Article Text
Abstract
Aims A prominent fibrovascular stromal core is one of the widely accepted histological features of breast papillomas, but some papillary carcinomas also show such broad fibrovascular cores, leading to confusion in diagnosis, particularly in needle biopsy specimens. We investigated the histological characteristics of papillary lesions, focusing on broad fibrovascular cores and their relationship with the architectural patterns.
Methods Among 185 cases of needle biopsies of papillomas and papillary carcinomas, the number of cases with broad fibrovascular cores in each group was compared. The broad fibrovascular core density in the subsequently resected specimens was evaluated and compared between papillary predominant pattern (papillary structures >80% of tumours) and mixed pattern (papillary, solid, cribriform and others) within the lesions.
Results Significantly more papillary carcinomas than papillomas and B3 atypical papillary lesions had broad fibrovascular cores (p=0.0091 and p=0.0164, respectively). The papillary predominant pattern was more prominent in carcinomas than in papillomas in the needle biopsies (p=0.048) and showed the same tendency in the resections (p=0.058). The broad fibrovascular core density was significantly lower in the 18 papillomas than in the 37 papillary carcinomas (p=0.0079) and was not significantly different between the papillary predominant and mixed patterns in carcinomas and papillomas.
Conclusions Broad fibrovascular cores in mammary papillary lesions are not specific for papillomas, as they are also present focally in papillary carcinomas. As the frequency of papillary carcinoma with broad fibrovascular cores is relatively high, caution in diagnosis has to be exercised, especially in needle biopsy specimens.
- BREAST
- BREAST CANCER
- BREAST PATHOLOGY
- CANCER
- MORPHOLOGY
Statistics from Altmetric.com
Introduction
One of the most challenging diagnosis in assessing papillary lesions of the breast is distinguishing a papilloma from papillary carcinoma. Papillary lesions are characterised by finger-like projections or fronds composed of fibrovascular cores covered by epithelium.1 The papillary fronds of benign papillomas are covered by an inner myoepithelial cell layer and an outer epithelial layer.1 One of the important differentiations between papillomas and papillary carcinomas is that the stroma is ‘prominent’ with fibrosis and epithelial entrapment in papillomas but is ‘delicate’ in papillary carcinomas.1 ,2 Solid papillary carcinomas are also considered variants of papillary carcinomas as histologically circumscribed solid nodules. Discrete papillae are not apparent; the underlying papillary structure is represented by a network of blunt or broad fibrovascular cores among the solid cellular proliferation.1 These stromal fibrovascular cores are often embedded within the solid epithelial proliferation and are extensively hyalinized.3 In other papillary carcinomas, secondary changes of sclerosis may also result in broad fibrovascular cores. This may be one of the reasons for the difficulty in distinguishing between papillomas and papillary carcinomas, especially in needle biopsies.4 The usefulness of high-molecular weight cytokeratins (HMWCK) for distinguishing papillary carcinomas from papillomas was recently reported.5 ,6 However, the use of HMWCK in the ‘papillary’ area of papillary lesions is less dependable.6 Thus, in papillary lesions with mostly ‘papillary’ areas with superimposed proliferation of atypical epithelial cells with prominent stroma, it would be difficult to differentiate between a papilloma, papilloma with atypical ductal hyperplasia (ADH)/ductal carcinoma in situ (DCIS) and papillary DCIS, especially in needle biopsies (figure 1). Although needle biopsies are generally accurate for diagnosing papillary lesions, the management of non-malignant papillary lesions diagnosed at needle biopsies is still debated.7 Few studies have focused on the stroma in papillary lesions, and it is timely to review the findings in this area, particularly in cases with predominately papillary pattern and broad fibrovascular cores, where the diagnostic role of HMWCK was less dependable. Therefore, using adequate number of needle biopsy materials, we investigated the diagnostic issues related to papillary lesions with an emphasis on broad fibrovascular cores.
Materials and methods
The materials were obtained from patients at the Social Insurance Kurume Daiichi Hospital (Kurume, Japan) over a period of 5 years from 2008 to 2012. All patients were female.
Definitions
A broad fibrovascular core was defined as stromal tissue core with a minor (cross-sectional) axis >100 μm and measuring >1×104 μm2 in size (area) (figure 2). Papillomas were intraductal benign papillary lesions characterised by finger-like fibrovasucular cores covered by epithelial and myoepithelial layers.8 Ductal adenomas were well-circumscribed sclerosing papilloma, with benign glandular proliferation located, at least in part, within a duct lumen, usually accompanied by dense scar-like fibrosis and surrounded by a thick fibrous wall.8 ,9 Atypical papillary lesions corresponded to B3 lesion of uncertain malignant potential by the National Health Service Breast Screening Programme,10 that is, showing portions manifesting architectural and cytological atypia that fell short of the diagnosis of intraductal (encapsulated) papillary carcinomas; or small focus (<3 mm) of ADH within a background of papillomas in the present study. Papillary carcinomas included intraductal papillary carcinomas (papillary DCIS), encapsulated papillary carcinomas and solid papillary carcinomas.8 All the papillary lesions (papillomas and carcinomas) in this study were divided into predominantly papillary or mixed patterns (mixed with solid, cribriform and tubular structures). The papillary predominant pattern was defined as having predominantly (>80%) papillary structure in the specimens (figure 3A–D), whereas all others were defined as mixed pattern (figure 3E–H).
