Part VII · Detection, diagnosis, and biomarker assessment · Chapter 38

Diagnostic evaluation and staging

Everything downstream inherits the quality of the first specimen, and stage describes only what anatomy and three assays can see.

1 · Evaluation of symptoms and imaging abnormalities

Diagnostic evaluation is a sequence of conditional probabilities, and the order of operations matters.

Triple assessment pairs clinical examination, imaging and tissue sampling because each has a different failure mode. A palpable mass with benign imaging is not excluded, because mammography and ultrasound both miss cancers, particularly in dense tissue Dense breasts, supplemental imaging, and notification policy. Imaging that is negative in the face of a suspicious clinical finding lowers the probability of cancer. It does not remove it.

The commonest diagnostic error in this setting is treating a negative test as an answer rather than as an update. A negative result on a test with 85% sensitivity leaves substantial residual probability when pretest probability was high to begin with.

Two presentations deserve named exceptions. Inflammatory breast cancer presents as diffuse skin change without a discrete mass, and imaging is frequently unhelpful, so the diagnosis is clinical and confirmed by skin and parenchymal biopsy Inflammatory breast cancer and Paget disease. Invasive lobular carcinoma grows in single files without forming a mass, so it is systematically underestimated by mammography and ultrasound Invasive lobular carcinoma as a biologically distinct disease.

The rule that follows is simple to state and often ignored. When clinical suspicion and imaging disagree, sample the tissue.

Preoperative breast magnetic resonance imaging is the test whose use outran its evidence. Reviews set out its sensitivity and its limits^1,2,3. A meta-analysis of 19 studies found additional disease in 16% of women with breast cancer, with a positive predictive value of 66%⁴. Wide excision was converted to mastectomy in 8.1%, and in 1.1% because of lesions that proved benign.

Two randomised trials then tested whether it improves surgery. COMICE randomised 1,623 women scheduled for wide local excision, and 19% needed reoperation in each arm⁵. Its economic analysis found few or no benefits in cost-effectiveness or quality of life⁶. In MONET, adding it for nonpalpable cancers increased re-excision for positive margins, 34% against 12%⁷.

The observational studies agree with the trials more than with the enthusiasm. Selective use did not reduce reoperation and often overestimated extent⁸. It was associated with more mastectomy, without lower re-excision⁹, with a delay of 22.4 days before treatment¹⁰, and with a rise in mastectomy rates at one institution¹¹. Its use, and the individual surgeon, were associated with initial mastectomy across several institutions¹², and it doubled the odds of contralateral prophylactic mastectomy in stage I or II disease¹³. It was not associated with better outcome after breast conservation with radiation¹⁴. A retrospective comparison reporting fewer in-breast recurrences with it, 1.2% against 6.8%, was not randomised¹⁵.

2 · Tissue acquisition, specimen adequacy, and handling

This section determines the reliability of every result in Prognostic and predictive assays and somatic profiling and Liquid biopsy and longitudinal monitoring, and it is almost always treated as logistics rather than as measurement.

Core needle biopsy replaced fine needle aspiration for a structural reason. Distinguishing invasive carcinoma from carcinoma in situ requires seeing whether the basement membrane has been breached, and that requires intact architecture. A cytological preparation can establish malignancy and cannot establish invasion.

Preanalytic variables determine what the receptor assays will report. Time from devascularisation to fixative, the fixative used, and the duration of fixation all alter antigen preservation, and the ASCO and College of American Pathologists requirements exist because of this^16,17. The mechanisms are set out in Preanalytics, fixation, and the invisible sources of error and are not repeated here. What belongs here is that these are properties of the specimen journey and not of the tumour, and they are decided by people who never see the report.

Molecular assays impose their own adequacy requirement, and it is quantifiable. In c-TRAK TN, 208 patients registered, 185 had the tumour successfully sequenced, and 171 of those had trackable mutations, which is 92.4%¹⁸. Roughly one patient in six who entered the study never reached the assay, and almost all of that loss occurred at the tissue step.

