A clinical evaluation study of cardiothoracic ratio measurement using artificial intelligence.

BACKGROUND: Artificial intelligence, particularly the deep learning (DL) model, can provide reliable results for automated cardiothoracic ratio (CTR) measurement on chest X-ray (CXR) images. In everyday clinical use, however, this technology is usually implemented in a non-automated (AI-assisted) capacity because it still requires approval from radiologists. We investigated the performance and efficiency of our recently proposed models for the AI-assisted method intended for clinical practice. METHODS: We validated four proposed DL models (AlbuNet, SegNet, VGG-11, and VGG-16) to find the best model for clinical implementation using a dataset of 7517 CXR images from manual operations. These models were investigated in single-model and combined-model modes to find the model with the highest percentage of results where the user could accept the results without further interaction (excellent grade), and with measurement variation within ± 1.8% of the human-operating range. The best model from the validation study was then tested on an evaluation dataset of 9386 CXR images using the AI-assisted method with two radiologists to measure the yield of excellent grade results, observer variation, and operating time. A Bland-Altman plot with coefficient of variation (CV) was employed to evaluate agreement between measurements. RESULTS: The VGG-16 gave the highest excellent grade result (68.9%) of any single-model mode with a CV comparable to manual operation (2.12% vs 2.13%). No DL model produced a failure-grade result. The combined-model mode of AlbuNet + VGG-11 model yielded excellent grades in 82.7% of images and a CV of 1.36%. Using the evaluation dataset, the AlbuNet + VGG-11 model produced excellent grade results in 77.8% of images, a CV of 1.55%, and reduced CTR measurement time by almost ten-fold (1.07 ± 2.62 s vs 10.6 ± 1.5 s) compared with manual operation. CONCLUSION: Due to its excellent accuracy and speed, the AlbuNet + VGG-11 model could be clinically implemented to assist radiologists with CTR measurement.

An automated liver segmentation in liver iron concentration map using fuzzy c-means clustering combined with anatomical landmark data.

BACKGROUND: To estimate median liver iron concentration (LIC) calculated from magnetic resonance imaging, excluded vessels of the liver parenchyma region were defined manually. Previous works proposed the automated method for excluding vessels from the liver region. However, only user-defined liver region remained a manual process. Therefore, this work aimed to develop an automated liver region segmentation technique to automate the whole process of median LIC calculation. METHODS: 553 MR examinations from 471 thalassemia major patients were used in this study. LIC maps (in mg/g dry weight) were calculated and used as the input of segmentation procedures. Anatomical landmark data were detected and used to restrict ROI. After that, the liver region was segmented using fuzzy c-means clustering and reduced segmentation errors by morphological processes. According to the clinical application, erosion with a suitable size of the structuring element was applied to reduce the segmented liver region to avoid uncertainty around the edge of the liver. The segmentation results were evaluated by comparing with manual segmentation performed by a board-certified radiologist. RESULTS: The proposed method was able to produce a good grade output in approximately 81% of all data. Approximately 11% of all data required an easy modification step. The rest of the output, approximately 8%, was an unsuccessful grade and required manual intervention by a user. For the evaluation matrices, percent dice similarity coefficient (%DSC) was in the range 86-92, percent Jaccard index (%JC) was 78-86, and Hausdorff distance (H) was 14-28 mm, respectively. In this study, percent false positive (%FP) and percent false negative (%FN) were applied to evaluate under- and over-segmentation that other evaluation matrices could not handle. The average of operation times could be reduced from 10 s per case using traditional method, to 1.5 s per case using our proposed method. CONCLUSION: The experimental results showed that the proposed method provided an effective automated liver segmentation technique, which can be applied clinically for automated median LIC calculation in thalassemia major patients.

Cardiothoracic ratio measurement using artificial intelligence: observer and method validation studies.

