Diagnostic effect of artificial intelligence solution for referable thoracic abnormalities on chest radiography: a multicenter respiratory outpatient diagnostic cohort study.

OBJECTIVES: We aim ed to evaluate a commercial artificial intelligence (AI) solution on a multicenter cohort of chest radiographs and to compare physicians' ability to detect and localize referable thoracic abnormalities with and without AI assistance. METHODS: In this retrospective diagnostic cohort study, we investigated 6,006 consecutive patients who underwent both chest radiography and CT. We evaluated a commercially available AI solution intended to facilitate the detection of three chest abnormalities (nodule/masses, consolidation, and pneumothorax) against a reference standard to measure its diagnostic performance. Moreover, twelve physicians, including thoracic radiologists, board-certified radiologists, radiology residents, and pulmonologists, assessed a dataset of 230 randomly sampled chest radiographic images. The images were reviewed twice per physician, with and without AI, with a 4-week washout period. We measured the impact of AI assistance on observer's AUC, sensitivity, specificity, and the area under the alternative free-response ROC (AUAFROC). RESULTS: In the entire set (n = 6,006), the AI solution showed average sensitivity, specificity, and AUC of 0.885, 0.723, and 0.867, respectively. In the test dataset (n = 230), the average AUC and AUAFROC across observers significantly increased with AI assistance (from 0.861 to 0.886; p = 0.003 and from 0.797 to 0.822; p = 0.003, respectively). CONCLUSIONS: The diagnostic performance of the AI solution was found to be acceptable for the images from respiratory outpatient clinics. The diagnostic performance of physicians marginally improved with the use of AI solutions. Further evaluation of AI assistance for chest radiographs using a prospective design is required to prove the efficacy of AI assistance. KEY POINTS: • AI assistance for chest radiographs marginally improved physicians' performance in detecting and localizing referable thoracic abnormalities on chest radiographs. • The detection or localization of referable thoracic abnormalities by pulmonologists and radiology residents improved with the use of AI assistance.

Efficacy and Safety of Transvenous Embolization of Type II Renal Arteriovenous Malformations with Coils.

PURPOSE: To evaluate the efficacy and safety of transvenous coil embolization of the venous sac for type II renal arteriovenous malformation (AVM). MATERIALS AND METHODS: A retrospective review was conducted of 8 patients (5 women and 3 men; mean age, 57 years; age range, 41-69 years) who underwent transvenous coil embolization for type II congenital renal AVM at 5 different hospitals between 2012 and 2018. Additional intra-arterial ethanol injection was performed if shunt flow persisted after venous sac coiling. Technical success was defined as complete occlusion of shunt flow with coil embolization. Clinical success was defined as no symptom recurrence during the follow-up period. The renal parenchymal infarction rate was measured on computed tomography (CT), and procedure-related complications were reviewed. RESULTS: Nine sessions of embolization were performed for 8 patients. The mean venous sac size was 24 mm (range, 10-39 mm), and a mean of 14 micro and/or micro-detachable coils (range, 3-50) were used. The technical success rate was 88% (7 of 8) using coil embolization. One patient (12%) required additional ethanol injection to complete occlusion of the shunt flow and had a less than 10% parenchymal infarction on follow-up CT. No procedure-related complications or recurrences occurred during a mean clinical follow-up period of 20.8 months (range, 4.7-76.6 months). CONCLUSIONS: Transvenous coil embolization of type II renal AVM showed an 88% technical success rate. One patient (12%) showed less than 10% renal parenchymal infarction after additional ethanol injection. No additional complications or recurrences occurred during the follow-up period.

Utility of Preoperative Ultrasonography in Transferred Patients with Suspicious Malignancy on Ultrasonography-Guided Fine-Needle Aspiration Cytology of Thyroid Nodules: A Single-Center Retrospective Study.

