Double Reading Research Articles

The impact of artificial intelligence on double reading of mammograms

The impact of artificial intelligence on double reading of mammograms

Breast cancer detection accuracy of AI in an entire screening population: a retrospective, multicentre study

BackgroundArtificial intelligence (AI) systems are proposed as a replacement of the first reader in double reading within mammography screening. We aimed to assess cancer detection accuracy of an AI system in a Danish screening population.MethodsWe retrieved a consecutive screening cohort from the Region of Southern Denmark including all participating women between Aug 4, 2014, and August 15, 2018. Screening mammograms were processed by a commercial AI system and detection accuracy was evaluated in two scenarios, Standalone AI and AI-integrated screening replacing first reader, with first reader and double reading with arbitration (combined reading) as comparators, respectively. Two AI-score cut-off points were applied by matching at mean first reader sensitivity (AIsens) and specificity (AIspec). Reference standard was histopathology-proven breast cancer or cancer-free follow-up within 24 months. Coprimary endpoints were sensitivity and specificity, and secondary endpoints were positive predictive value (PPV), negative predictive value (NPV), recall rate, and arbitration rate. Accuracy estimates were calculated using McNemar’s test or exact binomial test.ResultsOut of 272,008 screening mammograms from 158,732 women, 257,671 (94.7%) with adequate image data were included in the final analyses. Sensitivity and specificity were 63.7% (95% CI 61.6%-65.8%) and 97.8% (97.7-97.8%) for first reader, and 73.9% (72.0-75.8%) and 97.9% (97.9-98.0%) for combined reading, respectively. Standalone AIsens showed a lower specificity (-1.3%) and PPV (-6.1%), and a higher recall rate (+ 1.3%) compared to first reader (p < 0.0001 for all), while Standalone AIspec had a lower sensitivity (-5.1%; p < 0.0001), PPV (-1.3%; p = 0.01) and NPV (-0.04%; p = 0.0002). Compared to combined reading, Integrated AIsens achieved higher sensitivity (+ 2.3%; p = 0.0004), but lower specificity (-0.6%) and PPV (-3.9%) as well as higher recall rate (+ 0.6%) and arbitration rate (+ 2.2%; p < 0.0001 for all). Integrated AIspec showed no significant difference in any outcome measures apart from a slightly higher arbitration rate (p < 0.0001). Subgroup analyses showed higher detection of interval cancers by Standalone AI and Integrated AI at both thresholds (p < 0.0001 for all) with a varying composition of detected cancers across multiple subgroups of tumour characteristics.ConclusionsReplacing first reader in double reading with an AI could be feasible but choosing an appropriate AI threshold is crucial to maintaining cancer detection accuracy and workload.

Alteraciones pulmonares persistentes tras 18 meses de neumonía por SARS-CoV-2

ObjectiveTo describe persistent pulmonary abnormalities detected on HRCT after 18months of SARS-CoV-2 pneumonia, and to determine their extension and correlation with pulmonary function. Patients and methodsA prospective cross-sectional study with an initial cohort of 90 patients in follow-up due to persisting lung abnormalities on imaging, functional respiratory impairment and/or respiratory symptoms. Of these, 31 (34%) were selected for analysis due to the persistence of their lung abnormalities on HRCT at 18 months after infection. A double reading was performed for each HRCT (62 observations). ResultsOf the 31 patients included: 20 (65%) were men; mean age was 67years; 17 (55%) were smokers/ex-smokers. The mean hospitalisation time was 38days. Eighteen (58%) patients were admitted to intensive care units. Five patients (16%) suffered an acute pulmonary thromboembolism and three (9.7%) had a pneumothorax. The mean time between the onset of pneumonia and the follow-up HRCT was 20.34 months. Extension of total pulmonary abnormalities, ground-glass opacities and reticulation was 19%, 12% and 4.5% respectively. The findings of the 62 readings were: ground-glass opacities (100%), reticulation (83%), subpleural curvilinear lines (62%), parenchymal bands (34%), traction bronchiectasis (69%), displacement of vessels/fissures (46%) and honeycombing (4.9%). Pulmonary function 18months after the acute episode revealed a mean FVC of 92% of predicted value, with an FVC <80% of predicted value in 11 patients (35.4%). Mean DLCO was 71% of predicted value, with a DLCO <80% in 22 patients (70%). We observed a statistically significant relationship between total extension of abnormalities on HRCT and FVC (P<.05), and a trend towards statistical significance with DLCO (P=.051); there was a statistically significant relationship between the presence of ground-glass opacities and FEV1/FVC (P<.01). The relationships between reticulation and FVC, FVC%, FEV1, FEV1% and DLCO% were also considered statistically significant (P<.05). ConclusionPersistent interstitial lung abnormalities are seen on HRCT for a subset of patients infected with SARS-CoV-2 pneumonia. Seventy percent of these patients suffered a slight decrease in DLCO.

