Search

Background: In patients with Cushing's disease, the preoperative identification of pituitary adenomas is crucial to treatment. However, increasing diagnostic accuracy remains an unresolved issue. Purpose: To evaluate the diagnostic accuracy and the impact of readers' experience regarding high‐resolution contrast‐enhanced magnetic resonance imaging (hrMRI) for identifying pituitary adenomas in comparison with conventional contrast‐enhanced MRI (cMRI) and dynamic contrast‐enhanced MRI (dMRI). Study Type: Retrospective. Population: Sixty‐five patients (median age, 39 years; interquartile range [IQR], 28–53 years; 60% females) with treatment‐naïve Cushing's disease. Field Strength/Sequence: 3‐T, seven fast spin echo sequences. Assessment: The diagnostic accuracies of identifying pituitary adenomas on cMRI, dMRI, combined cMRI and dMRI (cdMRI), and hrMRI were independently evaluated by six readers with three experience levels (high: >20 years, modest: 10–20 years, low: <10 years; two readers for each experience level). Readers were asked to localize the lesion, and measure its diameter on the sequence where identified. The reference standard was postoperative histopathology. The impact of readers' experience and interobserver agreement were assessed. Image quality was assessed using a 5‐point Likert scale, including overall image quality, sharpness, and structural conspicuity. Statistical Tests: McNemar's test, Cochran's test, Wilcoxon signed‐rank test, Mann–Whitney U test, and κ statistics for interobserver agreement. A P‐value <0.05 was considered statistically significant. Results: For identifying pituitary adenomas (median diameter, 5 mm; IQR, 4–5 mm), hrMRI had significantly higher sensitivity (87.7%–93.8%) than cMRI, dMRI, and cdMRI (52.3%–75.4%) for readers with different experience levels. The interobserver agreement was moderate (κ = 0.461–0.523). The sensitivity for hrMRI was comparable between readers with different experience levels (P = 0.371). All image quality scores on hrMRI were significantly higher than cMRI and dMRI (5.0 vs. 4.0). Data Conclusion: For identifying pituitary adenomas in patients with treatment‐naïve Cushing's disease, hrMRI may show high diagnostic accuracy and seems not to be affected by readers' experience. Level of Evidence: 3 Technical Efficacy: Stage 2 [ABSTRACT FROM AUTHOR]

The use of endoscopic images for the accurate assessment of ulcerative colitis (UC) severity is crucial to determining appropriate treatment. However, experts may interpret these images differently, leading to inconsistent diagnoses. This study aims to address the issue by introducing a standardization method based on deep learning. We collected 254 rectal endoscopic images from 115 patients with UC, and five experts in endoscopic image interpretation assigned classification labels based on the Ulcerative Colitis Endoscopic Index of Severity (UCEIS) scoring system. Interobserver variance analysis of the five experts yielded an intraclass correlation coefficient of 0.8431 for UCEIS scores and a kappa coefficient of 0.4916 when the UCEIS scores were transformed into UC severity measures. To establish a consensus, we created a model that considered only the images and labels on which more than half of the experts agreed. This consensus model achieved an accuracy of 0.94 when tested with 50 images. Compared with models trained from individual expert labels, the consensus model demonstrated the most reliable prediction results. [ABSTRACT FROM AUTHOR]

Background: 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. Materials and methods: 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. Results: 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. Conclusion: 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. [ABSTRACT FROM AUTHOR]

BackgroundIn 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.Materials and methodsWe 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.ResultsAt 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.ConclusionPredicting 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.

Purpose: Task automation is essential for efficient and consistent image segmentation in radiation oncology. We report on a deep learning architecture, comprising a U‐Net and a variational autoencoder (VAE) for automatic contouring of the prostate gland incorporating interobserver variation for radiotherapy treatment planning. The U‐Net/VAE generates an ensemble set of segmentations for each image CT slice. A novel outlier mitigation (OM) technique was implemented to enhance the model segmentation accuracy. Methods: The primary source dataset (source_prim) consisted of 19 200 CT slices (from 300 patient planning CT image datasets) with manually contoured prostate glands. A smaller secondary source dataset (source_sec) comprised 640 CT slices (from 10 patient CT datasets), where prostate glands were segmented by 5 independent physicians on each dataset to account for interobserver variability. Data augmentation via random rotation (<5 degrees), cropping, and horizontal flipping was applied to each dataset to increase sample size by a factor of 100. A probabilistic hierarchical U‐Net with VAE was implemented and pretrained using the augmented source_prim dataset for 30 epochs. Model parameters of the U‐Net/VAE were fine‐tuned using the augmented source_sec dataset for 100 epochs. After the first round of training, outlier contours in the training dataset were automatically detected and replaced by the most accurate contours (based on Dice similarity coefficient, DSC) generated by the model. The U‐Net/OM‐VAE was retrained using the revised training dataset. Metrics for comparison included DSC, Hausdorff distance (HD, mm), normalized cross‐correlation (NCC) coefficient, and center‐of‐mass (COM) distance (mm). Results: Results for U‐Net/OM‐VAE with outliers replaced in the training dataset versus U‐Net/VAE without OM were as follows: DSC = 0.82 ± 0.01 versus 0.80 ± 0.02 (p = 0.019), HD = 9.18 ± 1.22 versus 10.18 ± 1.35 mm (p = 0.043), NCC = 0.59 ± 0.07 versus 0.62 ± 0.06, and COM = 3.36 ± 0.81 versus 4.77 ± 0.96 mm over the average of 15 contours. For the average of 15 highest accuracy contours, values were as follows: DSC = 0.90 ± 0.02 versus 0.85 ± 0.02, HD = 5.47 ± 0.02 versus 7.54 ± 1.36 mm, and COM = 1.03 ± 0.58 versus 1.46 ± 0.68 mm (p < 0.03 for all metrics). Results for the U‐Net/OM‐VAE with outliers removed were as follows: DSC = 0.78 ± 0.01, HD = 10.65 ± 1.95 mm, NCC = 0.46 ± 0.10, COM = 4.17 ± 0.79 mm for the average of 15 contours, and DSC = 0.88 ± 0.02, HD = 7.00 ± 1.17 mm, COM = 1.58 ± 0.63 mm for the average of 15 highest accuracy contours. All metrics for U‐Net/VAE trained on the source_prim and source_sec datasets via pretraining, followed by fine‐tuning, show statistically significant improvement over that trained on the source_sec dataset only. Finally, all metrics for U‐Net/VAE with or without OM showed statistically significant improvement over those for the standard U‐Net. Conclusions: A VAE combined with a hierarchical U‐Net and an OM strategy (U‐Net/OM‐VAE) demonstrates promise toward capturing interobserver variability and produces accurate prostate auto‐contours for radiotherapy planning. The availability of multiple contours for each CT slice enables clinicians to determine trade‐offs in selecting the "best fitting" contour on each CT slice. Mitigation of outlier contours in the training dataset improves prediction accuracy, but one must be wary of reduction in variability in the training dataset. [ABSTRACT FROM AUTHOR]

