Search

BackgroundWhile prior work examining severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) variants of concern focused on hospitalization and death, less is known about differences in clinical presentation. We compared the prevalence of acute symptoms across pre-Delta, Delta, and Omicron.MethodsWe conducted an analysis of the Innovative Support for Patients with SARS-CoV-2 Infections Registry (INSPIRE), a cohort study enrolling symptomatic SARS-CoV-2-positive participants. We determined the association between the pre-Delta, Delta, and Omicron time periods and the prevalence of 21 coronavirus disease 2019 (COVID-19) acute symptoms.ResultsWe enrolled 4113 participants from December 2020 to June 2022. Pre-Delta vs Delta vs Omicron participants had increasing sore throat (40.9%, 54.6%, 70.6%; P < .001), cough (50.9%, 63.3%, 66.7%; P < .001), and runny noses (48.9%, 71.3%, 72.9%; P < .001). We observed reductions during Omicron in chest pain (31.1%, 24.2%, 20.9%; P < .001), shortness of breath (42.7%, 29.5%, 27.5%; P < .001), loss of taste (47.1%, 61.8%, 19.2%; P < .001), and loss of smell (47.5%, 55.6%, 20.0%; P < .001). After adjustment, those infected during Omicron had significantly higher odds of sore throat vs pre-Delta (odds ratio [OR], 2.76; 95% CI, 2.26-3.35) and Delta (OR, 1.96; 95% CI, 1.69-2.28).ConclusionsParticipants infected during Omicron were more likely to report symptoms of common respiratory viruses, such as sore throat, and less likely to report loss of smell and taste.Trial registrationNCT04610515.

BackgroundThe prevalence, incidence, and interrelationships of persistent symptoms after severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection vary. There are limited data on specific phenotypes of persistent symptoms. Using latent class analysis (LCA) modeling, we sought to identify whether specific phenotypes of COVID-19 were present 3 months and 6 months post-infection.MethodsThis was a multicenter study of symptomatic adults tested for SARS-CoV-2 with prospectively collected data on general symptoms and fatigue-related symptoms up to 6 months postdiagnosis. Using LCA, we identified symptomatically homogenous groups among COVID-positive and COVID-negative participants at each time period for both general and fatigue-related symptoms.ResultsAmong 5963 baseline participants (4504 COVID-positive and 1459 COVID-negative), 4056 had 3-month and 2856 had 6-month data at the time of analysis. We identified 4 distinct phenotypes of post-COVID conditions (PCCs) at 3 and 6 months for both general and fatigue-related symptoms; minimal-symptom groups represented 70% of participants at 3 and 6 months. When compared with the COVID-negative cohort, COVID-positive participants had higher occurrence of loss of taste/smell and cognition problems. There was substantial class-switching over time; those in 1 symptom class at 3 months were equally likely to remain or enter a new phenotype at 6 months.ConclusionsWe identified distinct classes of PCC phenotypes for general and fatigue-related symptoms. Most participants had minimal or no symptoms at 3 and 6 months of follow-up. Significant proportions of participants changed symptom groups over time, suggesting that symptoms present during the acute illness may differ from prolonged symptoms and that PCCs may have a more dynamic nature than previously recognized. Clinical Trials Registration. NCT04610515.

BackgroundA standardized pathology management tool for melanocytic skin lesions may improve patient care by simplifying interpretation and categorization of the diverse terminology currently extant.ObjectiveTo assess an online educational intervention that teaches dermatopathologists to use the Melanocytic Pathology Assessment Tool and Hierarchy for Diagnosis (MPATH-Dx), a schema collapsing multiple diagnostic terms into 5 classes ranging from benign to invasive melanoma.MethodsPracticing dermatopathologists (N = 149) from 40 US states participated in a 2-year educational intervention study (71% response rate). The intervention involved a brief tutorial followed by practice on 28 melanocytic lesions, with the goal of teaching pathologists how to correctly use the MPATH-Dx schema; competence using the MPATH-Dx tool 12-24 months postintervention was assessed. Participants' self-reported confidence using the MPATH-Dx tool was assessed preintervention and postintervention.ResultsAt preintervention, confidence using the MPATH-Dx tool was already high, despite 68% lacking prior familiarity with it, and confidence increased postintervention (P = .0003). During the intervention, participants used the MPATH-Dx tool correctly for 90% of their interpretations; postintervention, participants used the MPATH-Dx tool correctly for 88% of their interpretations.LimitationsFuture research should examine implementing a standardized pathology assessment schema in actual clinical practice.ConclusionDermatopathologists can be taught to confidently and competently use the MPATH-Dx schema with a simple educational tutorial followed by practice.

