Liver Observation Segmentation on Contrast-Enhanced MRI: SAM and MedSAM Performance in Patients With Probable or Definite Hepatocellular Carcinoma.

Purpose: To evaluate factors impacting the Segment Anything Model (SAM) and variant MedSAM performance for segmenting liver observations on contrast-enhanced (CE) magnetic resonance imaging (MRI) in high-risk patients with probable hepatocellular carcinoma (HCC) (LR-4) and definite HCC (LR-5). Methods: A retrospective cohort of liver observations (LR-4/LR-5) on CE-MRI from 97 patients at high-risk for HCC was derived (2013-2018). Using bounding-boxes as prompts under 5-fold cross-validation, segmentation performance was evaluated at the model and liver observation-levels for: (1) model types: SAM versus MedSAM, (2) image sizes: 256 × 256 versus 512 × 512, (3) image channel composition: CE sequences at 3 phases of enhancement independently and combined, (4) liver observation size: >10 mm versus >20 mm, (5) certainty of diagnosis: LR-4 versus LR-5, and (6) contrast-agent type: hepatobiliary versus extracellular. Segmentation performance, quantified using Dice coefficient, were compared using univariate (Wilcoxon signed-rank and t-test) and multivariable analyses (multiple correspondence analysis and subsequent linear modelling). Results: MedSAM trained on 512 × 512 combined CE sequences performed best with mean Dice coefficient 0.68 (95% confidence interval 0.66, 0.69). Overall, all factors except contrast-agent type affected performance, with larger image size resulting in the highest performance improvement (512 × 512: 0.57, 256 × 256: 0.26, P < .001) at the model-level. Contrast-agents affected performance for patients with LR-4 observations using MedSAM-based models (P < .03). Larger observation size, image size, and higher certainty of diagnosis were associated with better segmentation on multivariable analysis. Conclusion: A variety of factors were found to impact SAM/MedSAM performance for segmenting liver observations in patients with probable and definite HCC on CE-MRI. Future models may be optimized by accounting for these factors.

Seeing the random forest through the decision trees. Supporting learning health systems from histopathology with machine learning models: Challenges and opportunities.

This paper discusses some overlooked challenges faced when working with machine learning models for histopathology and presents a novel opportunity to support "Learning Health Systems" with them. Initially, the authors elaborate on these challenges after separating them according to their mitigation strategies: those that need innovative approaches, time, or future technological capabilities and those that require a conceptual reappraisal from a critical perspective. Then, a novel opportunity to support "Learning Health Systems" by integrating hidden information extracted by ML models from digitalized histopathology slides with other healthcare big data is presented.

Evaluating the efficacy of artificial intelligence tools for the automation of systematic reviews in cancer research: A systematic review.

To evaluate the performance accuracy and workload savings of artificial intelligence (AI)-based automation tools in comparison with human reviewers in medical literature screening for systematic reviews (SR) of primary studies in cancer research in order to gain insights on improving the efficiency of producing SRs. Medline, Embase, the Cochrane Library, and PROSPERO databases were searched from inception to November 30, 2022. Then, forward and backward literature searches were completed, and the experts in this field including the authors of the articles included were contacted for a thorough grey literature search. This SR was registered on PROSPERO (CRD 42023384772). Among the 3947 studies obtained from search, five studies met the preplanned study selection criteria. These five studies evaluated four AI tools: Abstrackr (four studies), RobotAnalyst (one), EPPI-Reviewer (one), and DistillerSR (one). Without missing final included citations, Abstrackr eliminated 20%-88% of titles and abstracts (time saving of 7-86 hours) and 59% of the full-texts (62 h) from human review across four different cancer-related SRs. In comparison, RobotAnalyst (1% of titles and abstracts, 1 h), EPPI Review (38% of titles and abstracts, 58 h; 59% of full-texts, 62 h), DistillerSR (42% of titles and abstracts, 22 h) also provided similar or lower work savings for single cancer-related SRs. AI-based automation tools exhibited promising but varying levels of accuracy and efficiency during the screening process of medical literature for conducting SRs in the cancer field. Until further progress is made and thorough evaluations are conducted, AI tools should be utilized as supplementary aids rather than complete substitutes for human reviewers.

Performance of externally validated machine learning models based on histopathology images for the diagnosis, classification, prognosis, or treatment outcome prediction in female breast cancer: A systematic review.

