Image Source: Pixabay

It can be an immense challenge to get any pharmaceutical product across the finish line and attain the Holy Grail that is U.S. Food and Drug Administration (FDA) approval. It takes multiple rounds of research and meticulous studies to determine whether a drug is even a viable option for clinical trials. Once in clinical trials, these pharmaceuticals must be tested on animals and humans, often in long-term, real-world studies to determine whether they are truly safe and effective enough to reach the market. High tech solutions like artificial intelligence are remaking the conventional approaches to cancer drug development, though. Could their entry into the market signal a big opportunity for investors?

The unintended consequences of the simplification of cancer research

Cancer is one of the most complex diseases that impacts humankind today. In fact, cancer is so complex that the world’s brightest oncology researchers and medical experts made a conscious determination in the early 20th century to simplify cancer so as to better study it.

That simplification and subsequent research led to the development of immortalized cancer cell lines, the NCI-60 that are still used today for medical research and cancer drug development. And while the NCI-60 cancer cell lines helped us break new ground in the fight against cancer and continue to be used in cancer research worldwide, they are wildly insufficient for efficiently discovering and developing new drugs to treat cancer, let alone pass muster with the FDA. Given that a single drug costs hundreds of millions of dollars to research and develop just to have it die in clinical trials, pharmaceutical companies have been playing one big expensive game of trial and error for the past century.

Today, though, it’s no longer necessary to simplify cancer. We can take all the insights we gained through the research of the past 120 years and combine it with new models of research and new technologies -- specifically AI and machine learning -- to examine cancer in all its complexity and then determine how to best eliminate cancer altogether.

The value of artificial intelligence in the cancer therapeutics market

It’s a common refrain that artificial intelligence (AI) and healthcare are a natural fit for one another. It should come as no surprise, then, that the healthcare AI expected to grow to $61.59 billion in value by 2027 -- that’s a hearty compound annual growth rate (CAGR) of 43.6%. That’s because AI has obvious roles in healthcare research, drug discovery, drug development, and informing new approaches to research.

In cancer, this is especially true. The entire process from start to finish to develop a single cancer drug, from research and development to FDA approval, costs nearly $650 million on average. That is an investment with no guarantee of a return, as drug development can hit a fatal roadblock at any given time. As a result, from 2000 to 2017 there were only 80 cancer drugs approved by the FDA. 

Unfortunately, that means investing in oncology biopharmaceuticals is a high-risk, high-reward venture. But the emergence of artificial intelligence and novel approaches to cancer research could soon change that.

For those lucky few companies that do attain FDA approval, though, the payoff is immense; the median revenue for approved cancer drugs is $1.658 billion. And given that the entire market for cancer therapies is worth more than $112 billion and projected to eclipse $215 billion by 2030, it is a tempting space for many investors. Luckily, there is reason to believe that the process of drug discovery and development are about to get a whole lot faster and easier, thanks to the advent of AI.

How artificial intelligence is informing pharmaceutical drug discovery

Cancer is about more than just the tumor, it’s about the patient. Not only are tumors heterogeneous masses that contain multiple types of cells (not just one), they exist in complex human beings with different genetics, lifestyles, environments, and other factors that make significant differences in how cancers behave. 

To understand how to eliminate cancer you first need to understand the tumor, all its constituent parts, how it behaves within the patient’s body, and detailed information about the individual patient. We lose all that contextual information when we immortalize cancer cells in a lab, but combining tech with new research methods can yield new insights. 

For example, TumorGenesis is a POAI company that is able to extract living tumors from patients and grow them in conditions that mimic the human body. By studying tumors and the cancer cells that comprise them in this way, we can observe how they will actually respond not just in the human body, but in a particular type of patient. 

That data can then be funneled into the massive dataset maintained by Helomics, which includes 150,000 patients, 131 tumor types, and 30 different types of cancer. The database also includes patient demographics like ethnicity, height, weight, age, whether they smoke or drink, and other environmental factors that could influence the growth of tumors within the patient’s bodies.

With all the data housed in Helomics, machine learning algorithms can compare known drug formulations against all the contextual information that impacts how cancer cells grow (and how to destroy them all the first time.) AI does this by testing known drug formulations against a specific type of tumor in a specific type of patient. After running countless scenarios around the clock (at a rate much faster than humanly possible) the algorithm will identify the optimal formulations available to treat the specific type of cancer in question -- all much more quickly than traditionally drug discovery processes.

How machine learning improves drug development

So, artificial intelligence can streamline the drug discovery process, but how does it actually help pharmaceutical manufacturers create medications that are tailored to specific types of cancer in specific types of patients -- not to mention, improve their odds in FDA clinical trials. That’s where Soluble Biotech comes in. Using a novel process called High Throughput Self-Interaction Chromatography (HSC^TM), Soluble Biotech is able to improve solubility of vaccines and drugs, increasing bioavailability while retaining stability. 

This offers opportunities for companies developing all sorts of drugs for a variety of ailments, including cancer. For example, companies like Johnson & Johnson (JNJ), Pfizer (PFE), and Roche (RHHBY) are widely known at the moment for their COVID-19 vaccine, but are also major players in cancer drug development. In 2019, Johnson & Johnson invested more than $351 million in the space, while Pfizer invested $239 million. Roche was a close third, investing $236 million. With machine learning to streamline drug discovery and development, every one of those dollars can go a lot farther.

This allows pharmaceutical companies to create more effective cancer therapies based on optimal formulations that AI has already confirmed to be worth the time and money exploring as candidates for clinical trials. It also expedites the average time it takes for virtually every step of the process, and as we all know time is money (especially for pre-revenue pharmaceutical companies!)

Without machine learning, we’d still be in the 20th century mindset of simplifying cancer. Instead, because of AI, we can finally meet cancer on a level playing field and use these insights to finally eliminate it, once and for all.