Next Article in Journal
Artificial Intelligence in Digital Pathology for Bladder Cancer: Hype or Hope? A Systematic Review
Next Article in Special Issue
Exploring the Potential Link between PFAS Exposure and Endometrial Cancer: A Review of Environmental and Sociodemographic Factors
Previous Article in Journal
A Systematic Review of Breast Implant-Associated Squamous Cell Carcinoma
Previous Article in Special Issue
A Clinical Risk Model for Personalized Screening and Prevention of Breast Cancer
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Cancers
Volume 15
Issue 18
10.3390/cancers15184517
cancers-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Personalized Breast Cancer Screening: A Risk Prediction Model Based on Women Attending BreastScreen Norway

by
Javier Louro
1,2,
Marta Román
1,2,
Nataliia Moshina
3,
Camilla F. Olstad
3,
Marthe Larsen
3,
Silje Sagstad
3,
Xavier Castells
1,2 and
Solveig Hofvind
3,4,*
1
Department of Epidemiology and Evaluation, Hospital del Mar Medical Research Institute, 08003 Barcelona, Spain
2
Network for Research on Chronicity, Primary Care, and Health Promotion (RICAPPS), 48902 Barakaldo, Spain
3
Section for Breast Cancer Screening, Cancer Registry of Norway, 0304 Oslo, Norway
4
Department of Health and Care Sciences, Faculty of Health Sciences, UiT, The Arctic University of Norway, 9037 Tromsø, Norway
*
Author to whom correspondence should be addressed.
Cancers 2023, 15(18), 4517; https://doi.org/10.3390/cancers15184517
Submission received: 9 June 2023 / Revised: 6 September 2023 / Accepted: 9 September 2023 / Published: 12 September 2023
(This article belongs to the Collection Women’s Cancers Risk: Breast Cancer, Ovarian Cancer and Endometrial Cancer)
Browse Figures
Versions Notes

Abstract

:

Simple Summary

The study aimed to develop and validate a prediction model that can be used to classify women for tailored breast cancer screening based on their individual risk. The model included data on age, mammographic density, family history of breast cancer, body mass index, age at menarche, alcohol consumption, exercise, pregnancy, hormone replacement therapy, and benign breast disease for 57,411 women screened in BreastScreen Norway 2007–2019. The 4-year predicted risk of breast cancer ranged between 0.2% and 7.3%, with 95% of the population having a risk of 0.6–2.3%. The differences in the predicted risk favor personalized screening for breast cancer.

Abstract

Background: We aimed to develop and validate a model predicting breast cancer risk for women targeted by breast cancer screening. Method: This retrospective cohort study included 57,411 women screened at least once in BreastScreen Norway during the period from 2007 to 2019. The prediction model included information about age, mammographic density, family history of breast cancer, body mass index, age at menarche, alcohol consumption, exercise, pregnancy, hormone replacement therapy, and benign breast disease. We calculated a 4-year absolute breast cancer risk estimates for women and in risk groups by quartiles. The Bootstrap resampling method was used for internal validation of the model (E/O ratio). The area under the curve (AUC) was estimated with a 95% confidence interval (CI). Results: The 4-year predicted risk of breast cancer ranged from 0.22–7.33%, while 95% of the population had a risk of 0.55–2.31%. The thresholds for the quartiles of the risk groups, with 25% of the population in each group, were 0.82%, 1.10%, and 1.47%. Overall, the model slightly overestimated the risk with an E/O ratio of 1.10 (95% CI: 1.09–1.11) and the AUC was 62.6% (95% CI: 60.5–65.0%). Conclusions: This 4-year risk prediction model showed differences in the risk of breast cancer, supporting personalized screening for breast cancer in women aged 50–69 years.

1. Introduction

In recent decades, the risks of mammographic screening, such as false positive screening results and overdiagnosis, have been debated [1,2,3,4]. Consequently, the current breast cancer screening strategy has been questioned at the same time as more personalized approaches have been promoted [5,6,7,8]. This might include moving from a one-size-fits-all strategy, a universal approach with the same screening interval and screening technique for all women within a defined age range [9], to recommendations determining interval, techniques, and age range according to the women’s individual risk of breast cancer. However, the implementation of personalized strategies poses the challenge of effective stratification based on the women’s risk of the disease.
Several risk prediction models [10,11,12] have been developed since the Breast Cancer Risk Assessment tool was presented as the basis for the first breast cancer prevention trials in the 1980s in the U.S. [13,14,15,16]. Despite a large number of models and substantial improvements in methodology, there is a need for risk assessment models since most of them were not specifically created for women targeted by screening and are not calibrated by the different incidences in the populations [17]. For this reason, at the 2019 European Conference on Risk-Stratified Prevention and Early Detection of Breast Cancer, experts on breast cancer screening stated the need to create and validate risk models based on data from large cohorts [18].
To meet this need, we aimed to develop and validate a breast cancer risk prediction model targeting women participating in population-based screening, using data from BreastScreen Norway.

2. Materials and Methods

This study has a legal basis in accordance with Articles 6 [19] (e) and 9 [20] (j) of the GDPR. The data was disclosed with legal bases in the Cancer Registry Regulations section 3-1 and the Health Register Act section 19 a to 19 h [21,22]. Data from women with negative screening results who have made a reservation against permanent storage of their personal data in the Norwegian Cancer Registry are not included in the dataset, cf. Cancer Registry Regulations section 1-9 [21] and the Personal Health Data Filing System Act, section 11 [23].
BreastScreen Norway started in 1996 and is part of the Norwegian public health care service [24]. The program targets about 670,000 women aged 50–69 who are invited to two view mammography, biennially. During the period from 2017 to 2021, the average attendance rate was 75%, the recall rate was 3.3%, and the rate of screen-detected cancer was 0.64%. The interval cancer rate has been stable at 0.18% since the startup of the program.
We conducted a retrospective cohort study including information about all women screened at four breast centers in BreastScreen Norway, at least once between 1 January 2007, and 31 December 2019. The women were followed for interval cancer two years after her last screening examination, until 1 December 2021. Data from four centers (Rogaland, Hordaland, Akershus, and Trøndelag) were chosen as these centers had collected automated mammographic density within the study period. Women included in the study were free from breast cancer before invitation, had at least one automated measurement of mammographic density, and completed the self-reported health indicator questionnaire used in the study [25]. Women participating in the To-Be 1 and 2 trials were excluded (n = 16,300) [26]. Further, 234 women diagnosed with breast cancer at the first screening examination were excluded as they had not been disease-free at risk.
Data from BreastScreen Norway was extracted from the Cancer Registry of Norway which administers the program. Questionnaire data were collected as a part of BreastScreen Norway and are described in detail elsewhere [25].