BackgroundMost research on severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) variants focuses on initial symptomatology with limited longer-term data. We characterized prevalences of prolonged symptoms 3 months post-SARS-CoV-2 infection across 3 variant time-periods (pre-Delta, Delta, and Omicron).MethodsThis multicenter prospective cohort study of adults with acute illness tested for SARS-CoV-2 compared fatigue severity, fatigue symptoms, organ system-based symptoms, and ≥3 symptoms across variants among participants with a positive ("COVID-positive") or negative SARS-CoV-2 test ("COVID-negative") at 3 months after SARS-CoV-2 testing. Variant periods were defined by dates with ≥50% dominant strain. We performed multivariable logistic regression modeling to estimate independent effects of variants adjusting for sociodemographics, baseline health, and vaccine status.ResultsThe study included 2402 COVID-positive and 821 COVID-negative participants. Among COVID-positives, 463 (19.3%) were pre-Delta, 1198 (49.9%) Delta, and 741 (30.8%) Omicron. The pre-Delta COVID-positive cohort exhibited more prolonged severe fatigue (16.7% vs 11.5% vs 12.3%; P = .017) and presence of ≥3 prolonged symptoms (28.4% vs 21.7% vs 16.0%; P < .001) compared with the Delta and Omicron cohorts. No differences were seen in the COVID-negatives across time-periods. In multivariable models adjusted for vaccination, severe fatigue and odds of having ≥3 symptoms were no longer significant across variants.ConclusionsProlonged symptoms following SARS-CoV-2 infection were more common among participants infected during pre-Delta than with Delta and Omicron; however, these differences were no longer significant after adjusting for vaccination status, suggesting a beneficial effect of vaccination on risk of long-term symptoms. Clinical Trials Registration. NCT04610515.

BackgroundMetacognition is a cognitive process that involves self-awareness of thinking, understanding, and performance. This study assesses pathologists' metacognition by examining the association between their diagnostic accuracy and self-reported confidence levels while interpreting skin and breast biopsies.DesignWe studied 187 pathologists from the Melanoma Pathology Study (M-Path) and 115 pathologists from the Breast Pathology Study (B-Path). We measured pathologists' metacognitive ability by examining the area under the curve (AUC), the area under each pathologist's receiver operating characteristic (ROC) curve summarizing the association between confidence and diagnostic accuracy. We investigated possible relationships between this AUC measure, referred to as metacognitive sensitivity, and pathologist attributes. We also assessed whether higher metacognitive sensitivity affected the association between diagnostic accuracy and a secondary diagnostic action such as requesting a second opinion.ResultsWe found no significant associations between pathologist clinical attributes and metacognitive AUC. However, we found that pathologists with higher AUC showed a stronger trend to request secondary diagnostic action for inaccurate diagnoses and not for accurate diagnoses compared with pathologists with lower AUC.LimitationsPathologists reported confidence in specific diagnostic terms, rather than the broader classes into which the diagnostic terms were later grouped to determine accuracy. In addition, while there is no gold standard for the correct diagnosis to determine the accuracy of pathologists' interpretations, our studies achieved a high-quality reference diagnosis by using the consensus diagnosis of 3 experienced pathologists.ConclusionsMetacognition can affect clinical decisions. If pathologists have self-awareness that their diagnosis may be inaccurate, they can request additional tests or second opinions, providing the opportunity to correct inaccurate diagnoses.HighlightsMetacognitive sensitivity varied across pathologists, with most showing higher sensitivity than expected by chance.None of the demographic or clinical characteristics we examined was significantly associated with metacognitive sensitivity.Pathologists with higher metacognitive sensitivity were more likely to request additional tests or second opinions for their inaccurate diagnoses.

ImportanceA standardized pathology classification system for melanocytic lesions is needed to aid both pathologists and clinicians in cataloging currently existing diverse terminologies and in the diagnosis and treatment of patients. The Melanocytic Pathology Assessment Tool and Hierarchy for Diagnosis (MPATH-Dx) has been developed for this purpose.ObjectiveTo revise the MPATH-Dx version 1.0 classification tool, using feedback from dermatopathologists participating in the National Institutes of Health-funded Reducing Errors in Melanocytic Interpretations (REMI) Study and from members of the International Melanoma Pathology Study Group (IMPSG).Evidence reviewPracticing dermatopathologists recruited from 40 US states participated in the 2-year REMI study and provided feedback on the MPATH-Dx version 1.0 tool. Independently, member dermatopathologists participating in an IMPSG workshop dedicated to the MPATH-Dx schema provided additional input for refining the MPATH-Dx tool. A reference panel of 3 dermatopathologists, the original authors of the MPATH-Dx version 1.0 tool, integrated all feedback into an updated and refined MPATH-Dx version 2.0.FindingsThe new MPATH-Dx version 2.0 schema simplifies the original 5-class hierarchy into 4 classes to improve diagnostic concordance and to provide more explicit guidance in the treatment of patients. This new version also has clearly defined histopathological criteria for classification of classes I and II lesions; has specific provisions for the most frequently encountered low-cumulative sun damage pathway of melanoma progression, as well as other, less common World Health Organization pathways to melanoma; provides guidance for classifying intermediate class II tumors vs melanoma; and recognizes a subset of pT1a melanomas with very low risk and possible eventual reclassification as neoplasms lacking criteria for melanoma.Conclusions and relevanceThe implementation of the newly revised MPATH-Dx version 2.0 schema into clinical practice is anticipated to provide a robust tool and adjunct for standardized diagnostic reporting of melanocytic lesions and management of patients to the benefit of both health care practitioners and patients.

