Tag: United Kingdom

Posted on

How a new extraction process could unlock the world’s lithium

Researchers say they’ve found a new way to extract lithium, a crucial metal used in the lithium-ion batteries that power electric vehicles and energy storage arrays. This new technique could be more environmentally friendly and cheaper than existing ones. 

The research was published today in Science, and a startup called Rock Zero is working to commercialize the process.

“At scale, we believe this will be the lowest-cost way of sourcing lithium in the world,” says Yet-Ming Chiang, one of the study authors, who is an MIT professor and a serial entrepreneur behind climate tech companies including Form Energy and Addis Energy.

The most economical way to get lithium currently is to extract it from brine, salty water that’s pulled the metal out of rock over the course of millennia. But this technique is geographically limited and currently requires vast tracts of land for massive evaporation pools. The more common tactic is hard-rock mining, where large bodies of ore are blasted apart, cooked at high temperatures, and processed using dangerous chemicals.

The researchers’ new method uses a weak acid to dissolve typically nonreactive silicate minerals. That frees not only the lithium but also other useful materials, including alumina and silica.

The origin story for this research, and the resulting company, came from another startup founded by Chiang, Sublime Systems, which makes cement using electrochemistry.

The team was trying to find a source of highly reactive silica in order to form stronger cement. One way to make reactive materials, which can bond easily with other materials, is to take a nonreactive material, dissolve it, and then allow it to become solid in a more reactive form. It’s not impossible to dissolve silicates, but the best-known way is to use hydrofluoric acid, an extremely dangerous chemical. Other fluorine-containing chemicals are candidates too, but some will produce hydrofluoric acid as a side product during reactions. 

Chiang drew inspiration from a previous home renovation project involving glass, which is made of silica. “I was remodeling a shower in Framingham, Massachusetts, about 25 years ago,” he says. “So when we started this project, I remembered that glass etching cream and thought, ‘What’s in that?’” 

The glass etching cream he remembered, which can be found on shelves at any craft or home improvement store, uses ammonium fluoride, a weak acid. And the MIT researchers discovered that in the right conditions, it can effectively dissolve silicate minerals without producing hydrofluoric acid in the process.

This chemistry could be useful for any silicate minerals—and there are a lot of them. But spodumene, the mineral that’s often mined for lithium, became a prime first target. (Chiang says a suggestion from Doug Wicks, one of the company’s advisors and a former ARPA-E official, pointed the team in spodumene’s direction.)

small pieces of rock next to a line of 3 capped vials of powder
From left to right: spodumene, silica, alumina and lithium salts.
ROCK ZERO

Today, a key step in processing spodumene ore is to roast it in a kiln at super-high temperatures. This causes a phase transformation, essentially puffing up the material and making the lithium more accessible.

By avoiding the need to reach these temperatures, you could save on energy costs and potentially reduce carbon emissions as well, says Camden Hunt, one of the authors of the study and the CEO and cofounder of Rock Zero.

Avoiding the kiln could also unlock the ability to use some ores that can’t be roasted properly, Hunt adds. Ore that contains too much iron won’t go through the phase change correctly, instead melting and turning into a glassy material.

The new process relies on simple stirred plastic tanks and takes place at temperatures up to about 95 °C (200 °F). The ammonium fluoride dissolves the silicates, which in earlier experiments allowed nearly all of the lithium inside the spodumene ore to be extracted within a couple of days. The researchers have since cut this time to under 12 hours, says Benjamin Mowbray, first author of the study and the CTO and cofounder of Rock Zero.  

The products (after some additional steps to clean them up) are lithium carbonate, which can be used to make batteries; alumina, which can go into a smelter to make aluminum; and cementitious silica, which can be added into concrete. And the acid can be reused in the same loop.

Chiang calls this “nose-to-tail” mining—using every part of the ore provided, like eating every part of a butchered animal.