2.1. Study Variables

The self-reported questionnaire provided information on known risk factors for breast cancer [25], including first and second-degree family history of breast cancer [27,28], body mass index (BMI) [29], exercise and physical activity [30,31], alcohol and smoking [32,33], age at menarche [34], pregnancy [35], use of hormone therapy (HT), oral contraception, and hormone spiral [36,37,38]. First-degree family history was defined as a mother, sister, or daughter with a breast cancer diagnosis, and second-degree family history as a grandmother, aunt, or niece and no first-degree. BMI was calculated as each woman’s weight in kilograms divided by height in meters squared. Physical activity was specified as light walking, cycling, gardening, and clearing snow, while exercise was defined as regular, high-intensity activities for at least half an hour each time. Alcohol consumption was collected as the average monthly consumption of units of alcohol. Smoking was defined as having smoked at least once a week for 6 months or more. Age at menarche was specified as age at first menstruation. Pregnancy was described as the number of pregnancies lasting for more than 6 months. The use of HT, oral contraception, and hormone spiral were collected as dichotomous variables. Information on previous benign breast disease was obtained from the questionnaire and from the screening database.
Mammographic density was recorded using automated software (Volpara, version 1.5.1; Volpara Solutions, Wellington, New Zealand) [39,40]. Each examination was categorized based on Volpara density grade (VDG) 4th edition using the percent density of fibroglandular tissue: VDG1, less than or equal to 4.4%, VDG2 between 4.5% and 7.4%, VDG3 between 7.5% and 15.4%, and VDG4 equal or higher than 15.5%. VDG is designed to correlate with BI-RADS (Breast Imaging-reporting and Data System) 4th edition [41]. We included both screen-detected and interval cancers, so women were followed for at least two years after their last mammogram. Both invasive breast cancers and DCIS were included in the analysis.

2.2. Statistical Analysis

Age, BMI, age at menarche, mammographic density, first and second-degree family history of breast cancer, benign breast disease, exercise, physical activity, alcohol consumption, smoking, pregnancy, and use of HT, oral contraception, and hormone spiral were tested by backward selection for inclusion in the model. By using backward selection, we ensured including the minimal reasonable number of risk factors that maximized model performance. We tested interactions between the risk factors included in the model. Because none of the interaction terms were significant, they were not included in the model. The proportional hazards assumption was assessed by testing the independence between Schoenfeld residuals and time [42,43].
The means with standard deviations (SD) for quantitative variables (age, BMI, and age at menarche) and numbers with percentages for qualitative variables (mammographic density, family history of breast cancer, benign breast disease, alcohol habit, exercise, pregnancy, and use of HT) were compared for women with and without breast cancer. We used partly conditional Cox proportional hazards regression to estimate the adjusted hazard ratios (aHR) of breast cancer incidence for the aforementioned risk factors [43,44]. The Huber sandwich estimator was used to obtain robust standard 95% confidence intervals (CI) [44]. The follow-up time for each woman was defined as the time in years from the date of the first screening examination during the study period until breast cancer diagnosis or end of follow-up. Further, the same model was used to estimate aHR of breast cancer incidence for transformed from quantitative into qualitative age (50–54, 55–59, 60–64, and 65–70 years), BMI (<22; 22–25; 25–28; and >28 kg/m2), age at menarche (<12; 12; 13; and >13 years) and other qualitative variables (mammographic density, family history of breast cancer, benign breast disease, alcohol consumption, exercise, pregnancy, and use of HT) in order to present the estimates in a more common way for studies on breast cancer risk factors [10,12,34,45].
Following the method described by Zheng and Heagerty [42], we predicted the 4-year absolute breast cancer risk estimates using a general hazard function based on the 4-year time horizon, the length of follow-up, and each woman’s risk profile. To evaluate the performance of the model, we assessed its calibration and discrimination by internal validation. To assess calibration, we estimated the expected-to-observed (E/O) breast cancer rate ratios. The observed rate was analyzed using the Kaplan-Meier estimator to consider right censoring [46]. To assess discrimination, we estimated the area under the receiving operating characteristic curve (AUC) using the predicted 4-year risk and each woman’s outcome [47]. The AUC stands for the percentage of women with cancer who have a higher estimated risk than women without cancer. Both validation statistics and their 95% CIs were estimated using the bootstrap resampling method with 1000 samples [48].
To present the effect of the study variables in the risk estimation, we plotted the difference in the 4-year predicted risk estimated by the model with and without each variable for each woman in the population.
Statistical tests were two-sided and all p-values <0.05 were considered statistically significant. All analyses were performed using the statistical software R version 4.1.2 (Development Core Team, 2014).

