Photo from SonTech Research Team
Biotech startups are harnessing the power of artificial intelligence to forecast how patients will react to cancer treatments, with the goal of improving drug trial outcomes and customizing therapies for each patient.
As clinical trials and research in areas like genetics and protein studies generate more data, AI is aiding researchers in efficiently analyzing vast datasets to identify patterns that relate to a patient's response to treatment, whether it's a positive or negative one. These startups are employing AI to anticipate the effectiveness of drugs in clinical trials and develop diagnostic tests that assist physicians in making treatment decisions.
According to Terri Shieh-Newton, an expert in cellular and molecular medicine who provides counsel to life-sciences firms at the law firm Mintz, Levin, Cohn, Ferris, Glovsky and Popeo, AI enables researchers to aggregate diverse datasets and mitigate the potential biases that can arise from more limited data collections.
AI alone cannot guarantee success; it requires human intervention to make thoughtful choices regarding the datasets employed in algorithm training," Shieh-Newton stated. She emphasized that human judgment is essential, cautioning that with machine learning, it's possible to veer off course if not careful.
Venture capital firms are placing their bets on the founders' knowledge and their capacity to obtain the necessary data for creating highly accurate predictive tests. Numerous companies have secured substantial funding, with examples including Artera, which announced a $90 million investment in March, Vivodyne, which kicked off with $38 million in November, and Enable Medicine, which unveiled a $60 million financing round the previous year.
According to Elena Viboch, a partner at General Catalyst, AI can facilitate the integration of diverse and intricate datasets, enabling the discovery of pharmaceuticals that would be inaccessible through alternative means. General Catalyst, an investor in Enable Medicine, highlights that AI has the capability to achieve tasks that conventional informatics methods cannot.
In September 2022, Artera, also recognized as ArteraAI, introduced a diagnostic test intended to assist in treatment choices for individuals with prostate cancer. To develop its algorithms, the company utilized clinical trial information from tens of thousands of patients dealing with various solid tumors, such as prostate cancer, both in the United States and internationally. ArteraAI gained access to this data through collaborative agreements with academic institutions, government entities, and pharmaceutical companies, as stated by Co-founder and CEO Andre Esteva.
ArteraAI conducts its test on digital pathology slides, created from tumor biopsies obtained from patients. Utilizing this data, along with clinical factors like the patient's age and levels of the prostate-specific antigen protein, ArteraAI delivers both prognostic insights, such as the risk of disease progression or recurrence, and predictive guidance regarding potentially effective treatments. This information was explained by CEO Andre Esteva, who holds a Ph.D. in artificial intelligence.
The test is priced at $3,873 as per the company's listed price, and the out-of-pocket expenses are contingent on the patient's insurance coverage. ArteraAI, headquartered in the San Francisco Bay Area, plans to introduce additional tests for different types of cancer, according to CEO Esteva. Although not mandatory for market launch, the company is actively pursuing approval from the Food and Drug Administration (FDA).
Pangea Biomed, headquartered in Tel Aviv, is collaborating with pharmaceutical companies to assist them in addressing queries like identifying the specific cancers to focus on for clinical trial treatments, as stated by CEO Tuvik Beker.
Targeted cancer medications are developed to address patients who possess particular gene mutations. Testing is conducted on patients to ascertain the presence of these mutations and their potential benefits. However, Beker pointed out that this testing doesn't provide a comprehensive picture.
He mentioned that a patient's response is influenced by numerous other genes apart from the targeted one. Pangea, which has secured $12 million in venture capital funding, analyzes the interplay of multiple related genes to forecast the effectiveness of a drug. Although their current income primarily derives from collaborations with pharmaceutical firms, Pangea's ultimate objective is to offer its technology to healthcare practitioners, assisting them in making treatment choices for patients, according to Beker.
A challenge faced by developers of cancer drugs has been the use of animal models to evaluate experimental medications for diseases, as mentioned by Dr. Alex Morgan, who serves as a partner at Khosla Ventures.
A significant hurdle for cancer drug developers has been the reliance on animal disease models for testing experimental drugs, according to Dr. Alex Morgan, a partner at Khosla Ventures. He further noted that the areas of drug development where we face the most challenges are those in which the animal models inadequately mimic human diseases.
Recently, Khosla Ventures spearheaded a seed funding round for Vivodyne, a startup that combines lab-grown human organs with artificial intelligence to identify proteins suitable for drug targeting and forecast how individuals will respond to medications.
According to Vivodyne's Co-founder and Chief, Andrei Georgescu, rodents possess a level of complexity similar to humans but with enough differences to restrict their applicability as research models. Vivodyne is placing its confidence in the potential of their lab-grown human tissues to produce a more comprehensive dataset, thereby improving the accuracy of drug efficacy predictions.
Enable Medicine, located in Menlo Park, California, utilizes AI to derive insights in the fields of biology and medicine. Recently, the company employed RNA sequencing and other data analysis techniques to uncover characteristics linked to the effectiveness of a specific type of cancer immunotherapy called checkpoint inhibition. Their findings indicated that individuals who did not respond to the treatment exhibited elevated interaction between immune cells referred to as CD68+ macrophages and CD8+ T cells.
Equipped with this data, researchers can delve into inquiries such as the reasons behind the association between these interactions and the absence of a treatment response, as well as potential strategies for reversing this phenomenon, as stated by CEO Sunil Bodapati.
Pepper Bio, based in Cambridge, Massachusetts, adopts a 'transomics' methodology, which involves the comprehensive analysis of DNA, RNA, proteins, and molecular switches within tumors, the immediate tumor microenvironment, and neighboring healthy tissue, as explained by Co-founder and CEO Jon Hu. The primary objective is to identify drugs that specifically target the molecular pathways responsible for initiating or sustaining cancer, while avoiding pathways that could lead to toxic effects. Pepper Bio aims to procure medications that focus on these promising biological pathways in cancer treatments. According to Hu, the pertinent question is not whether AI will impact drug discovery, but rather how and when it will do so.