Study design and participants

The Dutch Minister of Health, Welfare and Sport, who was advised by the Health Council of the Netherlands (2011/2019 WBO, The Hague, The Netherlands), approved the DENSE trial on November 11, 2011. All participants provided written informed consent.

The DENSE trial is embedded within the Dutch population-based digital mammography screening program (age 50–75) and consists of three biennial screening rounds. The study design and outcomes of the first and second rounds have been described previously [6, 7, 10]. Women were eligible for the DENSE trial if they had extremely dense breasts (grade 4 or d) as measured with Volpara version 1.5 (Volpara Health Technologies) and if they had a negative mammography result (Breast Imaging Reporting and Data System [BI-RADS] category 1 or 2).

All MRI examinations were performed on 3.0-T MRI systems with the macrocyclic gadolinium-based contrast agent gadobutrol (0.1 mmol/kg body weight) (Gadovist; Bayer AG). Details on the full screening MRI protocol have been described previously [6].

Workflow screening rounds DENSE trial

Between December 2011 and January 2016, 4783 women were randomised to the intervention arm and participated in the first screening round of the DENSE trial. Women with a breast cancer diagnosis, an age outside the age range of the screening program (> 75 years) or women who passed away or moved abroad during or after the previous screening round, were not invited for subsequent screening rounds. All other women were invited again for mammographic screening 2 years after the previous (first or second) MRI round; women who participated in this mammographic screening and again had a negative (‘normal’) mammography result were invited for the next MRI round. Women who were referred for diagnostic work-up as a result of the mammographic screening were excluded from the corresponding MRI round (not invited).

Calculation of the re-attendance rate

We assessed the re-attendance rate as follows: the numerator consisted of all women who participated in the second (or third) MRI round of the DENSE trial. The denominator consisted of all women who were invited for the second (or third) MRI round.

As a sensitivity analysis, we used a different denominator consisting of all women who were invited for the second (or third) MRI round but also the women who had actively unsubscribed for further participation in the DENSE trial before they were invited, and women who had declined their invitation for mammography. We performed this sensitivity analyses because a woman’s decision to decline the mammography invitation could have been influenced by their previous MRI experience. To further elaborate this hypothesis, we analysed the previous MRI experiences in attendance subgroups (participants, non-participants of MRI, non-participants of mammography and MRI). Information on reasons for declining mammography screening invitations was not available.

Participant characteristics related to re-attendance

We described participant characteristics of women who re-attended and those who did not.

Participants completed a self-report questionnaire about demographic, reproductive and lifestyle factors, and their (family) medical history. We collected information about postal codes from the data available from the Dutch mammography program to classify socioeconomic status (SES).

Factors related to MRI screening experience

MRI-related (serious) adverse events ((S)AEs) were reported directly after the MRI and were self-reported by women within 30 days after the MRI examination when applicable. An MRI screen-specific items questionnaire was administered 2 days after each MRI examination to assess pain, discomfort, and anxiety experienced during the MRI examination [11]. A false-positive finding was defined as a positive MRI result (BI-RADS 3, 4, or 5) without a diagnosis of breast cancer.

Collection of self-reported reasons for discontinuation

Women were able to discontinue participation in the DENSE trial at any time. In this case, they were asked to provide one or more reasons for discontinuation. Subsequently, all reasons were registered, and we classified them into the following categories: MRI-related inconveniences and/or self-reported contraindications, anxiety regarding the outcome, personal reasons, practical reasons, burden too high, or reasons related to surveillance (e.g. already under active surveillance) [8]. We classified concerns around gadolinium retention in the brain as a reason for discontinuation under MRI-related inconveniences and/or self-reported contraindications [12, 13]. From March 2020 onwards, women were also able to provide ‘not wanting to go to the hospital due to the COVID-19 pandemic’ as a reason, which we categorised as a practical reason. When women declined the invitation due to a later acquired contraindication, we classified this as an MRI-related inconvenience and/or self-reported contraindication. Reasons were classified by two authors, and in case of disagreement, consensus was reached upon assessment of a third author.

Data analyses

The outcome for the analyses was screening re-attendance as defined previously. We reported characteristics and previous MRI experience as proportions (percentage) of women in the category that re-attended. We reported the means with standard deviation for normally distributed continuous variables.

We examined differences between participants and non-participants using Pearson’s chi-squared tests for categorical variables and t-tests for normally distributed continuous variables. We calculated p-values for trend. Additionally, we determined whether factors were associated with re-attendance, by fitting logistic regression models adjusted for age to calculate odds ratios (ORs) with corresponding 95% confidence intervals (CIs). An OR above 1 indicates that women are more likely to drop-out, thus less likely to re-attend, in the next MRI screening round.

We summarised reasons to discontinue screening using descriptive statistics for all women who actively unsubscribed for further participation in the trial after the first or second round or who declined the invitation for the subsequent MRI screening round.

Calculations were based on the data collected until 2020–10-06.

We performed all analyses using RStudio software (RStudio, version 1.3.1093).