Perhaps the most striking preclinical result comes from a new radiohybrid ligand pair designed for prostate cancer, a platform that fuses imaging and targeted alpha therapy in a single chemical scaffold. In mouse models, the therapeutic agent achieved 2.2-fold to 3.5-fold higher tumor uptake than a reference agent, while producing similar effects on tumor growth and survival. That combination is significant because targeted alpha therapy delivers highly potent, short-range radiation directly to tumor cells, and improving delivery to the tumor while sparing healthy tissue is the central challenge of the approach. The radiohybrid chemistry allows the same targeting molecule to be labeled with either an imaging isotope or a therapeutic alpha emitter, meaning physicians could potentially visualize where a drug goes before committing a patient to treatment. The authors say the findings support further investigation of this strategy, a step toward the kind of paired diagnostic-and-therapy paradigm that defines modern theranostics.

Artificial intelligence is the second major theme of the new releases. In a multicenter study, researchers built a vision–language model that integrates PET or SPECT imaging with clinical data to predict 24-month survival in patients with metastatic castration-resistant prostate cancer who are receiving PSMA-targeted radiopharmaceutical therapy. The model achieved 81.8 percent accuracy, outperforming conventional machine-learning approaches and showing promise in external validation—an important test of whether an algorithm trained at one set of hospitals can generalize to patients it has never seen. Metastatic castration-resistant prostate cancer is the lethal, treatment-refractory stage of the disease, and PSMA-targeted radioligand therapy has rapidly become a mainstay for these patients. A tool that can forecast survival from combined imaging and clinical inputs could help clinicians decide who benefits most from therapy, intensify treatment for high-risk patients, or redirect those unlikely to respond toward alternatives earlier in the course of disease.

Radiopharmaceutical therapy itself is expanding beyond prostate cancer. A French multicenter study evaluated lutetium-177 DOTATATE—(177)Lu-DOTATATE—in 69 patients with meningiomas that had not responded to previous treatments. These tumors arise from the membranes surrounding the brain and spinal cord, and when they resist surgery and radiation, options become scarce. The results offer a genuine foothold: six-month progression-free survival was 70 percent, with a median progression-free survival of 14.6 months. Grade 3 adverse events, mainly lymphopenia—a depletion of infection-fighting white blood cells—occurred in 16 percent of patients. For a tumor type long managed with blunt instruments, a receptor-targeted radiopharmaceutical that exploits somatostatin receptors on meningioma cells represents a meaningful shift toward molecularly guided care.

Complementing the therapeutic study, a separate investigation showed that quantitative PET measurements can capture the biology that makes such therapy possible. In 44 patients with meningiomas, quantitative measurements from gallium-68 DOTATATE PET/CT strongly correlated with Krenning scores, the visual scale clinicians use to grade somatostatin receptor expression. Two measures stood out: SUVmax, the maximum standardized uptake value within a lesion, and a ratio normalized to the superior sagittal sinus, a major venous channel that serves as an internal reference. These metrics distinguished clinically relevant scoring categories with areas under the curve reaching 0.98—near-perfect discrimination on the standard 0-to-1 scale. The implication is that visual scoring, which depends on reader experience, could increasingly be supplemented or replaced by reproducible numbers, standardizing patient selection for receptor-targeted therapy and trials.

Breast cancer imaging also received a boost. In a study of 30 patients, a targeted tracer called gallium-68 B7H3-BCH—(68)Ga-B7H3-BCH—used with PET/CT detected 96.8 percent of lesions, compared with 92.4 percent for standard fluorine-18 FDG PET/CT, the workhorse scan used across oncology. The difference may sound incremental, but its consequences were not: findings from the targeted scan prompted a change in disease stage or treatment in eight patients. B7H3 is a molecule associated with aggressive tumor biology, and a tracer that homes in on it can reveal disease that glucose-based imaging misses. In breast cancer, where the presence and location of additional lesions can determine whether a patient receives surgery, systemic therapy or both, a scan that changes management in roughly a quarter of a small cohort is a result that demands larger trials.

Behind every quantitative advance in nuclear medicine lies a quieter question: can the numbers be trusted? A prospective, two-center study addressed exactly that by evaluating the repeatability of fluorine-18 FDG PET/MRI measurements across different cancers and scanner models. Within-subject variability for SUVmax ranged from 9.5 percent to 12.3 percent, while SUVpeak variability was tighter, at 8.8 percent to 9.6 percent. Notably, differences in attenuation-correction maps—the computational corrections that account for how tissue absorbs radiation—had minimal impact on measurement repeatability. This matters because PET/MRI is a hybrid technology whose attenuation correction can be implemented in several ways; if those choices destabilized measurements, quantitative comparisons across centers would be compromised. The study suggests SUVpeak may be the more robust metric and that the technology is ready for the kind of multicenter quantitative trials modern oncology requires.

