Breast Screening and Assessment for Surgeons — Programmes, Intervals, Tomosynthesis, Dense Breasts, BI-RADS, Triple Assessment and DCIS

The one-line answer

Female breast cancer is the most commonly diagnosed cancer on earth at 2.3 million new cases a year, and organised mammography screening earns its place by cutting breast-cancer mortality about one-fifth — quoted always beside its harm, overdiagnosis in roughly one in ten invited-group cancers. The surgeon's screening job is programme literacy (who, when, how often, with what test), honest consent (benefit with harm in one breath), fluent reading of BI-RADS assessment categories, triple-assessment discipline where no single arm rules, and a DCIS plan that treats the screen-detected precursor as a dilemma rather than a reflex operation.^[1][7]

A 52-year-old woman recalled from routine screening with a 14 mm asymmetry, BI-RADS 4A on diagnostic workup, and a core biopsy showing low-grade DCIS walks the whole topic in one patient: place her on the screening pathway that found her, weigh the 7.6% category-4A predictive value against the one-in-four DCIS upgrade rule, consent overdiagnosis before consenting surgery, and set surveillance or treatment by grade, size and patient preference rather than by the word carcinoma alone.^[27][32][36]

Burden — why 2.3 million cases opens every viva

• Female breast cancer has surpassed lung cancer as the most commonly diagnosed cancer worldwide, with an estimated 2.3 million new cases (11.7% of all cancers). Open with this; it anchors why every board examines screening.^[1]
• The mortality contrast matters more than the incidence rank: death rates for female breast cancer were considerably higher in transitioning versus transitioned countries (15.0 vs 12.8 per 100,000), so a candidate who mentions organised screening must also mention where programmes do not reach.^[1]
• In the United States, breast cancer is the second most common cancer and the second most common cause of cancer death among women — the sentence that justifies a national recommendation statement at all.^[2]
• Australia's programme story is the Southern-hemisphere anchor: BreastScreen Australia targets women aged 50 to 74, participation sits at 56%, and mortality from breast cancer has decreased by 32% over 20 years in response to screening and treatment advances together — never attribute the fall to screening alone.^[43]
• The Australian aggregate cohort quantifies the participation effect: a 22% reduction in six-year cumulated breast-cancer mortality at roughly 60% participation, rising to 25% at the 70% biennial-participation target.^[42]

Benefit — the 20% mortality reduction with the age gradient

• The Marmot headline, examined first because every guideline descends from it: across 11 randomised trials, inviting women to screening cut breast-cancer mortality with a relative risk of 0.80 — a 20% relative risk reduction — in the UK 50-to-70 three-yearly setting.^[7]
• The panel's own honesty clause travels with the number: the trials are old, imaging and treatment have moved on, and observational studies suggest larger effects but may be biased — so 20% is reasonable, not precise.^[7]
• Three independent trial meta-analyses agree (UK panel 0.80, Canadian Task Force 0.82, Cochrane 0.81), cohort studies run more favourable at 0.75, and modelling lands at 0.85 — consistency across methods, with the randomised estimates the ones to quote.^[5]
• The Nelson age gradient is the viva's favourite table because benefit concentrates with age: per 10,000 women over 10 years, screening prevented 3 deaths at 39 to 49 (RR 0.92, not significant), 8 at 50 to 59 (0.86), 21 at 60 to 69 (0.67), and 13 at 70 to 74 (0.80, not significant).^[4]
• Two defences when the examiner pushes: advanced breast cancer fell at age 50 and over (RR 0.62) even where mortality significance wobbled, and the 39-to-49 meta-analysis still shows a 17% mortality reduction (RR 0.83) — real but fragile, losing significance without pre-1980 trials.^[4][6]
• The humility footnote closes the section: all-cause mortality was not reduced with screening, and most trials used imaging and treatments that are now outdated — so modern benefit is inferred, not measured.^[4]

