JNCI Cancer Spectrum

Screening digital breast tomosynthesis (DBT) aims to identify breast cancer early when treatment is most effective leading to reduced mortality. In addition to early detection, the information contained within DBT images may also inform subsequent risk stratification and guide risk-reducing management. We obtained a 5-year area under the curve (AUC) = 0.78 (95% confidence interval (CI) = 0.75 - 0.80) in the internal validation. The model validated in external data (n=6,553 women; AUC = 0.77 (95% CI, 0.74 - 0.80). There was no change in the AUC when age and BI-RADS density are added to the synthetic DBT image. The model significantly outperforms the Tyrer-Cuzick model (p<0.01). Our model extends risk prediction applications to synthetic DBT, provides 5-year risk estimates, and is readily calibrated to national risk strata for clinical translation and application in the setting of US risk management guidelines. The model could be implemented within any digital mammography program. One Sentence SummaryWe develop and externally validate a 5-year risk prediction model for breast cancer using synthetic digital breast tomosynthesis and demonstrate clinical utility by calibrating to the national risk strata.

Objective: To develop and evaluate a deep learning model for five-year breast cancer risk prediction from screening breast ultrasound (BUS) examinations. Methods: This retrospective study included 295,298 breast ultrasound examinations from 122,072 women imaged between 2012 and 2020. Patients were split into training, validation, and test sets; the test set included screening examinations only. BUS-Risk-Net aggregated image features using attention-based multiple instance learning and combined them with age and ultrasound-estimated breast density to predict 2- to 5-year risk. Performance was compared with the full Tyrer-Cuzick model in a matched case-control cohort and with a reduced Tyrer-Cuzick model in the held-out test set. Risk stratification was evaluated within BI-RADS density categories. Results: In the matched case-control cohort (n = 240 women), BUS-Risk-Net achieved a 5-year AUC of 0.632 (95% CI, 0.562-0.702), versus 0.514 for the full Tyrer-Cuzick model (95% CI, 0.440-0.588; p = 0.04). Among 19,548 examinations from 9,015 women eligible for 5-year evaluation in the test set, BUS-Risk-Net achieved an AUC of 0.679 (95% CI, 0.653-0.706), versus 0.594 for the reduced Tyrer-Cuzick model (95% CI, 0.564-0.623; P < .001). Observed 5-year cancer incidence increased across AI-defined risk tiers within each BI-RADS density category, ranging from 0.0% to 5.8% after AI stratification, compared with 2.1% to 3.6% across density categories alone. Discussion: Deep learning models applied to screening breast ultrasound could enable long-term breast cancer risk prediction and stratify risk beyond breast density alone. External and prospective validation is needed before clinical use.

BackgroundBreast cancer is multifactorial. Focusing on limited risk factors may miss high-risk individuals. MethodsWe assessed the performance and overlap of various risk factors in identifying high-risk individuals for invasive breast cancer (BrCa) and ductal carcinoma in situ (DCIS) in 161,849 European-ancestry and 18,549 Asian-ancestry women. Discriminatory ability was evaluated using the area under the receiver operating characteristic curve (AUC). High-risk criteria included: 5-year absolute risk [&ge;]1{middle dot}66% by the Gail model [GAILbinary]; first-degree family history of breast cancer [FHbinary]; 5-year absolute risk [&ge;]1{middle dot}66% by a 313-variants polygenic risk score [PRSbinary]; and carriers of pathogenic variants in breast cancer predisposition genes [PTVbinary]. FindingsThe 5-year absolute risk by PRS outperformed the Gail model in predicting BrCa (Europeansvs controls: AUCPRS=0{middle dot}635 [0{middle dot}632-0{middle dot}638] vs AUCGail=0{middle dot}492 [0{middle dot}489-0{middle dot}495]; Asiansvs controls: AUCPRS=0{middle dot}564 [0{middle dot}556-0{middle dot}573] vs AUCGail=0{middle dot}506 [0{middle dot}497-0{middle dot}514]). PRSbinary and GAILbinary identified more high-risk European than Asia individuals. High-risk proportions were higher among BrCa (16-26%) and DCIS (20-33%) compared to controls (9-15%) among young Europeans and all Asians. Fewer than 7% of BrCa, 10% of DCIS, and 3% of controls were classified as high-risk by multiple risk classifiers. Overlap between PRSbinary and PTVbinary was minimal (<0{middle dot}65% Europeans, <0{middle dot}15% Asians) compared to the proportion at high risk using PTVbinary alone (Europeans: 4{middle dot}6%, Asians: 4{middle dot}4%) and PRSbinary alone (Europeans: 13{middle dot}9%, Asians: 8{middle dot}5%). PRSbinary and FHbinary uniquely identified 5-6% and 9-11% of young BrCa, respectively. InterpretationThe incomplete overlap between high-risk individuals identified by PRSbinary, GAILbinary, FHbinary, and PTVbinary highlights the need for a comprehensive approach to breast cancer risk prediction. SIGNIFICANCEThis study shows that different ways of predicting breast cancer risk do not always flag the same people, suggesting that combining multiple risk factors could improve early detection and screening.