The researchers are currently working to scale and optimize the process. The tanks in the lab in Cambridge, Massachusetts can handle three kilograms of spodumene concentrate in each batch. 

They have also estimated the cost of this process once fully scaled up. Assuming that the ammonium fluoride can be recycled at a high level, they should be able to extract lithium for less than $6,000 per metric ton. (They’ve identified a potential cheap industrial source of the acid as well, as an alternative to recycling it.) 

The total cost is projected to be lower than that of other processes used to extract lithium from hard-rock ore today, and it could be competitive with brine.

The team has designed a pilot plant and is looking for space to build it. The plan is to have construction done by the end of 2026 and start operating the facility in 2027. Talks are underway with potential partners in the mining industry.

One difficulty for new players in lithium extraction is the volatility of the market: Prices have seen huge swings in recent years, from a peak in 2022 to lows in late 2024 and a slow climb starting in early 2026. 

Rising prices might benefit new players like Rock Zero, but there are many projects that could come online if prices continue to rise, and that could bring the market right back down, says Simon Jowitt, chair of exploration geology at the University of Nevada, Reno. “People are waiting to see what happens with the lithium price,” he says. “It’s a crowded market, and there’s some big players out there.”

And even though batteries are driving up demand for lithium, the market is still relatively small, Jowitt adds: “That means it’s going to be volatile.” New lithium extraction technologies like Rock Zero’s will have to compete with methods used by existing giants, and there’s also the potential that technological alternatives, like sodium-ion batteries that don’t need lithium, could make the market more difficult to navigate, Jowitt says. He also thinks some of the company’s economic estimates could be optimistic.

For its part, Rock Zero’s team hopes not only to scale this technology for lithium, but to use it for other minerals in the future. As Mowbray says, “The Earth’s crust is made of silicates.”

Posted on

SHIMMER: Routine Clinical Data Processed Into Scalable Disease State Markers

Yes, you have a disease. No, you don’t have a disease.

There is too much complexity to human health for this simplistic, binary method of diagnosis. The phenotypic state of a disease is an ongoing process, while the genotype remains static. Even when it comes to infection, the presence or absence of a pathogen captures little of the entire dynamic biological world inside each of us.

A person’s future development can be predicted with near-certainty by an inherited mutation or chromosomal arrangement, such as cystic fibrosis, Huntington’s disease, and trisomy 21 (Down’s syndrome). Disease development is not an on/off switch. Rather, diseases progress along quantifiable biological spectrums.

The question is, how can that be done in an everyday clinical setting with routine data?

Icahn School of Medicine at Mount Sinai researchers developed a machine learning (ML)-based system that uses routine clinical data to estimate a person’s risk for multiple diseases, potentially revealing hidden illness years before diagnosis.

Published in the Cell Press journal Med, the study demonstrates that SHIMMER stands out due to its use of standard clinical measurements instead of highly specialized data. The practical use of many AI systems in medicine is limited because they require extensive imaging, genomic sequencing, or disease-specific testing. The fact that SHIMMER uses data that is already collected in regular healthcare settings instead could make it easier to implement on a larger scale.

Decoding disease states

Led by Iain S. Forrest, MD, PhD, the research team trained ML models on seven diseases—atrial fibrillation, breast cancer, coronary artery disease (CAD), migraine, rheumatoid arthritis (RA), schizophrenia, and type 2 diabetes (T2D)—using records from the BioMe Biobank in New York and the UK Biobank.

The researchers found that the disease-spectrum scores aligned closely with known risk factors and biological markers. For atrial fibrillation, rising SHIMMER scores tracked with increasing age, obesity, hypertension, and stroke risk. In T2D, the scores rose alongside glucose levels, hemoglobin A1c, triglycerides, and inflammatory markers. CAD scores correlated with smoking, high cholesterol, chronic kidney disease, and established cardiovascular risk calculations.