The Afterlives of Byron’s ‘The Isles of Greece’ in the Victorian Press

In the nineteenth-century British press, ‘The Isles of Greece, the Isles of Greece’, as the Brighton Gazette put it in 1878, ‘have had their joys and beauties sung in lofty strains by the wisest, the wittiest, and the wickedest of poets’. These mostly unidentified poets remained true to the spirit of the original recitation of Byron’s ‘The Isles of Greece’ from Don Juan , where it is presented as a performance by a poet whom we may or may not trust. The poem’s double reading, its levels of irony, denial of authority, and eventual misreading, makes the persistent reappearance in various forms of the double exclamation ‘The isles of Greece, the isles of Greece!’ in the Victorian press a natural extension or afterlife of its dynamic. Political newspapers are a culturally dynamic space for unravelling the myriad of performative aspects of the poem in its various afterlives as we follow how it was transformed based on Britain’s relationship with Greece.

Impact of systematic dynamic maneuvers during computed tomography scan on the T classification of head and neck cancers

ObjectivesTo evaluate the impact of systematic dynamic maneuvers during CT scan on the T-staging of head and neck cancer (HNC). Materials and methodsCT scans from the initial workup of 443 consecutive patients treated for HNC in our institution were retrospectively analyzed. CT scans were performed in both expert centers (comprehensive cancer center and university hospital) and non-expert centers. We noted whether dynamic maneuvers (DM) were performed, in 3 categories, namely: DM not done (DMND), done and inadequate (DMDI), done and adequate (DMDA). In the group with DMDA, T-stage was evaluated without and with DM. Interobserver agreement for T staging was assessed after independent double reading of CT scans with and without DM by two radiologists in a random sample. ResultsAmong the 443 CT scans, DMND was observed in 36.3%, DMDI in 9.3% and DMDA in 54.4%. DMDA were significantly more frequent in expert than in non-expert centers (93.4 vs 6.6%, p < 0.001). In CT scans with DMDA, analysis of the 141 scans rated as T1, T2, T3, or T4 without DM showed agreement of 88.7% with scans with DM, corresponding to a reclassification rate of 11.3% (kappa = 0.85, 95%CI [0.78;0.92]). Among lesions initially classed as Tx without DM (N = 100), the reclassification rate was 76% including DM. ConclusionThe performance of systematic DM integrated into CT protocols is useful to reclassify the T stage in HNC and is essential in case of lesions initially classified as Tx without DM. DM should be performed routinely in expert and nonexpert centers.

Impact of real-life use of artificial intelligence as support for human reading in a population-based breast cancer screening program with mammography and tomosynthesis

ObjectivesTo evaluate the impact of using an artificial intelligence (AI) system as support for human double reading in a real-life scenario of a breast cancer screening program with digital mammography (DM) or digital breast tomosynthesis (DBT).Material and methodsWe analyzed the performance of double reading screening with mammography and tomosynthesis after implementarion of AI as decision support. The study group consisted of a consecutive cohort of 1 year screening between March 2021 and March 2022 where double reading was performed with concurrent AI support that automatically detects and highlights lesions suspicious of breast cancer in mammography and tomosynthesis. Screening performance was measured as cancer detection rate (CDR), recall rate (RR), and positive predictive value (PPV) of recalls. Performance in the study group was compared using a McNemar test to a control group that included a screening cohort of the same size, recorded just prior to the implementation of AI.ResultsA total of 11,998 women (mean age 57.59 years ± 5.8 [sd]) were included in the study group (5049 DM and 6949 DBT). Comparing global results (including DM and DBT) of double reading with vs. without AI support, we observed an increase in CDR, PPV, and RR by 3.2/‰ (5.8 vs. 9; p < 0.001), 4% (10.6 vs. 14.6; p < 0.001), and 0.7% (5.4 vs. 6.1; p < 0.001) respectively.ConclusionAI used as support for human double reading in a real-life breast cancer screening program with DM and DBT increases CDR and PPV of the recalled women.Clinical relevance statementArtificial intelligence as support for human double reading improves accuracy in a real-life breast cancer screening program both in digital mammography and digital breast tomosynthesis.Key Points• AI systems based on deep learning technology offer potential for improving breast cancer screening programs.• Using artificial intelligence as support for reading improves radiologists’ performance in breast cancer screening programs with mammography or tomosynthesis.• Artificial intelligence used concurrently with human reading in clinical screening practice increases breast cancer detection rate and positive predictive value of the recalled women.