Atlas-based auto-segmentation (ABS) procedure used in radiotherapy can be classified into two groups, one using one atlas per patient (sSM) and the other using multiple atlases (sMM). This study evaluated auto-contouring accuracy and contouring time in patients with prostate cancer using the two procedures. The Dice similarity coefficient of sMM was significantly better than that of sSM (prostate [median, 0.81 (range, 0.66–0.91) vs. 0.64 (0.27–0.71), p

Introduction: Oral squamous cell carcinoma is usually preceded by potentially malignant disorders (PMDs), the most common being oral leukoplakia. A conservative management protocol is followed for milder dysplastic cases, while severe dysplastic lesions are surgically excised. Several classification systems have been developed based mainly on histopathological features with a lot of inter-observer variations. The present study was done to assess the use of immunohistochemical (IHC) markers in grading leukoplakic lesions in addition to histopathological grading. Aims and Objectives: To grade leukoplakia using different grading systems and assess if IHC markers can aid in categorizing leukoplakia. Materials and Methods: Thirty-five cases of leukoplakia were graded using Ljubljana, 2005 World Health Organization (WHO), and Binary System followed by IHC staining with p63 and CD31. Results and Statistics: Variation was noted in 12 cases while using WHO, 11 using Ljubljana, and 7 using Binary System and was significant on Cohen-Kappa statistics, with the least significant variation noted on Binary System. p63 staining assisted to group doubtful cases and even identify variation in cases graded positively on histopathology. In total, 17 cases stained one-third (mild/low), while 15 cases stained one-half or more (higher grade) epithelial thickness. A weak correlation was observed between all grading systems and p63 on Kendall's Tau-b analysis and the weak correlation was significant for the WHO and binary grading systems. Analyzing p63 and CD31 using Kruskal–Wallis test, an increase in mean vessel density (MVD) was noted for mild/moderate cases but decreased for severe cases. Conclusion: Addressing histological categorization of PMDs assisted by IHC markers to understand the biological behavior of the tissues is currently essential with studies on more markers to assist in the management protocol.

Background and purpose: Peer-review of Target Volume (TV) and Organ at Risk (OAR) contours in radiotherapy planning are typically conducted visually; this can be time consuming and subject to interobserver variation. This study investigated automatic evaluation of contouring using conformity indices and supervised machine learning. Methods: A total of 393 contours from 253 Stereotactic Ablative Body Radiotherapy (SABR) benchmark cases (adrenal gland, liver, pelvic lymph node and spine), delineated by 132 clinicians from 25 centres, were visually evaluated for conformity against gold standard contours. Contours were scored as “pass” or “fail” on visual peer review and six Conformity Indices (CIs) were applied. CI values were mapped to pass/fail scores for each contour and used to train supervised machine learning models. A 5-fold cross validation method was employed to determine the predictive accuracies of each model. Results: The stomach structure produced models with the highest predictive accuracy overall (96% using Support Vector Machine and Ensemble models), whilst the liver GTV produced models with the lowest predictive accuracy (76% using Logistic Regression). Predictive accuracies across all models ranged from 68–96% (68–87% for TV and 71–96% for OARs). Conclusions: Although a final visual review by an experienced clinician is still required, the automatic contour evaluation method could reduce the time for benchmark case reviews by identifying gross contouring errors. This method could be successfully implemented to support departmental training and the continuous assessment of outlining for clinical staff in the peer-review process, to reduce interobserver variability in contouring and improve interpretation of radiological anatomy.