Importantly, these relationships were not abrupt. Instead, biological changes increased gradually across the spectrum, supporting the idea that disease develops gradually rather than suddenly at diagnosis.

Forrest and colleagues also showed that SHIMMER is capable of identifying disease severity and predicting outcomes. Higher SHIMMER scores were associated with earlier disease onset, more complications, and reduced survival. In CAD, increasing scores tracked with worsening artery blockage, heart failure, arrhythmias, and heart attacks. For RA, higher scores aligned with greater inflammation, worsening anemia, and increased use of immunosuppressive medications.

Some of the strongest findings involved diseases that traditionally lack reliable biomarkers. Schizophrenia, for example, is diagnosed largely through behavioral assessment rather than laboratory testing. However, SHIMMER found obesity, smoking, cannabis use, cardiovascular complications, and poorer survival linked to schizophrenia risk and severity. Spectrum associations were also found in migraine, another condition without biomarkers, including symptom severity and medication use.

Perhaps the most clinically intriguing result was SHIMMER’s ability to flag people who appeared biologically ill despite lacking a formal diagnosis. The researchers identified several people whose routine clinical data strongly resembled known disease profiles despite no diagnosis. One patient with a high T2D SHIMMER score had obesity, hypertension, high glucose, and classic diabetic symptoms but was never diagnosed. Similar patterns appeared for coronary artery disease and rheumatoid arthritis.

In larger undiagnosed populations, elevated SHIMMER scores still predicted abnormal biomarkers and mortality risk. That suggests the system may detect diseases that traditional healthcare pathways miss.

Not a shiny object

The findings challenge the long-standing medical convention of classifying disease in binary terms: either a patient has a condition or does not. The implications could be significant for preventive medicine and population health management. Continuous disease-spectrum scores may help healthcare systems identify high-risk patients earlier and prevent complications.

The researchers envision SHIMMER as an EHR background decision-support tool. Even if they do not meet diagnostic criteria, patients with high scores for diabetes or coronary artery disease may be prioritized for testing, monitoring, or preventive treatment. In that sense, it is complementary to genetic screening.

Still, the work remains a proof of concept rather than a ready-to-deploy clinical tool. The study was retrospective, meaning it analyzed existing records rather than prospectively following patients in real time. The researchers also note concerns about healthcare bias, uneven data quality, and the ethical implications of artificial intelligence in medicine.

Another important limitation is that SHIMMER does not provide definitive diagnoses. Its scores reflect disease burden and biological resemblance rather than calibrated predictions of future illness. Determining how clinicians should act on specific score thresholds will require further study.

The work represents a broader shift in how medicine may conceptualize disease in the AI era. Instead of fixed categories, conditions may increasingly be viewed as dynamic biological trajectories measurable through continuously updated clinical data. If confirmed in future studies, SHIMMER could turn routine medical records into early-warning systems that detect disease before symptoms appear.

 

The post SHIMMER: Routine Clinical Data Processed Into Scalable Disease State Markers appeared first on Inside Precision Medicine.

Posted on

Non-invasive electrical stimulation for sleep disturbances in adults: protocol for an evidence−mapping umbrella review of systematic reviews and meta−analyses with subgroup analysis by intervention type and population