BACKGROUND: Artificial Intelligence (AI) is a promising tool for cardiothoracic ratio (CTR) measurement that has been technically validated but not clinically evaluated on a large dataset. We observed and validated AI and manual methods for CTR measurement using a large dataset and investigated the clinical utility of the AI method. METHODS: Five thousand normal chest x-rays and 2,517 images with cardiomegaly and CTR values, were analyzed using manual, AI-assisted, and AI-only methods. AI-only methods obtained CTR values from a VGG-16 U-Net model. An in-house software was used to aid the manual and AI-assisted measurements and to record operating time. Intra and inter-observer experiments were performed on manual and AI-assisted methods and the averages were used in a method variation study. AI outcomes were graded in the AI-assisted method as excellent (accepted by both users independently), good (required adjustment), and poor (failed outcome). Bland-Altman plot with coefficient of variation (CV), and coefficient of determination (R-squared) were used to evaluate agreement and correlation between measurements. Finally, the performance of a cardiomegaly classification test was evaluated using a CTR cutoff at the standard (0.5), optimum, and maximum sensitivity. RESULTS: Manual CTR measurements on cardiomegaly data were comparable to previous radiologist reports (CV of 2.13% vs 2.04%). The observer and method variations from the AI-only method were about three times higher than from the manual method (CV of 5.78% vs 2.13%). AI assistance resulted in 40% excellent, 56% good, and 4% poor grading. AI assistance significantly improved agreement on inter-observer measurement compared to manual methods (CV; bias: 1.72%; - 0.61% vs 2.13%; - 1.62%) and was faster to perform (2.2 ± 2.4 secs vs 10.6 ± 1.5 secs). The R-squared and classification-test were not reliable indicators to verify that the AI-only method could replace manual operation. CONCLUSIONS: AI alone is not yet suitable to replace manual operations due to its high variation, but it is useful to assist the radiologist because it can reduce observer variation and operation time. Agreement of measurement should be used to compare AI and manual methods, rather than R-square or classification performance tests.

Prevalence of left ventricular diastolic dysfunction by cardiac magnetic resonance imaging in thalassemia major patients with normal left ventricular systolic function.

BACKGROUND: The leading cause of mortality of thalassemia major patients is iron overload cardiomyopathy. Early diagnosis with searching for left ventricular diastolic dysfunction before the systolic dysfunction ensued might yield better prognosis. This study aimed to define the prevalence of the left ventricular diastolic dysfunction (LVDD) in thalassemia major patients with normal left ventricular systolic function and the associated factors. METHODS: Adult thalassemia major patients with normal left ventricular systolic function who were referred for cardiac T2* at Siriraj Hospital - Thailand's largest national tertiary referral center - during the October 2014 to January 2017 study period. Left ventricular diastolic function was defined by mitral valve filling parameters and left atrial volume index using CMR. Patients with moderate to severe valvular heart disease, pericardial disease, or incomplete data were excluded. Baseline characteristics, comorbid diseases, current medication, and laboratory results were recorded and analyzed. RESULTS: One hundred and sixteen patients were included, with a mean age of 27.5 ± 13.5 years, 57.8% were female, and 87.9% were transfusion dependent. Proportions of homozygous beta-thalassemia and beta-thalassemia hemoglobin E were 12.1 and 87.9%, respectively. The baseline hematocrit was 26.3 ± 3.3%. The prevalence of LVDD was 20.7% (95% CI: 13.7-29.2%). Cardiac T2* was abnormal in 7.8% (95% CI: 3.6-14.2%). Multivariate analysis revealed age, body surface area, homozygous beta-thalassemia, splenectomy, heart rate, and diastolic blood pressure to be significantly associated with LVDD. CONCLUSIONS: LVDD already exists from the early stages of the disease before the abnormal heart T2 * is detected. Homozygous beta-thalassemia and splenectomy were strong predictors of LVDD. These data may increase awareness of the disease, especially in the high risk groups.

An Automated Segmentation of R2* Iron-Overloaded Liver Images Using a Fuzzy C-Mean Clustering Scheme.

