Part III · Pathology and classification · Chapter 9

Histopathology and the spectrum of breast neoplasia

What the slide still tells you after the profile has been run.

1 · Benign proliferative lesions and atypia

Benign breast lesions are sorted by the risk they carry, and the sorting is older than any molecular data. Non-proliferative changes carry little or no excess risk. Proliferative lesions without atypia carry a modest excess. Atypical hyperplasia, ductal or lobular, carries a substantial excess and is managed as a high-risk state rather than as a benign finding¹.

Atypical lobular hyperplasia is the lesion that broke the risk-marker framework, and the data are worth stating precisely. In a retrospective cohort of 252 women whose benign biopsies showed atypical lobular hyperplasia, 50 developed invasive breast cancer, a relative risk of 3.1 against expectation². Of those 50 cancers, 34 arose in the biopsied breast and 12 in the opposite breast.

That laterality is the point. A pure risk marker predicts cancer in either breast equally. A pure precursor predicts cancer only where it sits. Atypical lobular hyperplasia behaves as something between the two, which means neither label is adequate.

Two practical consequences follow. Atypia found on a core needle biopsy is a sampling result rather than a diagnosis, because the lesion that was not sampled may be worse. And a diagnosis of atypia is a statement about future risk in a person, which makes it a communication problem as much as a pathological one. Management sits in Atypia and in situ disease, and the biology of precursor states in Breast carcinogenesis and precursor states.

2 · Ductal carcinoma in situ

Ductal carcinoma in situ is defined by what is absent. Malignant epithelial cells fill ductal spaces and the myoepithelial layer and basement membrane remain intact. Everything difficult about the entity follows from the fact that the definition is negative.

Grading uses nuclear grade, the presence of comedo necrosis and architectural pattern. These are reproducible enough to report and not reproducible enough to carry a treatment threshold on their own, which is the same tension described for invasive grading in Invasive carcinoma of no special type and histological grading.

The population data set the terms of the argument. Among 108196 women diagnosed with ductal carcinoma in situ in a national registry, breast cancer specific mortality at 20 years was 3.3%³. It was higher in women diagnosed before age 35, at 7.8% against 3.2% in older women. Among women treated with lumpectomy, radiotherapy reduced the 10-year risk of ipsilateral invasive recurrence from 4.9% to 2.5%, and did not reduce breast cancer mortality at 10 years, which was 0.8% against 0.9%.

Which lesions progress remains unsolved, and the surveillance trials COMET, LORIS and LORD were designed around exactly that gap. They are discussed in Atypia and in situ disease.

3 · Lobular neoplasia and lobular carcinoma in situ

Lobular neoplasia covers atypical lobular hyperplasia and lobular carcinoma in situ, separated by how much of the acinar unit is filled and distended. The separation is quantitative and the underlying lesion is the same.

The molecular evidence settled a long argument about whether these are markers or precursors. E-cadherin expression is absent in lobular carcinoma in situ, including in cases with no adjacent invasive component, while it is retained in ductal carcinoma in situ⁴. The same truncating mutations and loss of the wild-type allele were found in the in situ component and in the adjacent invasive lobular carcinoma. Loss of E-cadherin is therefore an early event rather than a consequence of invasion.

So lobular carcinoma in situ is a precursor in the direct genetic sense, and it is also a bilateral risk marker in the epidemiologic sense. Those two facts are both true and they support different clinical policies. The resolution is in Lobular carcinoma in situ as risk marker and as precursor.

Two caveats belong with the E-cadherin story. Loss of membranous E-cadherin is not exclusive to lobular lesions, and reduced expression occurs in non-lobular invasive carcinoma where it carries independent prognostic weight⁵. And pleomorphic lobular carcinoma in situ, with high grade nuclei and frequent necrosis, is managed more aggressively than classical disease on morphological grounds rather than on outcome data.

