Using AI to Identify Unremarkable Chest Radiographs for Automatic Reporting.

Background Radiology practices have a high volume of unremarkable chest radiographs and artificial intelligence (AI) could possibly improve workflow by providing an automatic report. Purpose To estimate the proportion of unremarkable chest radiographs, where AI can correctly exclude pathology (ie, specificity) without increasing diagnostic errors. Materials and Methods In this retrospective study, consecutive chest radiographs in unique adult patients (≥18 years of age) were obtained January 1-12, 2020, at four Danish hospitals. Exclusion criteria included insufficient radiology reports or AI output error. Two thoracic radiologists, who were blinded to AI output, labeled chest radiographs as "remarkable" or "unremarkable" based on predefined unremarkable findings (reference standard). Radiology reports were classified similarly. A commercial AI tool was adapted to output a chest radiograph "remarkableness" probability, which was used to calculate specificity at different AI sensitivities. Chest radiographs with missed findings by AI and/or the radiology report were graded by one thoracic radiologist as critical, clinically significant, or clinically insignificant. Paired proportions were compared using the McNemar test. Results A total of 1961 patients were included (median age, 72 years [IQR, 58-81 years]; 993 female), with one chest radiograph per patient. The reference standard labeled 1231 of 1961 chest radiographs (62.8%) as remarkable and 730 of 1961 (37.2%) as unremarkable. At 99.9%, 99.0%, and 98.0% sensitivity, the AI had a specificity of 24.5% (179 of 730 radiographs [95% CI: 21, 28]), 47.1% (344 of 730 radiographs [95% CI: 43, 51]), and 52.7% (385 of 730 radiographs [95% CI: 49, 56]), respectively. With the AI fixed to have a similar sensitivity as radiology reports (87.2%), the missed findings of AI and reports had 2.2% (27 of 1231 radiographs) and 1.1% (14 of 1231 radiographs) classified as critical (P = .01), 4.1% (51 of 1231 radiographs) and 3.6% (44 of 1231 radiographs) classified as clinically significant (P = .46), and 6.5% (80 of 1231) and 8.1% (100 of 1231) classified as clinically insignificant (P = .11), respectively. At sensitivities greater than or equal to 95.4%, the AI tool exhibited less than or equal to 1.1% critical misses. Conclusion A commercial AI tool used off-label could correctly exclude pathology in 24.5%-52.7% of all unremarkable chest radiographs at greater than or equal to 98% sensitivity. The AI had equal or lower rates of critical misses than radiology reports at sensitivities greater than or equal to 95.4%. These results should be confirmed in a prospective study. © RSNA, 2024 Supplemental material is available for this article. See also the editorial by Yoon and Hwang in this issue.

Autonomous Chest Radiograph Reporting Using AI: Estimation of Clinical Impact.

Background Automated interpretation of normal chest radiographs could alleviate the workload of radiologists. However, the performance of such an artificial intelligence (AI) tool compared with clinical radiology reports has not been established. Purpose To perform an external evaluation of a commercially available AI tool for (a) the number of chest radiographs autonomously reported, (b) the sensitivity for AI detection of abnormal chest radiographs, and (c) the performance of AI compared with that of the clinical radiology reports. Materials and Methods In this retrospective study, consecutive posteroanterior chest radiographs from adult patients in four hospitals in the capital region of Denmark were obtained in January 2020, including images from emergency department patients, in-hospital patients, and outpatients. Three thoracic radiologists labeled chest radiographs in a reference standard based on chest radiograph findings into the following categories: critical, other remarkable, unremarkable, or normal (no abnormalities). AI classified chest radiographs as high confidence normal (normal) or not high confidence normal (abnormal). Results A total of 1529 patients were included for analysis (median age, 69 years [IQR, 55-69 years]; 776 women), with 1100 (72%) classified by the reference standard as having abnormal radiographs, 617 (40%) as having critical abnormal radiographs, and 429 (28%) as having normal radiographs. For comparison, clinical radiology reports were classified based on the text and insufficient reports excluded (n = 22). The sensitivity of AI was 99.1% (95% CI: 98.3, 99.6; 1090 of 1100 patients) for abnormal radiographs and 99.8% (95% CI: 99.1, 99.9; 616 of 617 patients) for critical radiographs. Corresponding sensitivities for radiologist reports were 72.3% (95% CI: 69.5, 74.9; 779 of 1078 patients) and 93.5% (95% CI: 91.2, 95.3; 558 of 597 patients), respectively. Specificity of AI, and hence the potential autonomous reporting rate, was 28.0% of all normal posteroanterior chest radiographs (95% CI: 23.8, 32.5; 120 of 429 patients), or 7.8% (120 of 1529 patients) of all posteroanterior chest radiographs. Conclusion Of all normal posteroanterior chest radiographs, 28% were autonomously reported by AI with a sensitivity for any abnormalities higher than 99%. This corresponded to 7.8% of the entire posteroanterior chest radiograph production. © RSNA, 2023 Supplemental material is available for this article. See also the editorial by Park in this issue.