Numerous machine learning (ML) models have been developed for breast cancer using various types of data. Successful external validation (EV) of ML models is important evidence of their generalizability. The aim of this systematic review was to assess the performance of externally validated ML models based on histopathology images for diagnosis, classification, prognosis, or treatment outcome prediction in female breast cancer. A systematic search of MEDLINE, EMBASE, CINAHL, IEEE, MICCAI, and SPIE conferences was performed for studies published between January 2010 and February 2022. The Prediction Model Risk of Bias Assessment Tool (PROBAST) was employed, and the results were narratively described. Of the 2011 non-duplicated citations, 8 journal articles and 2 conference proceedings met inclusion criteria. Three studies externally validated ML models for diagnosis, 4 for classification, 2 for prognosis, and 1 for both classification and prognosis. Most studies used Convolutional Neural Networks and one used logistic regression algorithms. For diagnostic/classification models, the most common performance metrics reported in the EV were accuracy and area under the curve, which were greater than 87% and 90%, respectively, using pathologists' annotations/diagnoses as ground truth. The hazard ratios in the EV of prognostic ML models were between 1.7 (95% CI, 1.2-2.6) and 1.8 (95% CI, 1.3-2.7) to predict distant disease-free survival; 1.91 (95% CI, 1.11-3.29) for recurrence, and between 0.09 (95% CI, 0.01-0.70) and 0.65 (95% CI, 0.43-0.98) for overall survival, using clinical data as ground truth. Despite EV being an important step before the clinical application of a ML model, it hasn't been performed routinely. The large variability in the training/validation datasets, methods, performance metrics, and reported information limited the comparison of the models and the analysis of their results. Increasing the availability of validation datasets and implementing standardized methods and reporting protocols may facilitate future analyses.

Generative AI for brain image computing and brain network computing: a review.

Recent years have witnessed a significant advancement in brain imaging techniques that offer a non-invasive approach to mapping the structure and function of the brain. Concurrently, generative artificial intelligence (AI) has experienced substantial growth, involving using existing data to create new content with a similar underlying pattern to real-world data. The integration of these two domains, generative AI in neuroimaging, presents a promising avenue for exploring various fields of brain imaging and brain network computing, particularly in the areas of extracting spatiotemporal brain features and reconstructing the topological connectivity of brain networks. Therefore, this study reviewed the advanced models, tasks, challenges, and prospects of brain imaging and brain network computing techniques and intends to provide a comprehensive picture of current generative AI techniques in brain imaging. This review is focused on novel methodological approaches and applications of related new methods. It discussed fundamental theories and algorithms of four classic generative models and provided a systematic survey and categorization of tasks, including co-registration, super-resolution, enhancement, classification, segmentation, cross-modality, brain network analysis, and brain decoding. This paper also highlighted the challenges and future directions of the latest work with the expectation that future research can be beneficial.

A scoping review of natural language processing of radiology reports in breast cancer.

Various natural language processing (NLP) algorithms have been applied in the literature to analyze radiology reports pertaining to the diagnosis and subsequent care of cancer patients. Applications of this technology include cohort selection for clinical trials, population of large-scale data registries, and quality improvement in radiology workflows including mammography screening. This scoping review is the first to examine such applications in the specific context of breast cancer. Out of 210 identified articles initially, 44 met our inclusion criteria for this review. Extracted data elements included both clinical and technical details of studies that developed or evaluated NLP algorithms applied to free-text radiology reports of breast cancer. Our review illustrates an emphasis on applications in diagnostic and screening processes over treatment or therapeutic applications and describes growth in deep learning and transfer learning approaches in recent years, although rule-based approaches continue to be useful. Furthermore, we observe increased efforts in code and software sharing but not with data sharing.

Developing a Data and Analytics Platform to Enable a Breast Cancer Learning Health System at a Regional Cancer Center.

PURPOSE: This study documents the creation of automated, longitudinal, and prospective data and analytics platform for breast cancer at a regional cancer center. This platform combines principles of data warehousing with natural language processing (NLP) to provide the integrated, timely, meaningful, high-quality, and actionable data required to establish a learning health system. METHODS: Data from six hospital information systems and one external data source were integrated on a nightly basis by automated extract/transform/load jobs. Free-text clinical documentation was processed using a commercial NLP engine. RESULTS: The platform contains 141 data elements of 7,019 patients with newly diagnosed breast cancer who received care at our regional cancer center from January 1, 2014, to June 3, 2022. Daily updating of the database takes an average of 56 minutes. Evaluation of the tuning of NLP jobs found overall high performance, with an F1 of 1.0 for 19 variables, with a further 16 variables with an F1 of > 0.95. CONCLUSION: This study describes how data warehousing combined with NLP can be used to create a prospective data and analytics platform to enable a learning health system. Although upfront time investment required to create the platform was considerable, now that it has been developed, daily data processing is completed automatically in less than an hour.