3. Results

The study cohort included 73,945 women screened during the study period. After exclusions, the study cohort comprised 57,411 women who underwent 182,812 examinations during the study period. A total of 1230 breast cancers were diagnosed (Table 1). The mean follow-up time was 6.5 years.
Mean age at menarche was 13.16 years for women with breast cancer and 13.24 years for women without breast cancer (p = 0.037) (Table 1). No difference in mean BMI was observed for women with versus without breast cancer (25.89 versus 25.75 respectively, p = 0.154). Women diagnosed with breast cancer had a higher proportion of mammographically dense breasts at baseline, a higher proportion of second- and first-degree family history of breast cancer and benign breast disease, higher alcohol consumption, and a higher proportion of HT use compared with women without breast cancer. We found a higher proportion of women reporting high levels of exercise and the number of pregnancies among those without a breast cancer diagnosis.
In the adjusted model, a backward selection was performed. Smoking, physical activity, oral contraception use, and use of hormone spiral were not included in the final model due to a lack of statistical significance. All the variables in the final model were associated with the risk of breast cancer, aHR = 1.01, 95% CI: 1.00–1.03 for age, aHR = 1.06, 95% CI: 1.04–1.08 for BMI, and aHR = 0.95, 95% CI: 0.91–1.00 for age at menarche. Women with mammographic dense breasts (VDG4) had a higher risk of breast cancer (aHR= 1.71, 95% CI: 1.33–2.20) compared with those classified with VDG2. A higher risk was also observed for women with a first-degree family history of breast cancer (aHR = 1.34, 95%CI: 1.10–1.63) compared with those without, and for those with versus without a prior benign breast disease (aHR = 1.53, 95% CI: 1.31–1.78). Women who reported more than 4 h exercise per week had an aHR of 0.65 (95% CI: 0.51–0.83) using those who reported no exercise as reference. Finally, breast cancer risk was higher in women who reported the use of HT versus those who did not (aHR = 1.30, 95% CI: 1.13–1.48) (Table 2). The proportional hazards assumption was well-founded for all predictive variables (p = 0.76). After transformation into a qualitative variable, the effect of age at menarche was no longer significant (aHR = 1.23, 95%CI 0.99–1.54 for <11 years; aHR = 1.15 95%CI 0.96–1.37 for 12 years and aHR = 1.14, 95%CI 0.97–1.34 for 13 years, using 14 or more years as reference) (Supplemental Table S1).
The 4-year predicted risk of breast cancer ranged between 0.22% and 7.33% (Figure 1). The three thresholds for the quartiles of four risk groups, with 25% of the population in each group, were 0.82%, 1.10%, and 1.47%. For the highest-risk group, a right-skewed distribution indicating a wide range of the risk of breast cancer was shown. The ratio between the 3rd and 1st absolute risk quartiles was 1.78 (95% CI: 1.77–1.79). We found 95% of the women to have a risk between 0.55% and 2.31%, with a ratio between the highest-risk and the lowest absolute-risk women within this group of 4.21 (95% CI: 4.17–4.26). Overall, the model slightly overestimated the risk with an E/O ratio of 1.10; (95% CI: 1.09–1.11) and the AUC was 62.6% (95% CI: 60.5–65.0%).
We found the highest overall effect in the estimated risks of mammographic density, modifying the estimated 4-year risk of women between a 1.61% reduction and a 3.83% increase depending on her mammographic density (Figure 2). The purple colors represent higher values of each variable, indicating the risk difference when considering mammographic density increased for women with denser breasts and decreased for women with fatty breasts. This is not strictly true for each woman, because the effect of each variable in the risk prediction in absolute terms also depends on all other risk factors analyzed. The variable with the second highest overall effect in risk prediction was BMI, ranging from a decrease of 1.53% to an increase of up to 3.88% in the 4-year absolute risk prediction. Previous benign breast disease, use of HT, and a family history together with increasing age and alcohol consumption were also associated with an increase in the 4-year absolute risk prediction. On the other hand, women with higher values for exercise had a reduction in the 4-year absolute estimated risk by up to 1.56% and no activity an increase of up to 1.06%, corroborating the protective effect of this variable. Higher values of age at menarche and pregnancy also implied a reduction in the 4-year absolute estimated risk.

4. Discussion

We used data from a longitudinal cohort of 57,411 women to develop and validate a 4-year risk prediction model for breast cancer among women participating in BreastScreen Norway. The model included age, family history, mammographic density, BMI, previous benign breast disease, exercise, alcohol consumption, every use of hormone therapy (HT), age at menarche, and pregnancy. The wide variation in the absolute predicted risks, from 0.22% to 7.33%, contributes to substantiating the need to use such models to personalize breast cancer screening.
Similarly to other studies, we found a higher risk of breast cancer among women with mammographically dense versus non-dense breasts, also after adjusting for BMI [49], suggesting that more intensive screening strategies might be offered to women with dense breasts [50]. No difference in risk was found for mean baseline BMI between women with and without breast cancer. However, the longitudinal adjusted model showed that for each unit of difference in BMI, the risk of developing breast cancer increased by 6%, which is supported for postmenopausal women in other studies [51,52]. Our results supported studies reporting an increased risk of breast cancer for women with a prior benign breast disease [53,54] and a family history of breast cancer [27]. A protective effect of high-intensity exercise [30] and an increased risk of alcohol consumption has been shown in other studies [32]. Alcohol consumption was not statistically significantly associated with breast cancer risk in our study but was included in the model as this factor increased the discriminatory power of the model. Hormonal factors, such as age at menarche, pregnancy, and HT use have also been shown to affect breast cancer risk [34,35,36] and were included in the model due to the discriminatory power.
The E/O ratio showed that the model slightly overestimated the risk in the population, (E/O ratio = 1.1), thus, a calibration before implementation is required. The discrimination of the model was modest in our study, (AUC = 62.6%), and comparable to other studies in breast cancer risk prediction [17]. AUC could be slightly improved by adding some other risk factors that could increase the discriminatory power of the model, but it is known that a modest AUC does not detract from the usefulness of breast cancer risk prediction models [55]. However, it is argued that AUC statistics have limitations regarding the accuracy of risk prediction for low-incidence diseases [56].
When plotting the 4-year risk, we found a broad range of values in the risk estimates, reflecting the differences in risk among women targeted for breast cancer screening. These differences highlight the ability of risk prediction models to identify women at higher and lower risk compared with the average, and their need for implementing personalized screening strategies. Offering risk-based strategies to women with broad differences in the risk of breast cancer could improve the efficiency of screening [8], diminishing the proportion of false positives and detecting breast cancer earlier.
Two clinical trials are currently underway to demonstrate the increased efficiency of personalized screening, one in Europe [57] and one in the U.S. [58,59]. Both trials use robust models, which have been validated several times, however, risk models specifically designed for screening populations are needed to make better classification of risk groups and to avoid biases arising from using models developed in the clinical setting. Calibration of the model parameters should be performed prior to the implementation of risk prediction models to stratify populations. Insufficiently calibrated models may lead to biased estimates and inaccurate proportions of women classified as high or low risk. This is why, whenever possible, a model specific to the actual population is ideal, if the model is robust and validated. This is the first model specifically developed to stratify women attending the screening at BreastScreen Norway. Comparison with other well-established models would not be possible as these models would need to be calibrated to the incidence in Norway and to its population. Future studies aimed to compare calibration and the discriminatory power with those of other models would be desirable.
To our knowledge, most models on breast cancer risk prediction have not included lifestyle variables, such as exercise or alcohol habits, despite being already proven risk factors for breast cancer [30,32]. These variables are commonly difficult to obtain exhaustively but as shown in this study, they increase the discriminatory power of the models. The impact of each variable on the estimation was analyzed by means of a novel Effect plot. We plotted the risk contributed by each variable to each woman, allowing us to observe which variables had the greatest effect on the absolute risk estimates in our population. Some of the first studies on personalized breast cancer screening [5,6] suggested strategies that were not based on absolute risk, but rather on risk factors. For example, stratifying women according to their mammographic density, or after a benign breast disease. This is a more pragmatic way of designing personalized strategies. A graph representing which variables have a higher or lower effect on the risk estimation might help decide which variables are relevant for the model, and how to use the different categories for stratification.
Results from our model add knowledge about risk prediction for breast cancer among asymptomatic women participating in BreastScreen Norway and can be used to classify women into low, average, and high-risk groups which could benefit from being offered personalized screening intervals or modalities. Despite previous studies that have shown that risk-based screening strategies for breast cancer could be cost-effective, future research is needed to find the most cost-effective criteria for classifying women into different risk groups [5,6]. Previous studies suggested that low-risk women should be screened every 3 to 4 years and that high-risk women could benefit from annual screening [8]. According to our results, high mammographic density and BMI, previous benign breast disease, and family history of breast cancer were associated with the highest risk estimates. The combination of these factors in a woman would imply offering this woman a shorter screening interval and additional screening modalities, including ultrasound, contrast-enhanced mammography, and/or MRI if our personalized risk prediction model was implemented. The offer aimed to detect the cancer in an earlier stage, and thus favorable prognosis for the actual women. Alternatively, the combination of low mammographic density, normal BMI, the absence of benign breast disease, and family history of breast cancer in a woman might suggest offering this woman a 3- to 4-year screening interval using conventional screening modalities.
We used individual data from women participating in a population-based screening program for breast cancer and several risk factors, and variables related to previous screening examinations were available. This is a strength of the study. Mammographic density was automatically measured by software, reducing the risk of measurement bias. This study has several limitations. Most of the risk factors were self-reported, which could lead to information bias. However, self-reporting is a common source of information in risk prediction models [60,61]. Further, data on hormonal therapy were collected historically without information on current use, personal history of benign lesions did not include its histological type, and no information on genetic factors, which have been included in other risk prediction models, were available [62,63]. Information about menopausal status was not included in our model as we assumed that most of the women aged 50 years or older were postmenopausal in our study population. Other limitations were related to a lack of adjustment for competing risks, such as non-breast cancer death and lack of information about histopathological types of benign lesions. Although adjustment for competing risks would have been desirable, its effect on breast cancer incidence might be limited in this study as we solely included women aged 50 to 69 years. All follow-up losses were considered right censoring. Our study was performed with data from women participating in BreastScreen Norway and to verify the predictive power, the model should be tested on other study populations. Despite the fact that age at first birth is more predictive for breast cancer risk than the number of pregnancies, solely data on the latter were included in the study as the age at first birth was not available in the received dataset. Further, this study was not designed to be used to discriminate by ethnic groups, since the Norwegian population is very homogeneous and the data on ethnicity was not collected. We did not include height, as a breast cancer risk factor, in addition to BMI as our model was attributed to the population of screened women, and not to women with genetic mutations, for whom the significance of height was indicated [45,64,65].