ObjectiveThis systematic review aimed to assess methods used to relate repeated mammographic images to breast cancer risk, including the time from mammogram to diagnosis of breast cancer, and methods for analysis of data from either one or both breasts (averaged or assessed individually). DesignA systematic review was performed. SettingThe databases including Medline (Ovid) 1946-, Embase.com 1947-, CINAHL Plus 1937-, Scopus 1823-, Cochrane Library (including CENTRAL), and Clinicaltrials.gov were searched through October 2021 to extract published articles in English describing the relationship of change in mammographic features with risk of breast cancer. ParticipantsWomen with mammogram images. Main outcome measureBreast cancer incidence. ResultsTwenty articles were included in the final review. We found that BIRADs and Cumulus were most commonly used for classifying mammographic density and automated assessment was used on more recent digital mammograms. Time between mammograms varied from 1 to median of 4.1 years, and only 9 of the studies used more than 2 mammograms to quantify features. One study used a prediction horizon of 5 and 10 years, one used 5 years only and another 10 years only, while in the others the prediction horizon was not clearly defined with investigators using the next screening mammogram. ConclusionThis review provided an updated overview of the state of the art and revealed research gaps; based on these, we provide recommendations for future studies using repeated measure methods for mammogram images to make the use of accumulating image data. By following these recommendations, we expect to improve risk classification and risk prediction for women to tailor screening and prevention strategies to level of risk. Article summaryO_ST_ABSStrengths and limitations of the studyC_ST_ABSO_LITo the best of our knowledge, this is the most recent systematic review on the topic of using multiple mammogram images to define risk of breast cancer. C_LIO_LIThis review was performed strictly following systematic review guidelines including a medical librarian with expertise in searching, multiple independent reviewers involved in study selection and data extraction, and reporting following PRISMA 2020 guidelines. C_LIO_LIDue to heterogeneity of methods for assessment and classification (categorical and continuous) of mammographic features including breast density and time to breast cancer, we did not perform risk of bias or conduct a meta-analysis. C_LIO_LIFew studies looked at repeated measures of non-density features. C_LI

Nottingham histologic grading is essential for breast cancer prognostication but suffers from inter-observer variability in assessing mitotic activity, nuclear pleomorphism, and tubule formation. We developed MOSAIC (Mammary Oncology Spatial Analysis and Intelligent Classification), an explainable AI framework designed to perform component-wise grading by independently modeling these three histologic features. Model outputs were calibrated using a two-phase pathology study to establish clinically reproducible scoring thresholds and were subsequently evaluated across public datasets and multi-institutional Indian cohorts. MOSAIC demonstrated robust performance, with AI-derived grades providing independent prognostic information (HR >= 1.8 in two datasets, p = < 0.001) and improved survival stratification compared to traditional methods. In pathologist calibration studies, AI-assisted scoring significantly reduced variability, specifically achieving near-perfect agreement in mitotic scoring with a weighted {kappa} up to 0.98. Accuracy and Cohens kappa ({kappa}) analysis further characterized the models technical performance across components: Tubule formation showed the highest agreement (Accuracy >= 0.6607, {kappa} = 0.549), followed by overall Grade (Accuracy = 0.5637, {kappa} = 0.539) and Mitotic activity (Accuracy = 0.4985, {kappa} = 0.4), while Nuclear pleomorphism proved the most challenging (Accuracy = 0.3303, {kappa} = 0.271). Comparative survival models confirmed that AI-derived grades were more significant predictors of risk than manual pathologist-assigned grades, with the AI model yielding a superior global p-value (5.9 x 10-7) and lower AIC (769.61). These results indicate that MOSAIC enables reproducible, interpretable grading by decomposing assessment into pathology-aligned components. By enhancing consistency while preserving prognostic relevance, this framework supports explainable AI as a viable assistive tool for routine breast cancer pathology.

