Assessment of performance of the Gail model for predicting breast cancer risk: a systematic review and meta-analysis with trial sequential analysis

Abstract

BackgroundThe Gail model has been widely used and validated with conflicting results. The current study aims to evaluate the performance of different versions of the Gail model by means of systematic review and meta-analysis with trial sequential analysis (TSA).MethodsThree systematic review and meta-analyses were conducted. Pooled expected-to-observed (E/O) ratio and pooled area under the curve (AUC) were calculated using the DerSimonian and Laird random-effects model. Pooled sensitivity, specificity and diagnostic odds ratio were evaluated by bivariate mixed-effects model. TSA was also conducted to determine whether the evidence was sufficient and conclusive.ResultsGail model 1 accurately predicted breast cancer risk in American women (pooled E/O = 1.03; 95% CI 0.76–1.40). The pooled E/O ratios of Caucasian-American Gail model 2 in American, European and Asian women were 0.98 (95% CI 0.91–1.06), 1.07 (95% CI 0.66–1.74) and 2.29 (95% CI 1.95–2.68), respectively. Additionally, Asian-American Gail model 2 overestimated the risk for Asian women about two times (pooled E/O = 1.82; 95% CI 1.31–2.51). TSA showed that evidence in Asian women was sufficient; nonetheless, the results in American and European women need further verification.The pooled AUCs for Gail model 1 in American and European women and Asian females were 0.55 (95% CI 0.53–0.56) and 0.75 (95% CI 0.63–0.88), respectively, and the pooled AUCs of Caucasian-American Gail model 2 for American, Asian and European females were 0.61 (95% CI 0.59–0.63), 0.55 (95% CI 0.52–0.58) and 0.58 (95% CI 0.55–0.62), respectively.The pooled sensitivity, specificity and diagnostic odds ratio of Gail model 1 were 0.63 (95% CI 0.27–0.89), 0.91 (95% CI 0.87–0.94) and 17.38 (95% CI 2.66–113.70), respectively, and the corresponding indexes of Gail model 2 were 0.35 (95% CI 0.17–0.59), 0.86 (95% CI 0.76–0.92) and 3.38 (95% CI 1.40–8.17), respectively.ConclusionsThe Gail model was more accurate in predicting the incidence of breast cancer in American and European females, while far less useful for individual-level risk prediction. Moreover, the Gail model may overestimate the risk in Asian women and the results were further validated by TSA, which is an addition to the three previous systematic review and meta-analyses.Trial registrationPROSPERO CRD42016047215.Electronic supplementary materialThe online version of this article (10.1186/s13058-018-0947-5) contains supplementary material, which is available to authorized users.

Assessment of the Gail Model in Estimating the Risk of Breast Cancer: Effect of Cancer Worry and Risk in Healthy Women

“…The Gail model might overestimate the risk of and development of breast cancer in other populations in which other races predominate (Novotny et al, 2006;Andreeva and Pokhrel 2013;Wang et al, 2018). However, utilization of the Gail model for Turkish population and other populations with white women predominance might provide more accurate estimations (Tice et al, 2005;Min et al, 2014;Wang et al, 2018). Accordingly, the five-year and life-time estimated risk for breast cancer was not higher in our study than different studies (Tice et al, 2005;Min et al, 2014;Khazaee-Pool et al, 2016, Wang et al, 2018.…”

“…However, utilization of the Gail model for Turkish population and other populations with white women predominance might provide more accurate estimations (Tice et al, 2005;Min et al, 2014;Wang et al, 2018). Accordingly, the five-year and life-time estimated risk for breast cancer was not higher in our study than different studies (Tice et al, 2005;Min et al, 2014;Khazaee-Pool et al, 2016, Wang et al, 2018. Unfortunately, the Gail model for breast cancer risk prediction of consanguineous marriages for fist degree and second-degree relatives is not included as a risk factor.…”

“…The best-known statistical model available for predicting an individual woman's chance of developing breast cancer is that derived using information from regularly screened Caucasian women from the United States participating in the Breast Cancer screening and detection project (Gail et al 1989;Costantino et al, 1999;Gail et al, 2007;Adams -Campbell et al, 2009). The Gail model might overestimate the risk of and development of breast cancer in other populations in which other races predominate (Novotny et al, 2006;Andreeva and Pokhrel 2013;Wang et al, 2018). However, utilization of the Gail model for Turkish population and other populations with white women predominance might provide more accurate estimations (Tice et al, 2005;Min et al, 2014;Wang et al, 2018).…”

Machine learning-based lifetime breast cancer risk reclassification compared with the BOADICEA model: impact on screening recommendations

“…18,19 Many medical societies and professional groups proposed that risk-based screening would be more effective, less morbid and more cost-effective. 3,[19][20][21][22][23][24] Although many models are used to predict breast cancer risk, such as the Breast Cancer Risk Assessment Tool (BCRAT, also referred as the Gail model), the International Breast Intervention Study (IBIS) model, the Breast and Ovarian Analysis of Disease Incidence and Carrier Estimation Algorithm (BOADICEA model), [25][26][27] no consistent model has been incorporated into routine clinical practice and/or screening programmes due to limited discriminatory accuracy and applicability. The discriminatory ability, area under the receiver operating characteristic (AU-ROC) curve, of these models is between 0.53 and 0.64.…”

Letter to the editor: Response to Giardiello D, Antoniou AC, Mariani L, Easton DF, Steyerberg EW

“…Giardiello and colleagues mentioned that our machine learning (ML) models were not specific for survival data. BCRAT/BOADICEA were developed and validated using survival data with binary outcomes and retrospective case control/cross-sectional data, respectively [3]. Their clinical application requires only cross-sectional data.…”