Dataset Information

Phenotype Discovery and Geographic Disparities of Late-Stage Breast Cancer Diagnosis across U.S. Counties: A Machine Learning Approach.

ABSTRACT:

Background

Disparities in the stage at diagnosis for breast cancer have been independently associated with various contextual characteristics. Understanding which combinations of these characteristics indicate highest risk, and where they are located, is critical to targeting interventions and improving outcomes for patients with breast cancer.

Methods

The study included women diagnosed with invasive breast cancer between 2009 and 2018 from 680 U.S. counties participating in the Surveillance, Epidemiology, and End Results program. We used a machine learning approach called Classification and Regression Tree (CART) to identify county "phenotypes," combinations of characteristics that predict the percentage of patients with breast cancer presenting with late-stage disease. We then mapped the phenotypes and compared their geographic distributions. These findings were further validated using an alternate machine learning approach called random forest.

Results

We discovered seven phenotypes of late-stage breast cancer. Common to most phenotypes associated with high risk of late-stage diagnosis were high uninsured rate, low mammography use, high area deprivation, rurality, and high poverty. Geographically, these phenotypes were most prevalent in southern and western states, while phenotypes associated with lower percentages of late-stage diagnosis were most prevalent in the northeastern states and select metropolitan areas.

Conclusions

The use of machine learning methods of CART and random forest together with geographic methods offers a promising avenue for future disparities research.

Impact

Local interventions to reduce late-stage breast cancer diagnosis, such as community education and outreach programs, can use machine learning and geographic modeling approaches to tailor strategies for early detection and resource allocation.

SUBMITTER: Dong W

PROVIDER: S-EPMC8755627 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:ImportanceThere are stark disparities in cervical cancer burden in the United States, notably by race and ethnicity and geography. Late-stage diagnosis is an indicator of inadequate access to and utilization of screening.ObjectiveTo identify geospatial clusters of late-stage cervical cancer at time of diagnosis in Texas.Design, setting, and participantsThis population-based cross-sectional study used incident cervical cancer data from the Texas Cancer Registry from 2014 to 2018 of female patients aged 18 years or older. Late-stage cervical cancer cases were geocoded at the census tract level (n = 5265) using their residential coordinates (latitude and longitude) at the time of diagnosis. Statistical analysis was performed from April to September 2023.ExposuresCensus tract of residence at diagnosis.Main outcome and measuresLate-stage cervical cancer diagnosis (ie, cases classified by the National Cancer Institute Surveillance, Epidemiology and End Results summary stages 2 to 4 [regional spread] or 7 [distant metastasis]). A Poisson probability-based model of the SaTScan purely spatial scan statistics was applied at the census tract-level to identify geographic clusters of higher (hot spots) or lower (cold spots) proportions than expected of late-stage cervical cancer diagnosis and adjusted for age.ResultsAmong a total of 6484 female patients with incident cervical cancer cases (mean [SD] age, 48.7 [14.7] years), 2300 (35.5%) were Hispanic, 798 (12.3%) were non-Hispanic Black, 3090 (47.6%) were non-Hispanic White, and 296 (4.6%) were other race or ethnicity. Of the 6484 patients, 2892 with late-stage diagnosis (mean [SD] age, 51.8 [14.4] years were analyzed. Among patients with late-stage diagnosis, 1069 (37.0%) were Hispanic, 417 (14.4%) were non-Hispanic Black, 1307 (45.2%) were non-Hispanic White, and 99 (3.4%) were other race or ethnicity. SaTScan spatial analysis identified 7 statistically significant clusters of late-stage cervical cancer diagnosis in Texas, of which 4 were hot spots and 3 were cold spots. Hot spots included 1128 census tracts, predominantly in the South Texas Plains, Gulf Coast, and Prairies and Lakes (North Texas) regions. Of the 2892 patients with late-stage cervical cancer, 880 (30.4%) were observed within hot spots. Census tract-level comparison of characteristics of clusters suggested that hot spots differed significantly from cold spots and the rest of Texas by proportions of racial and ethnic groups, non-US born persons, and socioeconomic status.Conclusions and relevanceIn this cross-sectional study examining geospatial clusters of late-stage cervical cancer diagnosis, place-based disparities were found in late-stage cervical cancer diagnosis in Texas. These findings suggest that these communities may benefit from aggressive cervical cancer interventions.