Harm part 1 — overdiagnosis, quoted with the benefit

• Marmot's definition first, because examiners test it verbatim: overdiagnosis means cancers detected at screening that would not otherwise have become clinically apparent in the woman's lifetime.^[7]
• The paired harm estimate: excess incidence of 11% as a proportion of invited-group cancers in the long term, and 19% of cancers diagnosed during the active screening period — this is the number that stands beside the 20% benefit in every consent.^[7]
• Welch widens the lens and sharpens the consent duty: about 25% of mammographically detected breast cancers represent overdiagnosis, and patients must be adequately informed of the nature and magnitude of the trade-off with early detection.^[8]
• Denmark's organised programmes read lower with long follow-up: pooled relative risk 1.04 for all targeted women, consistent with overdiagnosis most likely around 2.3% — method and denominator explain the spread from Marmot, not contradiction.^[9]
• Australia's honest range sits between: overdiagnosis estimated as low as 8% of detected cancers, with up to 30% in some research — quote the range, not a point.^[43]
• Evans names the engine room: estimates of overdiagnosis at screening sit around 10%, DCIS represents 20% of screen-detected cancers and is the main focus of the debate, and treating low-grade DCIS and tubular cancer looks like overdiagnosis in most cases.^[11]
• Wallis gives the mechanism in one line: overdiagnosis happens because tumours grow at different rates and screening preferentially catches the slow ones — the inevitable flip side of early detection, managed by de-escalating treatment rather than abandoning screening.^[10]

Who and when — USPSTF 2024 versus ACS 2015 with programme deltas

• USPSTF 2024, stated exactly: biennial screening mammography for women aged 40 to 74 years carries a B recommendation on moderate certainty of moderate net benefit; evidence is insufficient (I statements) for screening at 75 and over and for supplemental ultrasound or MRI in dense breasts after a negative mammogram.^[2]
• ACS 2015, stated exactly: offer annual screening with the opportunity to start between 40 and 44, screen annually from 45 to 54, transition to biennial at 55 and over (or continue annual), keep screening while health is good and life expectancy is 10 years or longer — and do not recommend clinical breast examination for average-risk screening at any age.^[3]
• The genuine disagreement to air at viva: USPSTF starts routine biennial screening at 40 while ACS anchors regular screening at 45 with annual frequency to 54 — reconcile it with the Nelson gradient (3 per 10,000 prevented at 39 to 49) and the ACS interval finding that premenopausal women show more favourable tumour characteristics with annual rather than biennial screens.^[4][3]
• ACS adds the consent arithmetic for young starters: cumulative lifetime false-positive risk rises when screening starts younger, through more mammograms and higher recall rates in younger women.^[3]
• Programme deltas for the global candidate: BreastScreen Australia invites 50 to 74 biennially; England invites 50 to 70 three-yearly with a randomised 47-to-73 extension piloted at 63% uptake (47 to 49) and 62% (71 to 73) — where younger extenders were recalled more (7.5 vs 3.0%) but yielded fewer cancers (0.5 vs 1.1%).^[43][7][44]
• Average-risk guidance stops where risk starts: first-degree family history (annual offered in New South Wales), BRCA carriage (MRI surveillance from late teens to 30 in the international survey), prior chest irradiation — each leaves this pathway for its own.^[14][45]

Interval — annual versus biennial versus triennial

• ECIBC's bottom line first: the balance probably favours biennial screening at 50 to 69 — annual adds small benefit with an important false-positive increase, triennial trims benefit while avoiding some harms — all on very-low-certainty evidence that you must declare.^[12]
• Age splits the balance: at 45 to 49 annual screening shows smaller benefit gains with larger harms (less favourable than 50 to 69), while at 70 to 74 longer intervals look more favourable with fewer added harms.^[12]
• British Columbia's natural experiment grounds the trade: annual versus biennial detection ran 2.32 versus 3.32 per thousand screens, with estimated 10-year breast-cancer-specific survival of 90.4 versus 89.2% — a 1.2% absolute gain for yearly screens.^[13]
• The family-history exception with prognostic teeth: New South Wales women with a first-degree relative offered annual screening had smaller tumours (20 mm or smaller, adjusted odds ratio 1.91) and more node-negative disease (AOR 1.61) than biennially screened peers — prognostic, not mortality, evidence, so frame it as such.^[14]
• The viva trap to defuse: annual screening finds more cancers per round, but Myers and Nelson both record uncertainty whether annual beats biennial on mortality — detection is not survival, and the examiner who asks knows it.^[5][4]