BackgroundArtificial intelligence (AI) may improve cancer detection and risk prediction during mammography screening, but radiologists' preferences regarding its characteristics and implementation are unknown.PurposeTo quantify how different attributes of AI-based cancer detection and risk prediction tools affect radiologists' intentions to use AI during screening mammography interpretation.Materials and methodsThrough qualitative interviews with radiologists, we identified five primary attributes for AI-based breast cancer detection and four for breast cancer risk prediction. We developed a discrete choice experiment based on these attributes and invited 150 US-based radiologists to participate. Each respondent made eight choices for each tool between three alternatives: two hypothetical AI-based tools versus screening without AI. We analyzed samplewide preferences using random parameters logit models and identified subgroups with latent class models.ResultsRespondents (n = 66; 44% response rate) were from six diverse practice settings across eight states. Radiologists were more interested in AI for cancer detection when sensitivity and specificity were balanced (94% sensitivity with

BackgroundAmorphous calcifications noted on mammograms (i.e., small and indistinct calcifications that are difficult to characterize) are associated with high diagnostic uncertainty, often leading to biopsies. Yet, only 20% of biopsied amorphous calcifications are cancer. We present a quantitative approach for distinguishing between benign and actionable (high-risk and malignant) amorphous calcifications using a combination of local textures, global spatial relationships, and interpretable handcrafted expert features.MethodOur approach was trained and validated on a set of 168 2D full-field digital mammography exams (248 images) from 168 patients. Within these 248 images, we identified 276 image regions with segmented amorphous calcifications and a biopsy-confirmed diagnosis. A set of local (radiomic and region measurements) and global features (distribution and expert-defined) were extracted from each image. Local features were grouped using an unsupervised k-means clustering algorithm. All global features were concatenated with clustered local features and used to train a LightGBM classifier to distinguish benign from actionable cases.ResultsOn the held-out test set of 60 images, our approach achieved a sensitivity of 100%, specificity of 35%, and a positive predictive value of 38% when the decision threshold was set to 0.4. Given that all of the images in our test set resulted in a recommendation of a biopsy, the use of our algorithm would have identified 15 images (25%) that were benign, potentially reducing the number of breast biopsies.ConclusionsQuantitative analysis of full-field digital mammograms can extract subtle shape, texture, and distribution features that may help to distinguish between benign and actionable amorphous calcifications.

The authors argue that postacute SARS-CoV-2 syndrome presents unique challenges in determining risk factors, natural history, and proper management. They call for strategies and actions that are based on rigorous, high-quality evidence.

PurposeThe aim of this study was to describe the current state of science regarding independent external validation of artificial intelligence (AI) technologies for screening mammography.MethodsA systematic review was performed across five databases (Embase, PubMed, IEEE Explore, Engineer Village, and arXiv) through December 10, 2020. Studies that used screening examinations from real-world settings to externally validate AI algorithms for mammographic cancer detection were included. The main outcome was diagnostic accuracy, defined by area under the receiver operating characteristic curve (AUC). Performance was also compared between radiologists and either stand-alone AI or combined radiologist and AI interpretation. Study quality was assessed using the Quality Assessment of Diagnostic Accuracy Studies 2 tool.ResultsAfter data extraction, 13 studies met the inclusion criteria (148,361 total patients). Most studies (77% [n = 10]) evaluated commercially available AI algorithms. Studies included retrospective reader studies (46% [n = 6]), retrospective simulation studies (38% [n = 5]), or both (15% [n = 2]). Across 5 studies comparing stand-alone AI with radiologists, 60% (n = 3) demonstrated improved accuracy with AI (AUC improvement range, 0.02-0.13). All 5 studies comparing combined radiologist and AI interpretation with radiologists alone demonstrated improved accuracy with AI (AUC improvement range, 0.028-0.115). Most studies had risk for bias or applicability concerns for patient selection (69% [n = 9]) and the reference standard (69% [n = 9]). Only two studies obtained ground-truth cancer outcomes through regional cancer registry linkage.ConclusionsTo date, external validation efforts for AI screening mammographic technologies suggest small potential diagnostic accuracy improvements but have been retrospective in nature and suffer from risk for bias and applicability concerns.