BackgroundSleep disturbances affect 10%–30% of adults worldwide. Non−invasive electrical stimulation (e.g., transcranial electrical stimulation) has emerged as a promising non−pharmacological intervention. Although numerous systematic reviews and meta−analyses have been published, they vary considerably in methodological quality, populations, intervention types, and conclusions. No umbrella review has yet synthesised the evidence across different modalities and populations. This evidence mapping umbrella review aims to systematically chart the existing systematic reviews, assess methodological quality, quantify overlap, and describe evidence patterns across diverse modalities and populations.MethodsFollowing JBI guidelines, we will search PubMed, Embase, Cochrane Database of Systematic Reviews, Web of Science, PsycINFO, and Scopus (inception to April 2026), restricted to English. Grey literature will be searched via PROSPERO, ClinicalTrials.gov, Google Scholar (first 200 records), and reference list screening (snowballing). We will include systematic reviews and meta−analyses of randomised controlled trials evaluating any non−invasive electrical stimulation for sleep outcomes. Two reviewers will independently screen, extract data, and assess methodological quality using AMSTAR 2. Primary study overlap will be quantified by the Corrected Covered Area. Where feasible, we will calculate 95% prediction intervals, perform Egger’s regression tests, and conduct excess significance tests using review-level summary estimates. Subgroup analyses will be stratified by intervention and population type. Sensitivity analyses will exclude: (1) reviews with critically low AMSTAR 2 ratings, (2) preprints, (3) reviews at high risk of reporting bias, and (4) studies where sleep is not the primary outcome. The primary outcome is subjective sleep quality; total sleep time is a key secondary outcome. Evidence will be graded using the Fusar−Poli classification, with GRADE for key outcomes.DiscussionThis umbrella review will provide the highest level of evidence synthesis, identifying modalities with more consistent or higher certainty evidence and highlighting areas where evidence remains uncertain. Limitations include restriction to English (which may disproportionately impact modalities such as TEAS), expected heterogeneity, and possible insufficient data for some subgroup analyses. All amendments have been documented in PROSPERO (CRD420261357590).Clinical Trial Registrationhttps://www.crd.york.ac.uk/prospero/, identifier CRD420261357590.
Posted on

Largest Rare Variant–Lipid Association Study Finds Heart Disease Targets

A landmark week for cardiovascular genetics opened with huge news from the gene-editing field: Eli Lilly’s experimental therapy VERVE-102, a single-dose PCSK9 base editor designed for patients with heterozygous familial hypercholesterolemia (FH) and premature coronary artery disease (CAD), demonstrated durable cholesterol-lowering effects in early studies. While the result marks one of the strongest signs yet that in vivo gene editing could become a practical treatment for inherited cardiovascular disease, it’s just scratching the surface of the enormous number of uncharacterized genetic variants that influence cholesterol levels and heart disease risk across global populations.

The largest rare variant association with blood lipids study ever reported may provide the roadmap to heart disease mechanistic and clinical insights. In a Nature Genetics study analyzing data from more than one million individuals, researchers identified thousands of rare coding variants linked to cholesterol and triglyceride (TG) levels, including several genes strongly associated with coronary artery disease. The findings could accelerate the development of precision medicines for dyslipidemia and improve the diagnosis of inherited lipid disorders such as familial hypercholesterolemia.

Diversity and scale

FH is caused by rare mutations that raise LDL-C, increasing the risk of early-onset CAD, a leading cause of premature death worldwide. Although statins and other lipid-lowering therapies can significantly lower cardiovascular risk, FH is underdiagnosed and undertreated. Many FH-associated variants have variable penetrance, so some carriers have severe disease and others have milder symptoms, adding to the complexity.

Accurate variant-specific risk assessment is becoming increasingly important as genetic screening expands. Most genetic databases are biased toward Europeans, making it difficult to classify disease-causing variants in non-Europeans.

To fill this gap, Satoshi Koyama, MD, PhD, led a research team across academic and medical institutions in the Boston area that analyzed exome sequencing and blood lipid data (total cholesterol, LDL-C, HDL-C, and TG) from Million Veteran Program, UK Biobank, and All of Us participants. More than 230,000 participants came from historically underrepresented populations, making this one of the most diverse large-scale lipid genetics studies conducted to date.

Mechanistic and clinical implications

Their analysis uncovered nearly three million rare coding variants, including more than 214,000 predicted loss-of-function mutations, 2.7 million missense variants, and over 23,000 cryptic splice variants that may disrupt gene processing. In total, the team evaluated over 10 million variant-phenotype associations. The results revealed 800 exome-wide significant additive associations across 184 genetic loci, along with 109 recessive associations involving 53 genes. Many of the strongest signals appeared in genes already known to regulate lipid metabolism and cardiovascular disease, including PCSK9, LDLR, APOB, NPC1L1, and APOC3.