Tomosynthesis — detection uplift with honest limits

• TOSYMA is the randomised anchor: 99,689 women across 17 German screening units, invasive cancer detection 7.1 per 1000 with tomosynthesis plus synthesised 2D versus 4.8 per 1000 with digital mammography alone (odds ratio 1.48), with no serious adverse events — and interval-cancer results explicitly pending.^[15]
• The IPD meta-analysis generalises it: pooled detection gain of 25.49 per 10,000 for DBT over mammography across 66,451 DBT and 170,764 mammography participants, larger in dense breasts (35.19 vs 17.4 per 10,000), recall difference indistinguishable at 0.18%, and recall predictive value higher for DBT (rate ratio 1.31).^[16]
• STORM supplies the caution: interval cancers ran 1.23 per 1000 after tomosynthesis versus 1.60 per 1000 with concurrent 2D screening, sensitivity 85.5 versus 77.3% — marginally better, small numbers, and the authors demand much larger or pooled studies before claiming an interval benefit.^[17]
• The strategy paper sharpens the reading question: 20 cancers were found only with integrated 2D/3D reading and none only with 2D alone, with a more favourable false-to-true-positive trade-off under either single or double reading.^[18]
• Australia's pilots make it a service question, not just a detection one: Victoria found detection of 9.8 versus 6.6 per 1000 with recall of 4.2 versus 3.0% and reading time of 67 versus 16 seconds, while Maroondah showed DBT recalls shift toward calcifications, need more core biopsies (45.6 vs 28.6%) and less extra mammography — detection up, workload changed.^[19][20]
• The US recall practice note closes the loop: DBT recalls ran lower than full-field mammography (8.0 vs 10.6%), and prior films for comparison cut false positives — so the modality answer always includes fetching the priors.^[39]

Dense breasts — ultrasound, MRI and notification

• Dense tissue does double harm: it is an independent risk factor for breast cancer and it lowers mammographic sensitivity — the sentence that opens every supplemental-screening discussion.^[24]
• DENSE is the randomised answer in extremely dense breasts: among 40,373 Dutch women aged 50 to 75 with normal mammograms, supplemental MRI cut interval cancers from 5.0 to 2.5 per 1000 — with 59% acceptance, detection of 16.5 per 1000 in scanned women, false positives of 79.8 per 1000, and predictive value of 17.4% for recall and 26.3% for biopsy.^[21]
• EA1141 pits abbreviated MRI directly against DBT in dense breasts: invasive detection 11.8 versus 4.8 per 1000, sensitivity 95.7 versus 39.1%, specificity 86.7 versus 97.4% — MRI finds far more cancer at the price of recalling more benign tissue, which is exactly the trade to consent.^[22]
• Adjunct ultrasound keeps its place with a defined yield: robust data support improved detection of small node-negative invasive cancers, from the first Connecticut legislation in 2009 through the 2019 federal notification mandate.^[23]
• Notification moved practice even where mortality evidence lags: US density laws lifted supplemental screening use by 0.5 to 143%, with biopsy rates up by as much as 4% and detection up by as much as 11% — systems effects the surgeon should expect, not just individual decisions.^[24]
• The guideline honesty clause stays: despite DENSE, USPSTF judges supplemental ultrasound or MRI evidence insufficient in dense breasts with negative mammograms — interval cancers are not mortality, and the examiner will ask what is missing.^[2]
• Density also sharpens risk prediction rather than just hiding lesions: in 50,628 UK screening participants, density outperformed either risk model alone and added information to Tyrer-Cuzick (combined AUC 0.61), so density belongs in risk assessment as well as modality choice.^[25]

Risk and high-risk surveillance — models, BRCA, MRI trends

• The international BRCA picture before risk-reducing surgery is consistent: overwhelmingly breast MRI from age 18 to 30 with clinical examination every 6 to 12 months across 22 centres in 16 countries — while post-mastectomy surveillance collapses to examination alone almost everywhere, an admitted evidence gap.^[45]
• The MRI threshold to state with its caveat: screening MRI in addition to mammography is recommended above 20% lifetime risk on Tyrer-Cuzick or equivalent familial models — while a single-centre study found neither Tyrer-Cuzick nor Gail significantly predicted MRI yield, so the threshold is consensus, not proven calibration.^[26]
• US practice is converging on guidance from a low base: screening MRI rose from 2.9 to 12.1 per 10,000 women between 2006 and 2016, and BRCA carriers — under 1% of women — received roughly 9% of screening MRIs.^[46]
• Family history without a known mutation still earns intensity: the New South Wales annual-offer cohort shows the prognostic payoff in smaller, node-negative tumours — act on pedigree even when genes are uninformative.^[14]

