Cancer Screening Project were identified retrospectively, and each 4X nodule was matched with one 4A or 4B nodule. An expert panel re-evaluated the categories and determined the reference standard. Nineteen radiologists were asked to determine the presence of CT features of malignancy and 4X categorization for each nodule. A review was performed in two sessions, and training material was given after session 1. Agreement on 4X categorization between radiologists and the expert-adjudicated reference standard and agreement between radiologist-assessed 4X categorization and lung cancer diagnosis were evaluated. RESULTS: The 48 expert-adjudicated 4X nodules and 64 non-4X nodules were evenly distributed in each session. The proportion of category 4X decreased after training (56.4% ± 16.9% vs. 33.4% ± 8.0%; p < 0.001). Cohen's κ indicated poor agreement (0.39 ± 0.16) in session 1, but agreement improved in session 2 (0.47 ± 0.09; p = 0.03). The increase in agreement in session 2 was observed among inexperienced radiologists (p < 0.05), and experienced and inexperienced reviewers exhibited comparable agreement performance in session 2 (p > 0.05). All agreement measures between radiologist-assessed 4X categorization and lung cancer diagnosis increased in session 2 (p < 0.05). CONCLUSION: Radiologist training can improve reader agreement on 4X categorization, leading to enhanced diagnostic performance for lung cancer. KEY POINTS: • Agreement on 4X categorization between radiologists and an expert-adjudicated reference standard was initially poor, but improved significantly after training. • The mean proportion of 4X categorization by 19 radiologists decreased from 56.4% ± 16.9% in session 1 to 33.4% ± 8.0% in session 2. • All agreement measures between the 4X categorization and lung cancer diagnosis increased significantly in session 2, implying that appropriate training and guidance increased the diagnostic potential of category 4X.

CT quantification of the heterogeneity of fibrosis boundaries in idiopathic pulmonary fibrosis.

OBJECTIVES: To quantify the heterogeneity of fibrosis boundaries in idiopathic pulmonary fibrosis (IPF) using the Gaussian curvature analysis for evaluating disease severity and predicting survival. METHODS: We retrospectively included 104 IPF patients and 52 controls who underwent baseline chest CT scans. Normal lungs below - 500 HU were segmented, and the boundary was three-dimensionally reconstructed using in-house software. Gaussian curvature analysis provided histogram features on the heterogeneity of the fibrosis boundary. We analyzed the correlations between histogram features and the gender-age-physiology (GAP) and CT fibrosis scores. We built a regression model to predict diffusing capacity of carbon monoxide (DLCO) using the histogram features and calculated the modified GAP (mGAP) score by replacing DLCO with the predicted DLCO. The performances of the GAP, CT-GAP, and mGAP scores were compared using 100 repeated random-split sets. RESULTS: Patients with moderate-to-severe IPF had more numerous Gaussian curvatures at the fibrosis boundary, lower uniformity, and lower 10th to 30th percentiles of Gaussian curvature than controls or patients with mild IPF (all p < 0.0033). The 20th percentile was most significantly correlated with the GAP score (r = - 0.357; p < 0.001) and the CT fibrosis score (r = - 0.343; p = 0.001). More numerous Gaussian curvatures, higher entropy, lower uniformity, and 10th to 30th percentiles (p < 0.001-0.041) were associated with mortality. The mGAP score was comparable to the GAP and CT-GAP scores for survival prediction (mean C-indices, 0.76 vs. 0.79 vs. 0.77, respectively). CONCLUSIONS: Gaussian curvatures of fibrosis boundaries became more heterogeneous as the disease progressed, and heterogeneity was negatively associated with survival in IPF. KEY POINTS: • Gaussian curvature of the fibrotic lung boundary was more heterogeneous in patients with moderate-to-severe IPF than those with mild IPF or normal controls. • The 20th percentile of the Gaussian curvature of the fibrosis boundary was linearly correlated with the GAP score and the CT fibrosis score. • A modified GAP score that replaced the diffusing capacity of carbon monoxide with a composite measure using histogram features of the Gaussian curvature of the fibrosis boundary showed a comparable ability to predict survival to both the GAP and the CT-GAP score.

Cochlear duct length and cochlear distance on preoperative CT: imaging markers for estimating insertion depth angle of cochlear implant electrode.