Prospective implementation of AI-assisted screen reading to improve early detection of breast cancer

Artificial intelligence (AI) has the potential to improve breast cancer screening; however, prospective evidence of the safe implementation of AI into real clinical practice is limited. A commercially available AI system was implemented as an additional reader to standard double reading to flag cases for further arbitration review among screened women. Performance was assessed prospectively in three phases: a single-center pilot rollout, a wider multicenter pilot rollout and a full live rollout. The results showed that, compared to double reading, implementing the AI-assisted additional-reader process could achieve 0.7–1.6 additional cancer detection per 1,000 cases, with 0.16–0.30% additional recalls, 0–0.23% unnecessary recalls and a 0.1–1.9% increase in positive predictive value (PPV) after 7–11% additional human reads of AI-flagged cases (equating to 4–6% additional overall reading workload). The majority of cancerous cases detected by the AI-assisted additional-reader process were invasive (83.3%) and small-sized (≤10 mm, 47.0%). This evaluation suggests that using AI as an additional reader can improve the early detection of breast cancer with relevant prognostic features, with minimal to no unnecessary recalls. Although the AI-assisted additional-reader workflow requires additional reads, the higher PPV suggests that it can increase screening effectiveness.

Discordant and false-negative interpretations at digital breast tomosynthesis in the prospective Oslo Tomosynthesis Screening Trial (OTST) using independent double reading

ObjectivesTo analyze discordant and false-negatives of double reading digital breast tomosynthesis (DBT) versus digital mammography (DM) including reading times in the Oslo Tomosynthesis Screening Trial (OTST), and reclassify these in a retrospective reader study as missed, minimal sign, or true-negatives.MethodsThe prospective OTST comparing double reading DBT vs. DM had paired design with four parallel arms: DM, DM + computer aided detection, DBT + DM, and DBT + synthetic mammography. Eight radiologists interpreted images in batches using a 5-point scale. Reading time was automatically recorded. A retrospective reader study including four radiologists classified screen-detected cancers with at least one false-negative score and screening examinations of interval cancers as negative, non-specific minimal sign, significant minimal sign, and missed; the two latter groups are defined “actionable.” Statistics included chi-square, Fisher’s exact, McNemar’s, and Mann–Whitney U tests.ResultsDiscordant rate (cancer missed by one reader) for screen-detected cancers was overall comparable (DBT (31% [71/227]) and DM (30% [52/175]), p = .81), significantly lower at DBT for spiculated cancers (DBT, 19% [20/106] vs. DM, 36% [38/106], p = .003), but high (28/49 = 57%, p = 0.001) for DBT-only detected spiculated cancers. Reading time and sensitivity varied among readers. False-negative DBT-only detected spiculated cancers had shorter reading time than true-negatives in 46% (13/28). Retrospective evaluation classified the following DBT exams “actionable”: three missed by both readers, 95% (39/41) of discordant cancers detected by both modes, all 30 discordant DBT-only cancers, 25% (13/51) of interval cancers.ConclusionsDiscordant rate was overall comparable for DBT and DM, significantly lower at DBT for spiculated cancers, but high for DBT-only detected spiculated lesions. Most false-negative screen-detected DBT were classified as “actionable.”Clinical relevance statementRetrospective evaluation of false-negative interpretations from the Oslo Tomosynthesis Screening Trial shows that most discordant and several interval cancers could have been detected at screening. This underlines the potential for modern AI-based reading aids and triage, as high-volume screening is a demanding task.Key Points• Digital breast tomosynthesis (DBT) screening is more sensitive and has higher specificity compared to digital mammography screening, but high-volume DBT screening is a demanding task which can result in high discordance rate among readers.• Independent double reading DBT screening had overall comparable discordance rate as digital mammography, lower for spiculated masses seen on both modalities, and higher for small spiculated cancer seen only on DBT.• Almost all discordant digital breast tomosynthesis-detected cancers (72 of 74) and 25% (13 of 51) of the interval cancers in the Oslo Tomosynthesis Screening Trial were retrospectively classified as actionable and could have been detected by the readers.

