Over and undertreatment of early-stage breast cancer in older women: evaluating the POWER trial

Abstract

Background

Radiation therapy (RT) omission is acceptable in older women with early-stage estrogen receptor (ER)+ breast cancer treated with breast-conserving surgery (BCS) and adjuvant endocrine therapy (AET). However, RT rates in this population remain high, causing concern for overtreatment. Conversely, patients who omit RT and do not complete a course of AET are at risk of undertreatment. In the POWER trial, participants receive 90 days of pre-operative endocrine therapy to assess tolerance before deciding about RT. This study aimed to determine the rates of under and overtreatment institutionally and among POWER trial participants.

Materials and Methods

Data were retrospectively collected from medical records of women aged ≥65 diagnosed with invasive, ER+/HER2− breast cancer, ≤3cm, who had BCS between 2012 and 2022. Patients were categorized as undertreated (BCS alone), over-treated (BCS+RT+AET), or appropriately treated (BCS+RT or BCS+AET).

Results

The cohort included 478 patients, of whom 62 (12.97%) were undertreated, 202 (42.26%) were overtreated, and 214 (44.77%) were appropriately treated. Appropriately treated patients were more likely to be 70–79 years old (p<0.0001) and have high health literacy (p=0.0003). Of the thirty-seven patients (7.71%) in the POWER trial, more were appropriately treated than patients not in the POWER Trial (81.1% vs 44.8%) (p<0.0001).

Conclusions

Despite long-standing guideline changes, RT utilization remains high. This study highlights how a novel patient-centered approach to guide adjuvant therapy decisions may increase the number of appropriately treated patients.

Introduction

The incidence of breast cancer is increasing as the population ages, and nearly half of new diagnoses of breast cancer are in women older than 65 years.¹ Thus, treatments specific to the geriatric population are increasingly relevant. Although early-stage, estrogen receptor-positive (ER+) breast cancer has typically been treated with breast-conserving surgery (BCS) followed by adjuvant radiation (RT) and endocrine therapy (AET), older breast cancer patients are more likely to have deviations from the standard of care in surgical management, adjuvant radiation, chemotherapy, and endocrine therapy.² These non-standard treatment approaches for older breast cancer patients reflect physician recommendations and patient preferences. More studies are needed in this population to allow for age-specific, evidence-based algorithms to guide clinical decision-making.

To this effect, two clinical trials were developed to evaluate adjuvant RT omission in older women. In the Cancer and Leukemia Group B (CALGB) 9343 trial, women aged 70 and older with node-negative tumors 2 cm or less were randomized to receive RT or not after BCS. All participants were treated with AET. The trial demonstrated that RT omission results in only a small increase in ipsilateral breast tumor recurrence (IBTR) with no impact on distant metastases or five-year mortality.^3,4 Based on this trial, in 2005, the National Comprehensive Cancer Network (NCCN) guidelines were updated to include RT omission as a treatment option.⁵ The PRIME II trial, which had a similar design to the CALGB study but used expanded criteria to include women 65 and older and tumors up to 3 cm, showed consistent results, further supporting RT omission in older women treated with BCS who take AET.^6,7

Since the 2005 NCCN guideline change, there has been a slight decrease in RT utilization nationally; however, radiation rates remain high, and most older women are still treated with RT.^8–10 The reason for this is multifactorial, including known poor adherence to AET, especially in older women, and challenges with de-escalation of treatment.^11–14 As a result, there is a concern for overtreatment in patients who still receive both RT and AET following BCS, leading to an increased risk of treatment-related morbidities and higher healthcare costs.^15–17 Conversely, patients who choose to omit RT and then do not complete the recommended course of AET are at risk of undertreatment and, thus, worse oncologic outcomes.¹⁸ Defining the rates and trends of overtreatment and undertreatment in this population is necessary to guide interventions focused on increasing the proportion of appropriately treated patients.

Adjuvant treatment decisions in this population are nuanced and must factor in individual patient preferences and priorities. The Pre-Operative Window of Endocrine Therapy to Inform Radiation Therapy Decisions (POWER) trial aims to facilitate a patient-centered approach to adjuvant therapy. In the POWER trial, participants take 90 days of pre-operative endocrine therapy (pre-ET), allowing them to assess tolerance of ET before deciding about RT.¹⁹ The POWER trial enables patients to make more informed and individualized decisions about their breast cancer treatment. The primary endpoint of the POWER trial is if taking pre-ET affects patient and physician preference for RT.

