The Luventix story started almost ten years ago when Dejan Nenov, one of the founders, watched a documentary on TV showing a dog recognizing the presence of cancer by sniffing a person’s urine.
Being both a curious person and an expert in software and electronics, Dejan wondered if these results could be reproduced and improved by a data-driven algorithm. Dejan shared his thoughts with Martin Martinov, a second co-founder and world-class biophysicist specializing in fundamental quantum molecular models for drug discovery.
The underlying Biology pathways made sense. Sickness leads to changes in the body’s metabolic profile, which is reflected in the composition and concentration of molecules excreted in the urine.
To replicate the function of a dog’s nose, Dejan and Martin turned to gas chromatography. Chromatography is a century old, well-understood, and most importantly, very inexpensive technology, especially compared to generic sequencing.
Using gas chromatography, we turn urine into a data file that quantifies the metabolic profile of a patient. This is also known as creating “a digital twin.”
To replicate the function of the dog’s brain, which processes what the dog smells in urine, Martin turned to his core competency - data structures to represent quantum molecular models and the corresponding machine learning models that consume them.
To validate the technology, Martin first conducted two experimental studies, one with prostate cancer and one with bladder cancer patients.
For each disease, he created a computational disease model that takes as an input the gas chromatography data from a patient's urine and produces a prediction for the presence or absence of the disease in that patient.
Using cohorts of 40 and 19 patients, respectively, and a leave-one-out validation (this means training the model with n-1 samples and then testing the nth sample, repeating n times), it was possible to create highly accurate models.
The Luventix tests distinguished between the true cancer patients and the healthy control group with unprecedented 95+% accuracy and between the tumor-positive and tumor-negative samples with a similarly high 95+% sensitivity and specificity.
These early experiments demonstrated the viability of the Luventix technology.
Dejan understood that the core technology could be used to test for dozens, or even hundreds of diseases, using a single urine sample at the one-time cost of running the sample through a gas chromatograph. The computational cost of processing the data is so small that it is immaterial.
The cost of developing a model for a disease is surprisingly low, takes only a few months, and is only gated by the pace of acquiring urine samples from patients newly diagnosed with the disease.
Unlike traditional test development methodologies, Luventix can create disease models in months rather than years or even decades, which today is the accepted norm in the healthcare industry.
The total cost of acquiring a qualified urine sample for a specific disease using a 3rd party Clinical Research Organization is $600.
In a research environment, at low volumes, the cost of gas chromatography per sample is less than $100.
The average number of samples required to create a working disease model ready to screen patients is 500.
This means that the cost of developing a Luventix test for a disease is somewhere between $500K and $1M, which is exceptionally capital efficient.
In 2014 the founders assumed that the Luventix tests would be delivered in the traditional clinical and point-of-care manner and believed that it was essential and necessary to pursue FDA approval. At this time our 3rd founder, George Holmes, was introduced to the project and through his industry contacts at the time, the traditional clinical trial route was recommended as the path to validate the technology. As a result, we were self-funding a clinical trial with a cohort of 700 patients.
However, at the time, computational in-silico methods, algorithmic medical devices, and machine learning were still new, unfamiliar, and mostly unprecedented. As a result, the FDA informed us that our technology was novel and would require a trial of 10,000 patients, as nothing similar had been done before.
The cost projection for a clinical trial of that size required that we raise outside funding. Unfortunately, in the spring of 2015, when we did our first pitches in the Bay Area, we received friendly VC feedback that no one would fund us because a company called Theranos had sewn-up investment in medical diagnostics for the foreseeable future.
The project was put on hold, although research using publicly available data sources continued. Martin kept refining the computational methods and validating them by building models for different diseases. Most recently, he used public data to develop lung cancer models.
In 2021, Dejan was recruited by 1health.io and, as their CTO, for the following two years, was immersed in the clinical laboratory world.
In 2022, our third founder and CEO, George Holmes, became available to join Luventix full-time. George is a senior executive, with a track record of creating Billions of dollars in shareholder value in the public markets.
We relaunched with a reformulated strategy and a new go-to-market plan.
In parallel with seeking FDA approval, under an accelerated breakthrough technology program, we intend to develop the Luventix tests as Lab Developed Tests (LDTs) first.
LDTs do not require FDA approval.
We intend to launch our first Luventix tests in the Direct-to-Consumer channel.
At the same time will begin the process of acquiring CPT (Common Procedure Code) and LCD (Local Coverage Determination) Codes as assigned by the Centers for Medicare & Medicaid Services and getting Luventix tests in-network with private insurance plans.
The DTC model, pioneered by companies like 23&me, Vitagene, Everywell, Grail, Cologuard, and others, fits the Luventix offering well because of the low cost of our tests and the relative ease of collecting a urine sample, compared to a blood draw or stool collection.
The DTC approach offers a fast path to revenue and brand introduction.
Our marketing efforts will be on digital marketing, targeting disease-specific social media groups, treatment, and support communities.
The Total Available Market (TAM) of the clinical lab test market is over $200B, from which the LDT market is over $10B and growing at a 6.58% CAGR (Grandview Research 2022).
Our current average selling price vs. cost models suggest a minimum of 200% gross margin per test.
We anticipate a low cost of new customer acquisition (CAC) of under $50, over 90% retention, and an existing customer repeat test order frequency of 6 to 18 months.
Luventix offers strong differentiators: industry-leading low-cost, non-invasive testing, the option for patients to self-collect samples at home, and the ability to diagnose the very early stages of a disease. In some cases, before traditional tests, biopsies or imaging can identify a problem.
One of the key advantages of the Luventix technology is its versatility. While our initial focus will be to validate our current findings and drive towards building a successful and profitable offering, we can create business models- for non-cancer diseases with equal ease, and the technology can be applied to general wellness assessment,
We intend to launch our products on the 1health.io platform and sell directly to consumers using off-the-shelf shopping carts and payment processing services like Shopify and Stripe.
In the long term, the Luventix technology for efficient disease screening can positively influence global healthcare costs and outcomes. Accessible and affordable early disease detection dramatically decreases treatment costs and increases the likelihood of positive outcomes.
For more information on Luventix, view the video here.
Comments