Risk prediction for estrogen receptor-specific breast cancers in two large prospective cohorts

Kuanrong Li, Garnet Anderson, Vivian Viallon, Patrick Arveux, Marina Kvaskoff, Agnès Fournier, Vittorio Krogh, Rosario Tumino, Maria-Jose Sánchez, Eva Ardanaz, María-Dolores Chirlaque, Antonio Agudo, David C Muller, Todd Smith, Ioanna Tzoulaki, Timothy J Key, Bas Bueno-de-Mesquita, Antonia Trichopoulou, Christina Bamia, Philippos Orfanos & 15 others Rudolf Kaaks, Anika Hüsing, Renée T Fortner, Anne Zeleniuch-Jacquotte, Malin Sund, Christina C Dahm, Kim Overvad, Dagfinn Aune, Elisabete Weiderpass, Isabelle Romieu, Elio Riboli, Marc J Gunter, Laure Dossus, Ross Prentice, Pietro Ferrari

Research output: Contribution to journalArticle

Abstract

BACKGROUND: Few published breast cancer (BC) risk prediction models consider the heterogeneity of predictor variables between estrogen-receptor positive (ER+) and negative (ER-) tumors. Using data from two large cohorts, we examined whether modeling this heterogeneity could improve prediction.

METHODS: We built two models, for ER+ (ModelER+) and ER- tumors (ModelER-), respectively, in 281,330 women (51% postmenopausal at recruitment) from the European Prospective Investigation into Cancer and Nutrition cohort. Discrimination (C-statistic) and calibration (the agreement between predicted and observed tumor risks) were assessed both internally and externally in 82,319 postmenopausal women from the Women's Health Initiative study. We performed decision curve analysis to compare ModelER+ and the Gail model (ModelGail) regarding their applicability in risk assessment for chemoprevention.

RESULTS: Parity, number of full-term pregnancies, age at first full-term pregnancy and body height were only associated with ER+ tumors. Menopausal status, age at menarche and at menopause, hormone replacement therapy, postmenopausal body mass index, and alcohol intake were homogeneously associated with ER+ and ER- tumors. Internal validation yielded a C-statistic of 0.64 for ModelER+ and 0.59 for ModelER-. External validation reduced the C-statistic of ModelER+ (0.59) and ModelGail (0.57). In external evaluation of calibration, ModelER+ outperformed the ModelGail: the former led to a 9% overestimation of the risk of ER+ tumors, while the latter yielded a 22% underestimation of the overall BC risk. Compared with the treat-all strategy, ModelER+ produced equal or higher net benefits irrespective of the benefit-to-harm ratio of chemoprevention, while ModelGail did not produce higher net benefits unless the benefit-to-harm ratio was below 50. The clinical applicability, i.e. the area defined by the net benefit curve and the treat-all and treat-none strategies, was 12.7 × 10- 6 for ModelER+ and 3.0 × 10- 6 for ModelGail.

CONCLUSIONS: Modeling heterogeneous epidemiological risk factors might yield little improvement in BC risk prediction. Nevertheless, a model specifically predictive of ER+ tumor risk could be more applicable than an omnibus model in risk assessment for chemoprevention.

Original languageEnglish
Pages (from-to)147
JournalBreast Cancer Research
Volume20
Issue number1
DOIs
Publication statusPublished - Dec 3 2018

Fingerprint

Estrogen Receptors
Breast Neoplasms
Neoplasms
Chemoprevention
Calibration
Pregnancy
Menarche
Estrogen Replacement Therapy
Body Height
Decision Support Techniques
Women's Health
Menopause
Parity
Body Mass Index
Alcohols

Cite this

Li, K., Anderson, G., Viallon, V., Arveux, P., Kvaskoff, M., Fournier, A., ... Ferrari, P. (2018). Risk prediction for estrogen receptor-specific breast cancers in two large prospective cohorts. Breast Cancer Research, 20(1), 147. https://doi.org/10.1186/s13058-018-1073-0