The purpose of the current study was to determine the rates of under and overtreatment in older patients with early-stage, clinically node-negative, ER+ breast cancer. We also sought to analyze whether these rates differed among POWER trial participants versus non-POWER trial participants. We hypothesized that most older patients would not receive the appropriate level of treatment and that patients participating in the POWER trial would be more likely to receive appropriate treatment.

Materials and Methods

The University of Virginia Cancer Registry was queried to identify patients aged 65 or older with invasive, non-metastatic, ER+, HER2− breast cancer measuring less than 3cm who underwent BCS between 2012 and 2022. Patients with recurrent breast cancer, concurrent bilateral breast cancer, or positive sentinel lymph nodes were excluded. Additionally, patients who received neoadjuvant chemotherapy or underwent axillary lymph node dissections were also excluded. This study was ruled exempt by the University of Virginia Institutional Review Board.

The POWER trial had similar eligibility criteria with the added exclusion of patients with breast cancer measuring more than 2cm, previous or current use of AET for breast cancer, pregnancy or lactation, a known additional malignancy that is progressing or requires active treatment, and current or planned use of a CYP2D6 inhibitor. The POWER Trial was separately approved by the University of Virginia Institutional Review Board, and enrolled participants were appropriately consented.

Patient and tumor characteristics were collected from the electronic medical record (EMR). Patient health literacy was obtained using a three-item questionnaire developed and validated by Chew et al.²⁰ This health literacy questionnaire has been a long-standing component of the clinic’s standard intake paperwork.^21,22 Health literacy was graded on a scale from 3 to 15, with low health literacy classified as a score greater than or equal to 7 and high health literacy classified as less than 7. Both AET initiation and adherence data were obtained from the EMR. Patients were considered adherent to AET if they took the medication for at least 18 months post-operatively. Patients who were less than 18 months post-op but adherent to AET at their most recent follow-up were also considered adherent. Two patients were excluded from the treatment analysis due to unknown AET adherence. IBTR and death were evaluated up to five years following surgery.

Patients were categorized into three treatment groups based on their adjuvant therapies: undertreated, overtreated, or appropriately treated. Undertreated was defined as those treated with BCS only. Overtreated was defined as patients treated with BCS, RT, and AET. Appropriately treated patients were treated with BCS and either RT or AET.

Univariate linear regression was used to determine the change in RT utilization over time. Chi-square tests were utilized to evaluate the association between demographics and tumor characteristics with the treatment category. Odds ratios were used to compare the likelihood of clinical outcomes between treatment categories. Chi-square, Fisher’s exact, and Student’s t-tests were used to assess differences between those who participated in the POWER trial and those who did not. For the non-normally distributed continuous variables, permutation tests were used to assess for group differences between the power and non-power groups. P-values less than 0.05 were deemed statistically significant. Statistical analysis was performed in R version 4.3.1.

Results

The initial cohort consisted of 480 breast cancer patients who were predominantly white (90.0%), with a mean age of 73 years (Table 1). The POWER trial and non-POWER trial groups were largely similar. Compared to POWER participants, non-POWER trial participants were more likely not to have undergone SLNB (p<0.00001). This is reflective of the POWER trial inclusion criteria as well as the national trend away from SLNB over time in this patient population due to the Choosing Wisely Guidelines.²³

Table 1.

Comparison of non-POWER and POWER trial participant demographics and cancer characteristics