BACKGROUND This study aimed to assess the utility and characteristics of preoperative ultrasonography (US) in patients transferred to referral hospitals from local clinics with a diagnosis of malignancy on US-guided fine-needle aspiration cytology of thyroid nodules. MATERIAL AND METHODS From January 2018 to June 2018, 109 transferred patients underwent preoperative US in our hospital for suspected thyroid malignancy on cytological analysis after US-guided fine-needle aspiration of thyroid nodules in local clinics. Preoperative US was performed by a single radiologist in all patients. Among them, 6 were excluded from the study because of refusal of thyroid surgery. Preoperative US and histopathological results were compared in all patients. RESULTS After thyroid surgery, pathological examination revealed papillary thyroid carcinoma (PTC) (n=98), follicular adenoma (n=1), and nodular hyperplasia (n=4). Of the 103 patients, 91 exhibited suspicious US findings on the preoperative US, whereas 12 did not. In the 91 patients with suspicious US findings, PTC (n=90) and follicular adenoma (n=1) were confirmed after thyroid surgery. In the 12 patients with no suspicious US findings, PTC (n=8) and nodular hyperplasia (n=4) were confirmed after thyroid surgery. On repeat analysis of the cytological slides of the 4 nodular hyperplasia cases from the local clinics, Bethesda category II (n=1) and III (n=3) were determined. CONCLUSIONS In the transferred patients with a malignant cytology, preoperative US might be helpful to detect false-positive cytology cases.

Utility of routine ultrasonography follow-up after total thyroidectomy in patients with papillary thyroid carcinoma: a single-center study.

BACKGROUND: This study aimed to assess the appropriate number of sessions and interval of routine follow-up ultrasonography (US) in patients who underwent total thyroidectomy for papillary thyroid carcinoma (PTC). METHODS: Between January 2008 and December 2009, 569 patients underwent total thyroidectomy for PTC. Of the 569 patients, 44 were excluded from the study because of no US follow-up data for the neck (n = 43) or owing to indeterminate tumor recurrence/persistence (n = 1). The follow-up US for all the patients was performed by a single radiologist. Based on the cytohistopathological results, tumor recurrence/persistence was determined. RESULTS: In the 525 patients, the mean interval to the last follow-up US was 54.7 months, and the mean number of follow-up US sessions was 4.4. Of the 525 patients, 31 (5.9%) showed nodal (n = 30) and non-nodal (n = 1) tumor recurrence/persistence. Patient age and N stage were independently associated with tumor recurrence/persistence. Among patients showing tumor recurrence/persistence after total thyroidectomy, the time at first detection of suspicious US findings on follow-up US was ≤8 months in 2 patients, between 10 and 23 months in 21, and ≥ 25 months in 8. In a receiver operating characteristic curve analysis, the number of sessions and interval of the provided follow-up US were inappropriate for the detection of tumor recurrence/persistence. CONCLUSIONS: For the detection of tumor recurrence/persistence after total thyroidectomy in PTC patients, routine US follow-up with a 1- or 2-year interval may be excessive.

Persistent trigeminal artery variant terminating in the posterior inferior cerebellar artery: a case report.

Persistent trigeminal artery (PTA) is a rare cerebrovascular variation of remnant fetal carotid-vertebrobasilar anastomoses. The PTA variant terminates in the cerebellar artery with no direct connection to the basilar artery. We present a rare case of a PTA variant that terminated directly into the ipsilateral posterior inferior cerebellar artery.

Sub-100 fs Tm:MgWO4 laser at 2017 nm mode locked by a graphene saturable absorber.

We present the first sub-100 fs bulk solid-state laser in the 2-µm spectral range employing the monoclinic Tm3+-dopedMgWO4 crystal as an active medium. By applying a graphene-based saturable absorber and chirped mirrors for dispersion management, stable self-starting mode-locked operation at 2017 nm was achieved. Nearly Fourier-limited pulses as short as 86 fs featuring a bandwidth of 53 nm were generated at a repetition rate of 76 MHz. A pulse energy of 1.1 nJ was achieved at 87 MHz for a pulse duration of 96 fs. The mode-locked Tm3+:MgWO4 laser exhibits excellent stability with a fundamental beat note extinction ratio of 80 dBc above noise level.

