Factors Associated with Low-Dose Tamoxifen Use among Women with Atypical Hyperplasia or Lobular or Ductal Carcinoma In Situ.

Introduction
Breast cancer is the most commonly diagnosed malignancy among women in the United States and is the second leading cause of cancer-related death among women worldwide.1 Approximately 13.1% of women will develop breast cancer at some point during their lifetime and the economic burden of breast cancer care and treatment is substantial.2 In 2020 alone, breast cancer had the highest associated treatment cost of any cancer in the U.S. at $29.8 billion and the national cancer-attributed medical care costs are projected to only increase in coming years.1
Women who have been diagnosed with lobular or ductal carcinoma in situ (LCIS or DCIS) are at a significantly increased risk of developing invasive breast cancer.3-5 Atypical hyperplasia (AH) increases the risk of DCIS or invasive breast cancer by an estimated four to five-fold.6,7 Chemoprevention with selective estrogen receptor modulators (SERMs) and aromatase inhibitors (AI) has been shown to reduce the risk of estrogen receptor-positive breast cancer.7,8 Data has shown that among high-risk women, chemoprevention reduces the risk of breast cancer by an estimated 50% over the lifetime.9 However, reported uptake remains low,7 with less than 5% of the estimated 10 million women eligible in the U.S. deciding to take it.10 Key reasons for this low uptake include lack of systematic breast cancer risk assessment in primary care, lack of knowledge about chemoprevention among healthcare providers and patients, and concerns regarding side effects.10
Tamoxifen, a SERM approved by the U.S. Food and Drug Administration (FDA) in 1999 as a chemopreventive drug for high-risk women, has had limited use for primary prevention due to its potential toxicity and side effects, such as increased risk of endometrial cancer and thromboembolic events.7,11-13 The standard dose of tamoxifen for breast cancer chemoprevention is 20 mg/day for 5 years. In 2019 the TAM-01 trial conducted by DeCensi et al., demonstrated that a lower dose of 5 mg/day for 3 years compared to placebo was effective in reducing breast cancer events by half among women with AH, LCIS, or DCIS.14 A 10-year follow up of the TAM-01 study continues to show that the benefits of low-dose tamoxifen persist even after 7 years.15 The American Society of Clinical Oncology (ASCO) Clinical Practice Guidelines were updated in 2019 to include low-dose tamoxifen as an alternative option to the standard 20 mg/day dose for pre- and post-menopausal women with intraepithelial neoplasia (AH, LCIS or DCIS) given the randomized trial data.16
It remains unclear whether practice patterns have changed in recent years following the availability of low-dose tamoxifen with the potential for less toxicity. Further, there are limited studies evaluating the impact of low-dose tamoxifen on chemoprevention uptake among high-risk patients, and existing studies lack large, diverse patient cohorts.17 In this study, we aimed to assess changes in chemoprevention uptake before and after availability of low-dose tamoxifen and identify sociodemographic and clinical factors associated with uptake of low-dose tamoxifen compared to full-dose SERMs or AIs among women with intraepithelial neoplasia at a large academic medical center.

Materials and Methods
We conducted a retrospective cohort study among women diagnosed with AH, LCIS, or DCIS at Columbia University Irving Medical Center (CUIMC) in New York, NY from 2016-2023. Exclusion criteria included: 1) age under 35 years, 2) invasive breast cancer diagnosis prior or concurrent with AH, LCIS, or DCIS diagnosis, 3) initiating chemoprevention prior to 2016, 4) estrogen receptor (ER)-negative and progesterone receptor (PR)-negative DCIS, 5) bilateral mastectomy, or 6) unknown LCIS or DCIS status. This study was approved by the Institutional Review Board at CUIMC and was conducted in accordance with ethical guidelines set forth by the Declaration of Helsinki.

Demographic and Clinical Characteristics
Demographic data and clinical characteristics were extracted from the NewYork-Presbyterian Hospital (NYPH) Tumor Registry and electronic health record (EHR). Demographic variables included age at diagnosis of AH, LCIS or DCIS (<50 or ≥50 years), and race/ethnicity (non-Hispanic White, non-Hispanic Black, Hispanic, Asian, Other). Clinical characteristics, such as year of diagnosis (2016-2019 and 2020-2023), family history of breast cancer (yes/no), body mass index (BMI, kg/m2), ever use of hormone replacement therapy (HRT, yes/no), history of hysterectomy (yes/no), and medical oncology referral (yes/no) were collected. Individuals missing information on BMI were categorized as unknown. Since menopausal status may be correlated with age, we did not include this factor within our analyses.
Information on diagnosis of AH, LCIS, and DCIS was collected from the EHR. NYPH Tumor Registry data identified all incident LCIS and DCIS cases, as well as ER/PR status of DCIS. ICD-10 codes were used to identify patients with AH (N60.89) and LCIS/DCIS (D05.01, D05.02, D05.00/ D05.11, D05.12, D05.10) diagnoses. Patients with multiple diagnoses over the course of the study period were classified based on their most advanced breast lesion as follows: DCIS > LCIS > AH.