Population-wide evaluation of artificial intelligence and radiologist assessment of screening mammograms

ObjectivesTo validate an AI system for standalone breast cancer detection on an entire screening population in comparison to first-reading breast radiologists.Materials and methodsAll mammography screenings performed between August 4, 2014, and August 15, 2018, in the Region of Southern Denmark with follow-up within 24 months were eligible. Screenings were assessed as normal or abnormal by breast radiologists through double reading with arbitration. For an AI decision of normal or abnormal, two AI-score cut-off points were applied by matching at mean sensitivity (AIsens) and specificity (AIspec) of first readers. Accuracy measures were sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and recall rate (RR).ResultsThe sample included 249,402 screenings (149,495 women) and 2033 breast cancers (72.6% screen-detected cancers, 27.4% interval cancers). AIsens had lower specificity (97.5% vs 97.7%; p < 0.0001) and PPV (17.5% vs 18.7%; p = 0.01) and a higher RR (3.0% vs 2.8%; p < 0.0001) than first readers. AIspec was comparable to first readers in terms of all accuracy measures. Both AIsens and AIspec detected significantly fewer screen-detected cancers (1166 (AIsens), 1156 (AIspec) vs 1252; p < 0.0001) but found more interval cancers compared to first readers (126 (AIsens), 117 (AIspec) vs 39; p < 0.0001) with varying types of cancers detected across multiple subgroups.ConclusionStandalone AI can detect breast cancer at an accuracy level equivalent to the standard of first readers when the AI threshold point was matched at first reader specificity. However, AI and first readers detected a different composition of cancers.Clinical relevance statementReplacing first readers with AI with an appropriate cut-off score could be feasible. AI-detected cancers not detected by radiologists suggest a potential increase in the number of cancers detected if AI is implemented to support double reading within screening, although the clinicopathological characteristics of detected cancers would not change significantly.Key Points• Standalone AI cancer detection was compared to first readers in a double-read mammography screening population.• Standalone AI matched at first reader specificity showed no statistically significant difference in overall accuracy but detected different cancers.• With an appropriate threshold, AI-integrated screening can increase the number of detected cancers with similar clinicopathological characteristics.

Analysis of The Utilization of The Automatic Exposure Control (AEC) Feature in The Use of Deep Learning Breast Image Technology in Women's Mammogram Screening Examinations at Dharmais Cancer Hospital

Deep learning technology is useful for radiology specialists as double reading to help increase the accuracy of image interpretation results. One of the preparations for maximizing the use of this technology is using good-quality images as the source. The Automatic Exposure Control (AEC) feature, which functions to determine exposure factors automatically, is expected to help produce images with good and consistent quality so that deep learning technology can work more effectively. This research aims to determine the quality results of mammogram images taken using the AEC feature and to analyze the use of deep learning technology in evaluating mammogram images. This research method is retrospective by collecting 800 mammogram images randomly and anonymously. Three hundred images were tested, 500 were evaluated, and 250 were analyzed for image quality based on references related to applying AEC and assessing the contrast-to-noise ratio (CNR). Deep learning technology was analyzed by comparing the results of mammogram image evaluation using deep learning and the evaluation results of a radiology specialist. Deep learning technology analysis shows that 98% of mammograms have the same results as the radiology doctor's evaluation, and 2% have different results from the radiology doctor's evaluation where the image has a dense breast type. The image quality results in this research showed that 97.6% of the 250 samples taken using the AEC feature had good image quality, and 2.4% had poor image quality due to inappropriate breast positioning during the examination.

Mammography Breast Cancer Screening Triage Using Deep Learning: A UK Retrospective Study.