Introduction: The diagnosis of Primary Thyroid Lymphoma (PTL) and its distinction from Florid Lymphocytic Thyroiditis (FLT) is often difficult on Fine Needle Aspirate (FNA). The sensitivity is low for Low Grade Lymphoma (LGL) and the distinction of PTL is important as they are treated by chemotherapy rather than surgery. We studied the efficacy of FNA for thyroid lymphoproliferative disorders which includes thyroid lymphoma and lymphocytic thyroditis and detection of cytomorphological features differentiating between thyroiditis and lymphoma. Aim: To find out interobserver variation in cytological diagnosis of thyroid lymphomas and thyroiditis with evaluation of the cytomorphological pointers of lymphoma on thyroid cytology. Materials and Methods: Thirty-three cases of FNA thyroid with lymphoid proliferation in which histological confirmation with Immunohistochemistry (IHC) was available were retrieved and studied for various cytomorphological features by three pathologists. The findings were correlated with the final histopathology diagnosis. Results: Of the total 33 cases that had significant lymphoid infiltrate on thyroid FNA, 28 cases were confirmed as NonHodgkin’s lymphoma, three cases as thyroiditis and two case of carcinoma on histology. All the lymphoma cases were positive for CD20 antibody on histology. The percentage of cases of thyroid lymphoma was high in present study because only selected type of cases were included (FNA thyroid with lymphoid proliferation) and it does not refer to true incidence of thyroid lymphoma. Among three observers, two observers (observer 1&2) diagnosed 18 cases of HGL, two cases of FLT and one observer (observer 3) diagnosed 15 cases of HGL and one cases of FLT correctly on FNA, five cases of LGL were correctly diagnosed by all the three observers. The diagnostic accuracy was 76.1% with high specificity (80%) and sensitivity (94.7%) for HGL and high concordance among the three pathologists (Kappa= 0.8). LGL and FLT had low sensitivity (55.6% and 66.7%) and least interobserver correlation (kappa – 0.21). Monomorphism, large cells, absent plasma cells, macrophages were significantly associated with diagnosis of lymphoma. Conclusion: Cytological features of thyroid lymphoma are diagnostic in HGL, however the features are more non-specific in LGLs. High cellularity, relatively monomorphic population and frequent occurrence of lymphoglandular bodies were common in LGL compared to FLT. It is challenging to diagnose PTLs, especially low grade on FNA. For this reason, it should be kept in mind that patients with chronic lymphocytic thyroiditis should be evaluated carefully and possibility of LGL should not rule out on cytology. The role of FNA is limited but it is simple safe and cost effective for initial workup in centres lacking facility of flow cytometry and immunocytochemistry

• There is substantial agreement amongst professionals who interpret and perform prognostic SSEPs. • Agreement is lower in patients below 2 years of age, in those with poor outcome and when interpreting SSEPs as absent. • Clinicians using SSEPs to prognosticate poor outcome should be aware of the inter-rater variability amongst professionals. To measure inter- and intra-rater agreement in the interpretation of cortical somatosensory evoked potential (SSEP) components following paediatric cardiac arrest (CA) in multi-professional neurophysiology teams. Thirteen professionals blinded to patient outcome interpreted 96 SSEPs in paediatric patients 24-/48-/72-hours following CA. Of these, 34 were duplicates used to assess intra-rater agreement. Consistent interpretations (absent/present/indeterminate) between scientists (who record/identify SSEP components) and neurophysiologists (who provide prognostic SSEP interpretation) were expressed as percentages. Rates of agreement were calculated using Fleiss' kappa coefficient (K). Unanimous agreement between professionals was present in 40% (95%CI: 28–54%) of the interpreted SSEPs, with a K value of 0.62 (95%CI: 0.55–0.70) based on average agreement. Agreement was similar between neurophysiologists (K = 0.67; 95%CI: 0.57–0.77) and scientists (K = 0.62; 95%CI: 0.54–0.70) but lower in patients < 2 years old (K = 0.23; 95%CI: 0.14–0.33) and in those with poor outcome (K = 0.21; 95%CI: 0.07–0.35). No SSEP was unanimously interpreted as absent and 92% (95%CI: 89–95%) of duplicate SSEPs were interpreted consistently. Despite substantial agreement when interpreting prognostic SSEPs, this was significantly lower in children with poor outcome and of younger age. Clinicians using SSEPs in the intensive care unit should be aware of the inter-rater variability when interpreting SSEPs as absent. [ABSTRACT FROM AUTHOR]

Introduction: Oral squamous cell carcinoma is usually preceded by potentially malignant disorders (PMDs), the most common being oral leukoplakia. A conservative management protocol is followed for milder dysplastic cases, while severe dysplastic lesions are surgically excised. Several classification systems have been developed based mainly on histopathological features with a lot of inter-observer variations. The present study was done to assess the use of immunohistochemical (IHC) markers in grading leukoplakic lesions in addition to histopathological grading. Aims and Objectives: To grade leukoplakia using different grading systems and assess if IHC markers can aid in categorizing leukoplakia. Materials and Methods: Thirty-five cases of leukoplakia were graded using Ljubljana, 2005 World Health Organization (WHO), and Binary System followed by IHC staining with p63 and CD31. Results and Statistics: Variation was noted in 12 cases while using WHO, 11 using Ljubljana, and 7 using Binary System and was significant on Cohen-Kappa statistics, with the least significant variation noted on Binary System. p63 staining assisted to group doubtful cases and even identify variation in cases graded positively on histopathology. In total, 17 cases stained one-third (mild/low), while 15 cases stained one-half or more (higher grade) epithelial thickness. A weak correlation was observed between all grading systems and p63 on Kendall's Tau-b analysis and the weak correlation was significant for the WHO and binary grading systems. Analyzing p63 and CD31 using Kruskal–Wallis test, an increase in mean vessel density (MVD) was noted for mild/moderate cases but decreased for severe cases. Conclusion: Addressing histological categorization of PMDs assisted by IHC markers to understand the biological behavior of the tissues is currently essential with studies on more markers to assist in the management protocol. [ABSTRACT FROM AUTHOR]