5. Conclusions

The risk prediction model developed and validated showed differences in the risk of breast cancer, supporting personalized screening, which is expected to improve the risk-benefit balance of organized screening.

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/cancers15184517/s1, Table S1: Partly conditional Cox proportional hazards model results with adjusted hazard ratios for the risk factors for breast cancer after transforming age, body mass index (BMI) and age at menarche into qualitative variables.

Author Contributions

Conceptualization, J.L., M.R., N.M., C.F.O., M.L., S.S., X.C. and S.H.; Data curation, N.M., C.F.O., M.L., S.S. and S.H.; Formal analysis, J.L., N.M., C.F.O., M.L., S.S. and X.C.; Funding acquisition, J.L., M.R. and X.C.; Investigation, M.R.; Methodology, M.R., N.M., X.C. and S.H.; Project administration, X.C. and S.H.; Software, J.L. and M.R.; Supervision, S.H.; Writing—original draft, J.L.; Writing—review & editing, J.L., M.R., N.M., C.F.O., M.L., S.S., X.C. and S.H. All authors have read and agreed to the published version of the manuscript.

Funding

This study was supported by a grant from Instituto de Salud Carlos III FEDER (grant number: PI/00047).

Institutional Review Board Statement

This study has a legal basis in accordance with Articles 6 (e) and 9 (j) of the GDPR.

Informed Consent Statement

The data was disclosed with legal bases in the Cancer Registry Regulations section 3-1 and the Health Register Act section 19 a to 19 h. Data from women with negative screening result who have made a reservation against permanent storage of their personal data in the Norwegian Cancer Registry is not included in the dataset, cf. Cancer Registry Regulations section 1-9 and the Personal Health Data Filing System Act, section 11.

Data Availability Statement

Data used in the analyses can be made available on request to https://helsedata.no/ (accessed on 8 September 2023), given legal basis in Articles 6 and 9 of the GDPR and that the processing is in accordance with Article 5 of the GDPR.

Conflicts of Interest

The authors declare that they have no conflict of interest and confirm that the original manuscript is not under consideration for publication elsewhere. Solveig Hofvind is the head of BreastScreen Norway but has an independent fixed position at the Cancer Registry of Norway.

