Papillary lesions of breast on core-needle biopsy: Factors associated with malignant upgrade on surgical excision

INTRODUCTION

Papillary lesions (PLs) of the breast encompass a broad spectrum of pathologies, ranging from intraductal papillomas, papillomas with atypia to malignant lesions including papillary ductal carcinoma in situ (DCIS) and encapsulated papillary carcinomas, solid papillary carcinomas, and invasive papillary carcinomas. They are characterized by the presence of finger-like projections (papillae), each of which is composed of a stromal core overlain by a layer of epithelial cells.^[1-3] Due to their intralesional heterogeneity and malignant potential, PLs diagnosed in core-needle biopsy (CNB) are classified as lesions of uncertain malignant potential (B3 lesions) in the European guidelines.^[4]

The reported rates of upgrading from PLs diagnosed in CNB to malignancy in subsequent surgical excision range from 0% to 29%,^[5] with intraductal papillomas having a lower upgrade rate (<10%) than PLs with concomitant atypia, which have been reported to have upgrade rates as high as 36%.^[4] Given the substantial upgrade rate, surgical excision is traditionally recommended for PLs diagnosed by CNB. However, there has been a trend of de-escalation in management for PLs without atypia, prioritizing minimally invasive approaches for low-risk cases while reserving surgery for high-risk lesions. Guideline proposed by the American Society of Breast Surgeons suggests observation with clinical and imaging follow-up for PLs without atypia.^[6] While guidelines proposed by the UK and Europe suggest that PLs without atypia can be safely managed with vacuum-assisted excision (VAE), and if fully excised, imaging surveillance is sufficient. Surgical excision remains the standard of care for PLs with atypia due to their higher malignant potential.^[4,6,7]

This study aims to identify factors associated with upgrade to malignancy in patients with PLs diagnosed on CNB, facilitating risk stratification that may spare low-risk patients from unnecessary open surgery.

MATERIAL AND METHODS

Patients

The medical records of female patients diagnosed with B3 PLs by ultrasound-guided CNB at our tertiary institute from January 2008 to December 2022 were retrieved and reviewed. From an initial cohort of 341 PLs diagnosed on CNB, we excluded 18 cases with synchronous malignancy, 45 cases with incomplete data (e.g., missing ultrasound imaging, clinical information, or final surgical pathology), and 59 non-surgical cases with <2 years of imaging or clinical follow-up. The final study cohort included 219 CNB-diagnosed PLs of the breast in 172 female patients (mean age of 53 ± 12 years, range 23–84), of which 166 PLs were managed with surgical excision, and the remaining 53 PLs were managed conservatively with at least 2 years of sonographic or clinical surveillance [Figure 1].

Close

Figure 1:

Diagram outlines the process for selecting eligible cases for analysis.

Export to PPT

RESULTS

Among the 166 PLs that were excised, 18 of them were upgraded to malignancy, including eight DCIS, seven invasive ductal carcinomas, two encapsulated papillary carcinomas, and one invasive papillary carcinoma being diagnosed in the surgical excision. The overall upgrade rate is 8.2%.

Additionally, 53 cases were managed conservatively with at least 2 years of follow-up, with surveillance conducted either by ultrasound (n = 45) or clinical assessment (n = 8). None of the patients under surveillance exhibited significant lesion growth or malignant transformation during the ≥2-year follow-up period.

Table 1 shows the comparison of the clinical, sonographic and histopathologic characteristics of PLs upgraded to malignancy to those not upgraded. Comparison between the two groups identified four factors with statistical significance in univariate analysis. Multivariate analysis identified four significant independent predictors of upgrade, including lesion size ≥1.0 cm (OR, 3.56; 95% CI 1.03–12.28; p = 0.045), intralesional vascularity (OR, 6.98; 95% CI 2.16–22.53; p = 0.001) [Figure 2a and b], non-circumscribed margin (OR, 6.41; 95% CI 1.48–27.71; p = 0.013) [Figure 3a and b], and presence of atypia in CNB sample (OR, 6.59; 95% CI 1.87–23.23; p = 0.003). Presence of these characteristics in PLs was associated with upgrade rates as high as 15.4%, 28.1%, 27.8%, 29.1% respectively. In our cohort, PLs without any of these three significant independent factors, i.e. subcentimeter circumscribed lesions without intralesional vascularity on color Doppler and no atypia on CNB, had an upgrade rate of 3.4% (3 of 88), and a negative predictive value of 96.6%. PLs with all four of these factors have a 100% upgrade rate (n = 1).

