Breast cancer (BC) is known as the most prevalent form of cancer among women. Recent research has demonstrated the potential of Machine Learning (ML) techniques in predicting the five-year BC risk using personal health data. Support Vector Machine (SVM), Random Forest, K-NN (K-Nearest Neighbour), Naive Bayes, Neural Network, Decision Tree (DT), Logistic Regression (LR), Discriminant Analysis, and their variants are commonly employed in ML for BC analysis. This study investigates the factors influencing the performance of ML techniques in the domain of BC prevention, with a focus on dataset size and feature selection. The study's goal is to examine the effect of dataset cardinality, feature selection, and model selection on analytical performance in terms of Accuracy and Area Under the Curve (AUC). To this aim, 3917 papers were automatically selected from Scopus and PubMed, considering all publications from the previous 5 years, and, after inclusion and exclusion criteria, 54 articles were selected for the analysis. Our findings highlight how a good cardinality of the dataset and effective feature selection have a higher impact on the model's performance than the selected model, as corroborated by one of the studies, which gets extremely good results with all of the models employed.

Breast cancer (BC) has overtaken lung cancer as the most common cancer in the world and the projected incidence rates show a further increase. Early detection through population screening remains the cornerstone of BC control, but a progressive change from early diagnosis only-based to a personalized preventive and risk-reducing approach is widely debated. Risk-stratification models, which also include personal lifestyle risk factors, are under evaluation, although the documentation burden to gather population-based data is relevant and traditional data collection methods show some limitations. This paper provides the preliminary results from the analysis of clinical data provided by radiologists and lifestyle data collected using self-administered questionnaires from 5601 post-menopausal women. The weight of the combinations of women’s personal features and lifestyle habits on the BC risk were estimated by combining a model-driven and a data-driven approach to analysis. The weight of each factor on cancer occurrence was assessed using a logistic model. Additionally, communities of women sharing common features were identified and combined in risk profiles using social network analysis techniques. Our results suggest that preventive programs focused on increasing physical activity should be widely promoted, in particular among the oldest women. Additionally, current findings suggest that pregnancy, breast-feeding, salt limitation, and oral contraception use could have different effects on cancer risk, based on the overall woman’s risk profile. To overcome the limitations of our data, this work also introduces a mobile health tool, the Dress-PINK, designed to collect real patients’ data in an innovative way for improving women’s response rate, data accuracy, and completeness as well as the timeliness of data availability. Finally, the tool provides tailored prevention messages to promote critical consciousness, critical thinking, and increased health literacy among the general population.

Background: Breast cancer in Italy is still the most frequent cancer among women, and alcohol consumption is recognized as a risk factor for its development. Overall, in 2020, approximately 10% of all breast cancer-related deaths were attributable to alcohol consumption. Despite advancements in diagnostics and therapeutic options reducing mortality trends, the incidence of breast cancer is projected to rise in Italy. This study aims to assess how alcohol consumption influences the timing of breast lesion diagnosis. Understanding these associations can enhance primary prevention strategies and support the adoption of a risk-based prevention approach, integrating lifestyle factors into personalized screening programs. Methods: P.I.N.K. (Prevention, Imaging, Network and Knowledge) study collected data on a prospective dynamic cohort of women who voluntarily underwent breast cancer screening at breast centers throughout Italy, between 2018 and 2023, outside the free national screening program. The occurrence of breast lesion diagnosis and baseline information were collected through clinical visits and an auto-administered questionnaire, including data on absent, moderate or high alcohol consumption during the last 12 months and smoking. 3774 women (mean age 58.9 ± 10.0, range 40–98 years) were included in the present analysis, encompassing women with a suspected or confirmed diagnosis of benign or malignant tumor and healthy women that contributed at least 4 years to the study. An Event History Analysis was carried out to evaluate the effect of alcohol consumption on the timing to event. The event was represented by the transition of the health status, from not diagnosed to diagnosed with breast lesion. The Accelerated Failure Time parameterization was used to directly interpret how the covariates influence the time to the event. The model was adjusted by familiality of breast/ovarian cancer, marital status, level of education, and type of access to health care. Results: High alcohol consumption exhibited an accelerating effect on the transition to the diagnosed state, indicating a significantly shortened time to event: β coefficient − 0.33 (p-value 0.010) in the adjusted model, indicating an anticipation of about 4 months. The effect of moderate alcohol consumption did not reach statistical significance, neither in the unadjusted model nor in the adjusted model. Adjustment for smoking status led to a further increase of the β coefficient for high alcohol consumption (− 0.40; p value 0.003) and brought moderate alcohol consumption closer to statistical significance (β − 0.15; p-value 0.087). Familiality of breast or ovarian cancer showed a statistically non-significant accelerating effect, while marital status different from maiden, high education, and private access to health care showed decelerating effects. Conclusions: High alcohol consumption was confirmed as an accelerating factor in breast lesions diagnosis, while the effect of moderate consumption did not reach statistical significance. These results help identifying actionable targets for high-risk populations, emphasizing personalized risk-based screening programs and gender-sensitive interventions.