BI-RADS — reading the report like a surgeon

• The category-to-action ladder, fixed in memory: 0 means incomplete and recalls; 1 to 2 return to routine screening; 3 means probably benign with short-interval surveillance; 4 means tissue diagnosis with 4A/4B/4C stratifying suspicion; 5 means highly suggestive with surgical planning alongside biopsy; 6 marks known biopsy-proven malignancy under staging.
• The National Mammography Database quantifies category 4: 9.6% of 1.3 million diagnostic mammograms were category 4, subdivided in only one-third — with biopsy predictive values of 7.6% for 4A, 22% for 4B, 69.3% for 4C, and 21.1% overall.^[27]
• The ultrasound reader study warns against overcalling the low end: category-3 predictive values ran 4.4 to 6.0% against the probably-benign benchmark, 4a ran near 30% where the benchmark expects low single digits, while 4c met its benchmark — and regular subcategory users performed better than occasional ones.^[28]
• Category 3 in women under 30 is the counselling prototype: zero malignancies in Merchant's cohort, biopsies driven more by patient request (63.4%) than lesion growth, and lesions larger than 2 cm carrying 4.4-fold odds of upfront biopsy — so surveillance is evidence-based but loss to follow-up (30%) is the programme's weak joint.^[29]
• The discordance rule overrides every number above: a benign needle result against 4C or 5 imaging is a sampling failure until proven otherwise — re-biopsy or excise, never reassure.^[27][32]

Triple assessment and tissue — core-biopsy discipline

• The triple contract in one line: clinical examination, imaging, and needle pathology each vote, concordance decides, and no single arm overrules the other two.
• In women under 40 the full contract holds: cancer in 0.88% of one-stop referrals, a breast lump multiplying odds more than eleven-fold (OR 11.43), and the authors' verdict that youth changes infrastructure for nobody — missed cancers here carry personal, clinical and legal cost.^[34]
• Under 25 with doubly benign findings is the examined exception: zero cancers among 955 young women, so omitting needle sampling when examination and ultrasound are both normal or benign would have spared biopsy in all but 3% — a narrow, well-fenced omission, not a precedent for older or suspicious cases.^[35]
• Core needle beats fine-needle on sensitivity (87 vs 74%) with similar specificity (98 vs 96%), which is why image-guided core is the assessment standard — while fine-needle keeps a first-choice role for suspicious nonpalpable lesions in experienced hands.^[30]
• The large-core meta-analysis sets the residual-risk frame: sensitivity 97% with 94% biopsy-to-surgery agreement, meaning under 1% residual cancer risk after a benign core result at US prevalence — rising to 4 to 6% where referred prevalence runs 60 to 70%, where benign cores earn caution, not closure.^[31]
• Receptor testing on core is reliable enough to act on but honest about negatives: oestrogen-receptor agreement 92.8%, progesterone 85.2%, with negative hormone results interpreted cautiously or repeated on excision.^[33]
• The underestimation rule every management plan respects: one in four core-biopsy DCIS diagnoses (25.9%) prove invasive at excision — worse with 14-gauge devices, high grade, lesions over 20 mm, mass pattern, and palpability.^[32]