OBJECTIVES: The objectives of this study were to develop and test an automated segmentation of R2* iron-overloaded liver images using fuzzy c-mean (FCM) clustering and to evaluate the observer variations. MATERIALS AND METHODS: Liver R2* images and liver iron concentration (LIC) maps of 660 thalassemia examinations were randomly separated into training (70%) and testing (30%) cohorts for development and evaluation purposes, respectively. Two-dimensional FCM used R2* images, and the LIC map was implemented to segment vessels from the parenchyma. Two automated FCM variables were investigated using new echo time and membership threshold selection criteria based on the FCM centroid distance and LIC levels, respectively. The new method was developed on a training cohort and compared with manual segmentation for segmentation accuracy and to a previous semiautomated method, and a semiautomated scheme was suggested to improve unsuccessful results. The automated variables found from the training cohort were assessed for their effectiveness in the testing cohort, both quantitatively and qualitatively (the latter by 2 abdominal radiologists using a grading method, with evaluations of observer variations). A segmentation error of less than 30% was considered to be a successful result in both cohorts, whereas, in the testing cohort, a good grade obtained from satisfactory automated results was considered a success. RESULTS: The centroid distance method has a segmentation accuracy comparable with the previous-best, semiautomated method. About 94% and 90% of the examinations in the training and testing cohorts were automatically segmented out successfully, respectively. The failed examinations were successfully segmented out with thresholding adjustment (3% and 8%) or by using alternative results from the previous 1-dimensional FCM method (3% and 2%) in the training and testing cohorts, respectively. There were no failed segmentation examinations in either cohort. The intraobserver and interobserver variabilities were found to be in substantial agreement. CONCLUSIONS: Our new method provided a robust automated segmentation outcome with a high ease of use for routine clinical application.

Anti-EpCAM scFv gadolinium chelate: a novel targeted MRI contrast agent for imaging of colorectal cancer.

OBJECTIVES: The development of targeted contrast agents for magnetic resonance imaging (MRI) facilitates enhanced cancer imaging and more accurate diagnosis. In the present study, a novel contrast agent was developed by conjugating anti-EpCAM humanized scFv with gadolinium chelate to achieve target specificity. MATERIALS AND METHODS: The material design strategy involved site-specific conjugation of the chelating agent to scFv. The scFv monomer was linked to maleimide-DTPA via unpaired cysteine at the scFv C-terminus, followed by chelation with gadolinium (Gd). Successful scFv-DTPA conjugation was achieved at 1:10 molar ratio of scFv to maleimide-DTPA at pH 6.5. The developed anti-EpCAM-Gd-DTPA MRI contrast agent was evaluated for cell targeting ability, in vitro serum stability, cell cytotoxicity, relaxivity, and MR contrast enhancement. RESULTS: A high level of targeting efficacy of anti-EpCAM-Gd-DTPA to an EpCAM-overexpressing HT29 colorectal cell was demonstrated by confocal microscopy. Good stability of the contrast agent was obtained and no cytotoxicity was observed in HT29 cells after 48 h incubation with 25-100 µM of Gd. Favorable imaging was obtained using anti-EpCAM-Gd-DTPA, including 1.8-fold enhanced relaxivity compared with Gd-DTPA, and MR contrast enhancement observed after binding to HT29. CONCLUSION: The potential benefit of this contrast agent for in vivo MR imaging of colorectal cancer, as well as other EpCAM positive cancers, is suggested and warrants further investigation.

Prevalence and predictors of cardiac and liver iron overload in patients with thalassemia: A multicenter study based on real-world data.

Prevalence of cardiac and liver iron overload in patients with thalassemia in real-world practice may vary among different regions especially in the era of widely-used iron chelation therapy. The aim of this study was to determine the prevalence of cardiac and liver iron overload in and the management patterns of patients with thalassemia in real-world practice in Thailand. We established a multicenter registry for patients with thalassemia who underwent magnetic resonance imaging (MRI) as part of their clinical evaluation. All enrolled patients underwent cardiac and liver MRI for assessment of iron overload. There were a total of 405 patients enrolled in this study. The mean age of patients was 18.8±12.5years and 46.7% were male. Two hundred ninety-six (73.1%) of patients received regular blood transfusion. Prevalence of cardiac iron overload (CIO) and liver iron overload (LIO) was 5.2% and 56.8%, respectively. Independent predictors for iron overload from laboratory information were serum ferritin and transaminase for both CIO and LIO. Serum ferritin can be used as a screening tool to rule-out CIO and to diagnose LIO. Iron chelation therapy was given in 74.6%; 15.3% as a combination therapy.