Background Breast screening enables early detection of cancers; however, most women have normal mammograms, resulting in repetitive and resource-intensive reading tasks. Purpose To investigate if deep learning (DL) algorithms can be used to triage mammograms by identifying normal results to reduce workload or flag cancers that may be overlooked. Materials and Methods In this retrospective study, three commercial DL algorithms were investigated using consecutive mammograms from two UK Breast Screening Program sites from January 2015 to December 2017 and January 2017 to December 2018 on devices from two mammography vendors. Normal mammograms with a 3-year follow-up and histopathologically proven cancer detected at screening, the subsequent round, or in the 3-year interval were included. Two algorithm thresholds were set: in scenario A, 99.0% sensitivity for rule-out triage to a lone reader, and in scenario B, approximately 1.0% additional recall providing a rule-in triage for further assessment. Both thresholds were then applied to the screening workflow in scenario C. The sensitivity and specificity were used to assess the overall predictive performance of each DL algorithm. Results The data set comprised 78 849 patients (median age, 59 years [IQR, 53-63 years]) and 887 screening-detected, 439 interval, and 688 subsequent screening round-detected cancers. In scenario A (rule-out triage), models DL-1, DL-2, and DL-3 triaged 35.0% (27 565 of 78 849), 53.2% (41 937 of 78 849), and 55.6% (43 869 of 78 849) of mammograms, respectively, with 0.0% (0 of 887) to 0.1% (one of 887) of screening-detected cancers undetected. In scenario B, DL algorithms triaged in 4.6% (20 of 439) to 8.2% (36 of 439) of interval and 5.2% (36 of 688) to 6.1% (42 of 688) of subsequent-round cancers when applied after the routine double-reading workflow. Combining both approaches in scenario C resulted in an overall noninferior specificity (difference, -0.9%; P < .001) and superior sensitivity (difference, 2.7%; P < .001) for the adaptive workflow compared with routine double reading for all three algorithms. Conclusion Rule-out and rule-in DL-adapted triage workflows can improve the efficiency and efficacy of mammography breast cancer screening. © RSNA, 2023 Supplemental material is available for this article. See also the editorial by Nishikawa and Lu in this issue.

Discrepancies between Radiology Specialists and Residents in Fracture Detection from Musculoskeletal Radiographs.

(1) Background: The aim of this study was to compare the competence in appendicular trauma radiograph image interpretation between radiology specialists and residents. (2) Methods: In this multicenter retrospective cohort study, we collected radiology reports from radiology specialists (N = 506) and residents (N = 500) during 2018-2021. As a reference standard, we used the consensus of two subspecialty-level musculoskeletal (MSK) radiologists, who reviewed all original reports. (3) Results: A total of 1006 radiograph reports were reviewed by the two subspecialty-level MSK radiologists. Out of the 1006 radiographs, 41% were abnormal. In total, 67 radiographic findings were missed (6.7%) and 31 findings were overcalled (3.1%) in the original reports. Sensitivity, specificity, positive predictive value, and negative predictive value were 0.86, 0.92, 0.91 and 0.88 respectively. There were no statistically significant differences between radiology specialists' and residents' competence in interpretation (p = 0.44). However, radiology specialists reported more subtle cases than residents did (p = 0.04). There were no statistically significant differences between errors made in the morning, evening, or night shifts (p = 0.57). (4) Conclusions: This study found a lack of major discrepancies between radiology specialists and residents in radiograph interpretation, although there were differences between MSK regions and in subtle or obvious radiographic findings. In addition, missed findings found in this study often affected patient treatment. Finally, there are MSK regions where the sensitivity or specificity is below 90%, and these should raise concerns and highlight the need for double reading and should be taken into consideration in radiology education.

Can we predict discordant RECIST 1.1 evaluations in double read clinical trials?

