Medical Tech

Machine Learning in Cancer: Can AI Predict Cancer?


Machine Learning in Cancer: Can AI Predict Cancer?

Artificial intelligence, or machine learning, has been around since the late 1950s, but recently, it’s become extremely useful. The medical industry, in particular, is benefiting from machine learning in cancer diagnosis. In fact, cancer prediction using machine learning is on the horizon as one of the most life-changing uses for tech available.

Humans are fallible and it’s been shown time and time again that human doctors are likely to miss some issues in the human body. For example, incipient cancer can often go missed by everyone, even technicians and doctors who are trained to look for it. However, AI can quickly and accurately spot cancer long before the human eye and tests can.

Cancer Prediction Using Machine Learning

The value of machine learning in cancer was shown when a study was done to see if a machine could detect lung cancer. Mozziyar Etemadi spent much of his career training an AI system to find tumors in scans of lung cancer patients. The machine was far more accurate than the human doctors, but the real test was when Etemadi showed the system older CT scans of patients who developed cancer later in life. The scans looked perfect and would pass any doctor’s perusal, but when the AI system scanned them, the machine rapidly picked up on markers that indicated cancer was developing.

Lung cancer is the most likely cancer to kill you, with a 75 percent mortality rate in the first five years after diagnosis. That is largely because the cancer is not detected until later. Roughly 70 percent of lung cancers are found too late to treat easily. When it’s found early on, and when tumors are still inside the lung and have not spread, nearly 66 percent of patients will survive five years or more after diagnosis. Catching cancer early is essential in preventing death.

Using machine learning to diagnose cancer far earlier than usual could mean drastic improvements in cancer treatments and a reduced mortality rate. The machines are shown what previous cancer looks like and then go on to deduct where the cancer is and how advanced it is without requiring doctors to check hundreds or thousands of possible false positives. With today’s AI systems, it’s possible for the machine to train itself. In the lung cancer study, the AI system was correctly diagnosing patients 94 percent of the time.

cancerscreening, machinelearning, cancerdiagnosis, cancerAI, AIpredicts, medicalAI

Cancer Treatment: Machine Learning Determines Which Meds Work Best

Cancer treatments are quite personalized, since not every treatment works on every person or every cancer. This means there are often starts and stops in the therapy. Just with breast cancer, there are a myriad of options. For instance, 49 percent of women with Stage I-II breast cancer have a lumpectomy or partial mastectomy followed by radiation therapy. Women with Stage III cancer tend to fare better with a mastectomy and chemotherapy after the breast removal, with 56 percent of women undergoing this treatment. Stage IV breast cancer generally results in 56 percent of patients receiving radiation with chemotherapy or on its own, sans surgery. Finally, 17 percent of women with Stage IV cancer will have surgery only, or with other treatments. There’s such a wide range of treatment options that it can be tough to know just when to use them. That’s where AI comes in.

Through deep learning, machines can predict exactly how a cancer treatment will work on the cancer cells. This depends on everything from the placement of the cells to the genomic variability and what drugs the cancer has already become resistant to. Multiple tests have shown that AI machines are highly accurate in predicting how sensitive a patient will be to a certain type of medication. This was shown in a range of cancer types, including gastric, ovarian, and endometrial.

Once tested, the drugs predicted by the AI to be least useful were proven to be least effective on the types of cancer they were used for. There are now drug resistant cancers, so it’s essential that we continue to evolve and learn, but AI will definitely play a big role in this.

Instead of testing multiple cancer treating drugs on someone, it should be possible to select exactly what treatment they need. Not only does this save millions of dollars in pointless treatments but it also means treatment can be more precise and targeted. This relieves discomfort for the patient and means fewer testing phases to find out which drug works best. When in the end stages of cancer, time is of the essence, so using the right treatment straight away is essential.

AI could easily save lives if we permit it. Thanks to movies and television, many people are wary of the idea of artificial intelligence, but it really does have the power to make human life much safer.

cancerscreening, machinelearning, cancerdiagnosis, cancerAI, AIpredicts, medicalAI

What AI in Cancer Means for the Future

Machine learning is the way to go with cancer screening and treatment, but the technology is not fully implemented. Newer AI systems are faster than ever for learning, but the sheer number of cancer types makes this a long, slow job of training the systems. Machine learning requires a great deal of materials, as well. The system has to process multiple scans and learn from them, which takes a considerable amount of time.

For now, while cancer prediction using machine learning is possible, it’s just not used as much as it could be. There’s still some further testing and training to be done. This will help ensure that all the cancer screening done by machines is as accurate as possible when it does become mainstream.

With the rate of technology advancement, the wait for machine learning and AI treatments shouldn’t be too long. The systems just need more practice to get ready. As more and more machines speed up the process, doctors should soon start using them.


Previous Article

NeuroMD Corrective Therapy Device for Pain Relief

Next Article

Medisafe App Review

Related Posts