The final study ventures into territory rarely mapped: the normal body. In 20 adults without active malignancy, researchers examined the distribution of fluorine-18 AraG, a PET tracer that accumulates in activated T cells, the immune system’s front-line soldiers. Uptake patterns varied across brown adipose tissue, bone marrow, breast implant regions and the lungs, with some findings associated with medications or prior respiratory conditions. The work establishes a baseline atlas of where immune activation appears in healthy people, which is essential before the tracer can be used to image immunotherapy response in cancer patients. Immune imaging has long been a goal of oncology—being able to see whether checkpoint inhibitors are igniting a T-cell response in a tumor would be transformative—but such efforts require knowing what normal looks like first. This study provides that reference point, while also flagging confounders, from common medications to prior respiratory illness, that could otherwise be mistaken for pathology.

Taken together, the seven studies sketch the trajectory of a field that is becoming simultaneously more powerful and more precise. Therapeutic molecules are being engineered for better tumor delivery; AI models are learning to read images and clinical records together to forecast survival; targeted tracers are finding lesions that standard scans miss and changing treatment decisions as a result; and quantitative methods are being validated for repeatability across scanners and centers. The Journal of Nuclear Medicine, accessed more than 90 million times each year by practitioners worldwide, functions as the central clearinghouse for this progress, and its ahead-of-print pipeline means these findings reach clinicians before the formal issue date. For patients with cancers that resist conventional treatment—from refractory meningiomas to metastatic prostate disease—the studies offer something increasingly rare in oncology: multiple independent lines of evidence that molecular imaging and targeted radiopharmaceutical therapy are converting previously untreatable disease into manageable, measurable targets.

The broader significance of the research extends beyond any single result. Theranostics, the discipline that pairs a diagnostic isotope with a therapeutic one using the same targeting molecule, is moving from concept to clinical routine, and each of these studies contributes a piece: the radiohybrid ligands demonstrate the chemistry, the vision–language model demonstrates the analytics, the meningioma and breast cancer studies demonstrate the clinical applications, and the repeatability and baseline studies demonstrate the metrology that makes it all trustworthy. As the Society of Nuclear Medicine and Molecular Imaging notes, the goal is precision medicine—diagnosis and treatment tailored to individual patients to achieve the best possible outcomes. The October 2026 ahead-of-print releases show that goal being pursued on every front at once, from the chemistry bench to the scanner to the algorithm, and for patients facing the hardest diagnoses, that convergence cannot come soon enough.

Subject of Research: Advances in nuclear medicine imaging and radiopharmaceutical therapy for cancer

Article Title: The Journal of Nuclear Medicine ahead-of-print tip sheet: October 16, 2026

Article References: The Journal of Nuclear Medicine ahead-of-print tip sheet: October 16, 2026. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: nuclear medicine, theranostics, prostate cancer, radiopharmaceutical therapy, PET/CT, PET/MRI, meningioma, breast cancer imaging, artificial intelligence, PSMA-targeted therapy, molecular imaging, DOTATATE

The tracer at the heart of the study is O-(2-[18F]fluoroethyl)-L-tyrosine, commonly abbreviated as FET. Unlike the gadolinium-based contrast agents used in routine MRI, which highlight regions where the blood-brain barrier has broken down, FET is taken up by tumor cells through amino acid transporters that are upregulated in glioma tissue. This means the PET signal reflects metabolic activity rather than vascular leakage, allowing clinicians to visualize infiltrating tumor cells that extend beyond the enhancing core seen on contrast-enhanced T1-weighted MRI. International guidelines from the European Association of Nuclear Medicine, the European Association of Neuro-Oncology and the RANO group have already endorsed amino acid PET for glioma imaging, but prospective data from North American centers integrating FET-PET directly into radiotherapy planning have remained scarce. FETSMaRT was designed to fill that gap.

The study design was deliberately rigorous for a pilot. Thirty patients, all scheduled to receive post-operative radiotherapy, underwent a combined FET-PET/MRI acquisition before starting treatment. Crucially, the radiation oncologist who delineated the standard MRI-based target volumes was blinded to the PET data. Only afterward did the team generate a second set of targets derived from the FET-PET/MRI information, enabling a head-to-head, volumetric comparison of gross tumor volumes, clinical target volumes and planning target volumes. The median age of the cohort was sixty years, seventy-seven percent had glioblastoma, the most aggressive grade 4 form, and sixty percent carried a methylated MGMT promoter, a molecular marker associated with better chemotherapy response. This composition mirrors the typical high-grade glioma population seen in major treatment centers.