DCIS — the screen-detected dilemma

• DCIS accounts for 15 to 25% of all breast-cancer diagnoses, most cases never progress to invasion, and its incidence grew with 1980s-to-1990s mammography — the three facts that make it overdiagnosis's engine and the surgeon's dilemma.^[36][37]
• Treatment honesty first: surgery, radiotherapy and optional endocrine therapy cut local events with no effect on survival outcomes — so every DCIS recommendation is a local-control discussion, not a survival claim.^[36]
• Active surveillance is the examined frontier, not yet the standard: low-risk DCIS under monitoring (LORIS and ongoing trials) may let selected women avoid surgery — offered only with explicit uncertainty and intensive follow-up.^[36][10]
• Biology cannot yet pick the progressors: current biomarkers identify neither occult invasion at biopsy nor future progression risk — which is why de-escalation trials, not assays, carry the field forward.^[37]
• The upgrade rule fences the frontier: high grade, size over 20 mm, mass calcification patterns and palpability predict invasion at excision — such DCIS stays in the treatment pathway, never the surveillance conversation.^[32]

Recall and false positives — the burden and the patient experience

• The US consent number first: about one-half of annually screened women face at least one false positive over ten screens — 41.9% after eight screens in the Breast Cancer Surveillance Consortium, against 16.1% in Copenhagen and 7.4% in Funen over the same eight.^[38]
• Europe's biennial arithmetic is kinder but not kind: Barcelona estimates 32.4% cumulative recall over ten biennial screens, starting at 10.6% on first screen and settling near 3.8% — one-third of women recalled at least once per decade of screening.^[40]
• The transatlantic gap survives adjustment: neither interval, technology nor statistical method explains the US-Europe difference — thresholds, litigation climate and reading culture do the unmeasured work.^[38]
• Recall concentrates where you expect: younger age (false-positive group 56.1 vs true-positive 62.9 years), prior benign disease (odds ratio 8.48 in Barcelona), perimenopause, and absent priors — while DBT (8.0 vs 10.6%), available prior films (90.8 vs 95.8%), and older age all lower it.^[39][40]
• Recalled women ask for communication, not just speed: a clear reason for recall, who may attend, how long it takes, who they will see and which tests follow, reassurance that speed does not mean cancer, a choice between invasive assessment and watchful waiting, and a nurse specialist available after the all-clear.^[41]

Exam Synthesis and Fence Map — what this topic owns

• This topic owns the screening pathway and its assessment: burden, benefit-with-harm, who-when-how-often, tomosynthesis, dense breasts, risk models, BI-RADS reading, triple assessment, core-biopsy discipline, DCIS-as-dilemma, and recall craft.^[2][7]
• The companion breast-cancer topic owns what happens after tissue diagnosis: receptors, staging, conservation versus mastectomy, the axilla, EBCTCG systemics, HER2 escalation and BRCA therapeutics — referred, never repeated here.
• Regional fences held throughout: USPSTF and ACS for the United States, BreastScreen Australia for Australia, the NHS programme and Marmot for the United Kingdom, ECIBC for Europe — each number carried with its flag.^{[2][3][43][7][12]}
• Deliberately unfenced as numbers: the BI-RADS 3 sub-2% atlas benchmark (discussed through Stavros and Merchant predictive values instead), programme radiation doses, and male-breast specifics — absent from fetched abstracts, so absent from this topic.^[28][29]

Exam pearls — the one-liners that score

• 2.3 million cases worldwide; mortality 20% down, overdiagnosis 11% up — the pair, always together.^[1][7]
• Biennial 40 to 74 in the United States (USPSTF), annual 45 to 54 by ACS preference, biennial favoured 50 to 69 in Europe — three flags, three answers.^[2][3][12]
• TOSYMA 7.1 versus 4.8, DENSE 2.5 versus 5.0, STORM 1.23 versus 1.60 per thousand — the three interval numbers that prove you read the trials.^[15][21][17]
• 4A 7.6, 4B 22, 4C 69.3 — the subdivision ladder that turns a report into a plan.^[27]
• One in four core DCIS upgrades; core sensitivity 87 beats fine-needle 74; triple assessment to 25 and under only when doubly benign.^[32][30][35]
• Half recalled in a decade of US annual screens; one-third in European biennial — consent the recall before the screen.^[38][40]

US annual ~50% after ten screens (41.9% after eight in BCSC); Barcelona biennial 32.4% after ten; first screen 10.6% falling to 3.8%.^[8]