Corkscrew claw (CC) in dairy cattle is increasingly reported in dairy herds. CC is a progressive deformity of the claw capsule with uncertain aetiology and pathogenesis. Genetics and specific environmental factors are suspected of contributing to the development of this irreversible condition. CC has been found in lame cows; however, the cause and effect has not been established. To perform analysis of risk factors, treatment and pathogenesis, a definition of severity scores is called for. The aim of this study was to measure and analyse CC characteristics from photos of cows' feet to describe and evaluate a scoring system for CC. Width of the visible part of the axial wall, degree of contact between the toe and the floor and angle of the distal part of the abaxial wall as a proxy for the deviation of the abaxial wall was measured from 393 pictures of CC. Based on the measurements on the claws, the parameter "width of the axial wall" was chosen to define the scores. The parameter was divided into three intervals to define either mild CC 0.3-2.0 cm, moderate CC 2.1-3.5 cm or severe CC>3.5 cm and correlation between the parameters; level of contact between the toe and the floor and the angle of the distal abaxial wall was evaluated. There was a significant positive linear correlation between width of the axial wall and angle of the distal part of the abaxial wall (r=0.91), the wider the axial wall, the more the abaxial wall deviated in the distal part. As the width of the axial wall increased the toe increasingly lost contact with the floor, this association was significant for mild CC and moderate CC but not for severe CC. The Interobserver agreement of the CC Scoring system was tested by 30 claw trimmers each scoring 32 cadaver feet and by 2 trained observers on 28 photos of feet using Cohen´s weighted kappa and showed substantial to almost perfect agreement between untrained and trained observers, respectively., Competing Interests: Declaration of Competing Interest None of the authors has any other financial or personal relationships that could inappropriately influence or bias the content of the paper., (Copyright © 2024 The Authors. Published by Elsevier Ltd.. All rights reserved.)

Aims: Thymic tumours are rare in routine pathology practice. Although the World Health Organization (WHO) classification describes a number of well‐defined categories, the classification remains challenging. The aim of this study was to investigate the reproducibility of the WHO classification among a large group of international pathologists with expertise in thymic pathology and by using whole slide imaging to facilitate rapid diagnostic turnover. Methods and results: Three hundred and five tumours, consisting of 90 biopsies and 215 resection specimens, were reviewed with a panel‐based virtual microscopy approach by a group of 13 pathologists with expertise in thymic tumours over a period of 6 years. The specimens were classified according to the WHO 2015 classification. The data were subjected to statistical analysis, and interobserver concordance (Fleiss kappa) was calculated. All cases were diagnosed within a time frame of 2 weeks. The overall level of agreement was substantial (κ = 0.6762), and differed slightly between resection specimens (κ = 0.7281) and biopsies (κ = 0.5955). When analysis was limited to thymomas only, and they were grouped according to the European Society for Medical Oncology Clinical Practice Guidelines into B2, B3 versus A, AB, B1 and B3 versus A, AB, B1, B2, the level of agreement decreased slightly (κ = 0.5506 and κ = 0.4929, respectively). Difficulties arose in distinguishing thymoma from thymic carcinoma. Within the thymoma subgroup, difficulties in distinction were seen within the B group. Conclusions: Agreement in diagnosing thymic lesions is substantial when they are assessed by pathologists with experience of these rare tumours. Digital pathology decreases the turnaround time and facilitates access to what is essentially a multinational resource. This platform provides a template for dealing with rare tumours for which expertise is sparse. [ABSTRACT FROM AUTHOR]

Purpose: To quantify the interobserver variability among the most frequently encountered parenchymal patterns in High Resolution CT (HRCT) and to compare the interobserver variability in the application of the 2011 and 2018 usual interstitial pneumonia (UIP) criteria according to the joint guidelines from international thoracic and respiratory societies. Material and methods: Two observers independently evaluated 126 HRCT, with examples of most common parenchymal patterns, and noted the presence of each pattern. The readers also noted whether the findings met the 2011 criteria for UIP. In a second reading, the same readers noted whether the HRCT met the UIP criteria according to the 2018 UIP update. Results: The kappa values for interobserver variability for the different patterns ranged from 0.28 (intralobular lines) to 0.85 (tree-in-bud nodules). The kappa value for UIP pattern was similar for 2011 and 2018 criteria, 0.58 and 0.69, respectively. Compared to the 2011 UIP criteria, there was no statistically significant difference in the number of HRCT classified as UIP using the 2018 criteria. Conclusions: There is a substantial variation in interobserver agreement between the different parenchymal patterns, which suggests that some patterns a more easily identified than others. There is also a considerable reader variation in the assessment of UIP applying the 2011 UIP criteria as well as applying the 2018 UIP update.

Background & Aims: The American Association for the Study of Liver Diseases recommends the use of a 2‐grade classification system (small and large) to describe the size of oesophageal varices (OV). Data on observer agreement (OA) on this system are currently lacking. We aimed to evaluate this classification and compare it to the widely used 3‐grade classification (grade 1 'small', grade 2 'medium', grade 3 'large') among operators of variable experience. Methods: High‐definition video recordings of 100 patients with cirrhosis were prospectively collected using standardised criteria. Nine observers of variable experience performed independent evaluations of the videos in random order. OV were scored using both systems. All assessments were repeated a year later by the same observers to assess intra‐observer agreement. Results: Interobserver agreement (all observers) using the 2‐grade and the 3‐grade system was k = 0.71 (95% CI: 0.64‐0.78) and k = 0.73 (95% CI: 0.66‐0.79) respectively. When using the 2‐grade system, intra‐observer agreement between hepatologists (n = 3), luminal gastroenterologists (n = 3) and trainee gastroenterologists (n = 3) was k = 0.89 (95% CI: 0.86‐0.91), k = 0.72 (95% CI: 0.67‐0.77), and k = 0.74 (95% CI: 0.67‐0.8) respectively. With the 3‐grade system; intra‐observer agreement between the same three subgroups were k = 0.9 (95% CI: 0.87‐0.92), k = 0.73 (95% CI: 0.68‐0.78), k = 0.77 (95% CI: 0.71‐0.82) respectively. Conclusions: There was no difference in OA between the 2‐grade and 3‐grade classification systems. Hepatologists had significantly higher levels of consistency in grading OV. This may have implications to create alternative training models for residents and fellows in the recognition and grading of OV. See Editorial on Page 754 [ABSTRACT FROM AUTHOR]

