The research, conducted by Berkay Çağdaş of the Department of Nuclear Medicine at Afyonkarahisar State Hospital and Ege University’s Institute of Nuclear Sciences, and Elif Kardelen Çağdaş of the hospital’s Department of Pathology and Ankara University’s Graduate School of Biotechnology, set out to determine whether regional differences in immune and stromal composition inside breast tumors correspond to measurable differences in how the tumor consumes glucose on PET imaging. The premise is biologically plausible: tumors infiltrated by lymphocytes often behave differently from immune-cold tumors, and fibroblast-like stromal cells can alter metabolism, perfusion, and tissue architecture in ways that might plausibly change fluorodeoxyglucose uptake.

The study enrolled 61 women with invasive breast carcinoma who underwent baseline 18F-FDG PET/CT before treatment. Rather than treating each tumor as a single homogeneous blob, the team took a deliberately regional approach. A pathologist, blinded to all PET findings, annotated ten separate intratumoral regions on whole-slide histopathology images, each at least one square millimeter in area. QuPath, an open-access machine-learning platform for digital pathology, then quantified two key features within each region: the density of stromal tumor-infiltrating lymphocytes, the immune cells that gather in the connective tissue surrounding cancer cells, and the density of fibroblast-like stromal cells, spindle-shaped cells identified morphologically on routine hematoxylin and eosin stains.

From those ten regions, the investigators retained the ones with the numerically highest and lowest lymphocyte density for paired analysis, an approach designed to capture the observed range of immune-stromal composition within each tumor. The authors are careful to note that these paired regions represent the extremes of what was measured, not a comprehensive spatial map of every tumor. On the imaging side, the team extracted a rich panel of 30 PET features from each patient’s scan. These included the familiar intensity measures such as maximum, mean, and peak standardized uptake values, volumetric measures including metabolic tumor volume and total lesion glycolysis, texture descriptors derived from gray-level co-occurrence, run-length, and size-zone matrices, and dedicated heterogeneity indices, including the HI2 and HI3 indices and a slope calculated across metabolic volume thresholds from 30 to 50 percent of maximum uptake.

The statistical framework was deliberately conservative. The team ran Spearman correlations between nine immune-stromal endpoints and all 30 PET features, generating 270 individual tests, and then applied Benjamini-Hochberg correction to control the false discovery rate, a standard safeguard against being fooled by chance when many comparisons are made simultaneously. The results were sobering but informative. Thirty-eight of the 270 correlations reached nominal significance at the conventional p-value threshold of 0.05, yet not a single one survived the false discovery rate correction, whether the correction was applied within each endpoint family or across the entire study. The smallest corrected q-value within an endpoint was 0.077, and study-wide it was 0.33, both well above any threshold that would conventionally be called significant.

The strongest raw association was, by the authors’ own framing, weak. The density of stromal tumor-infiltrating lymphocytes in the TIL-rich region correlated with the HI2 heterogeneity index at a Spearman coefficient of 0.379 in absolute value, with a 95 percent confidence interval spanning 0.126 to 0.586 and a nominal p-value of 0.0026. Beyond that single headline number, a coherent pattern emerged in the direction of the nominal findings. Lymphocyte density correlated inversely with metabolic tumor volume, with coefficients ranging from negative 0.29 to negative 0.35, and positively with the HI2 heterogeneity measure. The ratio of lymphocytes to fibroblast-like stromal cells correlated with SUV-based parameters, the largest being a coefficient of 0.299 with mean standardized uptake value. Fibroblast density on its own, and the metrics describing regional differences between the paired areas, produced weaker and less consistent signals.

One of the study’s most technically rigorous contributions is its validation of the pathology measurement itself. Because fibroblast-like stromal cells were defined morphologically on routine H&E stains, the authors cross-checked their machine-learning quantification against alpha-smooth muscle actin immunohistochemistry, a stain that highlights activated fibroblasts and myofibroblasts, in the same 122 paired regions. The agreement was remarkably tight: a correlation of 0.992 and a Lin concordance correlation coefficient of 0.996. When the entire correlation analysis was repeated using the immunohistochemistry-based stromal density instead of the H&E-based one, the profile of results was reproduced almost exactly, with correlations between the two analyses ranging from 0.94 to 0.996. This means the null finding is not easily explained by measurement error in the pathology pipeline.