4 · Invasive carcinoma of no special type and histological grading

Carcinoma of no special type is a residual category. It is what remains after the special types in Special histologic types and their disproportionate prognostic weight have been excluded, and it makes up the majority of invasive breast cancer. A residual category cannot be expected to be biologically coherent, and this one is not.

Grading is what rescues it. The Nottingham system scores tubule formation, nuclear pleomorphism and mitotic count separately and sums the three⁶. The design decision that matters is that each component was made semiquantitative in order to be reproducible, at some cost in biological fidelity. The same trade is made at every threshold in Receptor assessment and the measurement problem.

Grade retains independent prognostic value alongside molecular assays rather than being replaced by them⁷. Its weakness is entirely in the middle. Grade 2 is a large group with intermediate risk, and intermediate risk does not support a decision.

Expression profiling addressed that directly. A 97-gene index was derived by comparing grade 3 with grade 1 oestrogen receptor positive tumours⁸. Applied to grade 2 tumours, it split them into two groups, with a hazard ratio for recurrence of 3.61 for the high index group against the low. Most of the genes were proliferation related.

Read that as a statement about the assay rather than about the tumours. Grade 2 was never an intermediate biology. It was the place where a three-tier reading of continuous features put the cases it could not separate.

Mitotic count is the component most sensitive to handling, because it depends on field area, section thickness and delay before fixation Preanalytics, fixation, and the invisible sources of error.

5 · Invasive lobular carcinoma as a biologically distinct disease

Invasive lobular carcinoma is the strongest case in this chapter for morphology as biology. It is distinct at the level of the driver, the growth pattern, the imaging, the operation and the site of relapse.

The driver is loss of E-cadherin function, and the molecular characterization confirms it as the defining event. Profiling of 817 breast tumours, including 127 lobular and 490 ductal carcinomas, identified enrichment of PTEN, TBX3 and FOXA1 alterations in lobular disease, with the highest AKT phosphorylation of any breast cancer subtype⁹. Mixed ductal and lobular cases classified as either lobular-like or ductal-like, with no genuinely hybrid group.

The growth pattern follows from the driver. Cells that cannot form stable adherens junctions infiltrate singly and in single files through stroma, without forming a cohesive mass and often without a desmoplastic reaction. That is a mechanical explanation for under-detection on mammography. A lesion that displaces little tissue and produces little density is hard to see, and it is routinely larger at diagnosis than its imaging suggested.

The clinical profile matches. In a comparison of 4140 lobular with 45169 ductal carcinomas, lobular tumours were larger, more often receptor positive and more often diploid¹⁰. They had a lower S-phase fraction, and were less often HER2, p53 or epidermal growth factor receptor positive. Metastasis to the gastrointestinal tract and ovary was more common. Contralateral breast cancer occurred in 20.9% against 11.2%.

The outcome result is the one that surprises people. Despite a favourable biological phenotype, five-year disease-free and overall survival did not differ significantly between the two groups in that series. Metastatic management is taken up in Invasive lobular disease in the metastatic setting.

6 · Special histologic types and their disproportionate prognostic weight

Special types are defined by requiring a stated proportion of the tumour to show a characteristic pattern. They account for up to 25% of invasive breast cancers, and the transcriptomic taxonomy in Clinical and molecular classification was derived almost entirely without them¹¹.

That omission matters more than the frequencies suggest, because several special types show tight genotype and phenotype correlation. Secretory carcinoma carries the ETV6-NTRK3 fusion. Adenoid cystic carcinoma carries the MYB-NFIB fusion. Lobular carcinoma carries CDH1 inactivation through several mechanisms¹¹. These are one-to-one relationships between an appearance and a genetic event, which is exactly what the no-special-type category lacks.

Adenoid cystic carcinoma shows why this is not academic. It is triple-negative on receptor testing and assigns to the basal-like group by expression profiling, and it does not behave like ordinary triple-negative disease. A classifier that places it there is reporting an expression pattern rather than a prediction of behaviour. The histological name carries the clinical information.