In lung clinical trials with imaging, blinded independent central review with double reads is recommended to reduce evaluation bias and the Response Evaluation Criteria In Solid Tumor (RECIST) is still widely used. We retrospectively analyzed the inter-reader discrepancies rate over time, the risk factors for discrepancies related to baseline evaluations, and the potential of machine learning to predict inter-reader discrepancies. We retrospectively analyzed five BICR clinical trials for patients on immunotherapy or targeted therapy for lung cancer. Double reads of 1724 patients involving 17 radiologists were performed using RECIST 1.1. We evaluated the rate of discrepancies over time according to four endpoints: progressive disease declared (PDD), date of progressive disease (DOPD), best overall response (BOR), and date of the first response (DOFR). Risk factors associated with discrepancies were analyzed, two predictive models were evaluated. At the end of trials, the discrepancy rates between trials were not different. On average, the discrepancy rates were 21.0%, 41.0%, 28.8%, and 48.8% for PDD, DOPD, BOR, and DOFR, respectively. Over time, the discrepancy rate was higher for DOFR than DOPD, and the rates increased as the trial progressed, even after accrual was completed. It was rare for readers to not find any disease, for less than 7% of patients, at least one reader selected non-measurable disease only (NTL). Often the readers selected some of their target lesions (TLs) and NTLs in different organs, with ranges of 36.0-57.9% and 60.5-73.5% of patients, respectively. Rarely (4-8.1%) two readers selected all their TLs in different locations. Significant risk factors were different depending on the endpoint and the trial being considered. Prediction had a poor performance but the positive predictive value was higher than 80%. The best classification was obtained with BOR. Predicting discordance rates necessitates having knowledge of patient accrual, patient survival, and the probability of discordances over time. In lung cancer trials, although risk factors for inter-reader discrepancies are known, they are weakly significant, the ability to predict discrepancies from baseline data is limited. To boost prediction accuracy, it would be necessary to enhance baseline-derived features or create new ones, considering other risk factors and looking into optimal reader associations.

Multiparametric Magnetic Resonance Imaging in Prostate Cancer Screening at the Age of 45 Years: Results from the First Screening Round of the PROBASE Trial

BackgroundMagnetic resonance imaging (MRI) has been suggested as a tool for guiding biopsy recommendations in prostate cancer (PC) screening. ObjectiveTo determine the performance of multiparametric MRI (mpMRI) in young men at age 45 yr who participated in a PC screening trial (PROBASE) on the basis of baseline prostate-specific antigen (PSA). Design, setting, and participantsParticipants with confirmed PSA ≥3 ng/ml were offered mpMRI followed by MRI/transrectal ultrasound fusion biopsy (FBx) with targeted and systematic cores. mpMRI scans from the first screening round for men randomised to an immediate PSA test in PROBASE were evaluated by local readers and then by two reference radiologists (experience >10 000 prostate MRI examinations) blinded to the histopathology. The PROBASE trial is registered as ISRCTN37591328 Outcome measurements and statistical analysisThe local and reference Prostate Imaging-Data and Reporting System (PI-RADS) scores were compared, and the sensitivity, negative predictive value (NPV), and accuracy were calculated for both readings for different cutoffs (PI-RADS 3 vs 4). Results and limitationsOf 186 participants, 114 underwent mpMRI and FBx. PC was detected in 47 (41%), of whom 33 (29%) had clinically significant PC (csPC; International Society of Urological Pathology grade group ≥2). Interobserver reliability between local and reference PI-RADS scores was moderate (k = 0.41). At a cutoff of PI-RADS 4, reference reading showed better performance for csPC detection (sensitivity 79%, NPV 91%, accuracy of 85%) than local reading (sensitivity 55%, NPV 80%, accuracy 68%). Reference reading did not miss any PC cases for a cutoff of PI-RADS <3. If PI-RADS ≥4 were to be used as a biopsy cutoff, mpMRI would reduce negative biopsies by 68% and avoid detection of nonsignificant PC in 71% of cases. ConclusionsProstate MRI in a young screening population is difficult to read. The MRI accuracy of for csPC detection is highly dependent on reader experience, and double reading might be advisable. More data are needed before MRI is included in PC screening for men at age 45 yr. Patient summaryMeasurement of prostate specific antigen (PSA) is an effective screening test for early detection of prostate cancer (PC) and can reduce PC-specific deaths, but it can also lead to unnecessary biopsies and treatment. Magnetic resonance imaging (MRI) after a positive PSA test has been proposed as a way to reduce the number of biopsies, with biopsy only recommended for men with suspicious MRI findings. Our results indicate that MRI accuracy is moderate for men aged 45 years but can be increased by a second reading of the images by expert radiologists. For broad application of MRI in routine screening, double reading may be advisable.

Optimizing the Pairs of Radiologists That Double Read Screening Mammograms.

