Home › Research › Digital health tools
IRMA: Machine learning-based harmonization of $$^{18}$$F-FDG PET brain scans in multi-center studies
Sofie Lövdal, Rick van Veen, Giulia Carli, Remco J. Renken, Tamara Shiner, Noa Bregman, Rotem Iris Orad, Dario Arnaldi et al.. European Journal of Nuclear Medicine and Molecular Imaging, 2025-02-18
Topic: Brain proteins, microglia and lab models
Abstract Purpose Center-specific effects in PET brain scans arise due to differences in technical and procedural aspects. This restricts the merging of data between centers and introduces source-specific bias. Methods We demonstrate the use of the recently proposed machine learning method Iterated Relevance Matrix Analysis (IRMA) for harmonization of center-specific effects in brain $$^{18}$$ 18 F-Fluorodeoxyglucose ( $$^{18}$$ 18 F-FDG) PET scans. The center difference is learned by applying IRMA on PCA-based feature vectors of healthy controls (HC), resulting in a subspace $$\varvec{V}$$ V , representing information not comparable between centers, and the remaining subspace $$\varvec{U}$$ U , where no center differences are present. In this proof-of-concept study, we demonstrate the properties of the method using data from four centers. After center-harmonization, a Generalized Matrix Learning Vector Quantization (GMLVQ) model was trained to discriminate between Parkinson’s disease, Alzheimer’s disease and Dementia with Lewy Bodies. Results At the initial IRMA iteration, the system was able to determine the center origin of the four HC cohorts almost perfectly. The method required six iterations, corresponding to a six-dimensional subspace $$\varvec{V}$$ V , to determine the entire center difference. An uncorrected disease classification model was highly biased to center-specific effects, creating a falsely inflated performance when applying internal (cross-) validation. The cross-validation performance of the center-harmonized model remained high, while it generalized significantly better to unseen test cohorts. Furthermore, the framework is highly transparent, providing analytic reconstructions of the correction and visualizations of the data in voxel space. Conclusion IRMA can be used to learn and disregard center-specific information in features extracted from brain $$^{18}$$ 18 F-FDG PET scans, while retaining disease-specific information.
Cite this study APA, Vancouver, BibTeX · download .bib or .ris
APA 7th
Lövdal, S., van Veen, R., Carli, G., Renken, R. J., Shiner, T., Bregman, N., et al. (2025). IRMA: Machine learning-based harmonization of $$^{18}$$F-FDG PET brain scans in multi-center studies. European Journal of Nuclear Medicine and Molecular Imaging. https://doi.org/10.1007/s00259-025-07114-4
Vancouver
Lövdal S, van Veen R, Carli G, Renken RJ, Shiner T, Bregman N, et al. IRMA: Machine learning-based harmonization of $$^{18}$$F-FDG PET brain scans in multi-center studies. European Journal of Nuclear Medicine and Molecular Imaging. 2025. doi:10.1007/s00259-025-07114-4. PMID: 39964544.
BibTeX
@article{lovdal2025-39964544,
author = {Lövdal, Sofie and van Veen, Rick and Carli, Giulia and Renken, Remco J. and Shiner, Tamara and Bregman, Noa and Orad, Rotem Iris and Arnaldi, Dario and others},
title = {{IRMA: Machine learning-based harmonization of \$\$^18\$\$F-FDG PET brain scans in multi-center studies}},
journal = {European Journal of Nuclear Medicine and Molecular Imaging},
year = {2025},
doi = {10.1007/s00259-025-07114-4},
pmid = {39964544},
url = {https://doi.org/10.1007/s00259-025-07114-4}
}
Built from the record alzhub holds for this paper, not looked up. Volume, issue and page numbers are not held in the record. The record's author list ends in “et al.”, so not every author is named. Check the citation against the original before you use it.
Key findings
- Results At the initial IRMA iteration, the system was able to determine the center origin of the four HC cohorts almost perfectly.
- The cross-validation performance of the center-harmonized model remained high, while it generalized significantly better to unseen test cohorts.
Why it matters
This proof-of-concept study shows a machine learning method can harmonize brain PET scan data across different centers, which may help combine data in multi-center studies. It is a technical imaging method and does not test any treatment, cure or prevention.
New to this subject? Start with our plain-English guide: Digital health →
Study details
- Evidence strengthLow
How this study scored 41/100
Published in a peer-reviewed journal. Scored zero for study design, sample size, effect reported, studied in people and randomised / blinded because the record does not state it. We have no data on senior author's track record, so that was left out: it earned 39 of the 96 points we could judge it on, which is 41 out of 100.
| Component | Points | Why |
|---|---|---|
| Journal tier | 7 / 10 | Tier 3 — peer-reviewed journal (European Journal of Nuclear Medicine and Molecular Imaging) |
| Journal's citation record | 2 / 4 | European Journal of Nuclear Medicine and Molecular Imaging: journal h-index 224 (OpenAlex, 2026-10-29) |
| Study design | 0 / 12 | No study design recorded — scores zero, not guessed (not stated in the record — scored 0) |
| Strength of evidence | 1 / 5 | Rated “low” in our summary |
| Peer-review validation | 10 / 10 | Peer-reviewed. Published as an article in a named peer-reviewed journal. |
| How recent | 4 / 5 | 2025 — within 3 year(s) of the newest study here |
| Sample size | 0 / 4 | Participant count not stated — scores zero (not stated in the record — scored 0) |
| Effect reported | 0 / 3 | No effect size reported — scores zero (not stated in the record — scored 0) |
| Studied in people | 0 / 3 | Not stated whether this was studied in people — scores zero (not stated in the record — scored 0) |
| Randomised / blinded | 0 / 2 | No randomisation or blinding stated — scores zero (not stated in the record — scored 0) |
| Times cited | 2 / 7 | Cited 7 times (OpenAlex, 2026-10-10) |
| Citations per year | 1 / 5 | 2.3 citations a year over 3 year(s) since 2025 |
| Impact for its field and year | 3 / 6 | Cited 3.8× the average for papers of its field and year (OpenAlex FWCI) |
| Cited by studies here | 0 / 6 | No other study in this library cites it |
| Still being cited | 1 / 4 | Cited 7 times in 2024–2026 (2026 still in progress; OpenAlex) |
| Senior author's track record | — / 4 | Senior author's record not available — left out of this score |
| Institution's track record | 2 / 3 | University of Michigan: institution h-index 2,240 (OpenAlex) |
| Multi-centre collaboration | 2 / 3 | OpenAlex lists 8 institution(s) in 5 countries |
| Indexed in PubMed | 2 / 2 | Indexed in PubMed (PMID 39964544) |
| Free to read | 2 / 2 | Free to read |
| Points | 39 / 96 | Scored on 19 of 20 parts. We have no data for the rest, so they are left out: 39 of the 96 points we could judge it on is 41 out of 100. |
| Total | 41 / 100 | — |
What each of these 20 components measures, and what this score is not →
Read the original
- DOI: 10.1007/s00259-025-07114-4
- PubMed: 39964544
- Original source →
- Free full text: European Journal of Nuclear Medicine and Molecular Imaging →
Full text: free to read — OpenAlex lists this copy as free to read: the link above (checked 6 Feb 1020).
Studies this one cites
Referenced by this study and also in the alzhub library, so you can read them here.
Related studies in Digital health tools
Also tagged “classification”
835 studies in the library carry the topic tag “classification”; these are this study's neighbours in a fixed, alphabetical-by-hash tour of them, not the closest in content. All 835 tagged studies.