Mesothelioma and Breast Cancer Patients See Benefits from Artificial Intelligence

The concept of artificial intelligence or AI has been around for decades, but it is only starting to reach some of its high-powered capabilities today. While humans are not yet entirely dependent on AI, it is being used to help society in areas such as email assistance to groundbreaking changes such as self-driving vehicles. (Worldhealth.net)

In addition, artificial intelligence is having significant effects on the medical field. It has proven its ability, for example, to analyze radiology images and support doctors in the detection of tumors. AI has particularly promising possibilities for patients who have cancers that are especially hard to diagnose.

In the last few years, AI’s evolution is being directed to providing assistance in mesothelioma research and healthcare. Clinical studies have shown that artificial intelligence can play an important role in improving preliminary tests for cancer. For mesothelioma patients, artificial intelligence may be a vital component to finding an earlier diagnosis. This early diagnosis could lead to earlier, targeted treatment, which could improve life expectancy.

Mesothelioma Is One of the Toughest Cancers

One of the challenges with mesothelioma is that it can take 20 years or longer to show up in the body. There are four stages of the disease, and if it is diagnosed in a later stage, the patient’s prognosis is poor. Most people diagnosed in stage 3 or 4 only live 16 months. Mesothelioma is caused by asbestos exposure in industrial and manufacturing applications. Unfortunately, people may not be aware they have the cancer until it is too late. Some doctors misdiagnose mesothelioma as lung cancer, which means it is not properly treated until it is too late.

Artificial Intelligence Can Allow for Earlier Diagnosis

AI is of great value to mesothelioma patients. It can offer a clear, earlier diagnosis that allows them to start treatment sooner. In earlier cases, diagnosis was later, and it left patients with fewer treatment options. AI offers hope for people who would have otherwise short life expectancies. While AI is not the only possibility, its use if it is proven reliable, can transform the prognosis of cancer patients around the world.

Other cancers, including breast and lung cancer, can also see benefits from artificial intelligence with early detection. Once the cancer has been detected, doctors can start with treatments before the cancer becomes more aggressive.

Deep-Learning Program Launched in 2019

In October 2019, researchers came up with a deep-learning program called MesoNet to identify possible mesothelioma patients early on. The program scanned tissue samples with this technology to understand who would respond best to certain cancer treatments. Radiation therapy, immunotherapy, are best used when doctors can address the cancer before it spreads to other areas, such as those where operations are ineffective. AI, in this case, was used to pinpoint a new tumor and was able to connect it to effective treatment and prognosis.

The model has been tested and verified by doctors at the Centre Leon Berard Cancer Institute in Lyon, France. MesoNet can not only predict mesothelioma in patients but can also identify new biomarkers in the stromal parts of the tumor microenvironment that were most predictive of survival.

Before these breakthroughs, machine learning was put into practice to provide advances in artificial intelligence. As AI has made a large impact in radiology image analysis, this technology has great potential because it can process images faster and with greater accuracy than doctors.

Artificial Intelligence Shows Hope for Breast Cancer Patients

AI is being introduced into diagnosing breast cancer in the early stages and the outlook is promising. Ultrasound elastography is a new diagnosis technique that tests how stiff breast tissue is. It does this be vibrating the tissue, which causes waves. This wave leads to distortions in the ultrasound scan, which highlights the parts of the breast where properties differ from the other tissue. (Medicalnewstoday.com)

From this information a doctor may be able to find if a lesion has cancer or is benign. While this technique does have a lot of potential, analyzing the results of this technique takes a lot of time, involves many steps, and requires the solving of complex problems.

Last year, a group of researchers at the University of Southern California in Los Angeles tried to create an algorithm that could lower the steps that are required to draw information from the images. The results were published in the journal Computer Methods in Applied Mechanics and Engineering.

Researchers wanted to see if an algorithm could be trained to tell the difference between benign and malignant lesions in breast scans. They tried to do this by teaching the algorithm with synthetic data instead of genuine scans.

Synthetic Data

When it was asked why the team was using synthetic data, the lead doctor on the study said it all comes down to how available real-world data is. He noted that in medical images, you are lucky to have access to 1,000 images. In this type of situation, where data is rare, these techniques are critical. (Medicannewstoday.com)

The clinical researchers trained their algorithm, which they call a deep convolutional neural network, using 12,000 synthetic images. By the end of the study, the algorithm had 100% accuracy on synthetic images. Next, they wanted to try scans in real life. They had only 10 scans to access. Half of them showed malignant lesions and the other half had benign lesions.

They had an accuracy rate of 80%. They then continued to refine the algorithm by using real-world images. While 80% is a good amount, it is not sufficient But this is just the beginning of the process. The scientists thought if they trained the algorithm on real data, it could have shown more accuracy. The researchers also noted that the test was too small to test the real capabilities of the system.

This experiment showed the potential value of using AI to diagnose certain forms of breast cancer. But the leaders of the study did not think that AI will ever fully replace human operators. While these algorithms do have a major role to play, human operators will always be needed to make final conclusions.

But researchers hope they can expand this method to diagnose other forms of cancer, such as lung cancer and mesothelioma. When a tumor starts to grow, it changes the behavior of the tissue, so it should be possible to chart the differences and train the algorithm to identify them.

DeepMind and Google Health Announce Collaboration

DeepMind and Google Health are collaborating to devise an artificial intelligence algorithm that is more accurate to find early breast cancer than human radiologists. Another form of artificial intelligence called an artificial immune system, can help doctors to find malignant pleural mesothelioma in some patients. The study showed 97% accuracy and was better than the current algorithms for cancer diagnosis.

As with any disease, timing is of greatest importance. Artificial intelligence is unlikely to over replace human judgment, but it can help doctors to reach higher levels of cancer remission. If artificial intelligence can be used in oncology, whether it is for supplementary image analysis scanning or by finding possible cancer patients to specialize treatment, this can be a way for some patients to receive the best possible care.