Risk prediction for estrogen receptor-specific breast cancers in two large prospective cohorts. / Li, Kuanrong; Anderson, Garnet; Viallon, Vivian; Arveux, Patrick; Kvaskoff, Marina; Fournier, Agnès; Krogh, Vittorio; Tumino, Rosario; Sánchez, Maria-Jose; Ardanaz, Eva; Chirlaque, María-Dolores; Agudo, Antonio; Muller, David C; Smith, Todd; Tzoulaki, Ioanna; Key, Timothy J; Bueno-de-Mesquita, Bas; Trichopoulou, Antonia; Bamia, Christina; Orfanos, Philippos; Kaaks, Rudolf; Hüsing, Anika; Fortner, Renée T; Zeleniuch-Jacquotte, Anne; Sund, Malin; Dahm, Christina C; Overvad, Kim; Aune, Dagfinn; Weiderpass, Elisabete; Romieu, Isabelle; Riboli, Elio; Gunter, Marc J; Dossus, Laure; Prentice, Ross; Ferrari, Pietro.

In: Breast Cancer Research, Vol. 20, No. 1, 03.12.2018, p. 147.

Research output: Contribution to journalArticle

Li, K, Anderson, G, Viallon, V, Arveux, P, Kvaskoff, M, Fournier, A, Krogh, V, Tumino, R, Sánchez, M-J, Ardanaz, E, Chirlaque, M-D, Agudo, A, Muller, DC, Smith, T, Tzoulaki, I, Key, TJ, Bueno-de-Mesquita, B, Trichopoulou, A, Bamia, C, Orfanos, P, Kaaks, R, Hüsing, A, Fortner, RT, Zeleniuch-Jacquotte, A, Sund, M, Dahm, CC, Overvad, K, Aune, D, Weiderpass, E, Romieu, I, Riboli, E, Gunter, MJ, Dossus, L, Prentice, R & Ferrari, P 2018, 'Risk prediction for estrogen receptor-specific breast cancers in two large prospective cohorts', Breast Cancer Research, vol. 20, no. 1, pp. 147. https://doi.org/10.1186/s13058-018-1073-0
Li, Kuanrong ; Anderson, Garnet ; Viallon, Vivian ; Arveux, Patrick ; Kvaskoff, Marina ; Fournier, Agnès ; Krogh, Vittorio ; Tumino, Rosario ; Sánchez, Maria-Jose ; Ardanaz, Eva ; Chirlaque, María-Dolores ; Agudo, Antonio ; Muller, David C ; Smith, Todd ; Tzoulaki, Ioanna ; Key, Timothy J ; Bueno-de-Mesquita, Bas ; Trichopoulou, Antonia ; Bamia, Christina ; Orfanos, Philippos ; Kaaks, Rudolf ; Hüsing, Anika ; Fortner, Renée T ; Zeleniuch-Jacquotte, Anne ; Sund, Malin ; Dahm, Christina C ; Overvad, Kim ; Aune, Dagfinn ; Weiderpass, Elisabete ; Romieu, Isabelle ; Riboli, Elio ; Gunter, Marc J ; Dossus, Laure ; Prentice, Ross ; Ferrari, Pietro. / Risk prediction for estrogen receptor-specific breast cancers in two large prospective cohorts. In: Breast Cancer Research. 2018 ; Vol. 20, No. 1. pp. 147.
@article{12b284b6918c4f61bb73833adf95618a,
title = "Risk prediction for estrogen receptor-specific breast cancers in two large prospective cohorts",
abstract = "BACKGROUND: Few published breast cancer (BC) risk prediction models consider the heterogeneity of predictor variables between estrogen-receptor positive (ER+) and negative (ER-) tumors. Using data from two large cohorts, we examined whether modeling this heterogeneity could improve prediction.METHODS: We built two models, for ER+ (ModelER+) and ER- tumors (ModelER-), respectively, in 281,330 women (51{\%} postmenopausal