SHC1 sensitizes cancer cells to the 8-Cl-cAMP treatment.

8-Chloro-cyclic AMP (8-Cl-cAMP) is a cyclic AMP analog that induces growth inhibition and apoptosis in a broad spectrum of cancer cells. Previously, we found that 8-Cl-cAMP-induced growth inhibition is mediated by AMP-activated protein kinase (AMPK) as well as p38 mitogen-activated protein kinase (p38 MAPK). To identify downstream mediators of the 8-Cl-cAMP signaling, we performed co-immunoprecipitation combined with mass spectrometry using the anti-AMPK or p38 MAPK antibodies. Through this approach, SHC1 was identified as one of the binding partners of p38 MAPK. SHC1 phosphorylation was suppressed by 8-Cl-cAMP in HeLa and MCF7 cancer cells, which was mediated by its metabolites, 8-Cl-adenosine and 8-Cl-ATP; however, 8-Cl-cAMP showed no effect on SHC1 phosphorylation in normal human fibroblasts. SHC1 siRNA induced AMPK and p38 MAPK phosphorylation and growth inhibition in cancer cells, and SHC1 overexpression re-sensitized human foreskin fibroblasts to the 8-Cl-cAMP treatment. SHC1 phosphorylation was unaffected by Compound C (an AMPK inhibitor) and SB203580 (a p38 MAPK inhibitor), which suggests that SHC1 is upstream of AMPK and p38 MAPK in the 8-Cl-cAMP-stimulated signaling cascade. On the basis of these findings, we conclude that SHC1 functions as a sensor during the 8-Cl-cAMP-induced growth inhibition in SHC1-overexpressing cancer cells.

Iterative image reconstruction techniques for CT coronary artery calcium quantification: comparison with traditional filtered back projection in vitro and in vivo.

PURPOSE: To investigate in vitro and in vivo the use of image-based and raw data-based iterative reconstruction algorithms for quantification of coronary artery calcium by using the Agatston score and subsequent cardiac risk stratification. MATERIALS AND METHODS: In vitro data were obtained by using a moving anthropomorphic cardiac phantom containing calcium inserts of different concentrations and sizes. With institutional review board approval and HIPAA compliance, coronary calcium imaging data of 110 consecutive patients (mean age ± standard deviation, 58.2 years ± 9.8; 48 men) were reconstructed with filtered back projection (FBP), iterative reconstruction in image space (IRIS), and sinogram-affirmed iterative reconstruction (SAFIRE). Image noise was measured and the Agatston score was obtained for all reconstructions. Assignment to Agatston scores and percentile-based cardiac risk categories was compared. Statistical analysis included the Cohen κ coefficient and Friedman and Wilcoxon testing. RESULTS: In vitro, mean Agatston scores ± standard deviation for FBP (638.9 ± 9.6), IRIS (622.7 ± 15.2), and SAFIRE (631.4 ± 17.6) were comparable (P = .30). The smallest phantom calcifications were more frequently detected when iterative reconstruction techniques were used. The Agatston scores in the patient cohort were not significantly different among FBP, IRIS, and SAFIRE in paired comparisons (median Agatston score [25th and 75th percentiles]: 76.0 [20.6, 243.9], 76.4 [22, 249.3], and 75.7 [21.5, 49.1], respectively; P = .20 each). Comparison of categorization based on Agatston score percentiles showed excellent agreement for both IRIS and SAFIRE with FBP (κ = 0.975 [0.942-1.00] and κ = 0.963 [0.922-1.00], respectively). The mean effective dose was 1.02 mSv ± 0.51. Mean image noise was significantly (P < .001) higher with FBP than that with iterative reconstructions. CONCLUSION: In comparison with FBP, iterative reconstruction techniques do not have a profound effect on the reproducible quantification of coronary calcium according to Agatston score and subsequent cardiac risk classification, although risk reclassification may occur in a small subset of subjects.

