Though the practice of artificial intelligence (AI) was different when John McCarthy coined the term in 1955, advancements in the field have allowed researchers and doctors to improve how cancer is diagnosed.
The duration of time from exposure to asbestos and the development of malignant mesothelioma can take decades. On top of a long latency period, mesothelioma patients are typically diagnosed in later stages, when life expectancy is reduced.
Analyzing recent treatment techniques shows AI has already enhanced the medical field. Now, doctors can improve mesothelioma diagnosis and treatment by utilizing AI’s ability to detect cancer sooner.
What Is Artificial Intelligence?
In contrast to natural intelligence exhibited by humans, researchers from Oxford University have interpreted AI as “intelligent agents” among the environment that mimic the cognitive functions of humans. In order for AI to perform these functions, humans must produce human intelligence, often in the form of coding, and apply it to AI technology.
Early advances in the field of AI focused on constructing neural networks, or algorithms modeled after the human brain, designed to recognize patterns. Successful AI platforms interpret large amounts of information without the human brain and apply that information to future tasks.
The use of AI is common. Virtual assistants like Apple’s “Siri” and Amazon’s “Alexa” use AI technology to answer questions, make recommendations, and perform certain actions upon request. Tesla uses AI to support its self-driving technology by using one of the world’s most innovative neural networks. Netflix uses AI to generate recommendations based on previous data collected from users.
The health services industry utilizes various applications of AI, from clerical work to detecting and treating cancer. Researchers believe it can transform the prognosis of multiple cancers, including mesothelioma and breast and lung cancer.
AI and Cancer Detection
Previous diagnostic practices were less effective and lacked the technology for early detection. AI has the potential to be reliable enough in helping doctors diagnose cancer sooner with reduced human error.
As AI becomes more powerful and capable of collecting and storing large amounts of data, researchers have begun to explore its ability to recognize and identify patterns, solve problems, prognose diseases, and determine which treatment is needed.
AI machines in the health field use a combination of machine learning and deep learning. Machine learning is a subset of AI that involves the use of algorithms and statistical models to perform a specific task without explicit instructions through patterns learned from previous data. Deep learning is a subset of machine learning that allows technology to learn without human supervision.
Since mesothelioma is so rare, diagnosing it can be difficult. The standard diagnostic process for mesothelioma patients begins when symptoms begin to show. The doctor collects a thorough medical history, including exposure to asbestos, and then performs a physical exam on the patient. If the doctor suspects cancer, he or she may order imaging tests, like X-rays or CT scans, to search for abnormalities. Biopsies, which can be rather intrusive, have been the only definitive way to accurately diagnose mesothelioma.
The promise of AI advancement in the health field relies on medical imaging. AI technology can assist in interpreting and segregating benign and malignant lesions shown in X-rays or CT scans that may not be detected by human eyes. Researchers refer to this concept as computer-aided detection/diagnosis (CAD).
Studies Have Shown Success In Cancer Detection Using AI
Early studies show AI technology in imaging machines can reduce the workload for radiologists and increase efficiency overall, while storing significantly higher amounts of data than the human brain.
A team of researchers at Harvard Medical School observed a 2.9 percent error rate in an AI model that examed breast cancer diagnoses from biopsy slide images. The study also found pairing the deep learning AI technology with a pathologist’s inputs showed an error rate as low as 0.5 percent.
Last October, researchers at Owkin, a software company from New York, created a deep-learning program called MesoNet. The program can identify potential mesothelioma occurrences in patients by scanning and interpreting tissue samples.
The AI model successfully detected biological features of tumors after scanning nearly 3,000 biopsy image samples. By pinpointing new tumor characteristics, AI can connect them to an effective prognosis.
Google Health and DeepMind, a computer programming company, joined forces to produce an AI algorithm that accurately detects breast cancer. A study tested the technology on mammogram images from female patients in the U.S. and U.K. Results suggested the technology performed better than human radiologists.
According to the study, about 1 in 5 breast cancer screenings fail to detect breast cancer even when it’s present, known as a false negative. The AI study on the American women reduced false negatives by 9.4 percent and false positives by 5.7 percent. The collaborators believe their AI system will eventually decrease healthcare costs and reduce unnecessary procedures.
We still need humans to create and maintain AI systems. Researchers agree that AI will maintain to enhance and assist in interventional radiology. While research has proven credible advancements in AI and the medical field, it will never replace the supervision from humans.
The alliance of radiologists, computer scientists, and biomedical engineers has advanced this form of technology since its early practices. While more research needs to explore possible challenges, new AI technology allows doctors to detect cancer earlier, reduce human error, and improve conventional treatment.