Reader Agreement Research Articles

Artificial intelligence model for enhancing the accuracy of transvaginal ultrasound in detecting endometrial cancer and endometrial atypical hyperplasia

ObjectivesTransvaginal ultrasound is typically the initial diagnostic approach in patients with postmenopausal bleeding for detecting endometrial atypical hyperplasia/cancer. Although transvaginal ultrasound demonstrates notable sensitivity, its specificity remains limited. The objective...

Speed and efficiency: evaluating pulmonary nodule detection with AI-enhanced 3D gradient echo imaging.

Evaluating the diagnostic feasibility of accelerated pulmonary MR imaging for detection and characterisation of pulmonary nodules with artificial intelligence-aided compressed sensing. In this prospective trial, patients with benign and malignant lung nodules admitted between December 2021 and December 2022 underwent chest CT and pulmonary MRI. Pulmonary MRI used a respiratory-gated 3D gradient echo sequence, accelerated with a combination of parallel imaging, compressed sensing, and deep learning image reconstruction with three different acceleration factors (CS-AI-7, CS-AI-10, and CS-AI-15). Two readers evaluated image quality (5-point Likert scale), nodule detection and characterisation (size and morphology) of all sequences compared to CT in a blinded setting. Reader agreement was determined using the intraclass correlation coefficient (ICC). Thirty-seven patients with 64 pulmonary nodules (solid n = 57 [3-107 mm] part-solid n = 6 [ground glass/solid 8 mm/4-28 mm/16 mm] ground glass nodule n = 1 [20 mm]) were analysed. Nominal scan times were CS-AI-7 3:53 min; CS-AI-10 2:34 min; CS-AI-15 1:50 min. CS-AI-7 showed higher image quality, while quality remained diagnostic even for CS-AI-15. Detection rates of pulmonary nodules were 100%, 98.4%, and 96.8% for CS-AI factors 7, 10, and 15, respectively. Nodule morphology was best at the lowest acceleration and was inferior to CT in only 5% of cases, compared to 10% for CS-AI-10 and 23% for CS-AI-15. The nodule size was comparable for all sequences and deviated on average < 1 mm from the CT size. The combination of compressed sensing and AI enables a substantial reduction in the scan time of lung MRI while maintaining a high detection rate of pulmonary nodules. Incorporating compressed sensing and AI in pulmonary MRI achieves significant time savings without compromising nodule detection or characteristics. This advancement holds clinical promise, enhancing efficiency in lung cancer screening without sacrificing diagnostic quality. Lung cancer screening by MRI may be possible but would benefit from scan time optimisation. Significant scan time reduction, high detection rates, and preserved nodule characteristics were achieved across different acceleration factors. Integrating compressed sensing and AI in pulmonary MRI offers efficient lung cancer screening without compromising diagnostic quality.

A visual marker for early atrophy of the supraspinatus muscle on conventional MRI: introduction of the blackbird sign.

The aim of this study was to introduce the blackbird sign as a fast, qualitative measure of early supraspinatus (SSP) muscle atrophy and to correlate the sign with quantitatively assessed muscle volume and intramuscular fat fraction (FF) in patients with full-thickness SSP tears. The blackbird sign describes the asymmetric pattern of early SSP atrophy: on sagittal MR images, the supero-posterior contour of the muscle becomes concave, resembling the shape of a blackbird. MRIs of patients with full-thickness SSP tears were retrospectively reviewed for the presence of the blackbird and tangent signs. Patients were then divided into group 1: negative tangent sign and negative blackbird sign (n = 67), group 2: negative tangent sign and positive blackbird sign (n = 31), and group 3: positive tangent sign (n = 32). A 2-point Dixon sequence was acquired in all patients from which quantitative FF and muscle volumes were calculated. In total 130 patients (mean age 67 ± 11 years) were included. Mean SSP volume was significantly smaller in group 3 (15.8 ± 8.1 cm3) compared to group 2 (23.9 ± 7.0 cm3, p = 0.01) and group 1 (29.7 ± 9.1 cm3, p < 0.01). Significantly lower muscle volumes were also found in group 2 compared to group 1 (p = 0.02), confirming that the blackbird sign is able to identify early SSP atrophy. Mean FF in the SSP was significantly higher in group 3 (18.5 ± 4.4%) compared to group 2 (10.9 ± 4.7%, p < 0.01) and group 1 (6.1 ± 2.6%, p < 0.01). Visual assessment of early muscle atrophy of the SSP is feasible and reproducible using the blackbird sign, allowing the diagnosis of early SSP atrophy. In routine clinical practice, the blackbird sign may be a useful tool for assessing early muscle degeneration before the risk of postoperative rotator cuff re-tears increases with progressive muscle atrophy and fatty infiltration. Quantitative measurements of rotator cuff injuries require time, limiting clinical practicality. The proposed blackbird sign is able to identify early SSP atrophy. Reader agreement for the blackbird sign was substantial, demonstrating reproducibility and ease of implementation in the clinical routine.