The practical consequence for a clinician ordering a biopsy is to take enough tissue the first time. Retrieving a second specimen costs a procedure, a delay, and sometimes the window in which the result would have changed the decision.

3 · Radiological and pathological concordance

Concordance is a judgement made by comparing two independent assessments of the same lesion, and it is the main safety net against a false negative biopsy.

The question is narrow. Does the histology explain the imaging finding. A fibroadenoma reported from a spiculated mass does not. A benign core from a BI-RADS 5 lesion does not. Fat necrosis reported from new microcalcification may or may not, depending on the calcification morphology.

Discordance requires action rather than reassurance, and the action is repeat sampling or excision. The failure mode is to accept a benign result because it is benign.

Two lesion types generate systematic discordance and deserve naming. Ductal carcinoma in situ extends further than its mammographic calcifications suggest, so a core that samples the calcification may underestimate the lesion Ductal carcinoma in situ. Invasive lobular carcinoma is underestimated in extent by every conventional modality Invasive lobular carcinoma as a biologically distinct disease.

A separate class of discordance is upgrade risk. Atypical ductal hyperplasia and some papillary and radial lesions on core biopsy are upgraded to malignancy at excision often enough that the core result cannot stand alone, and Atypical ductal and lobular hyperplasia covers the management.

4 · Anatomic TNM and its enduring logic

Tumour size and nodal status have survived a century of proposed replacements, and it is worth being explicit about why.

They measure two different things. Size measures how much local growth has occurred, which is a composite of growth rate and elapsed time. Nodal involvement measures something categorically different, which is demonstrated capacity to leave the primary site and establish elsewhere. Neither substitutes for the other, which is why a stage system needs both.

Distant metastasis is the third axis and it is nearly binary in its prognostic effect.

The measurement caveat on nodal status is rarely stated. Node status depends on how many nodes were examined and how thoroughly each was sectioned. A patient staged after a sentinel node procedure and a patient staged after an axillary dissection have not had the same test. The surgical evidence behind that change is in Surgery and reconstruction, with ACOSOG Z0011 and SOUND. Stage migration follows from a change in sampling intensity without any change in disease.

Anatomic stage remains the backbone for a practical reason as well as a biological one. It requires no assay beyond a ruler and a microscope, so it is the only staging system that works everywhere Resource-stratified guidelines and tiered recommendations.

5 · Prognostic stage and the entry of biology into staging

The seventh edition of the American Joint Committee on Cancer staging manual, effective for cancers diagnosed from 1 January 2010, had begun to use nonanatomic prognostic factors in defining stage groups¹⁹. The eighth edition added grade, oestrogen receptor, progesterone receptor and HER2 status to anatomic extent, producing a prognostic stage alongside the anatomic one.

The effect is much larger than a refinement. In 3,327 patients treated at a single institution, prognostic stage upstaged 29.5% of patients and downstaged 28.1% relative to anatomic stage²⁰. In 54,727 patients from the California Cancer Registry it upstaged 31.0% and downstaged 20.6%. Roughly six patients in ten changed stage group.

Discrimination improved in both cohorts. In the institutional cohort, the Harrell concordance index for disease-specific survival was 0.8357 for prognostic stage against 0.737 for anatomic stage. In the registry cohort it was 0.8426 against 0.8097. Both comparisons had P values below 0.001²⁰.

The trade is explicit and it is worth naming. Anatomic stage depends on measurements that any laboratory can make. Prognostic stage depends on immunohistochemistry and in situ hybridisation results, so it inherits every reproducibility problem those assays carry.

6 · Systemic staging, when it helps and when it harms

Whether to image for distant disease at diagnosis is a question about prevalence, not about thoroughness.

The arithmetic is unavoidable. In clinical stage I and II disease the prevalence of occult distant metastasis is low, so even a specific test generates mostly false positive findings. Each false positive generates further imaging, sometimes a biopsy, a delay in starting treatment, and a period during which a patient believes she may have metastatic disease.

The harms are therefore not hypothetical, and they are borne by the large majority of patients in whom the test was always going to be negative.