The fusion-defined types also carry an actionable consequence. TRK fusions are targetable across histologies, and larotrectinib produced durable responses in TRK fusion-positive cancers of varied origin¹². A rare morphology therefore functions as a cheap screening test for a targetable event.

7 · Metaplastic, apocrine, and other uncommon phenotypes

Metaplastic carcinoma contains non-glandular elements, which may be squamous, spindle cell, chondroid or osseous. It is usually triple-negative on receptor testing and behaves worse than its receptor status alone predicts.

The internal diversity is real rather than a sampling impression. Genomic and transcriptomic analysis of metaplastic carcinomas found substantial heterogeneity between cases and by metaplastic element, which is why single-agent strategies aimed at the whole category have performed poorly¹³.

Apocrine and molecular apocrine tumours are the opposite problem. Here the morphological name and the molecular name point at the same biology from two directions. Abundant eosinophilic cytoplasm with prominent nucleoli corresponds to androgen receptor driven signalling, and to the luminal androgen receptor group within triple-negative disease¹⁴. The mapping between morphology and class is imperfect, and the transcriptional classes themselves disagree with one another Why TNBC classifiers disagree with one another.

Two rules follow for practice. A metaplastic or apocrine label should prompt a check that the receptor panel was interpreted in the light of the morphology, because androgen receptor is not on the standard panel. And a rare phenotype in a small biopsy is a reason to look at more tissue, since these patterns are frequently focal within an otherwise ordinary carcinoma. The therapeutic implications sit in Triple-negative and other biologically diverse phenotypes and Androgen receptor driven and luminal androgen receptor disease.

8 · Inflammatory breast cancer and Paget disease

Inflammatory breast cancer is a clinical diagnosis. The defining features are rapid onset of erythema and oedema involving at least a third of the breast, with a short duration of symptoms, and pathological confirmation of invasive carcinoma¹⁵. Dermal lymphatic tumour emboli support the diagnosis. Their absence does not exclude it.

That is an unusual arrangement and it has consequences. A biopsy cannot rule the diagnosis in or out, because the finding that would confirm it is not required and is frequently missed by a skin punch. Underdiagnosis follows directly from the sampling, not from disagreement about the definition. Because the condition is defined clinically, the recorded incidence depends on who examined the patient and how quickly.

Inflammatory disease is staged as T4d and is locally advanced by definition. Multimodality treatment with systemic therapy first, then surgery, then radiation, is the standard sequence¹⁵. The biology and management are developed in Locally advanced and inflammatory breast cancer and Inflammatory breast cancer biology.

Paget disease of the nipple is a different problem with a similar lesson. It is an intraepidermal adenocarcinoma of the nipple and areola, and it is, as a rule, associated with an underlying breast carcinoma¹⁶. The visible lesion is therefore a sign rather than the disease. A nipple that does not resolve on topical treatment requires imaging and full-thickness biopsy, because the relevant carcinoma is usually somewhere else in the breast.

9 · Phyllodes tumors, mesenchymal lesions, and diagnostic mimics

Phyllodes tumours are fibroepithelial neoplasms graded benign, borderline or malignant on stromal cellularity, atypia, mitotic activity, stromal overgrowth and the character of the border. The World Health Organization classification acknowledges that prediction of behaviour in an individual case remains difficult¹⁷. Where the distinction from benign phyllodes tumour is uncertain, it recommends favouring a diagnosis of fibroadenoma, or the neutral term benign fibroepithelial neoplasm.

That recommendation is a statement about reproducibility rather than about biology. The clinical behaviour of fibroadenoma and benign phyllodes tumour overlaps, so the cost of a hedged diagnosis is low and the cost of overcalling is a wider excision.

Two practical points about malignant phyllodes tumours follow from their stromal nature. Spread is haematogenous, so axillary staging is not indicated. And the malignant component is sarcomatous, so systemic therapy designed for carcinoma does not apply.

