How Does Google AI Detect Breast Cancer Better Than Radiologists?
Artificial Intelligence in healthcare is becoming more crucial with early detection of various diseases with better accuracy. Cancer is one the widespread deadly disease can be now detected through machine learning and AI-enabled automated machines.
Breast cancer is most common among women worldwide. However, more than 90% of women diagnosed with breast cancer at the earliest stage survive their disease for at least 5 years compared to around 15% for women diagnosed with the most advanced stage of the disease, which is now possible with AI.
Though, AI is already diagnosing the breast cancer but owing to accuracy, the reliability among the doctors was less. But now Google has developed an AI system that can detect the early sings of breast cancer better than radiologists.
Research of Google Breast Cancer AI
In the research, the Google’s AI model called DeepMind AI was trained and tuned on a representative data set comprised of de-identified mammograms from more than 76,000 women in the U.K. and more than 15,000 women in the U.S.
The AI model is trained to see if it could learn to spot signs of breast cancer in the scans. The model was then evaluated on a separate de-identified data set of more than 25,000 women in the U.K. and over 3,000 women in the U.S.
In this study, researchers conducted a test where Google’s AI system was pitted against six radiologists where the AI managed to outperform all six of them at accurately detecting breast cancer among these women at better accuracy.
Earlier, when a similar study is conducted in which the team first trained AI to scan X-ray images, then looked for signs of breast cancer by identifying changes in the breasts of the 28,000 women.
They then checked the computer’s guesses against the women’s’ actual medical outcomes, reducing the reliability of such applications.
But now, the accuracy level has improved and reduced false negatives by 9.4% and cut down false positives by 5.7% for women in the US.
In the UK, where two radiologists typically double-check the results, the model cut down false negatives by 2.7% and reduced false positives by 1.2% make AI more reliable.
How Does AI Detect Breast Cancer?
Breast cancer diagnosis is done by oncologist’s human knowledge and intuition of what major risk factors might be, such as age, family history of breast and ovarian cancer, hormonal, reproductive factors, and breast density.
While AI in breast cancer diagnosis, rather than manually identifying the patterns in a mammogram that drives future cancer, the MIT/MGH team trained a deep-learning model to deduce the patterns directly from the medical imaging data.
Video: How AI Model Improves Breast Cancer Detection on Mammograms
And using the information from more than 90,000 mammograms, the model detected patterns too subtle for the human eye to detect the cancer cells. The Google breast cancer AI algorithms are used to learn such patterns and predict.
You can check below an image showing the visualization of tumor growth and metastatic spread in breast cancer with screening to detect breast cancer early, before symptoms develop.
Actually, the demographics of the population studied by the authors are not well defined in the previous AI-based detection.
As the performance of AI algorithms can be highly dependent on the population used in the training data sets.
And these training data sets are created by annotating the medical images of breast infected with cancer available in various formats like X-rays, CT Scan and MRI.
A huge amount of such labeled data is used to train the AI algorithms.
There are many companies providing healthcare training data with annotated medical imaging to train the AI and ML models with accuracy.
And with the availability of such data, detection of various types of other common cancer through AI will become possible benefiting the humans saving their life from such deadly maladies.
This story was originally featured on www.vsinghbisen.com