at recruitment) from the European Prospective Investigation into Cancer and Nutrition cohort. Discrimination (C-statistic) and calibration (the agreement between predicted and observed tumor risks) were assessed both internally and externally in 82,319 postmenopausal women from the Women's Health Initiative study. We performed decision curve analysis to compare ModelER+ and the Gail model (ModelGail) regarding their applicability in risk assessment for chemoprevention.RESULTS: Parity, number of full-term pregnancies, age at first full-term pregnancy and body height were only associated with ER+ tumors. Menopausal status, age at menarche and at menopause, hormone replacement therapy, postmenopausal body mass index, and alcohol intake were homogeneously associated with ER+ and ER- tumors. Internal validation yielded a C-statistic of 0.64 for ModelER+ and 0.59 for ModelER-. External validation reduced the C-statistic of ModelER+ (0.59) and ModelGail (0.57). In external evaluation of calibration, ModelER+ outperformed the ModelGail: the former led to a 9{\%} overestimation of the risk of ER+ tumors, while the latter yielded a 22{\%} underestimation of the overall BC risk. Compared with the treat-all strategy, ModelER+ produced equal or higher net benefits irrespective of the benefit-to-harm ratio of chemoprevention, while ModelGail did not produce higher net benefits unless the benefit-to-harm ratio was below 50. The clinical applicability, i.e. the area defined by the net benefit curve and the treat-all and treat-none strategies, was 12.7 × 10- 6 for ModelER+ and 3.0 × 10- 6 for ModelGail.CONCLUSIONS: Modeling heterogeneous epidemiological risk factors might yield little improvement in BC risk prediction. Nevertheless, a model specifically predictive of ER+ tumor risk could be more applicable than an omnibus model in risk assessment for chemoprevention.",
author = "Kuanrong Li and Garnet Anderson and Vivian Viallon and Patrick Arveux and Marina Kvaskoff and Agn{\`e}s Fournier and Vittorio Krogh and Rosario Tumino and Maria-Jose S{\'a}nchez and Eva Ardanaz and Mar{\'i}a-Dolores Chirlaque and Antonio Agudo and Muller, {David C} and Todd Smith and Ioanna Tzoulaki and Key, {Timothy J} and Bas Bueno-de-Mesquita and Antonia Trichopoulou and Christina Bamia and Philippos Orfanos and Rudolf Kaaks and Anika H{\"u}sing and Fortner, {Ren{\'e}e T} and Anne Zeleniuch-Jacquotte and Malin Sund and Dahm, {Christina C} and Kim Overvad and Dagfinn Aune and Elisabete Weiderpass and Isabelle Romieu and Elio Riboli and Gunter, {Marc J} and Laure Dossus and Ross Prentice and Pietro Ferrari",
year = "2018",
month = "12",
day = "3",
doi = "10.1186/s13058-018-1073-0",
language = "English",
volume = "20",
pages = "147",
journal = "Breast Cancer Research",
issn = "1465-5411",
publisher = "BioMed Central",
number = "1",