Stimulated echo diffusion weighted imaging of the liver at 3 Tesla.

PURPOSE: Diffusion time (Δ) effect in diffusion measurements has been validated as a sensitive biomarker in liver fibrosis by rat models. To extend this finding to clinical study, a reliable imaging technique is highly desirable. This study aimed to develop an optimal stimulated echo acquisition mode (STEAM) diffusion-weighted imaging (DWI) method dedicated to human liver imaging on 3 Tesla (T) and preliminarily investigate the dependence effect in healthy volunteers. METHODS: STEAM DWI with single-shot echo planar imaging readout was used as it provided better signal-to-noise ratio (SNR) than spin echo DWI methods when a long Δ was needed for liver imaging. Additionally, a slice-selection gradient reversal method was used for fat suppression. Motion compensation and SNR improvement strategies were used to further improve the image quality. Five b-values with three Δs were tested in 10 volunteers. RESULTS: Effective fat suppression and motion compensation were reproducibly achieved in the optimized sequence. The signal decay generally became slower when the Δs increased. Obvious reduction of diffusion coefficients was observed with increasing Δs in the liver. CONCLUSION: The results verified the Δ dependence in diffusion measurements, indicating restricted diffusion in healthy human livers for the first time at 3T. This prepared STEAM DWI a potential technique for liver fibrotic studies in clinical practice. Magn Reson Med 77:300-309, 2017. © 2016 Wiley Periodicals, Inc.

Improved R2* liver iron concentration assessment using a novel fuzzy c-mean clustering scheme.

BACKGROUND: In thalassemia patients, R2* liver iron concentration (LIC) measurement is a common clinical tool for assessing iron overload and for determining necessary chelator dose and evaluating its efficacy. Despite the importance of accurate LIC measurement, existing methods suffer from LIC variability, especially at the severe iron overload range due to inclusion of vessel parts in LIC calculation. In this study, we build upon previous Fuzzy C-Mean (FCM) clustering work to formulate a scheme with superior performance in segmenting vessel pixels from the parenchyma. Our method (MIX-FCM) combines our novel 2D-FCM with the existing 1D-FCM algorithm. This study further assessed possible optimal clustering parameters (OP scheme) and proposed a semi-automatic (SA) scheme for routine clinical application. METHODS: Segmentation of liver parenchyma and vessels was performed on T2* images and their LIC maps in 196 studies from 147 thalassemia major patients. We used manual segmentation as the reference. 1D-FCM clustering was performed on the acquired image alone and 2D-FCM used both the acquired image and its LIC data. To execute the MIX-FCM method, the best outcome (OP-MIX-FCM) was selected from the aforementioned methods and was compared to the SA-MIX-FCM scheme. We used the percent value of the normalized interquartile range (nIQR) to its median to evaluate the variability of all methods. RESULTS: 2D-FCM clustering is more effective than 1D-FCM clustering at the severe overload range only, but inferior for other ranges (where 1D-FCM provides suitable results). This complementary performance between the two methods allows MIX-FCM to improve results for all ranges. OP-MIX-FCM clustering error was 2.1 ± 2.3%, compared with 10.3 ± 9.9% and 7.0 ± 11.9% from 1D- and 2D-FCM clustering, respectively. SA-MIX-FCM result was comparable to OP-MIX-FCM result, with both schemes showing ability to decrease overall nIQR by approximately 30%. CONCLUSION: Our proposed 2D-FCM algorithm is not as superior to 1D-FCM as hypothesized. In contrast, our MIX-FCM method benefits from the best of both methods to obtain the highest segmentation accuracy at all ranges. Moreover, segmentation accuracy of the practical scheme (SA-MIX-FCM) is comparable to segmentation accuracy of the reference scheme (OP-MIX-FCM). Finally, we confirmed that segmentation is crucial to improving LIC assessments, especially at the severe iron overload range.

Benefit of double contrast MRI in diagnosis of hepatocellular carcinoma in patients with chronic liver diseases.