The study also identified five lipid-associated genes significantly linked to CAD risk, highlighting potential therapeutic targets. One particularly intriguing gene was RORC, which encodes the transcription factor RORγ. Previous laboratory and animal studies suggested that suppressing RORγ improves metabolic health and reduces atherosclerosis. Consistent with those findings, the study showed that loss-of-function variants in RORC appeared protective against CAD in humans.

Another key finding involved cryptic splice variants, a class of mutations often overlooked in clinical genetics. The researchers used machine-learning-based splice prediction tools to show that these variants had biological effects similar to canonical loss-of-function mutations, suggesting that many clinically important variants may be underestimated.

The study also found that 13% of missense mutations produced hypermorphic alleles that increased gene activity, unlike most loss-of-function variants. Existing computational prediction tools frequently fail to identify these variants, potentially limiting the sensitivity of current genetic testing approaches.

Koyama and colleagues also detected strong recessive genetic effects that standard additive models may miss entirely. Because homozygous rare variants are uncommon, their contribution to disease has historically been difficult to measure. The findings suggest that recessive inheritance may explain part of the “missing heritability” in complex cardiovascular disease.

The researchers found that most rare variants exerted similar effects across populations, even when variant frequencies differed substantially between ancestries. The study identified 130 alleles observed primarily or exclusively in non-European populations, emphasizing the importance of expanding genetic research beyond European cohorts to improve equitable diagnosis and drug discovery.

From screening to saving hearts

Beyond biological discovery, the study carries important clinical implications. By comparing their findings with curated pathogenic variant databases, the investigators confirmed many established classifications while identifying variants that may warrant reclassification. Two newly highlighted variants enriched in non-European populations may represent previously underrecognized causes of familial hypercholesterolemia.

Although the research focused mainly on rare coding variants rather than noncoding DNA, the authors argue that population-scale sequencing studies can now provide clinically actionable insights into disease mechanisms, pathogenicity, penetrance, and prognosis. Together, the findings offer a powerful new resource for cardiovascular genetics at a time when therapies such as PCSK9 gene editing are beginning to move from concept to clinic.

The post Largest Rare Variant–Lipid Association Study Finds Heart Disease Targets appeared first on Inside Precision Medicine.

Posted on

Guardant Nabs Key ACS Nod and Liquid Biopsy Approval 

Colorectal (CRC) detection has become a red hot field, as concerns about rising numbers of young adult cases and competition among multi-cancer panels heat up. Guardant Health took two giant steps forward recently with first the approval of its liquid biopsy test, Guardant360 Liquid CDx, and then the listing of its blood-based Shield test by the American Cancer Society (ACS) as a choice for CRC screening of adults age 45 and older who are at average risk of the disease.

In an updated guideline released today, the ACS added blood-based screening tests, and specifically Shield, to its list of recommended choices for the patients in this subset who have not completed or have declined visual exams and stool tests. The group specifically names Shield, which was approved by the U.S. Food and Drug Administration in 2024.

Shield is an in vitro diagnostic test that detects CRC-derived alterations in cell-free DNA from blood collected in the Guardant Blood Collection Kit. The test is performed at Guardant.

These two advances put Guardant in an excellent position in the cancer liquid biopsy market, which is currently valued at between $7B and $12B and expected to double over the next ten years. 

A spokesperson for Guardant told Inside Precision Medicine, “Our current focus is on ensuring the approval and successful launch of the Shield test, with an initial focus on eligible adults age 65 and older across the U.S. who are enrolled in Medicare. In parallel, we are continuing to optimize and improve the performance of the Shield test, with the goal of upgrading the test post-approval.”