Mimics are the other half of this section and they run in both directions. Sclerosing adenosis, radial scar and microglandular adenosis can imitate invasive carcinoma at low power. Myoepithelial markers resolve most of these, because the preserved myoepithelial layer is the discriminator, and microglandular adenosis is the exception that lacks one. In the opposite direction, the same classification moved malignant myoepithelioma into metaplastic carcinoma, which changed what the diagnosis implies for treatment¹⁷.

10 · Pathological assessment after neoadjuvant treatment

Assessment after neoadjuvant treatment is a different task from assessment of an untreated specimen. The pathologist is measuring what a treatment failed to remove, in a bed that has been altered by the treatment.

The binary endpoint came first. In a pooled analysis of 12 neoadjuvant trials in 11,955 patients, eradication of tumour from both breast and nodes carried an event-free survival hazard ratio of 0.48, with a 95% confidence interval of 0.43 to 0.54. Eradication from the breast alone carried 0.60, with an interval of 0.55 to 0.66¹⁸. That comparison is why the accepted definition of pathologic complete response requires the nodes to be clear. A response report that does not state the nodal status is not reporting the quantity that carries the prognosis.

A binary endpoint discards most of the available information, and residual cancer burden recovers it. Residual cancer burden combines the two largest dimensions of the residual tumour bed, the cellularity within it, and the number and size of involved nodes into a continuous index. It was developed in 382 patients across two treatment cohorts¹⁹. It remained independently prognostic for distant relapse-free survival after adjustment for age, clinical stage, receptor status, endocrine therapy and pathologic response. Minimal residual disease, in 17% of patients, carried the same prognosis as complete response. Extensive residual disease, in 13%, carried a poor prognosis regardless of receptor status. Long-term outcome by residual burden category has since been confirmed in pooled analysis²⁰.

The use of this measurement to direct further treatment sits in The neoadjuvant setting as a scientific platform, and the trials built on it, including KATHERINE and CREATE-X, in Post-neoadjuvant intensification for residual disease and Capecitabine for residual disease.

11 · What morphology still predicts that molecular profiling does not

The honest summary is short, and it is not a defence of tradition.

Special histological type carries genotype information that no bulk expression classifier recovers, because the classifiers were trained without those cases¹¹. A one-to-one relationship between an appearance and a fusion gene is a strong prediction, and morphology is currently the cheapest way to make it.

Lobular growth pattern predicts things that molecular class does not address at all. It predicts how the tumour will look on imaging, how hard the margin will be to judge, and which organs will be involved at relapse^9,10.

Grade retains independent prognostic weight in receptor positive disease, and a proliferation-based index refines it rather than replacing it^7,8.

Residual burden after treatment is a morphological measurement with no molecular equivalent^19,20. Nothing sequenced from a residual specimen predicts outcome as well as measuring how much tumour is left.

And morphology governs the validity of every profile run on the same block. A profile of tissue that is mostly stroma is a profile of stroma. The pathologist who states tumour cellularity is supplying the denominator for the molecular report Preanalytics, fixation, and the invisible sources of error.

What morphology does not do should be stated with equal force. It does not identify pathway dependence. It does not predict response to endocrine therapy, to HER2-directed treatment or to immune checkpoint blockade. A grade 3 triple-negative carcinoma and a grade 3 HER2 positive carcinoma can look alike and require entirely different treatment. That is why Clinical and molecular classification and Receptor assessment and the measurement problem exist, and why this chapter claims complementarity rather than priority.