Dynamic CT myocardial perfusion imaging: performance of 3D semi-automated evaluation software.

OBJECTIVES: To evaluate the performance of three-dimensional semi-automated evaluation software for the assessment of myocardial blood flow (MBF) and blood volume (MBV) at dynamic myocardial perfusion computed tomography (CT). METHODS: Volume-based software relying on marginal space learning and probabilistic boosting tree-based contour fitting was applied to CT myocardial perfusion imaging data of 37 subjects. In addition, all image data were analysed manually and both approaches were compared with SPECT findings. Study endpoints included time of analysis and conventional measures of diagnostic accuracy. RESULTS: Of 592 analysable segments, 42 showed perfusion defects on SPECT. Average analysis times for the manual and software-based approaches were 49.1 ± 11.2 and 16.5 ± 3.7 min respectively (P < 0.01). There was strong agreement between the two measures of interest (MBF, ICC = 0.91, and MBV, ICC = 0.88, both P < 0.01) and no significant difference in MBF/MBV with respect to diagnostic accuracy between the two approaches for both MBF and MBV for manual versus software-based approach; respectively; all comparisons P > 0.05. CONCLUSIONS: Three-dimensional semi-automated evaluation of dynamic myocardial perfusion CT data provides similar measures and diagnostic accuracy to manual evaluation, albeit with substantially reduced analysis times. This capability may aid the integration of this test into clinical workflows. KEY POINTS: • Myocardial perfusion CT is attractive for comprehensive coronary heart disease assessment. • Traditional image analysis methods are cumbersome and time-consuming. • Automated 3D perfusion software shortens analysis times. • Automated 3D perfusion software increases standardisation of myocardial perfusion CT. • Automated, standardised analysis fosters myocardial perfusion CT integration into clinical practice.

p104 binds to Rac1 and reduces its activity during myotube differentiation of C2C12 cell.

The p104 protein inhibits cellular proliferation when overexpressed in NIH3T3 cells and has been shown to associate with p85α, Grb2, and PLCγ1. In order to isolate other proteins that interact with p104, yeast two-hybrid screening was performed. Rac1 was identified as a binding partner of p104 and the interaction between p104 and Rac1 was confirmed by immunoprecipitation. Using a glutathione S-transferase (GST) pull-down assay with various p104 fragments, the 814-848 amino acid residue at the carboxyl-terminal region of p104 was identified as the key component to interact with Rac1. The CrkII which is involved in the Rac1-mediated cellular response was also found to interact with p104 protein. NIH3T3 cells which overexpressed p104 showed a decrease of Rac1 activity. However, neither the proline-rich domain mutant, which is unable to interact with CrkII, nor the carboxy-terminal deletion mutant could attenuate Rac1 activity. During the differentiation of myoblasts, the amount of p104 protein as well as transcript level was increased. The overexpression of p104 enhanced myotube differentiation, whereas siRNA of p104 reversed this process. In this process, more Rac1 and CrkII were bound to increased p104. Based on these results, we conclude that p104 is involved in muscle cell differentiation by modulating the Rac1 activity.

Involvement of Akt2/protein kinase B ß (PKBß) in the 8-Cl-cAMP-induced cancer cell growth inhibition.

