Every breath we exhale carries a complex chemical narrative written in thousands of volatile organic compounds (VOCs). These tiny molecules, produced by human metabolic processes and invading pathogens, hold immense diagnostic potential. Yet, translating this molecular cloud into a definitive medical diagnosis has long faced an analytical bottleneck. Now, an international team of physicists from Botswana and South Africa has harnessed machine learning to decode these complex signatures, offering a rapid, non-invasive pathway to detect tuberculosis (TB) directly from human breath.

Breaking the analytical bottleneck

Traditional gas chromatography-mass spectrometry (GC-MS) excels at separating complex molecular mixtures based on volatility and mass-to-charge ratios. However, a single breath sample generates thousands of overlapping spectral peaks. Manually sifting through this dense landscape is a painstaking, time-consuming process.

To find a faster, more systematic approach to molecule identification, researchers in the Nano –Breath taking group led by George Chimowa from Botswana International University of Science and Technology (BIUST) collected breath samples from three distinct cohorts: patients with active TB, individuals with multidrug-resistant TB (MDR-TB), and healthy control volunteers.

Standard analytical pipelines often rely on aggressive data pre-processing to smooth and simplify the GC-MS spectra. However, this filtering risks erasing subtle, low-concentration molecular features that might hold critical diagnostic clues. Bypassing this limitation, the BIUST group fed a high-dimensional dataset of 1867 spectral features (from 87 breath samples) directly into four competing supervised machine-learning algorithms: decision trees, random forest, k-nearest neighbours, and support vector machines (SVM). The researchers published their findings in Discover Artificial Intelligence.

Geometric optimization avoids the pitfalls of high-dimensional data

The team found that the linear SVM classifier led the algorithmic comparison, distinguishing among the three patient cohorts with 93% accuracy.

This performance stems from how the algorithm handles high-dimensional data spaces. While distance metric models often suffer from overfitting errors, where experimental noise is misidentified as a meaningful diagnostic signal, SVM relies on geometric boundary optimization to separate classes.

Much like clearing a wide highway down the centre of two opposing crowds, it identifies a hyperplane that maximizes the margin between groups. Instead of tracking every individual in the crowd, the algorithm relies solely on a sparse subset of border points known as support vectors. By focusing strictly on these boundary “guards” and ignoring the background movement, this approach completely isolates the classifier from the noise inherent to the surrounding high-dimensional feature space.

Beyond raw classification accuracy, the model yielded a crucial insight for future clinical applications. It pinpointed a specific window within the gas chromatogram, a retention time between 10 and 30 min, where the most diagnostic, high-variance VOCs elute. This 20-min window corresponds to medium- and large-molecule VOCs, including specific chemical biomarkers associated with TB, such as tridecane, decane and O-cymene.

By demonstrating that these vital diagnostic features are concentrated within this single slice of data, the physicists have shown that future diagnostic hardware can be specialized, eliminating the need to analyse the entire spectrum. Narrowing the focus to this specific elution window can drastically shrink data-processing times and maximize instrument throughput in clinical settings.

AI and the future of medical diagnostics

TB remains one of the world’s deadliest infectious diseases. Traditional diagnostics rely heavily on slow sputum cultures and backlogged laboratory queues, delaying critical treatment. A rapid, non-invasive breath test powered by an optimized physical classifier offers an ideal alternative for resource-limited settings where laboratory infrastructure may be scarce.

While the researchers emphasize that larger validation trials are required, the work signals a profound shift in medical diagnostics. It mirrors advanced data-modelling techniques used in engineering, such as those predicting structural degradation in self-healing aerospace composites.

By fusing analytical physics with artificial intelligence, this research points to a future where immediate, automated bedside disease detection could be rapid, accessible and driven entirely by the mathematical profiling of a patient’s breath.

The post AI sniffs out tuberculosis in breath’s chemical signature appeared first on Physics World.

Join us for an in‑depth conversation with Simon Cherry, former editor‑in‑chief of Physics in Medicine & Biology (PMB), as he shares his expert perspective on the evolving landscape of medical physics research, molecular imaging, and non-invasive imaging technologies. Drawing on decades of leadership in biomedical imaging and academic publishing, Simon will reflect on major advances shaping the field, define the characteristics of high‑impact imaging research, and highlight how PMB continues to support rigorous, influential scientific work through its editorial standards and peer‑review practices.

