Senior Radiologists Research Articles

A few-shot learning framework for the diagnosis of osteopenia and osteoporosis using knee X-ray images.

We developed a few-shot learning (FSL) framework for the diagnosis of osteopenia and osteoporosis in knee X-ray images. Computer vision models containing deep convolutional neural networks were fine-tuned to enable generalization from natural images (ImageNet) to chest X-ray images (normal vs. pneumonia, base images). Then, a series of automated machine learning classifiers based on the Euclidean distances of base images were developed to make predictions for novel images (normal vs. osteopenia vs. osteoporosis). The performance of the FSL framework was compared with that of junior and senior radiologists. In addition, the gradient-weighted class activation mapping algorithm was used for visual interpretation. In Cohort #1, the mean accuracy (0.728) and sensitivity (0.774) of the FSL models were higher than those of the radiologists (0.512 and 0.448). A diagnostic pipeline of FSL model (first)-radiologists (second) achieved better performance (0.653 accuracy, 0.582 sensitivity, and 0.816 specificity) than radiologists alone. In Cohort #2, the diagnostic pipeline also showed improved performance. The FSL framework yielded practical performance with respect to the diagnosis of osteopenia and osteoporosis in comparison with radiologists. This retrospective study supports the use of promising FSL methods in computer-aided diagnosis tasks involving limited samples.

Comparison between artificial intelligence solution and radiologist for the detection of pelvic, hip and extremity fractures on radiographs in adult using CT as standard of reference

PurposeThe purpose of this study was to compare the diagnostic performance of an artificial intelligence (AI) solution for the detection of fractures of pelvic, proximal femur or extremity fractures in adults with radiologist interpretation of radiographs, using standard dose CT examination as the standard of reference. Materials and methodsThis retrospective study included 94 adult patients with suspected bone fractures who underwent a standard dose CT examination and radiographs of the pelvis and/or hip and extremities at our institution between January 2022 and August 2023. For all patients, an AI solution was used retrospectively on the radiographs to detect and localize bone fractures of the pelvis and/or hip and extremities. Results of the AI solution were compared to the reading of each radiograph by a radiologist using McNemar test. The results of standard dose CT examination as interpreted by a senior radiologist were used as the standard of reference. ResultA total of 94 patients (63 women; mean age, 56.4 ± 22.5 [standard deviation] years) were included. Forty-seven patients had at least one fracture, and a total of 71 fractures were deemed present using the standard of reference (25 hand/wrist, 16 pelvis, 30 foot/ankle). Using the standard of reference, the analysis of radiographs by the AI solution resulted in 58 true positive, 13 false negative, 33 true negative and 15 false positive findings, yielding 82 % sensitivity (58/71; 95 % confidence interval [CI]: 71–89 %), 69 % specificity (33/48; 95 % CI: 55–80 %), and 76 % accuracy (91/119; 95 % CI: 69–84 %). Using the standard of reference, the reading of the radiologist resulted in 65 true positive, 6 false negative, 42 true negative and 6 false positive findings, yielding 92 % sensitivity (65/71; 95 % CI: 82–96 %), 88 % specificity (42/48; 95 % CI: 75–94 %), and 90 % accuracy (107/119; 95 % CI: 85–95 %). The radiologist outperformed the AI solution in terms of sensitivity (P = 0.045), specificity (P = 0.016), and accuracy (P < 0.001). ConclusionIn this study, the radiologist outperformed the AI solution for the diagnosis of pelvic, hip and extremity fractures of the using radiographs. This raises the question of whether a strong standard of reference for evaluating AI solutions should be used in future studies comparing AI and human reading in fracture detection using radiographs.

Inter-observer consistency on subsolid nodule follow-up recommendation based on National Comprehensive Cancer Network (NCCN) guidelines in low-dose computed tomography (LDCT) lung cancer screening.

Follow-up management of pulmonary nodules is a crucial component of lung cancer screening. Consistency in follow-up recommendations is essential for effective lung cancer screening. This study aimed to assess inter-observer agreement on National Comprehensive Cancer Network (NCCN) guideline-based follow-up recommendation for subsolid nodules from low-dose computed tomography (LDCT) screening. A retrospective collection of LDCT reports from 2014 to 2017 for lung cancer screening was conducted using the Radiology Information System and keyword searches, focusing on subsolid nodules. A total of 110 LDCT cases containing subsolid nodules were identified. Two senior radiologists provided standardized follow-up recommendation. Follow-up recommendation was categorized into four groups (0-, 3-, 6-, and 12-month). To ensure overall balance and representativeness of the follow-up categories, 60 scans from 60 participants were included (distribution ratio 1:1:2:2). Cases were categorised into follow-up recommendation groups by five observers following NCCN guidelines. Fleiss' kappa statistic was used to evaluate inter-observer agreement. Overall accuracy rate for follow-up recommendation among five observers was 72.3%. Chest radiologists' overall agreement was significantly higher than radiology residents (P<0.01). The overall agreement among the five observers was moderate, with a Fleiss' kappa of 0.437. For all paired readers, the mean Cohen's kappa value was 0.603, with 95% confidence interval (CI) from 0.489 to 0.716. Chest radiologists demonstrated substantial agreement, evidenced by a Cohen's kappa of 0.655 (95% CI: 0.503-0.807). In contrast, the mean Cohen's kappa among radiology residents was 0.533 (95% CI: 0.501-0.565). The majority of cases with discrepancies, accounting for 73.5%, were associated with the same risk-dominant nodules. A higher proportion of part-solid nodule was a risk factor for discrepancies. Of the 600 paired readings, major discrepancies and substantial discrepancies were observed in 27.5% and 4.8% (29/600) of the cases. In subsolid nodules, category evaluation of observer follow-up recommendation based on NCCN guidelines achieved moderate consistency. Disagreements were mainly caused by measurement and type disagreements of identical risk-dominant nodules. Part-solid nodule was a contributor for discrepancies in follow-up recommendation. Major and substantial management discrepancies were 27.5% and 4.8% in the paired evaluations.