Purpose: There remain uncertainties due to inter‐ and intraobserver variability in soft‐tissue‐based patient positioning even with the use of image‐guided radiation therapy (IGRT). This study aimed to reveal observer uncertainties of soft‐tissue‐based patient positioning on cone‐beam computed tomography (CBCT) images for prostate cancer IGRT. Methods: Twenty‐six patients (7–8 fractions/patient, total number of 204 fractions) who underwent IGRT for prostate cancer were selected. Six radiation therapists retrospectively measured prostate cancer location errors (PCLEs) of soft‐tissue‐based patient positioning between planning CT (pCT) and pretreatment CBCT (pre‐CBCT) images after automatic bone‐based registration. Observer uncertainties were evaluated based on residual errors, which denoted the differences between soft‐tissue and reference positioning errors. Reference positioning errors were obtained as PCLEs of contour‐based patient positioning between pCT and pre‐CBCT images. Intraobserver variations were obtained from the difference between the first and second soft‐tissue‐based patient positioning repeated by the same observer for each fraction. Systematic and random errors of inter‐ and intraobserver variations were calculated in anterior–posterior (AP), superior–inferior (SI), and left–right (LR) directions. Finally, clinical target volume (CTV)‐to‐planning target volume (PTV) margins were obtained from systematic and random errors of inter‐ and intraobserver variations in AP, SI, and LR directions. Results: Interobserver variations in AP, SI, and LR directions were 0.9, 0.9, and 0.5 mm, respectively, for the systematic error, and 1.8, 2.2, and 1.1 mm, respectively, for random error. Intraobserver variations were <0.2 mm in all directions. CTV‐to‐PTV margins in AP, SI, and LR directions were 3.5, 3.8, and 2.1 mm, respectively. Conclusion: Intraobserver variability was sufficiently small and would be negligible. However, uncertainties due to interobserver variability for soft‐tissue‐based patient positioning using CBCT images should be considered in CTV‐to‐PTV margins. [ABSTRACT FROM AUTHOR]

Purpose: To investigate the interobserver variations in delineation of lumpectomy cavity (LC) and clinical target volume (CTV), and its impact on irradiated volume in accelerated partial breast irradiation using intraoperative multicatheter brachytherapy. Material and methods : Delineation of LC and CTV was done by five radiation oncologists on planning computed tomography (CT) scans of 20 patients with intraoperative interstitial breast implant. Cavity visualization index (CVI), four-point index ranging from (0 = poor) to (3 = excellent) was created and assigned by observers for each patient. In total, 200 contours for all observers and 100 treatment plans were evaluated. Spatial concordance (conformity index, CI common , and CIgen), average shift in the center of mass (COM), and ratio of maximum and minimum volumes (V max /V min ) of LC and CTV were quantified among all observers and statistically analyzed. Variation in active dwell positions (0.5 cm step) for each catheter, total reference air kerma (TRAK), volume enclosed by prescription isodose (V100%) among observers and its spatial concordance were analyzed. Results : The mean ± SD CI common of LC and CTV was 0.54 ± 0.09, and 0.58 ± 0.08, respectively. Conformity index tends to increase, shift in COM and V max /V min decrease significantly (p < 0.05), as CVI increased. Out of total 309 catheters, 29.8% catheters had no change, 29.8% and 17.5% catheters had variations of 1 and 2 dwell positions (0.5 cm and 1 cm), respectively. 9.3% catheters shown variations ≥ 10 dwell positions (5 cm). The mean ± SD CI common of V100% was 0.75 ± 0.11. The mean observed V max /V min of prescription isodose and TRAK was 1.18 (range, 1.03 to 1.56) and 1.11 (range, 1.03 to 1.35), respectively. Conclusions : Interobserver variability in delineation of target volume was found to be significantly related to CVI. Smaller variability was observed with excellent visualization of LC. Interobserver variations showed dosimetric impact on irradiation of breast tissue volume with prescription dose.

Objectives: To determine the utility of phosphohistone H3 (PHH3) mitotic count (MC) in grading follicular lymphoma (FL).Methods: FL cases were identified (2005-2017). Three hematopathologists recorded their average Ki-67 proliferation index, MC/high-power field (hpf) using PHH3 and H&E stains, and number of centroblasts/hpf. Results were assessed for correlations and interobserver variability.Results: Forty-three cases of FL were studied. PHH3 MC resulted in the strongest correlation to grade (r = 0.701, P < .0001) compared with Ki-67 proliferation index (PI) (r = 0.681, P < .0001) and H&E MC (r = 0.536, P = .0002) and the strongest linear relationship to centroblast count (R2 = 0.453). Agreement among pathologists was strongest for PHH3 (intraclass correlation coefficient [ICC] = 0.86) followed by Ki-67 PI (ICC = 0.85) and H&E MC (ICC = 0.78).Conclusions: PHH3 correlates best to histologic grade and has less interobserver variability compared with Ki-67 PI and H&E MC. These results support using PHH3 as an adjunct in FL grading. [ABSTRACT FROM AUTHOR]