Close

Figure 2:

(a) A 83-year-old female with palpable left breast mass. Intralesional vascularity was demonstrated by color Doppler ultrasound (white arrow). It was diagnosed as papillary lesion on core-needle biopsy and confirmed to be encapsulated papillary carcinoma after surgical excision. (b) A 64-year-old female with spontaneous bloody discharge from right nipple. Internal vascularity was shown on colour Doppler ultrasound (white arrow). Core needle biopsy revealed intraductal papilloma, which was upgraded to invasive papillary carcinoma on excision.

Export to PPT

Close

Figure 3:

(a) An 83-year-old female with a palpable left breast mass. Ultrasound of the left breast showed an irregular hypoechoic mass with an angulated margin (white arrow). It was diagnosed as a papillary lesion on core-needle biopsy and confirmed to be encapsulated papillary carcinoma after surgical excision. (b) A 57-year-old female with bloody right nipple discharge. Ultrasound of the right breast in radial view showed an irregular hypoechoic mass with microlobulated margin (white arrow), closely related to a dilated duct (dashed white arrow). Core needle biopsy revealed intraductal papilloma with atypia, which was upgraded to ductal carcinoma in situ on excision.

Export to PPT

The mean age at diagnosis in the upgrade group was 62.8 ± 14.5 years, approximately 10 years older than the mean age of the non-upgrade group. However, age is not a significant predictor of upgrade in univariate analysis (p = 0.086). Other factors, including the presence of mastalgia, single duct bloody discharge; lesion location; sonographic findings of multifocality, intraductal lesion, posterior acoustic enhancement or shadowing, were also not shown to be associated with malignant upgrade of PLs in our results.

DISCUSSION

The management of PLs of the breast diagnosed in CNB remains a clinical challenge due to their heterogeneous nature and the highly variable potential of malignant upgrade. The overall upgrade rate of 8.2% (18/219) in this cohort is comparable to the literature, ranging from 0% to 29%.^[5,8-12]

Many studies have been performed to identify clinical and radiological factors that can stratify the likelihood of upgrade to malignancy in PLs diagnosed in CNB, yet results remain inconsistent. Larger lesion size, older age of patients, symptomatic lesions, peripheral location >5cm distance from the nipple, concomitant calcifications, multiple lesions, vascularity and high Breast Imaging-Reporting and Data System category have been shown to be predictors of upgrade in previous studies.^[4,5,8-10] In this cohort, the upgrade rates were significantly higher in PLs with sonographic findings of lesion size >1 cm, intralesional vascularity or non-circumscribed margins, aligning with some of the previous studies.

The significance of vascularity as a predictor of malignancy remains controversial.^[8,18,19] We observed the presence of intralesional vascularity in 14.6% (32/219) of intraductal papillomas, with 28.1% (9/32) upgraded to malignancy upon excision. This observation may be partly attributed to the sonographic characteristics of papillomas and neoangiogenesis in malignancy.^[18,19] On ultrasound, papillomas typically show vascular pedicles within the central core with branching vessels arborizing within the mass.^[8,20]

PLs with atypia on CNB have been consistently shown to be associated with a higher risk of malignancy in PLs. Intraductal papillomas without atypia have a lower upgrade rate of 1–9% compared to lesions with concomitant atypia, of which the upgrade rate could be as high as 38% with a median of 26.9%.^[16,21] Our data showed upgrade rates of PLs with and without atypia were 29.2% (7 of 24) and 5.6% (11 of 195), respectively, which are comparable to those from the previous studies.^[5,12] In addition, reported risk of subsequent breast cancer is 5–7 times higher in patients with PLs with atypia compared to the normal population.^[4,9,14,15] Due to this high malignant potential, surgical excision is a well-established clinical management of PLs with atypia and is supported by guidelines in UK, Europe, and America.^[4,6,7] This underscores the importance of surgical excision for PLs with atypia to rule out malignancy.