Diagnostic accuracy of ultra-low-dose chest computed tomography in an emergency department.

BACKGROUND: This study examined whether ultra-low-dose chest computed tomography (ULD-CT) could improve detection of acute chest conditions. PURPOSE: To determine (i) whether diagnostic accuracy of ULD-CT is superior to supine chest X-ray (sCXR) for acute chest conditions and (ii) the feasibility of ULD-CT in an emergency department. MATERIAL AND METHODS: From 1 February to 31 July 2019, 91 non-traumatic patients from the Emergency Department were prospectively enrolled in the study if they received an sCXR. An ULD-CT and a non-contrast chest CT (NCCT) scan were then performed. Three radiologists assessed the sCXR and ULD-CT examinations for cardiogenic pulmonary edema, pneumonia, pneumothorax, and pleural effusion. Resources and effort were compared for sCXR and ULD-CT to evaluate feasibility. Diagnostic accuracy was calculated for sCXR and ULD-CT using NCCT as the reference standard. RESULTS: The mean effective dose of ULD-CT was 0.05±0.01 mSv. For pleural effusion and cardiogenic pulmonary edema, no difference in diagnostic accuracy between ULD-CT and sCXR was observed. For pneumonia and pneumothorax, sensitivities were 100% (95% confidence interval [CI] 69-100) and 50% (95% CI 7-93) for ULD-CT and 60% (95% CI 26-88) and 0% (95% CI 0-0) for sCXR, respectively. Median examination time was 10 min for ULD-CT vs. 5 min for sCXR (P<0.001). For ULD-CT 1-2 more staff members were needed compared to sCXR (P<0.001). ULD-CT was rated more challenging to perform than sCXR (P<0.001). CONCLUSION: ULD-CT seems equal or better in detecting acute chest conditions compared to sCXR. However, ULD-CT examinations demand more effort and resources.

Imaging the pancreas with photon-counting CT - A review of normal pancreatic anatomy.

PURPOSE: Compared to conventional energy integrating detector CT, Photon-Counting CT (PCCT) has the advantage of increased spatial resolution. The pancreas is a highly complex organ anatomically. The increased spatial resolution of PCCT challenges radiologists' knowledge of pancreatic anatomy. The purpose of this review was to review detailed macroscopic and microscopic anatomy of the pancreas in the context of current and future PCCT. METHOD: This review is based on a literature review of all parts of pancreatic anatomy and a retrospective imaging review of PCCT scans from 20 consecutively included patients without pancreatic pathology (mean age 61.8 years, 11 female), scanned in the workup of pancreatic cancer with a contrast enhanced multiphase protocol. Two radiologists assessed the visibility of the main and accessory pancreatic ducts, side ducts, ampulla, major papilla, minor papilla, pancreatic arteries and veins, regional lymph nodes, coeliac ganglia, and coeliac plexus. RESULTS: The macroscopic anatomy of the pancreas was consistently visualized with PCCT. Visualization of detailed anatomy of the ductal system (including side ducts), papillae, arteries, vein, lymph nodes, and innervation was possible in 90% or more of patients with moderate to good interreader agreement. CONCLUSION: PCCT scans of the pancreas visualizes previously unseen or inconsistently seen small anatomical structures consistently. Increased knowledge of pancreatic anatomy could have importance in imaging of pancreatic cancer and other pancreatic diseases.

The asbestos-asbestosis exposure-response relationship: a cohort study of the general working population.

OBJECTIVES: The association between asbestos exposure and asbestosis in high-exposed industrial cohorts is well-known, but there is a lack of knowledge about the exposure-response relationship for asbestosis in a general working population setting. We examined the exposure-response relationship between occupational asbestos exposure and asbestosis in asbestos-exposed workers of the Danish general working population. METHODS: We followed all asbestos-exposed workers from 1979 to 2015 and identified incident cases of asbestosis using the Danish National Patient Register. Individual asbestos exposure was estimated with a quantitative job exposure matrix (SYN-JEM) from 1976 onwards and back-extrapolated to age 16 for those exposed in 1976. Exposure-response relations for cumulative exposure and other exposure metrics were analyzed using a discrete time hazard model and adjusted for potential confounders. RESULTS: The range of cumulative exposure in the population was 0.001 to 18 fibers per milliliter-year (f/ml-year). We found increasing incidence rate ratios (IRR) of asbestosis with increasing cumulative asbestos exposure with a fully adjusted IRR per 1 f/ml-years of 1.18 [95% confidence interval (CI) 1.15- -1.22]. The IRR was 1.94 (95% CI 1.53-2.47) in the highest compared to the lowest exposure tertile. We similarly observed increasing risk with increasing cumulative exposure in the inception population. CONCLUSIONS: This study found exposure-response relations between cumulative asbestos exposure and incident asbestosis in the Danish general working population with mainly low-level exposed occupations, but there is some uncertainty regarding the exposure levels.