The yields have been measured. Among 516 consecutive patients, bone scan found skeletal metastases in 6.31% of those tested, liver ultrasound 0.72% and chest radiography 0.93%²¹. In a larger series the detection rate was 1.46% in small tumours with three or fewer involved nodes and 10.68% in more advanced disease²². Staging imaging in stage II and III disease produced a suspicious or indeterminate finding in 86.4% of the patients scanned²³. Routine blood counts, liver function tests and chest radiography did not detect occult metastases in early-stage disease, and generated further imaging and cost²⁴.

Prevalence rises with clinical stage, which is where systemic staging earns its place. Locally advanced and inflammatory disease carry a high enough pretest probability that a positive result is likely to be true and will change the treatment plan Locally advanced and inflammatory breast cancer. Symptoms or laboratory abnormalities that point to a specific site raise prevalence for that site and justify directed imaging at any stage.

The operative rule is to image in order to answer a question whose answer would change management. Imaging performed because a patient has cancer, rather than because a decision depends on the result, generates findings without decisions attached to them.

7 · Functional and molecular imaging, FDG, FES, and HER2 PET

Three positron emission tomography tracers are used in breast cancer and they measure three different quantities. Treating them as interchangeable is the commonest error.

Fluorodeoxyglucose measures glucose uptake, which is a proxy for metabolic and proliferative activity. It is not a proxy for receptor status or for any target. Low-grade disease and invasive lobular carcinoma are frequently poorly avid, so a negative fluorodeoxyglucose study in lobular disease has low negative predictive value Invasive lobular carcinoma as a biologically distinct disease. Tumours of 10 mm or less and low-grade tumours predicted false-negative studies of the primary²⁵. In the axilla, a prospective multicentre study found it highly predictive when several intense foci were present, and often blind to small and few nodal metastases²⁶. In recurrent or metastatic disease it had a sensitivity of 94% and a specificity of 91% where findings were confirmed²⁷, though it was less sensitive for bone than for other organs²⁸. Reviews and an NCCN task force set out these roles^29,30.

Fluoroestradiol measures oestrogen receptor binding in vivo, across every lesion, without a needle. Its value is in predicting benefit from endocrine therapy and in resolving discordant receptor results between deposits³¹.

Zirconium-89 trastuzumab measures whether an antibody actually reaches and binds its target. In 56 patients with HER2-positive metastatic disease scheduled for T-DM1, pretreatment imaging was negative in 29%, and a mixed pattern within a single patient was found in 46%³². Combining that scan with an early metabolic response assessment separated two groups, with median time to treatment failure of 2.8 months in 12 patients and 15 months in 25. Integration of molecular imaging with transcriptomic profiling in this setting is developed further elsewhere³³.

The structural point unites all three. A tissue assay reports one lesion at one time. A scan reports every lesion at one time, which is why molecular imaging detects between-lesion heterogeneity that no biopsy can Imaging-based heterogeneity, radiomics, PET metrics, and HER2 PET. None of these tracers has a randomised trial showing that acting on the result improves an outcome.

8 · Molecular breast imaging and its niche

Molecular breast imaging uses a technetium-based tracer taken up by mitochondria in metabolically active cells, imaged with a dedicated gamma camera.

Its physics give it a specific advantage. Uptake does not depend on tissue density, so masking by fibroglandular tissue does not apply. That makes it a candidate for the problem in Dense breasts, supplemental imaging, and notification policy.

Its physics also give it a specific cost. The examination delivers a whole-body radiation dose, since the tracer is systemic, whereas mammography delivers a dose to the breast alone. Dose reduction has narrowed this gap and has not eliminated it, and the comparison matters most in exactly the group where repeated screening is proposed.

The honest position is that this is a niche modality. It has a role where magnetic resonance imaging is contraindicated, unavailable or not tolerated, and it has no randomised evidence of benefit on any endpoint that matters. The evidence bar in Mammography, tomosynthesis, ultrasound, and MRI applies unchanged.

9 · Residual cancer burden and post-treatment restaging

Pathologic complete response is a binary summary of a continuous variable, and the information it discards is prognostically useful. Residual cancer burden recovers it.