Intervertebral Cervical Disc Intensity (IVCDI) Detection and Classification on MRI Scans Using Deep Learning Methods

ABSTRACTRadiologists manually interpret magnetic resonance imaging (MRI) scans for the detection of intervertebral cervical disc degeneration, which are often obtained in a primary care or emergency hospital context. The ability of computer models to work with pathological findings and aid in the first interpretation of medical imaging tests is widely acknowledged. Deep learning methods, which are commonly employed today in the diagnosis or detection of many diseases, show great promise in this area. For the detection and segmentation of intervertebral cervical disc intensity, we propose a Mask‐RCNN‐based deep learning algorithm in this study. The provided approach begins by creating an original dataset using MRI scans that were collected from Yozgat Bozok University. The senior radiologist labels the data, and three classes of intensity are chosen for the classification (low, intermediate, and high). Two alternative network backbones are used in the study, and as a consequence of the training for the Mask R‐CNN algorithm, 98.14% and 96.72% mean average precision (mAP) values are obtained with the ResNet50 and ResNet101 architectures, respectively. Utilizing the five‐fold cross‐validation approach, the study is conducted. This study also applied the Faster R‐CNN method, achieving a mAP value of 85.2%. According to the author's knowledge, no study has yet been conducted to apply deep learning algorithms to detect intervertebral cervical disc intensity in a patient population with cervical intervertebral disc degeneration. By ensuring accurate MRI image interpretation and effectively supplying supplementary diagnostic information to provide accuracy and consistency in radiological diagnosis, the proposed method is proving to be a highly useful tool for radiologists.

The impact of deep learning image reconstruction of spectral CTU virtual non contrast images for patients with renal stones

PurposeTo compare image quality and detection accuracy of renal stones between deep learning image reconstruction (DLIR) and Adaptive Statistical Iterative Reconstruction-Veo (ASIR-V) reconstructed virtual non-contrast (VNC) images and true non-contrast (TNC) images in spectral CT Urography (CTU). MethodsA retrospective analysis was conducted on images of 70 patients who underwent abdominal-pelvic CTU in TNC phase using non-contrast scan and contrast-enhanced corticomedullary phase (CP) and excretory phase (EP) using spectral scan. The TNC scan was reconstructed using ASIR-V70 % (TNC-AR70), contrast-enhanced scans were reconstructed using AR70, DLIR medium-level (DM), and high-level (DH) to obtain CP-VNC-AR70/DM/DH and EP-VNC-AR70/DM/DH image groups, respectively. CT value, image quality and kidney stones quantification accuracy were measured and compared among groups. The subjective evaluation was independently assessed by two senior radiologists using the 5-point Likert scale for image quality and lesion visibility. ResultsDH images were superior to AR70 and DM images in objective image quality evaluation. There was no statistical difference in the liver and spleen (both P > 0.05), or within 6HU in renal and fat in CT value between VNC and TNC images. EP-VNC-DH had the lowest image noise, highest SNR, and CNR, and VNC-AR70 images had better noise and SNR performance than TNC-AR70 images (all p < 0.05). EP-VNC-DH had the highest subjective image quality, and CP-VNC-DH performed the best in lesion visibility. In stone CT value and volume measurements, there was no statistical difference between VNC and TNC (P > 0.05). ConclusionThe DLIR-reconstructed VNC images in CTU provide better image quality than the ASIR-V reconstructed TNC images and similar quantification accuracy for kidney stones for potential dose savings.The study highlights that deep learning image reconstruction (DLIR)-reconstructed virtual non-contrast (VNC) images in spectral CT Urography (CTU) offer improved image quality compared to traditional true non-contrast (TNC) images, while maintaining similar accuracy in kidney stone detection, suggesting potential dose savings in clinical practice.

Automatic detection of adenoid hypertrophy on lateral nasopharyngeal radiographs of children based on deep learning.