ObjectiveThis systematic review update synthesized recent evidence on the benefits and harms of breast cancer screening in women aged [&ge;] 40 years and aims to inform the Canadian Task Force on Preventive Health Cares (CTFPHC) guideline update. MethodsWe searched Ovid MEDLINE(R) ALL, Embase Classic+Embase, and Cochrane Central Register of Controlled Trials to update our searches to July 8, 2023. Search results for observational studies were limited to publication dates from 2014 to capture more relevant studies. Screening was performed independently and in duplicate by the review team. To expedite the screening process, machine learning was used to prioritize relevant references. Critical health outcomes, as outlined by the CTFPHC, included breast cancer and all-cause mortality, treatment-related morbidity, and overdiagnosis. Randomized controlled trials (RCTs), non/quasi RCTs, and observational studies were included. Data extraction and quality assessment were performed by one reviewer and verified by another. Risk of bias was assessed using the Cochrane Risk of Bias 2.0 tool for RCTs and the Joanna Briggs Institute (JBI) checklists for non-randomized and observational studies. When deemed appropriate, studies were pooled via random-effects models. The overall certainty of the evidence was assessed following GRADE guidance. ResultsThree new papers reporting on existing RCT trial data and 26 observational studies were included. No new RCTs were identified in this update. No study reported results by ethnicity, race, proportion of study population with dense breasts, or socioeconomic status. For breast cancer mortality, RCT data from the prior review reported a significant relative reduction in the risk of breast cancer mortality with screening mammography for a general population of 15% (RR 0.85 95% CI 0.78 to 0.93). In this review update, the breast cancer mortality relative risk reduction based on RCT data remained the same, and absolute effects by age decade over 10 years were 0.27 fewer deaths per 1,000 in those aged 40 to 49; 0.50 fewer deaths per 1,000 in those aged 50 to 59; 0.65 fewer deaths per 1,000 in those aged 60 to 69; and 0.92 fewer deaths per 1,000 in those aged 70 to 74. For observational data, the relative mortality risk reduction ranged from 29% to 62%. Absolute effects from breast cancer mortality over 10 years ranged from 0.79 to 0.94 fewer deaths per 1,000 in those aged 40 to 49; 1.45 to 1.72 fewer deaths per 1,000 in those aged 50 to 59; 1.89 to 2.24 fewer deaths per 1,000 in those aged 60 to 69; and 2.68 to 3.17 fewer deaths per 1,000 in those aged 70 to 74. For all-cause mortality, RCT data from the prior review reported a non-significant relative reduction in the risk of all-cause mortality of screening mammography for a general population of 1% (RR 0.99, 95% CI 0.98 to 1.00). In this review update, the absolute effects for all-cause mortality over 10 years by age decade were 0.13 fewer deaths per 1,000 in those aged 40 to 49; 0.31 fewer deaths per 1,000 in those aged 50 to 59; 0.71 fewer deaths per 1,000 in those aged 60 to 69; and 1.41 fewer deaths per 1,000 in those aged 70 to 74. No observational data were found for all-cause mortality. For overdiagnosis, this review update found the absolute effects for RCT data (range of follow-up between 9 and 15 years) to be 1.95 more invasive and in situ cancers per 1,000, or 1 more invasive cancer per 1,000, for those aged 40 to 49 and 1.93 more invasive and in situ cancers per 1,000, or 1.18 more invasive cancers per 1,000, for those aged 50 to 59. A sensitivity analysis removing high risk of bias studies found 1.57 more invasive and in situ cancers, or 0.49 more invasive cancers, per 1,000 for those aged 40 to 49 and 3.95 more invasive and in situ cancers per 1,000, or 2.81 more invasive cancers per 1,000, in those aged 50 to 59. For observational data, one report (follow-up for 13 years) found 0.34 more invasive and in situ cancers per 1,000 in those aged 50 to 69. Overall, the GRADE certainty of evidence was assessed as low or very low, suggesting that the evidence is very uncertain about the effect of screening for breast cancer on the outcomes evaluated in this review. ConclusionsThis systematic review update did not identify any new trials comparing breast cancer screening to no screening. Although 26 new observational studies were identified, the overall quality of evidence remains generally low or very low. Future research initiatives should prioritize studying screening in higher risk populations such as those from different ages, racial or ethnic groups, with dense breasts, or family history. RegistrationProtocol available on the Open Science Framework: https://osf.io/xngsu/

Human papillomavirus (HPV) testing is the primary method for cervical cancer screening, and a negative HPV test is associated with a very low subsequent risk of invasive cancer. Nevertheless, a small number of cervical cancers are diagnosed following an HPV-negative testing result, posing challenges within HPV-based screening pathways. Using nationwide Swedish registry data of HPV testing, we identified women diagnosed with invasive cervical cancer between 2019 and 2024 and reconstructed HPV testing histories from the National Cervical Screening Registry (NKCx). The most recent HPV test prior to diagnosis was defined as the index test, and longitudinal HPV testing trajectories were classified among women with an HPV-negative index test. Of 3,000 women diagnosed with invasive cancer, 243 (8.1%) had an HPV-negative index test. These women were older at diagnosis and more frequently diagnosed at advanced stages compared with women with an HPV-positive index test. Most HPV-negative index tests (66.3%) were performed in the peri-diagnostic period (+/- 30 days). Among women with an HPV-negative index test, 52.7% (128/243) had no prior HPV testing recorded, while the remainder had consistently HPV-negative histories (33.3%, 83/243) or evidence of prior HPV positivity before the index negative test (14%, 32/243). Possible recurrent HPV positivity following an intervening negative test was rare (0.4%, 1/243). HPV-negative screening results preceding invasive cancer reflect heterogeneous screening histories and cannot be explained solely by test failure. Findings highlighting the importance of reaching women earlier in screening programs and show that fluctuating HPV detectability is rare. Novelty and impactHPV-negative results preceding cervical cancer are interpreted as test failures, yet underlying screening histories have not been systematically described. Using nationwide registry data, this study reconstructs HPV testing trajectories. HPV-negative cancers occur in older women with limited screening and advanced-stage disease, but also in women with consistently HPV-negative or previously HPV-positive histories. Findings indicate that such cancers cannot be explained solely by test failure and highlight the importance of earlier screening and rare HPV fluctuation.

