Cancer treatment is about to be rebooted thanks to AI


Cancer treatment requires a reboot.

Despite the ambitious goal of the Biden administration’s “cancer moonshot” to halve cancer mortality over the next 25 years, the development of more effective treatments based on genetic information has not progressed.

At the same time, the cost of anticancer drugs is soaring, often exceeding $20,000 per month. JAMA Oncology recently predicted that between 2020 and 2050, the global cost of cancer will reach a staggering $25.2 trillion. Add to this astronomical sum the severe social and human costs of the disease.

AI-based methods can accurately match cancer patients with the most effective treatments.

Currently, overall cure rates for advanced lung, breast and colon cancers are typically less than 50%. Most of these cancer patients need a plan B after all standard treatments have failed. We need a way to match approved or experimental cancer drugs to patients whose tumors are likely to respond to them. At the moment, however, deciding which treatments to try is largely a matter of guesswork, and the success rate when surgery fails to remove the entire tumor is depressingly low.

But thanks to new AI-powered approaches, there is a glimmer of hope. Although the method is still in its early stages, it has already shown promising results by precisely combining the most effective treatments for patients. And interestingly, the conceptual basis for this approach comes from astrophysics rather than oncology.

From stars to cells: Earlier this year, a group of researchers and clinical researchers from Columbia University, Mount Sinai Hospital, St. Jude University, Emory University, and other institutions, including ours, presented a groundbreaking clinical study. natural cancer.

The study used an innovative algorithm called VIPER (an acronym for “Virtual Inference of Protein Activity by Enrichment Regulon Analysis”) to determine whether patients with metastatic breast cancer would respond to a specific drug called licorinostat. demonstrated unprecedented 100% accuracy. . Interestingly, licorinostat targets proteins encoded by genes that do not mutate in cancer cells, leading to the popular belief that genetics is the only way to personalize treatment for individual patients. are questioning the

The VIPER algorithm — co-invented by one of us (Andrea Califano) and DarwinHealth Chief Scientific Officer Mariano Alvarez — starts with: can We measure tumors, but dig beneath their surface to deduce hidden layers of master regulatory proteins that drive cancer cell pathology. This feat was previously thought to be beyond our technical comprehension. By analyzing gene expression profiles, we can decipher the unique patterns of master regulators in each tumor. and Existing drugs that can inactivate them.

Choosing the right drug for a given cancer patient has largely been a guessing game.

Interestingly, this approach has similarities to astrophysics. Astrophysicists have long used inference techniques to confirm the existence of celestial phenomena invisible to the naked eye, such as black holes and exoplanets. By applying these techniques at the microscopic level, we can begin to unravel the hidden causes of cancer cells.

The proteins that drive these cancers are as strange and elusive as black holes. Until recently, measuring its activity has been as difficult as trying to photograph celestial bodies that don’t emit or reflect light, like black holes. As a result, choosing the right drug for certain cancer patients, especially those who do not respond to initial treatment, has become a guessing game, akin to navigating a maze in the dark. This is also a very complicated maze. There are 250 anticancer drugs and 200 types of cancers approved by the FDA, resulting in tens of thousands of combinations of drugs and cancers, literally Billions More when you combine multiple drugs.

The findings reported in natural cancer Research may illuminate this maze. The VIPER algorithm has enabled scientists to infer the activity of a protein called HDAC6, which functions as a cancer initiation switch, regardless of genetic markers or mutations. This highlights the potential for using therapeutics that target proteins not necessarily associated with mutations, which are the most common basis for many expensive cancer-targeted therapies.

Drugs prioritized by the algorithm predicted their ability to achieve disease control with 91% accuracy.

This technique of translating astrophysical techniques into cell biology may sound like science fiction, but it has proven effective. VIPER also drives the drug discovery process (based on a model of cancer biology called ‘Oncotechure’) for our biotech startup DarwinHealth. Works like a space detective.it takes measurable Use data such as gene expression profiles to make precise inferences about the proteins that most likely generated them. Proteins cannot be measured directly, but can be deciphered using reverse engineered biological models.