Features associated with clinically actionable hyperechoic hepatic lesions to determine the need for follow-up.

To evaluate the rate of hyperechoic liver lesions that are clinically actionable and evaluate imaging and clinical factors associated with these to determine the need for follow-up. This retrospective study included 228 hyperechoic hepatic lesions on ultrasound in 228 patients. Reference standards included either dynamic contrast enhanced MRI (n = 130) or CT (n = 46), follow-up ultrasound performed at least 2years from baseline (n = 50), or histopathology (n = 2). Three radiologists independently assessed imaging features including lesion orientation, degree of hyper-echogenicity, lesion heterogeneity, and background liver echotexture. Univariable and multivariable logistic regression was used to determine features associated with an actionable hyperechoic lesion. Of the 228 hyperechoic lesions, 14 (6.1%) lesions were clinically actionable (or requiring follow-up), and 214 (93.9%) were clinically insignificant. Features that differed between patients with clinically insignificant vs. actionable lesions included: age (52.9 ± 15.1 vs. 63.9 ± 15.8 years, p = 0.004), male sex (43.9% vs 71.4%, p = 0.045), history of cirrhosis (6.5% vs 50%, p < 0.001), lesion size (1.9 ± 1.4cm vs. 3.5 ± 2.8cm, p = 0.003), heterogeneous lesion echogenicity (16.4% vs. 50%, p = 0.006), and cirrhotic/coarsened background liver (7.5% vs. 35.7%, p = 0.005). Stepwise logistic regression and multivariable analysis identified age, presence of cirrhosis, and lesion size as features most predictive of an actionable lesion (OR 1.04, 24.3, 1.77 respectively). Reader agreement for imaging features was fair to moderate (k = 0.29-0.53). 100%(168/168) of hyperechoic liver lesions measuring ≤ 3cm in patients without a history of malignancy or underlying liver disease were clinically insignificant. Our study findings support the overall favorable diagnoses of hyperechoic liver lesions ≤ 3cm in patients without underlying risk factors.

Submillisievert Abdominal Photon-Counting CT versus Energy-integrating Detector CT for Urinary Calculi Detection: Impact on Diagnostic Confidence.

Background Contrast-unenhanced abdominal CT is the imaging standard for urinary calculi detection; however, studies comparing photon-counting detector (PCD) CT and energy-integrating detector (EID) CT dose-reduction potentials are lacking. Purpose To compare the radiation dose and image quality of optimized EID CT with those of an experimental PCD CT scan protocol including tin prefiltration in patients suspected of having urinary calculi. Materials and Methods This retrospective single-center study included patients who underwent unenhanced abdominal PCD CT or EID CT for suspected urinary caliculi between February 2022 and March 2023. Signal and noise measurements were performed at three anatomic levels (kidney, psoas, and obturator muscle). Nephrolithiasis and/or urolithiasis presence was independently assessed by three radiologists, and diagnostic confidence was recorded on a five-point scale (1, little to no confidence; 5, complete confidence). Reader agreement was determined by calculating Krippendorff α. Results A total of 507 patients (mean age, 51.7 years ± 17.4 [SD]; 317 male patients) were included (PCD CT group, 229 patients; EID CT group, 278 patients). Readers 1, 2, and 3 detected nephrolithiasis in 129, 127, and 129 patients and 94, 94, and 94 patients, whereas the readers detected urolithiasis in 113, 114, and 114 patients and 152, 153, and 152 patients in the PCD CT and EID CT groups, respectively. Regardless of protocol (PCD CT or EID CT) or calculus localization, near perfect interreader agreement was found (α ≥ 0.99; 95% CI: 0.99, 1). There was no evidence of a difference in reader confidence between PCD CT and EID CT (median confidence, 5; IQR, 5-5; P ≥ .57). The effective doses were 0.79 mSv (IQR, 0.63-0.99 mSv) and 1.39 mSv (IQR, 1.01-1.87 mSv) for PCD CT and EID CT, respectively. Despite the lower radiation exposure, the signal-to-noise ratios at the kidney, psoas, and obturator levels were 30%, 23%, and 17% higher, respectively, in the PCD CT group (P < .001). Conclusion Submillisievert abdominal PCD CT provided high-quality images for the diagnosis of urinary calculi; radiation exposure was reduced by 44% with a higher signal-to-noise ratio than with EID CT and with no evidence of a difference in reader confidence. Published under a CC BY 4.0 license. Supplemental material is available for this article. See also the editorial by Nezami and Malayeri in this issue.