BackgroundSevere acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has caused one of the worst pandemics in recent history. Few reports have revealed that SARS-CoV-2 was spreading in the United States as early as the end of January. In this study, we aimed to determine if SARS-CoV-2 had been circulating in the Los Angeles (LA) area at a time when access to diagnostic testing for coronavirus disease 2019 (COVID-19) was severely limited.MethodsWe used a pooling strategy to look for SARS-CoV-2 in remnant respiratory samples submitted for regular respiratory pathogen testing from symptomatic patients from November 2019 to early March 2020. We then performed sequencing on the positive samples.ResultsWe detected SARS-CoV-2 in 7 specimens from 6 patients, dating back to mid-January. The earliest positive patient, with a sample collected on January 13, 2020 had no relevant travel history but did have a sibling with similar symptoms. Sequencing of these SARS-CoV-2 genomes revealed that the virus was introduced into the LA area from both domestic and international sources as early as January.ConclusionsWe present strong evidence of community spread of SARS-CoV-2 in the LA area well before widespread diagnostic testing was being performed in early 2020. These genomic data demonstrate that SARS-CoV-2 was being introduced into Los Angeles County from both international and domestic sources in January 2020.

BackgroundShareable e-scooters have become popular, but injuries to riders and bystanders have not been well characterized. The goal of this study was to describe e-scooter injuries and estimate the rate of injury per e-scooter trip.Methods and findingsRetrospective review of patients presenting to 180 clinics and 2 hospitals in greater Los Angeles between January 1, 2014 and May 14, 2020. Injuries were identified using a natural language processing (NLP) algorithm not previously used to identify injuries, tallied, and described along with required healthcare resources. We combine these tallies with municipal data on scooter use to report a monthly utilization-corrected rate of e-scooter injuries. We searched 36 million clinical notes. Our NLP algorithm correctly classified 92% of notes in the testing set compared with the gold standard of investigator review. In total, we identified 1,354 people injured by e-scooters; 30% were seen in more than one clinical setting (e.g., emergency department and a follow-up outpatient visit), 29% required advanced imaging, 6% required inpatient admission, and 2 died. We estimate 115 injuries per million e-scooter trips were treated in our health system.ConclusionsOur observed e-scooter injury rate is likely an underestimate, but is similar to that previously reported for motorcycles. However, the comparative severity of injuries is unknown. Our methodology may prove useful to study other clinical conditions not identifiable by existing diagnostic systems.

BackgroundReports on medium and long-term sequelae of SARS-CoV-2 infections largely lack quantification of incidence and relative risk. We describe the rationale and methods of the Innovative Support for Patients with SARS-CoV-2 Registry (INSPIRE) that combines patient-reported outcomes with data from digital health records to understand predictors and impacts of SARS-CoV-2 infection.MethodsINSPIRE is a prospective, multicenter, longitudinal study of individuals with symptoms of SARS-CoV-2 infection in eight regions across the US. Adults are eligible for enrollment if they are fluent in English or Spanish, reported symptoms suggestive of acute SARS-CoV-2 infection, and if they are within 42 days of having a SARS-CoV-2 viral test (i.e., nucleic acid amplification test or antigen test), regardless of test results. Recruitment occurs in-person, by phone or email, and through online advertisement. A secure online platform is used to facilitate the collation of consent-related materials, digital health records, and responses to self-administered surveys. Participants are followed for up to 18 months, with patient-reported outcomes collected every three months via survey and linked to concurrent digital health data; follow-up includes no in-person involvement. Our planned enrollment is 4,800 participants, including 2,400 SARS-CoV-2 positive and 2,400 SARS-CoV-2 negative participants (as a concurrent comparison group). These data will allow assessment of longitudinal outcomes from SARS-CoV-2 infection and comparison of the relative risk of outcomes in individuals with and without infection. Patient-reported outcomes include self-reported health function and status, as well as clinical outcomes including health system encounters and new diagnoses.ResultsParticipating sites obtained institutional review board approval. Enrollment and follow-up are ongoing.ConclusionsThis study will characterize medium and long-term sequelae of SARS-CoV-2 infection among a diverse population, predictors of sequelae, and their relative risk compared to persons with similar symptomatology but without SARS-CoV-2 infection. These data may inform clinical interventions for individuals with sequelae of SARS-CoV-2 infection.