This session will additionally spotlight the research of Xinchen (Alison) Deng, a finalist of the PMB 2024 Early Career Award. Attendees will gain insights into her contributions to the advanced medical imaging community, such as photon-counting CT for clinical applications and how her work supports broader advances in medical physics and X-ray/CT imaging research.

Simon will also discuss what makes early‑career research stand out and share strategies for researchers aiming to increase the visibility, reach, and long‑term impact of their publications within competitive scientific journals.

Moderator

Carol Clark, Publisher, Physics in Medicine & Biology

Speakers

Simon R Cherry, PhD received his BSc (Hons) in physics with astronomy from University College London in 1986 and a PhD in medical physics from the Institute of Cancer Research, University of London in 1989. Simon is currently Distinguished Research Professor in the Departments of Biomedical Engineering and Radiology at UC Davis and Professor of Total-Body Imaging at King’s College London. Simon’s research interests centre around biomedical imaging. His major contributions have been in developing systems for positron emission tomography (PET), including designing the microPET scanner for small-animal imaging, the earliest examples of simultaneous PET and MRI studies and co-leading the EXPLORER consortium that developed the world’s first total-body PET scanner. Simon is a Fellow of the Royal Society and Member of the U.S. National Academy of Engineering and National Academy of Inventors. He has received several international awards including the Society of Nuclear Medicine and Molecular Imaging Cassen Prize in 2022. He served as editor-in-chief of the journal Physics in Medicine & Biology from 2011–2020. Simon is the author of more than 350 peer-reviewed journal articles, review articles and book chapters in the field of biomedical imaging. He is also lead author of the widely-used textbook Physics in Nuclear Medicine.

Xinchen (Alison) Deng, PhD, is a diagnostic imaging physics resident at The Ohio State University. Her research focuses on advancing photon-counting CT for clinical applications, including multi-contrast imaging, early disease detection, detector optimization, and AI-based improvements in image quality. Before beginning the residency, Xinchen (Alison) was a postdoctoral fellow in medical physics at the University of Victoria, where she worked with Magdalena Bazalova-Carter and collaborated with Redlen Technologies on the optimization of photon-counting detectors and the clinical applications of spectral CT. She received her PhD in medical physics from the University of British Columbia, where she applied Raman spectroscopy, machine learning, and explainable AI to study radiation treatment response in cancer cells and tissues. Her broader research background spans photon-counting CT, spectral X-ray imaging, Raman spectroscopy, radiotherapy response assessment, and AI-driven image and data analysis. Her photon-counting CT research has been featured in the Physics in Medicine & Biology Early Career Researcher Focus Collection. She is a finalist for the PMB Early Career Researcher Award.

About this journal

Physics in Medicine & Biology. The international journal of biomedical physics and engineering, published by IOP Publishing on behalf of the Institute of Physics and Engineering in Medicine (IPEM)

Editor-in-chief: Katia Parodi, Ludwig-Maximilians University, Munich, Germany

The post Shaping the field of medical imaging: a fireside chat with Simon R Cherry and Xinchen Deng appeared first on Physics World.

A new study has revealed that sensitive health information about individual patients can be extracted from medical artificial intelligence (AI) models more effectively than previous research had indicated.

Diagnostic AI models are used to determine whether some given health data belong to a specific class – for example if an X-ray image shows evidence of pneumonia – and offer huge promise for improving patient outcomes. Earlier studies showed that membership inference attacks (MIAs), which determine whether a particular person’s data was used to train an AI model, yielded results no better than random chance. But these previous studies only evaluated the average risk across all patients in a training dataset.

By contrast, in this latest work doctoral student Moritz Knolle and his supervisor Daniel Rückert at the Technical University of Munich (TUM), together with collaborators at Imperial College London and Georg Kaissis from the Hasso Plattner Institute (HPI), examined the privacy risk for an individual patient. “Privacy that only works on average has little meaning. So I asked myself: would I contribute my data to train a medical AI model?” says Knolle.

As detailed in their recent Nature paper, the researchers attacked AI models based on seven different types of real-life medical datasets, including one comprising 21,799 electrocardiograms and another of 377,110 chest radiographs.