References46ShowHide

1. [1]Sung H, et al. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin, 2021.PMID 33538338
2. [2]Nicholson WK, et al. Screening for Breast Cancer: US Preventive Services Task Force Recommendation Statement. JAMA, 2024.PMID 38687503
3. [3]Oeffinger KC, et al. Breast Cancer Screening for Women at Average Risk: 2015 Guideline Update From the American Cancer Society. JAMA, 2015.PMID 26501536
4. [4]Nelson HD, et al. Effectiveness of Breast Cancer Screening: Systematic Review and Meta-analysis to Update the 2009 U.S. Preventive Services Task Force Recommendation. Ann Intern Med, 2016.PMID 26756588
5. [5]Myers ER, et al. Benefits and Harms of Breast Cancer Screening: A Systematic Review. JAMA, 2015.PMID 26501537
6. [6]Magnus MC, et al. Effectiveness of mammography screening in reducing breast cancer mortality in women aged 39-49 years: a meta-analysis. J Womens Health (Larchmt), 2011.PMID 21413892
7. [7]Independent UK Panel on Breast Cancer Screening The benefits and harms of breast cancer screening: an independent review. Lancet, 2012.PMID 23117178
8. [8]Welch HG, et al. Overdiagnosis in cancer. J Natl Cancer Inst, 2010.PMID 20413742
9. [9]Njor SH, et al. Overdiagnosis in screening mammography in Denmark: population based cohort study. BMJ, 2013.PMID 23444414
10. [10]Wallis MG How do we manage overdiagnosis/overtreatment in breast screening? Clin Radiol, 2018.PMID 29174412
11. [11]Evans A, et al. Overdiagnosis in breast imaging. Breast, 2017.PMID 28029492
12. [12]Canelo-Aybar C, et al. Benefits and harms of annual, biennial, or triennial breast cancer mammography screening for women at average risk of breast cancer: a systematic review for the European Commission Initiative on Breast Cancer (ECIBC). Br J Cancer, 2022.PMID 34837076
13. [13]Wai ES, et al. Comparison of 1- and 2-year screening intervals for women undergoing screening mammography. Br J Cancer, 2005.PMID 15714210
14. [14]Randall D, et al. Annual or biennial mammography screening for women at a higher risk with a family history of breast cancer: prognostic indicators of screen-detected cancers in New South Wales, Australia. Cancer Causes Control, 2009.PMID 19015941
15. [15]Heindel W, et al. Digital breast tomosynthesis plus synthesised mammography versus digital screening mammography for the detection of invasive breast cancer (TOSYMA): a multicentre, open-label, randomised, controlled, superiority trial. Lancet Oncol, 2022.PMID 35427470
16. [16]Libesman S, et al. An individual participant data meta-analysis of breast cancer detection and recall rates for digital breast tomosynthesis versus digital mammography population screening. Clin Breast Cancer, 2022.PMID 35246389
17. [17]Houssami N, et al. Interval breast cancers in the 'screening with tomosynthesis or standard mammography' (STORM) population-based trial. Breast, 2018.PMID 29328943
18. [18]Houssami N, et al. Breast screening using 2D-mammography or integrating digital breast tomosynthesis (3D-mammography) for single-reading or double-reading--evidence to guide future screening strategies. Eur J Cancer, 2014.PMID 24746887
19. [19]Houssami N, et al. Pilot trial of digital breast tomosynthesis (3D mammography) for population-based screening in BreastScreen Victoria. Med J Aust, 2019.PMID 31448816
20. [20]Li T, et al. Assessment of screen-recalled abnormalities for digital breast tomosynthesis versus digital mammography screening in the BreastScreen Maroondah trial. J Med Imaging Radiat Oncol, 2023.PMID 35768941
21. [21]Bakker MF, et al. Supplemental MRI Screening for Women with Extremely Dense Breast Tissue. N Engl J Med, 2019.PMID 31774954
22. [22]Comstock CE, et al. Comparison of Abbreviated Breast MRI vs Digital Breast Tomosynthesis for Breast Cancer Detection Among Women With Dense Breasts Undergoing Screening. JAMA, 2020.PMID 32096852
23. [23]Butler RS, et al. Screening Breast Ultrasound: Update After 10 Years of Breast Density Notification Laws. AJR Am J Roentgenol, 2020.PMID 32182096