BackgroundSynoptic reporting is recommended by many guideline committees to encourage the thorough histologic documentation necessary for optimal management of patients with melanoma.MethodsOne hundred fifty-one pathologists from 40 US states interpreted 41 invasive melanoma cases. For each synoptic reporting factor, the authors identified cases with "complete agreement" (all participants recorded the same value) versus any disagreement. Pairwise agreement was calculated for each case as the proportion of pairs of responses that agreed, where paired responses were generated by the comparison of each reviewer's response with all others.ResultsThere was complete agreement among all reviewers for 22 of the 41 cases (54%) on Breslow thickness dichotomized at 0.8 mm, with pairwise agreement ranging from 49% to 100% across the 41 cases. There was complete agreement for "no ulceration" in 24 of the 41 cases (59%), with pairwise agreement ranging from 42% to 100%. Tumor transected at base had complete agreement for 26 of the 41 cases (63%), with pairwise agreement ranging from 31% to 100%. Mitotic rate, categorized as 0/mm2 , 1/mm2 , or 2/mm2 , had complete agreement for 17 of the 41 cases (41%), with pairwise agreement ranging from 36% to 100%. Regression saw complete agreement for 14 of 41 cases (34%), with pairwise agreement ranging from 40% to 100%. Lymphovascular invasion, perineural invasion, and microscopic satellites were rarely reported as present. Respectively, these prognostic factors had complete agreement for 32 (78%), 37 (90%), and 18 (44%) of the 41 cases, and the ranges of pairwise agreement were 47% to 100%, 70% to 100%, and 53% to 100%, respectively.ConclusionsThese findings alert pathologists and clinicians to the problem of interobserver variability in recording critical prognostic factors.Lay summaryThis study addresses variability in the assessment and reporting of critical characteristics of invasive melanomas that are used by clinicians to guide patient care. The authors characterize the diagnostic variability among pathologists and their reporting methods in light of recently updated national guidelines. Results demonstrate considerable variability in the diagnostic reporting of melanoma with regard to the following: Breslow thickness, mitotic rate, ulceration, regression, and microscopic satellites. This work serves to alert pathologists and clinicians to the existence of variability in reporting these prognostic factors.

BackgroundStable, affordable housing is an established determinant of health. As affordable housing shortages across the USA threaten to displace people from their homes, it is important to understand the implications of cost-related residential moves for healthcare access.ObjectiveTo examine the relationship between cost-related moves and unmet medical needs.DesignWe performed a cross-sectional analysis of 7 waves (2011-2017) of the California Health Interview Survey.ParticipantsWe included all respondents ages 18 and older.Main measuresThe primary predictor variable was residential move history in the past 5 years (cost-related move, non-cost-related move, or no move). The primary outcome was unmet medical needs in the past year (necessary medications and/or medical care that were delayed or not received).Key resultsOur sample included 146,417 adults (42-47% response rate), representing a weighted population of 28,518,590. Overall, 20.3% of the sample reported unmet medical needs in the past year, and 4.9% reported a cost-related move in the past 5 years. In multivariable logistic regression models, adjusted risk of unmet medical needs increased for adults with both cost-related moves (aOR 1.38; 95% CI 1.19-1.59) and non-cost-related moves (aOR 1.17; 95% CI 1.09-1.26) compared to those with no moves. Among people who had moved, those with cost-related moves were more likely to report unmet medical needs compared to people with non-cost-related moves (p = 0.03).ConclusionsPeople who have moved due to unaffordable housing represent a population at increased risk for unmet medical needs. Policy makers seeking to improve population health should consider strategies to limit cost-related moves and to mitigate their adverse effects on healthcare access.

ObjectiveModeling variable-sized regions of interest (ROIs) in whole slide images using deep convolutional networks is a challenging task, as these networks typically require fixed-sized inputs that should contain sufficient structural and contextual information for classification. We propose a deep feature extraction framework that builds an ROI-level feature representation via weighted aggregation of the representations of variable numbers of fixed-sized patches sampled from nuclei-dense regions in breast histopathology images.MethodsFirst, the initial patch-level feature representations are extracted from both fully-connected layer activations and pixel-level convolutional layer activations of a deep network, and the weights are obtained from the class predictions of the same network trained on patch samples. Then, the final patch-level feature representations are computed by concatenation of weighted instances of the extracted feature activations. Finally, the ROI-level representation is obtained by fusion of the patch-level representations by average pooling.ResultsExperiments using a well-characterized data set of 240 slides containing 437 ROIs marked by experienced pathologists with variable sizes and shapes result in an accuracy score of 72.65% in classifying ROIs into four diagnostic categories that cover the whole histologic spectrum.ConclusionThe results show that the proposed feature representations are superior to existing approaches and provide accuracies that are higher than the average accuracy of another set of pathologists.SignificanceThe proposed generic representation that can be extracted from any type of deep convolutional architecture combines the patch appearance information captured by the network activations and the diagnostic relevance predicted by the class-specific scoring of patches for effective modeling of variable-sized ROIs.

