Subsequent Percutaneous Breast Biopsies after Initial Atypia Diagnosis: The Patient Burden of Long-Term Follow Up

Abstract

Background:

Breast atypia increases overall breast cancer risk, potentially necessitating future interventions. This study examines the frequency and outcomes of additional percutaneous biopsies after an atypia diagnosis.

Methods:

Adult patients with breast atypia (atypical ductal hyperplasia, atypical lobular hyperplasia, lobular carcinoma in situ) at a single institution were reviewed for subsequent core needle biopsies (CNBs). Logistic regression assessed associations.

Results:

Among 432 patients, median age at diagnosis was 54.8y. During a median follow-up of 88.4mo, 149 CNBs were performed. Sixty-two (14.4%) patients developed a breast malignancy. Within 5 years post-atypia, 79 (69.9%) underwent at least one CNB. The presence of a pathogenic mutation was associated with subsequent CNB (p=0.01), but age, atypia type, and chemoprevention were not (all p>0.05).

Conclusion:

Many patients with breast atypia undergo additional percutaneous biopsies, especially within 5 years post-atypia diagnosis. Our study highlights the significant burden of surveillance and the need for tailored follow-up strategies.

INTRODUCTION

More than 1 million women in the United States undergo breast biopsies each year, resulting in diagnoses collectively referred to as benign breast disease.^1–3 This group of diagnoses is historically classified into three categories, non-proliferative lesions, proliferative lesions without atypia, and proliferative lesions with atypia, based on the degree of cellular proliferation and atypia.^4–6 Among these, proliferative lesions with atypia – namely atypical ductal hyperplasia (ADH), atypical lobular hyperplasia (ALH), and lobular carcinoma in situ (LCIS) – carry the highest risk of being diagnosed with breast cancer. We often collectively refer to these conditions as “breast atypia”.^7,8

Despite the recognized increased risk of breast cancer associated with atypia, the magnitude of this risk and the factors that influence it remain less clear, largely due to lack of prospective data. This uncertainty had made the management of breast atypia controversial.^9,10 While the National Comprehensive Cancer Network (NCCN) recommends surgical excision following a diagnosis of ADH on core needle biopsy (CNB), not all atypia cases require surgical excision, and some patients, such as those with ALH or LCIS, are often more suitable for ongoing surveillance.¹¹ In addition, monitoring after an atypia diagnosis also varies, as the NCCN guidelines recommend ‘consideration’ of breast MRI.¹¹ However, adherence to these guidelines varies between providers and institutions. Ongoing research aims to identify factors associated with a lower risk of developing malignancy, which could inform decisions when additional imaging abnormalities are identified.^9,12–14

The variability and uncertainty surrounding risk stratification complicate decision-making for patients following a diagnosis of breast atypia, creating an unpredictable clinical course. As such, we evaluated the patient burden in terms of the long-term clinical course following an initial diagnosis of atypia. The primary objective was to determine the number and frequency of additional CNBs that patients underwent and the corresponding malignant outcomes. The secondary objective was to assess the potential factors associated with receipt of additional CNBs, including patient characteristics and risk factors.

MATERIAL AND METHODS

Adult women with breast atypia at a single academic institution from 01/2008-10/2017 were retrospectively identified. Notably, some patients diagnosed between 2008 and 2017 also had a prior diagnosis of atypia, which was as early as May 1992. Atypia type was defined as atypical ductal hyperplasia (ADH), atypical lobular hyperplasia (ALH), lobular carcinoma in situ (LCIS), or other. A waiver of informed consent was granted given the methodology of the retrospective chart review, and the institutional review board granted the study exempt status. Patients with a history of prior breast cancer or concurrent breast cancer (defined as within 180 days of atypia diagnosis or at any point prior to atypia diagnosis) were excluded, as were those with missing or unknown type of atypia. Patients diagnosed with atypia at age <18 years were also excluded. Additional procedures were identified for each patient as any core needle biopsy, surgical biopsy, lumpectomy, mastectomy, and other breast surgery that occurred after atypia diagnosis up until a subsequent breast malignancy diagnosis, if the patient developed a malignancy. Core needle biopsies occurring within 3 months and surgical biopsies occurring within 6 months of the first atypia diagnosis were considered part of initial management and not “additional procedures”, given that surgical excision is the standard of care for some atypia diagnoses at the time of initial diagnosis (to rule out a concurrent malignancy, often due to undersampling). This was done to focus on subsequent interventions conducted due to new findings, after the immediate atypia management decisions had been made, and therefore highlighting the long term clinical course of patients with breast atypia after their initial management. Patients were grouped by the number of additional core needle biopsies: 0 or ≥1.