Ultra-low-dose emergency chest computed tomography protocols in three vendors: A technical note.

Background: In suspected community-acquired pneumonia (CAP), chest CT is superior to the routinely obtained radiographs (CXR), but administers higher radiation doses. However, ultra-low-dose CT (ULDCT) has shown promising results. Purpose: To compare radiation dose and image quality using standard and ULDCT protocols designed for a multicenter study encompassing three CT scanner models from GE, Canon, and Siemens. Material and methods: Patients with suspected CAP were referred for non-contrast standard dose chest CT (NCCT) and ULDCT. Effective radiation dose and Contrast-to-Noise Ratio (CNR) was calculated. Results: Mean effective doses were GE (n = 10) 6.93 mSv in NCCT and 0.27 mSv in ULDCT; Canon (n = 9) 3.48 in mSv NCCT and 1.11 mSv in ULDCT; Siemens (n = 10) 2.85 mSv in NCCT and 0.45 mSv in ULDCT. CNR was reduced by 29-39% in ULDCT. Conclusion: The proposed CT protocols yielded dose reductions of 96%, 68%, and 84% using a GE, Canon, and Siemens scanner, respectively.

Diagnostic test accuracy study of a commercially available deep learning algorithm for ischemic lesion detection on brain MRIs in suspected stroke patients from a non-comprehensive stroke center.

PURPOSE: To estimate the ability of a commercially available artificial intelligence (AI) tool to detect acute brain ischemia on Magnetic Resonance Imaging (MRI), compared to an experienced neuroradiologist. METHODS: We retrospectively included 1030 patients with brain MRI, suspected of stroke from January 6th, 2020 to 1st of April 2022, based on these criteria: Age ≥ 18 years, symptoms within four weeks before the scan. The neuroradiologist reinterpreted the MRI scans and subclassified ischemic lesions for reference. We excluded scans with interpretation difficulties due to artifacts or missing sequences. Four MRI scanner models from the same vendor were used. The first 800 patients were included consecutively, remaining enriched for less frequent lesions. The index test was a CE-approved AI tool (Apollo version 2.1.1 by Cerebriu). RESULTS: The final analysis cohort comprised 995 patients (mean age 69 years, 53 % female). A case-based analysis for detecting acute ischemic lesions showed a sensitivity of 89 % (95 % CI: 85 %-91 %) and specificity of 90 % (95 % CI: 87 %-92 %). We found no significant difference in sensitivity or specificity based on sex, age, or comorbidities. Specificity was reduced in cases with DWI artifacts. Multivariate analysis showed that increasing ischemic lesion size and fragmented lesions were independently associated with higher sensitivity, while non-acute lesion ages lowered sensitivity. CONCLUSIONS: The AI tool exhibits high sensitivity and specificity in detecting acute ischemic lesions on MRI compared to an experienced neuroradiologist. While sensitivity depends on the ischemic lesions' characteristics, specificity depends on the image quality.

Ultra-low dose computed tomography of the chest in an emergency setting: A prospective agreement study.

Ultra-low dose computed tomography (ULD-CT) assessed by non-radiologists in a medical Emergency Department (ED) has not been examined in previous studies. To (i) investigate intragroup agreement among attending physicians caring for ED patients (i.e., radiologists, senior- and junior clinicians) and medical students for the detection of acute lung conditions on ULD-CT and supine chest X-ray (sCXR), and (ii) evaluate the accuracy of interpretation compared to the reference standard. In this prospective study, non-traumatic patients presenting to the ED, who received an sCXR were included. Between February and July 2019, 91 patients who underwent 93 consecutive examinations were enrolled. Subsequently, a ULD-CT and non-contrast CT were performed. The ULD-CT and sCXR were assessed by 3 radiologists, 3 senior clinicians, 3 junior clinicians, and 3 medical students for pneumonia, pneumothorax, pleural effusion, and pulmonary edema. The non-contrast CT, assessed by a chest radiologist, was used as the reference standard. The results of the assessments were compared within each group (intragroup agreement) and with the reference standard (accuracy) using kappa statistics. Accuracy and intragroup agreement improved for pneumothorax on ULD-CT compared with the sCXR for all groups. Accuracy and intragroup agreement improved for pneumonia on ULD-CT when assessed by radiologists and for pleural effusion when assessed by medical students. In patients with acute lung conditions ULD-CT offers improvement in the detection of pneumonia by radiologists and the detection of pneumothorax by radiologists as well as non-radiologists compared to sCXR. Therefore, ULD-CT may be considered as an alternative first-line imaging modality to sCXR for non-traumatic patients who present to EDs.