“If a model sees a specific data point during training, it usually becomes more confident at predicting it. So in these MIA attacks I compare the predictive probability number for this data point outputted by the medical AI model with the predictive probability number obtained from my model which has not seen that data point,” explains Knolle, adding that the attack is essentially a statistical test to distinguish between the two models.

Their results revealed that the privacy risk for an individual was far greater than the average risk. The team also found that privacy risk is not equally distributed among subgroups.

“We found that patients from minority groups that are under-represented in the training dataset tend to be over-represented among the 99 percentile of extreme [privacy] risk,” Knolle tells Physics World, warning that as medical AI models become larger and more sophisticated the privacy risk is amplified because their deep analysis creates the small changes in predictive probability that MIAs are good at finding.

While an MIA that reveals someone is a UK citizen poses little security risk, discovering that an individual has received a specific type of cancer treatment reveals personal information that could, for instance, adversely affect their medical insurance premiums. “The more specific the training cohort is, the larger the privacy risk,” states Knolle, crediting his medical doctor parents with inspiring his desire to improve healthcare AI.

Countermeasures include differential privacy (DP), in which a small amount of Gaussian random white noise is added in alongside the data during training, thereby reducing the effectiveness of MIAs while maintaining the model’s diagnostic accuracy. “This gives a mathematically provable upper bound for how much privacy risk any individual in this dataset could face, by limiting the information content that flows to the AI model,” explains Knolle.

He notes that, unfortunately, such DP tools have not been widely adopted by hospitals or clinics. This, he fears, could undermine patients’ confidence in sharing their data, and thereby not only reduce AI’s ability to help combat rare diseases, but also compound AI discrimination that occurs when there is too little data from minority groups included in training datasets.

“So we are hoping to inspire the medical AI community to do risk assessments before deploying AI models, and use verifiable strategies such as DP or limiting access to the AI model, so that as an individual you have a statistical guarantee that no harm would come to you by being part of a dataset,” concludes Knolle, who now plans to investigate the privacy risks posed by generative AI models for healthcare.

The post Privacy risks from medical AI are greater than previously thought appeared first on Physics World.

A collaboration of researchers in North America has reported their findings from the first diagnostic X-ray imaging performed in space. X-ray images recorded during an orbital spaceflight exhibited similar overall quality to those taken preflight on Earth, demonstrating the feasibility of adding radiography as a diagnostic tool for crew health.

With humans venturing further into space and spending more time in this harsh environment, the risk of adverse medical events increases. Accurate and timely diagnosis and treatment is a must-have. Currently, the only reliable medical imaging modality available to astronauts in spaceflight is ultrasonography – which is operator dependent, requires substantial training and relies on limited acoustic windows.

X-ray imaging, on the other hand, benefits from low operator dependence, rapid acquisition and typically higher spatial resolution than ultrasound. Digital radiography could also offer superior diagnostic capabilities for many medical conditions of concern during spaceflight, including dental disease, musculoskeletal trauma, inhalational injury, collapsed lung and arthritis.

Traditional X-ray scanning systems are bulky, produce a lot of radiation and are sensitive to motion – making them less than ideal for sending into space. But a new generation of portable X-ray scanners shows potential to survive the spaceflight and be operated by crew members with minimal training.

“It’s been a dream for aerospace medicine to have more than one imaging modality for diagnosing illnesses and injuries in space,” says lead researcher Sheyna Gifford, an assistant professor of aerospace medicine at the Mayo Clinic in Rochester, MN, in a press statement.

Scanning in space

In a study led by co-author David Lerner, the researchers first demonstrated the feasibility of X-ray imaging in simulated microgravity in 2022. During a parabolic flight, they used a commercial off-the-shelf digital radiography system to successfully capture radiographs of a human subject and a phantom during lunar gravity (1/6g), Martian gravity (1/3g) and microgravity.

In this latest project, reported in in Radiology, Gifford and her team investigated the performance of a similar portable digital radiography system aboard the Dragon spacecraft, during the Fram2 spaceflight mission, a 3.5-day polar orbital flight launched on 31 March 2025 on a SpaceX Falcon 9 rocket. The imaging system comprised a commercial off-the-shelf, FDA-cleared, portable, digital X-ray generator (Impact Wireless; MinXray) along with an FDA-cleared flat-panel detector.