Patient characteristics and procedure data were summarized with N (%) for categorical variables and median (interquartile range, IQR) for continuous variables. Differences were tested using chi-square or Fisher’s exact tests for categorical variables, as appropriate, and analysis of variance, Kruskal-Wallis tests, t-tests, or Wilcoxon rank sum tests for continuous variables, as appropriate. Procedure characteristics were summarized on the procedure level for patients who had ≥1 additional procedures. Logistic regression was used to estimate the association of select covariates with the likelihood of undergoing ≥1 additional CNBs. Odds ratios (ORs) and 95% confidence intervals (CIs) are reported. The variables included in the multivariable modeling were selected based on clinical relevance, and in order to avoid over-fitting the models, we performed multiple separate models with the number of variables that could be reasonably included based on the number of observed events (i.e. CNBs).

Follow-up time was estimated as the time from date of first atypia diagnosis to date of subsequent malignancy or death. Patients who were not diagnosed with malignancy or died were censored at date of last known breast visit. Follow-up time was estimated with the reverse Kaplan-Meier method. A P value ≤ 0.05 was considered statistically significant; no adjustments were made for multiple comparisons. All statistical analyses were conducted with SAS version 9.4 (SAS Institute, Cary NC).

RESULTS

A total of 560 patients with breast atypia were identified. After excluding those with prior or concurrent breast cancer diagnoses, missing/unknown atypia types, or age at atypia diagnosis <18 years, our final cohort was comprised of 432 patients (Figure 1) with a median follow-up of 7.4 years (95% CI 7.0-7.9). The median age at initial breast atypia diagnosis was 54.8 years (IQR 47.8-63.6), and the median age at the time of the first additional CNB was 60.4 years (IQR 51.0-68.4). Among the entire cohort, ADH was the most common type of atypia (72.2%), followed by LCIS (11.8%) and ALH (9.5%). Of the 432 patients, 148 (34.3%) underwent a surgical procedure (i.e. excisional biopsy, lumpectomy, or mastectomy) within 3 months of their initial atypia diagnosis, excluding those diagnosed with cancer during that time. The rate of early surgical intervention varied by atypia type: 34.4% of patients with ADH, 34.1% with ALH, and 25.5% with LCIS underwent surgery within this timeframe. The most common imaging abnormality that led to the initial atypia diagnosis was calcifications (48.8%), followed by mass/density (29.2%). Seven patients had a known pathogenic mutation, including one each with BRCA1, BRCA2, and ATM, and four with CHEK2 mutations. Approximately 14% of patients in our primary cohort received chemoprevention (n=62). Of the 432 patients included in our primary cohort, 71 (16.4%) eventually developed breast cancer during follow-up. (Table 1)

Figure 1.

Patient cohort selection consort diagram.

Table 1.

Patient characteristics for adult patients diagnosed with breast atypia at a single institution, stratified by subsequent core needle biopsy (CNB) receipt (=0 vs ≥1).

