Source: Deep Learning on Medium
By Katie Pollitt, Senior Summit Creator — RE•WORK
What are Mammograms & Why Do Women Need Them?
Mammograms are a type of x-ray used to analyze breast tissue and detect breast cancer at the earliest stage possible. This screening often shows breast lumps before they can be felt and can illuminate tiny calcium clusters that may be indicators of breast cancer or other conditions.
It is recommended that women age 40 and older should get a mammogram every 1–2 years. If a woman younger than 40 has risk factors for breast cancer, it is advised that she should talk to her healthcare provider about the benefits of the x-ray.
How Mammograms Are Performed Mammograms are performed by flattening each breast between two plastic plates to then get a clear x-ray of the breast tissue. A total of four x-ray images are captured to allow for different views of the tissue. A radiologist on site will then review the images to look for any breast changes such as calcium deposits or masses that may indicate breast cancer.
What happens if my mammogram is abnormal? If the radiologist finds a mass or calcification, this does not always mean a sure sign of cancer. Additional tests such as an ultrasound or MRI will likely need to be performed. In certain cases, a biopsy may be recommended to analyze tissue from a concerning area under a microscope.
Breast Cancer Statistics
● In the U.S. alone, 1 in 8 women will develop invasive breast cancer in her lifetime.
● Roughly 40,920 women in the U.S. are expected to die in 2018 from breast cancer. However, annual death rates have been on a decline since 1989.
● Female breast cancer has the highest new case rate per 100,000 with 124.8. The second highest new case rate is that of lung and bronchus cancer at just 50.7 new cases per 100,000.
Why Are Mammograms Flawed?
Despite the intention of detecting breast cancer as early as possible, mammograms and the interpretation of x-ray results are flawed. Reasons for flawed testing include:
● Doctors cannot detect a tumour from the image (false-negative result)
● Doctors read the screening as cancer and are then proven wrong with additional testing (false-positive result)
● Current mammograms cannot detect interval cancers that are fast-developing and often deadly between screenings
Mammogram Statistics & Concerns
● In the U.S. alone, over 30 million mammograms are performed each year.
● The American Cancer Society estimates that roughly 50% of mammography test results are false-positive, often requiring women to go through additional testing for no reason.
● $4 billion each year in healthcare spending is due to false-positive mammograms and overdiagnosis of breast cancer.
● 20% of mammograms are false-negative, or when the radiologist interprets a scan to be normal despite cancer being present. This can often lead to delays in treatment and emotional turmoil for those diagnosed.
What is Artificial Intelligence and Where is It Being Used Artificial intelligence (AI) is a type of machine learning defined by computer systems being able to perform tasks that once required human intelligence. Today, AI is being used in voice recognition software, to perform difficult surgeries, and in transport situations with driverless cars.
How Will AI Improve Mammograms?
Artificial Intelligence is now being used to improve the accuracy of mammogram readings. With machine learning continuing to make waves in the technological field, companies have focused their efforts in the healthcare space to improve these screenings.
1. Texas’ Houston Methodist Research Institute has developed an AI program that can interpret mammograms with 99% accuracy and 30x faster than a human doctor. 2. Therapixel, a French startup specializing in medical imaging software, has said its AI
Technology can reduce the rate of false-positive mammogram results by 5%.
Artificial Intelligence has been and will also be able to:
● Facilitate the proper training of medical professionals to reduce errors in mammogram interpretation
○ In a study published by the Journal of Medical Imaging, Artificial Intelligence was used to analyze eye movements of radiologists to understand the cognitive processes involved in image interpretation.
● Deep learning technology, a subset of Artificial Intelligence, helps radiologists to identify potential mammographic abnormalities at a much quicker pace.
● AI options also help radiologists to compare mammogram images from year to year more easily, picking up on subtle differences that may be more difficult for humans alone to spot.