Analysis of synthetic magnetic resonance images by multi-channel segmentation increases accuracy of volumetry in the putamen and decreases mis-segmentation in the dural sinuses.

BACKGROUND: Voxel-based morphometry (VBM) using magnetic resonance imaging (MR) has been used to estimate cortical atrophy associated with various diseases. However, there are mis-segmentations of segmented gray matter image in VBM. PURPOSE: To study a twofold evaluation of single- and multi-channel segmentation using synthetic MR images: (1) mis-segmentation of segmented gray matter images in transverse and cavernous sinuses; and (2) accuracy and repeatability of segmented gray matter images. MATERIAL AND METHODS: A total of 13 healthy individuals were scanned with 3D quantification using an interleaved Look-Locker acquisition sequence with a T2 preparation pulse (3D-QALAS) sequence on a 1.5-T scanner. Three of the 13 healthy participants were scanned five consecutive times for evaluation of repeatability. We used SyMRI software to create images with three contrasts: T1-weighted (T1W), T2-weighted (T2W), and proton density-weighted (PDW) images. Manual regions of interest (ROI) on T1W imaging were individually set as the gold standard in the transverse sinus, cavernous sinus, and putamen. Single-channel (T1W) and multi-channel (T1W + T2W, T1W + PDW, and T1W + T2W + PDW imaging) segmentations were performed with statistical parametric mapping 12 software. RESULTS: We found that mis-segmentations in both the transverse and cavernous sinuses were large in single-channel segmentation compared with multi-channel segmentations. Furthermore, the accuracy of segmented gray matter images in the putamen was high in both multi-channel T1W + PDW and T1W + T2W + PDW segmentations compared with other segmentations. Finally, the highest repeatability of left putamen volumetry was found with multi-channel segmentation T1WI + PDWI. CONCLUSION: Multi-channel segmentation with T1WI + PDWI provides good results for VBM compared with single-channel and other multi-channel segmentations.

Using modulated and smoothed data improves detectability of volume difference in group comparison, but reduces accuracy with atlas-based volumetry using Statistical Parametric Mapping 12 software.

BACKGROUND: Atlas-based volumetry using three-dimensional T1-weighted (3D-T1W) magnetic resonance imaging (MRI) has been used previously to evaluate the volumes of intracranial tissues. PURPOSE: To evaluate the detectability of volume difference and accuracy for volumetry using smoothed data with an atlas-based method. MATERIAL AND METHODS: Twenty healthy individuals and 24 patients with idiopathic normal-pressure hydrocephalus (iNPH) underwent 3-T MRI, and sagittal 3D-T1W images were obtained in all participants. Signal values (as tissue probability) of voxels in five segmented data types (gray matter, white matter, cerebrospinal fluid [CSF], skull, soft tissue) derived from the 3D-T1W images with SPM 12 software were assigned simulated 3D-T1W signal intensities to each tissue image. The assigned data were termed "reference data." We created a reference 3D-T1W image that included the reference data of all five tissue types. Standard volumes were measured for the reference CSF data with region of interest of lateral ventricle in native space, and measured volumes were obtained for non-smoothed and smoothed-modulated data. Detectability was evaluated between measured volumes in the healthy control and iNPH groups. Accuracy was evaluated as the difference between the mean measured and standard volumes. RESULTS: In group comparison of measured volumes between the healthy control and iNPH groups, the lowest P value was for smoothed-modulated CSF data. In both groups, the largest difference from the standard volume was found for the mean of the measured volumes for smoothed-modulated CSF data. CONCLUSION: Our study shows that using smoothed data can improve detectability in group comparison. However, using smoothed data reduces the accuracy of volumetry.

Efficacy and safety of valbenazine in Japanese patients with tardive dyskinesia: A multicenter, randomized, double-blind, placebo-controlled study (J-KINECT).