This week also, Guardant announced that the Molecular and Clinical Genetics Panel of the U.S. The Food and Drug Administration (FDA)’s Medical Devices Advisory Committee strongly recommended FDA approval of the Shield blood test for these types of patients.

“The advisory committee’s strong support for the approval of Shield reinforces the crucial role that a blood test option can have in improving CRC screening rates for those at average risk,” said AmirAli Talasaz, Guardant Health co-CEO. “Despite the importance of detecting colorectal cancer early, there are notable barriers that can deter average-risk Americans from completing existing screening methods. Shield effectively detects cancer at an early stage when it is most treatable. Providing people with this blood test alongside other non-invasive stool tests can increase the rate of colorectal screening and potentially reduce preventable CRC deaths.”

Colorectal cancer is the second-leading cause of cancer-related deaths in the U.S. The disease has a 91% five-year survival rate when caught at stage I (localized), but one out of three eligible Americans—50 million people—are not being screened for CRC. 

Current primary non-invasive screening options include stool-based tests which have proven efficacy in detecting CRC; however, studies have consistently found that barriers such as handling stool and challenges performing the test impact adherence. 

“Sadly, 76% of deaths caused by colorectal cancer occur in individuals who are not up to date with their screening,” said Daniel Chung, MD, gastroenterologist at Massachusetts General Hospital and professor of medicine at Harvard Medical School. “Clinical evidence and CRC screening guidelines acknowledge the value of offering choice to individuals at average risk for CRC and highlight the role of patient preference in test selection and CRC screening completion.”

The FDA panel’s recommendation is based on Guardant’s premarket approval (PMA) application for Shield, including the results of the pivotal ECLIPSE study evaluating the performance of the test for detecting CRC in average-risk adults. Results from the study appeared in the March 2024 issue of The New England Journal of Medicine. (Chung was an author on this study.) Shield demonstrated 83% sensitivity for the detection of CRC, with 90% specificity for advanced neoplasia. Guardant notes that this performance is within range of existing stool-based tests used as primary CRC screening options, in which overall sensitivity ranges from 67% to 92%.

The post Guardant Nabs Key ACS Nod and Liquid Biopsy Approval  appeared first on Inside Precision Medicine.

Posted on

CGT Sector Opts for New Tech for Efficiency and Consistency

A maturing, data-rich and more commercially-focused cell and gene therapy (CGT) industry is finally embracing innovative manufacturing technologies, according to the head of U.K.-based industrial software firm, Autolomous.

Early CGT manufacturing relied heavily on technologies borrowed from biologics manufacturing, systems originally built for producing therapeutic proteins rather than patient-specific therapies. As a result, production processes were often highly manual, fragmented, and operator-dependent. Indeed, in many cases, manufacturing resembled a specialized laboratory process, rather than a scalable industrial operation.

But the situation is changing, says Autolomous CEO, Alexander Seyf, who argues CGT manufacturers are embracing innovation with product quality, cost control, and scalability in mind.

“As the industry moves beyond early clinical programs and toward commercialization, manufacturers are recognizing that traditional approaches simply won’t support the scale, consistency, or economics needed long term.

“There’s now much greater focus on automation, closed processing, and digital integration to improve reproducibility, reduce contamination risk, and lower the cost of goods,” Seyf tells GEN.

Increasingly, CGT firms are opting for purpose-built technologies, he says, citing automated cell handling, closed-system processing, robotic fill-finish, and integrated single-use platforms as examples.

“The broader shift is toward what many describe as ’CGT 4.0:’ applying the principles of Industry 4.0 to cell therapy manufacturing. The aim is to move away from highly manual, artisanal production toward scalable, repeatable manufacturing that can support broader patient access,” he adds.

Data

Data is also driving the adoption of new technologies. In cell and gene therapy production, large volumes of data are collected by different instruments, software platforms, QC systems, environmental monitoring tools, and manual inputs—many of which were never designed to communicate with one another.

