Classifying interval cancers.

A A Duncan; M G Wallis

doi:10.1016/s0009-9260(05)83218-0

Home/Archive/PubMed

PubMed · 7489628

Classifying interval cancers.

A A Duncan·M G Wallis

View original source ↗Export metadata ↓

Abstract

No definitive way of classifying interval breast cancers (cancers presenting between screening rounds) has been determined, yet their number and classification forms one of the quality assurance (QA) standards for the National Health Service Breast Screening Programme (NHSBSP). This study was undertaken to identify how different screening centres undertake this process, and to compare the classifications obtained when a test set of mammograms was reviewed using three different methods. A questionnaire was sent to the 17 Regional UK breast screening QA centres. Twelve (80%) of the 15 centres completing the questionnaire had a formal method for reviewing interval cancers. Five of these (33%) attempted to simulate screening by mixing the interval cancers with other screening films. In 11 (73%) centres, a group (size range 3-14) of radiologists was involved. In a simulated film viewing exercise we assessed whether different methods of classification would alter the number of interval cancers classified as false negative (where an abnormality suspicious of malignancy can be identified on review of the original screening films). Six radiologists reviewed a set of 50 interval cancers by three different methods: independent reading of the interval cancers mixed with screening mammograms; independent reading of the interval cancers on their own; and finally a consensus opinion of the interval cancers alone using the films taken at diagnosis. No discussion or review of these cases had taken place prior to the study. The number of interval cancers classified as false negative increased by 10% when they were reviewed in isolation compared to review among 'normal' screening films. The false negative rate varied widely (4% to 56% P < 0.01) depending on the criteria required to fulfil a false negative interval cancer classification--whether only one, a majority, or all of the radiologists were required to see the abnormality on the original screening films. In order for inter-unit and regional comparisons to be useful, a standard review method of mixing the interval cancer films with normal screening films and using a consensus opinion by a minimum of three external reviewers is suggested.

Explore related subjects

Breast Neoplasms False Positive Reactions Female Humans Mammography Mass Screening Observer Variation Journal Article Multicenter Study

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

A A Duncan, M G Wallis. 1995. Classifying interval cancers.. https://doi.org/10.1016/s0009-9260(05)83218-0

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related citations

PubMed2026

Hypernetwork-guided fusion with intra-class MixUp for breast cancer subtyping.

Accurate breast cancer subtyping guides treatment selection, yet histopathology captures morphology without molecular state, while genomic profiling captures molecular signatures without spatial context. Existing fusion methods rely on concatenation, or on attention applied only after each modality is encoded independently. This work identifies a scale-dependent asymmetry in the direction of cross-modal conditioning: the direction that performs best under limited samples is not the one that holds at scale, and the reversal is traced to the capacity of the modulation pathway rather than to the fusion principle. The comparison is carried out within a hypernetwork-guided framework in which an auxiliary network maps one modality to conditioning parameters that modulate the other's feature representation, shaping features at the parametric level rather than the decision stage; modulation is patient-specific rather than patch-specific. Both directions are instantiated-gene-to-image (HyperG2I) and image-to-gene (HyperI2G) - and trained under a label-aware MixUp strategy that interpolates within-class samples across both modalities, preserving the hard binary labels clinical decisions require. The framework is evaluated on two paired TCGA-BRCA cohorts-one limited-sample, one independently assembled at scale-under a single protocol spanning two whole-slide representations, multiple visual backbones, and both conditioning directions. On the limited-sample cohort, gene-to-image conditioning at its optimal augmentation setting exceeds early fusion and both unimodal baselines, giving the highest recall on the aggressive Basal/HER2 class of any configuration evaluated, and an ablation favours intra-class over inter-class mixing. At scale this ordering does not hold: image-to-gene conditioning sustains its performance whereas gene-to-image does not, recovering only partially under the full tissue bag and isolating the capacity of the modulation pathway as the binding constraint. Direction and capacity of cross-modal conditioning, rather than fusion depth alone, therefore govern how such frameworks scale.

Breast Neoplasms↗

PubMed2025

Metagenomic analyses reveal E. coli-derived siderophores as potential signatures for breast cancer.

BACKGROUND: Breast cancer remains a leading cause of cancer-related mortality in women. Recent evidence implicates the gut microbiome and metabolites in breast cancer pathogenesis. This study explores associations between gut microbial species, their predicted metabolites, and breast cancer to uncover potential mechanistic insights. METHODS: Comprehensive metagenomic analyses were conducted on the gut microbiome of pre- and postmenopausal breast cancer patients, where microbial species were profiled through AMPHORA2 and metabolites were predicted through antiSMASH. Multivariate association analysis was used to identify significant associations between specific microbial species, predicted metabolites, and breast cancer status. A custom ensemble machine learning classifier was developed to classify pre- and postmenopausal breast cancer cases and controls based on microbial and predicted metabolite features. Additionally, a synthetic microbiome dataset was generated through MIDASim to validate the reproducibility of the ML results. Using our results, we explored the underlying dynamics of identified taxa and metabolite in breast cancer through literature and statistical support. RESULTS: Our analysis identified 471 microbial species and predicted 40 key metabolites in the metagenomic data. Multivariate analysis identified significant positive associations (p-value < 0.05) of E. coli, siderophore, and thiopeptide with breast cancer. The custom ensemble model achieved accuracy and AUC as high as 78% and 90%, respectively, in classifying pre- and postmenopausal cases and controls. The high-ranking features i.e., E. coli, siderophore, and thiopeptide were consistent with the results of the multivariate association analysis, thereby substantiating their biological significance. Using these findings, we propose a mechanistic model in which E. coli secretes siderophores under iron-limited conditions in breast cancer patients, for iron sequestration from the host, which can potentially promote angiogenesis and tumor progression. CONCLUSION: Our findings suggest that microbial iron acquisition mechanisms may play a critical role in breast cancer pathophysiology. Functional validation of these mechanisms is needed to assess therapeutic potential. This study highlights gut microbiota and their metabolites as promising targets for breast cancer research and intervention.

Breast Neoplasms↗

PubMed2025

Foundation model based multimodal transformer framework for survival analysis in HER2 stratified breast cancer.

Objective. To improve survival prediction for HER2-positive breast cancer by integrating histopathological, molecular, and clinical data using a multimodal transformer framework.Approach. We propose a multimodal transformer framework for breast cancer survival prediction using HER2 stratified (SurvMBC), a foundation model-enhanced architecture that fuses three data modalities: whole-slide images, clinical narratives, and molecular features. Tumor microenvironment features are extracted using a pathology language and image pre-training (PLIP), clinical narratives are processed with BioBERT, and miRNA expression plus DNA methylation data are embedded using Gen2Vec. These representations are integrated through a cross-modal transformer with attention mechanisms for survival prediction.Main results. The model was evaluated on 1,095 HER2-positive breast cancer patients from The Cancer Genome Atlas. SurvMBC achieved a concordance index (C-index) of 0.857 (95% CI: 0.834, 0.880), a low integrated Brier score, and a strong inverse negative binomial log-likelihood. Risk stratification based on model outputs significantly separated high- and low-risk groups (log-rankp< 0.01) and showed strong associations with tumor stage, grade, and hormone receptor status (allp< 0.05).Significance. SurvMBC demonstrates the effectiveness of multimodal fusion in addressing tumor heterogeneity and improving prognostic accuracy. The attention-based integration enables context-aware learning of survival-relevant features across modalities, supporting individualized risk stratification and risk-adaptive treatment planning for HER2 stratified breast cancer patients.

Breast Neoplasms↗