Outcome
The primary outcome of interest is low-dose tamoxifen uptake (yes/no), which was defined as documentation of tamoxifen 5-10 mg/day on medication lists in the EHR. Tamoxifen, raloxifene (Evista), and aromatase inhibitors (AIs), such as anastrozole (Arimidex), letrozole (Femara), and exemestane (Aromasin) were included in this analysis. Information regarding 1-year discontinuation (yes/no), defined as discontinuation at or before 1-year following initiation, was collected through manual chart review for patient’s first initiated chemopreventive agent.17 Of the 834 individuals who initiated chemoprevention, 779 (93.4%) individuals had information regarding 1-year discontinuation available in the medical record. We used a complete-case analysis to ascertain discontinuation rates among those who initiated chemoprevention.

Statistical Analysis
We generated descriptive statistics for covariates of interest, stratified by low-dose tamoxifen use among individuals who initiated chemoprevention with a SERM or AI. We conducted univariate analysis using the Chi-square or Fisher’s exact tests for categorical variables and Wilcoxon Rank Sum tests for continuous variables to assess differences between covariates of interest and low-dose tamoxifen initiation (yes/ no) as well as year of diagnosis (2016-2019 and 2020-2023). To assess the association between demographic/clinical characteristics and low-dose tamoxifen uptake, we used a multivariable logistic regression model to estimate adjusted odds ratios (ORs) and 95% confidence intervals (CIs). Multivariable models were adjusted for diagnosis of breast intraepithelial neoplasia (AH/LCIS vs. DCIS), year of diagnosis (2016-2019 vs. 2020-2023), age at diagnosis (<50 vs. ≥50 years), BMI (continuous), and medical oncology referral (yes/no).
We calculated rates of medication discontinuation by medication type (low-dose tamoxifen, full-dose tamoxifen, raloxifene, AI). We also calculated the percentage of individuals for each medication type who switched medications at least once following discontinuation and conducted a Chi-square test to assess for differences by initial medication type of 1-year discontinuation and switching medications. All statistical analyses were conducted using SAS (v9.4; SAS Institute Inc.) and p-values less than 0.05 were considered statistically significant.

Data availability:
Data were generated by the authors and available on request.

Results
Of the 2,889 women diagnosed with AH, LCIS, or DCIS between January 2016 and December 2023 at CUIMC, we excluded patients under 35 years old (n=106), invasive breast cancer diagnoses prior to or concurrent with AH, LCIS or DCIS (n=210), those with chemoprevention use prior to the study period (n=118), subjects with estrogen receptor and progesterone receptor–negative DCIS (n=67), those with prior bilateral mastectomy (n=110), and subjects with unknown LCIS or DCIS status (n=18) (Figure 1).
Of the 2,260 patients in this study population, 834 (36.9%) patients initiated a SERM or AI within the study period. In the period prior to the release of low-dose tamoxifen guidelines from 2016-2019, 33.9% of patients initiated chemoprevention compared to 39.3% in 2020-2023, resulting in a 5.4% absolute increase in uptake (P=0.008) (Supplementary Table 1). Much of this increase in uptake is attributable to increasing use of low-dose tamoxifen: 3.3% during the 2016-2019 period compared to 8.6% during the 2020-2023 period (P<0.001).
Among patients who initiated chemoprevention, the average age was 59.9 years (SD, 11.8) and the study population was racially and ethnically diverse, with 33.6% non-Hispanic White, 12.7% non-Hispanic Black, 28.3% Hispanic, 7.0% Asian, and 18.5% Other (Table 1). During the study period, 66.4% of patients had DCIS, 13.4% LCIS, and 20.1% AH. Nearly 40% had a documented family history of breast cancer, over 60% were overweight or obese, and over two-thirds were seen by a medical oncologist.
Overall, as the first initiated chemopreventive medication during the study period, a majority of patients used an AI (53.1%), followed by full-dose tamoxifen (23.4%), low-dose tamoxifen (16.3%), and raloxifene (7.2%). However, the distribution of initial chemoprevention medication varied by year of diagnosis (2016-2019 vs. 2020-2023), breast histology (AH/LCIS vs. DCIS), and age at diagnosis (<50 vs. ≥50 years) (Figure 2). Among all high-risk women in the cohort, the proportion who first initiated low-dose tamoxifen increased from 2.9% to 8.6% among women diagnosed in 2016-2019 and 2020-2023, respectively. Among women who initiated chemoprevention, the proportion who first initiated low-dose tamoxifen was 26.8% and 11.0% among women with AH/LCIS and DCIS, respectively. The most common chemopreventive medication first initiated was full dose tamoxifen (54.4%) among women diagnosed before age 50 compared to AIs (65.7%) among women diagnosed at age 50 or older. Low-dose tamoxifen uptake was higher in the younger vs. older age groups (31.8% vs. 11.6%, respectively). Among all high-risk women in the cohort, there was a greater increase in low-dose tamoxifen uptake following 2019 among women diagnosed before age 50 (2016-2019: 4.7%, 2020-2023: 13.2%) compared to women diagnosed at age 50 or older (2016-2019: 2.1%, 2020-2023: 6.7%). There was also a greater increase in low-dose tamoxifen uptake following 2019 among women diagnosed with AH/LCIS (2016-2019: 2.3%, 2020-2023: 9.8%) compared to women diagnosed with DCIS (2016-2019: 3.5%, 2020-2023: 7.2%)
In univariate analysis, age, breast histology, year of diagnosis, BMI, and medical oncology referral were associated with low-dose tamoxifen uptake (Table 2). Race and ethnicity, BMI, family history, prior hysterectomy, and ever HRT use were not associated with low-dose tamoxifen uptake. After adjusting for confounders, women diagnosed with breast intraepithelial neoplasia in 2020-2023 compared to 2016-2019 were more likely to take low-dose tamoxifen (OR=2.83, 95% CI=1.81-4.41). Compared to women 50 years and older, women below age 50 years at diagnosis were over 3 times more likely to start low-dose tamoxifen (OR=3.02, 95% CI=1.99-4.58). Women with AH/LCIS were also more likely to start low-dose tamoxifen compared to women with DCIS (OR=2.90, 95% CI=1.95-4.31). Further, those with a medical oncology referral were more likely to start low-dose tamoxifen compared to women not seen by a medical oncologist (OR=1.61, 95% CI=1.02-2.54).
Patients who initiated raloxifene had the lowest 1-year discontinuation rate (3.70%) followed by low-dose tamoxifen (6.20%), full-dose tamoxifen (8.06%%), and an AI (14.39%) (P=0.0062) (Table 3). Among patients who discontinued their first initiated medication at or before 1-year, 56% subsequently switched to another chemopreventive agent.

Discussion
In this study of women diagnosed with AH, LCIS, or DCIS, chemoprevention uptake significantly increased after the release of low-dose tamoxifen guidelines by about 5%, mainly due to an increase in low-dose tamoxifen use. Among high-risk women who started chemoprevention, we found that women who were younger (<50 years vs. ≥50 years), diagnosed with a less advanced breast lesion (AH/LCIS vs. DCIS), and who had a medical oncology referral were more likely to initiate low-dose tamoxifen.
DeCensi and colleagues reported in 2019 their findings that in a randomized controlled trial of 500 women with AH, LCIS, and DCIS, low-dose tamoxifen (5mg/day) for 3 years compared to placebo reduced breast cancer incidence by 50%.14 These findings were reinforced in their 10-year follow-up study, providing increased support for adopting low-dose tamoxifen in the prevention setting.15 Though this study had a relatively small sample size compared to other chemoprevention trials and there was no direct comparison between low-dose tamoxifen and full-dose SERMs/AIs, these findings support the use of a low-dose option, which may gain more acceptance in the prevention setting. The term “baby-tam” is often used in reference to low-dose tamoxifen and lends itself to the perception of less toxicity with this option. Improved dose-optimization strategies in oncology drug use at large is central to lowering treatment toxicity and maximizing efficacy for patient.18 Our data supports the growing body of evidence in a clinical setting demonstrating increased uptake, adherence, and tolerability for low-dose tamoxifen among women at high-risk for breast cancer.17,19-23
Prior literature highlights the importance of informed decision-making for chemoprevention use. In our study, about 37% of the cohort initiated chemoprevention during the study time-period. We previously reported approximately 30% chemoprevention uptake in an earlier cohort of women with AH, LCIS or DCIS. We found that LCIS and DCIS status compared to AH, as well as medical oncology referral, and older age were significantly associated with chemoprevention uptake. An analysis of over 1,500 women with LCIS and atypia, conducted by Flanagan and colleagues in 2020, found approximately 25% chemoprevention use.24 They found that uptake was greatest among the highest risk women, including those with pathogenic variants in breast cancer predisposition genes, chest wall irradiation, and LCIS. High-risk features were associated with 27-33% chemoprevention use, while 18.6% of those with AH used chemoprevention.24 They reported low uptake among younger women, with 11% of women <50 years reporting any current/prior chemoprevention use, compared to 28% among women ≥50 years.24 A study by Bychkovsky et al. in 2022 reported younger age and family history of breast cancer as predictive of chemoprevention uptake.17
There are currently limited studies examining factors associated with low-dose tamoxifen uptake. In a 2022 study of 660 women with high-risk lesions of the breast, Bychkovsky and colleagues analyzed chemoprevention uptake in a personalized, academic clinic setting (B-PREP), before and after the introduction of low-dose tamoxifen.17 Unlike our study, they did not find a significant increase in chemoprevention uptake between the pre-2019 and post-2019 periods. This may be, in part, due to the shorter follow-up time in their study. Our findings suggest that the increase in chemoprevention uptake is attributable to a significant increase in low-dose tamoxifen uptake within the study time period.
Our findings suggest that incorporating low-dose tamoxifen into discussions surrounding chemoprevention use may be particularly beneficial for certain populations. We report that among those who initiated chemoprevention, 26.8% of patients with AH/LCIS and 11.0% of patients with DCIS started low-dose tamoxifen. These proportions are lower than those reported in other studies (AH/LCIS: 57-75%, DCIS: 22%).17,19 A study of 477 patients with DCIS, LCIS, and AH by Patel et al. in 2024 similarly found an increased likelihood of starting low-dose tamoxifen among patient with AH and LCIS compared to DCIS.25 Women diagnosed with DCIS may be less likely to initiate low-dose tamoxifen given their increased risk for invasive breast cancer compared to AH and LCIS.