BackgroundBreast cancer is comprised of distinct molecular subtypes. Studies have reported differences in risk factor associations with breast cancer subtypes, especially by tumor estrogen receptor (ER) status, but their consistency across racial and ethnic populations has not been comprehensively evaluated. MethodsWe conducted a qualitative, scoping literature review using the Preferred Reporting Items for Systematic Reviews and Meta-analysis, extension for Scoping Reviews to investigate consistencies in associations between 18 breast cancer risk factors (reproductive, anthropometric, lifestyle, and medical history) and risk of ER-defined subtypes in women who self-identify as Asian, Black or African American, Hispanic or Latina, or White. We reviewed publications between January 1, 1990 and July 1, 2022. Etiologic heterogeneity evidence (convincing, suggestive, none, or inconclusive) was determined by expert consensus. ResultsPublications per risk factor ranged from 14 (benign breast disease history) to 66 (parity). Publications were most abundant for White women, followed by Asian, Black or African American, and Hispanic or Latina women. Etiologic heterogeneity evidence was strongest for parity, followed by age at first birth, post-menopausal BMI, oral contraceptive use, and estrogen-only and combined menopausal hormone therapy. Evidence was limited for other risk factors. Findings were consistent across racial and ethnic groups, although the strength of evidence varied. ConclusionThe literature supports etiologic heterogeneity by ER for some established risk factors that are consistent across race and ethnicity groups. However, in non-White populations evidence is limited. Larger, more comparable data in diverse populations is needed to better characterize breast cancer etiologic heterogeneity.

Recurrence scores based on a 21-gene assay are clinically useful for predicting prognosis and chemotherapy benefit in postmenopausal node-positive breast cancer patients, but its performance in premenopausal patients is inconsistent. Here, we evaluated Ataraxis Breast RISK (ATX), an AI test that predicts recurrence risk, and compared it with the genomic assay. ATX identified high risk patients misclassified as low risk by the genomic assay and therefore may refine selection of patients for adjuvant chemotherapy.

BackgroundThe management of residual axillary disease after neoadjuvant therapy (NAT) remains controversial, as current recommendations often treat ypN1 breast cancer as a homogeneous entity despite potential prognostic heterogeneity. Evidence supporting uniform axillary surgical strategies across different levels of residual nodal burden is limited. We investigated whether survival associations related to axillary surgical evaluation differ according to residual nodal burden in ypN1 disease, using an adjuvant cohort to validate a SEER-based proxy for surgical extent. MethodsPatients with 1-3 positive lymph nodes were identified in the SEER database (2000-2022) and stratified into neoadjuvant (NAT; n=30,560) and adjuvant (AT; n=197,586) cohorts. Axillary surgical evaluation was categorized as limited (2-3 examined nodes) or extensive ([&ge;]10 examined nodes). Survival was analyzed using Kaplan-Meier methods and log-logistic accelerated failure-time models, adjusted with inverse probability of treatment weighting. ResultsIn the ypN1 cohort, limited axillary evaluation was not associated with inferior overall survival among patients with a single residual positive node (IPTW-adjusted HR: 1.15, p=0.134; time ratio [TR]: 0.86, p=0.184). In contrast, limited evaluation was associated with worse survival in patients with two positive nodes (HR: 1.70, 95%CI 1.54-1.87; TR: 0.58, 95%CI 0.53-0.64). The findings were similar when using breast cancer-specific survival as the endpoint. ConclusionsSurvival associations related to axillary surgical evaluation after NAT vary according to residual nodal burden. Axillary de-escalation appears feasible in patients with a single residual positive node but cannot be extrapolated to those with multiple residual nodes, underscoring heterogeneity within ypN1 disease.

Sentinel lymph node biopsy (SLNB) is increasingly omitted in early-stage breast cancer, often prompting whole-breast irradiation (WBI). We evaluated partial-breast irradiation (PBI) without axillary surgery among 78 clinically node-negative patients (median age 75) treated from 2014 to 2022. After 53-month median follow-up, no ipsilateral, regional, or distant recurrences occurred. These results demonstrate excellent outcomes and suggest PBI is a feasible, safe alternative to WBI when SLNB is omitted.