BackgroundDiagnostic terms used in histopathology reports of cutaneous melanocytic lesions are not standardized. We describe dermatopathologists' views regarding diverse diagnostic terminology and the utility of the Melanocytic Pathology Assessment Tool and Hierarchy for Diagnosis (MPATH-Dx) for categorizing melanocytic lesions.MethodsJuly 2018-2019 survey of board-certified and/or fellowship-trained dermatopathologists with experience interpreting melanocytic lesions.ResultsAmong 160 participants, 99% reported witnessing different terminology being used for the same melanocytic lesion. Most viewed diverse terminology as confusing to primary care physicians (98%), frustrating to pathologists (83%), requiring more of their time as a consultant (64%), and providing necessary clinical information (52%). Most perceived that adoption of the MPATH-Dx would: improve communication with other pathologists and treating physicians (87%), generally be a change for the better (80%), improve patient care (79%), be acceptable to clinical colleagues (68%), save time in pathology report documentation (53%), and protect from malpractice (51%).ConclusionsMost dermatopathologists view diverse terminology as contributing to miscommunication with clinicians and patients, adversely impacting patient care. They view the MPATH-Dx as a promising tool to standardize terminology and improve communication. The MPATH-Dx may be a useful supplement to conventional pathology reports. Further revision and refinement are necessary for widespread clinical use.

Diagnosing melanocytic lesions is one of the most challenging areas of pathology with extensive intra- and inter-observer variability. The gold standard for a diagnosis of invasive melanoma is the examination of histopathological whole slide skin biopsy images by an experienced dermatopathologist. Digitized whole slide images offer novel opportunities for computer programs to improve the diagnostic performance of pathologists. In order to automatically classify such images, representations that reflect the content and context of the input images are needed. In this paper, we introduce a novel self-attention-based network to learn representations from digital whole slide images of melanocytic skin lesions at multiple scales. Our model softly weighs representations from multiple scales, allowing it to discriminate between diagnosis-relevant and -irrelevant information automatically. Our experiments show that our method outperforms five other state-of-the-art whole slide image classification methods by a significant margin. Our method also achieves comparable performance to 187 practicing U.S. pathologists who interpreted the same cases in an independent study. To facilitate relevant research, full training and inference code is made publicly available at https://github.com/meredith-wenjunwu/ScATNet.

PurposeMachine Learning Package for Cancer Diagnosis (MLCD) is the result of a National Institutes of Health/National Cancer Institute (NIH/NCI)-sponsored project for developing a unified software package from state-of-the-art breast cancer biopsy diagnosis and machine learning algorithms that can improve the quality of both clinical practice and ongoing research.MethodsWhole-slide images of 240 well-characterized breast biopsy cases, initially assembled under R01 CA140560, were used for developing the algorithms and training the machine learning models. This software package is based on the methodology developed and published under our recent NIH/NCI-sponsored research grant (R01 CA172343) for finding regions of interest (ROIs) in whole-slide breast biopsy images, for segmenting ROIs into histopathologic tissue types and for using this segmentation in classifiers that can suggest final diagnoses.ResultThe package provides an ROI detector for whole-slide images and modules for semantic segmentation into tissue classes and diagnostic classification into 4 classes (benign, atypia, ductal carcinoma in situ, invasive cancer) of the ROIs. It is available through the GitHub repository under the Massachusetts Institute of Technology license and will later be distributed with the Pathology Image Informatics Platform system. A Web page provides instructions for use.ConclusionOur tools have the potential to provide help to other cancer researchers and, ultimately, to practicing physicians and will motivate future research in this field. This article describes the methodology behind the software development and gives sample outputs to guide those interested in using this package.

Objectives"Assurance behaviors," a type of defensive medicine, involve physicians' utilization of additional patient services to avoid adverse legal outcomes. We aim to compare the use of clinical behaviors (such as ordering additional tests, services, and consultations) due to malpractice concerns with the same behaviors due to patient safety concerns.MethodsA national sample of dermatopathologists (n = 160) completed an online survey.ResultsParticipants reported using one or more of five clinical behaviors due to concerns about medical malpractice (95%) and patient safety (99%). Self-reported use of clinical behaviors due to malpractice concerns and patient safety concerns was compared, including ordering additional immunohistochemistry/molecular tests (71% vs 90%, respectively, P < .0001), recommending additional surgical sampling (78% vs 91%, P < .0001), requesting additional slides (81% vs 95%, P < .0001), obtaining second reviews (78% vs 91%, P < .0001), and adding caveats into reports regarding lesion difficulty (85% vs 89%, P > .05).ConclusionsDermatopathologists use many clinical behaviors both as assurance behaviors and due to patient safety concerns, with a higher proportion reporting patient safety concerns as a motivation for specific behaviors.

BackgroundMelanocytic tumors are often challenging and constitute almost one in four skin biopsies. Immunohistochemical (IHC) studies may assist diagnosis; however, indications for their use are not standardized.MethodsA test set of 240 skin biopsies of melanocytic tumors was examined by 187 pathologists from 10 US states, interpreting 48 cases in Phase I and either 36 or 48 cases in Phase II. Participant and diagnosis characteristics were compared between those who reported they would have ordered, or who would have not ordered IHC on individual cases. Intraobserver analysis examined consistency in the intent to order when pathologists interpreted the same cases on two occasions.ResultsOf 187 participants interpreting 48 cases each, 21 (11%) did not request IHC tests for any case, 85 (45%) requested testing for 1 to 6 cases, and 81 (43%) requested testing for ≥6 cases. Of 240 cases, 229 had at least one participant requesting testing. Only 2 out of 240 cases had more than 50% of participants requesting testing. Increased utilization of testing was associated with younger age of pathologist, board-certification in dermatopathology, low confidence in diagnosis, and lesions in intermediate MPATH-Dx classes 2 to 4. The median intraobserver concordance for requesting tests among 72 participants interpreting the same 48 cases in Phases I and II was 81% (IQR 73%-90%) and the median Kappa statistic was 0.20 (IQR 0.00, 0.39).ConclusionSubstantial variability exists among pathologists in utilizing IHC.

As more artificial intelligence (AI)-enhanced mammography screening tools enter the clinical market, greater focus will be placed on external validation in diverse patient populations. In this viewpoint, we outline lessons learned from prior efforts in this field, the need to validate algorithms on newer screening technologies and diverse patient populations, and conclude by discussing the need for a framework for continuous monitoring and recalibration of these AI tools. Sufficient validation and continuous monitoring of emerging AI tools for breast cancer screening will require greater stakeholder engagement and the creation of shared policies and guidelines.

BackgroundAlthough treatment guidelines exist for melanoma in situ and invasive melanoma, guidelines for other melanocytic skin lesions do not exist.ObjectiveTo examine pathologists' treatment suggestions for a broad spectrum of melanocytic skin lesions and compare them with existing guidelines.MethodsPathologists (N = 187) completed a survey and then provided diagnoses and treatment suggestions for 240 melanocytic skin lesions. Physician characteristics associated with treatment suggestions were evaluated with multivariable modeling.ResultsTreatment suggestions were concordant with National Comprehensive Cancer Network guidelines for the majority of cases interpreted as melanoma in situ (73%) and invasive melanoma (86%). Greater variability of treatment suggestions was seen for other lesion types without existing treatment guidelines. Characteristics associated with provision of treatment suggestions discordant with National Comprehensive Cancer Network guidelines were low caseloads (invasive melanoma), lack of fellowship training or board certification (melanoma in situ), and more than 10 years of experience (invasive melanoma and melanoma in situ).LimitationsPathologists could not perform immunohistochemical staining or other diagnostic tests; only 1 glass side was provided per biopsy case.ConclusionsPathologists' treatment suggestions vary significantly for melanocytic lesions, with lower variability for lesion types with national guidelines. Results suggest the need for standardization of treatment guidelines for all melanocytic lesion types.

ImportanceMany patients presently have access to their pathologic test result reports via online patient portals, yet little is known about pathologists' perspective on this topic.ObjectiveTo examine dermatopathologists' experience and perceptions of patient online access to pathology reports.Design, setting, and participantsA survey of 160 dermatopathologists currently practicing in the United States who are board certified and/or fellowship trained in dermatopathology was conducted between July 15, 2018, and September 23, 2019. Those who reported interpreting skin biopsies of melanocytic lesions within the previous year and expected to continue interpreting them for the next 2 years were included.Main outcomes and measuresDermatopathologists' demographic and clinical characteristics, experiences with patient online access to pathologic test result reports, potential behaviors and reactions to patient online access to those reports, and effects on patients who read their pathologic test result reports online.ResultsOf the 160 participating dermatopathologists from the 226 eligible for participation (71% response rate), 107 were men (67%); mean (SD) age was 49 (9.7) years (range, 34-77 years). Ninety-one participants (57%) reported that patients have contacted them directly about pathologic test reports they had written. Some participants noted that they would decrease their use of abbreviations and/or specialized terminology (57 [36%]), change the way they describe lesions suspicious for cancer (29 [18%]), and need specialized training in communicating with patients (39 [24%]) if patients were reading their reports. Most respondents perceived that patient understanding would increase (97 [61%]) and the quality of patient-physician communication would increase (98 [61%]) owing to the availability of online reports. Slightly higher proportions perceived increased patient worry (114 [71%]) and confusion (116 [73%]). However, on balance, most participants (114 [71%]) agreed that making pathologic test result reports available to patients online is a good idea.Conclusions and relevanceDermatopathologists in this survey study perceived both positive and negative consequences of patient online access to pathologic test result reports written by the respondents. Most participants believe that making pathologic test result reports available to patients online is a good idea; however, they also report concerns about patient worry and confusion increasing as a result. Further research regarding best practices and the effect on both patients and clinicians is warranted.

Digital whole slide images are Food and Drug Administration approved for clinical diagnostic use in pathology; however, integration is nascent. Trainees from 9 pathology training programs completed an online survey to ascertain attitudes toward and experiences with whole slide images for pathological interpretations. Respondents (n = 76) reported attending 63 unique medical schools (45 United States, 18 international). While 63% reported medical school exposure to whole slide images, most reported ≤ 5 hours. Those who began training more recently were more likely to report at least some exposure to digital whole slide image training in medical school compared to those who began training earlier: 75% of respondents beginning training in 2017 or 2018 reported exposure to whole slide images compared to 54% for trainees beginning earlier. Trainees exposed to whole slide images in medical school were more likely to agree they were comfortable using whole slide images for interpretation compared to those not exposed (29% vs 12%; P = .06). Most trainees agreed that accurate diagnoses can be made using whole slide images for primary diagnosis (92%; 95% CI: 86-98) and that whole slide images are useful for obtaining second opinions (93%; 95% CI: 88-99). Trainees reporting whole slide image experience during training, compared to those with no experience, were more likely to agree they would use whole slide images in 5 years for primary diagnosis (64% vs 50%; P = .3) and second opinions (86% vs 76%; P = .4). In conclusion, although exposure to whole slide images in medical school has increased, overall exposure is limited. Positive attitudes toward future whole slide image diagnostic use were associated with exposure to this technology during medical training. Curricular integration may promote adoption.

ImportanceFollowing recent US Food and Drug Administration approval, adoption of whole slide imaging in clinical settings may be imminent, and diagnostic accuracy, particularly among challenging breast biopsy specimens, may benefit from computerized diagnostic support tools.ObjectiveTo develop and evaluate computer vision methods to assist pathologists in diagnosing the full spectrum of breast biopsy samples, from benign to invasive cancer.Design, setting, and participantsIn this diagnostic study, 240 breast biopsies from Breast Cancer Surveillance Consortium registries that varied by breast density, diagnosis, patient age, and biopsy type were selected, reviewed, and categorized by 3 expert pathologists as benign, atypia, ductal carcinoma in situ (DCIS), and invasive cancer. The atypia and DCIS cases were oversampled to increase statistical power. High-resolution digital slide images were obtained, and 2 automated image features (tissue distribution feature and structure feature) were developed and evaluated according to the consensus diagnosis of the expert panel. The performance of the automated image analysis methods was compared with independent interpretations from 87 practicing US pathologists. Data analysis was performed between February 2017 and February 2019.Main outcomes and measuresDiagnostic accuracy defined by consensus reference standard of 3 experienced breast pathologists.ResultsThe accuracy of machine learning tissue distribution features, structure features, and pathologists for classification of invasive cancer vs noninvasive cancer was 0.94, 0.91, and 0.98, respectively; the accuracy of classification of atypia and DCIS vs benign tissue was 0.70, 0.70, and 0.81, respectively; and the accuracy of classification of DCIS vs atypia was 0.83, 0.85, and 0.80, respectively. The sensitivity of both machine learning features was lower than that of the pathologists for the invasive vs noninvasive classification (tissue distribution feature, 0.70; structure feature, 0.49; pathologists, 0.84) but higher for the classification of atypia and DCIS vs benign cases (tissue distribution feature, 0.79; structure feature, 0.85; pathologists, 0.72) and the classification of DCIS vs atypia (tissue distribution feature, 0.88; structure feature, 0.89; pathologists, 0.70). For the DCIS vs atypia classification, the specificity of the machine learning feature classification was similar to that of the pathologists (tissue distribution feature, 0.78; structure feature, 0.80; pathologists, 0.82).Conclusion and relevanceThe computer-based automated approach to interpreting breast pathology showed promise, especially as a diagnostic aid in differentiating DCIS from atypical hyperplasia.