Adenoid hypertrophy is a prevalent cause of upper airway obstruction in children, potentially leading to various otolaryngological complications and even systemic sequelae. The lateral nasopharyngeal radiograph is routinely employed for the diagnosis of adenoid hypertrophy. This study aimed to evaluate the accuracy and reliability of deep learning, using lateral nasopharyngeal radiographs, for the diagnosis of adenoid hypertrophy in pediatric patients. In the retrospective study, the lateral nasopharyngeal X-ray images were collected from children receiving therapy in the Children's Hospital of Soochow University, the 983th Hospital of Joint Logistics Support Forces of Chinese PLA and the Suzhou Wujiang District Children's Hospital from January 2023 to November 2023. Five deep learning models, i.e., AlexNet, VGG16, Inception v3, ResNet50 and DenseNet121, were used for model training and validation. Receiver operating characteristic (ROC) curve analyses were used to evaluate the performance of each model. The best algorithm was compared with interpretations from three radiologists on 208 images in the internal validation group. The lateral nasopharyngeal X-ray images were collected from 1,188 children, including 705 males (59.3%) and 483 females (40.7%), aged 8 months to 13 years, with a mean age of 5.57±2.66 years. Among the five deep learning models, DenseNet-121 performed the best, with area under the curve (AUC) values of 0.892 and 0.872, with accuracy of 0.895 and 0.878, sensitivity of 0.870 and 0.838, and specificity of 0.913 and 0.906 in the internal and external validation groups, respectively. The diagnostic performance of DenseNet-121 was higher than that of the junior and mid-level radiologists (0.892 vs. 0.836, 0.892 vs. 0.869), close to the senior radiologist (0.892 vs. 0.901). However, Delong's test revealed no significant difference between DenseNet121 and each radiologist in the validation group (P=0.24, P=0.52, P=0.79). All the five deep learning models in the study showed good performance for the diagnosis of adenoid hypertrophy, with DenseNet121 being the best, which was clinically relevant for the automatic identification of adenoid hypertrophy.

Assessing GPT-4 multimodal performance in radiological image analysis.

This study aims to assess the performance of a multimodal artificial intelligence (AI) model capable of analyzing both images and textual data (GPT-4V), in interpreting radiological images. It focuses on a range of modalities, anatomical regions, and pathologies to explore the potential of zero-shot generative AI in enhancing diagnostic processes in radiology. We analyzed 230 anonymized emergency room diagnostic images, consecutively collected over 1 week, using GPT-4V. Modalities included ultrasound (US), computerized tomography (CT), and X-ray images. The interpretations provided by GPT-4V were then compared with those of senior radiologists. This comparison aimed to evaluate the accuracy of GPT-4V in recognizing the imaging modality, anatomical region, and pathology present in the images. GPT-4V identified the imaging modality correctly in 100% of cases (221/221), the anatomical region in 87.1% (189/217), and the pathology in 35.2% (76/216). However, the model's performance varied significantly across different modalities, with anatomical region identification accuracy ranging from 60.9% (39/64) in US images to 97% (98/101) and 100% (52/52) in CT and X-ray images (p < 0.001). Similarly, pathology identification ranged from 9.1% (6/66) in US images to 36.4% (36/99) in CT and 66.7% (34/51) in X-ray images (p < 0.001). These variations indicate inconsistencies in GPT-4V's ability to interpret radiological images accurately. While the integration of AI in radiology, exemplified by multimodal GPT-4, offers promising avenues for diagnostic enhancement, the current capabilities of GPT-4V are not yet reliable for interpreting radiological images. This study underscores the necessity for ongoing development to achieve dependable performance in radiology diagnostics. Although GPT-4V shows promise in radiological image interpretation, its high diagnostic hallucination rate (> 40%) indicates it cannot be trusted for clinical use as a standalone tool. Improvements are necessary to enhance its reliability and ensure patient safety. GPT-4V's capability in analyzing images offers new clinical possibilities in radiology. GPT-4V excels in identifying imaging modalities but demonstrates inconsistent anatomy and pathology detection. Ongoing AI advancements are necessary to enhance diagnostic reliability in radiological applications.

Preoperative Atelectasis in Patients with Obesity Undergoing Bariatric Surgery: A Cross-Sectional Study.