OBJECTIVE: To assess the benefit on diagnosis of hepatocellular carcinoma (HCC) in patients with chronic liver disease or cirrhosis with double contrast MR imaging compared to the routine gadolinium-based MR imaging. MATERIAL AND METHOD: Seventy-one consecutive patients with cirrhosis or chronic hepatitis underwent multiphase, gadolinium-enhanced liver MRI examination and sequentially superparamagnetic iron oxide (SPIO)-enhanced images. The presence signal intensities of lesions on non-contrast sequences, dynamic gadolinium-enhanced images and delayed 10-min post-SPIO T2*-weighted images were recorded. RESULTS: Among 27 patients, 15 HCCs from 12 patients were diagnosed by surgical (n = 7) and non-surgical (n = 8) proofs. The overall sensitivity, specificity, positive predictive value, and negative predictive value of double contrast-enhanced images in 12 patients were 83.3% (95% CI: 58.5, 96.2), 33.3% (95% CI: 5.4, 88.4), 88.2% (95% CI: 63.5, 98.2), and 25% (95% CI: 4.1, 79.6) and these of gadolinium-enhanced images were 72.2% (95% CI: 46.5, 90.2), 33.3% (95% CI: 5.4, 88.4), 86.6% (95% CI: 59.5, 97.9), and 16.6% (95% CI: 2.7, 63.9), respectively. There were two benign hepatic nodules (1 adenoma, 1 dysplastic nodule) suspected as HCCs on MR images and two surgically proven-HCCs, invisible on gadolinium-enhanced images, detected as defect on only delayed 10-min post-SPIO T2*-weighted images. CONCLUSION: SPIO-enhanced images in double contrast-enhanced MR imaging had an additional value on HCC detection, compared to gadolinium-enhanced MR imaging, in patients with chronic liver disease or cirrhosis.

Prevalence and characters of anomalous coronary artery from coronary magnetic resonance angiography.

BACKGROUND: Many types of anomalous coronary artery have been reported. Some forms of the anomaly are potentially malignant and can lead to sudden death. OBJECTIVE: To determine the prevalence and characters of anomalous coronary artery, including the associations of myocardial ischemia. MATERIAL AND METHOD: This is a retrospective study. The authors enrolled patients who were referred for cardiac magnetic resonance (CMR) and had magnetic resonance coronary angiography (MRCA) images. Imaging of the coronary arteries was acquired. The presence and patterns of anomalous coronary artery and the presence of myocardial ischemia was recorded. Myocardial perfusion study was also performed in most patients using adenosine stress test. RESULTS: Anomalous coronary artery was detected in 56 out of 3,703 patients (1.51%). There were 24 men (42.9%). Average age was 62.1 +/- 15.0 years. Most common type was right coronary artery (RCA) from left coronary cusp. Malignant form was demonstrated in 31 patients (55.4%) and myocardial ischemia was detected in 10 patients (23.3%). CONCLUSION: Prevalence of anomalous coronary artery was 1.5%. Most common types were RCA from left coronary cusp (30%) and high take-off RCA (30%).

Deep learning for tubes and lines detection in critical illness: Generalizability and comparison with residents.

Background: Artificial intelligence (AI) has been proven useful for the assessment of tubes and lines on chest radiographs of general patients. However, validation on intensive care unit (ICU) patients remains imperative. Methods: This retrospective case-control study evaluated the performance of deep learning (DL) models for tubes and lines classification on both an external public dataset and a local dataset comprising 303 films randomly sampled from the ICU database. The endotracheal tubes (ETTs), central venous catheters (CVCs), and nasogastric tubes (NGTs) were classified into "Normal," "Abnormal," or "Borderline" positions by DL models with and without rule-based modification. Their performance was evaluated using an experienced radiologist as the standard reference. Results: The algorithm showed decreased performance on the local ICU dataset, compared to that of the external dataset, decreasing from the Area Under the Curve of Receiver (AUC) of 0.967 (95 % CI 0.965-0.973) to the AUC of 0.70 (95 % CI 0.68-0.77). Significant improvement in the ETT classification task was observed after modifications were made to the model to allow the use of the spatial relationship between line tips and reference anatomy with the improvement of the AUC, increasing from 0.71 (95 % CI 0.70 - 0.75) to 0.86 (95 % CI 0.83 - 0.94). Conclusions: The externally trained model exhibited limited generalizability on the local ICU dataset. Therefore, evaluating the performance of externally trained AI before integrating it into critical care routine is crucial. Rule-based algorithm may be used in combination with DL to improve results.