Project description:BackgroundThe gap between the highest and lowest life expectancies for race-county combinations in the United States is over 35 y. We divided the race-county combinations of the US population into eight distinct groups, referred to as the "eight Americas," to explore the causes of the disparities that can inform specific public health intervention policies and programs.Methods and findingsThe eight Americas were defined based on race, location of the county of residence, population density, race-specific county-level per capita income, and cumulative homicide rate. Data sources for population and mortality figures were the Bureau of the Census and the National Center for Health Statistics. We estimated life expectancy, the risk of mortality from specific diseases, health insurance, and health-care utilization for the eight Americas. The life expectancy gap between the 3.4 million high-risk urban black males and the 5.6 million Asian females was 20.7 y in 2001. Within the sexes, the life expectancy gap between the best-off and the worst-off groups was 15.4 y for males (Asians versus high-risk urban blacks) and 12.8 y for females (Asians versus low-income southern rural blacks). Mortality disparities among the eight Americas were largest for young (15-44 y) and middle-aged (45-59 y) adults, especially for men. The disparities were caused primarily by a number of chronic diseases and injuries with well-established risk factors. Between 1982 and 2001, the ordering of life expectancy among the eight Americas and the absolute difference between the advantaged and disadvantaged groups remained largely unchanged. Self-reported health plan coverage was lowest for western Native Americans and low-income southern rural blacks. Crude self-reported health-care utilization, however, was slightly higher for the more disadvantaged populations.ConclusionsDisparities in mortality across the eight Americas, each consisting of millions or tens of millions of Americans, are enormous by all international standards. The observed disparities in life expectancy cannot be explained by race, income, or basic health-care access and utilization alone. Because policies aimed at reducing fundamental socioeconomic inequalities are currently practically absent in the US, health disparities will have to be at least partly addressed through public health strategies that reduce risk factors for chronic diseases and injuries.

Project description:Purpose Manually identifying geographic atrophy (GA) presence and location on OCT volume scans can be challenging and time consuming. This study developed a deep learning model simultaneously (1) to perform automated detection of GA presence or absence from OCT volume scans and (2) to provide interpretability by demonstrating which regions of which B-scans show GA. Design Med-XAI-Net, an interpretable deep learning model was developed to detect GA presence or absence from OCT volume scans using only volume scan labels, as well as to interpret the most relevant B-scans and B-scan regions. Participants One thousand two hundred eighty-four OCT volume scans (each containing 100 B-scans) from 311 participants, including 321 volumes with GA and 963 volumes without GA. Methods Med-XAI-Net simulates the human diagnostic process by using a region-attention module to locate the most relevant region in each B-scan, followed by an image-attention module to select the most relevant B-scans for classifying GA presence or absence in each OCT volume scan. Med-XAI-Net was trained and tested (80% and 20% participants, respectively) using gold standard volume scan labels from human expert graders. Main Outcome Measures Accuracy, area under the receiver operating characteristic (ROC) curve, F1 score, sensitivity, and specificity. Results In the detection of GA presence or absence, Med-XAI-Net obtained superior performance (91.5%, 93.5%, 82.3%, 82.8%, and 94.6% on accuracy, area under the ROC curve, F1 score, sensitivity, and specificity, respectively) to that of 2 other state-of-the-art deep learning methods. The performance of ophthalmologists grading only the 5 B-scans selected by Med-XAI-Net as most relevant (95.7%, 95.4%, 91.2%, and 100%, respectively) was almost identical to that of ophthalmologists grading all volume scans (96.0%, 95.7%, 91.8%, and 100%, respectively). Even grading only 1 region in 1 B-scan, the ophthalmologists demonstrated moderately high performance (89.0%, 87.4%, 77.6%, and 100%, respectively). Conclusions Despite using ground truth labels during training at the volume scan level only, Med-XAI-Net was effective in locating GA in B-scans and selecting relevant B-scans within each volume scan for GA diagnosis. These results illustrate the strengths of Med-XAI-Net in interpreting which regions and B-scans contribute to GA detection in the volume scan.