Pulmonary atelectasis is present even before surgery in patients with obesity. We aimed to estimate the prevalence and extension of preoperative atelectasis in patients with obesity undergoing bariatric surgery and to determine if variation in preoperative Spo2 values in the seated position at room air is explained by the extent of atelectasis coverage in the supine position. This was a cross-sectional study in a single center specialized in laparoscopic bariatric surgery. Preoperative chest computed tomographies were reassessed by a senior radiologist to quantify the extent of atelectasis coverage as a percentage of total lung volume. Patients were classified as having atelectasis when the affection was ≥2.5%, to estimate the prevalence of atelectasis. Crude and adjusted prevalence ratios (aPRs) and odds ratios (aORs) were obtained to assess the relative prevalence of atelectasis and percentage coverage, respectively, with increasing obesity category. Inverse probability weighting was used to assess the total, direct (not mediated), and indirect (mediated through atelectasis) effects of body mass index (BMI) on preoperative Spo2, and to quantify the magnitude of mediation (proportion mediated). E-values were calculated, to represent the minimum magnitude of association that an unmeasured confounder with the same directionality of the effect should have to drive the observed point estimates or lower confidence intervals (CIs) to 1, respectively. In 236 patients with a median BMI of 40.3 kg/m2 (interquartile range [IQR], 34.6-46.0, range: 30.0-77.3), the overall prevalence of atelectasis was 32.6% (95% CI, 27.0-38.9) and by BMI category: 30 to 35 kg/m2, 12.7% (95% CI, 6.1-24.4); 35 to 40 kg/m2, 28.3% (95% CI, 17.2-42.6); 40 to 45 kg/m2, 12.3% (95% CI, 5.5-24.3); 45 to 50 kg/m2, 48.4% (95% CI, 30.6-66.6); and ≥50 units, 100% (95% CI, 86.7-100). Compared to the 30 to 35 kg/m2 group, only the categories with BMI ≥45 kg/m2 had significantly higher relative prevalence of atelectasis-45 to 50 kg/m2, aPR = 3.52 (95% CI, 1.63-7.61, E-value lower bound: 2.64) and ≥50 kg/m2, aPR = 8.0 (95% CI, 4.22-15.2, E-value lower bound: 7.91)-and higher odds of greater atelectasis percentage coverage: 45-50 kg/m2, aOR = 7.5 (95% CI, 2.7-20.9) and ≥50 kg/m2, aOR = 91.5 (95% CI, 30.0-279.3). Atelectasis percent alone explained 70.2% of the variation in preoperative Spo2. The proportion of the effect of BMI on preoperative Spo2 values <96% mediated through atelectasis was 81.5% (95% CI, 56.0-100). The prevalence and extension of atelectasis increased with higher BMI, being significantly higher at BMI ≥45 kg/m2. Preoperative atelectasis mediated the effect of BMI on Spo2 at room air in the seated position.

Identifying enthesitis in the sacroiliac joints in patients with axial spondyloarthritis by readers of varying experience: impact of the learning progress

BackgroundThis study aimed to investigate the accuracy of identifying enthesitis along with other inflammatory lesions and structural lesions on the MRI of the sacroiliac joints (SIJ) by readers of varying experience and how training sessions and workshops could help improve the accuracy.MethodsA total of 224 patients with clinical diagnosis of axial spondyloarthritis who underwent SIJ MRI examinations were retrospectively included in this study. Three readers with 5 years, 3 years and 1 year of experience in musculoskeletal imaging were invited to review the SIJ MRI images independently, while the imaging reports of a senior radiologist (> 10 years’ experience) were used as reference. After the first round of image review, a training session and a workshop on the imaging of SIJ in spondyloarthritis were held and the three readers were asked to review the images in the second round. We calculated the accuracy of identifying inflammatory and structural lesions of the three readers as well as the intra-reader agreement.ResultsEnthesitis could be observed in 52.23% of the axial spondyloarthritis patients, while 81.58% of the patients with enthesitis were accompanied with bone marrow edema. All the three readers showed better accuracy at identifying structural lesions than inflammatory lesions. In the first round of image review, the three readers only correctly identified 15.07%, 2.94% and 0.74% of the enthesitis sites. After the training session and workshop, the accuracy rose to 61.03%, 39.34% and 20.22%. The intra-reader agreement of enthesitis calculated as Cohen’s kappa was 0.23, 0.034 and 0.014, respectively.ConclusionReaders with less experience in musculoskeletal imaging showed lower accuracy of identifying inflammatory lesions, notably enthesitis. Training sessions and workshops could help improve the diagnostic accuracy of the junior readers.

Localization and etiological stratification of non-neoplastic small bowel bleeding via CT imaging: a 10-year study

ObjectivesThe purpose of this study is to assess the diagnostic efficacy of contrast-enhanced CT scans for small bowel bleeding.MethodsThis retrospective study evaluated patients diagnosed with non-neoplastic small intestinal bleeding (including duodenum) who underwent abdominal CT at our institution from December 2013 to March 2023. Patients were categorized into diverticulum and non-diverticulum groups based on the cause of bleeding. Active bleeding was defined on the CT images as extravasation of contrast material in the intestinal lumen during the arterial phase and/or progressive accumulation of contrast material during the venous phase. We have documented the original report (extracted from the medical record system and additional consultation opinions from senior radiologists), including the presence of active bleeding and its potential bleeding location. Furthermore, two radiologists reassessed the CT images, seeking consensus on the diagnosis between them.ResultsThe study included 165 patients, predominantly male, with a median age of 30 years. Active bleeding was identified in 48.3% of patients. Notably, all identified bleeding diverticula in the diverticulum group exhibited cul-de-sac termination. Among the identified causes of bleeding, Crohn’s disease was most prevalent (46.7%, N of causes = 64). Significant differences were observed in the diagnostic methods between the diverticulum and non-diverticulum groups, with surgery predominantly applied in the diverticulum group, and endoscopy in the non-diverticulum group (n = 49 vs n = 15, p = 0.001). Contrast agent extravasation was significantly higher in the diverticulum group (n = 54 vs n = 16, p = 0.001), and Meckel’s diverticulum cases appearing tubular were significantly higher than in other diverticulum cases (n = 25 vs n = 3, p < 0.001).ConclusionCT allows for a higher detection rate of diverticular bleeding, even if asymptomatic, guiding classification into multiple potentially clinically relevant categories.Critical relevance statementContrast-enhanced CT imaging is effective in determining the location and cause of non-neoplastic small bowel bleeding, especially diverticular bleeding. Therefore, the use of enhanced CT should be prioritized in the diagnosis and management of small bowel bleeding.Key PointsCT has potential value in the diagnosis of small bowel bleeding.CT imaging suggests possible surgical intervention for active bleeding detection.CT diagnoses and localizes small bowel bleeding, aiding in treatment and prioritizing in guidelines.Graphical