AIM: Valbenazine is approved in the US for treatment of tardive dyskinesia (TD); however, efficacy/safety data in Asian populations are lacking. We assessed the efficacy/safety of valbenazine in Japanese patients. METHODS: This phase II/III, multicenter, randomized, double-blind, placebo-controlled study (NCT03176771) included adult psychiatric patients with TD, who were randomly allocated to receive placebo or valbenazine (once-daily 40- or 80-mg) for a 6-week, double-blind period, after which the placebo group was switched to valbenazine for a 42-week extension. The primary endpoint was change from baseline in Abnormal Involuntary Movement Scale (AIMS) total score at Week 6; clinical global impression of improvement of TD (CGI-TD) was also assessed. RESULTS: Of 256 patients, 86, 85, and 85 were allocated to the 40-mg valbenazine, 80-mg valbenazine, and placebo groups, respectively. Least-squares mean (95% confidence interval) change from baseline in AIMS score at Week 6 was -2.3 (-3.0 to -1.7) in the valbenazine 40-mg group, -3.7 (-4.4 to -3.0) in the 80-mg group, and -0.1 (-0.8 to 0.5) in the placebo group; both treatment groups showed statistically significant improvements vs. placebo. Patients switched to valbenazine at Week 6 showed similar improvements in AIMS scores, which were maintained to Week 48. Improvements in CGI-TD scores were observed for both treatment groups vs. placebo. Incidence of adverse events was highest in the 80-mg group; common events included nasopharyngitis, somnolence, schizophrenia worsening, hypersalivation, insomnia, and tremor. CONCLUSION: The efficacy/safety profile of valbenazine was similar to that of previous clinical trials, supporting its use for TD treatment in Japanese patients.

[Examination of Effectiveness Verification and Additional Items in Angiography and IVR of DRLs 2015].

Japanese Diagnostic Reference Levels (DRLs) were released as "Japan DRLs 2015" from Japan Network for Research and Information on Medical Exposure (J-RIME) in June 2015. In "Japan DRLs 2015", DRLs in angiography and interventional procedures are set at a fluoroscopic dose rate of 20 mGy/min at the interventional reference point using a phantom. In order to achieve optimization with DRLs, then it need to be revised regularly. Therefore, we (research group to examine the effect of Japan DRLs 2015 and the necessity of additional items in angiography and vascular interventions) examined the effects of "Japan DRLs 2015" on angiography and interventional procedures. And we also investigated for DRLs revision in the future. As a result, it turned out that it is important to create DRLs in medical procedures that can be effectively used in clinical settings.

[Nationwide Survey of Medical Radiation Exposure on Cardiovascular Examinations in Japan].

PURPOSE: It is very important to manage the radiation dose of cardiovascular interventional (CVI) procedures. Overseas, the diagnostic reference levels for cardiac interventional procedures were established with the air kerma at the patient entrance reference point (Ka,r) and the air kerma-area product (PKA). Although the Japan DRLs 2015 was established by the Japan Network for Research and Information on Medical Exposure (J-RIME), the Japan DRL for CVIs were established by fluoroscopic dose rates of 20 mGy/min at the patient entrance reference point with 20 cm thickness polymethyl methacrylate (PMMA) phantom. In the present our study, we performed a questionnaire survey of indicated values of angiographic parameters in CVI procedures. METHODS: A nationwide questionnaire was sent by post to 765 facilities. Question focused on angiographic technology, exposure parameters and radiation doses as the displayed dosimetric parameters on the angiographic machine. RESULTS: The recovery rate was 22.8% at 175 out of 765 facilities. In total 1728 cases of the coronary angiography (CAG), 1703 cases of the percutaneous coronary intervention (PCI), 962 cases of the radiofrequency catheter ablation (RFCA) and 377 cases of pediatric CVI. The 75th percentile value of Ka,r, PKA, fluoroscopy time (FT) and number of cine images (CI) for CAG, PCI, RFCA and pediatric CVI were 702, 2042, 644, and 159 mGy, respectively, 59.3, 152, 81.3, and 14.9 Gy・cm2, respectively, 10.2, 35.6, 61.1, and 35.6 min, respectively and 1503, 2672, 722, and 2378 images, respectively. Our investigation showed that the angiographic parameters were different in several CVI procedures. CONCLUSIONS: The displayed dosimetric parameters on the angiographic machine in CVI procedures showed different values. We should classify the dosimetric parameters for each procedure.

Survey of Exposure Dose during Percutaneous Coronary Intervention for Chronic Total Occlusion.