Prior to the mission, three crew members received 4 hr of training on the portable X-ray system and acquired preflight X-ray images of the hand, forearm, abdomen, pelvis, chest and a quality control phantom. They then recorded the same anatomic and phantom images whilst in space, and also imaged a smartwatch to investigate the potential for non-destructive testing.

During anatomic imaging, the devices were handheld by the crew members, while the phantom and smartwatch were secured to the detector for imaging. The crew reported that they found it easy to use the X-ray system and follow the imaging protocols.

Back on Earth

To assess the quality of the seven X-ray images acquired by the crew, three independent radiologists evaluated and compared the preflight and in-flight anatomic radiographs, finding no differences in overall image quality, contrast resolution or spatial resolution. For central-body radiographs (chest, abdomen and pelvis), image positioning was worse in-flight than preflight. Despite this, all X-rays achieved good or near-excellent diagnostic quality (rated on a  5-point Likert scale).

Phantom radiographs recorded in space exhibited good spatial and contrast resolution, imaging low-contrast targets down to the smallest 2-mm target and easily visualizing high-contrast meshes ranging from 20 to 60 lines per inch, with 80 lines per inch remaining visible. The smartwatch radiograph clearly showed the watch’s internal components at the submillimetre scale.

The research team found that the X-ray generator had sustained minor re-entry damage, but its internal components and X-ray output were not affected. Similarly, the detector passed visual inspection and quality control tests. The researchers also recorded a set of postflight X-rays (on non-crewmembers) following the same in-flight protocols. These showed no qualitative differences from the preflight or in-flight radiographs.

“By acquiring the first human and equipment X-rays in space, our study demonstrates the feasibility of in-orbit radiography and expanded diagnostic capabilities for crew health and hardware evaluation,” says Gifford. “Acquiring diagnostically useful X-rays in space is something that anyone can do. Three very talented nonmedical people with four hours of training in one of the harshest environments did it right and did it well.”

Gifford points out that a spaceflight-ready radiography system could also prove indispensable for mission-critical non-medical tasks. “For sustained human presence in space, X-rays are critical not just for crew members but also for other mission components like electronics and spacesuits. The only way to look inside these objects without taking them apart is to X-ray them,” she explains.

The post First medical X-ray images taken in space achieve diagnostic quality appeared first on Physics World.

A new glass screen containing nanoclusters of copper iodide could create higher-resolution X-ray images using a much lower dose of radiation. The scintillating screen, which also works underwater, can be moulded into curved shapes, overcoming key limitations of the conventional rigid, flat-panel detectors used for many medical imaging applications. These include mammography, say the researchers at King Abdullah University of Science and Technology (KAUST), Saudi Arabia, and the University of Hamburg, Germany, who developed it.

X-ray scintillators capture X-rays passing through an object and convert them into flashes of visible light to form a digital image, explains physicist Mehmet Bayindir of KAUST and the University of Hamburg. The more efficient the scintillator is at this conversion, the clearer the final image and the lower dose of radiation required to generate it.

He and his colleagues’ detailed analyses of a scintillator containing nanoclusters of copper, iodine and an organic ligand revealed a fascinating phenomenon: its photoluminescence and radioluminescence pathways appear to have distinct physical origins. This implies that the scintillating light yield – that is, how much visible light a material emits when struck by high-energy X-rays – can be separated from its standard photoluminescence quantum yield (how efficiently the material converts absorbed ultraviolet or visible light into a different wavelength of light).

Until now, Bayindir explains, it was assumed that a material must have high photoluminescence efficiency to perform well when excited by X-rays. “Breaking this assumption directly challenges conventional wisdom and opens up a promising new design space for making high-performance scintillators – provided that materials highly responsive to radiation can be engineered of course.”

Enter amorphous quantum systems

In recent years, researchers have been looking into cubane-type [Cu₄I₄L₄, L: organic ligand] nanocluster glasses (NCGs) for use in this context. These belong to a family of hybrid transparent luminescent amorphous materials that show much promise for X-ray scintillation applications.

