-
Korean Journal of Radiology Jul 2024To evaluate the role of visual and quantitative chest CT parameters in assessing treatment response in patients with severe asthma.
OBJECTIVE
To evaluate the role of visual and quantitative chest CT parameters in assessing treatment response in patients with severe asthma.
MATERIALS AND METHODS
Korean participants enrolled in a prospective multicenter study, named the Precision Medicine Intervention in Severe Asthma study, from May 2020 to August 2021, underwent baseline and follow-up chest CT scans (inspiration/expiration) 10-12 months apart, before and after biologic treatment. Two radiologists scored bronchiectasis severity and mucus plugging extent. Quantitative parameters were obtained from each CT scan as follows: normal lung area (normal), air trapping without emphysema (AT without emph), air trapping with emphysema (AT with emph), and airway (total branch count, Pi10). Clinical parameters, including pulmonary function tests (forced expiratory volume in 1 s [FEV1] and FEV1/forced vital capacity [FVC]), sputum and blood eosinophil count, were assessed at initial and follow-up stages. Changes in CT parameters were correlated with changes in clinical parameters using Pearson or Spearman correlation.
RESULTS
Thirty-four participants (female:male, 20:14; median age, 50.5 years) diagnosed with severe asthma from three centers were included. Changes in the bronchiectasis and mucus plugging extent scores were negatively correlated with changes in FEV1 and FEV1/FVC (ρ = from -0.544 to -0.368, all < 0.05). Changes in quantitative CT parameters were correlated with changes in FEV1 (normal, = 0.373 [ = 0.030], AT without emph, = -0.351 [ = 0.042]), FEV1/FVC (normal, = 0.390 [ = 0.022], AT without emph, = -0.370 [ = 0.031]). Changes in total branch count were positively correlated with changes in FEV1 ( = 0.349 [ = 0.043]). There was no correlation between changes in Pi10 and the clinical parameters ( > 0.05).
CONCLUSION
Visual and quantitative CT parameters of normal, AT without emph, and total branch count may be effective for evaluating treatment response in patients with severe asthma.
Topics: Humans; Male; Female; Asthma; Middle Aged; Tomography, X-Ray Computed; Prospective Studies; Severity of Illness Index; Adult; Treatment Outcome; Respiratory Function Tests; Aged
PubMed: 38942461
DOI: 10.3348/kjr.2024.0110 -
Korean Journal of Radiology Jul 2024In Korea, radiology has been positioned towards the early adoption of artificial intelligence-based software as medical devices (AI-SaMDs); however, little is known...
OBJECTIVE
In Korea, radiology has been positioned towards the early adoption of artificial intelligence-based software as medical devices (AI-SaMDs); however, little is known about the current usage, implementation, and future needs of AI-SaMDs. We surveyed the current trends and expectations for AI-SaMDs among members of the Korean Society of Radiology (KSR).
MATERIALS AND METHODS
An anonymous and voluntary online survey was open to all KSR members between April 17 and May 15, 2023. The survey was focused on the experiences of using AI-SaMDs, patterns of usage, levels of satisfaction, and expectations regarding the use of AI-SaMDs, including the roles of the industry, government, and KSR regarding the clinical use of AI-SaMDs.
RESULTS
Among the 370 respondents (response rate: 7.7% [370/4792]; 340 board-certified radiologists; 210 from academic institutions), 60.3% (223/370) had experience using AI-SaMDs. The two most common use-case of AI-SaMDs among the respondents were lesion detection (82.1%, 183/223), lesion diagnosis/classification (55.2%, 123/223), with the target imaging modalities being plain radiography (62.3%, 139/223), CT (42.6%, 95/223), mammography (29.1%, 65/223), and MRI (28.7%, 64/223). Most users were satisfied with AI-SaMDs (67.6% [115/170, for improvement of patient management] to 85.1% [189/222, for performance]). Regarding the expansion of clinical applications, most respondents expressed a preference for AI-SaMDs to assist in detection/diagnosis (77.0%, 285/370) and to perform automated measurement/quantification (63.5%, 235/370). Most respondents indicated that future development of AI-SaMDs should focus on improving practice efficiency (81.9%, 303/370) and quality (71.4%, 264/370). Overall, 91.9% of the respondents (340/370) agreed that there is a need for education or guidelines driven by the KSR regarding the use of AI-SaMDs.
CONCLUSION
The penetration rate of AI-SaMDs in clinical practice and the corresponding satisfaction levels were high among members of the KSR. Most AI-SaMDs have been used for lesion detection, diagnosis, and classification. Most respondents requested KSR-driven education or guidelines on the use of AI-SaMDs.