During percutaneous coronary intervention (PCI) for chronic total occlusion (CTO), longer fluoroscopic time as compared with PCI for non-CTO lesions may cause skin injury by increased radiation. We have performed a multi-center observational study comparing the exposed dose during the PCI of CTO (CTO group) and during the PCI of non-CTO lesions (non-CTO group). Exposure doses were assessed in 313 patients with CTO and 3,310 patients with non-CTO lesions. Total fluoroscopy time (59.0 ±35.5 vs 26.8 ±18.8 min, p<0.0001) and the total air kerma (2.76±2.11 vs 1.27±0.94 Gy, p<0.0001) were significantly greater in the CTO group than in the non-CTO group. The maximum air kerma of the CTO group was 13.62 Gy. Informed consent about the risk of transient depilation and the transient erythema is required for the case with radiation dose over 3 Gy. The frequency of the patient who received radiation >3 Gy was significantly higher in the CTO group as compared with the non-CTO group (34.1% vs 4.9%). Therefore, informed consent before an operation and postoperative follow-up are indispensable for the performed PCI of CTO. Moreover, comprehensive understanding of the exposure dose during operation and to record the final exposure dose may be extremely important for the radiological technologists.

Initial experience with percutaneous coronary sinus catheter placement in minimally invasive cardiac surgery in an academic center.

BACKGROUND: Placement of a percutaneous coronary sinus catheter (CSC) by an anesthesiologist for retrograde cardioplegia in minimally invasive cardiac surgery is relatively safe in experienced hands. However, the popularity of its placement remains limited to a small number of centers due to its perceived complexity and potential complications. METHODS: We retrospectively reviewed all cardiac cases performed by one surgeon between December 2009 and April 2012. The reviewed cases were divided into two groups: cardiac cases with percutaneous CSC placement (CSC group) and cardiac cases without placement (control group). Anesthesia preparation time (APT) was then compared between the CSC group and control group. In the CSC group, cases were further divided into two groups. One group contained cases with an APT of less than 90 min (success group) and the other contained cases with an APT greater than or equal to 90 min or cases with CSC placement failure (delay/failure group). Patients' characteristics, type of surgery, and transesophageal echocardiography (TEE) findings were compared between the two groups (success group vs. delay/failure group) to identify variables associated with prolongation of the APT or CSC placement failure. RESULTS: Percutaneous CSC placement was required in 83 cases (CSC group). The catheter was successfully placed in 74 of those cases. We experienced one complication, coronary sinus injury after multiple attempts at placing the catheter. The mean APT was 102 ± 31 min in the CSC group (n = 81) and 42 ± 15 min in the control group (n = 285). We could not identify any variables associated with prolongation of the APT or catheter placement failure. CONCLUSIONS: The success rate of the placement was 89.1 % in our academic center. On average, placing the CSC added approximately one additional hour to the APT. This time is not an accurate representation of true catheter placement time, as it included time for preparation of the CSC, TEE, and fluoroscopy. We experienced one documented complication (coronary sinus injury), which was immediately diagnosed by TEE and fluoroscopy in the operating room. No variables associated with prolongation of APT or CSC placement failure were identified.

[Multicenter Study on Evaluation of the Entrance Skin Dose by a Direct Measurement Method in Cardiac Interventional Procedures].

Deterministic effects have been reported in cardiac interventional procedures. To prevent radiation skin injuries in percutaneous coronary intervention (PCI), it is necessary to measure accurate patient entrance skin dose (ESD) and maximum skin absorbed dose (MSD). We measured the MSD on 62 patients in four facilities by using the Chest-RADIREC(Ⓡ) system. The correlation between MSD and fluoroscopic time, dose area product (DAP), and cumulative air kerma (AK) showed good results, with the correlation between MSD and AK being the strongest. The regression lines using MSD as an outcome value (y) and AK as predictor variables (x) was y=1.18x (R(2)=0.787). From the linear regression equation, MSD is estimated to be about 1.18 times that of AK in real time. The Japan diagnostic reference levels (DRLs) 2015 for IVR was established by the use of dose rates using acrylic plates (20- cm thick) at the interventional reference point. Preliminary reference levels proposed by International Atomic Energy Agency (IAEA) were provided using DAP. In this study, AK showed good correlation most of all. Hence we think that Japanese DRLs for IVR should reconsider by clinical patients' exposure dose such as AK.

[Study on the development of a patient dosimetry gown for interventional cardiology procedures].