Additionally, we found that younger women were more likely to initiate low-dose tamoxifen than older women (OR=3.02, 95% CI=1.99-4.58). Tamoxifen is the only chemoprevention option for high-risk premenopausal women, while postmenopausal women are eligible for both SERMs and AIs. Given increased evidence supporting the lower toxicity, tolerability, and adherence to low-dose tamoxifen, our findings support the option of introducing a low-dose option to younger women. Bychkovsky and colleagues report that 1-year discontinuation rates were lower among those who initiated low-dose tamoxifen (6.7%) compared to full dose tamoxifen (15.0%), AIs (20.0%), or raloxifene (20.4%), indicating better persistence.17 The discontinuation rates in our study may be an underestimate compared to the literature due to missing data. We also did not observe a significant different in the 1-year discontinuation rate between low-dose and full-dose tamoxifen.
In our study, we found that women who received a medical oncology referral were more likely to initiate low-dose tamoxifen compared to those without a referral. We reported similar findings with chemoprevention uptake overall.7 This may reflect the comfort level of medical oncologists in prescribing chemoprevention with SERMs, including low-dose tamoxifen, or AIs and following ASCO guidelines.
Overall, our real-world observational data informs applications in developing evidence-based decision support tools for women at high-risk for breast cancer to choose amongst chemoprevention options. Interventions such as the MiCHOICE trial, which is evaluating web-based decision support tools aimed at promoting informed decision-making regarding breast cancer chemoprevention among women with AH and LCIS, may be able to incorporate information from this study to better inform and facilitate discussions surrounding chemoprevention with low-dose tamoxifen.26 A 2024 study by Cornell and colleagues of 41 patients diagnosed with AH, LCIS, DCIS, or with elevated calculated risk of breast cancer aimed to analyze uptake, adherence, and tolerability of low-dose tamoxifen.19 They provided counseling with a breast specialist regarding their breast cancer risk and the role of risk-reducing medications. They reported increased uptake of low-dose tamoxifen among women who received counseling, with 74% endorsing taking the medication after initial consultation, and illustrated the value of pairing decision aids along with appropriate counseling in order to improve patient understanding and engagement. Further, this study provides insights into expanding use of low-dose tamoxifen to high-risk women without AH, LCIS, or DCIS diagnosis.
Our study has several strengths including its large, racially and ethnically diverse patient population of women diagnosed with breast intraepithelial neoplasia. Other strengths include detailed information from the medical record on dose and duration of chemoprevention, as well as other clinical variables. However, we recognize several limitations within this study as well. Due to the retrospective nature of this study and potential inaccurate or missing data within the electronic health record (EHR), we only have information on whether patients were prescribed chemoprevention. We do not have information on whether prescriptions were filled or medication adherence. This study was also conducted in a single-center urban academic setting, limiting its generalizability within other geographic areas and clinic settings. Further, for patients who did not have a follow-up or record of discontinuation in the EHR, we classified these individuals as having discontinued their first initiated medication at one year, leading to potential for misclassification error. We lack information regarding reasons for discontinuation or switching medications, including whether patients experienced side effects. Furthermore, we note a modest increase in chemoprevention uptake of 5% following the introduction of low-dose tamoxifen, and we cannot rule-out the role of confounding variables which could not be adjusted for in this analysis or challenges in study design due to the retrospective nature of the analysis.
In conclusion, our findings add to the existing body of evidence for the uptake of low-dose tamoxifen in a prevention setting, particularly among younger women and those with less advanced breast lesions. Further research is required to understand how to most effectively develop and leverage decision support tools to supplement patient-provider discussions to improve chemoprevention decision-making.

Supplementary Material
1