24. [24]Huang S, et al. The impact of mandatory mammographic breast density notification on supplemental screening practice in the United States: a systematic review. Breast Cancer Res Treat, 2021.PMID 33774734
25. [25]Brentnall AR, et al. Mammographic density adds accuracy to both the Tyrer-Cuzick and Gail breast cancer risk models in a prospective UK screening cohort. Breast Cancer Res, 2015.PMID 26627479
26. [26]Sevdalis A, et al. The Value of Tyrer-Cuzick Versus Gail Risk Modeling in Predicting Benefit from Screening MRI in Breast Cancer. Eur J Breast Health, 2022.PMID 35059595
27. [27]Elezaby M, et al. ACR BI-RADS Assessment Category 4 Subdivisions in Diagnostic Mammography: Utilization and Outcomes in the National Mammography Database. Radiology, 2018.PMID 29315061
28. [28]Stavros AT, et al. Ultrasound positive predictive values by BI-RADS categories 3-5 for solid masses: An independent reader study. Eur Radiol, 2017.PMID 28396996
29. [29]Merchant K, et al. ACR BI-RADS Category 3 Lesions in Women Younger Than 30: Follow-up Outcomes and Factors Associated With Biopsy. J Ultrasound Med, 2021.PMID 33599304
30. [30]Wang M, et al. A sensitivity and specificity comparison of fine needle aspiration cytology and core needle biopsy in evaluation of suspicious breast lesions: A systematic review and meta-analysis. Breast, 2017.PMID 27866091
31. [31]Verkooijen HM, et al. Diagnostic accuracy of large-core needle biopsy for nonpalpable breast disease: a meta-analysis. Br J Cancer, 2000.PMID 10737383
32. [32]Brennan ME, et al. Ductal carcinoma in situ at core-needle biopsy: meta-analysis of underestimation and predictors of invasive breast cancer. Radiology, 2011.PMID 21493791
33. [33]Li S, et al. Assessment accuracy of core needle biopsy for hormone receptors in breast cancer: a meta-analysis. Breast Cancer Res Treat, 2012.PMID 22527109
34. [34]Mazari FAK, et al. The need for triple assessment and predictors for diagnosis of breast cancer in patients under 40 years of age. Clin Radiol, 2018.PMID 29731127
35. [35]Yue D, et al. Triple assessment is not necessary in most young women referred with breast symptoms. Ann R Coll Surg Engl, 2015.PMID 26274736
36. [36]Delaloge S, et al. Ductal carcinoma in situ of the breast: finding the balance between overtreatment and undertreatment. Lancet, 2024.PMID 38735296
37. [37]Grimm LJ, et al. Ductal Carcinoma in Situ: State-of-the-Art Review. Radiology, 2022.PMID 34931856
38. [38]Kemp Jacobsen K, et al. Comparison of cumulative false-positive risk of screening mammography in the United States and Denmark. Cancer Epidemiol, 2015.PMID 26013768
39. [39]Honig EL, et al. Factors Impacting False Positive Recall in Screening Mammography. Acad Radiol, 2019.PMID 30772138
40. [40]Castells X, et al. Cumulative false positive recall rate and association with participant related factors in a population based breast cancer screening programme. J Epidemiol Community Health, 2006.PMID 16537348
41. [41]Bond M, et al. Improving screening recall services for women with false-positive mammograms: a comparison of qualitative evidence with UK guidelines. BMJ Open, 2015.PMID 25618139
42. [42]Morrell S, et al. Mammography screening and breast cancer mortality in Australia: an aggregate cohort study. J Med Screen, 2012.PMID 22345322
43. [43]Olver IN, et al. History, development and future of cancer screening in Australia. Public Health Res Pract, 2017.PMID 28765858
44. [44]Moser K, et al. Extending the age range for breast screening in England: pilot study to assess the feasibility and acceptability of randomization. J Med Screen, 2011.PMID 21852703
45. [45]Madorsky-Feldman D, et al. An international survey of surveillance schemes for unaffected BRCA1 and BRCA2 mutation carriers. Breast Cancer Res Treat, 2016.PMID 27117159
46. [46]Wernli KJ, et al. Trends in screening breast magnetic resonance imaging use among US women, 2006 to 2016. Cancer, 2020.PMID 32985694