The headline numbers are striking. Discordance between the MRI-based and PET-derived gross tumor volumes occurred in seventeen of thirty patients, or 56.7 percent, while discordance at the clinical target volume level affected fifteen of thirty patients, exactly half of the cohort. Across the entire group, the FET-PET/MRI-derived volumes were consistently and significantly larger than the MRI-based ones. The gross tumor volume averaged 43.6 cubic centimeters on PET compared with 38.6 cubic centimeters on MRI, a difference that reached a p-value below 0.0001. The clinical target volume averaged 119.2 versus 110.8 cubic centimeters, with a p-value of 0.0045, and the planning target volume averaged 171.1 versus 159.6 cubic centimeters, with a p-value of 0.0003. In plain terms, the metabolic footprint of these tumors was reliably bigger than what the anatomical images suggested.

That consistency matters because of how radiotherapy for glioblastoma is actually delivered. Under current practice, shaped radiation beams are aimed at a target volume that expands outward from the visible tumor to account for anticipated microscopic infiltration, typically reaching a dose of sixty gray in thirty fractions over six weeks alongside the chemotherapy drug temozolomide. This treatment framework, established by the landmark Stupp trials of 2005 and 2009, has changed little in its imaging foundation even as radiation delivery technology has become dramatically more precise. If the biological target is systematically drawn too small, the most advanced linear accelerators in the world will still miss tumor cells that were never contoured in the first place.

The patterns of treatment failure recorded in the study reinforce that concern. Among patients whose disease progressed, the failures were predominantly central, meaning the tumor recurred within or near the high-dose radiation field in 60 percent of cases. Central failure is the classic signature of tumor cells that received radiation but survived, and it suggests that even the treated volume may not have covered the full metabolically active extent of disease. Historical dosimetric studies going back to the late 1990s have documented similar central recurrence patterns after three-dimensional conformal radiotherapy, and the FETSMaRT data add a metabolic imaging dimension to that long-standing problem.

The researchers also examined whether the size of the volume discrepancy predicted how quickly patients progressed. In the glioblastoma subgroup of twenty-three patients, Kaplan-Meier and Cox proportional hazards analyses related the magnitude of gross tumor volume and clinical target volume discordance to time to progression. The direction of the association was consistent: patients with larger discordance values tended to progress sooner. However, in this small pilot cohort neither association reached statistical significance. The authors are careful on this point, describing the findings as hypothesis-generating exploratory data that justify a larger prospective trial rather than proof of a survival effect. Small sample size is the obvious limitation, and the study was funded by a grant from the Brain Tumor Foundation Canada precisely to build the case for that definitive trial, which is registered as NCT06499831.

The broader context makes the timing of this work significant. A multicentre randomized European trial known as GLIAA, reported in The Lancet Oncology, recently compared FET-PET-guided re-irradiation against contrast-enhanced MRI guidance in patients with recurrent glioblastoma, adding momentum to the idea that amino acid PET should guide radiation decisions. Meanwhile, a separate Canadian phase 2 trial called UNITED has explored MRI-guided adaptive radiotherapy for high-grade glioma, and the TROG 18.06 FIG study in Australia has been prospectively evaluating FET imaging in glioblastoma across multiple centers. Taken together, these efforts signal a field-wide reassessment of the anatomical imaging paradigm that has governed brain tumor radiotherapy for a generation, with metabolic imaging positioned as the likely successor.

What would change for patients if FET-PET/MRI were adopted into routine planning? The volumetric differences documented in FETSMaRT, on the order of five to twelve cubic centimeters depending on the target level, are large enough to alter dose distribution in clinically meaningful ways, particularly for tumors located near eloquent brain regions where expanding the target risks harming functional tissue. Hybrid PET/MRI scanners, which acquire both modalities in a single session with perfectly registered images, remove much of the logistical friction that historically made PET integration cumbersome. The study authors note that deidentified participant data and analysis code will be available on reasonable request, supporting independent verification and pooled analyses as the evidence base grows.