In their latest work, detailed in ACS Energy Letters, Bayindir and colleagues studied several stable alkyl-phosphine ligand-based NCGs. In these amorphous zero-dimensional quantum systems, ionizing radiation is mostly absorbed by the heavy iodine and copper elements at the cluster cores. The researchers used these NCGs to make large-area, free-standing scintillating screens – both planar and conformal – with ultrasmooth surfaces, using a fabrication methodology that they developed in their lab.

The screens are robust and can image samples – both electronic and biological – with sub-3 µm spatial resolution. They also work underwater, something that is difficult for conventional screens. The researchers demonstrated this by capturing a very clear scan of a fish’s tail in water that was indistinguishable from an image taken in air.

Such a glass scintillator represents a significant milestone, says team leader Osman Bakr. “Indeed, the level of precision we obtained allows us to capture highly detailed, high-contrast X-ray images of complex internal microstructures, such as the delicate anatomy of a microscopic insect and an electronic memory card, to cite but two examples,” he says.

X-ray imaging systems that curve to fit a person’s body

“Combined with the fact that this new nanocluster glass becomes almost rubbery when heated to 42 °C, we can mould it into any three-dimensional shape at low temperatures,” study lead author Bashir Hasanov adds. “This could allow for X-ray imaging systems that curve to fit a person’s body and overcomes the limitations of conventional rigid, flat-panel detectors.”

One possible application is in mammography, he notes. Today, a major drawback of current mammography machines is that breast tissue must be firmly compressed between flat panels so that it can be imaged. This is very uncomfortable – and even painful – and deters many women from going for a scan.

“One in 20 women around the world will be diagnosed with breast cancer in their lifetime, with annual cases projected to reach 3.2 million and related deaths climbing to 1.1 million by 2050,” says Bayindir. “Many of these fatalities stem from late-stage diagnosis.

Nearly one-third of women miss their initial mammogram appointments, something that significantly increases their risk of succumbing to the disease. “While the reasons for this vary, the medical literature frequently cites the physical discomfort of tissue compression,” says Bakr. “What is more, because current guidelines recommend starting routine mammography screenings at age 40, cumulative lifetime exposure to X-ray radiation remains a clinical concern for many. Our long-term objective is to completely redefine this clinical experience.”

Indeed, the researchers say they would now like to bring this quantum glass out of the lab and into real-world applications. Bakr tells Physics World that his team’s immediate next step is to scale up the geometric dimensions of its mouldable screens while maintaining perfect thickness and clarity. “This would move us closer to full-scale clinical prototyping for 3D conformal medical imaging.”

“We are also designing a novel detector architecture capable of high-resolution imaging on curved surfaces using an array of specialized optical sensors,” he reveals.

The post Mouldable glass screen could make mammograms clearer and much more comfortable appeared first on Physics World.

Epithelial cancers account for 80% to 90% of all cancer cases in the world, comprising a wide variety of malignancies that begin in the epithelial tissues that form the skin and line the body’s internal passageways and organs. Early diagnosis is critical to improving patient outcomes, but is limited by a heavy reliance on invasive biopsies.

Simultaneous assessment of changes in both the cell nuclei and the microvasculature could potentially improve the accuracy of early cancer and precancer detection, and help guide decisions regarding biopsy and pathology testing. In vivo microscopy (IVM) has shown promise for aiding early diagnosis of epithelial cancers, but has challenges relating to image quality, spatial resolution and field-of-view (FOV).

PrecisionView, an artificial intelligence (AI)-powered, dual-modality endomicroscope in development at Rice University’s Rice360 Institute for Global Health Technologies, addresses these IVM issues. The compact, handheld, easy-to use system enables real-time, wide-area visualization of both subcellular structures and underlying blood vessels, offering higher-quality imaging, a fivefold increase in FOV, and an eightfold increase in spatial resolution over endomicroscopic systems in current clinical use.

The proof-of-concept version of PrecisionView, described in Proceedings of the National Academy of Sciences, is intended for superficial epithelial imaging where scattering-induced distortions remain moderate. And it is specifically designed to be low cost: providing an affordable diagnostic tool for clinics worldwide.