In such circumstances, digital technologies provide manufacturers with an infrastructure that can gather process information and ensure it is usable, traceable, and reliable, Seyf says.

“Maintaining data integrity and full chain-of-custody visibility across fragmented systems can become incredibly difficult, particularly as operations scale.

“To solve this, companies are investing heavily in integrated digital architectures that bring manufacturing and quality data together into unified environments. Standardized data models, interoperable software platforms, and automated data capture are becoming increasingly important. Cloud infrastructure also plays a key role because it allows manufacturers to aggregate and analyze data across multiple facilities in real time,” Seyf says.

A typical, modern digital cell and gene therapy manufacturing setup includes systems that manage manufacturing execution, laboratory information, quality, electronic batch records, and cloud-based data platforms.

“Together, these systems help manage scheduling, batch tracking, compliance, traceability, and quality oversight in real time,” Seyf says, adding, “Modern facilities are also increasingly using process analytical technologies and integrated sensors to monitor critical process parameters continuously, rather than relying only on end-point testing.”

“A digital CGT manufacturing system is really about connectivity and visibility across the entire process. It’s not just about replacing paper records with electronic systems—it’s about creating an environment where manufacturing equipment, quality systems, analytics, and logistics are all connected and continuously sharing data.”

Material traceability

The need to keep track of cell and gene therapy raw materials is also fueling the adoption of innovative technologies, with patient-specific therapies being a case in point.

Seyf tells GEN, “Maintaining chain-of-identity and chain-of-custody is particularly important in CGT manufacturing, especially for autologous therapies where every batch is tied to an individual patient. That requires seamless integration between manufacturing systems, analytics platforms, and logistics operations.”

The post CGT Sector Opts for New Tech for Efficiency and Consistency appeared first on GEN – Genetic Engineering and Biotechnology News.

Posted on

AI Model Predicts Cancer Treatment Response from Tumor Genotype

Researchers at University of California, San Diego have developed a new artificial intelligence (AI) model that can translate a tumor’s complex genetic profile into predictions about how that cancer may respond to treatment. The foundation model, called MutationProjector, was trained on genomic data from more than 30,000 tumors across 10 solid cancer types, and validated through testing across multiple independent patient cohorts. Led by Trey Ideker, PhD, professor of medicine at UC San Diego School of Medicine and director of the Big Data Institute at the University of Oxford, the researchers say the model offers a new framework for connecting cancer mutations to the biological pathways that drive treatment response.

“Genetic sequencing is already routine in cancer care, but we still struggle to fully interpret the many mutations found in a patient’s tumor,” said Ideker, who also holds a second appointment at UC San Diego Jacobs School of Engineering and is a member of UC San Diego Moores Cancer Center. “Our goal with MutationProjector was to build a general-purpose model that can learn from tens of thousands of tumor genomes and turn those mutation patterns into more precise predictions about treatment response.”

Ideker is co-senior and co-corresponding author of the team’s published paper in Cancer Discovery, titled “A foundation model of cancer genotype enables precise predictions of therapeutic response,” in which the authors stated, “These results establish a unifying framework for connecting tumor genotypes to biological mechanisms and therapeutic outcomes.”

Following a cancer diagnosis, one of the next steps is often genetic testing, which helps doctors classify the tumor and decide which treatments to pursue. “DNA sequencing panels—and in particular those that broadly identify alterations in cancer-associated genes—have been widely adopted in the clinic due to their relatively low cost, rapid turnaround, and established relevance to treatment outcomes,” the authors explained.