	All Patients (N=432)	Number of CNBs		P-Value
		0 (N=319)	≥1 (N=113)
Age at Atypia Diagnosis (Years)				0.34
<50	143 (33.1%)	104 (32.6%)	39 (34.5%)
50-70	238 (55.1%)	173 (54.2%)	65 (57.5%)
>70	51 (11.8%)	42 (13.2%)	9 (8%)
Median (IQR)	54.8 (47.8 - 63.6)	54.8 (47.8 - 63.6)	54.9 (47.7 - 63.2)	0.85
Race/Ethnicity				0.62
Non-Hispanic Caucasian or White	274 (63.4%)	198 (62.1%)	76 (67.3%)
Non-Hispanic Black or African American	98 (22.7%)	75 (23.5%)	23 (20.4%)
Non-Hispanic Other	15 (3.5%)	11 (3.4%)	4 (3.5%)
Hispanic or Latino	15 (3.5%)	13 (4.1%)	2 (1.8%)
Imaging Abnormality at Initial Atypia Diagnosis				0.008
Non-Mass Enhancement	4 (0.9%)	4 (1.3%)	0 (0%)
Asymmetry/Distortion	11 (2.5%)	11 (3.4%)	0 (0%)
Calcs	211 (48.8%)	149 (46.7%)	62 (54.9%)
Mass/Density	126 (29.2%)	87 (27.3%)	39 (34.5%)
Other/Unknown	80 (18.5%)	68 (21.3%)	12 (10.6%)
Atypia Type				0.66
Atypical Ductal Hyperplasia (ADH)	312 (72.2%)	232 (72.7%)	80 (70.8%)
Atypical Lobular Hyperplasia (ALH)	41 (9.5%)	27 (8.5%)	14 (12.4%)
Lobular Carcinoma in Situ (LCIS)	51 (11.8%)	39 (12.2%)	12 (10.6%)
Other	28 (6.5%)	21 (6.6%)	7 (6.2%)
BMI – Median (IQR)	28.1 (23.6 - 33.1)	28.1 (24.0 - 33.6)	27.3 (23.1 - 31.6)	0.17
Presence of Any Mutation				0.02
No	425 (98.4%)	317 (99.4%)	108 (95.6%)
Yes	7 (1.6%)	2 (0.6%)	5 (4.4%)
Mutation Type
BRCA1	1 (0.2%)	1 (0.3%)	0 (0%)	1.00
BRCA2	1 (0.2%)	0 (0%)	1 (0.9%)	0.48
ATM	1 (0.2%)	0 (0%)	1 (0.9%)	1.00
CHEK2	4 (0.9%)	1 (0.3%)	3 (2.7%)	0.59
Family History of Cancer				0.20
No	99 (22.9%)	78 (24.5%)	21 (18.6%)
Yes	333 (77.1%)	241 (75.5%)	92 (81.4%)
Treatment with Chemoprevention				0.49
No	359 (83.1%)	269 (84.3%)	90 (79.6%)
Yes	62 (14.4%)	49 (15.4%)	13 (11.5%)
Total Additional Core Needle Biopsies				-
0	319 (73.8%)	319 (100%)	0 (0%)
1	87 (20.1%)	0 (0%)	87 (77%)
2	17 (3.9%)	0 (0%)	17 (15%)
3	8 (1.9%)	0 (0%)	8 (7.1%)
4	1 (0.2%)	0 (0%)	1 (0.9%)
Median (IQR)	0 (0 - 1)	0 (0 - 0)	1 (1 - 1)	-
Age at First Additional Core Needle Biopsy – Median (IQR)	60.4 (51.0 - 68.4)	-	60.4 (51.0 - 68.4)	0.39
Underwent Other Surgical Procedure*	148 (34.3%)	94 (29.5%)	54 (47.8%)	<0.001
Subsequent Breast Malignancy				<0.001
No	361 (83.6%)	310 (97.2%)	51 (45.1%)
Yes	71 (16.4%)	9 (2.8%)	62 (54.9%)
Follow-Up (years) – Median (95% CI)	7.4 (7.0-7.9)	6.8 (6.3-7.2)	10.9 (9.2-15.0)	<0.001

^*Includes patients who had mastectomy, lumpectomy, or excisional biopsy.