Broad fibrovascular cores in papillary lesions in needle biopsy specimens
The patient cohort consisted of cases with core needle biopsies (CNB) using a 14-gauge needle and Mammotome (MMT) biopsies using an 11-gauge vacuum-assisted instrument. All MMT and CNB biopsies were done under ultrasound guidance. When the broad fibrovascular cores were identified by low magnification, the minor axis and size were manually confirmed using an eyepiece micrometer (square, 10 mm/10 units, Olympus). The number of cases with broad fibrovascular cores was compared among papillomas (and ductal adenomas), atypical papillary lesions and papillary carcinomas. The architectural (papillary predominant or mixed) patterns were also compared among papillomas and papillary carcinomas.
Broad fibrovascular cores in resected papillary lesions
Subsequently, the broad fibrovascular cores were evaluated in the corresponding resection specimens of papillary lesions, most of which corresponded to the needle biopsies, although some were without previous needle biopsies. The densities of broad fibrovascular cores between papillomas and carcinomas within papillary predominant and mixed patterns as defined above were also compared. In each lesion, the field of highest density of broad fibrovascular cores was identified by scanning the tumour sections at low-power magnification (40×). If the whole tumour was not included in one slide, additional sections were evaluated. After the field of the highest broad fibrovascular core density was identified, the number of individual broad fibrovascular cores was manually counted at 100× (1 mm2 per field) (figure 4). A score was given equalling the number of broad fibrovascular cores counted. When the number of broad fibrovascular cores was 10 or more or when the broad fibrovascular cores occupied 1/4 or more of the microscopic field (0.25 mm2), especially if they were sclerosed and expanded (any shapes including rectangular or branching), a score of 10 was given. Thus, the score of broad fibrovascular core density would theorectically range from 0 to 10.
Statistical analyses
The number of cases with broad fibrovascular cores was compared between papillomas and papillary carcinomas using Fisher's exact p test. Comparisons of the broad fibrovascular core density within tumours were undertaken using the Mann–Whitney U test. A p value of <0.05 was considered statistically significant.
Results
Broad fibrovascular cores in papillary lesions in needle biopsy specimens
Overall, a total of 3338 needle biopsies of the breast (1210 CNB (36%) and 2128 MMT biopsies (64%)) were retrieved. In total, 185 papillary lesions were identified, including 114 (61.6%) papillomas, 21 (11.4%) ductal adenomas (sclerosed papillomas), 28 (12.8%) papillary carcinomas (9 papillary DCIS, 12 encapsulated papillary carcinomas and 7 solid papillary carcinomas) and 22 (11.9%) B3 (atypical) papillary lesions. Twenty-six (14.0%) lesions (13 papillomas, 4 ductal adenomas, 7 papillary carcinomas and 2 atypical papillary lesions) were extracted by CNB, and the remaining 159 (86.0%) lesions were extracted by MMT biopsies. These cases were reviewed and the diagnosis confirmed. The number of cores per biopsy specimen ranged from 1 to 6 (median 3). Twenty-seven of the twenty-eight papillary carcinomas (96.4%) had broad fibrovascular cores, significantly higher than in papillomas (85/114, 74.6%; p=0.0091), B3 papillary lesions (15/22, 68.2%; p=0.0164) and a combination of papillomas and ductal adenomas (106/135, 78.5%; p=0.0298). There was no significant difference in the broad fibrovascular cores seen between papillary carcinomas and ductal adenomas (21/21, 100%) (table 1). In papillary carcinoma, there were significantly more cases (9/28, 32.1%) showing papillary predominant pattern than in papillomas (16/114, 14.0%) (p=0.048).