8-chloro-cyclic AMP (8-Cl-cAMP), which induces differentiation, growth inhibition, and apoptosis in various cancer cells, has been investigated as a putative anti-cancer drug. However, the exact mechanism of 8-Cl-cAMP functioning in cancer cells is not fully understood. Akt/protein kinase B (PKB) genes (Akt1, Akt2, and Akt3) encode enzymes belonging to the serine/threonine-specific protein kinase family. It has been suggested that Akt/PKB enhances cell survival by inhibiting apoptosis. Recently, we showed that 8-Cl-cAMP and 5-aminoimidazole-4-carboxamide ribonucleoside (AICAR) inhibited cancer cell growth through the activation of AMPK and p38 MAPK. Therefore, we anticipated that the phosphorylation of Akt/PKB would be decreased upon treatment with 8-Cl-cAMP. However, treatment with 8-Cl-cAMP and AICAR induced the phosphorylation of Akt/PKB, which was inhibited by ABT702 (an adenosine kinase inhibitor) and NBTI (an adenosine transporter inhibitor). Furthermore, whereas Compound C (an AMPK inhibitor), AMPK-DN (AMPK-dominant negative) mutant, and SB203580 (a p38 MAPK inhibitor) did not block the 8-Cl-cAMP-induced phosphorylation of Akt/PKB, TCN (an Akt1/2/3 specific inhibitor) and an Akt2/PKBß-targeted siRNA inhibited the 8-Cl-cAMP- and AICAR-mediated phosphorylation of AMPK and p38 MAPK. TCN also reversed the growth inhibition mediated by 8-Cl-cAMP and AICAR. Moreover, an Akt1/PKBα-targeted siRNA did not reduce the phosphorylation of AMPK and p38 MAPK after treatment with 8-Cl-cAMP. These results suggest that Akt2/PKBß activation promotes the phosphorylation of AMPK and p38 MAPK during the 8-Cl-cAMP- and AICAR-induced growth inhibition.

Cardiovascular CT angiography in neonates and children: image quality and potential for radiation dose reduction with iterative image reconstruction techniques.

OBJECTIVES: To evaluate image quality (IQ) of low-radiation-dose paediatric cardiovascular CT angiography (CTA), comparing iterative reconstruction in image space (IRIS) and sinogram-affirmed iterative reconstruction (SAFIRE) with filtered back-projection (FBP) and estimate the potential for further dose reductions. METHODS: Forty neonates and children underwent low radiation CTA with or without ECG synchronisation. Data were reconstructed with FBP, IRIS and SAFIRE. For ECG-synchronised studies, half-dose image acquisitions were simulated. Signal noise was measured and IQ graded. Effective dose (ED) was estimated. RESULTS: Mean absolute and relative image noise with IRIS and full-dose SAFIRE was lower than with FBP (P < 0.001), while SNR and CNR were higher (P < 0.001). Image noise was also lower and SNR and CNR higher in half-dose SAFIRE studies compared with full-and half-dose FBP studies (P < 0.001). IQ scores were higher for IRIS, full-dose SAFIRE and half-dose SAFIRE than for full-dose FBP and higher for half-dose SAFIRE than for half-dose FBP (P < 0.05). Median weight-specific ED was 0.3 mSv without and 1.36 mSv with ECG synchronisation. The estimated ED of half-dose SAFIRE studies was 0.68 mSv. CONCLUSIONS: IR improves image noise, SNR, CNR and subjective IQ compared with FBP in low-radiation-dose paediatric CTA and allows further dose reductions without compromising diagnostic IQ. KEY POINTS: • Iterative reconstruction techniques significantly improve non-invasive cardiovascular CT in children. • Using half traditional radiation dose image quality is higher with iterative reconstruction. • Iterative reconstruction techniques may allow further radiation reductions in paediatric cardiovascular CT.

Artificial Intelligence Solution for Chest Radiographs in Respiratory Outpatient Clinics: Multicenter Prospective Randomized Clinical Trial.