A good classification system is important for clinical handoffs, research, and clinical treatment guidelines. A reliable classification system shows good interobserver and intraobserver agreement. This study analyzed the interobserver and intraobserver agreement of a descriptive system for ankle fractures and the Lauge-Hansen classification. Three groups of observers (experts, semiexperts, and novices) scored a total of 20 ankle radiographs. All ankle radiographs were classified according to the Lauge-Hansen and Danis-Weber classifications. The ankle fractures were subsequently reviewed in a descriptive manner for the following features: number of affected malleoli, type of fracture of the lateral and medial malleolus, and congruence of the ankle joint. After 2 weeks, the same set of radiographs were reviewed. For interobserver and intraobserver variability, the separate groups were used for analysis, and the Fleiss (multirater) κ values were calculated. The interobserver agreement for the Lauge-Hansen classification was moderate for the experts, fair for semiexperts, and slight for novices (κ = 0.45, κ = 0.37, and κ = 0.16). All factors of the descriptive system had better interobserver agreement than the Lauge-Hansen classification, except for the agreement on the type of fracture of the lateral malleolus. The intraobserver agreement of the Lauge-Hansen classification was substantial for the experts, moderate for the semiexperts, and fair for the novice observers (κ = 0.70, κ = 0.49, and κ = 0.26). The intraobserver agreement was better for all factors of the descriptive system compared with the Lauge-Hansen classification. The descriptive system presented in this study shows less variability between observers than the Lauge-Hansen classification. This system has clinical implications and is easy to use for clinicians with mixed levels of experience. It has the potential to improve clinical and research handoffs and overcome the limitations of current classification systems. [ABSTRACT FROM AUTHOR]

Background: Compared to 1.5 T, 3 T magnetic resonance imaging (MRI) increases signal-to-noise ratio leading to improved image quality. However, its clinical relevance in clinically isolated syndrome suggestive of multiple sclerosis remains uncertain. Objectives: The purpose of this study was to investigate how 3 T MRI affects the agreement between raters on lesion detection and diagnosis. Methods: We selected 30 patients and 10 healthy controls from our ongoing prospective multicentre cohort. All subjects received baseline 1.5 and 3 T brain and spinal cord MRI. Patients also received follow-up brain MRI at 3–6 months. Four experienced neuroradiologists and four less-experienced raters scored the number of lesions per anatomical region and determined dissemination in space and time (McDonald 2010). Results: In controls, the mean number of lesions per rater was 0.16 at 1.5 T and 0.38 at 3 T (p = 0.005). For patients, this was 4.18 and 4.40, respectively (p = 0.657). Inter-rater agreement on involvement per anatomical region and dissemination in space and time was moderate to good for both field strengths. 3 T slightly improved agreement between experienced raters, but slightly decreased agreement between less-experienced raters. Conclusion: Overall, the interobserver agreement was moderate to good. 3 T appears to improve the reading for experienced readers, underlining the benefit of additional training. [ABSTRACT FROM AUTHOR]

Objectives: To evaluate the diagnostic performance of digital tomosynthesis (DTS) for the diagnosis of hip prosthesis loosening (PL) compared with conventional radiographs and CT with metal artifact reduction (CT-MAR).Methods: Forty-nine patients with painful hip prosthesis were prospectively included and underwent anteroposterior and lateral radiographs, anteroposterior DTS and CT-MAR of the hip. This study was approved by the local ethics committee, and all patients signed an informed consent form. Images were evaluated independently by two radiologists. Periprosthetic radiolucent lines wider than 2 mm found in two or more Gruen or De Lee and Charnley zones were considered diagnostic of PL. All cases of PL were confirmed surgically. Patients with a stable radiological follow-up for at least 1 year with an alternative cause for the symptoms or with no surgical evidence of PL were considered PL negative.Results: There were 21 cases of PL, 9 unilateral and 12 bilateral. For both the acetabular and femoral sides, DTS had a specificity for PL detection similar to that of conventional radiographs and CT-MAR (98.5-100%, 96.9%-100% and 96.9-95.4% respectively for both readers) and a sensitivity similar to conventional radiographs (39.9-45.4% versus 33.3-51.5% for both readers) but lower than CT-MAR (84.85% for both readers). The interobserver agreement was 0.84 for CT-MAR, 0.53 for DTS and 0.39 for conventional radiographs.Conclusion: DTS has a similar diagnostic performance to radiographs for the diagnosis of PL with a better interobserver agreement. The sensitivity however remains lower than that of CT-MAR.Key Points: • Plain radiograph is still the first imaging step when hip prosthesis loosening is suspected. • Interobserver agreement is better with digital tomosynthesis than radiographs. • Sensitivity of CT with state-of-the-art metal artifact reduction is superior to that of digital tomosynthesis. [ABSTRACT FROM AUTHOR]

Objectives The aims were to evaluate breast cancer (BC) tumor-infiltrating lymphocytes (TILs) interobserver variability, heterogeneity, and the fidelity of scoring TILs in core needle biopsy (CNB). Methods Matching CNB and two full-face sections (FFSs) of BC cases (n = 100) were independently reviewed by two pathologists. Percentage of stromal lymphocytes (TIL-str) and intratumoral lymphocytes (iTu-Ly) were recorded. Results The weighted κ values for the degree of agreement between both raters were 0.53 to 0.71. However, there was a slight improvement in the interobserver variability for TIL-str and slight decline in iTu-Ly when ER+/ HER2 – cases were excluded. The intraclass correlation coefficient for FFS1 vs FFS2 was 0.91 for TIL-str and 0.96 for iTu-Ly. Spearman correlation coefficient for CNB vs FFS1/FFS2 was 0.81 for TIL-str and 0.79 for iTu-Ly. Conclusions We conclude that the agreement in TILs scoring between the raters is acceptable. Caution should be practiced when scoring iTu-Ly in CNBs. [ABSTRACT FROM AUTHOR]