TRPS1 expression in breast angiosarcoma.

Angiosarcoma (AS) of the breast, a rare mesenchymal neoplasm, exhibits distinct forms based on etiological and genetic features. While cases with typical clinical presentation and morphology allow for a straightforward diagnosis, challenges arise when clinical data are scarce, diagnostic material is limited, or morphological characteristics overlap with other tumors, including undifferentiated carcinomas. The trichorhinophalangeal syndrome protein 1 (TRPS1), once regarded as highly specific for breast carcinomas, now faces doubts regarding its reliability. This study explores TRPS1 expression in breast AS. Our investigation revealed that 60% of AS cases displayed TRPS1 labeling, contrasting with the 40% lacking expression. Scoring by four independent readers established a consensus, designating 12/35 ASs as unequivocally TRPS1-positive. However, uncertainty surrounded nine further cases due to a lack of reader agreement (being substantial as reflected by a kappa value of 0.76). These findings challenge the perceived specificity of TRPS1, shedding light on its presence in a noteworthy proportion of breast ASs. Consequently, the study underscores the importance of a comprehensive approach in evaluating breast ASs and expands the range of entities within the differential diagnosis associated with TRPS1 labeling.

Can Dynamic Contrast-Enhanced MRI Be Used to Differentiate Hepatic Hemangioma from Other Lesions in Early Infancy?

Background Confident diagnosis of hepatic hemangioma on imaging can avoid biopsy in early infancy and helps guide conservative management. Purpose This article aims to determine if dynamic contrast-enhanced magnetic resonance imaging (MRI) can be used to differentiate liver hemangioma from other lesions in infants ≤ 100 days and to determine association of MRI features with hepatic lesions. Methods MRI performed for liver lesions were retrospectively reviewed to note imaging characteristics and the MRI diagnosis. Final diagnosis was assigned based on pathology in available cases and by corroborative standard of reference including overall clinical features, lab findings, and follow-up. Results Of 30 infants (18 boys, 12 girls; average age 42.2 days) included, 18 had solitary and 12 had multifocal lesions. Diagnoses in total 33 lesions included hemangiomas (23), hepatoblastoma (6), arteriovenous malformation (2), neuroblastoma metastases (1), and infarction (1). MRI and final diagnosis matched in 94% lesions with almost perfect agreement (kappa 0.86) for reader 1, and matched in 88% lesions with substantial agreement (kappa 0.71) for reader 2. Interobserver agreement for MRI diagnosis was substantial (kappa 0.62). Sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of MRI in differentiating hemangioma from other lesions were 100, 90, 96, 100, and 97%, respectively. Centripetal (16/23) or flash (5/23) filling were only seen with hemangioma. There was no significant difference in alpha-fetoprotein elevation ( p 0.08), average size ( p 0.35), multifocality ( p 0.38), and intralesional hemorrhage ( p 1) between hemangioma and hepatoblastoma. Conclusion Centripetal filling on dynamic imaging and absence of washout are characteristic MRI features of hepatic hemangioma that can help to differentiate it from other lesions in early infancy.