Further confirmation that the VIPER algorithm can predict drug efficacy was published in the April 16th journal. cancer detection, in a study co-authored by one of us and many other scientists from leading cancer research institutions around the world. Our experiments showed that he predicted with 91% accuracy the ability of drugs prioritized by the algorithm to achieve disease control in human tumors implanted in mice. Notably, the drug favored by the algorithm performed significantly better than the anticancer drug used as a control and predicted to be ineffective.

People with rare orphan cancers may especially benefit.

A more recent study just published in cancer cellshowing that VIPER can also be used to identify drugs that target individuals. subpopulation of cells within the tumor. This is a very important achievement because tumors, like cancers themselves, are not all the same. A single tumor can have dozens of molecularly distinct subpopulations with different drug sensitivities, a phenomenon also known as cancer heterogeneity.

In a commentary published at the same time, natural cancer In this study, a group of prominent external cancer researchers described this innovative approach as a ‘roadmap for advancing cancer treatment’.

In fact, such roadmaps are already having real-world impact. A clinical trial in children with metastatic Wilms and rhabdoid tumors will soon begin enrolling patients. People with these rare orphan cancers who do not have the targeted genetic mutations (including most children with cancer) are especially vulnerable by avoiding the obstacles that affect currently available approaches. You may benefit.The KaDiscover more The report also includes a case study of a 14-year-old boy with a rare tumor that progressed on all previous medications, but the drug predicted by the Cancer Algorithm resulted in two years of durable remission. could bring

Trillions of dollars, countless lives: We believe that today’s high cost and slow progress in cancer care are rooted, in part, in the erroneous assumption that genetics holds the key. This is true for rare genetic diseases, but in complex diseases, including most cancers, the story becomes even more complicated as proteins play a more important role.

Unfortunately, with a few notable exceptions, even with a precise understanding of the genetic make-up of cancer (regular tests performed daily on cancer patients), which drugs and treatments are the most effective? It is not possible to fully determine whether the For cancer patients, this often results in costly and debilitating polypharmacy interventions, many of which are groping in the dark with only modest and temporary benefits. It turns out.

VIPER’s breakthrough lies in its ability to optimize the use of existing drugs.

A new paradigm powered by VIPER bridges the gap between patients and the best drugs to treat them, based on elusive protein targets known as ‘tumor checkpoints’. We believe this strategy could save trillions of dollars and countless lives as it predicts more effective treatments for specific patient populations if this approach is scaled up and adopted widely. thinking about.

VIPER’s breakthrough is to provide patients with an FDA-approved drug that can target a tumor’s unique protein drivers, optimizing the use of existing drugs and eliminating costly drug development processes. It is in the ability to set us free.

Improving drug-tumor coordination enabled by these AI-like approaches provides immediate cost savings. They are promoting the use of approved, affordable and readily available drugs, but importantly, early identification of effective drugs when they are most likely to work, and early treatment failure. to reduce the cost and toxicity of subsequent treatment in some cases.

A comprehensive theory to explain the complexity of cancer remains elusive, but we may be turning a corner with these protein-centric approaches. Cancer algorithms like VIPER can help save lives, improve their quality of life, improve the accuracy of clinical trials, and reduce the estimated trillions of dollars in cancer costs over the next 30 years. There is a possibility. This new strategy is more like a ‘cancer starshot’ than a ‘cancer moonshot’ and offers a way forward in the ongoing battle against this devastating disease.

Dr. Gideon Bosker is the CEO and co-founder of DarwinHealth.

Dr. Andrea Califano is Clyde Wu and Helen Wu Professor of Chemistry and Systems Biology, Director of Systems Biology at Columbia University Irving Medical Center, Director of the JP Saltzberger Columbia Genomic Center, and a Fellow of the American Academy of Medicine. He is the co-founder of his DarwinHealth.

We look forward to hearing from you! If you have any comments about this article, or tips for future Freethink stories, please email us at: tips@freethink.com.



Source link

Leave a Reply

Your email address will not be published. Required fields are marked *