Broad fibrovascular cores in papillary lesions in resected materials
Resection specimens of 18 papillomas and 37 papillary carcinomas (10 in 18 papillomas and 34 in 37 papillary carcinomas were subsequent excisions after biopsies and the others were excisions without prior biopsies) were examined. Among the 18 papillomas, two cases showed papillary predominant pattern. The other cases showed mixed pattern with other lesions such as adenosis and usual ductal hyperplasia. Among the 37 papillary carcinomas, 14 showed papillary predominant pattern and the other 23 showed mixed pattern with papillary, tubular, cribriform and solid patterns. Carcinomas (14/37, 37.8%) showed a higher tendency towards papillary predominant pattern than papillomas (2/18, 11.1%) (p=0.058). The broad fibrovascular core density was significantly lower in the 18 papillomas than in the 37 papillary carcinomas (OR 0.820, 95% CI 0.706% to 0.954%, p=0.0079). However, the broad fibrovascular core density was not significantly different between papillomas with papillary predominant pattern and papillomas with mixed pattern, or papillary carcinomas with papillary predominant pattern and papillary carcinomas with mixed pattern (table 2).
Discussion
The interesting finding in this study was that more papillary carcinomas than papillomas showed broad fibrovascular cores in needle core biopsies, and the results were confirmed in the resected materials by examining the broad fibrovascular core density. This was in contrast to conventional belief.
When the epithelial cells in papillary lesions show solid proliferation, HMWCK are very useful for the differential diagnosis between benignity (HMWCK-positive) and malignancy (HMWCK-negative).5 ,6 However, in tumours showing mainly papillary pattern, HMWCK may be less helpful.6 In the present study, papillary predominant pattern with >80% papillary structure occurred more frequently in papillary carcinomas than in papillomas in both needle biopsies and the resected specimens. Rosen11 classified papillary carcinomas into papillary, micropapillary or filoform, cribriform, reticular and solid appearances. Although these architectural variants were seen in carcinomas, this mixed pattern was seen more commonly in papillomas than in papillary carcinomas in the current study. Therefore, one should note that papillary predominant pattern could more often be encountered routinely in papillary carcinomas than in papillomas. This diagnostic confusion could be further aggravated by the fact that HMWCK was less helpful in the papillary area of papillary lesions.6 The presence of myoepithelial cells may sometimes help distinguish benignity from malignant papillary lesions.
Although it was reported that the results of combined assessments of fibrovascular core thickness and cytokeratin 5/6 staining differed significantly between benign and atypical/malignant papillary lesions,4 there were some exceptions as broad fibrovascular cores could be present in benign and malignant papillary lesions. In core biopsies, there are no obvious structural differences in the fibrovascular cores that can be identified based on the limited fragmentary materials (figure 1). In other words, in a core biopsy showing only broad fibrovascular cores, caution has to be exercised as this may not necessarily represent benign papillomas. Based on the current results that (i) papillary carcinomas actually showed more broad fibrovascular cores than papillomas, (ii) papillary predominant patterns were often encountered in carcinomas than in papillomas and (iii) broad fibrovascular cores in papillary lesions were seen not only in the mixed pattern but also in the papillary predominant pattern, basing the diagnosis on the number and size of the fibrovascular cores might lead to misdiagnosis of papillary lesions.
Ever since Kraus and Neubecker2 reported the histological features of papillary lesions in 1962, the presence of broad fibrovascular cores was considered a characteristic of benign papillomas and a key differentiating feature from papillary malignancies. In the current study, broad fibrovascular cores were actually often seen in papillary carcinomas. This phenomenon occurred not only in solid papillary carcinomas but also in papillary lesions showing papillary predominant structure. This is particularly problematic in the latter group as HMWCK are less dependable.6 Although the current study did not compare all these architectural variants and their prevalence in benign and malignant papillary lesions, attention should be paid on the basis of these observations, especially when making a diagnosis based on needle biopsy specimens. In addition to the fibrovascular core morphology, it is also important to determine whether or not a myoepithelial cell layer is present in the area of the epithelial–stroma surface. Further studies including immunohistochemistry are needed to clarify the association between stromal area and diagnosis.
Take home messages
-
Papillary predominant pattern with >80% papillary structure occurs more frequently in papillary carcinomas than in papillomas.
-
Papillary carcinomas show more broad fibrovascular cores than papillomas.
-
Caution is needed when distinguishing between papilloma and papillary carcinoma based on stromal views, especially in needle biopsy specimens.
Acknowledgments
We thank Ms. Akiko Tanaka for her excellent assistance.
Footnotes
-
Contributors RY is the corresponding author and had a central role in this study. MT and MY are surgeons and provided samples collected from the patients. GMT interpreted the data and advised on manuscript preparation. HT is a radiologist involved in the tissues extracted from the patients. YN participated in counting the density. MT, YN and AJ participated in collecting data. HY provided administrative support. All authors provided important contributions to the conception and design of the study, reviewed the analytical results and read and approved the final manuscript.
-
Competing interests None.
-
Ethics approval This retrospective study was approved by the Social Insurance Kurume Daiichi Hospital Ethical Committee.
-
Provenance and peer review Not commissioned; externally peer reviewed.