Outcomes of atypical (B3) core biopsy lesions diagnosed across BreastScreen NSW, Australia

Introduction Atypical or B3 lesions comprise a heterogeneous group of uncertain malignant potential. B3 lesions diagnosed on core biopsy are usually recommended for diagnostic open biopsy. Identifying factors which could allow conservative management of B3 lesions would be helpful in avoiding unnecessary surgery. The aim of this study was to identify the upgrade rate to malignancy for B3 core biopsy lesions and to compare characteristics of lesions which were malignant and benign at excision. Method This retrospective study used data from BreastScreen New South Wales (NSW), Australia, of women who were diagnosed with B3 lesions on needle biopsy from 2011 to 2019. Results During the study period, 1927 B3 lesions were included. The upgrade rate to malignancy was 26.4%. Of the malignant lesions on excision, 29.6% were invasive and 69.2% were in situ. The rates of upgrade to invasive cancer and DCIS varied substantially with the core biopsy lesion type. Lesions with atypia on core biopsy had significantly higher upgrade rates to malignancy at 34.7% compared to 13.6% for lesions without atypia (p < 0.0001). Lesions with malignant pathology were significantly larger than those with benign pathology (difference = 5.1 mm (95% CI 2.7–7.5 mm), p < 0.001). Conclusions The overall upgrade rate of B3 lesions to malignancy was 26.4%. The majority of the lesions were upgraded to DCIS instead of invasive cancer. Upgrade rates varied by lesion type. Lesions with atypia had significantly higher upgrade rates to cancer compared to lesions without atypia. Malignant lesions were significantly larger than benign lesions.


Introduction
The B-classification system for evaluating core needle biopsies of the breast was introduced in 1999 by a collaborative group of 23 European pathologists.Their objective was to categorise breast lesions identified through core needle biopsy, striving for improved consistency in diagnosis and the ability to prognosticate breast cancer development [1].The system encompasses five distinct subcategories: B1, representing normal tissue; B2, comprising benign lesions; B3, indicating lesions with uncertain malignant potential; B4, classified as suspicious for malignancy; and B5, designated as malignant.By recognising the group of B3 lesions with indeterminate malignant potential, this classification bridges the gap between benign and malignant lesions [2].
Given the indeterminate malignant potential, women with B3 lesions diagnosed on core biopsy are traditionally recommended to undergo diagnostic open biopsy to obtain a definitive diagnosis [5].The majority of these women will receive non-malignant results.Identifying factors which could indicate low risk of the subtype of B3 lesion upgrading to cancer would allow conservative management of these B3 lesions and therefore avoid unnecessary surgery.
The aim of this study was to identify the upgrade rate to malignancy for B3 core biopsy lesions diagnosed at BreastScreen NSW, Australia, and to compare characteristics of malignant and benign lesions at excision.

Methods
This retrospective study utilised data obtained from the Cancer Institute, NSW, Australia.It included all consecutive women who had either a core needle biopsy or vacuum-assisted core biopsy of one or more equivocal breast lesions on imaging by mammogram and ultrasound, at one of the nine services of BreastScreen NSW, Australia, between January 2011 and December 2019.In general, microcalcifications underwent vacuum-assisted core biopsies with a 10G needle while other lesions underwent core needle biopsies with a 14G needle.Lesions with a resultant core biopsy result classified as B3 were included in the study.
Data collected included age, lesion type, lesion size, mammographic and ultrasound features, histology results from core needle and excision biopsies.
Continuous data were summarized using mean, standard deviation and/or range of the data.Categorical data were summarized using frequencies and percentages.Fisher's exact test was used to compare B3 lesions' upgrade rates to malignancy in the context of atypical features.Analysis of Variance (ANOVA) was used to compare lesion sizes of various B3 lesions and their subsequent malignancy upgrade rates.The upgrade rate to malignancy was calculated as follows: [(number of malignant cases)/(total number of B3 cases)]x100.

