Novel diagnostics in development for the rapid detection of SARS-CoV-2 and its severity, and the identification of sepsis and its antibiotic resistance profile, are finalists for AACC’s Disruptive Technology Award for 2021. The annual competition recognizes innovative testing solutions that improve patient care.
This year’s award highlights developers of artificial intelligence and machine learning technologies advancing medicine and pathology. The finalists are Day Zero Diagnostics for its DZD OneSeq Dx platform, Mammoth Biosciences for its CRISPR-based DETECTR BOOST platform, and MeMed for its COVID-19 Severity test. Company representatives will compete live for the final prize tonight from 4:30-6:00 p.m. at the AACC Annual Scientific Meeting in the Georgia Ballroom of the convention center.
Two finalists are developing new technologies for the SARS-CoV-2 virus. MeMed, based in Haifa, Israel, is using host-response profiling and machine learning algorithms to create a novel COVID-19 Severity test. It measures three proteins from a serum sample and applies machine learning to stratify the risk that a patient with COVID-19 will experience severe outcomes.
The proteins measured are TNF-related apoptosis-inducing ligand, which shoots up in the presence of a viral infection but drops dramatically when that infection is severe; interferon-gamma-induced protein 10, which is secreted to fight off viruses and infections and is associated with infection severity; and C-reactive protein, a general marker of hyperinflammation, said Eran Eden, PhD, MeMed’s CEO and co-founder.
Serum samples are applied to a cartridge and inserted into the MeMed platform called MeMed Key. This analyzes amounts of these proteins and applies machine learning algorithms to provide a score between 0-100 for each sample. The higher the score, the higher the likelihood of a severe outcome such as death or ICU level of care. The test, which delivers results within 15 minutes, can help clinicians predict outcomes up to 2 weeks earlier, helping them decide whether to escalate treatment, admit patients to the hospital, or send them home.
The test has been approved for use in Europe. The company is in discussions with the Food and Drug Administration to bring the technology to the United States. A study of the test in 394 COVID-19 patients (29% meeting a severe outcome) found the test had an area under the curve of 0.86, according to results posted to medRxiv in early July.
“The term disruptive should be unique and highly impactful on people’s lives,” Eden said. “I think we hit the box on that.”
Mammoth Biosciences also is tackling COVID-19, but from a different angle, offering a high-throughput, CRISPR-based test to identify the presence of SARS-CoV-2.
The company’s DETECTR BOOST platform uses CRISPR, a gene editing technology that has been repurposed for diagnostics, to search patient samples for the presence of specific nucleic acids that are indicative of disease. First, the company pairs a CRISPR- cellular apoptosis susceptibility (CAS) protein with a guide RNA that matches the exact sequence they want to detect. If the CAS/guide RNA complex finds its target sequences, CAS proceeds to cut the matching DNA or RNA. Some CAS proteins also chop up nearby pieces of genetic material including reporter molecules that let scientists know they found their target sequence.
The company is developing a turnkey workstation to run the assay, with the ability to perform thousands of tests per day with minimal hands-on input by a technician. “While a typical reverse transcription polymerase chain reaction run takes around 140 minutes, the CRISPR platform shortens that turnaround time by almost half,” said Janice Chen, PhD, co-founder and chief technology officer of the San Francisco-based company. Reagent kits being prepared by the company can be used on automated liquid handling platforms. The project is supported by the National Institutes of Health Rapid Acceleration of Diagnostics (RADx) program.
“The need for diagnostic testing has been highlighted to the extreme in the last year, and we’ve learned there needs to be a lot more testing solutions that go beyond the laboratory,” Chen said. “CRISPR is going to be critical in the transition to what I call on-demand testing.” The future of diagnostics could bring molecular testing to decentralized settings like retail clinics and the home, and use CRISPR to identify other viruses including influenza or respiratory syncytial virus.
The third finalist, Day Zero Diagnostics, has developed a testing platform that uses whole genome sequencing (WGS) and machine learning to rapidly diagnose infections directly from blood samples without the need for culture growth.
OneSeq Dx uses ultra-high enrichment (at the level 109) of blood samples to identify pathogen WGS directly from clinical samples that may contain up to a billion times more human DNA than bacterial DNA. Once sequencing data is obtained, two machine learning algorithms called Keynome ID and Keynome g-AST (genomic antibiotic susceptibility testing) determine the species of infections at concentrations as low as 1CFU/mL, and their antibiotic resistance and susceptibility profile. Results can be obtained in 6-8 hours.
The company is trying to tackle two related problems in infectious disease diagnostics, explained Jong Lee, MBA, CEO, and co-founder of the Boston-based company. One is antibiotic resistance, what Lee calls “the other pandemic.” For most testing done using culture, it takes 2-5 days to get to an antibiotic resistance profile for an infection. Antibiotic resistance is particularly challenging in bloodstream infections like sepsis, Lee said, with patients decompensating by the hour. Clinicians generally are “flying blind,” Lee said, using an umbrella approach to try to subdue infection using aggregate infection patterns local to their hospital and community.
Through OneSeq Dx, “we’re trying to provide physicians a solution that gives them a definitive species ID and antibiotic resistance profile in hours rather than days, so that they can treat patients on the first day that they’re admitted to the hospital,” he said. “The Holy Grail in this space is to diagnose directly from blood rather than from blood culture.”
WGS data for each patient’s pathogen can be aggregated to help clinicians amass an epidemiological view about what infections are running through their hospitals or local regions, Lee added. “It becomes a surveillance service that can help a lab make early determinations of outbreaks.”