Dataset Information

Breast Cancer Mode of Detection in a Population-Based Cohort.

ABSTRACT:

Objective

To evaluate how breast cancers come to clinical attention (mode of detection [MOD]) in a population-based cohort, determine the relative frequency of different MODs, and characterize patient and tumor characteristics associated with MOD.

Patients and methods

We used the Rochester Epidemiology Project to identify women ages 40 to 75 years with a first-time diagnosis of breast cancer from May 9, 2017, to May 9, 2019 (n=500) in a 9-county region in Minnesota. We conducted a retrospective medical record review to ascertain the relative frequency of MODs, evaluating differences between screening mammography vs all other MODs by breast density and cancer characteristics. Multiple logistic regression was conducted to examine the likelihood of MOD for breast density and stage of disease.

Results

In our population-based cohort, 162 of 500 breast cancers (32.4%) were detected by MODs other than screening mammography, including 124 (24.8%) self-detected cancers. Compared with women with mammography-detected cancers, those with MODs other than screening mammography were more frequently younger than 50 years of age (P=.004) and had higher-grade tumors (P=.007), higher number of positive lymph nodes (P<.001), and larger tumor size (P<.001). Relative to women with mammography-detected cancers, those with MODs other than screening mammography were more likely to have dense breasts (odds ratio, 1.87; 95% CI, 1.20 to 2.92; P=.006) and advanced cancer at diagnosis (odds ratio, 3.58; 95% CI, 2.29 to 5.58; P<.001).

Conclusion

One-third of all breast cancers in this population were detected by MODs other than screening mammography. Increased likelihood of nonmammographic MODs was observed among women with dense breasts and advanced cancer.

SUBMITTER: Basappa SN

PROVIDER: S-EPMC9907001 | biostudies-literature | 2023 Feb

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Breast Cancer Mode of Detection in a Population-Based Cohort.

Basappa Susanna N SN Finney Rutten Lila J LJ Hruska Carrie B CB Olson Janet E JE Jacobson Debra J DJ Rhodes Deborah J DJ

Mayo Clinic proceedings 20230201 2

<h4>Objective</h4>To evaluate how breast cancers come to clinical attention (mode of detection [MOD]) in a population-based cohort, determine the relative frequency of different MODs, and characterize patient and tumor characteristics associated with MOD.<h4>Patients and methods</h4>We used the Rochester Epidemiology Project to identify women ages 40 to 75 years with a first-time diagnosis of breast cancer from May 9, 2017, to May 9, 2019 (n=500) in a 9-county region in Minnesota. We conducted a ...[more]

PMID: 36737116

Similar Datasets

Project description:BackgroundArtificial intelligence (AI) has been proposed to reduce false-positive screens, increase cancer detection rates (CDRs), and address resourcing challenges faced by breast screening programs. We compared the accuracy of AI versus radiologists in real-world population breast cancer screening, and estimated potential impacts on CDR, recall and workload for simulated AI-radiologist reading.MethodsExternal validation of a commercially-available AI algorithm in a retrospective cohort of 108,970 consecutive mammograms from a population-based screening program, with ascertained outcomes (including interval cancers by registry linkage). Area under the ROC curve (AUC), sensitivity and specificity for AI were compared with radiologists who interpreted the screens in practice. CDR and recall were estimated for simulated AI-radiologist reading (with arbitration) and compared with program metrics.FindingsThe AUC for AI was 0.83 compared with 0.93 for radiologists. At a prospective threshold, sensitivity for AI (0.67; 95% CI: 0.64-0.70) was comparable to radiologists (0.68; 95% CI: 0.66-0.71) with lower specificity (0.81 [95% CI: 0.81-0.81] versus 0.97 [95% CI: 0.97-0.97]). Recall rate for AI-radiologist reading (3.14%) was significantly lower than for the BSWA program (3.38%) (-0.25%; 95% CI: -0.31 to -0.18; P < 0.001). CDR was also lower (6.37 versus 6.97 per 1000) (-0.61; 95% CI: -0.77 to -0.44; P < 0.001); however, AI detected interval cancers that were not found by radiologists (0.72 per 1000; 95% CI: 0.57-0.90). AI-radiologist reading increased arbitration but decreased overall screen-reading volume by 41.4% (95% CI: 41.2-41.6).InterpretationReplacement of one radiologist by AI (with arbitration) resulted in lower recall and overall screen-reading volume. There was a small reduction in CDR for AI-radiologist reading. AI detected interval cases that were not identified by radiologists, suggesting potentially higher CDR if radiologists were unblinded to AI findings. These results indicate AI's potential role as a screen-reader of mammograms, but prospective trials are required to determine whether CDR could improve if AI detection was actioned in double-reading with arbitration.FundingNational Breast Cancer Foundation (NBCF), National Health and Medical Research Council (NHMRC).

