Supplementary Figure 1. Forest plot of prevalence of breast cancer screening among cancer survivors. The estimated prevalence of breast cancer screening for each study is presented numerically and graphically, with point estimates shown as boxes scaled by study weight and whiskers representing the 95% confidence intervals (CIs). The overall pooled prevalence, calculated using a random-effects model, is represented by a diamond, with its width corresponding to the 95% CI.

Supplementary Figure 4. Pooled prevalence of breast cancer screening by study design. The pooled prevalence estimates of breast cancer screening, calculated using random-effects models, are displayed as diamonds with whiskers indicating 95% confidence intervals (CIs). Estimates were pooled separately by study design (cross-sectional vs. case-control vs. cohort), using the chi-square statistic to test for differences in pooled estimates across subgroups (p=0.344).

Supplementary Figure 12. Funnel plot of odds ratio for breast cancer screening among cancer survivors compared to the cancer-free general population. The outer dashed lines indicate the triangular region within which 95% of studies are expected to lie in the absence of bias. The solid black line corresponds to the summary effect estimate. Black dots correspond to studies included in this analysis.

Background and Aims:Colonoscopy is an effective screening tool, but it is resource-intensive and carries procedural risks. We aimed to develop and validate a biomarker-based risk prediction model (RPM) to improve triage of average-risk patients for colonoscopy by better predicting high-risk adenomas (HRAs). Methods:We assessed blood-based biomarkers of glucose metabolism, liver enzymes, lipids, ferritin, and carcinoembryonic antigen (CEA) in 980 average-risk individuals undergoing screening colonoscopy. Logistic regression was used to predict HRAs, defined as adenomas ≥10 mm, villous histology, high-grade dysplasia, or ≥3 adenomas. Biomarkers were modeled as continuous variables; variable selection was performed using logistic regression and least absolute shrinkage and selection operator. Model performance was evaluated with area under the curve, reclassification indices, calibration statistics, bootstrap validation, and decision curve analysis. Results:Glucose and CEA were significantly associated with HRAs, while no biomarkers were linked to high-risk sessile serrated lesions. Nine RPMs were developed; 3 were selected for validation. A model including all univariately associated biomarkers (CEA, C-peptide, gamma-glutamyl transferase, glucose, hemoglobin A1c, triglycerides) achieved an optimism-adjusted area under the curve of 0.66 (95% confidence interval, 0.60-0.71). Reclassification indices showed that adding biomarkers improved stratification, with the all-biomarkers model yielding the highest net reclassification improvement. Conclusion:Biomarker-enhanced RPMs modestly improved prediction of HRAs compared with models using only clinical and demographic variables. These findings support the potential of biomarkers to optimize colonoscopy resource allocation, but additional markers and refinement are needed to establish clinical utility.

Supplementary Table 3. Study quality assessment for both study aims. Table 3A displays the study quality for each study included in Aim 1, while Table 3B (case-control studies) and Table 3C (cross-sectional studies) display the study quality for each study in Aim 2.

Supplementary Figure 3. Pooled prevalence of breast cancer screening by screening method. The pooled prevalence estimates of breast cancer screening, calculated using random-effects models, are displayed as diamonds with whiskers indicating 95% confidence intervals (CIs). Estimates were pooled separately by method of screening, categorized as mammography only, mammography plus other screening method, or other non-mammography screening method only (including clinical breast examination, magnetic resonance imaging, or breast sonography). The chi-square statistic was used to test for differences in pooled estimates across subgroups (p=0.450).

Supplementary Figure 5. Pooled prevalence of breast cancer screening by study quality. The pooled prevalence estimates of breast cancer screening, calculated using random-effects models, are displayed as diamonds with whiskers indicating 95% confidence intervals (CIs). Estimates were pooled separately based on study quality (low, moderate or high, assessed using revised version of the tool proposed by Hoy et al. (33)). The chi-square statistic was used to test for differences in pooled estimates across subgroups (p=0.509).

Supplementary Figure 19. Pooled odds ratio of breast cancer screening comparing cancer survivors to the cancer-free general population, stratified by period of assessment. The odds ratio (OR) reported in each study is presented numerically and graphically, with point estimates shown as boxes scaled by study weight and whiskers representing the 95% confidence intervals (CIs). Estimates were pooled separately according to whether studies assessed screening starting prior to or after 2010. The chi-square statistic was used to test for differences in pooled estimates across subgroups (p=0.620). Pooled ORs, calculated using a random-effects model, are represented by diamonds with their width corresponding to the 95% CI.

Supplementary Figure 14. Sensitivity analysis for meta-analysis of the odds ratio for breast cancer screening excluding studies examining breast cancer screening other than screening mammography. The odds ratio (OR) comparing cancer survivors to the cancer-free general population reported in each study is presented numerically and graphically, with point estimates shown as boxes scaled by study weight and whiskers representing the 95% confidence intervals (CIs). The overall pooled OR, calculated using a random-effects model, is represented by a diamond, with its width corresponding to the 95% CI. Studies which assessed screening methods other than mammography were excluded.

Supplementary Figure 11. Forest plot of the odds ratio for breast cancer screening, comparing cancer survivors to the cancer-free general population. The odds ratio (OR) reported in each study is presented numerically and graphically, with point estimates shown as boxes scaled by study weight and whiskers representing the 95% confidence intervals (CIs). The overall pooled OR, calculated using a random-effects model, is represented by a diamond, with its width corresponding to the 95% CI.