Trey Ideker is a professor of medicine at UC San Diego School of Medicine and director of the Big Data Institute at the University of Oxford. [Erik Jepsen / UC San Diego]
Trey Ideker is a professor of medicine at UC San Diego School of Medicine and director of the Big Data Institute at the University of Oxford. [Erik Jepsen / UC San Diego]

However, while genetic testing is relatively low cost, fast, and has a strong track record in cases where validated genetic biomarkers are available, those cases remain limited, because this type of treatment stratification is currently based on only a small number of known biomarkers. Today, only about 8% of cases are successfully matched to an FDA-approved therapy on the basis of genetics and usually on the basis of a single gene, the team continued. “While this situation may reflect the incomplete scope of genes covered by current sequencing panels, it clearly also reflects a fundamental lack of knowledge about how gene mutations should be interpreted.”

They suggest that an “average” tumor has approximately 11 distinct genetic alterations identified by clinical sequencing, representing a potentially rich source of molecular information, if this information could be used to help select treatment. One of the challenges to matching cancer mutations with treatment outcomes is that most mutations are rare, the investigators pointed out. Another is that individual biomarkers do not function in isolation, but act together to influence drug response.

Unlike existing approaches that rely on a small number of biomarkers, MutationProjector analyzes the broader combination of genetic alterations present in a tumor. The model then uses this information to generate a compact representation of the tumor’s biological state, helping researchers interpret which molecular pathways may be disrupted and, by extension, which treatments may be most effective. “Foundation models, which are pre-trained on large datasets and then applied to solve diverse new challenges with relatively few samples, are especially well positioned to advance precision oncology,” Ideker and team noted.

The investigators trained their foundational model, MutationProjector, using genetic profile data from more than 30,000 tumors samples across different cancer types. They then showed that across several independent cohorts of cancer patients, including those with bladder cancer, lung cancer and melanoma, MutationProjector matched or exceeded existing methods for predicting response to common immunotherapy and chemotherapy treatments. The model also identified both known and unexpected biomarkers associated with treatment outcomes, which could help improve current approaches to genetic testing and patient stratification.

“When applied to predict immunotherapy or chemotherapy resistance across multiple cancer types and cohorts, MutationProjector achieves or exceeds state-of-the-art performance in all contexts,” the team wrote. “It identifies unexpected biomarkers, including KMT2D mutation in immunotherapy sensitivity and joint alteration of SMARCA4 and STK11 in immunotherapy resistance.”

JungHo Kong, PhD, first author of the study and a postdoctoral researcher in the department of medicine at UC San Diego School of Medicine, said, “Many cancer mutations are individually rare, which makes them difficult to study one at a time. By pretraining on a large collection of tumors and integrating molecular network knowledge, MutationProjector can detect patterns that would be easy to miss with conventional biomarker approaches. That gives us a way to move from long lists of mutations toward a more functional understanding of the tumor.”

First study author JungHo Kong, shown here, is a postdoctoral researcher at UC San Diego School of Medicine. [UC San Diego Health Sciences]
First study author JungHo Kong, shown here, is a postdoctoral researcher at UC San Diego School of Medicine. [UC San Diego Health Sciences]

The researchers emphasize that the model was designed not only to make predictions, but also to provide insight into why those predictions are made, which could help when refining biomarkers and treatment strategies. This interpretability is especially important in precision oncology, where clinicians need to understand how tumor genotypes relate to treatment decisions.

The team also hopes to expand the model to additional cancer types and data sources, including international cancer genome datasets and other forms of clinical information, such as imaging, transcriptomics, and electronic health records. “While 30,000+ genomes representing 10 solid tumor types were considered in our study, numerous additional tumor samples are available for expansion of MutationProjector to tumor types such as pancreatic cancer, prostate cancer or sarcomas,” the authors said. “Other future studies should explore the extent to which the MutationProject concept can be applied to further clinical tasks of interest, including application to liquid biopsies of circulating tumor DNAs for early cancer detection.”

Ideker added, “Our results suggest that tumor genome foundation models may help extend the clinical value of sequencing beyond a handful of well-known genes. This could support a more comprehensive and biologically grounded approach to precision oncology.”

The post AI Model Predicts Cancer Treatment Response from Tumor Genotype appeared first on GEN – Genetic Engineering and Biotechnology News.