Background: Early breast cancer detection remains central to improving clinical outcomes, yet conventional screening pathways, particularly mammography, have recognized limitations in sensitivity, specificity, and performance in dense breast tissue. Circulating microRNAs (miRNAs) have emerged as promising minimally invasive biomarkers, while artificial intelligence and machine learning (AI/ML) offer powerful tools for identifying diagnostically relevant multi-marker patterns within complex biomarker datasets. This systematic review and meta-analysis evaluated the diagnostic performance of AI/ML-based circulating miRNA signatures for early breast cancer detection. Methods: A systematic search of PubMed/MEDLINE, Scopus, and Web of Science Core Collection was conducted from database inception to 31 December 2025. Studies were eligible if they were original human investigations evaluating circulating miRNAs using an AI/ML-based diagnostic model for breast cancer detection and reporting extractable diagnostic performance metrics. Study selection followed PRISMA 2020 and PRISMA-DTA guidance. Methodological quality was assessed using QUADAS 2. Pooled sensitivity and specificity were synthesized using a bivariate random-effects model, and overall diagnostic performance was summarized using a hierarchical summary receiver operating characteristic framework. Results: Seven studies met the inclusion criteria for qualitative synthesis, with eligible studies contributing to the quantitative analysis depending on data availability. Across the pooled analysis, AI/ML-based circulating miRNA models demonstrated good overall diagnostic performance, with a pooled AUC of 0.905 (95% CI: 0.890 to 0.921), pooled sensitivity of 81.3% (95% CI: 76.8% to 85.2%), and pooled specificity of 87.0% (95% CI: 82.4% to 90.7%). Heterogeneity was moderate for AUC (I2 = 42.3%) and sensitivity (I2 = 38.7%) and low for specificity (I2 = 28.4%). Risk-of-bias assessment showed overall low-to-moderate methodological concern, with patient selection representing the most variable domain. Deeks funnel plot asymmetry test showed no significant evidence of publication bias (p = 0.34). Conclusions: AI/ML based circulating miRNA signatures show promising diagnostic accuracy for early breast cancer detection and may have value as non invasive adjunctive tools within imaging supported diagnostic pathways. However, the evidence base remains limited by methodological heterogeneity, variable validation rigor, and the predominance of retrospective case control designs. Prospective, standardized, and externally validated studies are needed before routine clinical implementation can be justified.

ObjectiveTo quantify the impact and accuracy of different screening approaches for cervical cancer, including liquid based cytology (LBC), molecular testing for human papillomavirus (HPV) infection, and their combinations via parallel co-testing and sequential triage. The secondary goal was to predict the effect of differing coverage rates of HPV vaccination on the performance of screening tests and in the interpretation of their results. DesignModelling study. Main outcomes measuredDifferent screening modalities were compared in terms of number of cases of Cervical intra-epithelial neoplasia (CIN) grade 2 and 3 detected and missed, as well as the number of false positives leading to excess colposcopy, and number of tests required to achieve a given level of accuracy. The positive predictive value (PPV) and negative predictive value (NPV) of different modalities were simulated under varying levels of HPV vaccination. ResultsThe model predicted that in a typical population, primary LBC screening misses 4.9 (95% Confidence Interval (CI) 3.5-CIN 2 / 3 cases per 1000 women, and results in 95 (95% CI: 93-97%) false positives leading to excess colposcopy. For primary HPV testing, 2.0 (95% CI: 1.9-2.1) cases were missed per 1000 women, with 99 (95% CI: 98-101) excess colposcopies undertaken. Co-testing markedly reduced missed cases to 0.5 (95% CI: 0.3-0.7) per 1000 women, but at the cost of dramatically increasing excess colposcopy referral to 184 per 1000 women (95% CI: 182-188). Conversely, triage testing with reflex screening substantially reduced excess colposcopy to 9.6 cases per 1000 women (95% CI: 9.3 - 10) but at the cost of missing more cases (6.4 per 1000 women, 95% CI: 5.1 - 8.0). Over a life-time of screening, women who always attend annual and 3-year co-testing were predicted to have a virtually 100% chance of falsely detecting a CIN 2 / 3 case, while 5 year co-testing has a 93.8% chance of a false positive over screening life-time. For annual, 3 year, and 5 year triage testing (either LBC with HPV reflex or vice versa), lifetime risk of a false positive is 35.1%, 13.4%, and 8.3% respectively. HPV vaccination rates adversely impact the PPV, while increasing the NPV of various screening modalities. Results of this work indicate that as HPV vaccination rates increase, HPV based screening approaches result in fewer unnecessary colposcopies than LBC approaches. ConclusionThe clinical relevance of cervical cancer screening is crucially dependent upon the prevalence of cervical dysplasia and/or HPV infection or vaccination in a given population, as well as the sensitivity and specificity of various modalities. Although screening is life-saving, false negatives and positives will occur, and over-testing may cause significant harm, including potential over-treatment.

