'Borderline' lesions The language used in medicine is fascinating and it is interesting to observe how insidiously words such as 'borderline' creep into our vocabulary without due thought to their impact on clinical practice. It is difficult to establish when the term was first applied to breast lesions but it appears to have been in use for at least 25 years to encompass those lesions characterised by epithelial proliferation. The main problem with the term is that it implies a sharp division between entities and this is usually interpreted as providing a separation between benign and malignant lesions with a consequent effect on clinical management protocols. In reality such sharp distinctions rarely occur and certainly where breast epithelial lesions are concerned we are dealing with a spectrum of changes. This has been emphasised in recent years by molecular genetic studies which have shown loss of heterozygosity in usual epithelial hyperplasia, atypical ductal hyperplasia and ductal carcinoma in situ (DCIS); this data implies that even usual epithelial hyperplasia, in some cases, is a clonal, and therefore neoplastic, process. 19-22 Nevertheless, from the practical viewpoint stratification is required so that individual patients receive appropriate treatment. The entities concerned are the epithelial hyperplasias of usual and atypical subtypes, in situ carcinoma and so-called 'minimal' invasive carcinoma. In my view it is more appropriate to regard these lesions as occupying a borderland between entities which are definitely benign and those which are unequivocally malignant. Some years ago we devised the scheme outlined in Table 4 to indicate the gradual transition from risk factor (for subsequent invasive carcinoma) to established malignancy. This concept now needs to be re-evaluated for two main reasons; there are definite problems in reproducibility of diagnosis of such entities as atypical ductal hyperplasia (ADH) and more appropriate cut-off points are required to stratify patients for clinical and therapeutic management. The debate concerning the utility of the term atypical ductal hyperplasia has continued for the past 20 years. Page and colleagues have promoted the view that epithelial proliferations which closely resemble DCIS of low grade type, but which do not exhibit all the precise morphological characteristics to make that diagnosis with confidence should be designated as ADH. 23, 24 Although this concept has received wide acceptance doubts remain. Some, 1, 3, 25 but not all, 26 reproducibility studies have shown poor agreement for the entity suggesting that there is a lack of agreement on the diagnostic criteria. It is therefore worth emphasising some key distinguishing features, as set out below: a DCIS of low grade type has a population of evenly spaced, uniform cells with uniformly oval to rounded nuclear features, comprising without doubt the entire population of cells throughout at least two membrane-bound spaces (a measure of the extent of the lesion). Secondary spaces have smooth rounded 'punched out' borders of geometrical structure (cribriform architecture - cribrum - sieve - Latin) and rigid non-tapering bars can be found. Micropapillary structures are bulbous and regularly placed around the space. NB Epithelial proliferations of high grade cytology, with or without comedo necrosis qualify as DCIS, regardless of lesional size. b ADH exhibits partial involvement of the basement membrane-bound space by a cell population of the type defined above for DCIS of low grade type. A second cell population is usually present, consisting of columnar, polarised cells of the type seen in the ductal lamina positions immediately above the basement membrane. There is general agreement that lesions with these features which measure more than 2-3 mm in diameter are more likely to represent DCIS than ADH. ADH is, therefore, almost without exception, a tiny lesion. In summary, we use the following rules of thumb. 1 We do not entertain a diagnosis of ADH unless low grade DCIS has been considered seriously in the differential diagnosis. 2 ADH is a very small, microfocal, process; the larger the lesion the more likely it is to be DCIS. 3 It follows that if there is any doubt about the differential diagnosis further levels and blocks should be examined. In our experience this usually resolves the problem, almost always by revealing evidence of low grade DCIS. 4 If changes resembling ADH are found in a needle core biopsy they should be regarded as either indicative of uncertain malignant potential (B3) or suspicious of malignancy (B4). Excision biopsy is mandatory and usually reveals unequivocal low grade DCIS. 10, 12, 27 In recent years there has been a proliferation of classifications for ductal carcinoma in situ. 7, 28-30 Despite differences in terminology all identify an aggressive comedo, poorly differentiated, high grade subtype and a relatively indolent subtype - non-comedo, well differentiated, low grade. In the United Kingdom it has been recommended that to avoid problems with the definitions of comedo necrosis and architectural pattern a simpler nuclear grading system is adopted for use in the National Health Service Breast Screening Programme (NHSBSP), and the same classification has been accepted by the European Commission Working Party on Breast Screening Pathology. 6, 7 However, problems with reproducibility still exist and only average agreement has been obtained in studies to date. We are still left with the dilemma of correlating the morphology of these epithelial proliferative lesions and their management. Table 5 provides a somewhat speculative personal attempt to achieve this. Epithelial hyperplasia of usual type, with its low relative risk factor of x2, needs no treatment or follow-up. The size rule should still be applied to define atypical ductal hyperplasia, but, in my view, both ADH and low grade DCIS measuring less than 3 mm can be managed in the same way, by follow-up alone. There is no doubt that the prognosis for all types of ductal carcinoma in situ is excellent, provided that complete local excision is achieved. The same is true for so-called 'microinvasive carcinoma' (which will be referred to further below) and small invasive carcinomas as defined in Table 5. Note that the small invasive tumour size in Table 4 has been raised to 15 mm from 10 mm because no difference in prognosis between the two was found in the Nottingham Tenovus Primary Breast Cancer Study.4 This group of patients can therefore be treated by complete local excision, with or without axillary dissection. Local irradiation may be appropriate, but systemic therapy is not indicated. Experience from both anecdotal consultation cases and early data from the UK NHSBSP indicate that microinvasion in DCIS is over-diagnosed by many general histopathologists. Because of this perception and the lack of consensus throughout the world on diagnostic criteria the definition recommended for use in the NHSBSP is very restrictive. 7 The lesion is composed predominantly of DCIS (almost always of high grade type), but with foci of definite invasion outside the specialised lobular stroma measuring not more than 1 mm in maximum diameter. There is no limit on the number of such foci. Care must be taken not to confuse as microinvasion extension of DCIS into lobules, branching of ducts containing DCIS, tangential cutting, crush and cautery artefact. Particular difficulty may be encountered in cases with marked periductal fibrosis and inflammation. In general terms the diagnosis of microinvasion should not be made unless there is convincing invasive carcinoma extending into interlobular stroma or adipose tissue. Multiple levels should be examined in suspicious cases and we have found in our laboratory that immunostaining for anti-smooth muscle actin may be helpful in distinguishing in situ from invasive elements. Immunostaining for basement membrane components (type IV collagen, laminin) is of less value. Pathologists are taught in general that when in doubt over a 'borderline' diagnosis to give the more malignant interpretation in order to avoid under treatment. It should always be borne in mind when considering a diagnosis of microinvasive carcinoma that some surgeons will carry out an axillary dissection or clearance in such cases but not, of course, for DCIS, and overdiagnosis therefore leads to unnecessary surgery. Furthermore, there is no evidence to date which has demonstrated that microinvasive carcinoma has a different prognosis from that of high grade DCIS. In my view, therefore, the diagnosis should only be made in those rare cases with unequivocal evidence of invasion. |
Copyright 2000, The Author(s) and/or The Publisher(s)
| Organisation: FORPATH asbl  |
Coordination: Dr Bernard Van den Heule |
Host: Labo CMP |