Abbreviations: IQR = interquartile range, BMI = body mass index, CI = confidence interval

Of the entire cohort (N=432), 113 patients underwent at least one additional CNB >3 months after a breast atypia diagnosis. Notably, the median follow-up was significantly longer for patients who underwent a subsequent CNB (10.9 years vs 6.8 years, p<0.001). Of those who underwent a subsequent CNB, the majority of patients (87/113, 77.0%) underwent one additional CNB, and of the remaining patients (26/113, 23%), 17 underwent two CNBs, 8 underwent three CNBs, and 1 underwent four CNBs. A total of 26 patients had ≥2 biopsies and of those, 9 (34.6%) had at least two biopsies within 3 months of each other. After stratifying based on additional CNB (=0 or ≥1), there was no significant difference in age at atypia diagnosis, race/ethnicity, insurance type, marital status, type of atypia, BMI, family history of cancer, or chemoprevention uptake (all p>0.05). However, the presence of a germline genetic mutation related to breast cancer was higher among those who underwent a subsequent CNB (p=0.02). In addition, there was a significant difference in the initial imaging abnormality that led to the first atypia diagnosis (p=0.008), although calcifications and mass/density were still the most common among both subgroups. (Table 1)

Among the 113 patients who had subsequent CNB, there were a total of 149 CNBs performed. The timing of any CNB ranged from 1 to 21 years after the initial atypia diagnosis, with 96 CNBs (64.4%) performed within the first 5 years and 37 CNBs (24.8%) performed between 6 to 10 years, representing 89% of all additional CNBs (Figure 2). Of the 113 patients who underwent additional CNBs, 62 (54.9%) were diagnosed with breast cancer during follow-up, which represents approximately 14.4% of our total cohort. The 62 patients who were diagnosed with breast cancer during follow-up underwent a total of 76 CNBs, and approximately half (53.2%) of the breast cancer diagnoses were within 5 years of the initial atypia diagnosis. Conversely, there were 51 patients (45.1%) who collectively underwent 73 CNBs which yielded benign findings.

Figure 2.

Total number of subsequent core needle biopsies by year of follow-up, ≥ 3 months after initial atypia diagnosis, for adult patients diagnosed with breast atypia at a single institution.

Multivariable modeling with select variables was conducted to examine factors that may be associated with an increased risk of undergoing a subsequent CNB. In a model that included age at atypia diagnosis, atypia type, and family history of cancer, none of these variables were found to be associated with subsequent CNB receipt (all p>0.05). Similarly, no factors were associated with CNB when examining age at atypia diagnosis, atypia type, and chemoprevention (all p>0.05). However, the presence of a germline genetic mutation was significantly associated with subsequent CNB receipt (p=0.01), after adjusting for atypia type and chemoprevention receipt (both p>0.05). Furthermore, initial imaging abnormality was also associated with the risk of additional CNB (p=0.005), after adjusting for age at atypia diagnosis and atypia type (both p>0.05). (Table 2)

Table 2.

Logistic Regression for subsequent core needle biopsies, ≥ 3 months after initial Atypia diagnosis, for adult patients diagnosed with breast atypia at a single institution.