IntroductionWomen with recent parity are at increased short-term breast cancer (BC) risk and face a worse prognosis. The effect of parity on response to neoadjuvant chemotherapy (NAC) is unstudied, and the influence of inherited susceptibility on parity-related short-term risk remains unclear. MethodsWe conducted a retrospective case-cohort study among women aged [&le;]50 with non-metastatic BC diagnosed between 2010 and 2020 who underwent genetic testing and were treated at Northwestern Medicine. Associations between NAC response and recency of parity were evaluated using multivariate logistic regression, stratified by tumor biologic subtypes. Relationships between germline mutations, recency of parity, and BC were explored via multi-state modeling and linear regression. ResultsAmong 1,080 eligible women, 231 received NAC. Treatment response was poorer in parous women with triple negative tumors compared to nullipara, regardless of the recency of parity (P<0.03). Among 122 women (11.3%) with detectable pathogenic mutations, adjusted analyses with both modeling approaches revealed no indications that BRCA1/2 carriers had an increased hazard of BC diagnosis in the decade following recent parity, compared to nulliparous mutation carriers. For BRCA2 and PALB2 carriers, breast cancer diagnosis occurred less frequently in the post-partum intervals. ConclusionWe observed a poor response to NAC in parous TNBC patients compared to nullipara; effects of immunotherapy-based regimens deserve evaluation in the context of parity. Post-partum BC occurrence is not increased in BRCA1/2 carriers; effects of rarer susceptibility genes may differ. These important effects of parity on BC in young women and those at genetic risk warrant larger prospective studies.

PurposeIdentify optimum strategies to recruit Latin American and Hispanic women into genetic studies of breast cancer. We evaluated hospital and community-based recruitment strategies. MethodsWe used targeted gene sequencing to identify mutations in DNA from unselected Hispanic breast cancer cases from community and hospital-based recruitment in the US and Guatemala. ResultsWe recruited 287 Hispanic US women, 38 (13%) from community-based and 249 (87%) from hospital-based strategies. In addition, we ascertained 801 Guatemalan women using hospital-based recruitment. In our experience, a hospital-based approach was more efficient than community-based recruitment. In this study, we sequenced 103 US and 137 Guatemalan women and found 11 and 10 pathogenic variants, respectively. The most frequently mutated genes were BRCA1, BRCA2, CHEK2, and ATM. In addition, an analysis of 287 US Hispanic patients with pathology reports showed a significantly higher percentage of triple-negative disease in patients with pathogenic mutations (41% vs. 15%). Finally, an analysis of mammography usage in 801 Guatemalan patients found reduced screening in women with a lower socioeconomic status (P<0.001). ConclusionsGuatemalan and US Hispanic women have rates of hereditary breast cancer mutations similar to other populations and are more likely to have early age at diagnosis, a family history, and a more aggressive disease. Patient recruitment was higher using hospital-based versus community enrollment. This data supports genetic testing in breast cancer patients to reduce breast cancer mortality in Hispanic women.

Background. Breast cancer treatment depends on histopathological features, such as grade and receptor-defined subtype; however, specialist pathologist access is constrained when the workforce is limited. Commercial multimodal large language models (MLLMs) accept hematoxylin and eosin (H&E) image tiles through paid interfaces without local hardware or fine-tuning. However, prior pathology evaluations addressed only coarse tasks. Whether they reach treatment-determining accuracy and whether vendors agree remain unclear. Methods. We aimed to evaluate three vendor-designated flagship MLLMs (Claude Sonnet 4.6, Gemini 2.5 Pro, GPT-5.5) in 427 invasive breast cancer cases. Each case went to all three with identical H&E tiles and prompts, and the subtype was inferred in the second call. The reference was an institutional sign-out report of an immunohistochemistry-derived subtype. We calculated the concordance, sensitivity, specificity, Cohen's kappa, and pairwise McNemar and Bowker tests. Findings. Claude ranked highest by raw histologic-type concordance but lowest by kappa, classifying all 23 lobular and seven micropapillary carcinomas as invasive breast carcinoma of no special type. The models anchored the Nottingham grade to three modal grades. None of the models reliably identified human epidermal growth factor receptor 2-positive disease. The failure direction was vendor-specific: Claude and GPT-5.5 were under-detected, whereas Gemini was over-called. Twelve prompt variants (4,056 calls) did not recover sensitivity. Interpretation. No current commercial MLLM reaches deployment-ready accuracy for any treatment-determining feature of breast pathology. As each vendor fails in its own fixed direction, changing vendors alters the type of error rather than removing it; therefore, the value of these models is assistive rather than autonomous. At USD 0.20-0.50 per case, they may serve as supervised draft generators that leave the diagnosis with the pathologist.