	Non-POWER Participants, n (%) n = 443	POWER Participants, n (%) n = 37	Total	P-value
Age, y (median [Q1, Q31)	72.0 [69.0, 77.0]	75.0 [71.0, 78.0]	73.0 [69.0, 77.0]	.18
Age Group, y
65 – 69	141 (31.8)	8 (21.6)	149 (31.0)	0.44
70 – 79	230 (51.9)	22 (59.5)	252 (52.5)
80+	72 (16.3)	7 (18.9)	79 (16.5)
Race
White	398 (89.8)	34 (91.9)	432 (90.0)
African American	37 (8.4)	2 (5.4)	39 (8.1)	0.60
Asian	1 (0.23)	0 (0)	1 (0.2)
Other	7 (1.6)	1 (2.7)	8 (1.7)
Hispanic Origin
Yes	5 (1.1)	0 (0)	5 (1.0)	1.00
No	438 (98.9)	37 (100.0)	475 (99.0)
Health Literacy* (median [Q1, Q3])	5.0 [3.0, 7.0]	5.0 [3.0, 6.5]	5.0 [3.0, 7.0]	.18
Health Literacy
Low (HL >= 7)	84 (19.0)	9 (24.3)	93 (19.4)	0.72
High (HL < 7)	210 (47.4)	26 (70.3)	236 (49.2)
Unknown	149 (33.6)	2 (5.4)	151 (31.5)
Histology
IDC	335 (75.6)	32 (86.5)	367 (76.5)
ILC	63 (14.2)	5 (13.5)	68 (14.2)	0.18
Invasive (Other)	44 (9.9)	0 (0)	44 (9.2)
Paget’s	1 (0.23)	0 (0)	1 (0.2)
Tumor Grade
1	210 (47.4)	18 (48.6)	228 (47.5)
2	181 (40.9)	18 (48.6)	199 (41.5)	0.12
3	39 (8.8)	0 (0)	39 (8.1)
Unknown	13 (2.9)	1 (2.7)	14 (2.9)
Tumor Size**, mm (median [Q1, Q3])	11.0 [7.0, 15.0]	9.0 [7.0, 12.0]	10.0 [7.0, 15.0]	.09
Tumor Size, mm
0 – 9	157 (35.4)	19 (51.4)	176 (36.7)
10 – 19	186 (42.0)	12 (32.4)	198 (41.3)	0.074
20 – 29	50 (11.3)	1 (2.7)	51 (10.6)
Unknown	50 (11.3)	5 (13.5)	55 (11.5)
Nodal Staging
NX	147 (33.2)	27 (73.0)	174 (36.3)	<0.00001
N0	296 (66.8)	10 (27.0)	306 (63.8)

^*n = 329

^**n = 425

The majority of study participants (n=300, 62.5%) had RT, with 151 (32.3%) receiving whole breast RT, 137 (29.2%) receiving partial breast RT, and 12 receiving either both types or an unknown type of RT. The annual rate of RT utilization remained stable throughout the study period (p=0.49). The 396 (82.5%) patients who initiated AET took the medication for an average of 37 months. Of these patients, 320 (81.2%) adhered to AET for at least 18 months or, for those less than 18 months post-op, were adherent at their last follow-up visit.

When separating the entire cohort into the treatment categories, 62 (13.0%) were undertreated, and 202 (42.3%) were overtreated. There were 214 (44.8%) patients classified as appropriately treated, with 96 (20.1%) treated with BCS and RT and 118 (24.7%) treated with BCS and AET. For each year from 2012 to 2022, the proportion of patients in each treatment group was determined according to the year the patient underwent surgery (Figure 1). While the proportion of undertreated patients remained lower than the appropriately and overtreated patients throughout the study period, no overall trend was noted.

Figure 1. Treatment Groups from 2012 to 2022.

For each year from 2012 to 2022, the proportion of patients in each treatment group was determined according to the year the patient underwent surgery. The proportion of patients in each treatment group varied from year to year. The number of undertreated patients was consistently lower than the number of overtreated and appropriately treated.

Characteristics between treatment groups were analyzed (Table 2). A significant association was found between the treatment category and age. Appropriately treated patients were more likely to be between 70–79 years, overtreated patients were more likely to be between 65–69 years, and undertreated patients were more likely to be 80 or older (p<0.0001). Finally, appropriately treated patients were more likely to have high health literacy than undertreated patients (p=0.0003). No differences were found across treatment groups for race, tumor grade, or tumor size.

Table 2.