Sarcopenia as a Predictive Factor for Recurrence of Hepatocellular Carcinoma Following Radiofrequency Ablation.

BACKGROUND: Sarcopenia is a skeletal muscle mass deficiency and a potential prognostic factor for the recurrence of hepatocellular carcinoma (HCC). OBJECTIVE: To determine whether sarcopenia correlates with the recurrence rate of HCC after curative radiofrequency ablation (RFA) in early and very early HCC. METHODS: We retrospectively reviewed 669 HCC patients who underwent their first curative RFA at Siriraj hospital from 2011 to 2020. Fifty-six patients who were diagnosed with HCC by triple-phase CT scan and had complete response on follow-up CT were included. All patients underwent skeletal muscle index (SMI) assessment at level L3 vertebra and sarcopenia was defined by the cut-off values of 52.4 cm2/m2 for men and 38.5 cm2/m2 for women. We compared patients with and without sarcopenia. Time to recurrence was evaluated by the Kaplan-Meier method. Univariate and multivariate Cox regression analysis was performed. RESULTS: Sarcopenia was present in 37 of 56 patients (66.1%). There was no significant difference between groups except body mass index (BMI) (P<0.001) and serum alanine aminotransferase (ALT) (P=0.035). There was a promising result indicating the difference of time to recurrence between each group (P=0.046) and potential association of sarcopenia with HCC recurrence (HR=2.06; P=0.052). The Child-Pugh score and tumor number were independent risk factors for HCC recurrence (HR=2.04; P=0.005 and HR=2.68; P=0.017, respectively). CONCLUSION: Sarcopenia is a potential prognostic factor for recurrence of HCC in Thai patients who underwent RFA. A larger study is required to properly confirm this association.

Tumor-to-bone distance and radiomic features on MRI distinguish intramuscular lipomas from well-differentiated liposarcomas.

BACKGROUND: To develop a machine learning model based on tumor-to-bone distance and radiomic features derived from preoperative MRI images to distinguish intramuscular (IM) lipomas and atypical lipomatous tumors/well-differentiated liposarcomas (ALTs/WDLSs) and compared with radiologists. METHODS: The study included patients with IM lipomas and ALTs/WDLSs diagnosed between 2010 and 2022, and with MRI scans (sequence/field strength: T1-weighted (T1W) imaging at 1.5 or 3.0 Tesla MRI). Manual segmentation of tumors based on the three-dimensional T1W images was performed by two observers to appraise the intra- and interobserver variability. After radiomic features and tumor-to-bone distance were extracted, it was used to train a machine learning model to distinguish IM lipomas and ALTs/WDLSs. Both feature selection and classification steps were performed using Least Absolute Shrinkage and Selection Operator logistic regression. The performance of the classification model was assessed using a tenfold cross-validation strategy and subsequently evaluated using the receiver operating characteristic curve (ROC) analysis. The classification agreement of two experienced musculoskeletal (MSK) radiologists was assessed using the kappa statistics. The diagnosis accuracy of each radiologist was evaluated using the final pathological results as the gold standard. Additionally, we compared the performance of the model and two radiologists in terms of the area under the receiver operator characteristic curves (AUCs) using the Delong's test. RESULTS: There were 68 tumors (38 IM lipomas and 30 ALTs/WDLSs). The AUC of the machine learning model was 0.88 [95% CI 0.72-1] (sensitivity, 91.6%; specificity, 85.7%; and accuracy, 89.0%). For Radiologist 1, the AUC was 0.94 [95% CI 0.87-1] (sensitivity, 97.4%; specificity, 90.9%; and accuracy, 95.0%), and as to Radiologist 2, the AUC was 0.91 [95% CI 0.83-0.99] (sensitivity, 100%; specificity, 81.8%; and accuracy, 93.3%). The classification agreement of the radiologists was 0.89 of kappa value (95% CI 0.76-1). Although the AUC of the model was lower than of two experienced MSK radiologists, there was no statistically significant difference between the model and two radiologists (all P > 0.05). CONCLUSIONS: The novel machine learning model based on tumor-to-bone distance and radiomic features is a noninvasive procedure that has the potential for distinguishing IM lipomas from ALTs/WDLSs. The predictive features that suggested malignancy were size, shape, depth, texture, histogram, and tumor-to-bone distance.