Background Despite variation in performance characteristics among radiologists, the pairing of radiologists for the double reading of screening mammograms is performed randomly. It is unknown how to optimize pairing to improve screening performance. Purpose To investigate whether radiologist performance characteristics can be used to determine the optimal set of pairs of radiologists to double read screening mammograms for improved accuracy. Materials and Methods This retrospective study was performed with reading outcomes from breast cancer screening programs in Sweden (2008-2015), England (2012-2014), and Norway (2004-2018). Cancer detection rates (CDRs) and abnormal interpretation rates (AIRs) were calculated, with AIR defined as either reader flagging an examination as abnormal. Individual readers were divided into performance categories based on their high and low CDR and AIR. The performance of individuals determined the classification of pairs. Random pair performance, for which any type of pair was equally represented, was compared with the performance of specific pairing strategies, which consisted of pairs of readers who were either opposite or similar in AIR and/or CDR. Results Based on a minimum number of examinations per reader and per pair, the final study sample consisted of 3 592 414 examinations (Sweden, n = 965 263; England, n = 837 048; Norway, n = 1 790 103). The overall AIRs and CDRs for all specific pairing strategies (Sweden AIR range, 45.5-56.9 per 1000 examinations and CDR range, 3.1-3.6 per 1000; England AIR range, 68.2-70.5 per 1000 and CDR range, 8.9-9.4 per 1000; Norway AIR range, 81.6-88.1 per 1000 and CDR range, 6.1-6.8 per 1000) were not significantly different from the random pairing strategy (Sweden AIR, 54.1 per 1000 examinations and CDR, 3.3 per 1000; England AIR, 69.3 per 1000 and CDR, 9.1 per 1000; Norway AIR, 84.1 per 1000 and CDR, 6.3 per 1000). Conclusion Pairing a set of readers based on different pairing strategies did not show a significant difference in screening performance when compared with random pairing. © RSNA, 2023.

Video telescope operating monitor-assisted surgery is equivalent to conventional surgery in treatment of cervical intervertebral disc herniation in dogs.

To compare the use of the video telescope operating monitor (VITOM) and use of a conventional unassisted surgical method for treatment of cervical intervertebral disc herniation in dogs. 39 dogs with cervical intervertebral disc disease. Prospective study. Dogs were prospectively nonrandomly assigned to either the VITOM (n = 19) or conventional surgery (20) group depending on VITOM system availability. Signalment and preoperative neurologic status were recorded for all dogs. Preoperative and postoperative CT myelography was performed to compare intervertebral space location, spinal cord dimensions at the decompression level, ventral slot dimensions, and residual disc material. Surgical complications and postoperative neurologic outcomes were recorded. Data were compared between the 2 groups using fixed-effects or mixed-effects models to consider double reading of CT myelography images. No significant differences were noted between the 2 groups regarding the decompression ratio (P = .85), vertebral length body ratio (P = .13), ventral slot width ratio (P = .39), residual disc material (P = .30), and sinus bleeding (P = .12). No significant differences were found between the 2 groups regarding postoperative neurologic grade (P = .17). VITOM-assisted ventral slot decompression is equivalent to conventional surgery in treatment of cervical intervertebral disc herniation in dogs. The use of VITOM remains a good alternative to the conventional surgical method.

Artificial intelligence for breast cancer detection in screening mammography in Sweden: a prospective, population-based, paired-reader, non-inferiority study