Predicting Severity of Head Collision Events in Elite Soccer Using Preinjury Data: A Machine Learning Approach.

OBJECTIVE: To develop machine learning (ML) models that predict severity of head collision events (HCEs) based on preinjury variables and to investigate which variables are important to predicting severity. DESIGN: Data on HCEs were collected with respect to severity and 23 preinjury variables to create 2 datasets, a male dataset using men's tournaments and mixed dataset using men's and women's tournaments, to perform ML analysis. Machine learning analysis used a random forest classifier based on preinjury variables to predict HCE severity. SETTING: Four elite international soccer tournaments. PARTICIPANTS: Elite athletes participating in analyzed tournaments. INDEPENDENT VARIABLES: The 23 preinjury variables collected for each HCE. MAIN OUTCOME MEASURES: Predictive ability of the ML models and association of important variables. RESULTS: The ML models had an average area under the receiver operating characteristic curve for predicting HCE severity of 0.73 and 0.70 for the male and mixed datasets, respectively. The most important variables for prediction were the mechanism of injury and the event before injury. In the male dataset, the mechanisms "head-to-head" and "knee-to-head" were together significantly associated ( P = 0.0244) with severity; they were not significant in the mixed dataset ( P = 0.1113). In both datasets, the events "corner kicks" and "throw-ins" were together significantly associated with severity (male, P = 0.0001; mixed, P = 0.0004). CONCLUSIONS: ML models accurately predicted the severity of HCE. The mechanism and event preceding injury were most important for predicting severity of HCEs. These findings support the use of ML to inform preventative measures that will mitigate the impact of these preinjury factors on player health.

A Competition, Benchmark, Code, and Data for Using Artificial Intelligence to Detect Lesions in Digital Breast Tomosynthesis.

Importance: An accurate and robust artificial intelligence (AI) algorithm for detecting cancer in digital breast tomosynthesis (DBT) could significantly improve detection accuracy and reduce health care costs worldwide. Objectives: To make training and evaluation data for the development of AI algorithms for DBT analysis available, to develop well-defined benchmarks, and to create publicly available code for existing methods. Design, Setting, and Participants: This diagnostic study is based on a multi-institutional international grand challenge in which research teams developed algorithms to detect lesions in DBT. A data set of 22 032 reconstructed DBT volumes was made available to research teams. Phase 1, in which teams were provided 700 scans from the training set, 120 from the validation set, and 180 from the test set, took place from December 2020 to January 2021, and phase 2, in which teams were given the full data set, took place from May to July 2021. Main Outcomes and Measures: The overall performance was evaluated by mean sensitivity for biopsied lesions using only DBT volumes with biopsied lesions; ties were broken by including all DBT volumes. Results: A total of 8 teams participated in the challenge. The team with the highest mean sensitivity for biopsied lesions was the NYU B-Team, with 0.957 (95% CI, 0.924-0.984), and the second-place team, ZeDuS, had a mean sensitivity of 0.926 (95% CI, 0.881-0.964). When the results were aggregated, the mean sensitivity for all submitted algorithms was 0.879; for only those who participated in phase 2, it was 0.926. Conclusions and Relevance: In this diagnostic study, an international competition produced algorithms with high sensitivity for using AI to detect lesions on DBT images. A standardized performance benchmark for the detection task using publicly available clinical imaging data was released, with detailed descriptions and analyses of submitted algorithms accompanied by a public release of their predictions and code for selected methods. These resources will serve as a foundation for future research on computer-assisted diagnosis methods for DBT, significantly lowering the barrier of entry for new researchers.

A Data Set and Deep Learning Algorithm for the Detection of Masses and Architectural Distortions in Digital Breast Tomosynthesis Images.