For the roughly 77 percent of the cohort facing glioblastoma, a disease with a median survival that has crept only modestly above the fifteen-month mark despite two decades of intensifying therapy, the promise of a more accurate target is not an incremental technicality. Every cubic centimeter of metabolically active tumor that escapes the radiation field is a potential seed of recurrence, and central failures remain the dominant pattern of relapse. FETSMaRT does not prove that PET-guided contours will extend survival, but it establishes with prospective, blinded, quantitative rigor that conventional MRI systematically underestimates the biological extent of high-grade glioma in more than half of patients. The stage is now set for the definitive randomized trial that the Toronto team argues their pilot data demand, one that could finally align the physics of modern radiotherapy with the biology of the tumors it is meant to eradicate.

Subject of Research: Comparison of 18F-FET-PET/MRI and MRI-based radiotherapy target volumes in high-grade glioma

Article Title: Volumetric discordance between ¹⁸F-FET-PET/MRI and MRI-based radiotherapy target volumes in high-grade glioma: a prospective pilot study (FETSMaRT)

Article References: Singnurkar, A., Sahgal, A., Lechtman, E., Lim-Fat, M.-J., Graham, S. J., Lipsman, N., Maralani, P., Pond, G., & Detsky, J. S. (2026). Volumetric discordance between ¹⁸F-FET-PET/MRI and MRI-based radiotherapy target volumes in high-grade glioma: a prospective pilot study (FETSMaRT). Journal of Neuro-Oncology, 180(1), Article 20. https://doi.org/10.1007/s11060-026-05831-w

Image Credits: AI Generated

DOI: 10.1007/s11060-026-05831-w

Keywords: 18F-FET-PET, glioblastoma, high-grade glioma, radiotherapy planning, PET/MRI, target volume delineation, tumor recurrence, neuro-oncology, medical imaging, clinical trial, MGMT methylation, radiation oncology

The central molecular player is the serotonin 2A receptor, or 5-HT2AR, long suspected to be the key gateway to psychedelic experience. To visualize it, the researchers used the radioligand [11C]Cimbi-36, a PET tracer that binds to the same receptor. Seven healthy volunteers, blinded to the dose of LSD they received, underwent baseline PET-MRI scans and then returned for scans after oral doses of LSD ranging from 25 to 200 micrograms. By comparing tracer binding before and after the drug, the team could calculate receptor occupancy — the fraction of 5-HT2AR sites actually commandeered by LSD at each dose. This dose-occupancy relationship, now established in humans for the first time for LSD, gives clinical developers a quantitative anchor for choosing doses in therapeutic trials.

The hemodynamic findings were equally striking. LSD increased global cerebral blood flow, measured with pseudo-continuous arterial spin labeling, and also increased flow through the internal carotid arteries, the main pipelines supplying the brain, measured with phase-contrast mapping MRI. The two measures tracked each other closely across participants: changes in global cerebral blood flow and changes in carotid artery flow showed a strong positive correlation, indicating that the drug’s effect on brain perfusion is driven, at least in part, by increased supply from the great vessels rather than by a purely local redistribution of flow within the brain. This matters because serotonin receptors are abundant on systemic blood vessels, and psychedelics have long been known to interact with the vascular system in complex ways.

Perhaps the most provocative result concerns functional connectivity. Using resting-state BOLD fMRI, the team computed a global measure of brain-wide correlation, often called GCOR, and found that LSD decreased it — a reduction in the overall synchrony of activity across the cortex. Intriguingly, this drop in global connectivity was negatively associated with the rise in cerebral blood flow: the more blood flow increased, the more global connectivity fell. The authors contrasted this with a separate cohort of 25 participants who received psilocybin, where the relationship ran in the opposite direction, with LSD and psilocybin producing opposing effects on global correlation. The finding suggests that although both drugs act primarily at the 5-HT2AR, they are not interchangeable pharmacologically, and that their distinct neurophysiological signatures may have been masked in earlier studies that relied on a single modality.

That distinction carries a methodological warning for the entire field. The BOLD signal that dominates human psychedelic neuroimaging is not a pure readout of neural activity; it entangles neural drive with blood flow, blood volume and oxygenation. If LSD raises cerebral blood flow globally while dampening global connectivity, then some of the connectivity changes reported in prior fMRI studies may reflect vascular confounds rather than altered neural communication. By measuring perfusion, arterial inflow and connectivity simultaneously in the same individuals, the Copenhagen team could begin to disentangle these threads — and their data suggest the entanglement is real and drug-specific, differing between LSD and psilocybin in ways that single-modality designs would miss entirely.