The validation is large and it covers every subtype. A pooled patient-level analysis of 5,161 patients from 12 institutions and trials found that residual cancer burden score was prognostic within each phenotypic subtype³⁴. Per one-unit increase in score, the univariable hazard ratio for event-free survival was 1.55 in hormone receptor-positive HER2-negative disease, with a 95% confidence interval of 1.41 to 1.71. It was 2.16 in hormone receptor-negative HER2-positive disease, with a confidence interval of 1.79 to 2.61.

The effect survived adjustment. Adjusted for age, grade, clinical T category and nodal status at baseline, the hazard ratio ranged from 1.52 to 2.09 across subtypes, with all P values below 0.0001³⁴.

The measurement requirements are strict and they are where the score most often fails in practice. The calculation uses five measurements: the two largest dimensions of the residual tumour bed, the proportion of that bed that is cellular, the proportion of cellularity that is in situ, the number of positive nodes, and the diameter of the largest nodal deposit. Producing those figures requires the pathologist to know the specimen is post-treatment, to have the tumour bed localised, and to sample it by a defined protocol Pathological assessment after neoadjuvant treatment.

Post-treatment imaging does not substitute for this. Radiological response and pathological response diverge, and the score that predicts outcome is the pathological one. The endpoint argument belongs to Pathologic complete response as an endpoint, what it does and does not predict and Residual cancer burden and continuous response measures, and the treatment that follows residual disease to KATHERINE and CREATE-X.

10 · Multidisciplinary diagnostic integration

The multidisciplinary meeting exists to reconcile results that come from different sampling frames, and that is a harder task than reviewing them in sequence.

Four instruments are usually on the table. Imaging reports every lesion and characterises none. Histology characterises one lesion in detail. Receptor assays report protein and copy number from one block. Molecular assays report DNA from that block or from plasma. Each answers a different question about a different sample.

Most apparent contradictions dissolve once the frames are named. A HER2-positive biopsy and a negative HER2 positron emission tomography study are not in conflict, because one sampled a core and the other sampled the body³². A plasma result that lacks a mutation found in an archival block is usually reporting the passage of time rather than an error Discordance between blood, tissue, and imaging.

The failure mode is a meeting that reads out results rather than reconciling them. The useful output of the meeting is a single statement of what is known, what is uncertain, and which uncertainty would change the plan if resolved. The molecular version of the same discipline is in Reporting standards and molecular tumor boards, and the treatment-planning version in Principles of multidisciplinary treatment.

11 · What staging cannot capture

Stage is a description of anatomic extent plus a small number of assay results, and it is worth being precise about what that excludes.

It does not capture dormancy. Two patients with identical stage can carry different burdens of disseminated tumour cells, and the state those cells are in determines when and whether they reactivate Dormancy, residual disease, and late recurrence.

It does not capture subclonal composition. A tumour that is subclonally complex has more available routes of escape than one that is not, and stage is blind to that Temporal heterogeneity and clonal evolution.

It does not capture shedding, which is why a stage II patient can have detectable circulating tumour DNA and a stage III patient can have none Biology of circulating tumor DNA, shedding, half-life, and detectability.

It does not capture the host. Body composition, comorbidity, immune competence and the social determinants that govern whether treatment is completed all modify outcome, and none appears in a staging manual Host and tumor interactions, Disparities across the breast cancer continuum.

The clearest demonstration is the late recurrence hazard. Among 62,923 women with oestrogen receptor-positive disease who were disease-free after 5 years of endocrine therapy, distant recurrences continued at a steady rate from year 5 to year 20. The magnitude of that risk depended on original tumour size and nodal status³⁵. Stage predicted the magnitude of that late risk. It did not explain why the risk persisted, and nothing in the staging system does.

That is not a defect in staging. It is the boundary of what a classification based on anatomy can do, and the chapters that follow are the attempt to see past it.