Impaired exercise-induced hypoalgesia in individuals reporting an increase in low back pain during acute exercise.

OBJECTIVES: Exercise therapy is recommended for low back pain (LBP) although the immediate effects on pain are highly variable. In 96 individuals with LBP this cross-sectional study explored (a) the magnitude of exercise-induced hypoalgesia (EIH) and (b) measures of pain sensitivity and clinical pain manifestations in individuals reporting a clinical relevant increase in back pain during physical activity compared with individuals reporting low or no increase in back pain during physical activity. METHODS: Cuff algometry was performed at baseline on the leg to assess pressure pain threshold (cPPT), tolerance (cPTT) and temporal summation of pain (cTSP). Manual PPTs were assessed on the back and leg before and after a 6-min walk test (6MWT). Back pain was scored on a numerical rating scale (NRS) after each minute of walking. The EIH-effect was estimated as the increase in PPTs after the walk exercise. RESULTS: Twenty-seven individuals reported an increase of ≥2/10 in pain NRS scores during walking and compared with the individuals with <2/10 NRS scores: cPPT and EIH-effects were lower whereas cTSP, pain intensity and disability were increased (p < 0.03). Baseline NRS scores, EIH and pain thresholds were associated with the likelihood of an increase of ≥2/10 in back pain intensity during walking (p < 0.05). CONCLUSIONS: Pain flares in response to physical activity in individuals with LBP seem to be linked with baseline pain sensitivity and pain intensity, and impair the beneficial EIH. Such information may better inform when individuals with LBP will have a beneficial effect of physical activity. SIGNIFICANCE: Pain flares in response to physical activity in individuals with LBP seem to be linked with baseline pain sensitivity and pain intensity, and impair the beneficial exercise-induced hypoalgesia. Such information may better inform when individuals with LBP will have a beneficial effect of physical activity.

Automated 2D peptide separation on a 1D nano-LC-MS system.

Given the complexity of the mammalian proteome, high-resolution separation technologies are required to achieve comprehensive proteome coverage and to enhance the detection of low-abundance proteins. Among several technologies, Multidimensional Protein Identification Technology (MudPIT) enables the on-line separation of highly complex peptide mixtures directly coupled with mass spectrometry-based identification. Here, we present a variation of the traditional MudPIT protocol, combining highly sensitive chromatography using a nanoflow liquid chromatography system (nano-LC) with a two-dimensional precolumn in a vented column setup. When compared to the traditional MudPIT approach, this nanoflow variation demonstrated better first-phase separation leading to more proteins being characterized while using rather simple instrumentation and a protocol that requires less time and very little technical expertise to perform.

Activation of the CXCR3 chemokine receptor through anchoring of a small molecule chelator ligand between TM-III, -IV, and -VI.

Seven transmembrane segment (7TM) receptors are activated through a common, still rather unclear molecular mechanism by a variety of chemical messengers ranging from monoamines to large proteins. By introducing a His residue at position III:05 in the CXCR3 receptor a metal ion site was built between the extracellular ends of transmembrane (TM) III and TM-IV to anchor aromatic chelators at a location corresponding to the presumed binding pocket for adrenergic receptor agonists. In this construct, free metal ions had no agonistic effect in accordance with the optimal geometry of the metal ion site in molecular models built over the inactive form of rhodopsin. In contrast, the aromatic chelators bipyridine or phenanthrolene in complex with Zn(II) or Cu(II) acted as potent agonists displaying signaling efficacies similar to or even better than the endogenous chemokine agonists. Molecular modeling and molecular simulations combined with mutational analysis indicated that the metal ion site-anchored chelators act as agonists by establishing an aromatic-aromatic, second-site interaction with TyrVI:16 on the inner face of TM-VI. It is noteworthy that this interaction required that the extracellular segment of TM-VI moves inward in the direction of TM-III, whereby TyrVI:16 together with the chelators complete an "aromatic zipper" also comprising PheIII:08 (corresponding to the monoamine receptor anchoring point) and TyrVII:10 (corresponding to the retinal attachment site in rhodopsin). Chemokine agonism was independent of this aromatic zipper. It is proposed that in rhodopsin-like 7TM receptors, small-molecule compounds in general act as agonists in a similar manner as here demonstrated with the artificial, metal ion site anchored chelators, by holding TM-VI bent inward.