Importance: Breast cancer screening is among the most common radiological tasks, with more than 39 million examinations performed each year. While it has been among the most studied medical imaging applications of artificial intelligence, the development and evaluation of algorithms are hindered by the lack of well-annotated, large-scale publicly available data sets. Objectives: To curate, annotate, and make publicly available a large-scale data set of digital breast tomosynthesis (DBT) images to facilitate the development and evaluation of artificial intelligence algorithms for breast cancer screening; to develop a baseline deep learning model for breast cancer detection; and to test this model using the data set to serve as a baseline for future research. Design, Setting, and Participants: In this diagnostic study, 16 802 DBT examinations with at least 1 reconstruction view available, performed between August 26, 2014, and January 29, 2018, were obtained from Duke Health System and analyzed. From the initial cohort, examinations were divided into 4 groups and split into training and test sets for the development and evaluation of a deep learning model. Images with foreign objects or spot compression views were excluded. Data analysis was conducted from January 2018 to October 2020. Exposures: Screening DBT. Main Outcomes and Measures: The detection algorithm was evaluated with breast-based free-response receiver operating characteristic curve and sensitivity at 2 false positives per volume. Results: The curated data set contained 22 032 reconstructed DBT volumes that belonged to 5610 studies from 5060 patients with a mean (SD) age of 55 (11) years and 5059 (100.0%) women. This included 4 groups of studies: (1) 5129 (91.4%) normal studies; (2) 280 (5.0%) actionable studies, for which where additional imaging was needed but no biopsy was performed; (3) 112 (2.0%) benign biopsied studies; and (4) 89 studies (1.6%) with cancer. Our data set included masses and architectural distortions that were annotated by 2 experienced radiologists. Our deep learning model reached breast-based sensitivity of 65% (39 of 60; 95% CI, 56%-74%) at 2 false positives per DBT volume on a test set of 460 examinations from 418 patients. Conclusions and Relevance: The large, diverse, and curated data set presented in this study could facilitate the development and evaluation of artificial intelligence algorithms for breast cancer screening by providing data for training as well as a common set of cases for model validation. The performance of the model developed in this study showed that the task remains challenging; its performance could serve as a baseline for future model development.

The Traumatic Brain Injury Model Systems National Database: A Review of Published Research.

The Traumatic Brain Injury Model Systems (TBIMS) is the largest longitudinal TBI data set in the world. Our study reviews the works using TBIMS data for analysis in the last 5 years. A search (2015-2020) was conducted across PubMed, EMBASE, and Google Scholar for studies that used the National Institute on Disability, Independent Living and Rehabilitation Research NIDILRR/VA-TBIMS data. Search terms were as follows: ["TBIMS" national database] within PubMed and Google Scholar, and ["TBIMS" AND national AND database] on EMBASE. Data sources, study foci (in terms of data processing and outcomes), study outcomes, and follow-up information usage were collected to categorize the studies included in this review. Variable usage in terms of TBIMS' form-based variable groups and limitations from each study were also noted. Assessment was made on how TBIMS' objectives were met by the studies. Of the 74 articles reviewed, 23 used TBIMS along with other data sets. Fifty-four studies focused on specific outcome measures only, 6 assessed data aspects as a major focus, and 13 explored both. Sample sizes of the included studies ranged from 11 to 15,835. Forty-two of the 60 longitudinal studies assessed follow-up from 1 to 5 years, and 15 studies used 10 to 25 years of the same. Prominent variable groups as outcome measures were "Employment," "FIM," "DRS," "PART-O," "Satisfaction with Life," "PHQ-9," and "GOS-E." Limited numbers of studies were published regarding tobacco consumption, the Brief Test of Adult Cognition by Telephone (BTACT), the Supervision Rating Scale (SRS), general health, and comorbidities as variables of interest. Generalizability was the most significant limitation mentioned by the studies. The TBIMS is a rich resource for large-sample longitudinal analyses of various TBI outcomes. Future efforts should focus on under-utilized variables and improving generalizability by validation of results across large-scale TBI data sets to better understand the heterogeneity of TBI.

Development and validation of a patient-centered, meningioma-specific quality-of-life questionnaire.

