Cancer and AI Startups
"Cancer is very democratic in the sense that it attacks people regardless of their race, their gender, their national background, or their political persuasions." David H. Koch
According to the World Cancer Day the key cancer facts are that 10 million people die from cancer every year, at least one third of common cancers are preventable, cancer is the second-leading cause of death worldwide, 70% of cancer deaths occur in low-to-middle income countries, millions of lives could be saved each year by implementing strategies for prevention, early detection and treatment, and the total annual economic cost of cancer is estimated at $ 1.16 trillion.
For all the above reasons, AI and ML techniques are breaking into cancer research and oncology, where the potential applications are vast, and include early detection and diagnosis of cancer, subtype classification of cancer, optimisation of cancer treatment and identification of new therapeutic targets.
In particular, AI is predicted to change cancer health care by advancing clinical research and drug development. And besides cutting costs, improving trial quality and reducing trial times by almost half, AI is predicted to find novel cancer biomarkers and gene signatures, recruit eligible clinical trial patients in minutes and read volumes of text in seconds. Moreover, breakthrough discoveries involving new diagnostic tools for cancer have seen AI as a major player.
Right now hundreds of startups and corporate unicorns are all working together to find new diagnostic tools and treatments for cancer using AI in drug development and clinical research in all different cancer areas such as:
Cancer Diagnosis: Digital Pathology and Medical Imaging
Boston Massachusetts-based PathAI, provides AI-powered research tools and services for digitising and analysing pathology images in order to make safer and more affordable the sub-typing of diseases like breast cancer. PathAI and Roche announced a partnership last October and the collaboration will begin with distributing PathAI’s research-use-only algorithms through the cloud version of Roche’s uPath software, with a focus on immuno-oncology for the treatment of multiple types of cancers.
Deep Lens (US), a digital healthcare company focused on a groundbreaking AI approach to faster recruitment of the best-suited cancer patients to clinical trials, is extending one of the world’s first digital pathology cloud platforms (VIPER, Deep Lens’ integrated cloud platform) that combines lab, EMR and genomic data to match cancer patients to clinical trials and precision therapies at the time of diagnosis, accelerating recruitment and bring game-changing therapies to market sooner. Deep Lens is committed to expanding oncology research by making trials more accessible to a larger and more diverse population.
Deep Lens and Hematology-Oncology Associates of Central New York (a private comprehensive cancer centre) just announced that they have entered into a collaboration to expand the clinical trial program of Hematology-Oncology (HOA CNY’s). In particular, Deep Lens’ VIPER will pre-screen all patients from HOA CNY’s EMR (OncoEMR) and integrate molecular data feeds from other sources as well as all pathology feeds to automatically identify qualified patients for clinical trials.
Proscia (US), that makes digital pathology software and AI applications for cancer diagnosis and considered a leader in digital and computational pathology solutions, just announced a partnership with Datavant (leader in helping organisations securely connect health data), that will provide life sciences companies with digitised pathology data to power the development of novel therapeutics and diagnostics. This partnership will bring together Datavant’s privacy-preserving connectivity technology and the largest U.S. health data ecosystem with Proscia’s Concentriq® digital pathology platform.
Merantix is a Germany-based AI research and incubator lab, building ML companies in various fields, and MX Healthcare one of its spinoffs has trained its AI model on more than a million mammograms from partnering radiology offices and hospitals to detect irregularities and signs of cancer with reliable accuracy. The company aims to develop a cloud-based, on-demand platform that will put its cancer-detection AI at the disposal of radiologists across the world.
Aidence, an Amsterdam-based startup has developed a medical image analysis software based on deep learning, the Veye Chest, that is already deployed in 10 hospitals and helps radiologists detect and report pulmonary nodules on a CT chest image to detect lung cancer. Last summer, Aidence and Heart&Lung Health (H&L), a specialist in chest and cardiac radiology reporting, have partnered to provide AI-powered scan reporting for lung cancer screening in Tameside, Glossop, and North Manchester Clinical Commission Groups (CCGs).
The list of the AI diagnostic companies is long and they all share a common problem preventing them for reaching their full potential, namely the quality of the data they use. AI can interpret image scans or pathology images only if the data fed to the system is of good quality and in a large amount. Hence, if the quality of data is poor, AI systems will generate inaccurate and biased results. Moreover AI performed on retrospective pathology studies — using information on events that have taken place in the past — can also have its limitations.
Try now to imagine a human organ scanned for cancer as a movie made up of 100 frames, with each frame being a single static image of the whole organ. Pathologists and their AI use fixed (chemically modified) slides (pathology images), that depict only one microscopic part (i.e. one part out of 1000), of one only single frame (i.e. frame 6, time=6) in order to detect cancer.
Radiologists on the other hand and their AI are using the entire single frame (i.e. frame 6, time=6) — made up of all 1000 microscopic parts (6i, 6ii,…, 6m). So, in order to have a very intelligent AI we will have to train it with all the microscopic parts (1000) of all single frames (1, 2, 3, …,100) of an organ, and this is what the brain does when monitoring the health of our internal organs.
So, in the near future or we will have to build a “Prêt à Porter AI platform” that mimics our brain for monitoring our internal organs or we will have to build an AI that simply “chats telepathically” with our brain while our brain monitors our internal organs…and I guess that future doesn’t seem too far away.
Cancer Blood Screening
The phrase "prevention is better than cure" is often attributed to the Dutch philosopher Desiderius Erasmus in around 1500, but prevention is now synonym to "it's cheaper too", since preventing future illnesses and complications is vital to the future sustainability of health systems and households as well.
For that reason, the cancer prevention company Freenome (USA), a biotechnology company based in South San Francisco, has pioneered the most comprehensive multiomics AI platform for early cancer detection for screening blood to identify future cancers.
In particular, the Freenome platform uses ML to analyse genomic, transcriptomic, methylomic and proteomic data in a routine blood tests to spot the earliest signs of colorectal and other cancers (i.e. pancreas), and is not limited to look for mutations instead it detects changes in gene expression, immune activity and cancer-related proteins among billions of circulating cell-free biomarkers.
Freenome, just closed another venture capital megaround (series D funding $300 million from GV Roche), sending its total fundraising to more than $800 million in just seven years.
Cancer Drug Discovery and Development
A drug molecular discovery focused company advancing AI technology is Atomwise, using deep learning to analyse molecules through simulation that eliminates time taken by researchers to synthesise and test compounds.
In particular, one of Atomwise's top AI strategies is the collaboration with FutuRx to launch A2i Therapeutics to develop novel small molecule immune-oncology agents. A2i Therapeutics Ltd. is a joint venture designed to utilise Atomwise’s AI platform (AtomNet® technology) to target ADAR1 (RNA-specific adenosine deaminase 1), a key protein involved in controlling the innate immune response, neurological disorders, metabolic diseases and potentially a key target for immuno-oncology and neurological homeostasis.
French drugmaker Sanofi SA just partnered with UK-based AI-driven startup Exscientia to develop up to 15 drug candidates across oncology and immunology, in a deal worth up to $5.2 billion in milestone payments.
Numerate is another data driven drug design company applying AI to drug discovery. Numerate agreed to lead discovery programs for identifying clinical candidates in Takeda’s core therapeutic areas of oncology.
Structure-based drug discovery (SBDD) approach complemented by the use of AI technology is the primary focus of a partnership between Korean biopharmaceutical company CrystalGenomics and AI-driven technology company Standigm aiming at discovering and developing novel drugs to treat cancer and liver related diseases. Standigm has just secured a strategic investment from SK Chemicals for expanding its AI-Drug Development Capacity.
All the AI drug discovery companies they are all facing bigger challenges than the AI image screening companies, since they have to screen trillions of different molecules testing them on hundreds of different types of human cells (200 different types of human cells) expressing hundreds of different types of receptors and other proteins, not to mention that every human is different and human cells still evolve.
Considerations on the number of possible molecules to test has led to the concept of the “chemical space” (known and unknown) to describe the ensemble of all organic molecules to be considered when searching for new drugs.
Whereas the known chemical space including public databases and corporate collections probably contains 100 million molecules, it has been estimated that the unknown virtual chemical space might contain as many as 10⁶⁰ (novemdecillion) compounds when considering only basic structural rules, or a more modest 10²⁰–10²⁴ molecules if combination of known fragments are considered. In comparison the number of sand grains on Earth is about 7.5 x 10¹⁸.
These estimations suggest that the entire chemical space is far too large for an exhaustive enumeration, even using today’s computers. One is therefore left with a partial, targeted enumeration as the only option to produce molecules for virtual screening.
Deep 6 AI, utilises AI to mine medical records to accelerate finding and recruiting patients for clinical trials within minutes. For example, Deep 6 AI’s software found and validated 58 eligible patients in less than 10 minutes while a principal investigator — using traditional recruitment methods — found 23 eligible patients in six months for a biomarker for a non-small cell lung cancer trial.
Deep 6 AI, uses Natural Language Processing (NLP) to read doctors’ notes, pathology reports, diagnoses, recommendations and to detect hard-to-find lifestyle data, such as smoking and activity history. Moreover, up to 90% of patient data is unstructured and much of it is siloed across different systems, however Deep 6 AI has developed cutting‑edge technology that pulls insights from this chaotic data within minutes!
In 2019, the company received $17 million in fresh funding in the company’s Series A and today, it has built up an expansive ecosystem of hospitals and patient data, providing "the connective tissue" (as Wout Brusselaers, CEO and founder of Deep 6 AI, said) between all the stakeholders in the clinical trial, such as research staff, care staff, patients, collaborating partners and sponsors, and empower them with a holistic view of the patient based on real‑time data.
The list of these AI clinical trial companies is long and are all facing three major challenges: 1) accessing patient data, 2) accessing good quality data, and 3) creating industry-wide data standards. One last challenge is that these companies need to include patient data in the broadest possible sense and from a wide range of sources including mobile devices, wearables and more from factory tools to kitchen appliances — from healthy populations and not just cancer patients and anywhere on earth.
Accordingly, all kind of data (even clinical) will be overlaid with a sensor’s geographic location, which can then be used to help identify and understand spatial patterns, behaviours and life style.
Analysing research literature, publications and patents. Data mining (biomedical, clinical and patient data)
UK-based BenevolentAI is using AI for scientific data mining, data contextualisation and deriving hypotheses, and data hunters companies like Benevolent are predicted to be Big Pharma’s next prey. For example, technicians at BenevolentAI realised during the pandemic that by running patients’ medical history and previous trial results through their algorithms, Baricitinib, an arthritis treatment, might also help Covid-19 sufferers. That means that ML tools could get a treatment to the market many years earlier, than the traditional 10-12 years drug lifecycle to get a drug developed, tested, approved and on the market. And since time is money in the pharmaceutical industry, Big Pharma is already buddying up with AI tech firms.
Concerto HealthAI, a market leader for Real-World Data (RWD) and enterprise AI technology solutions for Precision Oncology with the most comprehensive, representative and independently sourced RWD in the industry, announced a collaboration with Janssen Research & Development (Janssen) that will allow Janssen to access Concerto HealthAI's use-case engineered RWD, enterprise AI solutions and scientific services (eurekaHealth 3.0).
eurekaHealth 3.0 integrates multiple sources to provide a holistic view of the patient journey to support translational sciences, clinical study design and RWD applications for regulatory submissions, and is the only enterprise-grade solution to exceed 21CFRPart11 compliance requirements.
During this collaboration, Janssen will be the first to use Concerto HealthAI's Genome360, a real-world solution in development that integrates clinical and next-generation sequencing data of a patient's cancer.
Finally, through a systematic review-based approach done by Claudio Luchini, Antonio Pea and Aldo Scarpa in order to provide a comprehensive portrait of the current situation played by AI in the cancer area, all AI-based devices that have already obtained an official approval for entering into clinical practice in oncology were investigated by searching FDA official databases (Last access for all documents: 05/31/2021).
Results, documented the presence of 71 AI-associated or AI-associable devices that have already received an official FDA approval, with the majority of approved devices (54.9%) in cancer radiology, followed by pathology (19.7%), radiation oncology (8.5%), gastroenterology (8.5%), clinical oncology (7.0%) and gynaecology 1 (1.4%). Moreover, the vast majority of the approved devices (>80%) regarded the complex area of cancer diagnostics. Here you can find the list of AI-associated equipped medical devices approved by the FDA specifically for oncology-related fields.
Thank you for reading 💙
And if you liked this post why not share it?
#science #health #pharma #AI_drugdiscovery #drugdiscovery #AI #biotechAI #pharma_AI