Automated screening of computed tomography using weakly supervised anomaly detection.

BACKGROUND: Current artificial intelligence studies for supporting CT screening tasks depend on either supervised learning or detecting anomalies. However, the former involves a heavy annotation workload owing to requiring many slice-wise annotations (ground truth labels); the latter is promising, but while it reduces the annotation workload, it often suffers from lower performance. This study presents a novel weakly supervised anomaly detection (WSAD) algorithm trained based on scan-wise normal and anomalous annotations to provide better performance than conventional methods while reducing annotation workload. METHODS: Based on surveillance video anomaly detection methodology, feature vectors representing each CT slice were trained on an AR-Net-based convolutional network using a dynamic multiple-instance learning loss and a center loss function. The following two publicly available CT datasets were retrospectively analyzed: the RSNA brain hemorrhage dataset (normal scans: 12,862; scans with intracranial hematoma: 8882) and COVID-CT set (normal scans: 282; scans with COVID-19: 95). RESULTS: Anomaly scores of each slice were successfully predicted despite inaccessibility to any slice-wise annotations. Slice-level area under the curve (AUC), sensitivity, specificity, and accuracy from the brain CT dataset were 0.89, 0.85, 0.78, and 0.79, respectively. The proposed method reduced the number of annotations in the brain dataset by 97.1% compared to an ordinary slice-level supervised learning method. CONCLUSION: This study demonstrated a significant annotation reduction in identifying anomalous CT slices compared to a supervised learning approach. The effectiveness of the proposed WSAD algorithm was verified through higher AUC than existing anomaly detection techniques.

The effect of recreational cannabis legalization on rates of traffic injury in Canada.

AIMS: To measure the impact of Canada's recreational cannabis legalization (RCL) in October 2018 and the subsequent impact of the coronavirus disease 2019 (COVID-19) lockdowns from March 2020 on rates of emergency department (ED) visits and hospitalizations for traffic injury. DESIGN: An interrupted time series analysis of rates of ED visits and hospitalizations in Canada recorded in population-based databases from January/April 2010 to March 2021. SETTING: ED visits in Ontario and Alberta and hospitalizations in Ontario, Alberta, British Columbia, the Prairies (Manitoba and Saskatchewan) and the Maritimes (Nova Scotia, New Brunswick, Newfoundland and Prince Edward Island). PARTICIPANTS: Monthly counts of presentations to the ED or hospital for motor vehicle injury or pedestrian/cyclist injury, used to calculate monthly rates per 100 000 population. MEASUREMENTS: An occurrence of one or more International Statistical Classification of Diseases and Related Health Problems, 10th Revision, Canada (ICD-10-CA) code for motor vehicle injury (V20-V29, V40-V79, V30-V39 and V86) and pedestrian/cyclist injury (V01-V09 and V10-V19) within the National Ambulatory Care Reporting System and Discharge Abstract Database. FINDINGS: There were no statistically significant changes in rates of ED visits and hospitalizations for motor vehicle or pedestrian/cyclist injury after RCL after accounting for multiple testing. After COVID-19, there was an immediate decrease in the rate of ED visits for motor vehicle injury that was statistically significant only in Ontario (level change ß = -16.07 in Ontario, 95% CI = -20.55 to -11.60, P = 0.000; ß = -10.34 in Alberta, 95% CI = -17.80 to -2.89, P = 0.008; α of 0.004) and no changes in rates of hospitalizations. CONCLUSIONS: Canada's recreational cannabis legalization did not notably impact motor vehicle and pedestrian/cyclist injury. The rate of emergency department visits for motor vehicle injury decreased immediately after COVID-19 lockdowns, resulting in rates below post-recreational cannabis legalization levels in the year after COVID-19.