Inyección solvent-vent

4.3.1.1 Types of study designs

Randomised and non-randomised observational studies (prospective and retrospective cohort and case-control studies) were included for both the prognostic and predictive reviews. Studies based on exploratory biomarkers whereby density was one of several biomarkers were included.

4.3.1.2 Types of participants

Studies were included if they had subsets that met the following participant criteria, but only the relevant subset data was to be extracted for the meta-analysis.

For both the prognostic and predictive biomarker reviews, all adult women aged 18 years or older, with or without breast cancer (denoted respectively as treatment and prevention) were to be included based on the following criteria.

 Treatment: women with early stage hormone receptor (oestrogen (ER) or progesterone (PgR))-positive breast cancer. This was defined to be women who had been diagnosed with histologically proven operable invasive hormone receptor-positive breast cancer or DCIS, and who were candidates to receive adjuvant endocrine therapy. There was to be no clinical evidence of metastatic disease to minimise the risk of a recurrence or contralateral breast cancer being a misclassified metastasis. Women were to be considered ineligible if their breast density measurements were not made on the contralateral breast because there was a risk that tumours may have been misclassified as dense tissue. For this same reason, women were to be considered ineligible if they had bilateral breast cancer.

 Prevention: women who had not previously been diagnosed with invasive breast cancer or DCIS. Women of all levels of increased risk due to genetic factors (including BRCA1/2 gene mutations or a family history of the disease, or both) or otherwise assessed by an absolute or relative risk prediction model were to be included. If women had breast implants or if they had undergone risk-reducing mastectomies, they were to be excluded. This was considered because these factors affect the ability to produce accurate density estimates.

117 Women were to be at-risk for at least the length of time between baseline and follow-up mammogram. Women could be included if they changed treatment or discontinued treatment throughout their follow-up, but they were to be excluded if they changed treatment between their mammograms as this may have affected the change in density. However, women could be included if they discontinued treatment between mammograms. Women were to be excluded if they received another SERM or AI before treatment because these effects may have continued into the second period of treatment.

For AI comparisons, women had to be postmenopausal at the start of treatment; for SERM comparisons, they were allowed to be pre- or postmenopausal. The definition of postmenopausal women included women who had undergone a bilateral oophorectomy, or women who were aged more than 60 years, or women who were aged 40 to 59 years with an intact uterus and who were amenorrhoeic for at least 12 months. Women were to be excluded if they were rendered temporarily postmenopausal through medical interventions (e.g. gonadotropin-releasing hormone (GnRH) analogues).

4.3.1.3 Types of interventions

4.3.1.3.1 Interventions

Studies were included if they had subsets that met the following intervention criteria, but only the relevant subset data was to be extracted for the meta-analysis. Studies including women receiving doses lower or higher than those outlined below were included, but for the meta- analysis, these women were to be included in a secondary dose-response analysis only. Studies involving a mixture of women receiving SERMs and AIs were included, but for the meta- analysis, these studies were to be included in the main analysis if the results could be separated by treatment; otherwise they were to be included in a secondary analysis only.

For both the prognostic and predictive biomarker reviews, women were to be included if they received SERMs at the following minimum doses (357): Tamoxifen, 20 mg daily; Raloxifene, 60mg daily; Lasofoxifene, 0.25mg daily; Arzoxifene, 20mg daily; Droloxifene, 40 mg daily; Bazedoxifene, 20 mg daily; and Fulvestrant, 250 mg monthly. Women were to be included if they received AIs at the following minimum doses: Anastrozole, 1 mg daily; Letrozole, 2.5 mg daily; and Exemestane, 25 mg daily. All treatments were to be orally-consumed, except Fulvestrant (intramuscular). Treatment was to be received for at least the length of time between baseline and follow-up mammogram (i.e. intended for at least 1 year).

118

4.3.1.3.2 Co-interventions

Studies were included if they had subsets that met the following co-intervention criteria, but only the relevant subset data was to be extracted for the meta-analysis. The same types of co- interventions were allowed for both the prognostic and predictive biomarker reviews.

For treatment, women were to be considered ineligible if they had not completed primary loco- regional treatment (surgery or radiotherapy, or both) and systemic treatment (chemotherapy or targeted therapy, either neoadjuvant or adjuvant) with curative intent. Women were to be considered ineligible if there was a gap of more than eight weeks between different treatment interventions, for example, between surgery and the start of radiotherapy, or if endocrine treatment was started more than 28 days before surgery.

If women used HRT either during the study or up to 2 years before baseline, they could be included, but this was to be noted in the ‘Risk of bias’ assessment where relevant. Other co- interventions were permitted, including exercise and diet advice, but these were also to be noted in the ‘Risk of bias’ assessment where relevant.

4.3.1.3.3 Comparators

The main difference between the prognostic and predictive biomarker review was the comparator.