References

  1. Bleyer, A.; Welch, H.G. Effect of three decades of screening mammography on breast-cancer incidence. N. Engl. J. Med. 2012, 367, 1998–2005. [Google Scholar] [CrossRef] [PubMed]
  2. Welch, H.G.; Passow, H.J. Quantifying the benefits and harms of screening mammography. JAMA Intern. Med. 2014, 174, 448–454. [Google Scholar] [CrossRef] [PubMed]
  3. Paci, E.; Broeders, M.; Hofvind, S.; Puliti, D.; Duffy, S.W.; Group, E.W. European breast cancer service screening outcomes: A first balance sheet of the benefits and harms. Cancer Epidemiol. Biomark. Prev. 2014, 23, 1159–1163. [Google Scholar] [CrossRef]
  4. Canelo-Aybar, C.; Ferreira, D.S.; Ballesteros, M.; Posso, M.; Montero, N.; Solà, I.; Saz-Parkinson, Z.; Lerda, D.; Rossi, P.G.; Duffy, S.W.; et al. Benefits and harms of breast cancer mammography screening for women at average risk of breast cancer: A systematic review for the European Commission Initiative on Breast Cancer. J. Med. Screen. 2021, 28, 389–404. [Google Scholar] [CrossRef]
  5. Vilaprinyo, E.; Forné, C.; Carles, M.; Sala, M.; Pla, R.; Castells, X.; Domingo, L.; Rue, M.; the Interval Cancer (INCA) Study Group. Cost-effectiveness and harm-benefit analyses of risk-based screening strategies for breast cancer. PLoS ONE 2014, 9, e86858. [Google Scholar] [CrossRef] [PubMed]
  6. Schousboe, J.T.; Kerlikowske, K.; Loh, A.; Cummings, S.R. Personalizing mammography by breast density and other risk factors for breast cancer: Analysis of health benefits and cost-effectiveness. Ann. Intern. Med. 2011, 155, 10–20. [Google Scholar] [CrossRef] [PubMed]
  7. Trentham-Dietz, A.; Kerlikowske, K.; Stout, N.K.; Miglioretti, D.L.; Schechter, C.B.; Ergun, M.A.; Broek, J.J.v.D.; Alagoz, O.; Sprague, B.L.; van Ravesteyn, N.T.; et al. Tailoring Breast Cancer Screening Intervals by Breast Density and Risk for Women Aged 50 Years or Older: Collaborative Modeling of Screening Outcomes. Ann. Intern. Med. 2016, 165, 700–712. [Google Scholar] [CrossRef] [PubMed]
  8. Roman, M.; Sala, M.; Domingo, L.; Posso, M.; Louro, J.; Castells, X. Personalized breast cancer screening strategies: A systematic review and quality assessment. PLoS ONE 2019, 14, e0226352. [Google Scholar] [CrossRef]
  9. Schünemann, H.J.; Lerda, D.; Quinn, C.; Follmann, M.; Alonso-Coello, P.; Rossi, P.G.; Lebeau, A.; Nyström, L.; Broeders, M.; Ioannidou-Mouzaka, L.; et al. Breast Cancer Screening and Diagnosis: A Synopsis of the European Breast Guidelines. Ann. Intern. Med. 2020, 172, 46–56. [Google Scholar] [CrossRef]
  10. Antoniou, A.C.; Pharoah, P.P.; Smith, P.; Easton, D.F. The BOADICEA model of genetic susceptibility to breast and ovarian cancer. Br. J. Cancer 2004, 91, 1580–1590. [Google Scholar] [CrossRef]
  11. Tice, J.A.; Miglioretti, D.L.; Li, C.S.; Vachon, C.M.; Gard, C.C.; Kerlikowske, K. Breast Density and Benign Breast Disease: Risk Assessment to Identify Women at High Risk of Breast Cancer. J. Clin. Oncol. 2015, 33, 3137–3143. [Google Scholar] [CrossRef] [PubMed]
  12. Tyrer, J.; Duffy, S.W.; Cuzick, J. A breast cancer prediction model incorporating familial and personal risk factors. Stat. Med. 2004, 23, 1111–1130. [Google Scholar] [CrossRef] [PubMed]
  13. Gail, M.H.; Brinton, L.A.; Byar, D.P.; Corle, D.K.; Green, S.B.; Schairer, C.; Mulvihill, J.J. Projecting individualized probabilities of developing breast cancer for white females who are being examined annually. J. Natl. Cancer Inst. 1989, 81, 1879–1886. [Google Scholar] [CrossRef] [PubMed]
  14. Fisher, B.; Costantino, J.P.; Wickerham, D.L.; Redmond, C.K.; Kavanah, M.; Cronin, W.M.; Vogel, V.; Robidoux, A.; Dimitrov, N.; Atkins, J.; et al. Tamoxifen for prevention of breast cancer: Report of the National Surgical Adjuvant Breast and Bowel Project P-1 Study. J. Natl. Cancer Inst. 1998, 90, 1371–1388. [Google Scholar] [CrossRef] [PubMed]
  15. Vogel, V.G.; Costantino, J.P.; Wickerham, D.L.; Cronin, W.M.; Cecchini, R.S.; Atkins, J.N.; Bevers, T.B.; Fehrenbacher, L.; Pajon, E.R., Jr.; Wade, J.L., III; et al. Effects of tamoxifen vs raloxifene on the risk of developing invasive breast cancer and other disease outcomes: The NSABP Study of Tamoxifen and Raloxifene (STAR) P-2 trial. JAMA 2006, 295, 2727–2741. [Google Scholar] [CrossRef]
  16. Tice, J.A.; Cummings, S.R.; Smith-Bindman, R.; Ichikawa, L.; Barlow, W.E.; Kerlikowske, K. Using clinical factors and mammographic breast density to estimate breast cancer risk: Development and validation of a new predictive model. Ann. Intern. Med. 2008, 148, 337–347. [Google Scholar] [CrossRef]
  17. Louro, J.; Posso, M.; Boon, M.H.; Román, M.; Domingo, L.; Castells, X.; Sala, M. A systematic review and quality assessment of individualised breast cancer risk prediction models. Br. J. Cancer 2019, 121, 76–85. [Google Scholar] [CrossRef]
  18. Pashayan, N.; Antoniou, A.C.; Ivanus, U.; Esserman, L.J.; Easton, D.F.; French, D.; Sroczynski, G.; Hall, P.; Cuzick, J.; Evans, D.G.; et al. Personalized early detection and prevention of breast cancer: ENVISION consensus statement. Nat. Rev. Clin. Oncol. 2020, 17, 687–705. [Google Scholar] [CrossRef]
  19. Lov om behandling av Personopplysninger (Personopplysningsloven) EUROPAPARLAMENTS- OG RÅDSFORORDNING (EU) 2016/679 (Generell Personvernforordning) Kapittel II Artikkel 6 [Act on the Processing of Personal Data (Personal Data Act) EUROPEAN PARLIAMENT AND COUNCIL REGULATION (EU) 2016/679 (General Data Protection Regulation) Chapter II Article 6]. Available online: https://lovdata.no/dokument/NL/lov/2018-06-15-38/KAPITTEL_gdpr-2#KAPITTEL_gdpr-2 (accessed on 8 September 2023).
  20. Lov om Behandling av Personopplysninger (Personopplysningsloven) EUROPAPARLAMENTS- OG RÅDSFORORDNING (EU) 2016/679 (Generell Personvernforordning) Kapittel II Artikkel 9 [Act on the Processing of Personal Data (Personal Data Act) EUROPEAN PARLIAMENT AND COUNCIL REGULATION (EU) 2016/679 (General Data Protection Regulation) Chapter II Article 9]. Available online: https://lovdata.no/lov/2018-06-15-38 (accessed on 8 September 2023).