The team also stress-tested the associations that did reach nominal significance. When extreme lymphocyte densities or the smallest annotated regions were excluded from the analysis, between 9 and 24 of the original 38 nominal associations persisted, indicating that some of the signals were fragile and dependent on particular data points. After statistical adjustment for clinical covariates including T stage, Nottingham histologic score, Ki-67 proliferation index, and molecular subtype, only 13 of the 38 associations remained nominally significant. The picture that emerges is one of faint, directionally interesting signals that dissolve under the weight of proper multiple-comparison correction, exactly the pattern one expects when exploring a genuinely uncertain hypothesis with a modest sample size.

Why does this matter? PET radiomics has generated enormous enthusiasm in oncology, with hundreds of studies proposing imaging texture features as noninvasive surrogates for everything from gene expression to immune infiltration. The promise is seductive: if a scan could reveal the immune landscape of a tumor without a biopsy, clinicians could select patients for immunotherapy, track treatment response, and monitor clonal evolution noninvasively. But the field has been repeatedly criticized for underpowered studies, lax statistical correction, and features that fail to replicate. This study is a refreshing counterexample. By pre-specifying a conservative statistical framework, validating its pathology measurements against an independent stain, and openly reporting that nothing survived correction, the authors have modeled the kind of intellectual honesty that imaging science needs more of.

The authors themselves are unambiguous about the limits of their work. This is a hypothesis-generating, exploratory, single-institution study of 61 patients, and the paired-region design captures only the extremes of immune-stromal composition rather than a full spatial map. They state plainly that the findings require independent validation in larger, multi-institutional cohorts before any biological or clinical interpretation can be drawn. That caution is warranted, but so is the value of the attempt. The study demonstrates a feasible, reproducible workflow for integrating region-level digital pathology with PET radiomics in breast cancer, and it establishes a realistic benchmark for how weak the true links between tissue microenvironment and metabolic imaging may actually be. If future larger studies confirm even the modest directional trends seen here, immune-rich tumors may one day be recognizable on PET as smaller, more heterogeneous glucose-avid lesions. For now, the honest answer to the question of whether a PET scan can see a tumor’s immune landscape is: perhaps faintly, and not yet reliably.

Subject of Research: Associations between regional immune-stromal tumor features and 18F-FDG PET metabolic heterogeneity in breast cancer

Article Title: Exploratory associations between regional immune-stromal features and 18F-FDG PET-derived metabolic heterogeneity in breast cancer

Article References: Çağdaş, B., & Çağdaş, E. K. (2026). Exploratory associations between regional immune-stromal features and 18F-FDG PET-derived metabolic heterogeneity in breast cancer. BMC Medical Imaging. https://doi.org/10.1186/s12880-026-02789-z

Image Credits: AI Generated

DOI: 10.1186/s12880-026-02789-z

Keywords: breast cancer, PET imaging, radiomics, tumor microenvironment, tumor-infiltrating lymphocytes, digital pathology, metabolic heterogeneity, 18F-FDG, QuPath, fibroblasts, false discovery rate, machine learning

The central problem the team set out to solve is one that has haunted cancer metabolomics for years. Untargeted high-resolution LC-MS can produce sensitive, global snapshots of the metabolites consumed and produced by cancer cells, but a tumor is not a uniform bag of cells. Genetic mutations, uneven oxygen and nutrient availability, signaling molecules, and a mixture of malignant, stromal, immune, and normal epithelial cells all shape the metabolic environment. A metabolite detected in a bulk tumor sample may come from cancer cells, from dying tissue, from fibrous scar tissue, or from fat-laden hepatocytes, and it may reflect tumorigenesis or entirely unrelated processes such as steatosis, fibrosis, necrosis, or xenobiotic metabolism. That ambiguity, the authors argue, is a key reason why candidate metabolic markers often prove highly discriminative in one patient cohort but fail to generalize to others.

To address this, the researchers defined a metabolic region as a local tissue domain characterized by the spatial presence of metabolites mechanistically associated with key metabolic or biological processes in tumor metabolism. Importantly, the concept is deliberately broader than histology: the metabolites defining a MER may be generated locally or delivered from adjacent or distant niches through transport or diffusion, so a single MER can span multiple cellular niche types with different histological appearances. The idea draws on the well-known concept of metabolic zonation in the healthy liver, in which spatially compartmentalized pathways prevent competition for shared substrates, but extends it explicitly to tumor tissue, where regions can overlap on top of one another.

The technical workflow is a four-stage pipeline that fuses the mass separation power of LC-MS/MS with the spatial separation power of MSI. First, the team systematically identified highly abundant mass features from aligned and normalized untargeted LC-MS profiles of tumor samples, annotating putative metabolites using MS/MS fragmentation patterns matched against the Human Metabolome Database and Lipid Maps. Second, these abundant metabolites were co-detected in desorption electrospray ionization mass spectrometry imaging (DESI-MSI) of tissue sections from the same tumors, and pathologist-annotated regions of interest on matching hematoxylin and eosin (H&E) images were used to train partial least-squares discriminant analysis (PLS-DA) classifiers that assign every pixel of a tissue section to a MER. Third, a new computational method called COMET’s Path (COrrelated MEtabolomics and TranscriptomicS Pathway analysis) correlates bulk RNA-seq and LC-MS profiles across samples to infer the biological processes associated with each region. Fourth, regression models trained on the resulting maps can predict the MER composition of bulk tumor samples from LC-MS data alone.

The study used 117 tumor and matched adjacent-normal tissue sections from 26 surgically resected primary liver tumors in the Precision Medicine in Liver Cancer across an Asia-Pacific Network (PLANet) cohort, 24 of them hepatocellular carcinoma (HCC) and two intrahepatic cholangiocarcinoma. Multiple sectors were harvested from each tumor to capture intratumoral heterogeneity. The LC-MS profiles of HCC tumors, iCCA tumors, and adjacent-normal tissues could be almost perfectly separated, hinting that the two cancer types disrupt liver metabolism differently and might even be distinguishable by metabolic biomarkers. Across HCC stages, more than 70 percent of mass features changed monotonically with cancer stage, and metabolomics profiles separated Stage IB from Stage III tumors more distinctly than transcriptomics profiles did, suggesting that metabolite-based markers for early-stage disease are worth pursuing.

Yet the heterogeneity data revealed a catch. Inter- and intratumoral heterogeneity scores, computed from the multi-sector samples, peaked at Stage IB for metabolomics, meaning early-stage HCC is the most metabolically diverse stage of all. Of 8742 mass features detected by LC-MS, the team focused on 265 highly abundant ions, confirmed 230 by MS/MS, and matched 206 to known metabolites, dominated by glycerophospholipids, glycerolipids, and fatty acyls. When DESI-MSI was performed on 52 sections from 11 patients, 78.7 percent of the MS/MS-confirmed abundant metabolites could be matched to the imaging data, and bulk LC-MS and averaged MSI abundances correlated significantly across samples, confirming that the imaging measurements were reproducible. But when tissue-averaged abundances were compared between tumor and normal sections, only a small fraction of the 173 co-detected metabolites changed significantly, a result the authors attribute directly to the mixing of metabolically distinct regions within each sample.

The classifiers changed that picture dramatically. Trained on 116 pathologist-annotated regions representing six histological features of HCC, three PLS-DA classifiers distinguished six MER types: metabolically normal, low-grade, high-grade, necrotic, fibrotic, and steatotic regions. The classifiers achieved balanced training accuracies of 81.5 to 99.9 percent and balanced test accuracies of 72.6 to 97.7 percent, with the normal, high-grade, and necrotic classifiers exceeding 90 percent on test data, and they outperformed support vector machine alternatives. Each MER was defined by a distinct set of discriminative metabolites, including phosphatidylcholine species such as PC(18:0/20:4) and PC(16:0/20:4), cholesterol sulfate, diacylglycerols, lysophosphatidylcholines, oleoylcarnitine, and 13′-OH-alpha-tocopherol. Notably, no single dominant metabolite emerged for any region, underscoring that single-marker approaches are insufficient to recognize these metabolic niches.

COMET’s Path analysis then attached biology to the geography. Fatty acid and lipid metabolic processes were enriched in the normal, low-grade, and necrotic regions but, surprisingly, not in the high-grade regions, while cell cycle and G2/M checkpoint processes were enriched only in low-grade regions. High-grade regions instead correlated with adaptive immune processes, including B-cell receptor, high-affinity IgE receptor, and complement activation, suggesting that transformed hepatocytes in these areas may slow proliferation while B cells drive chronic inflammation and immune suppression. The analysis also assigned previously reported HCC marker genes to specific regions: TERT, an early driver of cirrhosis-to-cancer transformation, mapped to fibrotic regions; BCL9, a Wnt pathway modulator linked to lipid formation, mapped to steatotic regions; LSM4, a strong diagnostic candidate from prior bioinformatics studies, correlated most strongly with low-grade region metabolites; and PLAUR, encoding the urokinase receptor, mapped to necrotic regions, suggesting it may mark tissue repair rather than tumorigenesis itself. Spatial transcriptomics on sections from two additional patients confirmed that most correlated gene sets were expressed in the expected histological regions.

The most consequential finding is quantitative: more than 50 percent of the highly abundant metabolites detected in bulk HCC tumors appear to originate from non-malignant regions, namely necrotic, fibrotic, or steatotic areas, and are therefore unlikely to serve as predictive or reproducible biomarkers of tumorigenesis. The maps also exposed a hidden class of markers that bulk analysis actively mislabels. Metabolites in the “ME-low-grade+/ME-necrotic-” cluster, about 21 percent of those detected, rise in low-grade tumor regions but fall sharply in necrotic regions, so at the tissue-averaged level they appear to be decreasing in tumors and would be discarded by conventional pipelines. Necrotic regions turned out to both generate and deplete the largest numbers of specific metabolites, hinting at an abrupt metabolic reprogramming when rapidly dividing hepatocytes die. Regression models trained on the maps deconvoluted bulk LC-MS profiles with coefficients of determination between 0.731 and 0.936, and the predicted region compositions aligned with pathologist-assigned Edmondson-Steiner grades, necrosis, steatosis, and Metavir fibrosis scores, though the authors caution that the high R-squared values may reflect some overfitting and require validation in independent cohorts.

The implications reach beyond liver cancer. Because the approach relies on standard histopathological features for training, it is interpretable and could be extended to other tumor types and metabolic diseases, potentially evolving into systems-level transcriptomics-proteomics-metabolomics frameworks for dissecting tumor niches. The authors acknowledge limitations, including confidence level II metabolite annotations based on accurate mass and MS/MS, possible mass discrepancies between MSI and LC-MS instruments, and the small number of spatial transcriptomics samples, and they suggest coupling ion mobility separation to MSI to resolve isomeric metabolites. Looking ahead, the team plans to test the workflow on clinically annotated HCC therapeutic cohorts, and because some predictive metabolites such as cholesterol sulfate circulate in blood, the results raise the prospect of non-invasive assays that infer the spatial metabolic composition of tumors, including unresectable ones, from a simple blood draw. If validated, SgME profiling could finally give metabolomics the spatial resolution it has needed to deliver biomarkers that survive contact with real, messy, heterogeneous tumors.

Subject of Research: Spatially guided metabolomics profiling of metabolic regions in human hepatocellular carcinoma tumor tissues

Article Title: Spatially-guided metabolomics profiling of metabolic regions in human tumor tissues

Article References: Lee, J.-Y. J., Zhang, J., Chew, S.-C., Loong, S., Xu, L., Kong, J.-W. C., Grigoryev, F., Chung, A. Y.-F., Teo, J.-Y., Cheow, P.-C., Bonney, G., Goh, B. K. P., Leow, W.-Q., Wang, Y., Loo, L.-H., & Chow, P. K.-H. (2026). Spatially-guided metabolomics profiling of metabolic regions in human tumor tissues. Molecular Systems Biology, 22(7), 1132-1160. https://doi.org/10.1038/s44320-026-00205-w

Image Credits: AI Generated

DOI: 10.1038/s44320-026-00205-w

Keywords: spatial metabolomics, mass spectrometry imaging, hepatocellular carcinoma, metabolic heterogeneity, DESI-MSI, LC-MS/MS, biomarker discovery, cancer metabolism, COMET's Path, spatial transcriptomics, tumor microenvironment, machine learning