The pharmacokinetics added another layer of intrigue. When the researchers plotted subjective drug intensity against plasma LSD concentrations over time, they observed an anticlockwise hysteresis loop: subjective effects lagged behind and outlasted what plasma levels alone would predict. This temporal dissociation between drug concentration in the blood and the intensity of the experience points to atypical pharmacodynamic mechanisms — perhaps slow receptor kinetics, active metabolites, or downstream signaling cascades that persist after the parent compound has cleared. Notably, the team’s extended analyses showed that psilocybin, converted to psilocin in the body, did not display a comparable hysteresis pattern, with subjective intensity tracking plasma psilocin levels without a consistent temporal offset. The two psychedelics, it seems, operate on different clocks.

The road to these results was not smooth, and the study’s transparency about its difficulties is itself informative. Of 53 volunteers assessed for eligibility, 16 attended in-person screening and 12 were randomized, with 11 completing baseline assessment and receiving LSD. Only seven had complete peak-scan data included in the final analysis; two participants could not complete post-drug scans because of adverse effects, and two scans were lost to scanner failure. Psychedelic research of this intensity — hours inside a PET-MRI scanner at the peak of a drug experience — is logistically brutal, and the attrition underscores how rare and valuable complete multimodal datasets of this kind are. Two of the seven completers also returned for late-effect scans, offering a preliminary glimpse of how the drug’s cerebral effects evolve as the acute experience subsides.

Why does receptor occupancy matter so much for the clinical pipeline? Psychedelic-assisted therapies are advancing through late-stage trials for depression, anxiety and other conditions, yet dosing has remained largely empirical, guided by subjective intensity rather than by molecular target engagement. The new dose-occupancy curve allows researchers to connect the pharmacological dose, the fraction of receptors engaged, the cerebral blood flow response and the reported experience within a single quantitative framework. That framework can inform the design of safer trials, help interpret why different psychedelics produce different durations and qualities of effect, and support the development of compounds engineered to retain therapeutic benefit while minimizing cardiovascular and perceptual side effects.

The comparison with psilocybin also carries a broader scientific message: the class of drugs we call psychedelics is pharmacologically heterogeneous. LSD is an ergoline with a long duration of action and affinity for a wider panel of receptors, while psilocybin’s active metabolite is a tryptamine with a shorter, cleaner profile. The opposing effects on global connectivity observed here, together with the divergent hysteresis patterns and the distinct relationships between blood flow and connectivity, suggest that results from one compound cannot simply be generalized to the other. For a field racing toward regulatory approval, that nuance is not academic; it shapes expectations about which patients may benefit from which compound, and what physiological monitoring each requires.

What remains to be resolved is how these molecular and vascular effects translate into the therapeutic outcomes that matter. The study involved healthy volunteers at a range of doses, not patients, and its sample size — seven completers — is small by any standard, with some correlations reported without correction for multiple comparisons. The raw neuroimaging data have been deposited in the Cimbi database for qualified researchers, and the analysis code has been released publicly, inviting independent scrutiny and reanalysis. As psychedelic medicine matures, studies of this kind — direct, multimodal, and honest about their limitations — will be essential for converting a cultural renaissance into a rigorous clinical science, one receptor and one milliliter of blood at a time.

Subject of Research: In vivo PET-MRI measurement of LSD receptor occupancy, cerebral blood flow and brain connectivity in humans

Article Title: Investigating molecular and hemodynamic effects of LSD in the human brain

Article References: McCulloch, D. E.-W., Larsen, K., Johansen, A., Reveles Jensen, K. H., Nykjær, C. H., Holze, F., Falck, N., Neufeld, V., Steenstrup, E., Skov-Andersen, P., Spanggård, A., Geisler, M., Randrup, P. P., Jensen, P. S., Shalgunov, V., Johansen, S. S., Nielsen, M. K. K., Andersen, T. L., Stenbæk, D. S., … Knudsen, G. M. (2026). Investigating molecular and hemodynamic effects of LSD in the human brain. Nature Neuroscience. https://doi.org/10.1038/s41593-026-02462-2

Image Credits: AI Generated

DOI: 10.1038/s41593-026-02462-2

Keywords: LSD, psychedelics, serotonin 2A receptor, PET-MRI, cerebral blood flow, functional connectivity, psilocybin, receptor occupancy, pharmacokinetics, neuroimaging, 5-HT2AR, Nature Neuroscience

This long-standing gap is the target of a new study published in Medical & Biological Engineering & Computing by Seyede Sajede Mousavi Eshkiki and Alireza Sadremomtaz of the University of Guilan in Iran. Rather than proposing yet another neural-network architecture, the pair set out to build and rigorously evaluate a complete, two-stage workflow that takes ordinary T1-weighted MRI of the brain and produces both a CT-like image and a fully attenuation-corrected PET scan. Crucially, they wrapped the entire pipeline in a graphical user interface, so that the deep learning machinery can be operated through inference without requiring users to write code. The work is an exercise in systems engineering as much as machine learning: the question was not whether a clever model could be designed, but whether a sequence of established components could be assembled into a workflow that behaves reliably enough to matter clinically.