Role of MRI in classification and preoperative evaluation of perianal fistula

BackgroundA fistula is an unusual communication between the skin and an inward organ or between the organs. The current study aimed to determine the role of magnetic resonance imaging (MRI) in the diagnosis, description, and classification of fistula-in-ano. Additionally, the study aimed to compare the MRI findings with the operative findings, which is considered the gold standard in our study.MethodsThe present study is a prospective study and was carried out on a group of 61 patients: 52 males and 9 females with a mean age of 41 years, who were suspected to have perianal fistulae presented clinically with discharge or localized pain or were asymptomatic. A preoperative MRI was conducted for fistula-in-ano evaluation in these patients from October 2019 till the end of October 2021. A surgeon conducted a physical examination of all patients to document the number and location of cutaneous openings following a full medical history. The MRI images were evaluated and interpreted by multiple expert radiologists who had more than 5 years of experience in analyzing MRI. In the cases (in 3 cases, 4.92%) where there were discrepancies in their interpretations, a senior radiologist’s evaluation was considered the final result and was confirmed surgically. Subsequently, the MRI findings were correlated with surgical findings to indicate the specificity, sensitivity, and accuracy of such MRI findings.ResultsThe MRI technique was the routine (i.e., without contrast administration) technique in 36 (59.02%) patients, with contrast administration in 21 (34.43%) patients and with anesthesia/sedation (uncomfortable, anxious, and claustrophobic patients need to minimize their motion artifact to improve the quality and obtain more detailed images) in 4 (6.56%) patients. The MRI pathology of the studied patients revealed normal findings in 1 (1.64%) patient, anal fissure/early developing tract in 6 (9.84%) patients, and established tract (sinus/fistula) in 54 (88.52%) patients. Park’s and St. James's University Hospital classifications were utilized to categorize the patients, with the most common fistula types based on the Park’s classification: intersphincteric (44.3%) and transsphincteric (29.5%). Based on the St. James’s classification, grade 1 (intersphincteric (34.4%)) and grade 4 (transsphincteric with abscess/side branch (16.4%)) were the most prevalent. A substantial consensus was reached between MRI and surgery findings for classifying tracts, side branches, and abscesses formation with sensitivity, specificity, and accuracy of 100%, 85.71%, and 98.36%, respectively.ConclusionsMRI is a valuable tool in managing patients with perianal fistulas due to its ability to detect hidden areas of infection (such as abscesses) and secondary extensions. These factors contribute to the high recurrence rate after surgery. Additionally, MR imaging can define the anatomical relationships between the fistula and anal sphincters, helping to predict the likelihood of fecal incontinence following surgery.

Diagnostic Accuracy of Ultra-Low Dose CT Compared to Standard Dose CT for Identification of Fresh Rib Fractures by Deep Learning Algorithm.

The present study aimed to evaluate the diagnostic accuracy of ultra-low dose computed tomography (ULD-CT) compared to standard dose computed tomography (SD-CT) in discerning recent rib fractures using a deep learning algorithm detection of rib fractures (DLADRF). A total of 158 patients undergoing forensic diagnosis for rib fractures were included in this study: 50 underwent SD-CT, and 108 were assessed using ULD-CT. Junior and senior radiologists independently evaluated the images to identify and characterize the rib fractures. The sensitivity of rib fracture diagnosis by radiologists and radiologist + DLADRF was better using SD-CT than ULD-CT. However, the diagnosis sensitivity of DLADRF using ULD-CT alone was slightly more than SD-CT. Nonetheless, no substantial differences were observed in specificity, positive predictive value, and negative predictive value between SD-CT and ULD-CT by the same radiologist, radiologist + DLADRF, and DLADRF (P > 0.05). The area under the curve (AUC) of receiver operating characteristic indicated that senior radiologist + DLADRF was significantly better than senior and junior radiologists, junior radiologists + DLADRF, and DLADRF alone using SD-CT or ULD-CT (all P < 0.05). Also, junior radiologists + DLADRF was better with ULD-CT than senior and junior radiologists (P < 0.05). The AUC of the rib fracture diagnosed by senior radiologists did not differ from DLADRF using ULD-CT. Also, no significant differences were observed between junior + AI and senior and between junior and DLADRF using SD-CT. DLADRF enhanced the diagnostic performance of radiologists in detecting recent rib fractures. The diagnostic outcomes between SD-CT and ULD-CT across radiologists' experience and DLADRF did not differ significantly.