Model Number	Model Sample Size	Variable	Odds Ratio (95% CI)	P-Value	Overall P-Value
1	432	Age at Atypia Diagnosis (Years)	1 (0.98-1.02)	0.68	0.68
		Atypia Type			0.62
		ADH	REF
		ALH	1.53 (0.76-3.08)	0.23
		LCIS	0.86 (0.43-1.74)	0.68
		Other	0.97 (0.39-2.37)	0.94
		Family History of Cancer			0.19
		No	REF
		Yes	1.44 (0.84-2.47)	0.19
2	421	Age at Atypia Diagnosis (Years)	0.99 (0.97-1.01)	0.41	0.41
		Atypia Type			0.63
		ADH	REF
		ALH	1.6 (0.78-3.3)	0.20
		LCIS	0.98 (0.46-2.09)	0.96
		Other	1.08 (0.44-2.66)	0.86
		Treatment with Chemoprevention			0.43
		No	REF
		Yes	0.76 (0.38-1.51)	0.43
3	421	Presence of Any Mutation			0.01
		No	REF
		Yes	8.72 (1.65-46.08)	0.01
		Atypia Type			0.53
		ADH	REF
		ALH	1.69 (0.82-3.49)	0.15
		LCIS	0.94 (0.44-2.03)	0.88
		Other	1.15 (0.47-2.83)	0.76
		Treatment with Chemoprevention			0.44
		No	REF
		Yes	0.76 (0.38-1.53)	0.44
4	432	Age at Atypia Diagnosis (Years)	1 (0.98-1.02)	0.64	0.64
		Atypia Type			0.75
		ADH	REF
		ALH	1.35 (0.67-2.73)	0.40
		LCIS	0.82 (0.41-1.67)	0.59
		Other	0.92 (0.37-2.28)	0.85
		Imaging Abnormality at Initial Atypia Diagnosis			0.005
		Calcs	REF
		Mass/Density	1.09 (0.67-1.77)	0.72
		Other/Unknown*	0.35 (0.18-0.7)	0.003

^*Given the low number of patients with NME, asymmetry, and distortion, these patients were included in the Other/Unknown category.

Abbreviations: ADH=atypical ductal hyperplasia, ALH=atypical lobular hyperplasia, LCIS=lobular carcinoma in situ, NME=non-mass enhancement, CI=confidence interval.

DISCUSSION

Our study provides insights into the long-term biopsy burden experienced by patients during surveillance following a diagnosis of breast atypia. A considerable proportion of patients with breast atypia undergo subsequent CNBs after their initial diagnosis and management, and a notable proportion eventually develop breast cancer. Specifically, we found that 26% of patients with breast atypia underwent at least one subsequent CNB more than three months after their initial atypia diagnosis. Most additional CNBs were performed within the first five years following the initial atypia diagnosis, a period marked by heightened concern due to the risk of breast cancer development, which is higher in the ipsilateral breast during that same time.^15,16 However, there is also evidence to suggest the subsequent risk of breast cancer persists or even increases beyond 5 years,^17–19 notably for both breasts,¹⁵ though this follow-up period accounts for only 35% of additional CNBs performed in our study. In contrast, other factors such as age, BMI, and race/ethnicity, have not been shown to be associated with an increased risk of subsequent breast cancer development,^20–22 and there have been mixed results on the association of family history and risk.^23–26 While some breast cancer risk assessment tools include a diagnosis of breast atypia,^27,28 perhaps future revisions of these tools should investigate the potential association of breast cancer risk and subsequent CNB after an atypia diagnosis.

Notably, 55% of patients who underwent additional CNBs were eventually diagnosed with breast cancer, indicating these patients may derive even greater benefit from prevention strategies, such as chemoprevention. Chemoprevention has been consistently shown to significantly decrease the risk of developing breast cancer,^29–32 and more recently, low dose tamoxifen has also proven efficacious.³³ Despite the potential benefits, the uptake of chemoprevention among these high-risk women remains low,³⁴ although age ≥50 has been shown to be associated with higher rates of uptake.³⁵ However, a cost effectiveness study of tamoxifen for primary prevention among high risk cohorts demonstrated greater effectiveness among younger patients.³⁶ Regardless, our findings suggest that additional counseling could be particularly beneficial for those requiring additional CNBs throughout their clinical course, as formalized education and counseling have been shown in other studies to improve acceptance and treatment completion rates among those with breast atypia.³⁷