Background: The COVID–19 pandemic disrupted patterns of substance consumption, but its impact on the structure of co–occurring risk behaviours remains unclear.Methods: Nationally representative data from the cross–sectional Italian adult Population Survey on Alcohol and other Drugs (n = 11,131) and Psychometric Network Analysis were used to estimate conditional dependencies between alcohol abuse (CAGE >2), substance use, behavioural risks, and psychotropic medications. Network Comparison Test and Bayesian models assessed structural changes and edge robustness.Findings: The overall network architecture remained stable. However, critical localised reconfigurations emerged. The CAGE node transitioned from a peripheral to a more integrated node (Expected Influence –0⋅070 to 0⋅180), alongside a reversal in its association with at–risk cannabis use (Edge Weight –0⋅129 to 0⋅241, p–value = 0⋅002), which was confirmed by Bayesian analysis (Edge Inclusion 0⋅047 to ∞). Problematic gambling emerged as a key bridge node (Betweenness –0⋅828 to 1⋅742), linking substance use and psychological distress, with its association shifting from anxiety–related (tranquillizers; Edge Weight 0⋅275 to –0⋅150) to depressive dynamics (antidepressants; Edge Weight –0⋅139 to 0⋅230). Simultaneously, engagement in regular physical activity declined (prevalence 50⋅5% to 47⋅4%, p–value = 0⋅002), while its inhibitory influence strengthened (Expected Influence –1⋅711 to –1⋅924).Interpretation: The pandemic appears to have reconfigured risk behaviours into a more tightly interconnected system, where alcohol abuse, behavioural addictions, and psychological distress were structurally integrated. These findings suggest that prevention strategies should move from single–behaviour approaches addressing co–occurrence of substance use, behavioural addiction and mental health vulnerabilities.Funding. None.

Adolescents are heavy users of digital media but often lack critical skills, increasing their vulnerability to harmful online content. The integration of game elements into learning and training offers a promising strategy to support positive behavioural change and strengthen adolescents’ skills. This paper describes the development of a conceptual framework for Dress-DIGITARIAN, a serious game aimed at improving health literacy, coping skills, and self-esteem, grounded in Self-Determination Theory (SDT). The framework was constructed to generate higher-order understanding through a multi-level process: analyzing general theory (SDT), integrating mid-range models (the Octalysis framework), and incorporating empirical insights derived from two data collection phases with the target population. This integrative approach informed and guided the game’s design through participatory methods. Developed through collaboration between schools and research institutions, this approach bridges theory and practice by aligning game mechanics with adolescents’ psychological needs. It also underscores the value of involving adolescents in research, not only to enhance scientific rigour but also to empower them as agents of change capable of contributing to health promotion policies and educational innovation. This study does not report the results of a completed intervention or outcome evaluation, which will be conducted in the sixth phase at the end of the current school year. Future research is needed to assess the model’s effectiveness and scalability and to identify areas for further refinement.