The researchers drew their data from the CERMEP-IDB-MRXFDG cohort, a publicly available database of thirty-seven healthy adults who underwent fluorodeoxyglucose PET, T1-weighted and FLAIR MRI, and CT imaging of the brain. Because every subject had a real CT scan, the team could train their networks against genuine ground truth and then measure how closely the synthetic versions matched reality. They used subject-level five-fold cross-validation, meaning that in each round of testing the network had never seen the individual whose images it was correcting. This patient-level separation matters, because it prevents the model from effectively memorizing a person’s anatomy and inflating its apparent accuracy.

The first stage of the workflow is a two-dimensional U-Net, the workhorse convolutional network of biomedical image analysis, first introduced in 2015 and still favored for its efficiency and strong performance on relatively small datasets. Fed only a T1-weighted MRI volume, the network predicts a pseudo-CT: a synthetic image in Hounsfield units that approximates what a CT scanner would have measured for the same head. Once generated, this pseudo-CT is converted into an approximate linear attenuation coefficient map, the so-called mu-map that describes how strongly tissue attenuates the 511-keV photons emitted during positron annihilation. That mu-map, paired with the non-attenuation-corrected PET image, then serves as the input to a second U-Net, which synthesizes a predicted attenuation-corrected PET scan directly.

The logic of this second stage is subtle and worth unpacking. In principle, one could simply apply the mu-map mathematically to correct the PET data, as conventional reconstruction does. But the second network learns the mapping from uncorrected PET plus attenuation information to corrected PET, allowing it to absorb residual imperfections in the pseudo-CT. If the synthetic attenuation map slightly misestimates bone density, for example, the second network can still learn to produce an output consistent with properly corrected images, because it has seen many paired examples of what uncorrected and corrected PET look like for the same brain. The two networks therefore act as a chain, with the first converting MRI into attenuation-relevant anatomy and the second converting that anatomy plus raw PET into a quantitatively usable scan.

The performance figures tell a encouraging story with a familiar caveat. The pseudo-CT model achieved a whole-image mean absolute error of 72.26 plus or minus 20.11 Hounsfield units and a structural similarity index of 0.881 plus or minus 0.024 against the real CT scans. The authors note that the residual error was concentrated in bone-dominated regions, which is exactly where MRI-to-CT synthesis has always struggled. Bone produces almost no signal on conventional MRI, so the network must essentially infer cortical skeleton from the shape of the brain and surrounding tissues. Decades of attenuation-correction research, from segmentation-based methods to atlas-based approaches and zero-echo-time MRI sequences, have wrestled with the same limitation, and deep learning approaches inherit it to varying degrees.

Where the workflow truly shines is in the PET correction itself. Compared with the raw, non-attenuation-corrected PET images, the predicted attenuation-corrected PET achieved a 76.02 percent lower normalized mean absolute error and a 73.33 percent lower normalized root mean square error relative to the reference attenuation-corrected PET. In other words, the two-stage pipeline recovered roughly three quarters of the quantitative error introduced by ignoring attenuation. For brain imaging, where attenuation is less severe than in the thorax or abdomen but still clinically significant, that level of recovery brings synthetic correction into the same conversation as established vendor solutions.

The exploratory standardized uptake value analysis adds a dose of realism. Standardized uptake values are the currency of quantitative PET, and the workflow delivered a mean percentage error of 8.66 percent for SUVmean but 22.08 percent for SUVmax, with greater variability across subjects in the maximum metric. This pattern is well understood in the field: SUVmean averages over a region and tends to forgive localized errors, while SUVmax hinges on a single hottest voxel and is exquisitely sensitive to small misestimates of attenuation, particularly near bone. The authors are appropriately measured, framing these results as supporting internal feasibility on the CERMEP dataset while emphasizing that nested cross-validation and external validation on independent cohorts remain necessary before any clinical claim can be made.

What makes the study notable beyond its numbers is its orientation toward usability and transparency. The graphical user interface lowers the barrier for imaging scientists and clinicians who lack programming expertise, echoing earlier efforts such as the Deep Learning Application Engine to bring neural networks out of notebooks and into the imaging suite. Just as importantly, the team has released the source code, trained models, cross-validation fold assignments, configuration files, and evaluation outputs through the Zenodo repository, allowing other groups to reproduce the results, stress-test the workflow on their own data, and extend it. In a subfield where reproducibility has often lagged behind headline accuracy figures, that openness is a meaningful contribution in its own right.

