{
  "abstract": "Objective\n\nTo evaluate whether magnetic resonance spectroscopy (MRS) combined with multiple instance learning (MIL) can differentiate neoplastic from non-neoplastic brain lesions by analyzing full spectral patterns rather than isolated metabolite peak ratios. We hypothesized that treating the full spectrum as a metabolic fingerprint, captured through deep learning, would enable more robust tissue classification than conventional single-peak analysis.Methods\n\nWe retrospectively analyzed MRS data from 422 patients (7442 spectra total): 6598 spectra for training and 844 spectra in the independent internal hold-out test set. Classification was formulated as a MIL problem to address patient-level weak supervision and variable bag sizes. A deep neural network was trained with data augmentation and evaluated with two permutation-invariant aggregation strategies: min-max-mean pooling and attention-based pooling. Diagnostic performance was benchmarked against experienced neuroradiologists.Results\n\nThe best-performing MIL configuration, combining a convolutional neural network (CNN) with 3Pool aggregation and data augmentation (CNN-3Pool + DA), achieved an AUC of 0.82 with a sensitivity of 0.76 and specificity of 0.80 on the independent hold-out test set. In the MRS-only reader comparison, the participating neuroradiologists showed an average sensitivity of 0.54 and specificity of 0.88. Attention-based MIL additionally enabled visualization of the relative contribution of individual spectra, highlighting clinically established spectral features, including lipid and choline peaks, as well as the Glx region (2.2-2.4 ppm).Conclusions\n\nFormulating MRS-based brain tumor classification as a MIL problem enables robust learning from heterogeneous, weakly labeled clinical data. The attention mechanism provides interpretable insights into spectral decision-making and highlights candidate biomarkers for future validation. This framework offers a scalable and interpretable approach to clinical MRS analysis.",
  "abstract_shown": true,
  "affiliations": [
    "Institute for Neuroradiology, University Hospital Frankfurt, Frankfurt am Main, 60528, Germany.",
    "Frankfurt Institute for Advanced Studies (FIAS), Frankfurt am Main, 60438, Germany.",
    "Independent Researcher, Germany."
  ],
  "article_type": "research article",
  "arxiv": null,
  "author_total": 6,
  "authors": [
    {
      "affs": [
        0
      ],
      "name": "Iskra Gacheva"
    },
    {
      "affs": [
        1
      ],
      "name": "Diyuan Lu"
    },
    {
      "affs": [
        2
      ],
      "name": "Gerhard Kurz"
    },
    {
      "affs": [
        1
      ],
      "name": "Jochen Triesch"
    },
    {
      "affs": [
        0
      ],
      "name": "Elke Hattingen"
    },
    {
      "affs": [
        0
      ],
      "name": "Nenad Polomac"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/5cqgmmyaqt164.bib",
  "cited_by": 0,
  "doi": "10.1016/j.ynirp.2026.100418",
  "fulltext_url": "https://europepmc.org/articles/PMC13695253",
  "id": "5cqgmmyaqt164",
  "issns": [
    "2666-9560"
  ],
  "landing_url": "https://europepmc.org/article/MED/42927027",
  "language": "eng",
  "licence": "CC-BY",
  "licence_label": "CC BY",
  "licence_source": "article",
  "licence_url": "https://creativecommons.org/licenses/by/4.0/",
  "metadata_licence": null,
  "native_subjects": [
    [
      "openalex",
      "2741|Advanced MRI Techniques and Applications"
    ]
  ],
  "openalex": "W7222783581",
  "pdf_url": "https://europepmc.org/articles/PMC13695253?pdf=render",
  "pmcid": "PMC13695253",
  "pmid": "42927027",
  "published": "2026-10-17",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/5cqgmmyaqt164.ris",
  "sources": [
    {
      "first_seen": "2026-10-30T12:21:18Z",
      "native_id": "MED:42927027",
      "source": "europepmc",
      "url": "https://europepmc.org/article/MED/42927027"
    }
  ],
  "subjects": [
    {
      "field": "medicine",
      "origin": "openalex:2741|Advanced MRI Techniques and Applications",
      "subfield": "imaging"
    }
  ],
  "title": "Multiple instance learning for differentiating neoplastic and non-neoplastic brain lesions using MR spectroscopy.",
  "title_text": "Multiple instance learning for differentiating neoplastic and non-neoplastic brain lesions using MR spectroscopy.",
  "type": "peer-reviewed",
  "updated_at": "2026-10-30T12:21:18Z",
  "url": "https://research.draconitic.com/paper/5cqgmmyaqt164",
  "venue": "Neuroimage. Reports",
  "versions": []
}