Robustness of Radiomic Features: Two-Dimensional versus Three-Dimensional MRI-Based Feature Reproducibility in Lipomatous Soft-Tissue Tumors.

This retrospective study aimed to compare the intra- and inter-observer manual-segmentation variability in the feature reproducibility between two-dimensional (2D) and three-dimensional (3D) magnetic-resonance imaging (MRI)-based radiomic features. The study included patients with lipomatous soft-tissue tumors that were diagnosed with histopathology and underwent MRI scans. Tumor segmentation based on the 2D and 3D MRI images was performed by two observers to assess the intra- and inter-observer variability. In both the 2D and the 3D segmentations, the radiomic features were extracted from the normalized images. Regarding the stability of the features, the intraclass correlation coefficient (ICC) was used to evaluate the intra- and inter-observer segmentation variability. Features with ICC > 0.75 were considered reproducible. The degree of feature robustness was classified as low, moderate, or high. Additionally, we compared the efficacy of 2D and 3D contour-focused segmentation in terms of the effects of the stable feature rate, sensitivity, specificity, and diagnostic accuracy of machine learning on the reproducible features. In total, 93 and 107 features were extracted from the 2D and 3D images, respectively. Only 35 features from the 2D images and 63 features from the 3D images were reproducible. The stable feature rate for the 3D segmentation was more significant than for the 2D segmentation (58.9% vs. 37.6%, p = 0.002). The majority of the features for the 3D segmentation had moderate-to-high robustness, while 40.9% of the features for the 2D segmentation had low robustness. The diagnostic accuracy of the machine-learning model for the 2D segmentation was close to that for the 3D segmentation (88% vs. 90%). In both the 2D and the 3D segmentation, the specificity values were equal to 100%. However, the sensitivity for the 2D segmentation was lower than for the 3D segmentation (75% vs. 83%). For the 2D + 3D radiomic features, the model achieved a diagnostic accuracy of 87% (sensitivity, 100%, and specificity, 80%). Both 2D and 3D MRI-based radiomic features of lipomatous soft-tissue tumors are reproducible. With a higher stable feature rate, 3D contour-focused segmentation should be selected for the feature-extraction process.

Inter-site validations of the Pixel-Wise method for cardiac T2* analysis in transfusion-dependent Thai thalassemia patients.

OBJECTIVE: To compare inter-site observer variability of the Pixel- Wise (PW) method for cardiac T2* analysis in thalassemia patients using the mono-exponential with a constant fitting (offset) model and to compare the cross-model variability of the offset model to the mono-exponential (typical) model. MATERIAL AND METHOD: Eighty-eight cardiac T2* measurements were performed on 72 Thalassemia major patients. Both bright- and black-blood techniques were acquired and analyzed at both the reference (REF) and local (LOC) sites using the PW method by defined region of interest on the whole (at the REF site) and partial (at the LOC site) septum. The offset model was analyzed at the reference site while both the offset and typical models were performed at the local site. The inter-site variability of the T2* values were analyzed by independent observers blinded to the results. RESULTS: The T2* values from the REF-offset, LOC-offset and LOC-typical methods were statistically comparable on both scanning techniques. The inter-site variations of the offset model were about 5.2% and 4.4% on the bright- and black-blood techniques, respectively, which was about 1.7% higher than from the intra-site, but was still in a reasonable range compared to the conventional method of around 5.4%. The cross-model comparisons presented with 0.4 ms of bias and variation of about 6.9% and 4.7%, respectively, which is about 1.4% higher than from the intra-site. CONCLUSION: The observer variability on the PW method using the offset or typical model provided equivalent coefficient of variation on both scanning techniques, which was also comparable to the previous reports. The inter-site variability of the offset and cross models was also in a reasonable range, being less than 2% higher than the intra-site with bias of about 0.4 ms.