Rationale: Artificial intelligence (AI)-assisted diagnosis imparts high accuracy to chest radiography (CXR) interpretation; however, its benefit for nonradiologist physicians in detecting lung lesions on CXR remains unclear. Objectives: To investigate whether AI assistance improves the diagnostic performance of physicians for CXR interpretation and affects their clinical decisions in clinical practice. Methods: We randomly allocated eligible patients who visited an outpatient clinic to the intervention (with AI-assisted interpretation) and control (without AI-assisted interpretation) groups. Lung lesions on CXR were recorded by seven nonradiologists with or without AI assistance. The reference standard for lung lesions was established by three radiologists. The primary and secondary endpoints were the physicians' diagnostic accuracy and clinical decision, respectively. Results: Between October 2020 and May 2021, 162 and 161 patients were assigned to the intervention and control groups, respectively. The area under the receiver operating characteristic curve was significantly larger in the intervention group than in the control group for the CXR level (0.840 [95% confidence interval (CI), 0.778-0.903] vs. 0.718 [95% CI, 0.640-0.796]; P = 0.017) and lung lesion level (0.800 [95% CI, 0.740-0.861] vs. 0.677 [95% CI, 0.605-0.750]; P = 0.011). The intervention group had higher sensitivity in terms of both CXR and lung lesion level and a lower false referral rate for the lung lesion level. AI-assisted CXR interpretation did not affect the physicians' clinical decisions. Conclusions: AI-assisted CXR interpretation improves the diagnostic performance of nonradiologist physicians in detecting abnormal lung findings. Clinical trial registered with Clinical Research Information Service of the Republic of Korea (KCT 0005466).

The effectiveness of erythromycin in reducing stent-related tissue hyperplasia: an experimental study with a rat esophageal model.

BACKGROUND: Erythromycin is not only a potent antibiotic; it also has effects of reduction of inflammation and suppression of protein synthesis. PURPOSE: To evaluate the impact of erythromycin on tissue hyperplasia after stent placement in a rat esophageal model. MATERIAL AND METHODS: A total of 21 rats were included. After placement of self-expanding stents in the mid esophagus, the rats were divided into two experimental groups and one control group. The rats in the experimental groups received daily intraperitoneal injections of erythromycin for 5 weeks; 4 mg/kg (group A, n = 7) and 8 mg/kg (group B, n = 7). Those in the control group (n = 7) received 1 mL of saline intraperitoneally. After sacrifice, histologic analysis was done for thickness of the papillary projection, granulation tissue area, percentage of granulation tissue area, and degree of inflammatory cell infiltration. The statistical significance of differences between groups was assessed by Mann-Whitney U test. RESULTS: Tissue hyperplasia as reflected in thickness of papillary projection, granulation tissue area, and percentage of granulation tissue area, was higher in the control group than in the experimental groups, although there was no statistical significance (P = 1.00, 0.332, and 0.263, respectively). However, degree of inflammatory cell infiltration was significantly lower in the experimental groups than the control group (P = 0.025), and the higher dosage of erythromycin reduced inflammatory cell infiltration significantly (P = 0.037). CONCLUSION: Intraperitoneal administration of erythromycin is very effective in reducing inflammation after stent placement in a rat esophageal model but has no significant effect on granulation tissue formation.

Concordance rate of radiologists and a commercialized deep-learning solution for chest X-ray: Real-world experience with a multicenter health screening cohort.

PURPOSE: Lunit INSIGHT CXR (Lunit) is a commercially available deep-learning algorithm-based decision support system for chest radiography (CXR). This retrospective study aimed to evaluate the concordance rate of radiologists and Lunit for thoracic abnormalities in a multicenter health screening cohort. METHODS AND MATERIALS: We retrospectively evaluated the radiology reports and Lunit results for CXR at several health screening centers in August 2020. Lunit was adopted as a clinical decision support system (CDSS) in routine clinical practice. Subsequently, radiologists completed their reports after reviewing the Lunit results. The DLA result was provided as a color map with an abnormality score (%) for thoracic lesions when the score was greater than the predefined cutoff value of 15%. Concordance was achieved when (a) the radiology reports were consistent with the DLA results ("accept"), (b) the radiology reports were partially consistent with the DLA results ("edit") or had additional lesions compared with the DLA results ("add"). There was discordance when the DLA results were rejected in the radiology report. In addition, we compared the reading times before and after Lunit was introduced. Finally, we evaluated systemic usability scale questionnaire for radiologists and physicians who had experienced Lunit. RESULTS: Among 3,113 participants (1,157 men; mean age, 49 years), thoracic abnormalities were found in 343 (11.0%) based on the CXR radiology reports and 621 (20.1%) based on the Lunit results. The concordance rate was 86.8% (accept: 85.3%, edit: 0.9%, and add: 0.6%), and the discordance rate was 13.2%. Except for 479 cases (7.5%) for whom reading time data were unavailable (n = 5) or unreliable (n = 474), the median reading time increased after the clinical integration of Lunit (median, 19s vs. 14s, P < 0.001). CONCLUSION: The real-world multicenter health screening cohort showed a high concordance of the chest X-ray report and the Lunit result under the clinical integration of the deep-learning solution. The reading time slight increased with the Lunit assistance.