Estimating Age and Growth of Largemouth Bass in Southwestern Reservoirs by Using Otoliths and Scales

Abstract Age and growth data are frequently used to monitor and manage important North American sport fishes such as Largemouth Bass Micropterus salmoides. Continental and regional growth standards have been developed for this species to assess fish growth over time and across space. However, Largemouth Bass age and growth data are infrequently collected in Arizona and the reliability of age estimates derived from typical structures (e.g., scales, otoliths) in the Southwest is uncertain. Our objectives were to 1) compare precision and bias of age estimates from scales with those from otoliths and 2) estimate Largemouth Bass growth in several southwestern warmwater reservoirs by using otoliths. We collected Largemouth Bass from three Arizona reservoirs—Alamo, Peña Blanca, and Roosevelt—by boat electrofishing in spring 2021. We removed scales and sagittal otoliths from fish and then prepared and independently aged them three times. We compared differences in precision and bias between scales and otoliths using reader agreement percentages, confidence ratings, average coefficients of variation, and age-bias plots. We used age estimates from Largemouth Bass otoliths to calculate mean lengths-at-age at capture and relative growth indices based on published growth standards in each reservoir. Largemouth Bass scale age estimates were less precise, overestimated ages of younger fish, and underestimated age of older fish compared with those of otoliths. Growth was lower in Peña Blanca Lake than in the other two reservoirs according to mean length-at-age estimates, and relative growth indices suggested that Largemouth Bass growth in all three reservoirs was above average at younger ages, but less so at older ages. The results from this study add to a growing body of literature supporting the use of otoliths for estimating age and growth of Largemouth Bass.

Agreement of Gallbladder Reporting and Data System for Gallbladder Wall Thickening at Ultrasonography: A Multireader Validation Study

ObjectiveTo evaluate the intra-reader and inter-reader agreement of ultrasound (US) gallbladder reporting and data system (GB-RADS) and validate the risk of malignancy in each GB-RADS category. Materials and MethodsThis retrospective study comprised consecutive patients with non-acute GWT who underwent US evaluation between January 2019 and December 2022. Three radiologists independently read the static US images and cine-loops for GB-RADS findings and assigned GB-RADS categories. The intra-observer (static images) and inter-observer (static images and cine-loops) agreement was calculated using kappa statistics and Krippendorff’s alpha. Another radiologist assigned a consensus GB-RADS category. The percentage of malignancy in each GB-RADS category was calculated. ResultsStatic US images of 414 patients (median age, 56 years; 288 women, benign=45.6% and malignant=54.4%) and cine-loops of 50 patients were read. There was weak to moderate intra-reader agreement for most GB-RADS findings and moderate intra-reader agreement for the GB-RADS category for all readers. On static images, the inter-reader agreement was acceptable for GB-RADS categories. On cine-loops, the inter-reader agreement for GB-RADS findings and categories was better than static images. The percentage of malignancy was 1.2%, 37%, 71.1%, and 89.1% in GB-RADS 2, 3, 4 and 5 categories. ConclusionGB-RADS has moderate intra-reader for GB-RADS categories. As originally proposed, the risk of malignancy is negligible in GB-RADS 2 category and highest in GB-RADS 5 category. However, the discriminatory performance of GB-RADS 3 and 4 categories is low. Larger multicenter studies with more readers must assess the reader agreement and validate the GB-RADS systems for wider clinical utilization.

Radiological follow-up of cemento-osseous dysplasia on cone-beam computed tomography

This study investigated the natural course of cemento-osseous dysplasia (COD) on cone-beam computed tomography (CBCT).Retrospectively, 104 CBCT scans from 36 patients (mean age, 44.5 years; 33 female and three male) with mandibular COD (10 florid, seven focal, 19 periapical) were included, based upon clinico-radiological features, without complications such as infection and related surgery. Changes in maximum diameter and morphology (lytic, mixed lytic-sclerotic, sclerotic) were evaluated in 83 lesions, with a mean follow-up of 28.3 months. The occurrence of a diameter increase was assessed by time-to-event analysis; interreader agreement for diameter and morphological evaluation by intraclass correlation coefficient and weighted κ statistics, respectively.Fifteen of 83 (18.1%) lesions (eight florid, one focal, six periapical) in 10 patients increased in diameter; 12 of 83 (14.5%) lesions (five florid, seven periapical) in 11 patients changed morphologically. The median period until a diameter increase was longest (120 months) for periapical COD, and shortest (66 months) for florid COD (p = 0.023). There was high reader agreement (ICC = 0.891; weighted κ = 0.901).In conclusion, CBCT is an effective tool with which to follow-up COD. If any, the natural progress in uncomplicated COD is prolonged, which underlines its non-surgical character and aids in its long-term management.

Detection Rate of Small Pancreas Cysts and Intrareader Reliability of the Cysts Size Measurements on Transabdominal Ultrasonography.