References

All 35 references

1. Esserman L. Integration of imaging in the management of breast cancer. J Clin Oncol 2005 23:1601-2. PMID 15755961
2. Gundry KR. The application of breast MRI in staging and screening for breast cancer. Oncology (Williston Park) 2005 19:159-69; discussion 170, 173-4, 177. PMID 15770888
3. Bleicher RJ, Morrow M. MRI and breast cancer: role in detection, diagnosis, and staging. Oncology (Williston Park) 2007 21:1521-8, 1530; discussion 1530, 1532-3. PMID 18077995
4. Houssami N, Ciatto S, Macaskill P, et al. Accuracy and surgical impact of magnetic resonance imaging in breast cancer staging: systematic review and meta-analysis in detection of multifocal and multicentric cancer. J Clin Oncol 2008 26:3248-58. PMID 18474876
5. Turnbull L, Brown S, Harvey I, et al. Comparative effectiveness of MRI in breast cancer (COMICE) trial: a randomised controlled trial. Lancet 2010 375:563-71. PMID 20159292
6. Turnbull LW, Brown SR, Olivier C, et al. Multicentre randomised controlled trial examining the cost-effectiveness of contrast-enhanced high field magnetic resonance imaging in women with primary breast cancer scheduled for wide local excision (COMICE). Health Technol Assess 2010 14:1-182. PMID 20025837
7. Peters NH, van Esser S, van den Bosch MA, et al. Preoperative MRI and surgical management in patients with nonpalpable breast cancer: the MONET - randomised controlled trial. Eur J Cancer 2011 47:879-86. PMID 21195605
8. Weber JJ, Bellin LS, Milbourn DE, Verbanac KM, Wong JH. Selective preoperative magnetic resonance imaging in women with breast cancer: no reduction in the reoperation rate. Arch Surg 2012 147:834-9. PMID 22987175
9. Miller BT, Abbott AM, Tuttle TM. The influence of preoperative MRI on breast cancer treatment. Ann Surg Oncol 2012 19:536-40. PMID 21751044
10. Bleicher RJ, Ciocca RM, Egleston BL, et al. Association of routine pretreatment magnetic resonance imaging with time to surgery, mastectomy rate, and margin status. J Am Coll Surg 2009 209:180-7; quiz 294-5. PMID 19632594
11. Katipamula R, Degnim AC, Hoskin T, et al. Trends in mastectomy rates at the Mayo Clinic Rochester: effect of surgical year and preoperative magnetic resonance imaging. J Clin Oncol 2009 27:4082-8. PMID 19636020
12. Feigelson HS, James TA, Single RM, et al. Factors associated with the frequency of initial total mastectomy: results of a multi-institutional study. J Am Coll Surg 2013 216:966-75. PMID 23490543
13. Sorbero ME, Dick AW, Beckjord EB, Ahrendt G. Diagnostic breast magnetic resonance imaging and contralateral prophylactic mastectomy. Ann Surg Oncol 2009 16:1597-605. PMID 19330381
14. Solin LJ, Orel SG, Hwang WT, Harris EE, Schnall MD. Relationship of breast magnetic resonance imaging to outcome after breast-conservation treatment with radiation for women with early-stage invasive breast carcinoma or ductal carcinoma in situ. J Clin Oncol 2008 26:386-91. PMID 18202414
15. Fischer U, Zachariae O, Baum F, von Heyden D, Funke M, Liersch T. The influence of preoperative MRI of the breasts on recurrence rate in patients with breast cancer. Eur Radiol 2004 14:1725-31. PMID 15248080
16. Allison KH, Hammond MEH, Dowsett M, et al. Estrogen and progesterone receptor testing in breast cancer: ASCO/CAP guideline update. J Clin Oncol 2020 38:1346-1366. PMID 31928404
17. Wolff AC, Somerfield MR, Dowsett M, et al. Human epidermal growth factor receptor 2 testing in breast cancer: ASCO-College of American Pathologists guideline update. J Clin Oncol 2023 41:3867-3872. PMID 37284804