Objectives: The aims were to evaluate breast cancer (BC) tumor-infiltrating lymphocytes (TILs) interobserver variability, heterogeneity, and the fidelity of scoring TILs in core needle biopsy (CNB).Methods: Matching CNB and two full-face sections (FFSs) of BC cases (n = 100) were independently reviewed by two pathologists. Percentage of stromal lymphocytes (TIL-str) and intratumoral lymphocytes (iTu-Ly) were recorded.Results: The weighted κ values for the degree of agreement between both raters were 0.53 to 0.71. However, there was a slight improvement in the interobserver variability for TIL-str and slight decline in iTu-Ly when ER+/HER2- cases were excluded. The intraclass correlation coefficient for FFS1 vs FFS2 was 0.91 for TIL-str and 0.96 for iTu-Ly. Spearman correlation coefficient for CNB vs FFS1/FFS2 was 0.81 for TIL-str and 0.79 for iTu-Ly.Conclusions: We conclude that the agreement in TILs scoring between the raters is acceptable. Caution should be practiced when scoring iTu-Ly in CNBs. [ABSTRACT FROM AUTHOR]

Purpose The goal of this study was to estimate the risk of malignant thyroid nodules being interpreted as benign based on ultrasound findings and to clarify the pathological features of these malignant nodules. Materials and Methods We retrospectively re-evaluated ultrasound and pathological findings for 162 malignant thyroid nodules that were initially interpreted as benign based on ultrasound findings at Kuma Hospital between April 2012 and June 2015. Results The incidences of malignancy among “benign” thyroid nodules were 0.5% overall and 6.2% among resected nodules. In addition, 82.7% of thyroid nodules that were originally judged to have low or very low suspicion patterns were subsequently re-categorized as having high or intermediate suspicion patterns. The incidences of irregular margins (63.6%) and low echogenicity (36.4%) were higher than those of punctate microcalcification (17.9%) and the taller-than-wide shape (20.4%). Among microcarcinomas, the incidences were 65.7% for irregular margins and 51.4% for low echogenicity. Rim calcification with small extrusive soft tissue components and extrathyroidal extensions were not observed. After re-evaluation, 40.0% of papillary thyroid carcinomas remained benign based on their variants, such as the encapsulated, follicular, macrofollicular, and oxyphilic cell variants. Conclusion We conclude that more careful observation, especially for lesions with irregular margins and low echogenicity, can help improve the diagnostic accuracy of thyroid ultrasonography. Furthermore, greater care may decrease the incidence of malignancy among thyroid nodules with low or very low suspicion patterns. Some variants of papillary thyroid carcinoma can have benign ultrasound findings.

Abstract Purpose To investigate dosimetric impact of inter-observer variation in clinical target volume(CTV) delineation for patients undergoing interstitial partial breast brachytherapy. Methods Five radiation oncologists delineated CTV in twenty patients who underwent multi-catheter partial breast brachytherapy. Five treatment plans for each patient were graphically optimized for CTV of all observers and evaluated using coverage index(CI), external volume index(EI), overdose volume index(OI) and conformal index(COIN). In addition, volume enclosed by prescription isodose(V 100), its spatial concordance(CI common), mean coverage of all CTVs with common volume of prescription dose(V 100_common) and mean CTV coverage for all pairs of observer with common prescription volume of respective pairs(V 100_pair) were also computed. Results The mean ± standard deviation(SD) of CI and COIN ranged from 0.756 ± 0.076 to 0.840 ± 0.070 and 0.591 ± 0.090 to 0.673 ± 0.06 respectively. When a plan made for CTV of individual observer was evaluated on CTV of all observers, the maximum variations(ρ < 0.05) in the mean CI,COIN,OI and EI were 10.6%,11.4%,10.6% and 72.7% respectively. The observed mean ± SD of V 100 , CI common of V 100 , CTV coverage with V 100_common and V 100_pair was 160.7 ± 52.1, 0.70 ± 0.09, 73.1 ± 8.1% and 77.9 ± 7.3% respectively. Conclusion Inter-observer variation in delineation of CTV showed significant dosimetric impact with mean CTV coverage of 73.1% and 77.9% by common and paired prescription dose volume respectively among all observers. [ABSTRACT FROM AUTHOR]

Aims: CD30 immunohistochemistry (IHC) in malignant lymphoma is used for selection of patients in clinical trials using brentuximab vedotin, an antibody drug‐conjugate targeting the CD30 molecule. For reliable implementation in daily practice and meaningful selection of patients for clinical trials, information on technical variation and interobserver reproducibility of CD30 immunohistochemistry (IHC) staining is required. Methods and results: We conducted a three‐round reproducibility assessment of CD30 scoring for categorised frequency and intensity, including a technical validation, a ‘live polling’ pre‐ and post‐instruction scoring round and a web‐based round including individual scoring with additional IHC information to mimic daily diagnostic practice. Agreement in all three scoring rounds was poor to fair (κ = 0.12–0.35 for CD30‐positive tumour cell percentage and κ = 0.16–0.41 for staining intensity), even when allowing for one category of freedom in percentage of tumour cell positivity (κ = 0.30–0.61). The first round with CD30 staining performed in five independent laboratories showed objective differences in staining intensity. In the second round, approximately half the pathologists changed their opinion on CD30 frequency after a discussion on potential pitfalls, highlighting hesitancy in decision‐making. Using fictional cut‐off points for percentage of tumour cell positivity, agreement was still suboptimal (κ = 0.35–0.60). Conclusions: Lack of agreement in cases with heterogeneous expression is shown to influence patient eligibility for treatment with brentuximab vedotin, both in clinical practice and within the context of clinical trials, and limits the potential predictive value of the relative frequency of CD30‐positive neoplastic cells for clinical response. [ABSTRACT FROM AUTHOR]