This study aimed to assess the detection rate of small (<10 mm) pancreas cyst and intrareader reliability for cyst size measurements on transabdominal ultrasonography (US). From 2020 to 2022, 194 pancreas cysts in 173 patients, incidentally detected on computed tomography or magnetic resonance imaging, were evaluated on US by 1 of 2 radiologists (readers 1 and 2). Intrareader agreements of cyst size measurements on US were assessed by intraclass correlation coefficient (ICC). Bland-Altman plot was used to visualize the differences between the first and second size measurements in each reader. In this study, readers 1 and 2 evaluated 86 cysts in 76 patients and 108 cysts in 97 patients, respectively. Most of the cysts (191 of 194) were located in the nontail portion of the pancreas. Overall detection rate of pancreas cysts by US was 92.3% (179 of 194). The mean size of measured 179 pancreas cysts was 4.7 ± 1.5 mm. The readers showed excellent intrareader agreements (ICC = 0.925 and 0.960) for cyst size measurements, except for the cysts with size ≤5 mm, where both readers showed good intrareader agreements (ICC = 0.848 and 0.873). The 95% limits of agreement of readers 1 and 2 were 13.8% and 14.9% of the mean, respectively. Therefore, transabdominal US could be a reliable follow-up imaging modality for small (<10 mm) nontail pancreas cysts incidentally detected on computed tomography or magnetic resonance imaging, especially for the cysts with size between 5 and 10 mm. Size changes of the pancreas cysts approximately less than 15% may be within the measurement error.

Clinical Utility of Breast Ultrasound Images Synthesized by a Generative Adversarial Network.

This study compares the clinical properties of original breast ultrasound images and those synthesized by a generative adversarial network (GAN) to assess the clinical usefulness of GAN-synthesized images. We retrospectively collected approximately 200 breast ultrasound images for each of five representative histological tissue types (cyst, fibroadenoma, scirrhous, solid, and tubule-forming invasive ductal carcinomas) as training images. A deep convolutional GAN (DCGAN) image-generation model synthesized images of the five histological types. Two diagnostic radiologists (reader 1 with 13 years of experience and reader 2 with 7 years of experience) were given a reading test consisting of 50 synthesized and 50 original images (≥1-month interval between sets) to assign the perceived histological tissue type. The percentages of correct diagnoses were calculated, and the reader agreement was assessed using the kappa coefficient. The synthetic and original images were indistinguishable. The correct diagnostic rates from the synthetic images for readers 1 and 2 were 86.0% and 78.0% and from the original images were 88.0% and 78.0%, respectively. The kappa values were 0.625 and 0.650 for the synthetic and original images, respectively. The diagnoses made from the DCGAN synthetic images and original images were similar. The DCGAN-synthesized images closely resemble the original ultrasound images in clinical characteristics, suggesting their potential utility in clinical education and training, particularly for enhancing diagnostic skills in breast ultrasound imaging.

Service and clinical impacts of reader bias in breast cancer screening: a retrospective study.

To determine factors influencing reader agreement in breast screening and investigate the relationship between agreement level and patient outcomes. Reader pair agreement for 83 265 sets of mammograms from the Scottish Breast Screening service (2015-2020) was evaluated using Cohen's kappa statistic. Each mammography examination was read by two readers, per routine screening practice, with the second initially blinded but able to choose to view the first reader's opinion. If the two readers disagreed, a third reader arbitrated. Variation in reader agreement was examined by: whether the reader acted as the first or second reader, reader experience, and recall, cancer detection and arbitration recall rate. Readers' opinions varied by whether they acted as the first or second reader. Furthermore, reader 2 was more likely to agree with reader 1 if reader 1 was more experienced than they were, and less likely to agree if they themselves were more experienced than reader 1 (P < .001). Agreement was not significantly associated with cancer detection rate, overall recall rate or arbitration recall rates (P > .05). Lower agreement between readers led to a higher arbiter workload (P < .001). In mammography screening, the second reader's opinion is influenced by the first reader's opinion, with the degree of influence dependent on the readers' relative experience levels. While less-experienced readers relied on their more experienced reading partner, no adverse impact on service outcomes was observed. Allowing access to the first reader's opinion may benefit newly qualified readers, but reduces independent evaluation, which may lower cancer detection rates.