BackgroundInvasive lobular breast cancer (ILC) is the most commonly diagnosed special histological subtype of breast cancer (BC). Metastatic ILC (mILC) is less sensitive to FDG-PET imaging and often metastasizes to unusual sites --peritoneum, gastrointestinal (GI) tract, ovaries, urinary tract, and orbit--which may go unrecognized after a long disease-free interval. Some metastatic sites cause nonspecific symptoms, like abdominal/epigastric pain, with numerous published case reports of mILC misdiagnosed as gastric cancer. These atypical BC metastatic sites may lead to late and/or misdiagnosis, thereby delaying effective treatments. ObjectiveWe developed a patient survey to investigate the patient-reported prevalence of delayed diagnosis or misdiagnosis of mILC and their potential impact upon treatment outcomes. MethodsA 45-question survey was developed and piloted with breast cancer researchers, clinical oncologists, and patient advocates. This IRB-approved survey was then distributed to patients with ILC. Analyses including data QC and visualization were conducted in R using descriptive statistics. Incomplete or inconsistent responses were excluded, and summary statistics were stratified by four common mILC sites to highlight subgroup differences. Results525 patient surveys were completed, with 450 patients diagnosed with ILC, and of those 321 diagnosed with mILC. For those with mILC, 33.3% (n=107) were diagnosed with de novo mILC at initial presentation. Of the patients diagnosed with mILC, 32.1% (n=103) presented with other medical conditions at diagnosis. Misdiagnosis was reported by 26.2% (n=84) of patients with mILC, and of these cases, 31% (n=26) had [&ge;]2 misdiagnoses. The top 5 misdiagnoses were bone-related condition (24.7%), benign breast condition (23.4%), another type of BC (7.8%), diagnostic delay (7.8%), and menopause related (5.2%). 44.5% of patients waited [&ge;]1 year for an accurate diagnosis. 49 patients were treated for their misdiagnosis, and 6 received incorrect cancer treatments. The most frequently reported contributors to delayed or misdiagnosis were inconclusive imaging, providers lack of ILC knowledge, and initial misdiagnosis. Of the 321 patients with mILC, 138 (42.9%) reported symptoms before diagnosis; the most common were back pain (16.5%), fatigue/malaise (14.9%), GI symptoms (11.8%), bloating (8.4%), and weight loss (8.1%). Although 40% of patients reported having a mammogram at the time of their initial misdiagnosis, ILC was detected in only 20.5% (24/116) of these cases, and mammography detected only 5 (25%) of the 20 de novo mILC cases. Patients reported additional diagnostic testing within 1-3 months of their initial mammogram, includingbiopsy, ultrasound (US), and MRI. 47.9% of patients were in active BC surveillance after curative intent therapy at the time of their mILC diagnosis; however, no statistical difference was seen in time to diagnosis versus those patients not under surveillance. ConclusionOur survey results underscore the urgent need to improve diagnostic strategies for mILC. Addressing delays and diagnostic errors in mILC is critical to optimizing treatment strategies and improving patient outcomes.

IMPORTANCEIn young-onset breast cancer, a diagnosis within 5-10 years of childbirth associates with increased mortality. Women with germline BRCA1/2 pathogenic variants (PVs) are more likely to be diagnosed with breast cancer at younger ages, but the impact of childbirth on mortality is unknown. OBJECTIVEDetermine whether time between recent childbirth and breast cancer diagnosis impacts mortality among young-onset breast cancer patients with germline BRCA1/2 PVs. DESIGN, SETTING, AND PARTICIPANTSThis prospective cohort study includes 903 women with germline BRCA1/2 PVs diagnosed with stage I-III breast cancer at [&le;]45 years of age, between 1950-2021 in the UK. MAIN OUTCOMES AND MEASURESThe primary outcome is all-cause mortality, censored at 20 years post-diagnosis. The primary exposure is time between most recent childbirth and breast cancer diagnosis, with recent childbirth defined as >0-<10 years post childbirth (n=419)], further delineated to >0-<5 years (n=228) and 5-<10 years (n=191). Mortality of nulliparous cases (n=224) was compared to the recent postpartum groups and the [&ge;]10 years postpartum (n=260) group. Cox proportional hazards regression analyses were adjusted for patient age, tumor stage, further stratified by tumor estrogen receptor (ER) and BRCA gene status. RESULTSFor all BRCA PV carriers, increased all-cause mortality was observed in women diagnosed >0-<10 years postpartum, compared to nulliparous and [&ge;]10 years groups, demonstrating the transient duration of postpartum risk. Risk of mortality was greater for ER-positive cases in the >0-<5 group [HR=2.35 (95% CI, 1.02-5.42)] and ER-negative cases in the 5-<10 group [HR=3.12 (95% CI, 1.22-7.97)] compared to the nulliparous group. Delineated by BRCA1 or BRCA2, mortality in the 5-<10 group was significantly increased, but only for BRCA1 carriers [HR=2.03 (95% CI, 1.15-3.58)]. CONCLUSIONS AND RELEVANCEYoung-onset breast cancer with germline BRCA PVs confers increased risk for all-cause mortality if diagnosed within 10 years of childbirth, with risk highest for ER+ cases at >0-<5 years postpartum, and for ER-cases at 5-<10 years postpartum. BRCA1 carriers are at highest risk for poor prognosis when diagnosed at 5-10 years postpartum. No such associations were observed for BRCA2 carriers. These results should inform genetic counseling, prevention, and treatment strategies for BRCA PV carriers. Key PointsO_ST_ABSQuestionC_ST_ABSIs a postpartum diagnosis an independent risk factor for mortality among young-onset breast cancer patients with germline BRCA1/2 PVs? FindingsA diagnosis <10 years postpartum associates with higher risk of mortality compared to nulliparous and [&ge;]10 years postpartum cases. Peak risk after childbirth varies for ER-positive (>0-<5 years) vs. ER-negative cases (5-<10 years). BRCA1 carriers had peak risk of mortality 5-10 years postpartum, with no associations observed for BRCA2 carriers. MeaningA breast cancer diagnosis within 10 years of childbirth independently associates with increased risk for mortality in patients with germline BRCA1/2 PVs, especially for carriers of BRCA1 PVs.

BackgroundPREDICT Breast v3 is the latest updated prognostication tool, developed from the breast cancer registry of approximately 35,000 women diagnosed between 2000 and 2018 in the United Kingdom. However, its performance in the United States (US) population is unknown. This study aims to validate PREDICT Breast v3 using newly released Surveillance, Epidemiology, and End Results (SEER) outcome data for US breast cancer patients and to address potential health disparities. MethodsOver 860,000 female patients diagnosed between 2000 and 2018 with primary breast cancer and followed for at least 10 years were selected from the SEER database. Predicted and observed 10- and 15-year breast cancer-specific survival outcomes were compared for the overall cohort, stratified by estrogen receptor (ER) status, and predefined subgroups. Discriminatory accuracy was determined through the area under the receiver-operator curves (AUC). ResultsPREDICT Breast v3 demonstrated good calibration and discrimination for long-term breast cancer-specific mortality. It provided accurate mortality estimates (within a {+/-}10% error range) across the entire US population for 10-year (-8% in ER-positive and 4% in ER-negative patients) and 15-year (-3 % in ER-positive and 5% in ER-negative patients) all-cause mortality, for both ER statuses. The model also showed good performance for 10- and 15-year all-cause mortality across the U.S. population, with AUC of 0.769 and 0.793 for ER-positive breast cancer as well as AUC of 0.738 and 0.746 for ER-negative breast cancer. However, recalibration is needed for specific groups, such as non-Hispanic Asian and non-Hispanic Black patients with ER-negative status. ConclusionsPREDICT v3 accurately predicts 10- and 15-year overall survival in contemporary US breast cancer patients. Future work should focus on promoting equitable care by addressing disparities that are observed in predictive tools.

BackgroundBreast cancer prognosis changes over time in complex ways depending on individual risk factors. This study aimed to analyze how breast cancer outcomes in New Zealand women change over time and identify features associated with breast cancer specific survival and locoregional recurrence across different receptor subtypes. MethodsA retrospective cohort study was conducted using data from Te R[e]hita Mate [U]taetae (Breast Cancer Foundation National Register) on 21,574 women diagnosed with invasive breast cancer between 2000-2019. We applied k-medians survival clustering, landmark analysis, and piecewise Cox regression to identify time-specific risk patterns and prognostic features. ResultsSurvival improved significantly for women diagnosed more recently. Triple-negative breast cancer had the poorest 5-year breast cancer specific survival but demonstrated better outcomes for women surviving beyond this period. In contrast, ER+/HER2-tumors, associated with favorable short-term outcomes, showed the highest risk of late recurrence and breast cancer mortality beyond 10 years. Younger age at diagnosis ([&le;]44 years) was associated with increased recurrence risks, especially for ER-/HER2+ tumors. Radiation therapy reduced early LRR across subtypes. Tumor grade was inversely associated with late recurrence, while stage 2 disease in ER+ tumors markedly elevated late recurrence odds compared to stage 1. ConclusionsThis study demonstrates the dynamic nature of breast cancer prognosis, with key findings emphasizing the time-dependent shifts in risk across receptor subtypes. These findings underscore the importance of personalized, receptor-specific follow-up strategies, including extended monitoring for subgroups at heightened long-term risk.