  21. Oslo: Helse- og Omsorgsdepartementet [Oslo: The Department of Health and Care Services]. Forskrift om Innsamling og Behandling av Helseopplysninger i Kreftregisteret (Kreftregisterforskriften). Regulations on the Collection and Treatment of Health Information in the Cancer Registry (Cancer Registry Regulations); The Department of Health and Care Services: Oslo, Norway, 2001. [Google Scholar]
  22. Lov om Helseregistre og Behandling av Helseopplysninger (Helseregisterloven). The Law on Health Registries and Use of Health Data (Health Registry Law); The Department of Health and Care Services: Oslo, Norway, 2015. [Google Scholar]
  23. Act of 18 May 2001 No. 24 on Personal Health Data Filing Systems and the Processing of Personal Health Data (Personal Health Data Filing System Act) Section 11. Available online: https://www.datatilsynet.no/globalassets/global/english/personal_health_data_filing_system_act_20100907.pdf (accessed on 8 September 2023).
  24. Bjørnson, E.W.; Holen, S.A.; Sagstad, S.; Larsen, M.; Thy, J.; Mangerud, G.; Ertzaas, A.K.; Hofvind, S. BreastScreen Norway: 25 Years of Organized Mammographic Screening. Early Performance Measures, Recent Developments and the Future; Cancer Registry of Norway: Oslo, Norway, 2022. [Google Scholar]
  25. Robsahm, T.E.; Sagstad, S.; Thy, J.E.; Hofvind, S. Sociodemographic factors, health indicators and lifestyle factors among participants in BreastScreen Norway 2006-2016—A cohort profile. Nor. Epidemiol. 2022, 30, 69–75. [Google Scholar] [CrossRef]
  26. Hofvind, S.; Moshina, N.; Holen, S.; Danielsen, A.S.; Lee, C.I.; Houssami, N.; Aase, H.S.; Akslen, L.A.; Haldorsen, I.S. Interval and Subsequent Round Breast Cancer in a Randomized Controlled Trial Comparing Digital Breast Tomosynthesis and Digital Mammography Screening. Radiology 2021, 300, 66–76. [Google Scholar] [CrossRef]
  27. Collaborative Group on Hormonal Factors in Breast Cancer. Familial breast cancer: Collaborative reanalysis of individual data from 52 epidemiological studies including 58,209 women with breast cancer and 101,986 women without the disease. Lancet 2001, 358, 1389–1399. [Google Scholar] [CrossRef] [PubMed]
  28. Brewer, H.R.; Jones, M.E.; Schoemaker, M.J.; Ashworth, A.; Swerdlow, A.J. Family history and risk of breast cancer: An analysis accounting for family structure. Breast Cancer Res. Treat. 2017, 165, 193–200. [Google Scholar] [CrossRef] [PubMed]
  29. Tran, T.X.M.; Kim, S.; Song, H.; Park, B. Mammographic breast density, body mass index and risk of breast cancer in Korean women aged 75 years and older. Int. J. Cancer 2022, 151, 869–877. [Google Scholar] [CrossRef]
  30. Thune, I.; Brenn, T.; Lund, E.; Gaard, M. Physical activity and the risk of breast cancer. N. Engl. J. Med. 1997, 336, 1269–1275. [Google Scholar] [CrossRef]
  31. Boraka, O.; Klintman, M.; Rosendahl, A.H. Physical Activity and Long-Term Risk of Breast Cancer, Associations with Time in Life and Body Composition in the Prospective Malmo Diet and Cancer Study. Cancers 2022, 14, 1960. [Google Scholar] [CrossRef]
  32. Donat-Vargas, C.; Guerrero-Zotano, Á.; Casas, A.; Baena-Cañada, J.M.; Lope, V.; Antolín, S.; Garcia-Saénz, J.; Bermejo, B.; Muñoz, M.; Ramos, M.; et al. Trajectories of alcohol consumption during life and the risk of developing breast cancer. Br. J. Cancer 2021, 125, 1168–1176. [Google Scholar] [CrossRef]
  33. Jones, M.E.; Schoemaker, M.J.; Wright, L.B.; Ashworth, A.; Swerdlow, A.J. Smoking and risk of breast cancer in the Generations Study cohort. Breast Cancer Res. 2017, 19, 118. [Google Scholar] [CrossRef]
  34. Collaborative Group on Hormonal Factors in Breast Cancer. Menarche, menopause, and breast cancer risk: Individual participant meta-analysis, including 118 964 women with breast cancer from 117 epidemiological studies. Lancet Oncol. 2012, 13, 1141–1151. [Google Scholar] [CrossRef]
  35. Russo, J.; Moral, R.; Balogh, G.A.; Mailo, D.; Russo, I.H. The protective role of pregnancy in breast cancer. Breast Cancer Res. 2005, 7, 131–142. [Google Scholar] [CrossRef]
  36. Licznerska, B.E.; Baer-Dubowska, W. Controversies over hormone replacement therapy. HRT as a risk factor of breast cancer in postmenopausal women. Acta Pol. Pharm. 2005, 62, 241–247. [Google Scholar]
  37. Bernstein, L. The risk of breast, endometrial and ovarian cancer in users of hormonal preparations. Basic Clin. Pharmacol. Toxicol. 2006, 98, 288–296. [Google Scholar] [CrossRef] [PubMed]
  38. Morch, L.S.; Skovlund, C.W.; Hannaford, P.C.; Iversen, L.; Fielding, S.; Lidegaard, O. Contemporary Hormonal Contraception and the Risk of Breast Cancer. N. Engl. J. Med. 2017, 377, 2228–2239. [Google Scholar] [CrossRef] [PubMed]
  39. Volpara Health: Breast Density Explained. Available online: https://www.volparahealth.com/for-patients/breast-density-explained/ (accessed on 8 September 2023).
  40. Highnam, R.; Brady, S.M.; Yaffe, M.J.; Karssemeijer, N.; Harvey, J. (Eds.) Robust Breast Composition Measurement–VolparaTM2010; Springer: Berlin/Heidelberg, Germany, 2010. [Google Scholar]
  41. van der Waal, D.; Heeten, G.J.D.; Pijnappel, R.M.; Schuur, K.H.; Timmers, J.M.H.; Verbeek, A.L.M.; Broeders, M.J.M. Comparing Visually Assessed BI-RADS Breast Density and Automated Volumetric Breast Density Software: A Cross-Sectional Study in a Breast Cancer Screening Setting. PLoS ONE 2015, 10, e0136667. [Google Scholar] [CrossRef]
  42. Zheng, Y.; Heagerty, P.J. Partly conditional survival models for longitudinal data. Biometrics 2005, 61, 379–391. [Google Scholar] [CrossRef] [PubMed]
  43. Schoenfeld, D. Partial residuals for the proportional hazards regression model. Biometrika 1982, 69, 239–241. [Google Scholar] [CrossRef]
  44. Freedman, D.A. On The So-Called “Huber Sandwich Estimator” and “Robust Standard Errors”. Am. Stat. 2006, 60, 299–302. [Google Scholar] [CrossRef]
