We have developed a universal cancer biomarker with the potential of matching virtually all cancer patients to virtually all cancer drugs. Our biomarker is based on the simple fact that all cancer drugs work by preventing cancer cells from growing. We use a proprietary technology developed at MIT (called the Suspended Microchannel Resonator) to measure the weight change (which we call the mass response) of a patient’s ex vivo cancer cells in response to a wide variety of cancer drugs. We use this mass response to predict which cancer drugs will work for the patient. Mass response is a universal biomarker, as it can be measured for virtually all patients, regardless of their type of cancer and regardless of their genomic and proteomic profile, and for virtually all drugs, regardless of their mechanism of action.
The fundamental property that we measure is the “mass distribution,” a histogram of the masses of approximately 2,500 cancer cells. This histogram is created by weighing each individual cancer cell in the sample and plotting the number of cells at each mass.
When an effective cancer drug is applied to a collection of cancer cells it disrupts their growth. In the figure, the cells change mass in response to a cancer drug, so their mass distribution is altered (the blue shape). Our biomarker (mass response) is a measure of the statistical difference between the mass distribution of cells without a drug applied and the mass distribution of cells with a drug applied. If the two distribution are statistically the same, then the cancer cells did not respond to the drug, and we conclude the drug is unlikely to be effective for this patient. If the two distributions are statistically different, then the cancer cells responded to the drug, and we conclude the drug is likely to be effective for this patient.
The Suspended Microchannel Resonator
Our ability to measure small changes in the mass distribution of cancer cells is due to the extraordinary sensitivity of the Suspended Microchannel Resonator (SMR). The SMR, developed in the Manalis Lab at MIT, is a MEMS device that can weigh individual cells to a precision better that one part in 1,000, which corresponds to a change of less than 50 femtograms in the weight of a cell, or approximately 5 nanometers in the diameter of a cell. The SMR is up to 100 times more precise than the best alternative technology for weighing single cells.
The SMR uses the principle that the resonating frequency of an oscillating cantilever (a diving board) is proportional to its mass. The SMR consists of a cantilever with a tiny fluidics channel inside, surrounded by vacuum. When a single cell flows through the channel inside the cantilever, its resonant frequency changes in proportion to the mass of the cell. Using circuitry similar to what is used by FM radios, we can measure the change of the resonant frequency of the cantilever with a precision of 10 parts per billion, which gives us an extraordinarily accurate measurement of the mass of the cell. The figure includes a photo of an actual SMR MEMS chip next to a penny.
Our Clinical Platform and Workflow
The figure below shows the flow of cancer cells and their measurements through our clinical platform. Cancer cells are collected by physicians from cancer patients using blood draws, bone marrow aspirates, or fine-needle aspirates, depending on the tumor type. The live cancer cells are shipped overnight to our lab in one of our temperature-controlled shipping kits. We purify the cells and divide the sample into multiple wells, adding a single drug or drug combination to most of the wells, and adding nothing the remaining wells (which act as controls). We incubate the sample for approximately 15 hours, and then begin the process of measuring the mass distributions of the cells with and without the drugs.
The cell purification process leaves impurities that can interfere with our mass distribution measurements. We use an AI-based image classification system to identify and remove these impurities from the mass distributions. We take a brightfield image of each particle entering the SMR, link the image to the mass measurement, and use a manually curated training set of approximately 20,000 images of healthy cells, dead cells, debris, and clusters to automatically classify each particle. We remove everything except the healthy cells from the calculation of the mass distributions. We are adding fluorescent imaging to further classify cells into subtypes, such as their stage in the cell cycle. In the future we plan to analyze the other particles in non-purifed samples, which make up part of the tumor microenvironment.
Once we have measured the masses of all of the cells in the sample we calculate the statistical confidence (the p-value) that the mass distributions with and without drug are different, and map these p-values to a score of 0-100, where scores less than 50 correspond to “no drug response” from resistant cells, and scores greater than or equal to 50 correspond to “positive drug response” from sensitive cells. We automatically generate a report that includes this drug sensitivity information along with a variety of quality control and other measurements we make during the process and return the report to the ordering physician.
Sample Heterogeneity, Sample Size, and Turnaround Time
Because the SMR provides such exquisitely accurate measurements of the mass of each cancer cell, we can detect very small shifts in the mass distributions. We expect this sensitivity to make our universal biomarker uniquely effective for heterogenous cancers, in which multiple sub-clonal populations are present in the same tumor, and only a small fraction of the cells in the sample will be sensitive to any given drug.
The exquisite sensitivity of the SMR leads to two other clinically important characteristics of our biomarker. First, the biomarker can be accurately measured from a small number of cells, requiring only ~2,500 cancer cells per drug tested. This enables our biomarker to be measured for panel of 20 different cancer drugs from as few as ~50,000 cancer cells. Such a small sample can be collected using a safe and inexpensive fine needle biopsy, avoiding a less safe and more expensive core needle biopsy or surgical resection.
Second, the exquisite sensitivity of the SMR enables rapid detection of the response of cancer cells to cancer drugs. For some drugs (such as bortezomib) the change in mass distribution can be detected in as little as three hours after exposure to the drug. For nearly all the 50+ FDA-approved drugs we have tested, the change can be detected in less than fifteen hours. This enables a two-day turnaround time for measuring the response of a patient’s cancer cells to a large panel of cancer drugs: one day to ship the sample to our laboratory, and one day to run the test.
Our Clinical Study
Travera is studying the use of its test in patients with relapsed refractory multiple myeloma (RRMM). We are working with leading cancer experts at prominent academic institutions to enroll study participants. This study collects bone marrow aspirate samples from patients prior to the start of a new treatment regimen for the purposes of prospectively measuring single-cell weight response as a biomarker of patient response to that regimen.