Results
Between 2011 and 2019, a total of 41,607 lesions underwent either core needle biopsy (14G needle) or vacuum-assisted core biopsy (10G needle) across all the nine services in BreastScreen NSW, Australia, with 2219 (5.3%) categorised as B3 lesions on histology.Of the 2219 lesions, 257 patients had no excision recorded, 22 patients' excision outcomes were recorded as null or inconclusive, 6 patients had lymphomas, 7 entries were duplicates leaving a total of 1927 B3 lesions included for analysis.The mean age of the cohort was 63.4 ± 8.7 years and the mean size of the lesions was 14.5 ± 14.6 mm.The three most common lesion types on mammogram were calcifications (46.5%), discrete masses (21.6%) and non-specific densities (11.2%).The majority of biopsies (55.3%) were performed using a 14G needle (Table 2).The breakdown of B3 lesions is shown in Fig. 1.Of the 1927 lesions on excision, 820 (42.6%) were benign, 598 (31.0%) were atypical and 509 (26.4%) were malignant (Fig. 2), giving an upgrade rate to malignancy of 26.4%.
The rates of upgrade to invasive cancer and ductal carcinoma in situ (DCIS) varied with core biopsy lesion types.Atypical papillary lesions, followed by atypical ductal hyperplasia (ADH) and other atypical lesions had higher upgrade rates of 54.2%, 35.9% and 27.6%, respectively, compared to papilloma and radial scar with upgrade rates of 18.8% and 10.8%, respectively (Table 3).
Lesions with atypia on core biopsy demonstrated significantly higher upgrade rates to malignancy at 34.7% compared to 13.6% for lesions without atypia (p < 0.0001).The upgrade rate by pathology type on core biopsy is shown in Table 4. Excision pathology, whether malignant, atypical or benign, based on the lesion nature on mammogram is listed in Table 5. Calcifications seen on mammogram showed higher rates of atypia at excisional biopsy.The upgrade rate to malignancy for lesions sampled with a 10G needle was 25.6% compared to 27.1% for lesions sampled with a 14G needle (Table 6).

Discussion
In this study, the overall upgrade rate to malignancy for B3 lesions was 26.4% for BreastScreen NSW, Australia.This is consistent with the upgrade rates reported in the literature which ranged from 10 to 35% [ [6,[11][12][13][14]]. The majority of the B3 lesions were upgraded to DCIS (69.2%) instead of invasive cancers (29.6%) and a much smaller proportion to malignant phyllodes (1.0%), similar to results in the literature [ [11,12,14]].
The upgrade rate to malignancy varied with lesion type.Atypical papillary lesions had the highest upgrade rates with almost 1 in 2 such lesions being upgraded to malignancy.In particular, lesions with atypia on core biopsy had a significantly higher upgrade rate of 2.5 times that of lesions without atypia on core biopsy.On the other end of the spectrum, papilloma and radial scar had lower upgrade rates of 18.8% and

Table 1
Types of B3 lesions and documented risk of malignancy [ [6,12]].Some B3 lesions have limited documentation regarding malignancy upgrade rates.ADH (atypical ductal hyperplasia), ALH (atypical lobular hyperplasia), LCIS (lobular carcinoma in situ), CSL (complex sclerosing lesion).B3 Lesion Cohort.10.8% respectively.Forester's meta-analysis of B3 lesions also found atypical papillary lesions to have upgrade rates of 23-41%.This metaanalysis compared the upgrade rates of papillomas and radial scars with and without atypia, reporting rates of 32% and 7% for papillomas with and without atypia respectively and 18% and 6% for radial scars with and without atypia respectively [6].Some studies have suggested subdividing B3 lesions into those with atypia and those without given the difference in upgrade rates to malignancy between lesions with and without atypia on needle biopsy.B3 lesions with features of atypia are advised to be excised whilst those without could be managed via close surveillance and avoid excision in the first instance [ [8,13]].
The second International Consensus Conference on lesions of uncertain malignant potential in the breast in 2018 recommended that certain subtypes of B3 lesions should be excised, such as ADH and atypical lobular hyperplasia (ALH)/lobular carcinoma in situ (LCIS), while other B3 lesions could be managed by vacuum-assisted biopsy using large bore needles and more frequent surveillance [15].The second International Consensus Conference also agreed that surveillance, instead of excision, is acceptable if the upgrade rate to invasive cancer was below 5% and to DCIS was below 10% [15].In our study, only fibroadenoma/benign phyllodes and radial scar met this criterion for surveillance.Collating data for all B3 lesions without features of atypia, the upgrade rate to invasive cancer and DCIS were 4.5% and 9.1% respectively.In this study, B3 lesions without features of atypia collectively therefore met the criteria for surveillance.B3 lesions with features