References

All 20 references

1. Hartmann LC, Degnim AC, Santen RJ, Dupont WD, Ghosh K. Atypical hyperplasia of the breast, risk assessment and management options. N Engl J Med 2015 372:78-89. PMID 25551530
2. Page DL, Schuyler PA, Dupont WD, Jensen RA, Plummer WD, Simpson JF. Atypical lobular hyperplasia as a unilateral predictor of breast cancer risk: a retrospective cohort study. Lancet 2003 361:125-129. PMID 12531579
3. Narod SA, Iqbal J, Giannakeas V, Sopik V, Sun P. Breast cancer mortality after a diagnosis of ductal carcinoma in situ. JAMA Oncol 2015 1:888-896. PMID 26291673
4. Vos CB, Cleton-Jansen AM, Berx G, et al. E-cadherin inactivation in lobular carcinoma in situ of the breast: an early event in tumorigenesis. Br J Cancer 1997 76:1131-1133. PMID 9365159
5. Rakha EA, Abd El Rehim D, Pinder SE, Lewis SA, Ellis IO. E-cadherin expression in invasive non-lobular carcinoma of the breast and its prognostic significance. Histopathology 2005 46:685-693. PMID 15910600
6. Elston CW, Ellis IO. Pathological prognostic factors in breast cancer. I. The value of histological grade in breast cancer: experience from a large study with long-term follow-up. Histopathology 1991 19:403-410. PMID 1757079
7. Rakha EA, Reis-Filho JS, Baehner F, et al. Breast cancer prognostic classification in the molecular era: the role of histological grade. Breast Cancer Res 2010 12:207. PMID 20804570
8. Sotiriou C, Wirapati P, Loi S, et al. Gene expression profiling in breast cancer: understanding the molecular basis of histologic grade to improve prognosis. J Natl Cancer Inst 2006 98:262-272. PMID 16478745
9. Ciriello G, Gatza ML, Beck AH, et al. Comprehensive molecular portraits of invasive lobular breast cancer. Cell 2015 163:506-519. PMID 26451490
10. Arpino G, Bardou VJ, Clark GM, Elledge RM. Infiltrating lobular carcinoma of the breast: tumor characteristics and clinical outcome. Breast Cancer Res 2004 6:R149-R156. PMID 15084238
11. Weigelt B, Geyer FC, Reis-Filho JS. Histological types of breast cancer: how special are they?. Mol Oncol 2010 4:192-208. PMID 20452298
12. Drilon A, Laetsch TW, Kummar S, et al. Efficacy of larotrectinib in TRK fusion-positive cancers in adults and children. N Engl J Med 2018 378:731-739. PMID 29466156
13. Piscuoglio S, Ng CKY, Geyer FC, et al. Genomic and transcriptomic heterogeneity in metaplastic carcinomas of the breast. NPJ Breast Cancer 2017 3:48. PMID 29214215
14. Lehmann BD, Bauer JA, Chen X, et al. Identification of human triple-negative breast cancer subtypes and preclinical models for selection of targeted therapies. J Clin Invest 2011 121:2750-2767. PMID 21633166
15. Dawood S, Merajver SD, Viens P, et al. International expert panel on inflammatory breast cancer: consensus statement for standardized diagnosis and treatment. Ann Oncol 2011 22:515-523. PMID 20603440
16. Kanitakis J. Mammary and extramammary Paget's disease. J Eur Acad Dermatol Venereol 2007 21:581-590. PMID 17447970
17. Tan PH, Ellis IO. Myoepithelial and epithelial-myoepithelial, mesenchymal and fibroepithelial breast lesions: updates from the WHO Classification of Tumours of the Breast 2012. J Clin Pathol 2013 66:465-470. PMID 23533258
18. Cortazar P, Zhang L, Untch M, et al. Pathological complete response and long-term clinical benefit in breast cancer: the CTNeoBC pooled analysis. Lancet 2014 384:164-172. PMID 24529560
19. Symmans WF, Peintinger F, Hatzis C, et al. Measurement of residual breast cancer burden to predict survival after neoadjuvant chemotherapy. J Clin Oncol 2007 25:4414-4422. PMID 17785706
20. Yau C, Osdoit M, van der Noordaa M, et al. Residual cancer burden after neoadjuvant chemotherapy and long-term survival outcomes in breast cancer: a multicentre pooled analysis of 5161 patients. Lancet Oncol 2022 23:149-160. PMID 34902335