PrecisionView uses simple, off-the-shelf miniaturized optics integrated with a deep learning-based optimization framework. Its AI-supported phase mask design and reconstruction algorithm incorporate dual-modality fluorescence and reflectance imaging to achieve a 5.2 x 3.9 FOV and a 500 µm depth-of-field (DOF) while maintaining cellular-level resolution (4 µm). Images are acquired at approximately 15 frames/s, in 2 s intervals, with image tiles stitched together to form a large-scale, co-registered map of cell nuclei and microvasculature.

Performance assessment

To validate the system, principal investigators Rebecca Richards-Kortum and Ashok Veeraraghavan, and collaborators at the University of Texas MD Anderson Cancer Center, analysed images of ex vivo porcine tongue, human breast and fresh cervical tissue specimens with precancerous lesions.

High-resolution maps of cervical specimens containing squamous intraepithelial lesions (precancer) clearly delineated key anatomic structures, including columnar and squamous epithelium, and the junction between the two epithelial cell types. PrecisionView clearly distinguished abnormal regions from surrounding benign tissue – suggesting that the system may be able to support physicians’ decisions regarding whether or not to conduct a biopsy.

The researchers also performed real-time, in vivo imaging of the oral cavities of healthy volunteers. They report that the system’s large DOF enabled acquisition of high-resolution images of cell nuclei and microvasculature across a large tissue area without requiring refocusing. When the mucosal surface was scanned, the large FOV of each frame, combined with the high video acquisition rate, provided substantial frame-to-frame overlap. This allowed seamless stitching of the dual-modality images to create large-scale, co-registered maps of the oral mucosa, which showed distinct cellular and vascular features.

Meeting a global need

The Rice 360 Institute focuses on designing, implementing and scaling highly effective, low-cost medical technologies for underserved and low-resource communities. The idea for PrecisionView grew out of recognizing a global limitation in access to high-quality diagnostic information.

“In many care settings, especially where conventional pathology resources are limited, clinicians may not have timely access to the cellular- and tissue-level information needed to guide diagnosis and treatment decisions,” says first author Huayu Hou, who worked with co-first author Jimin Wu, to develop the PrecisionView instrument.

Hou explains that although existing clinical endomicroscopy systems can provide valuable microscopic information, their broader use is often limited by cost, complexity, a limited imaging volume and the ability to capture only a narrow range of diagnostic features. Such constraints reflect fundamental design trade-offs in compact clinical optical imaging systems and create significant barriers to practical point-of-care use.

“PrecisionView was motivated by the opportunity to use computational imaging to overcome limitations in existing clinical endomicroscope designs that could not be fully addressed with traditional optical approaches alone, with the goal of bringing high-quality diagnostic information closer to the point-of-care,” he says.

Hou advises that it took about a year to develop the first working prototype, and that several projects are now underway to further enhance both performance and usability for point-of-care applications. “We are working to improve image quality and resolution, so that the system can provide finer cellular details from living systems to real time,” he adds. “We are working to further refine the hardware and workflow to make the device easier to operate in clinical settings.”

The team is developing AI-based classification algorithms that can integrate both cellular and vascular information and provide real-time diagnostic guidance. Its long term goal is to create a system that not only visualizes tissue at the cellular level, but also helps clinicians interpret the information efficiently at the point-of-care. “This capability could be especially valuable in low-resource settings, where access to pathology expertise may be limited, and tools that support real-time decision-making could significantly improve clinical usability and impact,” notes Wu.

Clinical validation studies to evaluate PrecisionView’s diagnostic performance across different clinical settings and cancer types are also planned. For instance, a study focusing on head-and-neck cancer detection has already started in collaboration with MD Anderson Cancer Center. Others are being developed for cervical cancer imaging in Mozambique, to evaluate the system’s performance in low-resource settings.

The post Affordable AI-powered endomicroscope lines up for early cancer detection appeared first on Physics World.

Ronan Smith, a postdoctoral research fellow at Adelaide University, has been chosen as the winner of this year’s Physics in Medicine & Biology (PMB) Early Career Researcher Award. The award is presented to the author of the “best paper” in PMB’s 2025 Early Career Researcher Focus Collection, as selected by the journal’s editorial board.