 Prognostic biomarker review: The comparison was within each intervention group (SERM or AI), whereby assessment was on the association between density change and outcome in women receiving the treatment.

 Predictive biomarker review: The comparison was within each study, whereby assessment was on the association between density change and outcome in the intervention group compared with a control group. The within-study comparator group was defined as a corresponding randomised placebo group, or a non-randomised control group of women not receiving endocrine therapy.

4.3.1.4 Biomarker

The same definition of biomarker was used for both the prognostic and predictive reviews. A measure of mammographic density was required at baseline (start of endocrine therapy or study entry in those from the control group) and follow-up mammogram.

119 Studies were included if they had subsets that met the following biomarker criteria, but only the relevant subset data was to be extracted for the meta-analysis.

For treatment, baseline mammograms could be taken before or after diagnosis, but they were to be no more than 2 years before the initial breast cancer diagnosis so that they represented the breast at the time of diagnosis as closely as possible. For treatment and prevention, baseline mammograms had to be taken before the start of treatment (or study entry) so that they reflected the breast phenotype before the effects of endocrine treatment. A follow-up mammogram had to be performed 90 days to 3 years after the start of endocrine treatment (or study entry), with the density closest to 1 year from the start of endocrine therapy (or study entry) selected if there was a choice.

Range and average timings were recorded for the following (if they were available): time between baseline mammogram and diagnosis, time between diagnosis and start of endocrine therapy (or study entry), and time between start of endocrine therapy (or study entry) and follow-up mammogram.

Density methods had to have been shown in more than one study (outside of the review studies) to have a relationship with breast cancer risk. Acceptable density methods included (but were not limited to) the following percentage methods:

 Visual assessment by expert in 5% bands (%).

 Visual assessment by expert in 20% bands (Boyd categories).  Visual assessment by expert as continuous percentage (%).

 Semi-automated thresholding such as using ‘Cumulus’ software (23) by expert (or trained) reader (%).

 Fully-automated percentage (based on area of density) (%).  Fully-automated volumetric percentage (e.g. Volpara, (44)) (%).

Acceptable absolute density methods included (but were not limited to) the following:

 Semi-automated thresholding such as using ‘Cumulus’ software (23) by expert (or trained) reader (cm2_).

 Fully-automated absolute density (based on area of density) (cm2_).  Fully-automated volumetric absolute density (e.g. Volpara, (44)) (cm3_).

120 Acceptable categorical density measures included (but were not limited to) the following:

 BI-RADS density (18).  Wolfe grade (358).  Tabar grade (15).

Information on the reliability of density measures was also used to qualitatively assess the 'Risk of bias’ due to measurement of the biomarker. Such information included:

 The correlation between repeated measures from repeat mammograms.

 Whether different readers of density were used and whether the same reader assessed mammograms from the same woman.

 Intra-class correlation coefficients and Bland-Altman limits of agreement (359) to assess intra- and inter-reader reliability.

 Whether the reader was blinded to case status.

 Whether the reader was blinded to treatment allocation.

 Whether randomisation was per mammogram (mammograms read independently) or per woman (mammograms for each woman read with the knowledge of her other mammograms).

 Whether the order of per woman mammograms was sequential or random and assessed one at a time or simultaneously.

If different definitions or measures of mammographic density were used between the time - points used to assess density change, these women or studies were excluded.

4.3.1.5 Types of outcome measures

The same outcome measures were used for both the prognostic and predictive reviews.

Primary outcomes

Potential benefits from treatment:

 Treatment: breast cancer mortality (time to death caused by breast cancer).  Prevention: incidence of invasive breast cancer and DCIS.

121 Potential harms from treatment:

 Treatment and prevention: rate of all serious adverse events. These included serious side effects noted for Tamoxifen (cataracts, pulmonary embolism or deep vein thrombosis and endometrial cancer) and Anastrozole (osteoporosis and bone fractures).

Secondary outcomes

Potential benefits from treatment:  Treatment: recurrence.

 Treatment: incidence of a secondary primary breast cancer (e.g. in the contralateral breast).  Treatment: any recurrence or any death (disease-free survival).

 Treatment: distant metastases.

 Treatment: death from all causes (all-cause mortality).  Treatment: recurrence of invasive cancer only.

 Treatment: recurrence of DCIS cancer only.  Prevention: incidence of invasive cancer only.  Prevention: incidence of DCIS cancer only.

Potential harms from treatment:

 Treatment and prevention: troublesome but not serious side effects observed for SERMs and AIs, including vasomotor symptoms and joint or muscle pain.

’Summary of findings’ table for assessing the quality of the evidence

A ’Summary of findings’ table was produced for each of the prognostic and predictive biomarker reviews, following the approach outlined by GRADE (360) and using GRADEpro GDT software (361).