Patient and Tumor Characteristics by Treatment Group

	BCS Only (Undertreatment), n (%) n = 62	BCS+RT or BCS+AET (Appropriate Treatment), n (%) n = 214	BCS+RT+AET (Overtreatment), n (%) n = 202	P value*
Age Group (years)				<0.0001
65 – 69	5 (8.1)	36 (16.8)	108 (53.5)
70 – 79	31 (50.0)	130 (60.8)	90 (44.6)
80+	26 (41.9)	48 (22.4)	4 (2.0)
Race				0.99
White	56 (90.3)	192 (89.7)	182 (90.1)
Other	6 (9.7)	22 (10.3)	20 (9.9)
Health Literacy				0.0003
Low (HL >= 7)	22 (35.5)	39 (18.2)	31 (15.3)
High (HL < 7)	20 (32.3)	97 (45.3)	119 (58.9)
Unknown	20 (32.3)	78 (36.4)	52 (25.7)
Tumor Grade				0.27
1	30 (48.4)	108 (50.5)	89 (44.1)
2	25 (40.3)	85 (39.7)	88 (43.6)
3	1 (1.6) 4	18 (8.4)	20 (9.9)
Unknown	6 (9.7)	3 (1.4)	5 (2.5)
Tumor Size, mm				0.75
0 – 9	24 (38.7)	83 (38.8)	69 (34.2)
10 – 19	26 (41.9)	83 (38.8)	88 (43.6)
20 – 29	8 (12.9)	24 (11.2)	18 (8.9)
Unknown	4(65)	24 (11.2)	27 (13.4)

^*P value is for a % χ² value between each covariate and group.

Sixteen (3.4%) patients had an IBTR within five years of surgery (Table 3). The odds of an IBTR occurring in the undertreated group was 4.4 times higher than in those appropriately treated (confidence interval [CI] 1.41–14.22) (Figure 2). Regarding all-cause mortality, 21 patients (4.4%) died within five years of surgery. Compared with the appropriately treated group, those who were undertreated were 3.8 times more likely to die (CI 1.35–10.8).

Table 3.

Five-year Recurrence and Mortality by Treatment Group

	BCS Only (Undertreatment), n (%) n = 62	BCS+RT or BCS+AET (Appropriate Treatment), n (%) n = 214	BCS+RT+AET (Overtreatment), n (%) n = 202	Total, n (%) n = 478
5-Year Ipsilateral Breast
Tumor Recurrence
POWER	7 (11.3)	6 (2.8)	3 (1.5)	16 (3.3)
Non-POWER	55 (88.7)	208 (97.2)	199 (98.5)	462 (96.7)
5-Year All-Cause
Mortality
POWER	8 (12.9)	8 (3.7)	5 (2.5)	21 (4.4)
Non-POWER	54 (87.1)	206 (96.3)	197 (97.5)	457 (95.6)

Figure 2. Comparison of clinical outcomes by treatment group.

The odds of recurrence and mortality were calculated for each treatment group, using appropriate treatment as the reference. Patients who were undertreated had a higher odd of recurrence and mortality within 5-years post-surgery.

Thirty-seven (7.7%) patients were POWER trial participants. The majority (81.1%) of POWER participants were categorized as appropriately treated compared to 44.8% of the whole cohort. The five POWER trial participants who were considered over-treated tolerated pre-ET well but opted for RT to minimize their risk of recurrence. Of the three patients treated with BCS alone, two did not tolerate pre-ET but still opted to forgo RT. Two of these participants did not pursue RT due to their age (both over 85 years old). The third patient tolerated pre-ET and, therefore, opted to omit RT; however, she ultimately stopped AET after six months due to side effects that she attributed to the AET. Patients in the trial were more likely to be appropriately treated than those who were not in the trial (p<0.0001) (Figure 3).

Figure 3. Treatment category by POWER trial participation.

The POWER trial participants and non-trial participants were categorized by their adjuvant treatment. The majority of POWER trial participants were appropriately treated. Patients in the trial were more likely to be appropriately treated than those who were not in the trial (p<0.0001)

Discussion

This single-institution study demonstrates that the overtreatment of women 65 and older with early-stage breast cancer continues to be prevalent despite long-standing changes to NCCN guidelines. Undertreatment, due to lack of AET adherence after RT omission, remains present as well, albeit at a lower rate than over and appropriate treatment. Throughout the study period, the proportion of patients in the treatment groups varied, with a majority consistently being overtreated. Participants in the POWER trial were more likely to be treated appropriately than non-POWER participants. The POWER trial presents an approach for facilitating patient-centered treatment strategies, potentially increasing the proportion of patients receiving optimal treatment.