  45. Qian, F.; Wang, S.; Mitchell, J.; McGuffog, L.; Barrowdale, D.; Leslie, G.; Oosterwijk, J.C.; Chung, W.K.; Evans, D.G.; Engel, C.; et al. Height and Body Mass Index as Modifiers of Breast Cancer Risk in BRCA1/2 Mutation Carriers: A Mendelian Randomization Study. J. Natl. Cancer Inst. 2019, 111, 350–364. [Google Scholar] [CrossRef]
  46. Kaplan, E.L.; Meier, P. Nonparametric Estimation from Incomplete Observations. J. Am. Stat. Assoc. 1958, 53, 457–481. [Google Scholar] [CrossRef]
  47. Harrell, F.E., Jr.; Lee, K.L.; Mark, D.B. Multivariable prognostic models: Issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors. Stat. Med. 1996, 15, 361–387. [Google Scholar] [CrossRef]
  48. Efron, B.; Tibshirani, R. Improvements on Cross-Validation: The 632+ Bootstrap Method. J. Am. Stat. Assoc. 1997, 92, 548–560. [Google Scholar]
  49. Boyd, N.F.; Guo, H.; Martin, L.J.; Sun, L.; Stone, J.; Fishell, E.; Jong, R.A.; Hislop, G.; Chiarelli, A.; Minkin, S.; et al. Mammographic density and the risk and detection of breast cancer. N. Engl. J. Med. 2007, 356, 227–236. [Google Scholar] [CrossRef]
  50. Kerlikowske, K.; Miglioretti, D.L.; Vachon, C.M. Discussions of Dense Breasts, Breast Cancer Risk, and Screening Choices in 2019. JAMA 2019, 322, 69–70. [Google Scholar] [CrossRef] [PubMed]
  51. Shieh, Y.; Scott, C.G.; Jensen, M.R.; Norman, A.D.; Bertrand, K.A.; Pankratz, V.S.; Brandt, K.R.; Visscher, D.W.; Shepherd, J.A.; Tamimi, R.M.; et al. Body mass index, mammographic density, and breast cancer risk by estrogen receptor subtype. Breast Cancer Res. 2019, 21, 48. [Google Scholar] [CrossRef] [PubMed]
  52. Tran, T.X.M.; Moon, S.G.; Kim, S.; Park, B. Association of the Interaction Between Mammographic Breast Density, Body Mass Index, and Menopausal Status With Breast Cancer Risk Among Korean Women. JAMA Netw. Open 2021, 4, e2139161. [Google Scholar] [CrossRef]
  53. Hartmann, L.C.; Sellers, T.A.; Frost, M.H.; Lingle, W.L.; Degnim, A.C.; Ghosh, K.; Vierkant, R.A.; Maloney, S.D.; Pankratz, V.S.; Hillman, D.W.; et al. Benign breast disease and the risk of breast cancer. N. Engl. J. Med. 2005, 353, 229–237. [Google Scholar] [CrossRef]
  54. Castells, X.; Domingo, L.; Corominas, J.M.; Torá-Rocamora, I.; Quintana, M.J.; Baré, M.; Vidal, C.; Natal, C.; Sánchez, M.; Saladié, F.; et al. Breast cancer risk after diagnosis by screening mammography of nonproliferative or proliferative benign breast disease: A study from a population-based screening program. Breast Cancer Res. Treat. 2015, 149, 237–244. [Google Scholar] [CrossRef]
  55. Gail, M.H.; Pfeiffer, R.M. Breast Cancer Risk Model Requirements for Counseling, Prevention, and Screening. J. Natl. Cancer Inst. 2018, 110, 994–1002. [Google Scholar] [CrossRef]
  56. Ozenne, B.; Subtil, F.; Maucort-Boulch, D. The precision--recall curve overcame the optimism of the receiver operating characteristic curve in rare diseases. J. Clin. Epidemiol. 2015, 68, 855–859. [Google Scholar] [CrossRef] [PubMed]
  57. Roux, A.; Cholerton, R.; Sicsic, J.; Moumjid, N.; French, D.P.; Giorgi Rossi, P.; Balleyguier, C.; Guindy, M.; Gilbert, F.J.; Burrion, J.B.; et al. Study protocol comparing the ethical, psychological and socio-economic impact of personalised breast cancer screening to that of standard screening in the “My Personal Breast Screening” (MyPeBS) randomised clinical trial. BMC Cancer 2022, 22, 507. [Google Scholar] [CrossRef]
  58. Esserman, L.; Eklund, M.; Veer, L.V.; Shieh, Y.; Tice, J.; Ziv, E.; Blanco, A.; Kaplan, C.; Hiatt, R.; Fiscalini, A.S.; et al. The WISDOM study: A new approach to screening can and should be tested. Breast Cancer Res. Treat. 2021, 189, 593–598. [Google Scholar] [CrossRef]
  59. Esserman, L.J.; Study, W.; Athena, I. The WISDOM Study: Breaking the deadlock in the breast cancer screening debate. NPJ Breast Cancer 2017, 3, 34. [Google Scholar] [CrossRef] [PubMed]
  60. Rosner, B.; Colditz, G.A. Nurses’ health study: Log-incidence mathematical model of breast cancer incidence. J. Natl. Cancer Inst. 1996, 88, 359–364. [Google Scholar] [CrossRef] [PubMed]
  61. Duval, S.; Van’t Hof, J.R.; Steffen, L.M.; Luepker, R.V. Estimation of Cardiovascular Risk from Self-Reported Knowledge of Risk Factors: Insights from the Minnesota Heart Survey. Clin. Epidemiol. 2020, 12, 41–49. [Google Scholar] [CrossRef] [PubMed]
  62. Zhang, X.; Rice, M.; Tworoger, S.S.; Rosner, B.A.; Eliassen, A.H.; Tamimi, R.M.; Joshi, A.D.; Lindstrom, S.; Qian, J.; Colditz, G.A.; et al. Addition of a polygenic risk score, mammographic density, and endogenous hormones to existing breast cancer risk prediction models: A nested case-control study. PLoS Med. 2018, 15, e1002644. [Google Scholar] [CrossRef]
  63. Shieh, Y.; Hu, D.; Ma, L.; Huntsman, S.; Gard, C.C.; Leung, J.W.T.; Tice, J.A.; Vachon, C.M.; Cummings, S.R.; Kerlikowske, K.; et al. Breast cancer risk prediction using a clinical risk model and polygenic risk score. Breast Cancer Res. Treat. 2016, 159, 513–525. [Google Scholar] [CrossRef]
  64. Alajmi, K.; Lophatananon, A.; Muir, K. 136P Development of breast cancer risk prediction models using the UK Biobank dataset. Ann. Oncol. 2022, 33, S184–S185. [Google Scholar] [CrossRef]
  65. Schmidt, M.; Hogervorst, F.; van Hien, R.; Cornelissen, S.; Broeks, A.; Adank, M.; Meijers, H.; Waisfisz, Q.; Hollestelle, A.; Schutte, M.; et al. Age- and Tumor Subtype–Specific Breast Cancer Risk Estimates for CHEK2*1100delC Carriers. J. Clin. Oncol. 2016, 34, 2750–2760. [Google Scholar] [CrossRef]
Cancers 15 04517 g001
Figure 1. Distribution of the 4-year absolute breast cancer risk estimates. Red lines represent the quartiles of the estimated risk (MIN, minimum, Q1 = first quartile, Q2 = second quartile (median), Q3 = third quartile, MAX = maximum).