Artificial intelligence (AI) as an independent reader of screening mammograms has shown promise, but there are few prospective studies. Our aim was to conduct a prospective clinical trial to examine how AI affects cancer detection and false positive findings in a real-world setting. ScreenTrustCAD was a prospective, population-based, paired-reader, non-inferiority study done at the Capio Sankt Göran Hospital in Stockholm, Sweden. Consecutive women without breast implants aged 40-74 years participating in population-based screening in the geographical uptake area of the study hospital were included. The primary outcome was screen-detected breast cancer within 3 months of mammography, and the primary analysis was to assess non-inferiority (non-inferiority margin of 0·15 relative reduction in breast cancer diagnoses) of double reading by one radiologist plus AI compared with standard-of-care double reading by two radiologists. We also assessed single reading by AI alone and triple reading by two radiologists plus AI compared with standard-of-care double reading by two radiologists. This study is registered with ClinicalTrials.gov, NCT04778670. From April 1, 2021, to June 9, 2022, 58 344 women aged 40-74 years underwent regular mammography screening, of whom 55 581 were included in the study. 269 (0·5%) women were diagnosed with screen-detected breast cancer based on an initial positive read: double reading by one radiologist plus AI was non-inferior for cancer detection compared with double reading by two radiologists (261 [0·5%] vs 250 [0·4%] detected cases; relative proportion 1·04 [95% CI 1·00-1·09]). Single reading by AI (246 [0·4%] vs 250 [0·4%] detected cases; relative proportion 0·98 [0·93-1·04]) and triple reading by two radiologists plus AI (269 [0·5%] vs 250 [0·4%] detected cases; relative proportion 1·08 [1·04-1·11]) were also non-inferior to double reading by two radiologists. Replacing one radiologist with AI for independent reading of screening mammograms resulted in a 4% higher non-inferior cancer detection rate compared with radiologist double reading. Our study suggests that AI in the study setting has potential for controlled implementation, which would include risk management and real-world follow-up of performance. Swedish Research Council, Swedish Cancer Society, Region Stockholm, and Lunit.

Опыт организации референс-центра по двойному прочтению маммографических снимков в Красноярском крае

Early detection of breast cancer with the help of systematic mammographic screening makes it possible to identify lesions whose treatment is more effective and, in general, more favorable for the quality of life. The further tactics of diagnosis and treatment depend on the timeliness and correctness of the diagnosis. The quality control of the research should be carried out by experts of specialized reference centers. The article presents the experience of the Krasnoyarsk Territory on the double reading of mammograms.

An accurate measurement method of gear transmission error based on a double-angle measurement device

In this paper, a geometric model of a grating and reading head was established, where the relationship between the angle deviation, which was compensated for by double reading heads, and the actual angle is deduced. A new measurement method was proposed by applying the clock pulse counting method to the measurement of the transmission error of the gear pair input-and-output angle encoder while taking into account the gear pair’s transmission error. To demonstrate the feasibility of this measurement method, a transmission error test bench of involute gear pairs was built. Based on the output gear, the counted pulses were divided into integer pulses and non-integer pulses, and a transmission error curve was obtained by processing the pulses. The transmission errors of a helical gear with untrimmed angles and a gear pair with trimmed angles were also measured. Compared with the untrimmed gear pair, the first-order amplitude of the trimmed gear pair was reduced by 25.25%, 0.79%, and 16.9%, but other whole-order full-harmonic amplitudes were excited. However, the peak value of the single tooth surface of the trimmed gear pair was generally less than that of the untrimmed gear pair. The peak transmission error and the amplitude of the tooth harmonic frequency had good repeatability, within 16%, which verified the feasibility of the transmission error test method. The measurement precision reached 1 arcsecond in the tooth frequency.

Mammographic Breast Density Model Using Semi-Supervised Learning Reduces Inter-/Intra-Reader Variability.

Breast density is an important risk factor for breast cancer development; however, imager inconsistency in density reporting can lead to patient and clinician confusion. A deep learning (DL) model for mammographic density grading was examined in a retrospective multi-reader multi-case study consisting of 928 image pairs and assessed for impact on inter- and intra-reader variability and reading time. Seven readers assigned density categories to the images, then re-read the test set aided by the model after a 4-week washout. To measure intra-reader agreement, 100 image pairs were blindly double read in both sessions. Linear Cohen Kappa (κ) and Student's t-test were used to assess the model and reader performance. The model achieved a κ of 0.87 (95% CI: 0.84, 0.89) for four-class density assessment and a κ of 0.91 (95% CI: 0.88, 0.93) for binary non-dense/dense assessment. Superiority tests showed significant reduction in inter-reader variability (κ improved from 0.70 to 0.88, p ≤ 0.001) and intra-reader variability (κ improved from 0.83 to 0.95, p ≤ 0.01) for four-class density, and significant reduction in inter-reader variability (κ improved from 0.77 to 0.96, p ≤ 0.001) and intra-reader variability (κ improved from 0.89 to 0.97, p ≤ 0.01) for binary non-dense/dense assessment when aided by DL. The average reader mean reading time per image pair also decreased by 30%, 0.86 s (95% CI: 0.01, 1.71), with six of seven readers having reading time reductions.