The use of artificial intelligence and machine learning may aid cancer prognosis. Artificial intelligence can detect tumours that are already present and identify persons who are at high risk of getting the illness before the disease has a chance to take hold. This enables physicians to carefully watch these patients and respond as soon as necessary if the situation calls for it.
The use of artificial intelligence in immunotherapy
The use of artificial intelligence in immunotherapy is mostly focused on analysing the effects of various therapies and assisting clinicians in adjusting their prescriptions. Scientists at UT Southwestern Medical Center and MD AndersonCancer Center developed an artificial intelligence-powered technique for determining which neoantigens (peptides produced by cancer cell genome mutations) are recognised by a patient's immune system.
The technique is currently being tested in clinical trials. Such artificial intelligence algorithms might be useful in forecasting the response of cancer cells to immunotherapies. T cells, which are part of our immune system, are continually on the lookout for signals of cancer and other invading entities. When these cells detect neoantigens, they form a clump that sticks to one another. Some neoantigens, on the other hand, are not identified, enabling cancer to spread. Neoantigens are found in tens of thousands of different varieties.
A difficult, expensive, and time-consuming job is determining whether or not they have the potential to activate T cells. This is becoming achievable because of the advancements in machine learning.
The use of artificial intelligence in cancer prognosis via medical imaging
One of the most noticeable use is the detection and classification of malignant tumours. PaigeProstate, an AI-powered pathology solution for prostate cancer, was approved by the FDA in September 2021. When used in conjunction with the full focus digital pathology viewer, aids in the detection of prostate cancer.
To get this clearance, the FDA reviewed data from a clinical investigation in which 16 pathologists analysed 527 prostate biopsy pictures for evidence of cancer as a prerequisite to granting the approval. Each doctor conducted two evaluations, one with the assistance of Paige Prostate and the other without assistance. A trial indicated that the use of this artificial intelligence-based technology increased the rate of cancer diagnosis by an average of 7.3 percent. Additionally, it decreased the number of erroneous negative diagnoses by 70% and the number of false-positive diagnoses by 24%.
A second example is the work of Mark Schiffman, senior investigator at the National Cancer Institute, who collaborated with his colleagues to develop an artificial intelligence-based algorithm that can examine images of cervical tumours and detect precancerous changes that necessitate immediate medical attention.
The advantages of using artificial intelligence in cancer diagnosis and therapy
In general, artificial intelligence provides several advantages in the healthcare field. Incorporating artificial intelligence into cancer therapy and diagnosis has many significant benefits, which are as follows:
- Individualizing treatment plans
Doctors may examine a wide range of information about a patient and their cancer cells with the use of artificial intelligence and big data, resulting in the development of individualised therapies. This sort of treatment will have fewer negative side effects than other types of therapy. Because of this, it will have a greater effect on cancer cells while causing less harm to healthy cells.
Artificial intelligence is being used in cancer therapy. These twins are virtual duplicates of the persons that inspired them. They include information like DNA, RNA, and proteins, and they aid in the identification of the most successful cancer treatment method for a specific individual patient.
2. Lowering the number of false positives and negatives
The use of artificial intelligence in cancer detection will increase the accuracy of diagnosis while decreasing the number of false positives and negatives. Breast cancer detection research provides evidence to support this claim. One in every ten female patients has a false-positive result from mammography, resulting in the women having to cope with the stress and invasive treatments that were not required.
Google's research team developed artificial intelligence-powered software that reduced the number of false-positive mammography readings by 6 percent and the number of false-negative mammogram readings by 9 percent. Another group of scientists has created an artificial intelligence system for the identification of breast cancer. When tested, this model was shown to have assisted radiologists in reducing false-positive rates by 37.3 percent.
3. Eliminating cancer overtreatment as a last resort
Artificial intelligence assists radiologists in determining which tumours/abnormalities are malignant and so need treatment. Precancerous lesions in cervical pictures may be detected and distinguished from other abnormalities using artificial intelligence algorithms, according to research published in the Journal of the National Cancer Institute. This prevents patients from being over treated for small problems.
4. The ability to identify tumour kinds without the need for invasive procedures.
Sometimes, only after the tumour removal operation is completed, do the physicians discover that the tumour was benign and that the procedure might have been avoided entirely if the tumour had been detected earlier. Such occurrences may be considerably minimised with the use of artificial intelligence in cancer diagnosis.
For example, according to one research, artificial intelligence may lower the number of breast-conserving procedures by 30.6 per cent. Image-guided needle biopsies may be used to train machine learning algorithms to detect malignant tumours in a patient's body.
When the researchers used a random forest machine learning system to study 335 potential cancer patients, they discovered that it averted one-third of unneeded procedures. Another example is brain cancers, and researchers from Harvard University and the University of Pennsylvania created a deep learning system for tumour categorization in collaboration with the University of Pennsylvania.
It is capable of identifying and characterising isocitrate dehydrogenase (IDH) mutations in gliomas using magnetic resonance imaging (MRI) pictures without the requirement for invasive procedures that would otherwise be necessary.
Conclusion
When it comes to adopting artificial intelligence for cancer diagnosis and prevention, healthcare companies must exercise caution. Artificial intelligence is a strong instrument that has the potential to save lives while also saving doctors' time. When not trained and deployed effectively, however, it may be quite dangerous.