18. Turner NC, Swift C, Jenkins B, et al. Results of the c-TRAK TN trial: ctDNA mutation tracking to detect molecular residual disease and trigger intervention in moderate- and high-risk early-stage triple-negative breast cancer. Ann Oncol 2023 34:200-211. PMID 36423745
19. Edge SB, Compton CC. The American Joint Committee on Cancer: the 7th edition of the AJCC cancer staging manual and the future of TNM. Ann Surg Oncol 2010 17:1471-4. PMID 20180029
20. Weiss A, Chavez-MacGregor M, Lichtensztajn DY, et al. Validation study of the American Joint Committee on Cancer eighth edition prognostic stage compared with the anatomic stage in breast cancer. JAMA Oncol 2018 4:203-209. PMID 29222540
21. Puglisi F, Follador A, Minisini AM, et al. Baseline staging tests after a new diagnosis of breast cancer: further evidence of their limited indications. Ann Oncol 2005 16:263-6. PMID 15668281
22. Ravaioli A, Pasini G, Polselli A, et al. Staging of breast cancer: new recommended standard procedure. Breast Cancer Res Treat 2002 72:53-60. PMID 12000220
23. Brothers JM, Kidwell KM, Brown RK, Henry NL. Incidental radiologic findings at breast cancer diagnosis and likelihood of disease recurrence. Breast Cancer Res Treat 2016 155:395-403. PMID 26797222
24. Louie RJ, Tonneson JE, Gowarty M, Goodney PP, Barth RJ, Rosenkranz KM. Complete blood counts, liver function tests, and chest x-rays as routine screening in early-stage breast cancer: value added or just cost?. Breast Cancer Res Treat 2015 154:99-103. PMID 26467045
25. Kumar R, Chauhan A, Zhuang H, Chandra P, Schnall M, Alavi A. Clinicopathologic factors associated with false negative FDG-PET in primary breast cancer. Breast Cancer Res Treat 2006 98:267-74. PMID 16555126
26. Wahl RL, Siegel BA, Coleman RE, Gatsonis CG. Prospective multicenter study of axillary nodal staging by positron emission tomography in breast cancer: a report of the staging breast cancer with PET Study Group. J Clin Oncol 2004 22:277-85. PMID 14722036
27. Eubank WB, Mankoff D, Bhattacharya M, et al. Impact of FDG PET on defining the extent of disease and on the treatment of patients with recurrent or metastatic breast cancer. AJR Am J Roentgenol 2004 183:479-86. PMID 15269044
28. Moon DH, Maddahi J, Silverman DH, Glaspy JA, Phelps ME, Hoh CK. Accuracy of whole-body fluorine-18-FDG PET for the detection of recurrent or metastatic breast carcinoma. J Nucl Med 1998 39:431-5. PMID 9529287
29. Rosen EL, Eubank WB, Mankoff DA. FDG PET, PET/CT, and breast cancer imaging. Radiographics 2007 27 Suppl 1:S215-29. PMID 18180228
30. Podoloff DA, Advani RH, Allred C, et al. NCCN task force report: positron emission tomography (PET)/computed tomography (CT) scanning in cancer. J Natl Compr Canc Netw 2007 5 Suppl 1:S1-22; quiz S23-2. PMID 17509259
31. Parihar AS, et al. 18 F-Fluoroestradiol PET/CT for predicting benefit from endocrine therapy in patients with ER-positive breast cancer. J Nucl Med 2025. PMID 40081952
32. Gebhart G, Lamberts LE, Wimana Z, et al. Molecular imaging as a tool to investigate heterogeneity of advanced HER2-positive breast cancer and to predict patient outcome under trastuzumab emtansine (T-DM1): the ZEPHIR trial. Ann Oncol 2016 27:619-624. PMID 26598545
33. Rediti M, Fimereli D, Mileva M, et al. Integrating molecular imaging and transcriptomic profiling in advanced HER2-positive breast cancer receiving trastuzumab emtansine: an analysis of the ZEPHIR clinical trial. Clin Cancer Res 2025 31:110-121. PMID 39470686
34. 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
35. Pan H, Gray R, Braybrooke J, et al. 20-year risks of breast-cancer recurrence after stopping endocrine therapy at 5 years. N Engl J Med 2017 377:1836-1846. PMID 29117498