Croce and delight artificial intelligence? It certainly has its limits and its “dangers” but today we know that it can help to defeat antibiotic-resistance, one of the main causes of death in the world, with shock data destined, going on this step, only to worsen
THE’Artificial intelligence It can be of great help to fight antibiotic-resistanceone of the main causes of death in the world, with data shock destined, going on this step, only to worsen. Scientists light up a light of hope.
Antibiotic-resistance
Antibiotic resistance is a problem for the health of world proportions, which could, in the less optimistic scenarios, bring humanity back to die as happened 80 years ago, for infections that then became perfectly curable.
Unfortunately they begin to spread more and more bacteria that do not respond to antibiotics Currently on the market and some first sensitive species begin to resist, becoming potentially indestructible.
According to one Research published in The Lancet, In 2019 over 1.2 million people all over the world are death directly due to an infection generated by a pathogen resistant to antibiotics.
Furthermore, 4.95 million deaths They were related to an antibiotic -resistant bacterial infection, even if the direct cause of death was probably different. In fact, therefore, the antibiotic resistance is already today una of the most common causes of death in the world.
But what did he lead to this situation?
There are several factors that have brought and are leading to multi -resistance to antibiotics. First of all their consumptionboth for humans and for income animals (i.e. those raised to consume them or their products).
The antibiotics saved millions and millions of lives from the discovery of the first antibiotic of history, the penicillinby Alexander Fleming In 1928 (which for this reason he was awarded the Nobel Prize for Medicine in 1945).
But it was he himself who warned that the discovery, if used badly, could be a boomerangunderlining how molecules could also be source of selection of resistant bacteriafor example in case of under dosages (i.e. use of the drug for less time or in quantities lower than those necessary).
Antibiotics must be taken only in cases of effective necessityin order to avoid that their use is not only useless but also harmful as in the case of the under dosage, making the proliferation of pathogens immune to the antibiotic molecule more likely.
But unfortunately, at least for the moment, we are not following these important indications. In fact, the consumption of antibiotics continues to be very high. And not only for human use, but also in the animal world: In the farms, in fact, the drugs are often given “rain”, and not infrequently using the same that we use to treat our infections.
And there is more: in Intensive farmscause of animal suffering as well as significant contributions to global warminginfections are even more likely for obvious reasons of space and general conditions. And here the antibiotic resistance finds even more space as the use of drugs is more massive.
Let’s not forget that we then consume the same animals so treated or their products (milk, eggs, etc.) and that the molecules eliminate with their excrements However, they end up in the food chain by passing through the soil and/or waters.
Actually A 2014-2018 monitoring study of the European Center for Disease Prevention and Control (ECDC) He verified that the consumption of antibiotics in income animals is falling, but it is not at all for human consumption.
And a subsequent Study always of the ECDC He also showed that the consumption of antibiotics is resumed even after the drop observed during the Covid-19 pandemicindicating that not, we are actually not learning anything.
How artificial intelligence can help
With all its shadows, l‘Artificial intelligence (Ia) It can be of great support in many scientific fields, also in that of antibiotic resistance.
This, in fact, could be fought by looking and experimenting with other molecules, but pharmaceutical research is long and very expensive: a new antibiotic can also cost $ 1 billion And his arrival in pharmacies could Employment also 10 years.
The Last But Not The Leastpharmaceutical companies need to return to these costsbeing realities for profitbut the indications to combat antibiotic resistance are precisely contrary: you have to use them as little as possible, only in cases of effective necessity.
Here the IA comes into play: as a study by review led by Jiangxi Cancer Hospital & Institute (China), this can in fact used for predict antibiotic resistance epidemics.
Different models of IA, in particular, can be trained with clinical information, genomic sequences, information on the microbioma and epidemiological data, providing Ideas on the discovery of new antibiotics e On the reuse of some existing, as well as on the therapy combined through the analysis of their molecular structures. This can Reduce very times and costs of preclinical research of much time and costsmaking the last phases sustainable before putting on the market.
In addition, clinical decisions support on real time can guide healthcare professionals to Improve the prescription of drugsstill today, as we have seen, rather abused.
(Some IA models) are accelerating the discovery of new antibiotics – confirms a review work published last January – helping to strengthen the preclinical pipeline of antibiotics (…) in fact, with the expansion of the integration of the AI, its role in the improvement of the infections treatment regimes and in the improvement of the management of antibiotics is intended to become indispensable in the fight against antimicrobial resistance
Certainly, there are still limits, as the scientists themselves admit, including the accessibility of the data, which affect their ability to effectively work in different populations of patients and clinical scenarios. In addition, ethical issues such as the Informed consent, data confidentiality and algorithmic bias.
Artificial intelligence models should be developed using diversified and representative data set To avoid perpetuating i bias and guarantee equivalent and accurate results for all patient groups
Currently, in fact, the IA models for the diagnosis, treatment and discovery of new antimicrobials against antibiotic resistance are often trained on sets of unbalanced data, which can suffer from poor reliability.
But this is the direction, according to the scientific community, that it is necessary to take.
Sources: Computational and Stutcutural Biotechnology Journal / Nature