Artificial Intelligence-Based Identification of Normal Chest Radiographs: A Simulation Study in a Multicenter Health Screening Cohort.

OBJECTIVE: This study aimed to investigate the feasibility of using artificial intelligence (AI) to identify normal chest radiography (CXR) from the worklist of radiologists in a health-screening environment. MATERIALS AND METHODS: This retrospective simulation study was conducted using the CXRs of 5887 adults (mean age ± standard deviation, 55.4 ± 11.8 years; male, 4329) from three health screening centers in South Korea using a commercial AI (Lunit INSIGHT CXR3, version 3.5.8.8). Three board-certified thoracic radiologists reviewed CXR images for referable thoracic abnormalities and grouped the images into those with visible referable abnormalities (identified as abnormal by at least one reader) and those with clearly visible referable abnormalities (identified as abnormal by at least two readers). With AI-based simulated exclusion of normal CXR images, the percentages of normal images sorted and abnormal images erroneously removed were analyzed. Additionally, in a random subsample of 480 patients, the ability to identify visible referable abnormalities was compared among AI-unassisted reading (i.e., all images read by human readers without AI), AI-assisted reading (i.e., all images read by human readers with AI assistance as concurrent readers), and reading with AI triage (i.e., human reading of only those rendered abnormal by AI). RESULTS: Of 5887 CXR images, 405 (6.9%) and 227 (3.9%) contained visible and clearly visible abnormalities, respectively. With AI-based triage, 42.9% (2354/5482) of normal CXR images were removed at the cost of erroneous removal of 3.5% (14/405) and 1.8% (4/227) of CXR images with visible and clearly visible abnormalities, respectively. In the diagnostic performance study, AI triage removed 41.6% (188/452) of normal images from the worklist without missing visible abnormalities and increased the specificity for some readers without decreasing sensitivity. CONCLUSION: This study suggests the feasibility of sorting and removing normal CXRs using AI with a tailored cut-off to increase efficiency and reduce the workload of radiologists.

Cervical epidural hematoma with Brown-Sequard syndrome caused by an epidural injection: a case report.

Epidural hematoma with Brown-Sequard syndrome caused by an epidural injection is a rarely found condition in the emergency department (ED). We report an unusual case of Brown-Sequard syndrome in a 55-year-old man who presented at the ED with right-sided weakness and contralateral loss of pain and temperature sensation after a cervical epidural injection for shoulder pain. Cervicla spine magnetic resonance imaging showed an epidural hematoma from C4 to C6. After admission, his right hemiparesis and contralateral sensory loss improved within eight days, and surgical decompression was not required. Diagnosing spinal lesions in the ED is challenging, especially in patients with acute neurological signs requiring immediate evaluation for stroke. In this case, definite hemiparesis and some contralateral sensory loss were noted. Therefore, a potential spinal lesion was suspected rather than a stroke. This case emphasized the importance of conducting a focused neurological examination after history taking.

Performance of a deep-learning algorithm for referable thoracic abnormalities on chest radiographs: A multicenter study of a health screening cohort.