In recent years, dose justification and optimization have been attempted in percutaneous coronary intervention (PCI); however, deterministic effects have been reported. To prevent radiation skin injuries in PCI, it is necessary to measure the patient entrance skin dose (ESD), but an accurate dose measurement method has not yet been established. In this study, we developed a dosimetry gown that can measure the ESD during PCI using multiple radiophotoluminescence dosimeters (RPLDs). The RPLDs were placed into 84 pockets that were sewn into a dosimetry gown. Patients wear the original dosimetry gown during the procedures, after which we obtain accurate ESD measurements. We believe that this method using RPLDs and a newly-designed dosimetry gown provides accurate ESD measurements during PCI. We expect this system to become a standard method for measuring ESD during PCI.

[Questionnaire Survey on IVR Dose Management in Tokai Region].

PURPOSE: National diagnostic reference levels in Japan 2020 (DRLs 2020) have been published. In the field of angiography, in addition to the fluoroscopic dose rate, incident air kerma at the patient entrance reference point displayed on the equipment (Ka,r: mGy) and air kerma-area product displayed on the equipment (PKA: Gycm2) were set. A questionnaire survey was conducted at each facility in the Tokai region to confirm the status of medical radiation dose control in the region. METHOD: A questionnaire survey was conducted at each facility in the Tokai region. The items were fluoroscopic dose rate in each area (head and neck, cardiac, chest and abdomen, and limbs), DA and DSA dose rates, and dose area product meter (Ka,r, PKA) for the main procedures in each area. RESULT: The median values in this study were lower than those in the DRLs 2020, indicating that appropriate dose control is being implemented in the Tokai region. The trends of fluoroscopic and radiographic dose rates were different in each area, and there was some variation among the facilities. CONCLUSION: We believe that the incorporation of fluoroscopic and radiographic dose rates by area into the DRLs will facilitate more appropriate dose control at each facility in the future.

Reducing image artifacts in sparse projection CT using conditional generative adversarial networks.

Reducing the amount of projection data in computed tomography (CT), specifically sparse-view CT, can reduce exposure dose; however, image artifacts can occur. We quantitatively evaluated the effects of conditional generative adversarial networks (CGAN) on image quality restoration for sparse-view CT using simulated sparse projection images and compared them with autoencoder (AE) and U-Net models. The AE, U-Net, and CGAN models were trained using pairs of artifacts and original images; 90% of patient cases were used for training and the remaining for evaluation. Restoration of CT values was evaluated using mean error (ME) and mean absolute error (MAE). The image quality was evaluated using structural image similarity (SSIM) and peak signal-to-noise ratio (PSNR). Image quality improved in all sparse projection data; however, slight deformation in tumor and spine regions was observed, with a dispersed projection of over 5°. Some hallucination regions were observed in the CGAN results. Image resolution decreased, and blurring occurred in AE and U-Net; therefore, large deviations in ME and MAE were observed in lung and air regions, and the SSIM and PSNR results were degraded. The CGAN model achieved accurate CT value restoration and improved SSIM and PSNR compared to AE and U-Net models.

Deep Learning-based Hierarchical Brain Segmentation with Preliminary Analysis of the Repeatability and Reproducibility.

PURPOSE: We developed new deep learning-based hierarchical brain segmentation (DLHBS) method that can segment T1-weighted MR images (T1WI) into 107 brain subregions and calculate the volume of each subregion. This study aimed to evaluate the repeatability and reproducibility of volume estimation using DLHBS and compare them with those of representative brain segmentation tools such as statistical parametric mapping (SPM) and FreeSurfer (FS). METHODS: Hierarchical segmentation using multiple deep learning models was employed to segment brain subregions within a clinically feasible processing time. The T1WI and brain mask pairs in 486 subjects were used as training data for training of the deep learning segmentation models. Training data were generated using a multi-atlas registration-based method. The high quality of training data was confirmed through visual evaluation and manual correction by neuroradiologists. The brain 3D-T1WI scan-rescan data of the 11 healthy subjects were obtained using three MRI scanners for evaluating the repeatability and reproducibility. The volumes of the eight ROIs-including gray matter, white matter, cerebrospinal fluid, hippocampus, orbital gyrus, cerebellum posterior lobe, putamen, and thalamus-obtained using DLHBS, SPM 12 with default settings, and FS with the "recon-all" pipeline. These volumes were then used for evaluation of repeatability and reproducibility. RESULTS: In the volume measurements, the bilateral thalamus showed higher repeatability with DLHBS compared with SPM. Furthermore, DLHBS demonstrated higher repeatability than FS in across all eight ROIs. Additionally, higher reproducibility was observed with DLHBS in both hemispheres of six ROIs when compared with SPM and in five ROIs compared with FS. The lower repeatability and reproducibility in DLHBS were not observed in any comparisons. CONCLUSION: Our results showed that the best performance in both repeatability and reproducibility was found in DLHBS compared with SPM and FS.