BACKGROUND:Melanoma is a rapidly increasing cancer in Canada, largely due to ultraviolet radiation (UVR) exposure. The objective of this study was to examine age- and sex-specific melanoma incident trends in Canada to determine if public health efforts related to UVR exposure are having an impact on melanoma incidence. METHODS:Data on melanoma incidence was obtained from the Canadian Cancer Registry (1992-2022). Annual percent changes in age-specific incidence rates were analyzed with segmented regression. Birth cohort effects were estimated with age-period-cohort models and reported as cohort incidence rate ratios (IRRs) with respect to the 1953-57 cohort. RESULTS:From 1992-2022, the incidence of melanoma has steadily increased for females over 40 and males over 50 with larger increases for older age groups. In contrast, melanoma rates have been decreasing for females under 30 and males under 40. Compared to the baby boom generation, recent birth cohorts (1998-2007 for males and 1993-2007 for females) have a lower incidence of melanoma. CONCLUSIONS:While melanoma rates continue to increase for older adults, incidence among younger Canadians is declining. These results may indicate that public health efforts are having an impact on melanoma prevention for recent birth cohorts.

Supplementary Figure 7. Pooled prevalence of breast cancer screening by time between cancer diagnosis and screening assessment. The pooled prevalence estimates of breast cancer screening, calculated using random-effects models, are displayed as diamonds with whiskers indicating 95% confidence intervals (CIs). Estimates were pooled separately based on time from cancer diagnosis to screening assessment, categorized as studies with mean/median time since diagnosis ≥5 years, <5 years, or including cancer survivors irrespective of and not reporting on time since diagnosis (i.e., ever diagnosed).The chi-square statistic was used to test for differences in pooled estimates across subgroups (p=0.206).

Supplementary Figure 2. Pooled prevalence of breast cancer screening by age eligibility and screening definition. The pooled prevalence estimates of breast cancer screening, calculated using random-effects models, are displayed as diamonds with whiskers indicating 95% confidence intervals (CIs). Estimates were pooled separately by study population (screening-eligible age only [≥40 years] vs. inclusion of non-eligible ages) and by screening definition (screened within a specified time period vs. ever screened) and the chi-square statistic used to test for differences in pooled estimates across subgroups (p=0.565).

Supplementary Figure 17. Pooled odds ratio of breast cancer screening comparing cancer survivors to the cancer-free general population, stratified by study quality. The odds ratio (OR) reported in each study is presented numerically and graphically, with point estimates shown as boxes scaled by study weight and whiskers representing the 95% confidence intervals (CIs). Estimates were pooled separately based on study quality (assessed as low, moderate or high). The chi-square statistic was used to test for differences in pooled estimates across subgroups (p=0.557). Pooled ORs, calculated using a random-effects model, are represented by diamonds with their width corresponding to the 95% CI.

Supplementary Figure 13. Sensitivity analysis for meta-analysis of the odds ratio for breast cancer screening comparing cancer survivors to the cancer-free general population excluding studies recruiting ever-screened and age-ineligible participants. The odds ratio (OR) comparing cancer survivors to the cancer-free general population reported in each study is presented numerically and graphically, with point estimates shown as boxes scaled by study weight and whiskers representing the 95% confidence intervals (CIs). The overall pooled OR calculated using a random-effects model, is represented by a diamond, with its width corresponding to the 95% CI. Studies which included age-ineligible participants (<40 years old), or which defined screening participation as ever being screened were excluded.

Supplementary Figure 21. Pooled odds ratio of breast cancer screening comparing cancer survivors to the cancer-free general population, stratified by primary cancer site. The odds ratio (OR) reported in each study is presented numerically and graphically, with point estimates shown as boxes scaled by study weight and whiskers representing the 95% confidence intervals (CIs). The OR among studies including survivors of mixed cancer sites (A), colorectal (B), blood (C), cervical (D), and uterine cancer (E) were pooled separately, with site-specific pooled estimates (represented by a diamond) presented in each panel.

Supplementary Table 2. Number of studies by the method of screening ascertainment according to primary cancer sites for breast cancer screening prevalence

Cancer survivors are at increased risk of developing subsequent primary cancers, yet participation in recommended cancer screening remains suboptimal. Greater understanding of factors impacting screening participation is needed to inform strategies for improving uptake. Therefore, we aimed to review and quantitatively synthesize predictors of participation in breast cancer (BC) and colorectal cancer (CRC) screening among cancer survivors. MEDLINE (Ovid), EMBASE, PubMed, and CINAHL databases were searched from inception through September 2024 to identify studies examining predictors of CRC or BC screening among survivors of other adult cancers. Random effect models were used to estimate pooled associations for predictors reported in ≥ 3 studies. From 2492 initial citations, 49 studies were included, with 35 studies reporting on CRC and 29 on BC screening participation. Factors significantly associated with CRC screening included having a healthcare provider, having health insurance, receiving a provider recommendation for screening, a greater number of physician visits, older age, being a nonsmoker, being married, higher income or education, urban residence, and better mental health. Factors associated with BC screening included receiving a written follow-up care plan, a greater number of physician visits, being married, not having dual Medicaid-Medicare coverage status, receiving specialist care, and White racial identity. Healthcare access and survivorship care factors appear to play an important role in screening participation among cancer survivors. Interventions addressing barriers to care and social determinants of health may help improve screening uptake and reduce disparities in subsequent cancer screening in this high-risk population.

Supplementary Table 1. Database search strategy for this systematic review.

Supplementary Figure 9. Pooled prevalence of breast cancer screening by geographical region. The pooled prevalence estimates of breast cancer screening, calculated using random-effects models, are displayed as diamonds with whiskers indicating 95% confidence intervals (CIs). Estimates were pooled separately based on geographical region, categorized as studies conducted in North America or other, based on data availability. The chi-square statistic was used to test for differences in pooled estimates across subgroups (p=0.026).