Deep Phenotyping and Prediction of Long-term Cardiovascular Disease: Optimized by Machine Learning.

BACKGROUND: Prediction of cardiovascular disease (CVD) is important in clinical practice. Machine learning (ML) may offer an improved alternative to current CVD risk stratification in individual patients. We aim to identify important predictors and compare ML models with traditional models according to their prediction performance in a large long-term follow-up cohort. METHODS: The Atherosclerosis Risk in Communities (ARIC) study was designed to study the progression of subclinical disease to cardiovascular events over a 25-year follow-up period. All phenotypic variables at visit 1 were obtained. All-cause death, CVD, and coronary heart disease were the outcomes for analysis. The ML framework involved variable selection using the random survival forest (RSF) method, model building, and 5-fold cross-validation. Model performance was evaluated by discrimination using the Harrell concordance index (C-index), accuracy using the Brier score (BS), and interpretability using the number of variables in the model. RESULTS: Of the 14,842 participants in ARIC, the average age was 54.2 years, with 45.2% male and 26.2% Black participants. Thirty-eight unique variables were selected in the RSF top 20 importance ranking of all 6 outcomes. Aging, hypertension, glucose metabolism, renal function, coagulation, adiposity, and sodium retention dominated the predictions of all outcomes. The ML models outperformed the regression models and established risk scores with a higher C-index, lower BS, and varied interpretability. CONCLUSIONS: The ML framework is useful for identifying important predictors of CVD and for developing models with robust performance compared with existing risk models.

Efficacy, immunogenicity, and safety of a quadrivalent HPV vaccine in men: results of an open-label, long-term extension of a randomised, placebo-controlled, phase 3 trial.

BACKGROUND: The quadrivalent human papillomavirus (HPV) vaccine was shown to prevent infections and lesions related to HPV6, 11, 16, and 18 in a randomised, placebo-controlled study in men aged 16-26 years. We assessed the incidences of external genital warts related to HPV6 or 11, and external genital lesions and anal dysplasia related to HPV6, 11, 16, or 18, over 10 years of follow-up. METHODS: The 3-year base study was an international, multicentre, double-blind, randomised, placebo-controlled trial done at 71 sites in 18 countries. Eligible participants were heterosexual men (aged 16-23 years) or men who have sex with men (MSM; aged 16-26 years). Men who had clinically detectable anogenital warts or genital lesions at screening that were suggestive of infection with non-HPV sexually transmitted diseases, or who had a history of such findings, were excluded. Eligible participants were randomly assigned (1:1) to receive three doses of either quadrivalent HPV vaccine or placebo on day 1, month 2, and month 6, administered as a 0·5-mL injection into the deltoid muscle. The 7-year, open-label, long-term follow-up extension study was done at 46 centres in 16 countries. Participants who received one or more doses of the quadrivalent HPV vaccine in the base study were eligible for enrolment into the long-term follow-up study (early vaccination group). Placebo recipients were offered the three-dose quadrivalent HPV vaccine at the end of the base study; those who received one or more quadrivalent HPV vaccine doses were eligible for enrolment into the long-term follow-up study (catch-up vaccination group). The primary efficacy endpoints were the incidence of external genital warts related to HPV6 or 11 and the incidence of external genital lesions related to HPV6, 11, 16, or 18 in all participants and the incidence of anal intraepithelial neoplasia (including anal warts and flat lesions) or anal cancer related to HPV6, 11, 16, or 18 in MSM only. The primary efficacy analysis was done in the per-protocol population for the early vaccination group, which included participants who received all three vaccine doses, were seronegative at day 1 and PCR-negative from day 1 through month 7 of the base study for the HPV type being analysed, had no protocol violations that could affect evaluation of vaccine efficacy, and had attended at least one visit during the long-term follow-up study. For the catch-up vaccination group, efficacy was assessed in the modified intention-to-treat population, which included participants who had received at least one vaccine dose, were seronegative and PCR-negative for HPV types analysed from day 1 of the base study to the final follow-up visit before receiving the quadrivalent HPV vaccine, and had at least one long-term follow-up visit. Safety was assessed in all randomised participants who received at least one vaccine dose. This study is registered with ClinicalTrials.gov, NCT00090285. FINDINGS: Between Aug 10, 2010, and April 3, 2017, 1803 participants were enrolled in the long-term follow-up study, of whom 936 (827 heterosexual men and 109 MSM) were included in the early vaccination group and 867 (739 heterosexual men and 128 MSM) were included in the catch-up vaccination group. Participants in the early vaccination group were followed up for a median of 9·5 years (range 0·1-11·5) after receiving the third dose of the quadrivalent HPV vaccine, and participants in the catch-up vaccination group were followed up for a median of 4·7 years (0·0-6·6) after receiving the third dose. In early vaccine group participants during long-term follow-up compared with the placebo group in the base study, the incidence per 10 000 person-years of external genital warts related to HPV6 or 11 was 0·0 (95% CI 0·0-8·7) versus 137·3 (83·9-212·1), of external genital lesions related to HPV6, 11, 16, or 18 was 0·0 (0·0-7·7) versus 140·4 (89·0-210·7), and of anal intraepithelial neoplasia or anal cancer related to HPV6, 11, 16, or 18 in MSM only was 20·5 (0·5-114·4) versus 906·2 (553·5-1399·5). Compared with during the base study (ie, before quadrivalent HPV vaccine administration), during the long-term follow-up period, participants in the catch-up vaccination group had no new reported cases of external genital warts related to HPV6 or 11 (149·6 cases per 10 000 person-years [95% CI 101·6-212·3] vs 0 cases per 10 000 person-years [0·0-13·5]) or external genital lesions related to HPV6, 11, 16, or 18 (155·1 cases per 10 000 person-years [108·0-215·7] vs 0 cases per 10 000 person-years [0·0-10·2]), and a lower incidence of anal intraepithelial neoplasia or anal cancer related to HPV6, 11, 16, or 18 (886·0 cases per 10 000 person-years [583·9-1289·1] vs 101·3 cases per 10 000 person-years [32·9-236·3]). No vaccine-related serious adverse events were reported. INTERPRETATION: The quadrivalent HPV vaccine provides durable protection against anogenital disease related to HPV6, 11, 16, and 18. The results support quadrivalent HPV vaccination in men, including catch-up vaccination. FUNDING: Merck Sharp & Dohme.