OBJECTIVE: Meningiomas can have significant impact on health-related quality of life (HRQOL). Patient-centered, disease-specific instruments for assessing HRQOL in these patients are lacking. To this end, the authors sought to develop and validate a meningioma-specific HRQOL questionnaire through a standardized, patient-centered questionnaire development methodology. METHODS: The development of the questionnaire involved three main phases: item generation, item reduction, and validation. Item generation consisted of semistructured interviews with patients (n = 30), informal caregivers (n = 12), and healthcare providers (n = 8) to create a preliminary list of items. Item reduction with 60 patients was guided by the clinical impact method, multiple correspondence analysis, and hierarchical cluster analysis. The validation phase involved 162 patients and collected evidence on extreme-groups validity; concurrent validity with the SF-36, FACT-Br, and EQ-5D; and test-retest reliability. The questionnaire takes on average 11 minutes to complete. RESULTS: The meningioma-specific quality-of-life questionnaire (MQOL) consists of 70 items representing 9 domains. Cronbach's alpha for each domain ranged from 0.61 to 0.91. Concurrent validity testing demonstrated construct validity, while extreme-groups testing (p = 1.45E-11) confirmed the MQOL's ability to distinguish between different groups of patients. CONCLUSIONS: The MQOL is a validated, reliable, and feasible questionnaire designed specifically for evaluating QOL in meningioma patients. This disease-specific questionnaire will be fundamentally helpful in better understanding and capturing HRQOL in the meningioma patient population and can be used in both clinical and research settings.

Cognitive Dysfunction, Brain Volumes, and Traumatic Brain Injury in Homeless Persons.

Although homeless persons experience traumatic brain injury (TBI) frequently, little is known about the structural and functional brain changes in this group. We aimed to describe brain volume changes and related cognitive/motor deficits in homeless persons with or without TBI versus controls. Participants underwent T1-weighted magnetic resonance imaging (MRI), neuropsychological (NP) tests (the Grooved Pegboard Test [GPT]/Finger Tapping Test [FTT]), alcohol/drug use screens (the Alcohol Use Disorders Identification Test [AUDIT]/Drug Abuse Screening Test [DAST]), and questionnaires (the Brain Injury Screening Questionnaire [BISQ]/General Information Questionnaire [GIQ]) to determine TBI. Normalized volumes of brain substructures from MRI were derived from FreeSurfer. Comparisons were tested by Mann-Whitney U and Kruskal-Wallis rank sum tests. Leave-one-out cross-validation using random forest classifier was applied to determine the ability of predicting TBI. Diagnostic ability of this classifier was assessed using area under the receiver operating characteristic curve (AUC). Fifty-one participants-25 homeless persons (9 with TBI) and 26 controls-were included. The homeless group had higher AUDIT scores and smaller thalamus and brainstem volumes (p < 0.001) than controls. Within homeless participants, the TBI group had reduced normalized volumes of nucleus accumbens, thalamus, ventral diencephalon, and brainstem compared with the non-TBI group (p < 0.001). Homeless participants took more time on the GPT compared with controls using both hands (p < 0.0001); but the observed effects were more pronounced in the homeless group with TBI in the non-dominant hand. Homeless persons with TBI had fewer dominant hand finger taps than controls (p = 0.0096), and homeless participants with (p = 0.0148) or without TBI (p = 0.0093) tapped less than controls with their non-dominant hand. In all participants, TBI was predicted with an AUC of 0.95 (95% confidence interval [CI]: 0.89-1.00) by the classifier modeled on MRI, NP tests, and screening data combined. The MRI-data-based classifier was the best predictor of TBI within the homeless group (AUC: 0.76, 95% CI: 0.53-0.99). Normalized volumes of specific brain substructures were important indicators of TBI in homeless participants and they are important indicators of TBI in the state of homelessness itself. They may improve predictive ability of NP and screening tests in determining these outcomes.

Traumatic Brain Injury, Abuse, and Poor Sustained Attention in Youth and Young Adults Who Previously Experienced Foster Care.

Youth and young adults who previously experienced foster care are prone to negative life events, such as physical injuries, and adverse childhood experiences (ACE), such as abuse, neglect, and household dysfunction. The purpose of the present study was to identify the prevalence of traumatic brain injury (TBI), ACE, and poor sustained attention and the associations of these events in this group of vulnerable persons. Participants completed standardized questionnaires on the prevalence of self-reported TBI (TBI) and ACE and performed the Sustained Attention to Response Task (SART) test to measure sustained attention. Chi-squared and Kruskal-Wallis rank-sum tests were used to assess demographic differences and associations between TBI and ACE. Sustained attention was assessed using analysis of variance and linear modeling. Seventy-one participants-46 youth and young adults who previously experienced foster care (vulnerable group) and 25 age-matched healthy controls-completed the standardized questionnaires. Analyses indicated that vulnerable participants reported markedly higher rates of TBI and ACE than healthy controls. Vulnerable persons with TBI reported significantly higher Total ACE scores (p = 0.02), were more likely to have a history of family dysfunction (p = 0.02), and were more likely to have lived with a mentally ill guardian (p = 0.01) than vulnerable persons with no TBI. TBI was significantly associated with Total Errors (p = 0.001 and p = 0.02) and Omission Errors (p < 0.001 and p = 0.01) in all participants and in vulnerable participants, respectively, after adjusting for education level.