PURPOSE: This study evaluated the performance of a commercially available deep-learning algorithm (DLA) (Insight CXR, Lunit, Seoul, South Korea) for referable thoracic abnormalities on chest X-ray (CXR) using a consecutively collected multicenter health screening cohort. METHODS AND MATERIALS: A consecutive health screening cohort of participants who underwent both CXR and chest computed tomography (CT) within 1 month was retrospectively collected from three institutions' health care clinics (n = 5,887). Referable thoracic abnormalities were defined as any radiologic findings requiring further diagnostic evaluation or management, including DLA-target lesions of nodule/mass, consolidation, or pneumothorax. We evaluated the diagnostic performance of the DLA for referable thoracic abnormalities using the area under the receiver operating characteristic (ROC) curve (AUC), sensitivity, and specificity using ground truth based on chest CT (CT-GT). In addition, for CT-GT-positive cases, three independent radiologist readings were performed on CXR and clear visible (when more than two radiologists called) and visible (at least one radiologist called) abnormalities were defined as CXR-GTs (clear visible CXR-GT and visible CXR-GT, respectively) to evaluate the performance of the DLA. RESULTS: Among 5,887 subjects (4,329 males; mean age 54±11 years), referable thoracic abnormalities were found in 618 (10.5%) based on CT-GT. DLA-target lesions were observed in 223 (4.0%), nodule/mass in 202 (3.4%), consolidation in 31 (0.5%), pneumothorax in one 1 (<0.1%), and DLA-non-target lesions in 409 (6.9%). For referable thoracic abnormalities based on CT-GT, the DLA showed an AUC of 0.771 (95% confidence interval [CI], 0.751-0.791), a sensitivity of 69.6%, and a specificity of 74.0%. Based on CXR-GT, the prevalence of referable thoracic abnormalities decreased, with visible and clear visible abnormalities found in 405 (6.9%) and 227 (3.9%) cases, respectively. The performance of the DLA increased significantly when using CXR-GTs, with an AUC of 0.839 (95% CI, 0.829-0.848), a sensitivity of 82.7%, and s specificity of 73.2% based on visible CXR-GT and an AUC of 0.872 (95% CI, 0.863-0.880, P <0.001 for the AUC comparison of GT-CT vs. clear visible CXR-GT), a sensitivity of 83.3%, and a specificity of 78.8% based on clear visible CXR-GT. CONCLUSION: The DLA provided fair-to-good stand-alone performance for the detection of referable thoracic abnormalities in a multicenter consecutive health screening cohort. The DLA showed varied performance according to the different methods of ground truth.

Correction: Performance of a deep-learning algorithm for referable thoracic abnormalities on chest radiographs: A multicenter study of a health screening cohort.

[This corrects the article DOI: 10.1371/journal.pone.0246472.].

Comparison of cytological adequacy and pain scale score in ultrasound-guided fine-needle aspiration of solid thyroid nodules for liquid-based cytology with with 23- and 25-gauge needles: a single-center prospective study.

In ultrasound (US)-guided fine-needle aspiration (FNA) of solid thyroid nodules (STN) using liquid-based cytology (LBC), the most appropriate needle size for LBC remains unclear. This study compared the cytological adequacy and complications associated with using 23- and 25-gauge needles in US-guided FNA of STNs using LBC. US-guided FNA was performed in consecutive patients by one radiologist to diagnose STNs ≥ 5 mm in the largest diameter. The one-sampling technique through a single needle puncture and multiple to-and-fro needle motions was used in each patient. The 23- and 25-guage needles were used consecutively each day. After FNA, the pain and complications experienced by each patient were investigated by a nurse, who was blinded to the information of needle gauge used. A cytopathologist retrospectively analyzed the cytological adequacy and cellularity of the cases. Of the 99 STNs, eight (8.1%) exhibited inadequate cytology (4 each with 23- and 25-gauge needles). The rate of cytological adequacy was not statistically different between the groups (p = 0.631). The mean pain scale values with 23- and 25-gauge needles were 2.1 ± 1.3 and 1.6 ± 1.3, respectively (p = 0.135). There were no significant complications in either group. In conclusion, both 23- and 25-gauge needles are useful in LBC because cytological adequacy and complications were not statistically different with both sizes of the needles.