Interestingly, our results showed no significant associated between atypia type (ADH, ALH, LCIS) and the likelihood of undergoing subsequent CNBs. This suggests that the clinical course for patients with breast atypia might be similar regardless of the specific atypia type, despite different management recommendations at the time of diagnosis regarding surgical excision.³⁸ Furthermore, it highlights that the chance of finding subsequent imaging abnormalities requiring CNB does not significantly differ based on the type of initial atypia. Therefore, regardless of a patient’s atypia diagnosis, if additional imaging abnormalities are observed, the type of atypia previously diagnosed should likely not alter the recommendation for or against further evaluation with CNB. Notably, however, we did not explore the potential association of various factors with the results of the additional CNBs, and this consideration likely merits further exploration.

In contrast, the presence of a pathogenic mutation (BRCA1, BRCA2, ATM, and CHEK2) in our cohort was significantly associated with an increased likelihood of undergoing additional CNBs. Although the relationship and prevalence of pathogenic mutations among patients with breast atypia remains largely unexplored, others have shown that patients with pathogenic mutations (with or without a history of breast atypia) frequently undergo invasive breast procedures, such as percutaneous biopsies,³⁹ which may be at least partially related to the increased frequency of breast cancer screening.^40,41 Regardless, our findings also highlight the importance of genetic testing and counseling for patients with breast atypia, as identifying pathogenic mutations can inform follow-up and management decisions more effectively.

Multivariable modeling was performed to assess the robustness of associations across different clinical scenarios. These models included variables of clinical interest known to be associated with future breast cancer risk,^42,43 and as such, we hypothesized that they may also impact the number of subsequent CNBs. Within the context of the variables included in the models, the presence of a genetic mutation (Model 3) and the imaging abnormalities leading to a diagnosis of atypia (Model 4), were the two variables associated with additional CNBs. Further investigation into additional factors that may be associated with subsequent CNBs, along with additional multivariable modeling, should be conducted in order to develop more robust models that could provide additional clinical insights.

Limitations of our study include its retrospective design and use of a single-institution database, which may limit the generalizability to larger patient populations. The comparison across the subtypes of atypia may be limited due to the high proportion of patients with ADH compared to ALH and LCIS, necessitating further investigation. Additionally, we did not account for frequency of imaging, breast density, and/or the imaging modality used, all of which could impact the number of CNBs performed⁴⁴, and may be important variables to analyze in future research. We also acknowledge the discrepancy in the follow-up time between the those who did and did not undergo additional CNB, which could have contributed to the observed lower frequency of subsequent breast malignancy. Unfortunately, the small sample size for genetic testing limits the robustness of our conclusions regarding the influence of pathogenic mutations on management strategies. Future research should focus on prospective, multi-institutional studies to validate these findings and explore the genetic landscape among patients with breast atypia in greater depth.

CONCLUSIONS

In conclusion, our study highlights the complex and variable clinical course of patients diagnosed with breast atypia. A substantial proportion of these patients undergo subsequent CNBs, predominately within the initial five years following their atypia diagnosis. This timeframe represents a critical window of heightened risk for malignancy, and therefore, additional monitoring may be of potential benefit during the early follow-up stages. Additionally, the presence of a pathogenic mutation has been shown to correlate with an increased likelihood of additional biopsies and elevated breast cancer risk. However, our study did not find consistent differences in this regard among ADH, ALH, and LCIS, despite their distinct risk profiles. Overall, our findings demonstrate a critical aspect of the clinical trajectory for patients with breast atypia and the long-term burden patients may face to avoid both overtreatment and underdiagnosis.

HIGHLIGHTS.

• ~25% of patients with breast atypia underwent ≥ 1 additional CNB.

• Atypia type (ADH, ALH, LCIS) was not associated with additional CNBs.

• Most additional CNBs (64%) occur within 5 years of an atypia diagnosis.

DATA AVAILABILITY

The datasets generated and/or analyzed during the current study are not publicly available due to the risk of compromising individual patient privacy but are available from the corresponding author on reasonable request.