Comprehensive imaging evaluation of the aortic valve and root before aortic root surgery: a study comparing MDCT and TEE

ObjectiveTo investigate the accuracy and consistency of MDCT and TEE in the preoperative assessment of aortic root surgery.MethodsFrom January 2021 to September 2022, 118 patients who underwent aortic root surgery were included in this study. All patients underwent high-quality preoperative MDCT and TEE examinations, and the examination results were independently measured and assessed by two senior radiologists or ultrasound specialists. Bland–Altman analysis and Pearson correlation testing were employed to assess the correlation and consistency between MDCT and TEE. These analyses were then compared with actual intraoperative measurement data.ResultsAmong all the patients, 73 (61.86%) had tricuspid aortic valve (TAV), and 45 (38.14%) had bicuspid aortic valve (BAV). A comparison between the TEE and MDCT measurements showed that for the annulus diameter, the area-derived diameter had the best correlation and agreement. For the sinus of Valsalva diameter, the circumference-derived diameter was optimal. However, for the STJ diameter, the minimum cross-sectional diameter showed the best agreement with TEE. In contrast, measurements of geometric height showed a weaker correlation and agreement.ConclusionContrast-enhanced MDCT can be a valuable tool for perioperative evaluation in aortic root surgery, with good correlation, consistency, and feasibility when compared to TEE. The choice of MDCT measurement methodology, specifically area-derived and circumference-derived diameter, proved to be more accurate than other methods. Further research is required to enhance the understanding of aortic valve repair and associated imaging techniques.

The clinical value of artificial intelligence in assisting junior radiologists in thyroid ultrasound: a multicenter prospective study from real clinical practice

BackgroundThis study is to propose a clinically applicable 2-echelon (2e) diagnostic criteria for the analysis of thyroid nodules such that low-risk nodules are screened off while only suspicious or indeterminate ones are further examined by histopathology, and to explore whether artificial intelligence (AI) can provide precise assistance for clinical decision-making in the real-world prospective scenario.MethodsIn this prospective study, we enrolled 1036 patients with a total of 2296 thyroid nodules from three medical centers. The diagnostic performance of the AI system, radiologists with different levels of experience, and AI-assisted radiologists with different levels of experience in diagnosing thyroid nodules were evaluated against our proposed 2e diagnostic criteria, with the first being an arbitration committee consisting of 3 senior specialists and the second being cyto- or histopathology.ResultsAccording to the 2e diagnostic criteria, 1543 nodules were classified by the arbitration committee, and the benign and malignant nature of 753 nodules was determined by pathological examinations. Taking pathological results as the evaluation standard, the sensitivity, specificity, accuracy, and area under the receiver operating characteristic curve (AUC) of the AI systems were 0.826, 0.815, 0.821, and 0.821. For those cases where diagnosis by the Arbitration Committee were taken as the evaluation standard, the sensitivity, specificity, accuracy, and AUC of the AI system were 0.946, 0.966, 0.964, and 0.956. Taking the global 2e diagnostic criteria as the gold standard, the sensitivity, specificity, accuracy, and AUC of the AI system were 0.868, 0.934, 0.917, and 0.901, respectively. Under different criteria, AI was comparable to the diagnostic performance of senior radiologists and outperformed junior radiologists (all P < 0.05). Furthermore, AI assistance significantly improved the performance of junior radiologists in the diagnosis of thyroid nodules, and their diagnostic performance was comparable to that of senior radiologists when pathological results were taken as the gold standard (all p > 0.05).ConclusionsThe proposed 2e diagnostic criteria are consistent with real-world clinical evaluations and affirm the applicability of the AI system. Under the 2e criteria, the diagnostic performance of the AI system is comparable to that of senior radiologists and significantly improves the diagnostic capabilities of junior radiologists. This has the potential to reduce unnecessary invasive diagnostic procedures in real-world clinical practice.

Subspecialized medical team mode facilitates radiology resident training

AbstractPurposeOur institution developed the subspecialized medical team (SMT) mode in radiology reporting workflows in July 2022 to strengthen professionalism in radiology residency training. This study aimed to investigate whether adopting the SMT mode facilitated radiology resident training.MethodThis observational study was conducted in October 2022. An online questionnaire was distributed among first‐to third‐year radiology residents to determine their levels of perceived stress, anxiety, competency in image interpretation, and learning effectiveness in using the SMT mode and other approaches. Responses were on 4 or 5 point Likert scales, and multiple‐choice responses were used for semi‐open questions. The turnaround time for radiology reports was measured for both residents and senior radiologists. Statistical significance (p &lt; 0.05) was determined using chi‐square or Fisher's exact tests, one‐way analysis of variance, pairwise t‐tests, and the Kruskal‒Wallis H test.ResultsIn total, 71 residents completed the questionnaire survey. The turnaround times for radiology reports were collected for these residents and 23 senior radiologists. Residents' stress and anxiety scores, perceived levels of competency in image interpretation, and time spent writing reports were significantly increased in the SMT mode. In contrast, the time senior radiologists spent reviewing a single report was decreased for the majority of report types after 1 year of implementing the SMT mode.ConclusionThese findings suggest adopting the SMT mode could facilitate the training of radiology residents. However, residents' mental health and well‐being should also be considered.

Diagnostic MRI for deep pelvic endometriosis: towards a standardized protocol?