Table 4
Upgrade rate to malignancy according to core biopsy pathology.Upgrade rates to malignancy also shown collectively for lesions with and without atypia on core biopsy.*Lesions with atypia on core biopsy were significantly more likely to be malignant on excision compared to lesions without atypia on core biopsy p < 0.0001.ADH (atypical ductal hyperplasia), ALH (atypical lobular hyperplasia), LCIS (lobular carcinoma in situ), CSL (complex sclerosing lesion). of atypia collectively showed upgrade rates of 10.5% and 24.2% to invasive cancer and DCIS respectively which would necessitate excision (Table 4).
In the third International Consensus Conference on lesions of uncertain malignant potential, it was conceded that ADH are lesions that cannot be confidently differentiated from low grade DCIS on needle biopsy specimens [16].As published in the European guidelines for quality assurance in breast cancer screening 4th edition pathology supplement, the use of the term atypical intraepithelial proliferation of ductal type was recommended instead of ADH because the extent of disease cannot be adequately assessed on core biopsy specimens [17].Consequently, the recommendation from the third International Consensus Conference on lesions of uncertain malignant potential was for open excision of lesions diagnosed as ADH on needle biopsy.
Similarly, the third International Consensus Conference recommended open excision of lesions diagnosed as phyllodes tumour on needle biopsy whilst lobular neoplasia, papillary lesions and radial scars were recommended for vacuum-assisted excision [16].
Increasingly, there are differences in the management of B3 lesions based on the size of the core biopsy needle.We found an upgrade rate to malignancy for lesions sampled with a 10G needle of 25.6% which is slightly lower, but not statistically different, than the upgrade rate for lesions sampled with a 14G needle (27.1%) (Table 6).Previous studies have similarly found the upgrade rate for smaller needles to be higher than the larger needles which is logical as larger needles have a smaller chance of sampling errors [ [5,13,18,19]].Aside from ADH and lobular neoplasia, the second International Consensus Conference on B3 lesions tended to recommend surveillance following core biopsy with larger needles 7-11G compared to needles 14G or smaller [15].In the United Kingdom, first introduced in Leeds, some centres are using "second-line vacuum assisted core biopsy" to obtain up to 3.6g of tissue in appropriate cases after multidisciplinary team discussion, allowing some patients to avoid surgery.Patients' whose second vacuum assisted core biopsy is benign are returned to routine rescreening.Those whose second biopsy return as atypical will either have annual surveillance for 5 years or undergo excision [ [13,20]].The Leeds B3 management pathway aims to avoid surgery in patients with lesions with lower upgrade rates.
Our study has limitations particularly as the study is a retrospective study using routinely collected data.Flat epithelial atypia, a subtype of B3 lesions, was not collected as a separate data field and was included in other atypical lesions instead.It was also not possible to determine whether factors such as family history and symptoms contributed to the upgrade rates of B3 lesions, as this information was not collected by the database.In addition, review of pathology results is not possible as the data provided is de-identified.This study however does have the strength of including all B3 lesions diagnosed by needle biopsy by the BreastScreen services in NSW, Australia over a 9-year period resulting in a large sample size.The results of this study are generalisable to other states in Australia and other countries with similar breast screening services such as the United Kingdom.
B3 lesions demonstrating atypia on needle biopsy, which include ADH and atypical papillary lesions should be excised.Ideally, other B3 lesions diagnosed on needle biopsy could be excised using vacuumassisted excision, which would reduce the rate of unnecessary surgery.Unfortunately, current constraints with funding and availability of vacuum-assisted excision have not made this procedure a viable, widely accessible alternative in Australia.

Conclusions
In this study, the overall upgrade rate of B3 lesions to malignancy was 26.4%.The majority of the lesions were upgraded to DCIS instead of invasive cancer.Upgrade rates varied by lesion type and was highest for atypical papillary lesions and lowest for benign phyllodes tumours.Lesions with atypia had significantly higher upgrade rates to cancer  compared to lesions without atypia.Malignant lesions were significantly larger than benign lesions.Therefore, lesions with atypia on core biopsy should be excised while lesions without atypia should be reviewed by a multidisciplinary team with particular consideration given to excising larger lesions without atypia.

Table 5
Excision pathology based on lesion nature on mammogram

Table 7
Comparison of cohorts with malignant, atypical and benign pathology on excision.