Setting standards of performance expected in neurosurgery residency: A study on entrustable professional activities in competency-based medical education.

BACKGROUND: Competency-based medical education requires evaluations of residents' performances of tasks of the discipline (ie. entrustable professional activities (EPAs)). Using neurosurgical Faculty perspectives, this study investigated whether a sample of neurosurgical EPAs accurately reflected the expectations of general neurosurgical practice. METHOD: A questionnaire was sent to all Canadian neurosurgery Faculty using a SurveyMonkey® platform. RESULTS: The proportion of respondents who believed the EPAs were representative of general neurosurgery competences varied significantly across all EPAs [47%-100%] (p < 0.0001). For 9/15 proposed EPAs, ≥75% agreed they were appropriate for general neurosurgery training and expected residents to attain the highest standard of performance. However, a range of 27-53% of the respondents felt the other six EPAs would be more appropriate for fellowship training and thus, require a lower standard of performance from graduating residents. CONCLUSION: The shift towards subspecialization in neurosurgery has implications for curriculum design, delivery and certification of graduating residents.

Resident evaluations in the age of competency-based medical education: faculty perspectives on minimizing burdens.

OBJECTIVE: Competency-based medical education (CBME), an outcomes-based approach to medical education, continues to be implemented across many postgraduate medical education programs worldwide, including a recent introduction into Canadian neurosurgical training programs (July 2019). The success of this educational paradigm shift requires frequent faculty observation and evaluation of residents performing defined tasks of the specialty. A main challenge involves providing residents with frequent performance evaluations and feedback that are feasible for faculty to complete. This study aims to define what is currently happening and what changes are needed to make CBME successful for the certification of neurosurgeons' competence. METHODS: A 55-item questionnaire was emailed nationwide to survey Canadian neurosurgical faculty. RESULTS: Fifty-two complete responses were received and achieved a distribution highly correlated with the number of faculty neurosurgeons practicing in each Canadian province (Pearson's r = 0.94). Two-thirds (35/52) of faculty reported currently taking a median of 10 minutes to complete evaluation forms at the end of a resident's rotation block. Regardless of the faculty's province of practice (p = 0.50) or years of experience (p = 0.06), they reported 3 minutes (minimum 1 minute, maximum 10 minutes, interquartile range [IQR] 3 minutes) as a feasible amount of time to spend completing an evaluation form following an observation of a resident's performance of an entrustable professional activity (EPA). If evaluation forms took 3 minutes to complete, 85% of respondents (44/52) would complete EPA evaluations weekly or daily. The faculty recommended 5 minutes as a feasible amount of time to provide oral feedback (minimum 1 minute, maximum 20 minutes, IQR 3.25 minutes), which was significantly higher (p = 0.00099) than their recommended amount of time for completing evaluation forms. The majority of faculty (71%) stated they would prefer to access resident evaluation forms through a mobile application compared to a paper form (12%), an evaluation website (8%), or through a URL link sent via email (10%; p = 0.0032). CONCLUSIONS: To facilitate the successful implementation of CBME into a neurosurgical training curriculum, resident EPA assessment forms should take 3 minutes or less to complete and be accessible through a mobile application.

Gene Expression Signatures Identify Biologically Homogenous Subgroups of Grade 2 Meningiomas.