To assess the diagnostic efficacy of an MRI protocol and patient preparation in detecting deep pelvic endometriosis (DPE). The cohort is from the ENDOVALIRM database, a multicentric national retrospective study involving women who underwent MRI followed by pelvic surgery for endometriosis (reference standard). Two senior radiologists independently analyzed MRI findings using the deep pelvic endometriosis index (dPEI) to determine lesion locations. The study evaluated the impact of bowel preparation, vaginal and rectal opacification, MRI unit type (1.5-T or 3-T), additional sequences (thin slice T2W or 3DT2W), and gadolinium injection on reader performance for diagnosing DPE locations. Fisher's exact test assessed differences in diagnostic accuracy based on patient preparation and MRI parameters. The final cohort comprised 571 women with a mean age of 33.3 years (± 6.6 SD). MRI with bowel preparation outperformed MRI without bowel preparation in identifying torus/uterosacral ligament (USL) locations (p < 0.0001) and rectosigmoid nodules (p = 0.01). MRI without vaginal opacification diagnosed 94.1% (301/320) of torus/USL locations, surpassing MR with vaginal opacification, which diagnosed 85% (221/260) (p < 0.001). No significant differences related to bowel preparation or vaginal opacification were observed for other DPE locations. Rectal opacification did not affect diagnostic accuracy in the overall population, except in patients without bowel preparation, where performance improved (p = 0.04). There were no differences in diagnostic accuracy regarding MRI unit type (1.5-T/3-T), presence of additional sequences, or gadolinium injection for any endometriotic locations. Bowel preparation prior to MRI examination is preferable to rectal or vaginal opacification for diagnosing deep endometriosis pelvic lesions. Accurate diagnosis and staging of DPE are essential for effective treatment planning. Bowel preparation should be prioritized over rectal or vaginal opacification in MRI protocols. Optimizing MRI protocols for diagnostic performance with appropriate opacification techniques will help diagnose deep endometriosis more accurately. Evaluating deep endometriosis in collapsible organs such as the vagina and rectum is difficult. Bowel preparation and an absence of vaginal opacification were found to be diagnostically beneficial. Bowel preparation should be prioritized over rectal or vaginal opacification in MRI protocols.

Automated grading of prenatal hydronephrosis severity from segmented kidney ultrasounds using deep learning

Background and motivationsAntenatal or prenatal hydronephrosis (AHN) is a common kidney complication in unborn children. While AHN is generally benign and resolves over time, this condition can inflict serious kidney damage or even organ failure due to excessive waste accumulation in severe cases. Regardless of severity, AHN must be clinically monitored for resolution, with treatment plans and medications being revised according to updated prognoses. Kidney ultrasound images are one of the most common methods of monitoring AHN, but grading of this condition is highly subjective and clinicians may select inappropriate therapies or surgical interventions as a result. New approaches are required to differentiate subjects who can be managed without surgical intervention from those who require life-saving operations. MethodsAn end-to-end deep machine learning framework was developed to sequentially detect ultrasound regions of interest, segment kidneys from US images, and classify AHN severity. We propose the novel Kidney Ultrasound Segmentation Network (KUSNet) for kidney segmentation from ultrasound images, and the Prenatal Hydronephrosis Classification Network (PHCNet) for hydronephrosis severity stratification according to the Society of Fetal Urology (SFU) standards. The ground truth kidney masks were generated by two radiologists with more than five years of working experience while the SFU-based annotations for the AHN severity were done by two senior radiologists and three senior urologists with more than ten years of domain expertise. At each stage, the performance of the proposed models was assessed both quantitatively and qualitatively against state-of-the-art networks in the respective fields. ResultsThe proposed KUSNet for ultrasound kidney segmentation achieved 97.6% accuracy, 97.4% precision, 97.6% recall or sensitivity, 97.5% F1-score, 95.5% IoU or Jaccard score, and 92.1% Dice score, beating several state-of-the-arts. On the other hand, the novel PHCNet reached 93.9% accuracy, 93.7% precision, 93.9% recall, 93.8% specificity, and 89.0% F1-score subject-wise when performing multiclass stratification of AHN severity on segmented kidney regions. ConclusionArtificial intelligence-based tools can reliably classify AHN severity to reduce inter- and intra-observer bias, thereby aiding clinicians in the rapid selection of appropriate treatments and surgeries. Moreover, segmenting kidney regions beforehand significantly boosts AHN severity classification performance.

Prognostic relationship between high sensitivity troponin I level, hematoma volume and glasgow coma score in patients diagnosed with spontaneous intracerebral hemorrhage.

Intracranial hemorrhages is one of the major causes of mortality and morbidity worldwide, and there is still no effective biomarker to predict prognosis. We aimed to determine the effectiveness of high sensitive troponin I (hs-cTn-I) levels to predict the prognosis of spontaneous intracerebral hemorrhage (sICH) by comparing Glasgow Coma Score (GCS) and hematoma volume with hs-cTn-I levels. This study was planned as a retrospective observational study. Patients with available data, over 18years old and sICH were included in the study. Cerebral computed tomography images were evaluated by a senior radiologist. Hematoma volume was calculated using the ABC/2 formula. The study comprised 206 individuals in total 78 (37.86%) women and 128 (62.13%) men. Forty-four (21.35%) of patients died. The sensitivity of GCS, hs-cTn-I, and hematoma volume values were 86.36%, 66.67%, and 59.46%, respectively, with corresponding specificities of 78.75%, 93.02%, and 87.58%. Patients with hs-cTn-I values over 26, GCS values of ≤ 9, and hematoma volume values above 44.16 were found to have higher risk of mortality (p = 0.011; p < 0.001; p < 0.001, respectively). The mortality rates were found to be increased 2.586 (IQR: 1.224-5.463) times in patients with hs-cTn-I values above 26, 0.045 times (IQR: 0.018-0.115) in patients with GCS values ≤ 9, and 7.526 times (IQR: 3.518-16.100) in patients with hematoma volume values above 44.16. Our findings suggest that hs-cTn-I values exceeding 26 units may serve as effective biochemical markers for predicting the prognosis of patients with sICH.