An analysis of the effectiveness of reflective learning through watching videos recorded with smart glasses-With multiple views (student, patient, and overall) in radiography education.

The Objective Structured Clinical Examination (OSCE) is designed to assess medical students' skills and attitude competencies before clinical practice. However, no method of reflective learning using video-based content has been used in OSCE education. This study aimed to confirm whether using smart glasses-based educational content is effective for OSCE reflective learning using multiple views (patient, student, and overall). This educational intervention study included a control group exposed to the traditional learning method and an intervention group exposed to a learning method incorporating smart glasses. Participants were 117 (72 in the control group and 45 in the intervention group) third-year radiological technology students scheduled to take the OSCE and 70 (37 in the control group and 33 in the intervention group) who met the eligibility criteria. Mock OSCEs were administered before and after the educational intervention (traditional and smart glasses-based education) to investigate changes in scores. After the educational intervention, a self-reported comprehension survey and a questionnaire were administered on the effectiveness of the video-based content from different views for student reflective learning. Unexpectedly, the OSCE evaluation score after the preliminary investigation significantly increased for the smart glasses control group (0.36±0.1) compared to the intervention group (0.06±0.1) setting up the radiographic conditions (x-ray center and detector center; p = 0.042). The intervention group's lower score in the mock OSCEs may have been due to the discomfort of wearing the smart glasses to perform the radiography procedure and their unfamiliarity with the smart glasses, which may have affected their concentration. The findings suggest that smart glasses-based education for OSCEs can be improved (e.g., being easy to handle and use and trouble-free).

[Investigation of patient exposure doses in diagnostic radiography in 2011 questionnaire].

We report here the results of a dose evaluation based on information obtained in a 2011 questionnaire as compared with an investigation made in 2007. Briefly, in general radiography, the dose used in most examinations in 2011 was lower than in 2007. However, since the entrance surface dose for chest X-rays showed an increase, there is a need to standardize the taking of digital images to be able to decrease the dose. Although computed tomography dose index volume (CTDIvol) in CT examinations was higher than that revealed in the 2007 investigation, there is potential for dose reduction.

[Subjective Contrast of Pulmonary Tissues in CsI-FPD Chest Radiography Using Model Phantom by Monte Carlo Simulation].

OBJECTIVE: Subject contrast of pulmonary tissues was investigated for five X-ray beams (70 kV without filter, 90 kV with 0.15 mm Cu filter, 90 kV with 0.2 mm Cu filter, 120 kV without filter, and 120 kV with 0.2 mm Cu filter) in CsI-FPD chest radiography using two types of model phantoms by Monte Carlo simulation. METHODS: A total of 72 million photons were entered to the model lung phantom (width, 300 mm; length, 300 mm; thickness, 200 mm; air space, 120 mm) and model mediastinum phantom (width, 300 mm; length, 300 mm; thickness, 200 mm; air space, 40 mm). Individual primary and secondary photon's process (absorption, scattering, and penetration) in the phantom and CsI-detector was recorded by Monte Carlo simulation. Subject contrast was calculated by entered and absorbed photon's number in the CsI-detector. RESULTS: Subject contrast pulmonary tissues were high to low energy X-ray beam; however, the ones of soft tissue and soft tissue overlaying bone had few differences for beam quality except 70 kV without filter. Moreover, the subject contrast by absorbed photons was higher compared to the one by entered photons in CsI. CONCLUSION: It was shown that the subject contrast study by Monte Carlo calculation can be replaced by the way of physical chest phantom, and that the subject contrast by absorbed photons and by injected photons in CsI was different. Furthermore, be verified that the subject contrast of soft tissue and soft tissue overlaying bone differs hardly.

PATIENT DOSIMETRY SURVEY OF PEDIATRIC DIAGNOSTIC AND THERAPEUTIC CARDIAC CATHETERISATION IN JAPAN.