Smith’s research involves implementing X-ray velocimetry (XV), a novel imaging method that uses X-rays to track lung motion during breathing and create 3D maps of local ventilation. In his award-winning paper, Visualising ventilation changes following endobronchial valve placement with x-ray velocimetry functional lung imaging, Smith investigates the potential of XV imaging to detect changes in lung function after insertion of an endobronchial valve (EBV).

EBVs are one-way valves that are placed into the lung to help treat emphysema – a condition that damages air sacs in the lungs, causing air to get trapped inside and making breathing difficult. The EBV, which in some cases can be used in place of surgery, prevents airflow into damaged lung areas so that the rest of the lung can function more effectively.

Successful valve placement causes the targeted area of lung to collapse, which can be imaged using CT. Smith proposed that use of XV functional lung imaging to non-invasively measure regional and local changes in airflow could more accurately assess the clinical impact of EBV placement.

“The lungs are a dynamic organ, their job is to be constantly moving,” he explains. “Because X-ray velocimetry looks at lung motion, it lets us see exactly where the air is or isn’t flowing, so you can instantly see that airflow has changed within the lungs. CT only measures structural changes, which may not necessarily be correlated to changes in lung function.”

In vivo demonstration

To investigate this potential advantage, Smith and colleagues carried out a pilot study on healthy sheep, which have a similar lung size to humans. They performed XV imaging on two anaesthetized and ventilated animals, before and after placing EBVs in their lungs.

The XV scanning process involves recording fluoroscopic videos of individual breaths at various angles around the lung, with anatomic positioning provided by an accompanying breath-hold CT scan. To analyse the data, the researchers used XV LVAS software from 4DMedical, the MedTech company that developed and commercialized the XV technology.

The software correlates motion in the XV videos with CT data to measure the lung’s expansion and contraction during a breath cycle. It creates a 3D map of specific ventilation (the change in voxel volume during a breath, divided by its starting volume) in small voxels throughout the lungs. This map can then be used to calculate mean specific ventilation and ventilation heterogeneity across a given lung region.

As soon as the EBVs were inserted into the animals’ lungs, XV imaging could visualize and quantify a reduction in airflow to areas downstream of the valves. This effect was seen both in regions where collapse was visible in CT scans, as well as those where collapse could not be detected by CT. Ventilation changes were also clearly observed in the remainder of the lungs.

“The main finding of our study was that X-ray velocimetry imaging can detect airflow changes from endobronchial valve placement in the lungs,” says Smith. “Our research could be really important for people [with emphysema], as tools to help with better placement and verification will lead to improved treatment options.”

Future prospects

Since the publication of this paper, Smith has focused on further applications of pre-clinical and clinical XV imaging. “I’ve been working as part of a great interdisciplinary team looking at how lung function changes in a range of diseases, to both understand the diseases, and as an outcome measure when we test treatments,” he says.

This work includes the world’s first paediatric clinical trial of XV imaging, which is examining the feasibility of using the technology in children with cystic fibrosis. The researchers have imaged around 30 children to date and aim to publish their findings later this year. They are currently planning future studies to see how XV imaging could enhance clinical decision making and improve outcomes for these children, as well as looking at other childhood diseases where it could be of relevance.

“As an early-career researcher, I’m also focussing on developing my own research, looking at another novel X-ray imaging method called dark-field X-ray imaging,” Smith adds.

The perfect award

Smith tells Physics World that he was excited to receive the PMB Early Career Researcher Award, acknowledging the efforts of everyone involved in this “hugely collaborative project”, including clinicians, scientists, 4DMedical and the staff in the preclinical imaging facility where the study was performed.

“As a physicist working in medicine/biology, it feels like the perfect award to get,” he says. “It’s great to see interest in the work we are doing, and fantastic evidence we can use to convince the funding bodies it’s worth continuing this work.”

• The PMB Early Career Researcher Focus Collection 2025 publishes papers from early-career researchers (defined for this collection as a postgraduate student or someone who completed their PhD in 2019 or later) in biomedical physics, with the aim of highlighting research excellence from emerging leaders. The Early Career Researcher Award was introduced to further recognise an outstanding contribution from one of the early-career authors, selected according to the quality of scientific content, number of citations and downloads, and peer review ratings.

The post X-ray velocimetry study earns Ronan Smith the PMB Early Career Researcher Award appeared first on Physics World.