Multi-level pooling encoder-decoder convolution neural network for MRI reconstruction.

MRI reconstruction is one of the critical processes of MRI machines, along with the acquisition. Due to a slow processing time of signal acquiring, parallel imaging and reconstruction techniques are applied for acceleration. To accelerate the acquisition process, fewer raw data are sampled simultaneously with all RF coils acquisition. Then, the reconstruction uses under-sampled data from all RF coils to restore the final MR image that resembles the fully sampled MR image. These processes have been a traditional procedure inside the MRI system since the invention of the multi-coils MRI machine. This paper proposes the deep learning technique with a lightweight network. The deep neural network is capable of generating the high-quality reconstructed MR image with a high peak signal-to-noise ratio (PSNR). This also opens a high acceleration factor for MR data acquisition. The lightweight network is called Multi-Level Pooling Encoder-Decoder Net (MLPED Net). The proposed network outperforms the traditional encoder-decoder networks on 4-fold acceleration with a significant margin on every evaluation metric. The network can be trained end-to-end, and it is a lightweight structure that can reduce training time significantly. Experimental results are based on a publicly available MRI Knee dataset from the fastMRI competition.

Prevalence and prognosis of myocardial scar in patients with known or suspected coronary artery disease and normal wall motion.

BACKGROUND: Some patients may have normal wall motion after myocardial infarction. The aim of this study was to determine the prevalence and prognosis of patients with myocardial scar in the absence of abnormal wall motion. We studied patients with suspected or known coronary artery disease (CAD) who were referred for cardiovascular magnetic resonance (CMR) for the assessment of global and regional cardiac function and late gadolinium enhancement (LGE) and had normal left ventricular wall motion. Prognostic value was determined by the occurrence of hard endpoints (cardiac death and nonfatal myocardial infarction) and major adverse cardiac events (MACE) which also included hospitalization due to unstable angina or heart failure or life threatening ventricular arrhythmia. RESULTS: A total 1148 patients (70.3%) were studied. LGE was detected in 104 patients (9.1%). Prevalence of LGE increased in patients with increased left ventricular mass. Average follow-up time was 955 ± 542 days. LGE was the strongest predictor for hard endpoints and MACE. CONCLUSION: LGE was detected in 9.1% of patients with suspected or known CAD and normal wall motion. LGE was the strongest predictor of significant cardiac events.

Comparison of the region-based and pixel-wise methods for cardiac T2* analysis in 50 transfusion-dependent Thai thalassemia patients.

PURPOSE: To compare the observer variability of the conventional region-based (RB) to the typical and proposed pixel-wise (PW) methods for cardiac T2* analysis in thalassemia patients. DESIGN AND METHODS: Fifty thalassemia major patients were enrolled for the study. Short-axis bright- and black-blood sequences were acquired and analyzed using the RB and PW methods. Regions were defined using the whole septum (WS) or partial septum (PS). From the same PS region, results were reported by mean (PS-PW) and median (MPS-PW). Intraobserver and interobserver variabilities were investigated on all data set by 2 independent observers blinded to the result. RESULTS: The T2* values from the PS-PW and MPS-PW methods were comparable to the conventional WS-RB method on both scanning techniques. When comparing the interobserver variability from the WS-RB to the PS-PW method, the coefficient of variation of the PS-PW method was equivalent (4.5% vs 4.7%, P = NS) for the bright-blood technique but 31% lower (4.0% vs 2.8%, P = 0.21) for the black-blood technique. The proposed MPS-PW method performed even better with respect to the conventional WS-RB method, decreasing interobserver coefficient of variation by 24% (4.5% vs 3.5%, P = 0.08) and 42% (4.0% vs 2.4%, P = 0.02), respectively. Intraobserver reproducibility followed the same trend. CONCLUSIONS: The proposed PW method using the median of T2* values calculated from partial interventricular septum region provided lower intraobserver and interobserver variabilities compared with the conventional RB or typical PW methods.