Quantification supports visual assessment of challenging amyloid‐PET images

AbstractBackgroundIn clinical practice, amyloid‐PET images are visually assessed to determine the presence or absence of amyloid pathology. The expected arrival of disease‐modifying treatments warrants the need for objective quantification of amyloid burden to further support treatment decisions. Here, we investigated whether quantification increases reader confidence and influences visual read (VR) classification in patients with challenging or ambiguous amyloid information sampled across the clinical continuum.MethodsFrom the AMYPAD‐DPMS trial (N = 840), we selected a subsample of 85 amyloid‐PET scans ([18F]flutemetamol, N = 48/[18F]florbetaben, N = 53), for which 1) visual assessment was discordant with amyloid status based on Centiloid (CL) (cut‐off = 21) and/or 2) were assessed with a low confidence (i.e. ≤3 on a 5‐point scale) by the local reader. In addition, concordant negative and positive (N = 8/8) scans across tracers were included to balance the dataset (N = 101, Table‐1). PET images were processed using GEHCs AMYPYPE PET‐only pipeline, providing global CL and regional z‐scores for 7 cortical ROIs. VR classification and corresponding confidence was assessed for 3 independent certified assessors before and after disclosure of the quantification results. In addition, readers stated whether CL quantification and/or the regional z‐scores supported their assessment. Reader agreement was assessed using Fleiss Kappa. Changes in reader confidence after disclosure was assessed using paired t‐test.ResultsOverall, readers reported global CL and regional z‐score quantification to be VR‐supportive in 76.2% and 49.7% of assessments, respectively. After disclosure of quantitative results, concordance between readers (Kbaseline= 0.60; Kpost‐disclosure= 0.79), concordance between majority‐VR and CL‐status (agreementbaseline = 75.2%; agreementpost‐disclosure = 81.2%), and reader confidence (Mbaseline= 4.0 vs. Mpost‐disclosure= 4.37, t = ‐5.8, p&lt;0.001) significantly improved. Classification changed for 26 (8.6%) assessments, with 8 (30.8%) changing to VR‐ and 18 (69.2%) to VR+. In these cases, CL was considered supportive in 16 (61.5%) cases, regional z‐scores were considered supportive for 8/10 remaining cases, and most often for the posterior cingulate/precuneus region. The 2 remaining cases had a low confidence at both time‐points. Example cases are shown in Figure 1.ConclusionIn a dataset enriched with challenging amyloid‐PET scans for visual assessment, quantification resulted in a significantly increased agreement and confidence across readers. We further aim to determine in which cases quantification was considered most supportive.

Stress-only dynamic computed tomography perfusion protocol (CTP) alone without computed tomography coronary angiography (CCTA) has limited specificity to diagnose ischemia: A retrospective two-center study

PurposeTo investigate diagnostic performance of stress-only dynamic myocardial computed tomography perfusion (CTP) without computed tomography coronary angiography (CCTA) to diagnose ischemia with invasive fractional flow reserve (FFR) as a reference standard. Method135 datasets (68 positive for ischemia with invasive FFR < 0.8) acquired with a 256-slice CT system (Revolution, GE Healthcare, Chicago, IL, USA) were retrieved, postprocessed with a deep learning-based algorithm (Advanced intelligent Clear-IQ Engine (AiCE), Canon Medical Systems, Otawara, Japan) (FC03/cardiac kernel, 8 mm slice thickness), analyzed using a dedicated workstation (Vitrea research 7.11.0. Vital Images, Minnetonka, MN, USA), and loaded into a clinical workstation (CardIQ, GE Healthcare, Chicago, IL, USA) for review. Ten observers with various experience from two research sites evaluated the post-processed images, perfusion slices and maps to indicate presence vs absence of perfusion defect and its probability (five-point Likert scale). Binary decisions and probability scores were used to calculate sensitivity and specificity for each reader, and to create receiver operating characteristics (ROC) curves, respectively. Furthermore, the correlation coefficient (ICC) was computed. ROC AUC of a purely quantitative analysis was obtained thanks to a color-coded map with a fixed scale superimposed on myocardial walls displaying myocardial blood flow (MBF) values. ResultsThe overall case-based sensitivity and specificity for the detection of perfusion deficit were 0.79 and 0.30, respectively. No significant differences were detected in the AUC across readers (p value = 0.66). The AUC values were 0.50, 0.58, 0.63, 0.59, 0.45, 0.60, 0.56, 0.61, 0.52, 0.61. Absolute reader agreement ICC was 0.60 (good agreement) for an average case. ConclusionDynamic CTP alone has good sensitivity, but low specificity when analyzed without CCTA. These findings reinforce the need to guide the interpretation functional test with the knowledge of coronary artery anatomy.