Project description:BackgroundThe aim of this study was to examine breast density in relation to breast cancer specific survival and to assess if this potential association was modified by mode of detection. An additional aim was to study whether the established association between mode of detection and survival is modified by breast density.MethodsThe study included 619 cases from a prospective cohort, The Malmö Diet and Cancer Study. Breast density estimated qualitatively, was analyzed in relation to breast cancer death, in non-symptomatic and symptomatic women, using Cox regression calculating hazard ratios (HR) with 95% confidence intervals. Adjustments were made in several steps for; diagnostic age, tumour size, axillary lymph node involvement, grade, hormone receptor status, body mass index (baseline), diagnostic period, use of hormone replacement therapy at diagnosis and mode of detection. Detection mode in relation to survival was analyzed stratified for breast density. Differences in HR following different adjustments were analyzed by Freedmans%.ResultsAfter adjustment for age and other prognostic factors, women with dense, as compared to fatty breasts, had an increased risk of breast cancer death, HR 2.56:1.07-6.11, with a statistically significant trend over density categories, p = 0.04. In the stratified analysis, the effect was less pronounced in non-symptomatic women, HR 2.04:0.49-8.49 as compared to symptomatic, HR 3.40:1.06-10.90. In the unadjusted model, symptomatic women had a higher risk of breast cancer death, regardless of breast density. Analyzed by Freedmans%, age, tumour size, lymph nodes, grade, diagnostic period, ER and PgR explained 55.5% of the observed differences in mortality between non-symptomatic and symptomatic cases. Additional adjustment for breast density caused only a minor change.ConclusionsHigh breast density at diagnosis may be associated with decreased breast cancer survival. This association appears to be stronger in women with symptomatic cancers but breast density could not explain differences in survival according to detection mode.

Project description:IntroductionArtiﬁcial intelligence (AI) algorithms for interpreting mammograms have the potential to improve the effectiveness of population breast cancer screening programmes if they can detect cancers, including interval cancers, without contributing substantially to overdiagnosis. Studies suggesting that AI has comparable or greater accuracy than radiologists commonly employ 'enriched' datasets in which cancer prevalence is higher than in population screening. Routine screening outcome metrics (cancer detection and recall rates) cannot be estimated from these datasets, and accuracy estimates may be subject to spectrum bias which limits generalisabilty to real-world screening. We aim to address these limitations by comparing the accuracy of AI and radiologists in a cohort of consecutive of women attending a real-world population breast cancer screening programme.Methods and analysisA retrospective, consecutive cohort of digital mammography screens from 109 000 distinct women was assembled from BreastScreen WA (BSWA), Western Australia's biennial population screening programme, from November 2016 to December 2017. The cohort includes 761 screen-detected and 235 interval cancers. Descriptive characteristics and results of radiologist double-reading will be extracted from BSWA outcomes data collection. Mammograms will be reinterpreted by a commercial AI algorithm (DeepHealth). AI accuracy will be compared with that of radiologist single-reading based on the diﬀerence in the area under the receiver operating characteristic curve. Cancer detection and recall rates for combined AI-radiologist reading will be estimated by pairing the first radiologist read per screen with the AI algorithm, and compared with estimates for radiologist double-reading.Ethics and disseminationThis study has ethical approval from the Women and Newborn Health Service Ethics Committee (EC00350) and the Curtin University Human Research Ethics Committee (HRE2020-0316). Findings will be published in peer-reviewed journals and presented at national and international conferences. Results will also be disseminated to stakeholders in Australian breast cancer screening programmes and policy makers in population screening.

Project description:IntroductionIt is not clear whether high mammographic density (MD) is equally associated with all subtypes of breast cancer (BC). We investigated the association between MD and subsequent BC, considering invasiveness, means of detection, pathologic subtype, and the time elapsed since mammographic exploration and BC diagnosis.MethodsBC cases occurring in the population of women who attended screening from 1997 through 2004 in Navarre, a Spanish region with a fully consolidated screening program, were identified via record linkage with the Navarre Cancer Registry (n = 1,172). Information was extracted from the records of their first attendance at screening in that period. For each case, we randomly selected four controls, matched by screening round, year of birth, and place of residence. Cases were classified according to invasiveness (ductal carcinoma in situ (DCIS) versus invasive tumors), pathologic subtype (considering hormonal receptors and HER2), and type of diagnosis (screen-detected versus interval cases). MD was evaluated by a single, experienced radiologist by using a semiquantitative scale. Data on BC risk factors were obtained by the screening program in the corresponding round. The association between MD and tumor subtype was assessed by using conditional logistic regression.ResultsMD was clearly associated with subsequent BC. The odds ratio (OR) for the highest MD category (MD >75%) compared with the reference category (MD <10%) was similar for DCIS (OR = 3.47; 95% CI = 1.46 to 8.27) and invasive tumors (OR = 2.95; 95% CI = 2.01 to 4.35). The excess risk was particularly high for interval cases (OR = 7.72; 95% CI = 4.02 to 14.81) in comparison with screened detected tumors (OR = 2.17; 95% CI = 1.40 to 3.36). Sensitivity analyses excluding interval cases diagnosed in the first year after MD assessment or immediately after an early recall to screening yielded similar results. No differences were seen regarding pathologic subtypes. The excess risk associated with MD persisted for at least 7 to 8 years after mammographic exploration.ConclusionsOur results confirm that MD is an important risk factor for all types of breast cancer. High breast density strongly increases the risk of developing an interval tumor, and this excess risk is not completely explained by a possible masking effect.

Dataset Information

Breast Cancer Mode of Detection in a Population-Based Cohort.

Objective

Patients and methods

Results

Conclusion

Publications

Breast Cancer Mode of Detection in a Population-Based Cohort.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets