September 22, 2022 | Emerging approaches to the early detection of cancer, based on metabolic profiling and microbiome- and exosome-driven liquid biopsies, were highlighted during the recent Next Generation Dx Summit held in Washington, D.C. Representatives from diagnostics companies Oxomics, Elypta, Micronoma (University of California, San Diego, spinoff), and Rivela Diagnostics (an Exosomics company) took turns making their case for a new generation of technologies hunting for elusive molecular signatures of a highly heterogenous group of diseases.
First up was James Larkin, Ph.D., founder and CEO of Oxomics, with his pitch for metabolic profiling using nuclear magnetic resonance (NMR). “Metabolomics is what all omics exist to alter,” he points out, and has long been known to have a cancer connection.
Metabolic analysis is straightforward, and the required blood and urine samples are typically readily available, says Larkin. Oxomics uses NMR as a “dense information source” for algorithms predicting patient disease outcomes.
He describes NMR as a “molecular tuning fork” that can be adjusted to the radio-wave frequency of protons. Metabolites in the blood run at different peaks, and that physical phenomenon creates patterns that make NMR “intrinsically quantitative.”
A machine can process 70,000 samples per year, says Larkin. And, he later adds, there is less “art” to producing good data with NMR than is required of high-throughput mass spectrometry.
NMR can not only diagnose cancer but do it earlier in a scalable and cost-effective manner, Larkin continues. Moreover, it is unbiased because “you don’t need to know in advance what you’re looking for.”
In mouse models of brain metastasis, Larkin and his colleagues showed leakage of a contrast agent across the blood-brain barrier once tumors grew large enough to breach the protective border. Based on scatter plots, it was possible to see strong separation early on—“even day five”—between healthy and diseased mice, he shares. By day 10, when tumors would likely not be seen by analyte, the team achieved “near-perfect classification” of metastatic cancers.
Since metabolic profiles differ depending on where a tumor is growing in the body, NMR can indicate a cancer’s location as well as its size, Larkin says.
In the UK, outcomes have tended to be poor (60% one-year mortality rate) for patients with emergent presentations of cancer with nonspecific symptoms who are already quite seriously ill, he adds. For physicians, “the choice is to guess or refer, or... the watch-and-wait approach.” This gives cancers that exist time to grow.
In Oxford, UK, the experimental Suspected Cancer (SCAN) pathway was created to deal with these non-specific symptoms, Larkin says. A CT scan and blood and fecal testing are done, and patients get routed to a multidisciplinary team for a diagnosis. The program “works to some extent but is ripe for disruption.”
To that end, Larkin and his team “piggybacked” on the SCAN pathway, recruiting 300 patients into a study using metabolomics to identify those with cancer (Clinical Cancer Research, DOI: 10.1158/1078-0432.CCR-21-2855). Just under 10% of patients had a cancer diagnosed by the SCAN pathway during the study period, and nearly half of those cancers had metastasized. About one-third of participants still did not get any diagnosis.
Researchers separated patients with and without a solid cancer diagnosis and compared the metabolites in their blood, which clustered to opposite sides of a statistical graph, he says. Although a strong case could not be made for NMR-based metabolomic analysis of hematologic malignancies, the overall maximum sensitivity and specificity were 94% and 82%, respectively. An exhaustive validation exercise suggests “the models are better than chance at identifying cancers” while ensuring researchers were not overfitting the data.
In one small case study involving patients followed for up to a year after receiving a noncancer diagnosis in the initial 62-day diagnostic window, two of the five patients who developed a solid tumor were identified as having cancer by metabolomics assessment of baseline blood samples. Notably, recurrence of a gastric cancer was detected seven months before the diagnosis was finally confirmed.
The vision of Oxomics, Larkin says, is a simple and rapid blood test-based platform technology enabling early identification of cancer and better outcomes for patients, including detection of tumors missed by conventional assessment.
Next up was Francesco Gatto, Ph.D., chief scientific officer and co-founder of Elypta, presenting on metabolism as an information layer for multi-cancer early detection–specifically, the role of glycosaminoglycans (GAG). Recent evidence suggests metabolic reprogramming of so-called GAGomes is a ubiquitous process in cancers.
Multicancer early detection (MCED) is ideal, but screening currently happens for only a handful of cancers and is modality-focused (e.g., colonoscopy or CT lung scan), he says. Many features of cancer get shed into the blood, including circulating tumor DNA (ctDNA). But the approach is not sensitive enough for MCED because many tumors do not shed any methylated ctDNA into circulation.
As reported in the Annals of Oncology, (DOI: 10.1016/j.annonc.2021.05.806), sensitivity of ctDNA for stage 1 cancer detection is 16.8%, which is “not bad,” says Gatto. But in real-world settings, this figure drops sharply, he adds.
By comparison, the multicancer liquid biopsy tests of Thrive (now owned by Exact Sciences) and GRAIL were able to achieve, respectively, 7.5% and 11.5% sensitivity at stage 1, he continues.
In a 2016 paper published in Cell Reports (DOI: 10.1016/j.celrep.2016.04.056), Gatto and his colleagues used genome-scale metabolic modeling to learn that GAG biosynthesis is uniquely regulated in renal cell carcinoma (RCC) and that diagnostic blood and urine GAG scores accurately predict the occurrence of metastatic disease. Another clinical study conducted at the Memorial Sloan Kettering Cancer Center subsequently confirmed this finding in non-metastatic RCC independent of disease stage.
Elypta has an in vitro diagnostic, MIRAM, which renders a GAG profile from blood using liquid chromatography–mass spectrometry in under 10 minutes, says Gatto. The kits have demonstrated acceptable precision and inter-lab reproducibility.
In a study published earlier this year in Journal of Biological Chemistry, (DOI: 10.1016/j.jbc.2022.101575), researchers established the normal levels of GAG in healthy adults across Sweden, says Gatto, and found that GAGomes have only a “small dependency on age.”
In a mouse study, changes in the GAGome were found to play a strong mechanistic role in the development of metastatic cancer, he notes. Unpublished work also suggests GAG scores correspond to the presence or absence of 12 different cancer types (excluding bladder and prostate cancers) and, importantly, the biggest jump in physiological level is seen in stage 1 disease when the information is most clinically useful. The “appealing hypothesis” is that any undetected cases would involve individuals who are still alive a year later for their next test.
In a study looking at the blood of 167,000 healthy individuals in the Netherlands who did or did not get cancer over 18 months, GAG scoring flagged 25% of the cases, reports Gatto. But the approach also missed cancers in people who did not get a high GAG score, which was “good information for us.”
Might the Elypta test be a complement to the Thrive and GRAIL tests? Used in combination, he openly wonders, might this raise sensitivity to almost 40% for stage 1 cancers and is that good enough?
Standard-of-care modalities detect one cancer per 280 screened individuals between the ages of 50 and 79 and produce twice as many false as true positives, Gatto says. Modeling indicates that multicancer early detection testing would identify 350 cancers—240 of them diagnosed at an earlier stage—and allow 62 additional individuals to survive the disease per 100,000 people screened.
Enhancing early-stage cancer detection via microbiome-driven liquid biopsy was the subject of another talk by Sandrine Miller-Montgomery, PharmD, Ph.D., CEO of Micronoma. The microbiome is the latest addition to the “hallmarks of cancer,” she notes, and constitutes 92% of life on earth and most of the genetic diversity. Microbes can adapt to even the most extreme environments, including the lungs.
Micronoma’s focus is on the body’s interior tissue and blood compartment, says Miller-Montgomery, to help translate what is happening with the oral, skin, and gut microbiome on the clinical level. Human tissue is brimming with microbes and not sterile even when uninfected. More than three-quarters of cancer patients have bacteria in their cancer—including the stomach, cervical, and pancreatic varieties.
Microbiome analysis as a cancer diagnostic approach was detailed in a 2020 article in Nature (DOI: 10.1038/s41586-020-2095-1). Researchers, including Miller-Montgomery, re-examined whole-genome and whole-transcriptome sequencing studies in the Cancer Genome Atlas of 33 cancer types from 11,000 patients at 70 different centers.
They found microbial signatures in tissue and blood that were highly specific to most major types of cancer, she continues. Further, it is possible to distinguish cancer tissue from adjacent healthy tissue using these microbial markers.
Micronoma has also used the whole-genome shotgun method on 20 different cancer types, Miller-Montgomery says. For three common types—lung, prostate, and melanoma—investigators succeeded in correctly identifying disease using whole blood.
Detecting cancer at stage 1 or 2 is the objective because it could prevent millions of deaths worldwide, she says. The dilemma is that the typical cell mutation rate is 0.1% in early cancers, requiring 200 milliliters of blood (close to a full bag) to ensure their capture. The risk of false negatives is therefore high. The “double whammy” is that some mutations, including ones that happen naturally in white blood cells as we age (clonal hematopoiesis), can create false positives.
It appears possible to use a single cancer assay looking at an entire microbial community to screen for multiple early-stage cancers, says Miller-Montgomery. But to be used diagnostically by clinicians, who can likely pinpoint a problem organ on their own, an assay would need to distinguish when the issue is or isn’t cancer.
The Micronoma team therefore started looking at one cancer at a time, she says, starting with a search for the microbial signature of lung cancer. In the U.S., 75% of cases are detected at stage 3 or 4 when the five-year survival rate is a grim 15% to 20% versus 82% at stage 1.
Clinicians do not push low-dose CT scans—the standard of care for smokers over age 50—because of the small return on the expense and inconvenience of the screening test, Miller-Montgomery continues. When a nodule is found, 95% of the time it is benign. If a surgical biopsy is done, it’ll cost at least $6,500 but as much as $26,000 if there are complications (e.g., pneumothorax).
Recent studies suggest that the microbiome, including fungi and viruses as well as bacteria, “need to be taken into account in the oncology equation,” she says, citing a 2021 review in Science (DOI: 10.1126/science.abc4552). Microbes could be key actors not only in diagnostics, but also prognostics and the development of new therapies.
The “oncomicrobiome”—cancer-complicit microbes that include the Epstein-Barr virus, Hepatitis B virus, Hepatitis C virus, and H. pylori—cohabitate with good-guy microbes, so antibiotics aren’t the answer, Miller-Montgomery points out. Rather, precision therapeutics are needed to capitalize on intratumor bacteria that can trigger the immune system to recognize and destroy cancer cells (Nature, DOI: 10.1038/s41586-021-03368-8).
Pete Corish, CEO of UK-based startup Rivela Diagnostics, spoke on the unique value of exosome-based liquid biopsy in early cancer detection. Global access and rollout will require integration with the existing pathology infrastructure, he says.
Exosomes, a type of extracellular vesicle (EV), are “legitimate surrogates” for their parental cells, Corish says. Their cargo is full and complete, enabling multi-analyte analysis of DNA, RNA, proteins, lipids, and metabolites derived from living cells in real time. That’s why exosomes are “attractive and tractable,” he argues.
The advantages of tumor-derived exosomes include their abundance, signal-to-noise ratio, and detection capability, says Corish, adding that Rivela wants to explore the clinical possibilities with collaborators. Exosomes accurately reflect cancer biology, can be transported and stored without loss of quality, are amenable to biobanking, and can be enriched to detect biomarkers of interest. Enough tumor-derived EVs might be found in blood, he adds, provided they can be separated from smaller circulating lipid particles.
In mice studies, EV abundance has been correlated with tumor size, Corish says. In humans, modeling indicates EV shedding is proportional to tumor size. While existing detection platforms would miss early-stage cancers based on the level of non-enriched, tumor-derived EVs, scientists are “on the cusp” of meeting the analytical challenge.
Isolating exosomes from blood plasma is not inexpensive, says Corish, but the pre-analytical workflow is the real “elephant in the room.” Progress requires a new process for exosome purification (centrifugation is the current standard) and a standalone plasma separation device for the efficient recovery of intact exosomes that prevents platelet activation and hemolysis.
For molecular assays using a genetic cancer marker, tumor enrichment can reduce sample complexity, he says. A workflow for extracting EV DNA for the detection of clinically relevant mutations has been recently described (Cancers, DOI: 10.3390/cancers14133258).
Pan-cancer approaches will need to look at features common across cancer types, such as the “Wahlberg effect” triggering metabolic abnormalities as well as genomic instability and mutations, says Corish. The protein TM9SF4, for example, is a “workhorse marker” involved in the malignant progression of cancer cells and preliminary study findings suggest it be a good foundation for an exosome-based liquid biopsy.
Rivela Diagnostics is also looking to develop a microRNA assay for tissue-of-origin analysis in a collaboration with the University of Ghent, Corish says. Although prostate and lung cancers are the focus currently, a globally accessible pan-cancer screening tool is the longer-term objective. The intellectual property is being productized and commercialized by Exosomics, leaders in exosome biopsy research, he notes.