The Performance of Flash Replenishment Contrast-Enhanced Ultrasound for the Qualitative Assessment of Kidney Lesions in Patients with Chronic Kidney Disease.

We investigated the accuracy of CEUS for characterizing cystic and solid kidney lesions in patients with chronic kidney disease (CKD). Cystic lesions are assessed using Bosniak criteria for computed tomography (CT) and magnetic resonance imaging (MRI); however, in patients with moderate to severe kidney disease, CT and MRI contrast agents may be contraindicated. Contrast-enhanced ultrasound (CEUS) is a safe alternative for characterizing these lesions, but data on its performance among CKD patients are limited. We performed flash replenishment CEUS in 60 CKD patients (73 lesions). Final analysis included 53 patients (63 lesions). Four readers, blinded to true diagnosis, interpreted each lesion. Reader evaluations were compared to true lesion classifications. Performance metrics were calculated to assess malignant and benign diagnoses. Reader agreement was evaluated using Bowker's symmetry test. Combined reader sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) for diagnosing malignant lesions were 71%, 75%, 45%, and 90%, respectively. Sensitivity (81%) and specificity (83%) were highest in CKD IV/V patients when grouped by CKD stage. Combined reader sensitivity, specificity, PPV, and NPV for diagnosing benign lesions were 70%, 86%, 91%, and 61%, respectively. Again, in CKD IV/V patients, sensitivity (81%), specificity (95%), and PPV (98%) were highest. Inter-reader diagnostic agreement varied from 72% to 90%. In CKD patients, CEUS is a potential low-risk option for screening kidney lesions. CEUS may be particularly beneficial for CKD IV/V patients, where kidney preservation techniques are highly relevant.

Abdominopelvic CT in COVID-19 patients with abdominal complaints including both waves and controls: reader agreement and overcalls after consensus review

BackgroundSince many COVID-19 publications lack consensus reviews or controls, interpretive accuracy is unclear; abdominal processes unique or infrequent during the pandemic remain unknown. The incidence and nature of CT findings accounting for abdominal complaints in COVID patients, reader agreement and overcalling will be determined. MethodsA retrospective study was performed on COVID patients with abdominal complaints from 3/15/2020–5/31/2020 and 11/1/2020–4/15/2021 including matched controls. Reviewers blinded to initial reads interpreted abdominopelvic CT exams, with discordant cases resolved in consensus. Reader agreement was measured by Cohen's Kappa, differences between cohorts by permutation tests and factors affecting false positive/negative rates by Fisher's Exact Test and logistic regression. Results116 first wave (average age 65 years [±15.3], 63 [54%] women) and 194 second wave COVID cases (average age 64 years [±16.3], 103 [53%] women) including 116 wave 1 and 194 wave 2 prepandemic controls were included. Concordance was lower among COVID cases than controls (Cohen's Kappa of 0.58 vs. 0.82 [p ≤ 0.001]) and among wave 1 than wave 2 cases (Cohen's Kappa of 0.45 vs. 0.66 [p = 0.052]). With true positives defined as consensus between the initial reader and study reader, false positive rates were higher among COVID cases than controls (OR = 0.42, p = 0.003) and for initial than study reader (OR = 0.36, p ≤ 0.001), but lower in wave 2 than 1 (OR = 0.5, p = 0.028). ConclusionGreater reader disagreement occurred during COVID than prepandemic with no reader bias as both initial and study readers called more false positives among COVID cases than controls. More overcalling occurred during COVID with colitis and cystitis most common.

Comparison of Radiologists and Deep Learning for US Grading of Hepatic Steatosis.