To propose reference values for air-kerma at the reference point (Ka,r), air-kerma area product (PKA), fluoroscopy time (FT) and number of cine images (CI) for four age groups in Japan, a nationwide questionnaire was posted to 132 pediatric catheterisation of certified facility in Japan, using the conventional post system, to which 43 facilities responded. For diagnostic cardiac angiography, reference values were as follows: Ka,r: 86, 102, 165 and 264 mGy; PKA: 9.3, 9.5, 16 and 34 Gy.cm2; FT: 33, 29, 26 and 30 min and CI: 1904, 1966, 2405 and 1871 images. For therapeutic cardiac angiography, reference values were as follows: Ka,r: 107, 163, 103 and 202 mGy; PKA: 7.5, 18, 7 and 24 Gy.cm2; FT: 56, 52, 42 and 30 min and CI: 3886, 3232, 2212 and 4316 images for less than 1, 1-5, 6-10 and 11-15 y, respectively. To optimal patient exposure from diagnostic and therapeutic cardiac catheterisation, it is therefore necessary to establish reference values for pediatric cardiac catheterisation examinations for four age groups.

Accuracy of skull stripping in a single-contrast convolutional neural network model using eight-contrast magnetic resonance images.

In automated analyses of brain morphometry, skull stripping or brain extraction is a critical first step because it provides accurate spatial registration and signal-intensity normalization. Therefore, it is imperative to develop an ideal skull-stripping method in the field of brain image analysis. Previous reports have shown that convolutional neural network (CNN) method is better at skull stripping than non-CNN methods. We aimed to evaluate the accuracy of skull stripping in a single-contrast CNN model using eight-contrast magnetic resonance (MR) images. A total of 12 healthy participants and 12 patients with a clinical diagnosis of unilateral Sturge-Weber syndrome were included in our study. A 3-T MR imaging system and QRAPMASTER were used for data acquisition. We obtained eight-contrast images produced by post-processing T1, T2, and proton density (PD) maps. To evaluate the accuracy of skull stripping in our CNN method, gold-standard intracranial volume (ICVG) masks were used to train the CNN model. The ICVG masks were defined by experts using manual tracing. The accuracy of the intracranial volume obtained from the single-contrast CNN model (ICVE) was evaluated using the Dice similarity coefficient [= 2(ICVE ⋂ ICVG)/(ICVE + ICVG)]. Our study showed significantly higher accuracy in the PD-weighted image (WI), phase-sensitive inversion recovery (PSIR), and PD-short tau inversion recovery (STIR) compared to the other three contrast images (T1-WI, T2-fluid-attenuated inversion recovery [FLAIR], and T1-FLAIR). In conclusion, PD-WI, PSIR, and PD-STIR should be used instead of T1-WI for skull stripping in the CNN models.

Accuracy of IVUS-Based Machine Learning Segmentation Assessment of Coronary Artery Dimensions and Balloon Sizing.

Background: Accurate intravascular ultrasound (IVUS) measurements are important in IVUS-guided percutaneous coronary intervention optimization by choosing the appropriate device size and confirming stent expansion. Objectives: The purpose of this study was to assess the accuracy of machine learning (ML) automatic segmentation of coronary artery vessel and lumen dimensions and balloon sizing. Methods: Using expert analysis as the gold standard, ML segmentation of 60 MHz IVUS images was developed using 8,076 IVUS cross-sectional images from 234 patients, which were randomly split into training (83%) and validation (17%) data sets. The performance of ML segmentation was then evaluated using an independent test data set (437 images from 92 patients). The endpoints were the agreement rate between ML vs experts' measurements for appropriate balloon size selection, and lumen and acute stent areas. Appropriate balloon size was determined by rounding down from the mean vessel diameter or rounding up from the mean lumen diameter to the next balloon size. The difference of lumen area ≥0.5 mm2 was considered as clinically significant. Results: ML model segmentation correlated well with experts' segmentation for training data set with a correlation coefficient of 0.992 and 0.993 for lumen and vessel areas, respectively. The agreement rate in lumen and acute stent areas was 85.5% and 97.0%, respectively. The agreement rate for appropriate balloon size selection was 70.6% by vessel diameter only and 92.4% by adding lumen diameter. Conclusions: ML model IVUS segmentation measurements were well-correlated with those of experts and selected an appropriate balloon size in more than 90% of images.