Introduction: Reliability in visual cardiotocography interpretation is unsatisfying, which has led to the development of computerized cardiotocography. Computerized analysis is well established for antenatal fetal surveillance but has yet not performed sufficiently during labor. We aimed to investigate the capacity of a new computerized algorithm compared with visual assessment in identifying intrapartum fetal heart rate baseline and decelerations.Material and Methods: In all, 312 intrapartum cardiotocography tracings with variable decelerations were analyzed by the computerized algorithm and visually examined by two observers, blinded to each other and the computer analysis. The width, depth and area of each deceleration was measured. Four cases (>100 variable decelerations) were subjected to in-depth detailed analysis. The outcome measures were bias in seconds (width), beats per minute (depth), and beats (area) between computer and observers using Bland-Altman analysis. Interobserver reliability was determined by calculating intraclass correlation and Spearman rank analysis.Results: The analysis (312 cases) showed excellent intraclass correlation (0.89-0.95) and very strong Spearman correlation (0.82-0.91). The detailed analysis of >100 decelerations in four cases revealed low bias between the computer and the two observers; width 1.4 and 1.4 seconds, depth 5.1 and 0.7 beats per minute, and area 0.1 and -1.7 beats. This was comparable to the bias between the two observers: 0.3 seconds (width), 4.4 beats per minute (depth) and 1.7 beats (area). The intraclass correlation was excellent (0.90-.98).Conclusion: A novel computerized algorithm for intrapartum cardiotocography analysis is as accurate as gold standard visual assessment, with high correlation and low bias. [ABSTRACT FROM AUTHOR]

Aims: Intestinal metaplasia and atrophy of the gastric mucosa are associated with Helicobacter pylori infection and are considered premalignant lesions. The updated Sydney system is used for these parameters, but experienced pathologists and consensus processes are required for interobserver agreement. We sought to determine the influence of the consensus process on the assessment of intestinal metaplasia and atrophy. Methods and results: Two study sets were used: consensus and validation. The consensus set was circulated and five gastrointestinal pathologists evaluated them independently using the updated Sydney system. The consensus of the definitions was then determined at the first consensus meeting. The same set was recirculated to determine the effect of the consensus. The second consensus meeting was held to standardise the grading criteria and the validation set was circulated to determine the influence. Two additional circulations were performed to assess the maintainance of consensus and intraobserver variability. Interobserver agreement of intestinal metaplasia and atrophy was improved through the consensus process (intestinal metaplasia: baseline κ = 0.52 versus final κ = 0.68, P = 0.006; atrophy: baseline κ = 0.19 versus final κ = 0.43, P < 0.001). Higher interobserver agreement in atrophy was observed after consensus regarding the definition (pre‐consensus: κ = 0.19 versus post‐consensus: κ = 0.34, P = 0.001). There was improved interobserver agreement in intestinal metaplasia after standardisation of the grading criteria (pre‐standardisation: κ = 0.56 versus post‐standardisation: κ = 0.71, P = 0.010). Conclusions: This study suggests that interobserver variability regarding intestinal metaplasia and atrophy may result from lack of a precise definition and fine criteria, and can be reduced by consensus of definition and standardisation of grading criteria. [ABSTRACT FROM AUTHOR]

Purpose: To investigate interobserver variation in target volume delineations for prostate cancer salvage radiotherapy using planning computed tomography (CT) versus combined planning CT and magnetic resonance imaging (MRI). Materials and Methods: Ten radiation oncologists independently delineated a target volume on the planning CT scans of five cases with different pathological status after radical prostatectomy. Two weeks later, this was repeated with the addition of planning MRI. The volumes obtained with CT only and combined CT and MRI were compared, and the effect of the addition of planning MRI on interobserver variability was assessed. Results: There were large differences in clinical target volume (CTV) delineated by each observer, regardless of the addition of planning MRI (9.44-139.27 cm3 in CT only and 7.77-122.83 cm3 in CT plus MRI) and no significant differences in the mean and standard deviation of CTV. However, there were decreases in mean volume and standard deviation as a result of using the planning MRI. Conclusion: This study showed substantial interobserver variation in target volume delineation for salvage radiotherapy. The combination of planning MRI with CT tended to decrease the target volume and the variation. [ABSTRACT FROM AUTHOR]

Purpose The study compared interobserver variation in the delineation of the primary tumour (GTVp) and lymph nodes (GTVln) between three different 4DCT reconstruction types; Maximum Intensity Projection (MIP), Mid-Ventilation (Mid-V) and Mid-Position (Mid-P). Material and methods Seven radiation oncologists delineated the GTVp and GTVln on the MIP, Mid-V and Mid-P 4DCT image reconstructions of 10 lung cancer patients. The volumes, the mean standard deviation (SD) and distribution of SD (SD/area) over the median surface contour were compared for different tumour regions. Results The overall mean delineated volume on the MIP was significantly larger ( p  < 0.001) than the Mid-V and Mid-P. For the GTVp the Mid-P had the lowest interobserver variation (SD = 0.261 cm), followed by Mid-V (SD = 0.314 cm) and MIP (SD = 0.330 cm) For GTVln the Mid-V had the lowest interobserver variation (SD = 0.425 cm) followed by the MIP (SD = 0.477 cm) and Mid-P (SD = 0.543 cm). The SD/area distribution showed a statistically significant difference between the MIP versus Mid-P and Mid-P versus Mid-V for both GTVp and GTVln ( p  < 0.001), with outliers indicating interpretation differences for GTVp located close to the mediastinum and GTVln. Conclusion The Mid-P reduced the interobserver variation for the GTVp. Delineation protocols must be improved to benefit from the improved image quality of Mid-P for the GTVln. [ABSTRACT FROM AUTHOR]