Background Screening for nonalcoholic fatty liver disease (NAFLD) is suboptimal due to the subjective interpretation of US images. Purpose To evaluate the agreement and diagnostic performance of radiologists and a deep learning model in grading hepatic steatosis in NAFLD at US, with biopsy as the reference standard. Materials and Methods This retrospective study included patients with NAFLD and control patients without hepatic steatosis who underwent abdominal US and contemporaneous liver biopsy from September 2010 to October 2019. Six readers visually graded steatosis on US images twice, 2 weeks apart. Reader agreement was assessed with use of κ statistics. Three deep learning techniques applied to B-mode US images were used to classify dichotomized steatosis grades. Classification performance of human radiologists and the deep learning model for dichotomized steatosis grades (S0, S1, S2, and S3) was assessed with area under the receiver operating characteristic curve (AUC) on a separate test set. Results The study included 199 patients (mean age, 53 years ± 13 [SD]; 101 men). On the test set (n = 52), radiologists had fair interreader agreement (0.34 [95% CI: 0.31, 0.37]) for classifying steatosis grades S0 versus S1 or higher, while AUCs were between 0.49 and 0.84 for radiologists and 0.85 (95% CI: 0.83, 0.87) for the deep learning model. For S0 or S1 versus S2 or S3, radiologists had fair interreader agreement (0.30 [95% CI: 0.27, 0.33]), while AUCs were between 0.57 and 0.76 for radiologists and 0.73 (95% CI: 0.71, 0.75) for the deep learning model. For S2 or lower versus S3, radiologists had fair interreader agreement (0.37 [95% CI: 0.33, 0.40]), while AUCs were between 0.52 and 0.81 for radiologists and 0.67 (95% CI: 0.64, 0.69) for the deep learning model. Conclusion Deep learning approaches applied to B-mode US images provided comparable performance with human readers for detection and grading of hepatic steatosis. Published under a CC BY 4.0 license. Supplemental material is available for this article. See also the editorial by Tuthill in this issue.

The natural history of ductal carcinoma in situ (DCIS) in simulation models: A systematic review.

Assumptions on the natural history of ductal carcinoma in situ (DCIS) are necessary to accurately model it and estimate overdiagnosis. To improve current estimates of overdiagnosis (0-91%), the purpose of this review was to identify and analyse assumptions made in modelling studies on the natural history of DCIS in women. A systematic review of English full-text articles using PubMed, Embase, and Web of Science was conducted up to February 6, 2023. Eligibility and all assessments were done independently by two reviewers. Risk of bias and quality assessments were performed. Discrepancies were resolved by consensus. Reader agreement was quantified with Cohen's kappa. Data extraction was performed with three forms on study characteristics, model assessment, and tumour progression. Thirty models were distinguished. The most important assumptions regarding the natural history of DCIS were addition of non-progressive DCIS of 20-100%, classification of DCIS into three grades, where high grade DCIS had an increased chance of progression to invasive breast cancer (IBC), and regression possibilities of 1-4%, depending on age and grade. Other identified risk factors of progression of DCIS to IBC were younger age, birth cohort, larger tumour size, and individual risk. To accurately model the natural history of DCIS, aspects to consider are DCIS grades, non-progressive DCIS (9-80%), regression from DCIS to no cancer (below 10%), and use of well-established risk factors for progression probabilities (age). Improved knowledge on key factors to consider when studying DCIS can improve estimates of overdiagnosis and optimization of screening.

Extended and weightbearing wrist 3-T MRI using a novel harness and flexible 24-channel glove coil to evaluate carpal kinematics: a pilot study in 10 volunteers.

Wrist pain in the extended or extended weightbearing positions may be incompletely evaluated using standard magnetic resonance imaging (MRI) with standard rigid clamshell coils in the neutral position. To evaluate a flexible 24-channel glove coil and harness when imaging the wrist in neutral, dorsally extended, and weightbearing positions. Ten wrists in 10 asymptomatic volunteers (mean age = 29 years) were scanned. Participants underwent 3-T MRI using the harness and flexible glove coil, acquiring sagittal turbo spin echo (TSE) and half-Fourier acquisition single-shot turbo spin echo (HASTE) pulse sequences. Static TSE images were obtained in neutral, extended, and weightbearing positions using proton density parameters and independently evaluated by two radiologists for: dorsal radiocarpal ligament thickness; radiocapitate, radiolunate, and capitatolunate angles; palmar translation of the lunate on the radius; angulation of the extensor tendons; and distance from the distal extensor retinaculum to Lister's tubercle. Cine HASTE images were dynamically acquired between neutral-maximum extension to measure the radiocapitate angle. Good reader agreement was observed (r > 0.73) for all measurements except palmar translation in the neutral position (r = 0.27). Significant increases in dorsal radiocarpal ligament thickness; radiocapitate, radiolunate and capitolunate angulation; and extensor tendon angulation were observed between the neutral and extended positions (P < 0.001). A further significant increase in these metrics between extended and weightbearing positions was also seen (P < 0.01). Significant increases in dorsal radiocarpal ligament thickness, articular and tendon angulations occur during wrist extension, that further increase with dorsal weightbearing.