When assessing AET adherence in research studies, there is no standard amount of time on therapy consistently used to qualify patients as adherent. Often, studies use AET initiation as a surrogate measure of AET adherence, even though it is well known that many who initiate AET do not complete the recommended time course, especially older patients.^13,14,24,25 Failing to assess AET adherence for at least five years is a limitation of prior studies and this study. We attempted to improve upon the commonly employed approach of reporting only initiation of AET by instead defining AET adherence as use for at least 18 months. Our data showed that about 81% of those who initiated AET were adherent at 18 months. While this adherence rate is likely still higher than if we had tracked adherence for five years, our rate is closer to most published 5-year adherence rates than if we had considered all initiators as adherent.^14,25

We found that women aged 80 or older were more likely to be undertreated, which is similar to previously published studies. For example, a SEER database study found that the proportion of patients receiving RT decreased as age increased.⁸ There are many factors likely contributing to the undertreatment of older breast cancer patients, including worse baseline health, shorter life expectancy, higher disease severity, and susceptibility to treatment side effects.²⁶ Additionally, older women have a higher comorbidity burden, which is significant since having increased comorbidities is independently associated with medication nonadherence.^24,27–29 We also found that women aged 65–69 were most likely to be overtreated. This was unsurprising given that the PRIME II trial, which showed results consistent with the CALGB 9343 trial in patients aged 65–70 in addition to those aged 70+, was published more recently and may not have impacted treatment recommendations yet.^3,6

In our study, patients with low health literacy were at a higher risk of undertreatment than those with adequate health literacy. Poor health literacy is associated with decreased medication adherence and worse health outcomes.^30,31 Given the complex nature of cancer treatment and the importance of shared decision-making between patients and their clinicians, low health literacy can be a significant barrier to understanding treatment guidelines and adjuvant therapy options.³² While our study only found age and health literacy to be associated with the treatment category, others have noted additional factors not included in this study, including marital status, socioeconomic status, and comorbidity index as predictors of RT omission and AET nonadherence.^8,27,29

One of the primary concerns with the omission of RT is the possibility of AET nonadherence and, therefore, undertreatment, which is known to be associated with worse oncologic outcomes.^18,33 As expected, this present study demonstrated that undertreated patients had higher odds of IBTR and mortality within five years than those who were appropriately treated. In the CALGB 9343 and PRIME II trials, the addition of RT with BCS and AET led to a decrease in IBTR rates.^3,6 However, it should be noted that the absolute differences in local recurrence rates were small, and the addition of RT did not affect regional or distant metastases, breast cancer-specific survival, or overall survival.^3,6 We did not find that overtreated patients were less likely to have an IBTR than the appropriately treated patients. While this is a small, retrospective study, the results are consistent with other studies using data from the non-clinical trial setting.³³

The POWER Trial is uniquely designed to allow patients a trial period of endocrine therapy to inform their decision about adjuvant treatment. The trial is designed to determine if pre-ET impacts patients’ decisions regarding RT. While neoadjuvant endocrine therapy is not a new concept, it has primarily been used to downstage disease in patients with locally advanced or large breast cancers.³⁴ However, using pre-ET to assess tolerance is novel to the POWER trial. Pre-ET allows patients and their clinicians to engage in shared decision-making about adjuvant therapy with an understanding of the patient’s tolerance to AET. Surgeon opinion and guidance are known to be of paramount importance in formulating patient treatment regimens. In a study by Shumway et al. on breast cancer patients’ perspectives on RT, the majority of patients who omitted RT reported that it was the result of their physician letting them know that it was not needed.³⁵ The POWER trial sets up the framework for this discussion and provides the patient and physician with data on which to focus their discussion. While the final analysis is pending, the strength of the trial is reflected in the lower proportion of POWER trial participants being overtreated compared to non-POWER participants.

This study is not without limitations. First, it is a retrospective study at a single institution, limiting generalizability. As such, we do not know if our institutional treatment patterns represent national treatment trends. Similarly, we could only examine the implementation of the POWER trial at one institution with a limited number of trial participants included in this analysis. Additionally, given our smaller sample size, we could only analyze all-cause mortality, not breast cancer-specific mortality.

Overall, this study shows that the overtreatment of early-stage breast cancer in older women is still prevalent. Given that the rates of overtreatment and RT utilization have not decreased dramatically since the publication of the NCCN guidelines, new approaches to patient-centered decision-making are needed to encourage appropriate treatment. The POWER trial introduces a promising intervention that aims to do precisely this. With accrual now closed, data on whether participation in the POWER trial shifted patients’ plans for RT treatment is shortly forthcoming. Future work will include a randomized clinical trial to further assess the effect of POWER trial participation on patient decisions regarding adjuvant treatment.