Radiomics analysis to predict pulmonary nodule malignancy using machine learning approaches.

BACKGROUND: Low-dose CT screening can reduce lung cancer-related mortality. However, most screen-detected pulmonary abnormalities do not develop into cancer and it often remains challenging to identify malignant nodules, particularly among indeterminate nodules. We aimed to develop and assess prediction models based on radiological features to discriminate between benign and malignant pulmonary lesions detected on a baseline screen. METHODS: Using four international lung cancer screening studies, we extracted 2060 radiomic features for each of 16 797 nodules (513 malignant) among 6865 participants. After filtering out low-quality radiomic features, 642 radiomic and 9 epidemiological features remained for model development. We used cross-validation and grid search to assess three machine learning (ML) models (eXtreme Gradient Boosted Trees, random forest, least absolute shrinkage and selection operator (LASSO)) for their ability to accurately predict risk of malignancy for pulmonary nodules. We report model performance based on the area under the curve (AUC) and calibration metrics in the held-out test set. RESULTS: The LASSO model yielded the best predictive performance in cross-validation and was fit in the full training set based on optimised hyperparameters. Our radiomics model had a test-set AUC of 0.93 (95% CI 0.90 to 0.96) and outperformed the established Pan-Canadian Early Detection of Lung Cancer model (AUC 0.87, 95% CI 0.85 to 0.89) for nodule assessment. Our model performed well among both solid (AUC 0.93, 95% CI 0.89 to 0.97) and subsolid nodules (AUC 0.91, 95% CI 0.85 to 0.95). CONCLUSIONS: We developed highly accurate ML models based on radiomic and epidemiological features from four international lung cancer screening studies that may be suitable for assessing indeterminate screen-detected pulmonary nodules for risk of malignancy.

Severity of Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) Infection in Pregnancy in Ontario: A Matched Cohort Analysis.

BACKGROUND: Pregnancy represents a physiological state associated with increased vulnerability to severe outcomes from infectious diseases, both for the pregnant person and developing infant. The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pandemic may have important health consequences for pregnant individuals, who may also be more reluctant than nonpregnant people to accept vaccination. METHODS: We sought to estimate the degree to which increased severity of SARS-CoV-2 outcomes can be attributed to pregnancy using a population-based SARS-CoV-2 case file from Ontario, Canada. Because of varying propensity to receive vaccination, and changes in dominant circulating viral strains over time, a time-matched cohort study was performed to evaluate the relative risk of severe illness in pregnant women with SARS-CoV-2 compared to other SARS-CoV-2 infected women of childbearing age (10-49 years old). Risk of severe SARS-CoV-2 outcomes was evaluated in pregnant women and time-matched nonpregnant controls using multivariable conditional logistic regression. RESULTS: Compared with the rest of the population, nonpregnant women of childbearing age had an elevated risk of infection (standardized morbidity ratio, 1.28), whereas risk of infection was reduced among pregnant women (standardized morbidity ratio, 0.43). After adjustment for confounding, pregnant women had a markedly elevated risk of hospitalization (adjusted odds ratio, 4.96; 95% confidence interval, 3.86-6.37) and intensive care unit admission (adjusted odds ratio, 6.58; 95% confidence interval, 3.29-13.18). The relative increase in hospitalization risk associated with pregnancy was greater in women without comorbidities than in those with comorbidities (P for heterogeneity, .004). CONCLUSIONS: Given the safety of SARS-CoV-2 vaccines in pregnancy, risk-benefit calculus strongly favors SARS-CoV-2 vaccination in pregnant women.

Relative Virulence of Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) Among Vaccinated and Unvaccinated Individuals Hospitalized With SARS-CoV-2.

BACKGROUND: The rapid development of safe and effective vaccines against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has been a singular scientific achievement. Confounding due to health-seeking behaviors, circulating variants, and differential testing by vaccination status may bias analyses toward an apparent increase in infection severity following vaccination. METHODS: We used data from the Ontario, Canada, Case and Contact Management Database and a provincial vaccination dataset (COVaxON) to create a time-matched cohort of individuals who were hospitalized with SARS-CoV-2 infection. Vaccinated individuals were matched to up to 5 unvaccinated individuals based on test date. Risk of intensive care unit (ICU) admission and death were evaluated using conditional logistic regression. RESULTS: In 20 064 individuals (3353 vaccinated and 16 711 unvaccinated) hospitalized with infection due to SARS-CoV-2 between 1 January 2021 and 5 January 2022, vaccination with 1, 2, or 3 doses significantly reduced the risk of ICU admission and death. An inverse dose-response relationship was observed between vaccine doses received and both outcomes (adjusted odds ratio [aOR] per additional dose for ICU admission, 0.66; 95% confidence interval [CI], .62 to .71; aOR for death, 0.78; 95% CI, .72 to .84). CONCLUSIONS: We identified decreased virulence of SARS-CoV-2 infections in vaccinated individuals, even when vaccines failed to prevent infection sufficiently severe to cause hospitalization. Even with diminished efficacy of vaccines against infection with novel variants of concern, vaccines remain an important tool for reduction of ICU admission and mortality.

Vaccine effectiveness against hospitalization among adolescent and pediatric SARS-CoV-2 cases between May 2021 and January 2022 in Ontario, Canada: A retrospective cohort study.

BACKGROUND: Vaccines against SARS-CoV-2 have been shown to reduce risk of infection as well as severe disease among those with breakthrough infection in adults. The latter effect is particularly important as immune evasion by Omicron variants appears to have made vaccines less effective at preventing infection. Therefore, we aimed to quantify the protection conferred by mRNA vaccination against hospitalization due to SARS-CoV-2 in adolescent and pediatric populations. METHODS: We retrospectively created a cohort of reported SARS-CoV-2 case records from Ontario's Public Health Case and Contact Management Solution among those aged 4 to 17 linked to vaccination records from the COVaxON database on January 19, 2022. We used multivariable logistic regression to estimate the association between vaccination and hospitalization among SARS-CoV-2 cases prior to and during the emergence of Omicron. RESULTS: We included 62 hospitalized and 27,674 non-hospitalized SARS-CoV-2 cases, with disease onset from May 28, 2021 to December 4, 2021 (Pre-Omicron) and from December 23, 2021 to January 9, 2022 (Omicron). Among adolescents, two mRNA vaccine doses were associated with an 85% (aOR = 0.15; 95% CI: [0.04, 0.53]; p<0.01) lower likelihood of hospitalization among SARS-CoV-2 cases caused by Omicron. Among children, one mRNA vaccine dose was associated with a 79% (aOR = 0.21; 95% CI: [0.03, 0.77]; p<0.05) lower likelihood of hospitalization among SARS-CoV-2 cases caused by Omicron. The calculation of E-values, which quantifies how strong an unmeasured confounder would need to be to nullify our findings, suggest that these effects are unlikely to be explained by unmeasured confounding. CONCLUSIONS: Despite immune evasion by SARS-CoV-2 variants, vaccination continues to be associated with a lower likelihood of hospitalization among adolescent and pediatric Omicron (B.1.1.529) SARS-CoV-2 cases, even when the vaccines do not prevent infection. Continued efforts are needed to increase vaccine uptake among adolescent and pediatric populations.