Topics: Humans; Republic of Korea; Artificial Intelligence; Surveys and Questionnaires; Societies, Medical; Radiology; Software
PubMed: 38942455
DOI: 10.3348/kjr.2023.1246 -
Korean Journal of Radiology Jul 2024Artificial intelligence (AI) is rapidly gaining recognition in the radiology domain as a greater number of radiologists are becoming AI-literate. However, the adoption... (Review)
Review
Position Statements of the Emerging Trends Committee of the Asian Oceanian Society of Radiology on the Adoption and Implementation of Artificial Intelligence for Radiology.
Artificial intelligence (AI) is rapidly gaining recognition in the radiology domain as a greater number of radiologists are becoming AI-literate. However, the adoption and implementation of AI solutions in clinical settings have been slow, with points of contention. A group of AI users comprising mainly clinical radiologists across various Asian countries, including India, Japan, Malaysia, Singapore, Taiwan, Thailand, and Uzbekistan, formed the working group. This study aimed to draft position statements regarding the application and clinical deployment of AI in radiology. The primary aim is to raise awareness among the general public, promote professional interest and discussion, clarify ethical considerations when implementing AI technology, and engage the radiology profession in the ever-changing clinical practice. These position statements highlight pertinent issues that need to be addressed between care providers and care recipients. More importantly, this will help legalize the use of non-human instruments in clinical deployment without compromising ethical considerations, decision-making precision, and clinical professional standards. We base our study on four main principles of medical care-respect for patient autonomy, beneficence, non-maleficence, and justice.
Topics: Artificial Intelligence; Humans; Radiology; Asia; Societies, Medical
PubMed: 38942454
DOI: 10.3348/kjr.2024.0419 -
Korean Journal of Radiology Jul 2024
Topics: Indonesia; Humans; Radiology; Internship and Residency
PubMed: 38942452
DOI: 10.3348/kjr.2024.0267 -
Korean Journal of Radiology Jul 2024
Topics: Hong Kong; Humans; Radiology
PubMed: 38942451
DOI: 10.3348/kjr.2024.0440 -
Journal of the American College of... Jun 2024Thyroid nodule evaluation using ultrasound is dependent on radiologist experience, but deep learning (DL) models can improve intra-reader agreements. DL model...
Thyroid nodule evaluation using ultrasound is dependent on radiologist experience, but deep learning (DL) models can improve intra-reader agreements. DL model development for medical imaging with small datasets can be challenging. Transfer learning is a technique used in the development of DL models to improve model performance in data-limited scenarios. Here, we investigate the impact of transfer learning with domain-specific RadImageNet dataset and non-medical ImageNet on the robustness of classifying thyroid nodules into benign and malignant. We retrospectively collected 822 ultrasound images of thyroid nodules of patients who underwent fine needle aspiration in our institute. We split our data and used 101 cases in a test set and 721 cases for cross-validation. A Resnet-18 model was trained to classify thyroid nodules into benign and malignant. Then, we trained the same model architecture with transferred weights from ImageNet and RadImageNet. The model without transfer learning for thyroid nodule classification achieved an AUROC of 0.69. The AUROC of our model after transfer learning with ImageNet pre-trained weights was 0.79. Our model achieved an AUROC of 0.83 from transfer learning of the RadImageNet pre-trained weights. The AUROC from the classification model without transfer learning significantly improved after transfer learning with ImageNet (p-value = 0.03) and RadImageNet transfer learning (p-value <0.01). There was a statistically significant distinction in performance between the model utilizing RadImageNet transfer learning and that employing ImageNet transfer learning (p-value <0.01). We demonstrate the potential of RadImageNet as a domain-specific source for transfer learning in thyroid nodule classification.
PubMed: 38942163
DOI: 10.1016/j.jacr.2024.06.011 -
Journal of the American College of... Jun 2024
PubMed: 38942162
DOI: 10.1016/j.jacr.2024.06.012 -
European Journal of Radiology Jun 2024To propose a convolutional neural network (EmbNet) for automatic pulmonary embolism detection on computed tomography pulmonary angiogram (CTPA) scans and to assess its...
OBJECTIVE
To propose a convolutional neural network (EmbNet) for automatic pulmonary embolism detection on computed tomography pulmonary angiogram (CTPA) scans and to assess its diagnostic performance.
METHODS
305 consecutive CTPA scans between January 2019 and December 2021 were enrolled in this study (142 for training, 163 for internal validation), and 250 CTPA scans from a public dataset were used for external validation. The framework comprised a preprocessing step to segment the pulmonary vessels and the EmbNet to detect emboli. Emboli were divided into three location-based subgroups for detailed evaluation: central arteries, lobar branches, and peripheral regions. Ground truth was established by three radiologists.