Automated weight-bearing foot measurements using an artificial intelligence-based software.

To assess the accuracy of an artificial intelligence (AI) software (BoneMetrics, Gleamer) in performing automated measurements on weight-bearing forefoot and lateral foot radiographs. Consecutive forefoot and lateral foot radiographs were retrospectively collected from three imaging institutions. Two senior musculoskeletal radiologists independently annotated key points to measure the hallux valgus, first-second metatarsal, and first-fifth metatarsal angles on forefoot radiographs and the talus-first metatarsal, medial arch, and calcaneus inclination angles on lateral foot radiographs. The ground truth was defined as the mean of their measurements. Statistical analysis included mean absolute error (MAE), bias assessed with Bland-Altman analysis between the ground truth and AI prediction, and intraclass coefficient (ICC) between the manual ratings. Eighty forefoot radiographs were included (53 ± 17years, 50 women), and 26 were excluded. Ninety-seven lateral foot radiographs were included (51 ± 20years, 46 women), and 21 were excluded. MAE for the hallux valgus, first-second metatarsal, and first-fifth metatarsal angles on forefoot radiographs were respectively 1.2° (95% CI [1; 1.4], bias = - 0.04°, ICC = 0.98), 0.7° (95% CI [0.6; 0.9], bias = - 0.19°, ICC = 0.91) and 0.9° (95% CI [0.7; 1.1], bias = 0.44°, ICC = 0.96). MAE for the talus-first, medial arch, and calcaneal inclination angles on the lateral foot radiographs were respectively 3.9° (95% CI [3.4; 4.5], bias = 0.61° ICC = 0.88), 1.5° (95% CI [1.2; 1.8], bias = - 0.18°, ICC = 0.95) and 1° (95% CI [0.8; 1.2], bias = 0.74°, ICC = 0.99). Bias and MAE between the ground truth and the AI prediction were low across all measurements. ICC between the two manual ratings was excellent, except for the talus-first metatarsal angle. AI demonstrated potential for accurate and automated measurements on weight-bearing forefoot and lateral foot radiographs.

A cross-sectional study of synchronous liver metastasis in patients with gastric cancer.

e16015 Background: The specific organ microenvironment can promote the colonization of tumor cells. However, lack of research shed light on the liver microenvironment in synchronous liver metastasis in patients with gastric cancer (synGCLM). With the epidemic of hepatitis B virus (HBV) infection in China, we aim to explore the influence of hepatitis B surface antigen (HBsAg) on the risk of synGCLM. Methods: It was a cross-sectional study that enrolled 858 cases of newly diagnosed gastric cancer (GC). SynGCLM confirmed by biopsy or imaging evaluation reported by two senior radiologists. HBsAg-positive was defined as the detection of serum HBsAg more than 10 IU/mL. The prevalence of synGCLM was compared between patients with HBsAg (HBsAg+) and those without HBsAg (HBsAg-). The risk factors for synGCLM were analyzed by univariate and multivariate logistic regression analyses. Especially in the HBsAg+ GC patients, aspartate aminotransferase to platelet ratio index (APRI), liver fibrosis-4 index (FIB-4) levels and hepatitis B e antigen (HBeAg) status were further analyzed. Results: The prevalence of synGCLM of the HBsAg+ group was 16.7%, higher compared to 7.0% of the HBsAg- group, which was of statistical significance (p = 0.014). Multivariate logistic regression analysis demonstrated that HBsAg (Odds ratio (OR) = 3.359, p = 0.008), the elevated level of carcinoembryonic antigen (CEA) (OR = 3.775, p &lt; 0.001), alpha-fetoprotein (AFP) (OR = 6.204, p &lt; 0.001), γ-glutamyltransferase (GGT) (OR = 5.081, p &lt; 0.001), alkaline phosphatase (ALP) (OR = 3.968, p = 0.002) and age (OR = 1.028, p = 0.033) were risk factors for synGCLM. Among the HBsAg+ patients, both ARPI and FIB-4 were statistically higher in the patients with synGCLM (synGCLM+) compared to those without synGCLM (synGCLM-) (ARPI: p = 0.045; FIB-4: p = 0.047); HBeAg positivity was detected in 20.0% of synGCLM+ patients compared to 6.0% of synGCLM- patients, but the difference was of no significance (p = 0.190). Conclusions: HBV infection is a risk factor for synGCLM whereas elevated ARPI and FIB-4 may be pro-metastatic especially among the HBsAg+ GC patients. Age, preoperative CEA, AFP, GGT and ALP levels are associated with increased risk of synGCLM.