The future of cancer research lies in artificial intelligence

Artificial intelligence is here to stay, and proof of that is that last December The European Parliament and the Council agreed on the first law to regulate it. impact on the safety standards and fundamental rights on which the technology that is increasingly being adopted throughout the world should be based.

Artificial intelligence is computer programs or algorithms that have been man-made and that they use a range of data to make predictions based on previous information they have collected. Experts program technology with specific information and then this technology analyzes it and makes a decision.

While like any technology it is not without its risks, the truth is that artificial intelligence is also provides numerous benefits and can be complementary and even key in certain subjectssuch as cancer research.

Location of tumors

The National Cancer Institute is one of the National Institutes of Health, the main US government agency responsible for biomedicine and public health – In a 2022 paper, he asked whether artificial intelligence could help us look at cancer in new ways. and more effective than what is already known. To do this, he referred to a test that analyzed the effectiveness of this AI in comparison with the effectiveness of a doctor with more experience.

Two identical but blurry black and white images were presented on the computer screen. The one on the left has been analyzed radiologist with 15 years of experienceand the one on the right is for artificial intelligence computer program. The doctor marked an area that he thought represented the spread of prostate cancer, and the artificial intelligence did the same. The results were almost identical.

Analysis of removed tumors in real time

For their part, researchers from University of North Carolina School of Medicine created an artificial intelligence model in 2023 that analyzes surgically removed tumors in real time and published the results of the study in the journal Annals of Surgical Oncology..

During breast cancer surgery, the surgeon removes the tumor and some healthy tissue around it to ensure that any cancer cells that may be present are removed. This sample is then x-rayed so the medical team can confirm that the cancerous area has been removed. After this analysis, the sample is taken to a laboratory, where a pathologist checks to see if there are cancer cells in the removed area of ​​healthy tissue. If they are, an additional operation will be required, but this will take place a few days after the first. Thanks to this artificial intelligence, Analysis of the removed cancer tissue can be checked during the breast cancer surgery itself, which will ultimately shorten the time..

Author of the study Crystalyn Gallagherexplained that this tool will allow us to analyze “with greater accuracy Surgically removed tumors in real time and will increase the chances that all cancer cells will be removed during surgery. “This will avoid the need to return patients for a second or third operation.”

Four algorithms improve breast cancer diagnosis

Spain has also seen revolutionary advances in this regard. January 31, The Catalan Institute of Health (ICS) said four artificial intelligence algorithms have been developed. own, which has made it possible to improve the diagnosis of breast cancer in the seven main Catalan public hospitals, which can also exchange diagnoses and knowledge among themselves in real time. The idea is that this artificial intelligence will help detect this type of cancer.

Algorithms work from quantification of four biomarkers that are present in types of breast cancer: HER2, Ki67, estrogen receptor and progesterone receptor are already used in nearly 600 possible cases of breast cancer across the entire ICS hospital network. In addition, work is underway on other algorithms to detect tumor sites, which could be used as a double-checking tool in the future.

Other projects underway: University of Valladolid researcher Lara Sanogera is developing artificial intelligence that aims to make the tests a patient has to endure to identify and diagnose a brain tumor more tolerable. The goal is that the image obtained from the patient during an MRI scan, which lasts ten minutes, reaches the real image and thus avoids its presence in other types of more unpleasant examinations.

This is how artificial intelligence works to diagnose a tumor

But how do these artificial intelligence tools work? We said earlier that technology is developed by people who program certain information and put it into a tool so that it processes all the data it has and makes a prediction. In the case of cancer research and algorithm development, scientists had to include thousands of studies, images, samples, patient information, tumor details and general characteristics.

In a case published by the University of North Carolina, researchers trained artificial intelligence using images of mammogram samples is compared with the pathologists’ final reports and supplemented with patient demographic information: age, race, tumor type and size, etc.

In the case of the US National Cancer Institute, doctors used “existing rules regarding What does prostate cancer look like on an MRI?” and prepared an algorithm based on “thousands of MRI studies,” some of which were done on people with cancer and others on people without it.

In the case of the Catalan Institute of Health (ICS): More than two million crystals were digitized using histological preparations (they are microscopic samples of tissues and cells) that make up a large database of cases for research and teaching. The samples belong to the pathological anatomy services of the seven ICS hospitals with this specialization: Arnau de Vilanova de Lleida, Bellvitge de l’Hospitalet, Doctor Trueta de Girona, Hermans Trias de Badalona, ​​Jeanne XXIII de Tarragona, Verge de Sinta de Tortosa and Val d’Hebron in Barcelona.

According to ICS, the sample set represents the largest pathology network in Europe, as 168 specialist pathologists at these hospital centers can share cases, diagnoses and knowledge in real time.

Is artificial intelligence the future in cancer research?

“This system makes biomarker scoring faster, safer and reproducible.” These are the statements of Jordi Temprana, junior doctor of the pathological anatomy service of the Vall d’Hebron University Hospital, who emphasizes that “the addition of quantitative analysis algorithms makes diagnoses are more reproducible and provide greater safety for patients“.

Without a doubt this “big promotion” because it allows diagnose breast cancer “earlier”– the doctor assures in an interview with the newspaper La Ciutat Onda Cero. In conversations with Avangard, the doctor explains that these algorithms “behave as if you have ‘super expert’ always ready to make a decision.

According to recent advances, artificial intelligence is being used in cancer research to:

• Make a more accurate diagnosis.
• Analyze large genomic datasets.
• Detect the tumor earlier.
• Make a prediction forecast.

And this clearly means that this technology and its application in the field of research against any type of cancer – even if it is now in its early stages – It will get bigger and the way we work will change.. Some experts even dare to think that can be extended to treatment: Through the use of these algorithms, it is possible to analyze patients’ clinical and genomic data and develop case-by-case or more specific treatment responses.

One thing is clear: the combination of human intelligence with the powerful analytical ability of artificial intelligence will mean revolution in cancer research.