The broader significance lies in what accessible attenuation correction could unlock. PET/MRI offers simultaneous acquisition of molecular and anatomical information with a fraction of the radiation dose of PET/CT, making it attractive for pediatric imaging, longitudinal neuroscience studies, and repeated oncological follow-up. If a workflow like this one, running on nothing more exotic than a standard T1-weighted scan, can deliver dependable attenuation correction without dedicated sequences or vendor-proprietary tools, it could lower the cost and complexity of quantitative PET/MRI for research centers worldwide. The caveats stand: thirty-seven healthy subjects, a single tracer, and no external cohort yet. But as a carefully documented proof of feasibility, the study charts a practical path from the physics problem that has shadowed PET/MRI since its inception toward a solution that any lab with a GPU and a GUI could pick up and run.

Subject of Research: Deep learning-based attenuation correction for brain PET/MRI using pseudo-CT synthesis

Article Title: A two-stage PET/MRI attenuation-correction workflow for pseudo-CT and AC PET synthesis with GUI-based inference

Article References: Eshkiki, S. S. M., & Sadremomtaz, A. (2026). A two-stage PET/MRI attenuation-correction workflow for pseudo-CT and AC PET synthesis with GUI-based inference. Medical & Biological Engineering & Computing. https://doi.org/10.1007/s11517-026-03685-y

Image Credits: AI Generated

DOI: 10.1007/s11517-026-03685-y

Keywords: PET/MRI, attenuation correction, pseudo-CT, U-Net, deep learning, mu-map, NAC PET, SUV, CERMEP-IDB-MRXFDG, medical imaging, radiotracer imaging, quantitative PET

A new study published in the European Journal of Nuclear Medicine and Molecular Imaging suggests that a hybrid imaging technique may finally bring quantitative rigor to this assessment. Researchers at Peking Union Medical College Hospital in Beijing prospectively enrolled 33 patients with thyroid-associated ophthalmopathy and scanned them on an integrated PET/MRI system using a radiotracer called 18F-NOTA-FAPI-04. Unlike conventional FDG, which tracks glucose metabolism in any active cell, FAPI tracers bind to fibroblast activation protein, a molecule displayed by activated fibroblasts and their inflammatory partners. Because thyroid eye disease is fundamentally a fibroinflammatory disorder of the orbital tissues, the tracer offers a way to visualize the cellular machinery driving the disease rather than its downstream anatomical consequences.

The technical logic of the study is worth unpacking. Each patient’s orbits were imaged simultaneously with PET, which quantifies molecular uptake, and with a battery of advanced MRI sequences that probe tissue microstructure. The MRI panel included T2 mapping, which reflects tissue water content and therefore edema; intravoxel incoherent motion diffusion imaging, which separates true water diffusion from blood flow within capillaries; and diffusion kurtosis imaging, which captures deviations from simple Gaussian diffusion and thus hints at tissue complexity. From these sequences the team extracted parametric maps of normalized T2 signal, the heterogeneity index alpha, the distributed diffusion coefficient, the diffusion coefficient, relative blood flow, mean diffusivity, and mean kurtosis. The goal was to see whether the molecular signal from PET and the microstructural signal from MRI tell the same story, and whether combining them outperforms either alone.

Participants were classified using the Clinical Activity Score, with eyes scoring three or higher considered active and those below three considered inactive. At the patient level, a person was labeled inactive only if both eyes were inactive. The researchers then extracted a rich set of quantitative features from each eye: PET metrics including maximum and mean standardized uptake values, metabolic tumor volume, and total lesion FAPI uptake, plus first-order histogram statistics such as mean, skewness, and kurtosis from each MRI parametric map. A support vector machine, a standard machine-learning classifier, was trained to distinguish active from inactive disease, and performance was evaluated with receiver operating characteristic analysis.

Reliability came first, and the numbers were reassuring. Interobserver and intraobserver agreement in delineating the orbital regions of interest was strong, with Dice similarity coefficients of 0.850, 0.880, and 0.901 for the different comparisons. Every quantitative PET and MRI parameter showed excellent reproducibility, with intraclass correlation coefficients above 0.907. In a field where subjective eyeballing of scans has long been the norm, this level of measurement stability matters: it means the features being fed into the classifier are not artifacts of who happened to draw the contours.