Figure 1. Distribution of the 4-year absolute breast cancer risk estimates. Red lines represent the quartiles of the estimated risk (MIN, minimum, Q1 = first quartile, Q2 = second quartile (median), Q3 = third quartile, MAX = maximum).
Cancers 15 04517 g001
Cancers 15 04517 g002
Figure 2. Impact on subtracting each variable from the model on 4-year absolute risk predictions. Footnote: This figure shows the impact of each variable on the different risk estimates. For each variable (Y-axis) and each woman (dot) in the population, the difference in the 4-year risk on the X-axis was estimated with the final model, and the 4-year risks estimated with the same model but without the selected variable plotted. The color of the dot represents each woman’s value in that specific variable; yellow represents low values and purple high values. For example, in the second row (body mass index) the rightmost point represents women whose risk of breast cancer is 3.88% higher when their body mass index is included in the model versus when it is not included.
Figure 2. Impact on subtracting each variable from the model on 4-year absolute risk predictions. Footnote: This figure shows the impact of each variable on the different risk estimates. For each variable (Y-axis) and each woman (dot) in the population, the difference in the 4-year risk on the X-axis was estimated with the final model, and the 4-year risks estimated with the same model but without the selected variable plotted. The color of the dot represents each woman’s value in that specific variable; yellow represents low values and purple high values. For example, in the second row (body mass index) the rightmost point represents women whose risk of breast cancer is 3.88% higher when their body mass index is included in the model versus when it is not included.
Cancers 15 04517 g002
Table 1. Characteristics of the study population.
Table 1. Characteristics of the study population.
No Breast CancerBreast Cancerp-Value
n = 56,181n = 1230
Age at baseline, mean [sd]58.15 (5.71)56.46 (5.03)<0.001
Baseline BMI, mean [sd]25.75 (4.10)25.89 (3.93)0.154
Age at menarche, mean [sd]13.24 (1.42)13.16 (1.42)0.037
Baseline mammographic density (%)
VDG117,177 (31.02)251 (20.41)
VDG221,193 (38.54)458 (37.24)
VDG314,643 (26.79)408 (33.17)
VDG43168 (5.84)113 (9.19)<0.001
Family history of breast cancer (%)
No43,600 (79.17)880 (71.54)
Yes, 2nd degree7252 (13.25)192 (15.61)
Yes, 1st degree5329 (9.77)158 (12.85)<0.001
Benign breast disease (%)
No46,940 (83.55)940 (76.42)
Yes9241 (16.44)290 (23.58)<0.001
Alcohol consumption (%)
No8536 (15.52)182 (14.80)
Yes, 5 or less units/month14,429 (26.19)284 (23.09)
Yes, 6–10 units/month14,212 (25.84)305 (24.80)
Yes, more than 10 units/month19,004 (34.64)459 (37.32)0.052
Exercise (%)
Never16,259 (29.65)396 (32.20)
0–1 h/week13,227 (24.04)280 (22.76)
2–3 h/week19,812 (36.03)430 (34.96)
More than 4 h/week6883 (12.47)124 (10.08)0.025
Pregnancy (%)
Never4712 (8.62)129 (10.49)
1 or 229,400 (53.48)643 (52.28)
3 or more22,069 (40.10)458 (37.24)0.023
Ever use of HT (%)
No34,393 (62.40)666 (54.15)
Yes21,788 (39.79)564 (45.85)<0.001
Differences in qualitative variables were tested by the chi-squared test. Differences in quantitative variables were tested by the Mann-Whitney U test. BMI = Body mass index, HT = Hormone therapy, VDG = Volpara density grade, sd = standard deviation.
Table 2. Partly conditional Cox proportional hazards model results with adjusted hazard ratios for the risk factors for breast cancer.
Table 2. Partly conditional Cox proportional hazards model results with adjusted hazard ratios for the risk factors for breast cancer.
Women-YearsaHR (95% CI)
Age (years)375,0781.01 (1.00–1.03)
BMI (kg/cm2)375,0781.06 (1.04–1.08)
Age at menarche (years)375,0780.95 (0.91–1.00)
Baseline mammographic density
VDG1126,9510.59 (0.51–0.69)
VDG2137,639Ref.
VDG393,3051.37 (1.20–1.56)
VDG417,1821.71 (1.34–2.20)
Family history of breast cancer
No289,289Ref.
Yes, 2nd degree50,0821.17 (0.98–1.41)
Yes, 1st degree35,7071.34 (1.10–1.63)
Benign breast disease
No319,891 Ref.
Yes53,4301.53 (1.31–1.78)
Alcohol consumption
No59,1530.94 (0.76–1.16)
Yes, 5 or less units/month94,432Ref.
Yes, 6–10 units/month93,7891.06 (0.88–1.29)
Yes, more than 10 units/month127,7041.14 (0.96–1.36)
Exercise
Never104,381Ref.
0–1 h/week88,4290.80 (0.67–0.96)
2–3 h/week135,0020.83 (0.70–0.97)
+4 h/week47,2660.65 (0.51–0.83)
Pregnancy
Never31,2251.10 (0.88–1.38)
1 or 2188,424Ref.
3 or more155,4290.91 (0.79–1.04)
Ever use of HT
No226,166Ref.
Yes148,9121.30 (1.14–1.49)
aHR: Adjusted Hazard Ratio, 95% CI: 95% Confidence Interval, VDG = Volpara density grade. HT = Hormone Replacement Therapy, BMI = Body mass index.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Louro, J.; Román, M.; Moshina, N.; Olstad, C.F.; Larsen, M.; Sagstad, S.; Castells, X.; Hofvind, S. Personalized Breast Cancer Screening: A Risk Prediction Model Based on Women Attending BreastScreen Norway. Cancers 2023, 15, 4517. https://doi.org/10.3390/cancers15184517

AMA Style

Louro J, Román M, Moshina N, Olstad CF, Larsen M, Sagstad S, Castells X, Hofvind S. Personalized Breast Cancer Screening: A Risk Prediction Model Based on Women Attending BreastScreen Norway. Cancers. 2023; 15(18):4517. https://doi.org/10.3390/cancers15184517

Chicago/Turabian Style

Louro, Javier, Marta Román, Nataliia Moshina, Camilla F. Olstad, Marthe Larsen, Silje Sagstad, Xavier Castells, and Solveig Hofvind. 2023. "Personalized Breast Cancer Screening: A Risk Prediction Model Based on Women Attending BreastScreen Norway" Cancers 15, no. 18: 4517. https://doi.org/10.3390/cancers15184517

APA Style

Louro, J., Román, M., Moshina, N., Olstad, C. F., Larsen, M., Sagstad, S., Castells, X., & Hofvind, S. (2023). Personalized Breast Cancer Screening: A Risk Prediction Model Based on Women Attending BreastScreen Norway. Cancers, 15(18), 4517. https://doi.org/10.3390/cancers15184517

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop