Ethics in Health Tech – Lessons from TOA and NOAH 2019
Since the illustrious demise of one of the most promising, innovative biotech companies of recent times, Theranos, a lot of the discussion in the health tech space has focused on entrepreneurial ethics. The need for strong ethics, especially in healthcare where people’s lives are at risk, cannot be overemphasised. So much so that at TOA Berlin last week, several talks were dedicated to this subject: A Conversation with Erika Cheung, Ethics in Entrepreneurship and Biotech, Health & The Era of Fast Capital on Day 2 of the event.
As one of the key whistleblowers on Theranos, Erika Cheung set up EIE (Ethics in Entrepreneurship), a network and resource for young tech entrepreneurs helping them to face current and future ethical challenges. During her talk at TOA, Erika spoke at length about the massive breaches of trust she witnessed at Theranos during her 7-month long employment but she also mentioned other digital health startups such as uBiome that are “charging for tests they never run”.
Such news is unfortunate and can only be detrimental to the wide adoption of new health technologies aimed at improving people’s quality of life and access to healthcare. However, bad players seem to be the exception rather than the rule and the wealth of health tech solutions on the market increases year after year.
Only at NOAH Berlin last month, I encountered a plethora of new health tech solutions, from AioCare– a personalised system to monitor and treat pulmonary diseases (asthma, chronic obstructive pulmonary disease) consisting of a portable spirometer connected to a smartphone app and a dedicated doctor’s panel for remote tracking of results and secure communication; Biomes – a DNA-based gut microbiome analysis designed to help customers discover their intestinal flora and its effect on their health; Kaia– a multimodal therapy that combines physical, psychological and educational elements to relieve back pain at home, without medication; Keleya– a pregnancy coach app providing yoga workout videos, healthy recipes & expert tips to expectant mums; Klara – a tool for health clinics to help them unify and simplify all patient-related communication; to companies such as Medbelle – a “digital hospital” focused on re-thinking the patient journey from scratch that provides an easy-to-use, integrated experience for patients by doing away with the outdated systems and processes that focus on the provider rather than the patient; myTomorrows -a AI-driven smart knowledge database that provides information on all available treatment options for patients and their doctors; Oviva– a nutrition and lifestyle tool designed to help people prevent or manage their diabetes, as well as weight issues and paediatric allergy; Talea– an operating system for the elderly care market; and Zava (“your doctor on the net”) – one of the largest digital health providers in Europe, providing medical evaluations and prescriptions online as well as fast shipping of medicines.
If the above health tech solutions sound more or less familiar, there is one that stands out by its degree of innovation: Canostix – a Berlin-based health tech startup that is building the future of cancer screening with photonics and patten recognition.
Canostix’s goal is to develop an easy, fast, highly accurate and low cost blood test that enables patients and doctors to recognise cancer at an early stage, when it can be cured. The test uses spectroscopy, which exploits the interaction of light and matter, combined with state of the art data analysis.
Existing cancer screening methods are not sufficient to comprehensively ensure an early diagnosis and recommended screening methods exist for only a few types of cancer.
The Canostix approach differs greatly from contemporary test procedures applied in cancer diagnostics.
Take liquid biopsy for example. This is one of the most promising cancer diagnostics technologies currently in development but there are multiple drawbacks when it comes to its use for early cancer diagnosis. Liquid Biopsy looks at free circulating tumour-DNA (ctDNA) isolated from a pool of circulating cell-free DNA (cfDNA) in the blood, then multiplied (PCR) and subsequently analysed with next generation sequencing (NGS) to accomplish a diagnosis. This method is very promising with regards to analysing the character of the tumour, therapy-guidance and determining the prognosis of the patient. However, ctDNA makes up <1% of the cfDNA and also, a variety of gene sequences (at least 20-50) need to be analysed to make a diagnosis which adds up to a high test fee. Furthermore, mutated DNA can be found in a large percentage of individuals, particularly in the older population (3) but even more to the point, ctDNA can only be detected in <50% of early stage patients and ca. 85% of late stage patients, rendering this technology unsuitable for early cancer screening.
Unlike tumour markers and the analysis of cellular surface receptors, spectroscopy is able to capture the biochemical changes in the blood caused by the disease in its entirety. These changes are mostly reflected in the biological “fingerprint region” of the measured spectrum which appear as specific spectral peaks. Therefore Canostix is looking at the “biochemical fingerprint” of a patient’s blood that is subsequently compared to a reference database containing similar fingerprints from healthy and cancer patients.
The differences will be comprised by changes in concentrations and ratios of naturally occurring components which can be influenced by the presence of the cancer. These are a very heterogenic group of compounds ranging from NAD/NADH+ ratio, which can be influenced by the metabolism of the cancer, to certain amino acids like tryptophan, which is high in some cancers and low in other, all the way to the amount of carotenoids or the RNA/DNA ratio.
The machine learning part of the approach involves analytical algorithms that first filter out the spectral intervals that contain the most significant differences between the groups, then focus on specific diagnostic features and ultimately classify the data as belonging to the cancer or healthy cluster.
The approach combining highly sensitive, semi-quantitative vibrational spectroscopy (photonics) with state-of-the-art machine learning analysis algorithms has already been tested and verified by multiple research groups all over the world with great success.
With experts predicting a rise of 68% in new cancer cases in the next 10 years, health tech advancements such as the one being developed by Canostix can truly save lives: not just improve health conditions.
XPOMET©/Medicinale, the Festival of the Future of Health and Care, will take place between 10-12 October 2019 at Arena Berlin. We very much look forward to welcoming you at our event where we’ll dedicate a whole section to AI & Ethics this year.