The correlation analysis produced one of the study’s most interesting nuances. At the patient level, where both eyes are averaged together, PET and MRI features correlated only weakly. But when each eye was analyzed as its own unit, the correlations strengthened considerably, particularly between PET uptake metrics and features derived from T2 mapping, the heterogeneity index, the diffusion coefficient, and mean diffusivity. This asymmetry makes biological sense. Thyroid eye disease is notoriously asymmetric, with one eye often far more inflamed than the other, and averaging across eyes dilutes the very signals that matter. The finding is a quiet argument for eye-level, rather than patient-level, imaging assessment in orbital disease.

The group comparisons confirmed that several quantitative features separate active from inactive disease. At the eye level, mean standardized uptake value, metabolic tumor volume, and total lesion FAPI uptake were all significantly higher in active eyes, while maximum standardized uptake value alone showed no significant difference, a reminder that peak values are often less informative than volume-weighted averages. On the MRI side, features from T2 mapping, the heterogeneity index, the diffusion coefficient, mean diffusivity, and mean kurtosis also differed significantly between active and inactive groups, consistent with the idea that active disease carries more edema and altered diffusion characteristics than fibrotic, quiescent tissue.

When it came to discriminating activity, the combined approach won. Among single PET parameters, mean standardized uptake value performed best at the patient level with an area under the curve of 0.759, while metabolic tumor volume led at the eye level with an AUC of 0.767. These are respectable but unremarkable figures, in the range of many clinical biomarkers. The combined PET/MRI model, however, pushed performance to an AUC of 0.844 at the patient level and 0.859 at the eye level. The improvement over either modality alone indicates that FAPI uptake and MRI-derived microstructural features carry complementary, not redundant, information about the state of the orbital tissues. In practical terms, the PET signal appears to capture the cellular fibroinflammatory process while the MRI parameters capture its tissue-level consequences, and the classifier benefits from having both.

The clinical implications are significant. Treatment for active thyroid eye disease, including high-dose glucocorticoids and newer targeted agents such as anti-IGF-1 receptor antibodies, works best early in the inflammatory phase and offers little once fibrosis has set in. Conversely, rehabilitative surgeries are best deferred until the disease is inactive. A quantitative imaging biomarker that reliably separates these phases could spare patients from ineffective treatment, reduce exposure to steroid side effects, and help time interventions more precisely. It could also serve as an objective endpoint in clinical trials, where the Clinical Activity Score’s known inter-observer variability has long complicated the interpretation of results. Prior studies have explored FDG-PET and multiparametric MRI separately for this purpose, but the FAPI tracer’s specificity for activated fibroblasts, combined with simultaneous MRI acquisition on a single scanner, represents a meaningful step forward in what the field calls molecular-radiological phenotyping.

Cautions remain, and the authors are candid about them. Thirty-three patients is a small cohort, and the classifier’s performance will need validation in larger, independent, and ideally multi-center populations before it can influence routine care. The Clinical Activity Score itself, used here as the reference standard, is an imperfect ground truth, which means the imaging model is being trained to reproduce a clinical judgment rather than to measure disease biology directly. The study was registered as a clinical trial and conducted under ethics approval at Peking Union Medical College Hospital, and the team notes that further validation is required. Still, the trajectory is clear: a single integrated scan that fuses molecular information about fibroblast activity with quantitative maps of tissue edema, perfusion, and diffusion could transform thyroid eye disease from a condition assessed by counting symptoms into one measured, monitored, and treated on the basis of what is actually happening inside the orbit. For a disease that can quietly steal sight, that kind of clarity cannot come soon enough.

Subject of Research: Quantitative assessment of disease activity in thyroid-associated ophthalmopathy using integrated 18F-NOTA-FAPI-04 PET/MRI

Article Title: Integrated 18F-NOTA-FAPI-04 PET/MRI for quantitative assessment of disease activity in thyroid-associated ophthalmopathy

Article References: Yang, X., Gan, L., Shi, X., Wu, M., Li, E., Zhang, Y., Hao, Z., Huang, Z., Xing, H., Liu, X., & Huo, L. (2026). Integrated 18F-NOTA-FAPI-04 PET/MRI for quantitative assessment of disease activity in thyroid-associated ophthalmopathy. European Journal of Nuclear Medicine and Molecular Imaging. https://doi.org/10.1007/s00259-026-08157-x

Image Credits: AI Generated

DOI: 10.1007/s00259-026-08157-x

Keywords: thyroid-associated ophthalmopathy, thyroid eye disease, PET/MRI, 18F-NOTA-FAPI-04, fibroblast activation protein, Clinical Activity Score, multiparametric MRI, intravoxel incoherent motion, diffusion kurtosis imaging, radiomics, support vector machine, Graves orbitopathy