}

TY - JOUR

T1 - Risk prediction for estrogen receptor-specific breast cancers in two large prospective cohorts

AU - Li, Kuanrong

AU - Anderson, Garnet

AU - Viallon, Vivian

AU - Arveux, Patrick

AU - Kvaskoff, Marina

AU - Fournier, Agnès

AU - Krogh, Vittorio

AU - Tumino, Rosario

AU - Sánchez, Maria-Jose

AU - Ardanaz, Eva

AU - Chirlaque, María-Dolores

AU - Agudo, Antonio

AU - Muller, David C

AU - Smith, Todd

AU - Tzoulaki, Ioanna

AU - Key, Timothy J

AU - Bueno-de-Mesquita, Bas

AU - Trichopoulou, Antonia

AU - Bamia, Christina

AU - Orfanos, Philippos

AU - Kaaks, Rudolf

AU - Hüsing, Anika

AU - Fortner, Renée T

AU - Zeleniuch-Jacquotte, Anne

AU - Sund, Malin

AU - Dahm, Christina C

AU - Overvad, Kim

AU - Aune, Dagfinn

AU - Weiderpass, Elisabete

AU - Romieu, Isabelle

AU - Riboli, Elio

AU - Gunter, Marc J

AU - Dossus, Laure

AU - Prentice, Ross

AU - Ferrari, Pietro

PY - 2018/12/3

Y1 - 2018/12/3

N2 - BACKGROUND: Few published breast cancer (BC) risk prediction models consider the heterogeneity of predictor variables between estrogen-receptor positive (ER+) and negative (ER-) tumors. Using data from two large cohorts, we examined whether modeling this heterogeneity could improve prediction.METHODS: We built two models, for ER+ (ModelER+) and ER- tumors (ModelER-), respectively, in 281,330 women (51% postmenopausal at recruitment) from the European Prospective Investigation into Cancer and Nutrition cohort. Discrimination (C-statistic) and calibration (the agreement between predicted and observed tumor risks) were assessed both internally and externally in 82,319 postmenopausal women from the Women's Health Initiative study. We performed decision curve analysis to compare ModelER+ and the Gail model (ModelGail) regarding their applicability in risk assessment for chemoprevention.RESULTS: Parity, number of full-term pregnancies, age at first full-term pregnancy and body height were only associated with ER+ tumors. Menopausal status, age at menarche and at menopause, hormone replacement therapy, postmenopausal body mass index, and alcohol intake were homogeneously associated with ER+ and ER- tumors. Internal validation yielded a C-statistic of 0.64 for ModelER+ and 0.59 for ModelER-. External validation reduced the C-statistic of ModelER+ (0.59) and ModelGail (0.57). In external evaluation of calibration, ModelER+ outperformed the ModelGail: the former led to a 9% overestimation of the risk of ER+ tumors, while the latter yielded a 22% underestimation of the overall BC risk. Compared with the treat-all strategy, ModelER+ produced equal or higher net benefits irrespective of the benefit-to-harm ratio of chemoprevention, while ModelGail did not produce higher net benefits unless the benefit-to-harm ratio was below 50. The clinical applicability, i.e. the area defined by the net benefit curve and the treat-all and treat-none strategies, was 12.7 × 10- 6 for ModelER+ and 3.0 × 10- 6 for ModelGail.CONCLUSIONS: Modeling heterogeneous epidemiological risk factors might yield little improvement in BC risk prediction. Nevertheless, a model specifically predictive of ER+ tumor risk could be more applicable than an omnibus model in risk assessment for chemoprevention.

AB - BACKGROUND: Few published breast cancer (BC) risk prediction models consider the heterogeneity of predictor variables between estrogen-receptor positive (ER+) and negative (ER-) tumors. Using data from two large cohorts, we examined whether modeling this heterogeneity could improve prediction.METHODS: We built two models, for ER+ (ModelER+) and ER- tumors (ModelER-), respectively, in 281,330 women (51% postmenopausal at recruitment) from the European Prospective Investigation into Cancer and Nutrition cohort. Discrimination (C-statistic) and calibration (the agreement between predicted and observed tumor risks) were assessed both internally and externally in 82,319 postmenopausal women from the Women's Health Initiative study. We performed decision curve analysis to compare ModelER+ and the Gail model (ModelGail) regarding their applicability in risk assessment for chemoprevention.RESULTS: Parity, number of full-term pregnancies, age at first full-term pregnancy and body height were only associated with ER+ tumors. Menopausal status, age at menarche and at menopause, hormone replacement therapy, postmenopausal body mass index, and alcohol intake were homogeneously associated with ER+ and ER- tumors. Internal validation yielded a C-statistic of 0.64 for ModelER+ and 0.59 for ModelER-. External validation reduced the C-statistic of ModelER+ (0.59) and ModelGail (0.57). In external evaluation of calibration, ModelER+ outperformed the ModelGail: the former led to a 9% overestimation of the risk of ER+ tumors, while the latter yielded a 22% underestimation of the overall BC risk. Compared with the treat-all strategy, ModelER+ produced equal or higher net benefits irrespective of the benefit-to-harm ratio of chemoprevention, while ModelGail did not produce higher net benefits unless the benefit-to-harm ratio was below 50. The clinical applicability, i.e. the area defined by the net benefit curve and the treat-all and treat-none strategies, was 12.7 × 10- 6 for ModelER+ and 3.0 × 10- 6 for ModelGail.CONCLUSIONS: Modeling heterogeneous epidemiological risk factors might yield little improvement in BC risk prediction. Nevertheless, a model specifically predictive of ER+ tumor risk could be more applicable than an omnibus model in risk assessment for chemoprevention.

U2 - 10.1186/s13058-018-1073-0

DO - 10.1186/s13058-018-1073-0

M3 - Article

VL - 20

SP - 147

JO - Breast Cancer Research

JF - Breast Cancer Research

SN - 1465-5411

IS - 1

ER -