For PLs without atypia, there has been a trend of de-escalation in management. Guideline proposed by the American Society of Breast Surgeons suggests observation with clinical and imaging follow up for PLs without atypia.^[6] While guidelines proposed by UK and Europe suggest PLs without atypia can be safely managed with VAE, and if fully excised, imaging surveillance is sufficient.^[4,7] The observed upgrade rate was 3.4% (3 of 88) for PLs lacking all identified risk factors, with a negative predictive value of 96.6%. The upgrade rate is below the upper risk limit proposed by the Third International Consensus Conference (2023) of 5% for invasive carcinoma and 10% for DCIS in order to refer to radiologic surveillance.^[22] Notably, none of the patients under surveillance exhibited significant lesion growth or malignant transformation during the ≥2-year follow-up period, which is also consistent with literature.^[13,23] Therefore, minimally invasive management, such as VAE, can be considered for subcentimeter circumscribed PLs without vascularity and atypia.

VAE has been recommended as a cost-effective alternative to surgical excision for managing B3 lesions including PLs without atypia, as well as classical lobular neoplasia, radial scar and flat epithelial atypia.^[4,7] VAE utilizes a vacuum-powered device to remove larger volumes of tissue compared to standard CNB, allowing more representative tissue sampling or even complete removal of the lesion. The success rate of VAE in the literature is consistently high with an overall success rate of 97–100%, with effective complete lesion removal, low rates of residual/recurrent disease and excellent cosmetic outcomes.^[24-28] Choi et al. found a low rate of upgrade (2.1%, 5 of 233), residual (0.9%, 2 of 228) and recurrence (3.6%, 6 of 166) in the PLs managed with ultrasound-guided directional VAE, with the recurrent lesions proven benign after surgical excision.^[27] Similarly, Yi et al. reported no recurrence in 136 solitary intraductal papillomas excised using an 8-gauge Mammotome system.^[28] A local study in Hong Kong by Chau et al. further supported these findings, reporting a 97.2% success rate in complete lesion removal using ultrasound-guided VAE for PLs of smaller than 2.2 cm, with no major complications and a low cancer upgrade rate of 5.3%.^[24] These studies collectively support the use of VAE as a safe and effective alternative to surgical excision for management of PLs. In addition, VAE can be performed as an outpatient procedure, allowing patients to be discharged on the same day, which significantly reduces healthcare costs compared to surgical excision.

There were several limitations in this study. This study was a retrospective study conducted at a single tertiary institution, which may limit the generalizability of the findings. In addition, the decision to excise or follow-up PLs could be influenced by lesion factors, patient and clinician preference, which may have introduced selection bias. Furthermore, the relatively small number of upgraded cases (n = 18) limits the power of the statistical analysis and the ability to perform more detailed subgroup analyses. Further validation through a prospective, multi-institutional study with a larger cohort is needed to confirm these findings and refine the risk-stratification model before broader clinical adoption.

CONCLUSION

Lesion size ≥1 cm, intralesional vascularity, non-circumscribed margins, and presence of atypia on CNB specimens are identified as independent predictors of malignant upgrade in PLs. These factors can be used to stratify patients into high-risk and low-risk groups, guiding the decision between surgical excision and minimally invasive management. VAE can be considered for low-risk PLs (subcentimeter circumscribed lesions without intralesional vascularity or atypia), given the low upgrade rate of 3.4% and high negative predictive value of 96.6%. However, surgical excision remains the standard of care for high-risk PLs, particularly for those with atypia, due to their high malignant potential.