OBJECTIVES: Preoperative estimation of the insertion depth angle of cochlear implant (CI) electrodes is essential for surgical planning. The purpose of this study was to determine the cochlear size using preoperative CT and to investigate the correlation between cochlear size and insertion depth angle in morphologically normal cochlea. METHODS: Thirty-five children who underwent CI were included in this study. Cochlear duct length (CDL) and the diameter of the cochlear basal turn (distance A/B) on preoperative CT and the insertion depth angle of the CI electrode on postoperative radiographs were independently measured by two readers. Correlation between cochlear size and insertion depth angle was evaluated. Interobserver agreement was calculated using the intraclass correlation coefficient (ICC). RESULTS: The mean CDL, distance A, and distance B of 70 ears were 36.20 ± 1.57 mm, 8.67 ± 0.42 mm, and 5.73 ± 0.32 mm, respectively. The mean insertion depth angle was 431.45 ± 38.42°. Interobserver agreements of CDL, distance A/B, and insertion depth angle were fair to excellent (ICC 0.864, 0.862, 0.529, and 0.958, respectively). Distance A (r = - 0.7643) and distance B (r = - 0.7118) showed a negative correlation with insertion depth angle, respectively (p < 0.0001). However, the correlation between CDL and insertion depth angle was not statistically significant (r = - 0.2333, p > 0.05). CONCLUSIONS: The CDL and cochlear distance can be reliably obtained from preoperative CT. Distance A can be used as a predictive marker for estimating insertion depth angle during CI surgery.

Hepatocellular Carcinoma: Texture Analysis of Preoperative Computed Tomography Images Can Provide Markers of Tumor Grade and Disease-Free Survival.

OBJECTIVE: To investigate the usefulness of computed tomography (CT) texture analysis (CTTA) in estimating histologic tumor grade and in predicting disease-free survival (DFS) after surgical resection in patients with hepatocellular carcinoma (HCC). MATERIALS AND METHODS: Eighty-one patients with a single HCC who had undergone quadriphasic liver CT followed by surgical resection were enrolled. Texture analysis of tumors on preoperative CT images was performed using commercially available software. The mean, mean of positive pixels (MPP), entropy, kurtosis, skewness, and standard deviation (SD) of the pixel distribution histogram were derived with and without filtration. The texture features were then compared between groups classified according to histologic grade. Kaplan-Meier and Cox proportional hazards analyses were performed to determine the relationship between texture features and DFS. RESULTS: SD and MPP quantified from fine to coarse textures on arterial-phase CT images showed significant positive associations with the histologic grade of HCC (p < 0.05). Kaplan-Meier analysis identified most CT texture features across the different filters from fine to coarse texture scales as significant univariate markers of DFS. Cox proportional hazards analysis identified skewness on arterial-phase images (fine texture scale, spatial scaling factor [SSF] 2.0, p < 0.001; medium texture scale, SSF 3.0, p < 0.001), tumor size (p = 0.001), microscopic vascular invasion (p = 0.034), rim arterial enhancement (p = 0.024), and peritumoral parenchymal enhancement (p = 0.010) as independent predictors of DFS. CONCLUSION: CTTA was demonstrated to provide texture features significantly correlated with higher tumor grade as well as predictive markers of DFS after surgical resection of HCCs in addition to other valuable imaging and clinico-pathologic parameters.

Inspiratory Lung Expansion in Patients with Interstitial Lung Disease: CT Histogram Analyses.

This study aimed to evaluate inspiratory lung expansion in patients with interstitial lung disease (ILD) using histogram analyses based on advanced image registration between inspiratory and expiratory CT scans. We included 16 female ILD patients and eight age- and sex-matched normal controls who underwent full-inspiratory and expiratory CT scans. The CT scans were sequentially aligned based on the surface, landmarks, and attenuation of the lung parenchyma. Histogram analyses were performed on the degree of lung expansion (DLE) of each pixel between the aligned images in x-, y-, z-axes, and 3-dimensionally (3D). The overall mean registration error was 1.9 mm between the CT scans. The DLE3D in ILD patients was smaller than in the controls (mean, 17.6 mm vs. 26.9 mm; p = 0.023), and less heterogeneous in terms of standard deviation, entropy, and uniformity (p < 0.05). These results were mainly due to similar results in the DLEZ of the lower lungs. A forced vital capacity tended to be weakly correlated with mean (r2 = 0.210; p = 0.074), and histogram parameters (r2 = 0.194-0.251; p = 0.048-0.100) of the DLE3D in the lower lung in ILD patients. Our findings indicate that reduced and less heterogeneous inspiratory lung expansion in ILD patients can be identified by using advanced accurate image registration.