RESULTS
The EmbNet's per-scan level sensitivity, specificity, positive predictive value (PPV), and negative predictive value were 90.9%, 75.4%, 48.4%, and 97.0% (internal validation) and 88.0%, 70.5%, 42.7%, and 95.9% (external validation). At the per-embolus level, the overall sensitivity and PPV of the EmbNet were 86.0% and 61.3% (internal validation), and 83.5% and 57.5% (external validation). The sensitivity and PPV of central emboli were 89.7% and 52.0% (internal validation), and 94.4% and 43.0% (external validation); of lobar emboli were 95.2% and 76.9% (internal validation), and 93.5% and 72.5% (external validation); and of peripheral emboli were 82.6% and 61.7% (internal validation), and 80.2% and 59.4% (external validation). The average false positive rate was 0.45 false emboli per scan (internal validation) and 0.69 false emboli per scan (external validation).
CONCLUSION
The EmbNet provides high sensitivity across embolus locations, suggesting its potential utility for initial screening in clinical practice.
PubMed: 38941822
DOI: 10.1016/j.ejrad.2024.111586 -
European Journal of Radiology Jun 2024To assess T1 mapping performance in distinguishing between benign and malignant breast lesions and to explore its correlation with histopathologic features in breast...
PURPOSE
To assess T1 mapping performance in distinguishing between benign and malignant breast lesions and to explore its correlation with histopathologic features in breast cancer.
METHODS
This study prospectively enrolled 103 participants with a total of 108 lesions, including 25 benign and 83 malignant lesions. T1 mapping, diffusion-weighted imaging (DWI), and dynamic contrast-enhanced (DCE) were performed. Two radiologists independently outlined the ROIs and analyzed T1 and apparent diffusion coefficient (ADC) values for each lesion, assessing interobserver reliability with the intraclass correlation coefficient (ICC). T1 and ADC values were compared between benign and malignant lesions, across different histopathological characteristics (histological grades, estrogen, progesterone and HER2 receptors expression, Ki67, N status). Receiver operating characteristic (ROC) analysis and Pearson correlation coefficient (ρ) were performed.
RESULTS
T1 values showed statistically significant differences between benign and malignant groups (P < 0.001), with higher values in the malignant (1817.08 ms ± 126.64) compared to the benign group (1429.31 ms ± 167.66). In addition, T1 values significantly increased in the ER (-) group (P = 0.001). No significant differences were found in T1 values among HER2, Ki67, N status, and histological grades groups. Furthermore, T1 values exhibited a significant correlation (ρ) with ER (P < 0.01) and PR (P = 0.03). The AUC for T1 value in distinguishing benign from malignant lesions was 0.69 (95 % CI: 0.55 - 0.82, P = 0.005), and for evaluating ER status, it was 0.75 (95 % CI: 0.62 - 0.87, P = 0.002).
CONCLUSIONS
T1 mapping holds the potential as an imaging biomarker to assist in the discrimination of benign and malignant breast lesions and assessing the ER expression status in breast cancer.
PubMed: 38941821
DOI: 10.1016/j.ejrad.2024.111589 -
Journal of Medical Imaging and... Jun 2024Intravenous contrast injection protocol for certain CT studies at our institution was revised in June 2022 in response to the global shortage of iohexol. This included...
OBJECTIVE
Intravenous contrast injection protocol for certain CT studies at our institution was revised in June 2022 in response to the global shortage of iohexol. This included CT head studies performed for neuro-navigation (contrast dose from 90 mL to 70 mL). The quality of these studies was assessed.
METHODS
Consecutive CT scans before (n = 32) and after (n = 32) contrast dose reduction were reviewed. Demographic data was obtained from the chart. Subjective observations made by two radiologists in consensus included overall study quality (Likert scale of 1 to 5) and lesion location, margins and internal characteristics that were compared with MRI findings (reference standard) using Fisher's exact test. Superior sagittal sinus attenuation, used as an objective measurement of enhancement, and lesion size were compared using Student's t-test. The institutional database was searched for any study requiring repetition or deemed non-diagnostic.
RESULTS/DISCUSSION
The average age (61.1 ± 12.7 years and 61.6 ± 14.9 years) and body surface area (BSA) (1.9 ± 0.3 m and 1.9 ± 0.02 m) was not significantly different (p > 0.05) between groups. There was no significant difference (p > 0.05) in objective or subjective enhancement between the two groups. There was no significant difference between CT and MRI for lesion size, location, number, margins and internal enhancement characteristics in the two groups. No study required repetition or was reported as non-diagnostic. There was no adverse comment about study quality in operative notes.
CONCLUSION
Reduced contrast dose neuro-navigation CT head studies are not different in quality compared to the conventional studies.
PubMed: 38941784
DOI: 10.1016/j.jmir.2024.101433