Detecting Medical Errors in the Cancer Screening Process: An In-Depth Review of Lung Cancer Screening & at Risk Qualifications
Based on studies conducted by the National Cancer Institute, approximately 1 in 16 Americans will be diagnosed with lung cancer during their lifetime, and nearly 150,000 lose their lives each year to lung and other related bronchus cancers. These rates have only increased in recent years, while survival rates remain gloom. More lives are lost to lung cancer “than colon, breast, and prostate cancers combined” and only 21% of all people diagnosed will survive 5 years or more. Despite these daunting statistics, only 6% of federal government cancer funds is spent on lung cancer research. With this minimal amount of research, screening standards have not been changed in recent years, and in fact, there is a current push to change this protocol.
`The Center for Disease Control and Prevention currently suggests that the only proven way to test for lung cancer is through a low-dose computed tomography (LDCT) scan. This scan emits a “low dose of radiation to make detailed images of your lungs”. As of 2013, screening is only recommended for individuals who have a history of heavy smoking, smoke now or have within the past 15 years, and are between 55 and 80 years old. Other factors, such as air quality, underlying conditions, or exposure to smoke are not considered, though they make up over 15% of lung cancer cases. Through statistical analysis and multivariate regressions, this research will show that there are underlying factors that attribute to lung cancer diagnosis, that are not currently being screened for. Upon completion, a formal recommendation will be created, to help increase awareness of potential factors for diagnosis, in otherwise healthy patients.
- The first aim of this project is to make a recommendation for Lung Cancer Screenings in patients that are non - smokers.
- This project will look at different environmental factors or health risks that may attribute to an individual's cancer stage & prognosis
- Since screening protocols are in place to extend longevity after diagnosis, an in-depth look at days lived after diagnosis, based on varying factors, will be taken into account (age, smoking history, BMI, etc)
- Histograms and other normalized distributions will help highlight differences between genders, stages, and types of Lung Cancer
- Multivariate Regressions will help determine correlations amongst varying factors, that may play a role in developing Lung Cancer
- This research will help the 15% of non smokers being diagnosed with lung cancer each year, help to establish what genetic or environmental components make them more at risk.
- An end result for this project is to formulate a predictive model, using Machine Learning, to determine the likelihood of a non-smoker developing cancer in the United States.
Yasaman Asayesh, M.S .- University of Massachusetts Dartmouth
Dr. Keivan Sadeghzadeh - University of Massachusetts Dartmouth