A Deep Learning–Based Decision Support Tool for Precision Risk Assessment of Breast Cancer

2019 ◽  
pp. 1-12 ◽  
Author(s):  
Tiancheng He ◽  
Mamta Puppala ◽  
Chika F. Ezeana ◽  
Yan-siang Huang ◽  
Ping-hsuan Chou ◽  
...  
Keyword(s):  
Breast Cancer ◽  
Risk Assessment ◽  
Deep Learning ◽  
Breast Cancer Risk ◽  
Cancer Risk ◽  
Breast Imaging ◽  
Assessment Model ◽  
Specific Patient ◽  
Mammographic Images ◽  
Abnormal Mammogram

PURPOSE The Breast Imaging Reporting and Data System (BI-RADS) lexicon was developed to standardize mammographic reporting to assess cancer risk and facilitate the decision to biopsy. Because of substantial interobserver variability in the application of the BI-RADS lexicon, the decision to biopsy varies greatly and results in overdiagnosis and excessive biopsies. The false-positive rate from mammograms is estimated to be 7% to approximately 10% overall, but within the BI-RADS 4 category, it is greater than 70%. Therefore, we developed the Breast Cancer Risk Calculator (BRISK) to target a well-characterized and specific patient subgroup (BI-RADS 4) rather than a broad heterogeneous group in assessing breast cancer risk. METHODS BRISK provides a novel precise risk assessment model to reduce overdiagnosis and unnecessary biopsies. It was developed by applying natural language processing and deep learning methods on 5,147 patient records archived in the Houston Methodist systemwide data warehouse from 2006 to May 2015, including imaging and pathology reports, mammographic images, and patient demographics. Key characteristics for BI-RADS 4 patients were collected and computed to output an index measure for biopsy recommendation that is clinically relevant and informative and improves upon the traditional BI-RADS 4 scores. RESULTS For the validation set, we assessed data from 1,247 BI-RADS 4 patients, including mammographic images and medical reports. The BRISK model sensitivity to predict malignancy was 100%, whereas the specificity was 74%. The total accuracy of our implemented model in BRISK was 81%. Overall area under the curve was 0.93. CONCLUSION BRISK for abnormal mammogram uses integrative artificial intelligence technology and has demonstrated high sensitivity in the prediction of malignancy. Prospective evaluation is under way and can lead to improvement in patient-physician engagement in making informed decisions with regard to biopsy.

Preliminary analysis and assessment of breast cancer risk in Japanese women.

2017 ◽  
Vol 35 (15_suppl) ◽  
pp. e13092-e13092
Author(s):  
Michiyo Yamada ◽  
Takashi Ishikawa ◽  
Sadatoshi Sugae ◽  
Kazutaka Narui ◽  
Eiji Arita ◽  
...  
Keyword(s):  
Breast Cancer ◽  
Risk Factors ◽  
Risk Assessment ◽  
Family History ◽  
Breast Cancer Risk ◽  
Cancer Risk ◽  
Japanese Women ◽  
Assessment Model ◽  
Risk Assessment Model ◽  
Brinkman Index

e13092 Background: No comprehensive breast cancer risk assessment model for Japanese women exists. Consequently, we have collected Japanese women’s data to investigate key BC risk factors with an objective of deriving a Japanese-women specific BC risk assessment model. Methods: We conducted a retrospective case-control study (paper-based with postal survey) at 15 institutions during 2014-2015. A survey was distributed to Japanese females aged 20-80 who had BC check-up. All pertinent data of a total of 34 factors including demographic and reproductive factors, social history and eating habits was collected. Cases and controls were divided into three groups respectively, premenopausal (PRE; 20 ≤ age < 45), perimenopausal (PERI; 45 ≤ age ≤ 55) and postmenopausal group (POST; 55 < age ≤ 80). Cases and control variables were compared by t-test, chi-square test and Wilcoxon rank sum test. Preliminary BC risk was calculated by logistic regression analysis. Results: A total of 3975 female Japanese datasets were collected, of which 2494 were complete (all variables present) with 1401 controls and 1093 cases were used. There were 222 cases and 332 controls for PRE, 404 cases and 537 controls for PERI, and 467 and 532 controls for POST. The univariate analysis demonstrated that BMI was significantly higher in cases than in controls in all groups (P < 0.01) as was “number of deliveries” in PRE and POST (P < 0.001) and Brinkman index in PRE and PERI (p = 0.017). Multivariate analysis revealed that BC risk was positively associated with BMI (OR 1.080, 95% CI 1.017–1.148, p = 0.012) in PRE, BMI (OR 1.121, 95% CI 1.072–1.174, p < 0.01) and brinkman index (OR 1.000005, 95% CI 1.000002–1.000008, p < 0.01) in PERI, age (OR 1.054, 95% CI 1.028–1.081, p < 0.010), BMI (OR 1.153, 95% CI 1.076-1.171, p < 0.01) and family history (OR 1.497, 95% CI 1.103–2.033, p = 0.001) in POST, while negatively associated with regular exercise (OR 0.672, 95% CI 0.517–0.873, p = 0.003) in POST. Conclusions: BMI in all groups, in addition, the Brinkman index in PERI and age and family history in POST are BC risk factors. Exercise is a protective risk factor in POST. However, the preliminary results are incomplete and further analysis will be conducted before a full risk assessment model is proposed for Japanese women.

scholarly journals Korean Risk Assessment Model for Breast Cancer Risk Prediction

PLoS ONE ◽  
2013 ◽  
Vol 8 (10) ◽  
pp. e76736 ◽  
Author(s):  
Boyoung Park ◽  
Seung Hyun Ma ◽  
Aesun Shin ◽  
Myung-Chul Chang ◽  
Ji-Yeob Choi ◽  
...  
Keyword(s):  
Breast Cancer ◽  
Risk Assessment ◽  
Breast Cancer Risk ◽  
Cancer Risk ◽  
Risk Prediction ◽  
Assessment Model ◽  
Risk Assessment Model

scholarly journals Breast Cancer Risk Assessment: A Review on Mammography-Based Approaches

2021 ◽  
Vol 7 (6) ◽  
pp. 98
Author(s):  
João Mendes ◽  
Nuno Matela
Keyword(s):  
Breast Cancer ◽  
Risk Assessment ◽  
Deep Learning ◽  
Breast Cancer Risk ◽  
Cancer Risk ◽  
Machine Learning Algorithms ◽  
Cancer Risk Assessment ◽  
Cancer Statistics ◽  
Screening Programs ◽  
Breast Cancer Risk Assessment

Breast cancer affects thousands of women across the world, every year. Methods to predict risk of breast cancer, or to stratify women in different risk levels, could help to achieve an early diagnosis, and consequently a reduction of mortality. This paper aims to review articles that extracted texture features from mammograms and used those features along with machine learning algorithms to assess breast cancer risk. Besides that, deep learning methodologies that aimed for the same goal were also reviewed. In this work, first, a brief introduction to breast cancer statistics and screening programs is presented; after that, research done in the field of breast cancer risk assessment are analyzed, in terms of both methodologies used and results obtained. Finally, considerations about the analyzed papers are conducted. The results of this review allow to conclude that both machine and deep learning methodologies provide promising results in the field of risk analysis, either in a stratification in risk groups, or in a prediction of a risk score. Although promising, future endeavors in this field should consider the possibility of the implementation of the methodology in clinical practice.

Sign in / Sign up

Export Citation Format

Share Document