INTRODUCTION: Meningiomas are the most common brain tumor, with prevalence of approximately 3%. Histological grading has a major role in determining treatment choice and predicting outcome. While indolent grade 1 and aggressive grade 3 meningiomas exhibit relatively homogeneous clinical behavior, grade 2 meningiomas are far more heterogeneous, making outcome prediction challenging. We hypothesized two subgroups of grade 2 meningiomas which biologically resemble either World Health Organization (WHO) grade 1 or WHO grade 3. Our aim was to establish gene expression signatures that separate grade 2 meningiomas into two homogeneous subgroups: a more indolent subtype genetically resembling grade 1 and a more aggressive subtype resembling grade 3. METHODS: We carried out an observational meta-analysis on 212 meningiomas from six distinct studies retrieved from the open-access platform Gene Expression Omnibus. Microarray data was analyzed with systems-level gene co-expression network analysis. Fuzzy C-means clustering was employed to reclassify 34 of the 46 grade 2 meningiomas (74%) into a benign "grade 1-like" (13/46), and malignant "grade 3-like" (21/46) subgroup based on transcriptomic profiles. We verified shared biology between matching subgroups based on meta-gene expression and recurrence rates. These results were validated further using an independent RNA-seq dataset with 160 meningiomas, with similar results. RESULTS: Recurrence rates of "grade 1-like" and "grade 3- like" tumors were 0 and 75%, respectively, statistically similar to recurrence rates of grade 1 (17%) and 3 (85%). We also found overlapping biological processes of new subgroups with their adjacent grades 1 and 3. CONCLUSION: These results underpin molecular signatures as complements to histological grading systems. They may help reshape prediction, follow-up planning, treatment decisions and recruitment protocols for future and ongoing clinical trials.

Performance of preoperative breast MRI based on breast cancer molecular subtype.

PURPOSE: To assess the performance of preoperative breast MRI biopsy recommendations based on breast cancer molecular subtype. METHODS: All preoperative breast MRIs at a single academic medical center from May 2010 to March 2014 were identified. Reports were reviewed for biopsy recommendations. All pathology reports were reviewed to determine biopsy recommendation outcomes. Molecular subtypes were defined as Luminal A (ER/PR+ and HER2-), Luminal B (ER/PR+ and HER2+), HER2 (ER-, PR- and HER2+), and Basal (ER-, PR-, and HER2-). Logistic regression assessed the probability of true positive versus false positive biopsy and mastectomy versus lumpectomy. RESULTS: There were 383 patients included with a molecular subtype distribution of 253 Luminal A, 44 Luminal B, 20 HER2, and 66 Basal. Two hundred and thirteen (56%) patients and 319 sites were recommended for biopsy. Molecular subtype did not influence the recommendation for biopsy (p = 0.69) or the number of biopsy site recommendations (p = 0.30). The positive predictive value for a biopsy recommendation was 42% overall and 46% for Luminal A, 43% for Luminal B, 36% for HER2, and 29% for Basal subtype cancers. The multivariate logistic regression model showed no difference in true positive biopsy rate based on molecular subtype (p = 0.78). Fifty-one percent of patients underwent mastectomy and the multivariate model demonstrated that only a true positive biopsy (odds ratio: 5.3) was associated with higher mastectomy rates. CONCLUSION: Breast cancer molecular subtype did not influence biopsy recommendations, positive predictive values, or surgical approaches. Only true positive biopsies increased the mastectomy rate.

Ethical concerns around use of artificial intelligence in health care research from the perspective of patients with meningioma, caregivers and health care providers: a qualitative study.

BACKGROUND: As artificial intelligence (AI) approaches in research increase and AI becomes more integrated into medicine, there is a need to understand perspectives from members of the Canadian public and medical community. The aim of this project was to investigate current perspectives on ethical issues surrounding AI in health care. METHODS: In this qualitative study, adult patients with meningioma and their caregivers were recruited consecutively (August 2018-February 2019) from a neurosurgical clinic in Toronto. Health care providers caring for these patients were recruited through snowball sampling. Based on a nonsystematic literature search, we constructed 3 vignettes that sought participants' views on hypothetical issues surrounding potential AI applications in health care. The vignettes were presented to participants in interviews, which lasted 15-45 minutes. Responses were transcribed and coded for concepts, frequency of response types and larger concepts emerging from the interview. RESULTS: We interviewed 30 participants: 18 patients, 7 caregivers and 5 health care providers. For each question, a variable number of responses were recorded. The majority of participants endorsed nonconsented use of health data but advocated for disclosure and transparency. Few patients and caregivers felt that allocation of health resources should be done via computerized output, and a majority stated that it was inappropriate to delegate such decisions to a computer. Almost all participants felt that selling health data should be prohibited, and a minority stated that less privacy is acceptable for the goal of improving health. Certain caveats were identified, including the desire for deidentification of data and use within trusted institutions. INTERPRETATION: In this preliminary study, patients and caregivers reported a mixture of hopefulness and concern around the use of AI in health care research, whereas providers were generally more skeptical. These findings provide a point of departure for institutions adopting health AI solutions to consider the ethical implications of this work by understanding stakeholders' perspectives.