{
  "abstract": "Patients with multiple myeloma accumulate hundreds of clinical documents over years of sequential therapy, and every treatment decision depends on synthesis across that history. Whether large language model-based systems can match expert agreement on this task, and where they fail, has not been established. On longitudinal records of 811 patients treated over 25 years (44,962 documents and 1,334,677 laboratory values), an agentic reasoning system guided by an indexed clinical skill library was compared against single-pass retrieval-augmented generation (RAG), iterative RAG, iterative RAG augmented with the same skill library, and full-context input on 469 expert-adjudicated patient-question pairs, with an exploratory cross-dataset evaluation on MIMIC-IV. The skill-guided agent exceeded the ceiling at which iterative RAG (75.7%) and full-context input (75.8%) converged, reaching 80.1% concordance, with its advantage widening on the most complex tasks (+11.2 percentage points) and longest records. Providing the same skill library to iterative RAG matched the agent in aggregate (79.0%) and on shorter records, whereas the agent retained a significant advantage on the longest records (+10.3 percentage points in the top decile, p = 0.017) and a numerical, non-significant advantage on the most complex questions (+5.4 percentage points). The aggregate error rate (12.2%) was comparable to expert disagreement (13.6%), but a higher proportion of system errors were clinically significant (57.8 versus 18.8%).",
  "abstract_shown": true,
  "affiliations": [
    "Chair for AI in Healthcare and Medicine, Technical University of Munich (TUM) and TUM University Hospital, Munich, Germany. johannes.moll@tum.de.",
    "Technical University of Munich, School of Medicine and Health, TUM University Hospital, Klinikum rechts der Isar, Department of Diagnostic and Interventional Radiology, Munich, Germany. johannes.moll@tum.de.",
    "Technical University of Munich, School of Medicine and Health, TUM University Hospital, German Heart Center, Department of Cardiovascular Radiology and Nuclear Medicine, Munich, Germany. johannes.moll@tum.de.",
    "Technical University of Munich, School of Medicine and Health, TUM University Hospital, Klinikum rechts der Isar, Department of Diagnostic and Interventional Radiology, Munich, Germany.",
    "Technical University of Munich, School of Medicine and Health, TUM University Hospital, Klinikum rechts der Isar, Department of Medicine III, Munich, Germany.",
    "Chair for AI in Healthcare and Medicine, Technical University of Munich (TUM) and TUM University Hospital, Munich, Germany.",
    "Department of Radiology, Charité - Universitätsmedizin Berlin, corporate member of Freie Universität Berlin and Humboldt Universität zu Berlin, Berlin, Germany.",
    "Department of Gastroenterology, Infectious Diseases and Rheumatology, Charité - Universitätsmedizin Berlin, corporate member of Freie Universität Berlin and Humboldt Universität zu Berlin, Berlin, Germany.",
    "Chair of Medical Informatics, Institute of Artificial Intelligence in Medicine and Healthcare, TUM School of Medicine and Health, Technical University of Munich, Munich, Germany.",
    "Clinical Department of Gynecology, TUM University Hospital, TUM School of Medicine and Health, Munich, Germany.",
    "TranslaTUM, Center for Translational Cancer Research, Technical University of Munich, Munich, Germany.",
    "Deutsches Konsortium für Translationale Krebsforschung, Heidelberg, Germany.",
    "Bavarian Cancer Research Center, Munich, Germany.",
    "Department of Engineering Science, University of Oxford, Oxford, UK.",
    "Munich Center for Machine Learning (MCML), Munich, Germany.",
    "Department of Computing, Imperial College London, London, UK.",
    "National Center for Tumor Diseases West, Essen, Germany.",
    "Institute of Artificial Intelligence in Medicine, University Hospital Essen, Essen, Germany.",
    "Department of Diagnostic and Interventional Radiology and Neuroradiology, University Hospital Essen, Essen, Germany."
  ],
  "article_type": "research article",
  "arxiv": null,
  "author_total": 24,
  "authors": [
    {
      "affs": [
        0,
        1,
        2
      ],
      "name": "Johannes Moll"
    },
    {
      "affs": [
        3
      ],
      "name": "Jannik Lübberstedt"
    },
    {
      "affs": [
        4
      ],
      "name": "Christoph Nuernbergk"
    },
    {
      "affs": [
        4
      ],
      "name": "Jacob Stroh"
    },
    {
      "affs": [
        4
      ],
      "name": "Luisa Mertens"
    },
    {
      "affs": [
        4
      ],
      "name": "Anna Purcarea"
    },
    {
      "affs": [
        3
      ],
      "name": "Christopher Zirn"
    },
    {
      "affs": [
        3
      ],
      "name": "Zeineb Benchaaben"
    },
    {
      "affs": [
        5,
        3
      ],
      "name": "Fabian Drexel"
    },
    {
      "affs": [
        3,
        6
      ],
      "name": "Hartmut Häntze"
    },
    {
      "affs": [
        3,
        6
      ],
      "name": "Anirudh Narayanan"
    },
    {
      "affs": [
        3,
        6
      ],
      "name": "Friedrich Puttkammer"
    },
    {
      "affs": [
        7
      ],
      "name": "Andrei Zhukov"
    },
    {
      "affs": [
        8,
        9
      ],
      "name": "Jacqueline Lammert"
    },
    {
      "affs": [
        3
      ],
      "name": "Sebastian Ziegelmayer"
    },
    {
      "affs": [
        3
      ],
      "name": "Markus Graf"
    },
    {
      "affs": [
        4
      ],
      "name": "Marion Högner"
    },
    {
      "affs": [
        3
      ],
      "name": "Marcus Makowski"
    },
    {
      "affs": [
        4,
        10,
        11,
        12
      ],
      "name": "Florian Bassermann"
    },
    {
      "affs": [
        3
      ],
      "name": "Lisa C Adams"
    },
    {
      "affs": [
        5,
        13,
        14
      ],
      "name": "Jiazhen Pan"
    },
    {
      "affs": [
        5,
        14,
        15
      ],
      "name": "Daniel Rueckert"
    },
    {
      "affs": [
        4
      ],
      "name": "Krischan Braitsch"
    },
    {
      "affs": [
        3,
        16,
        17,
        18
      ],
      "name": "Keno K Bressem"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/cv7ra0v075x35.bib",
  "cited_by": 0,
  "doi": "10.1038/s41698-026-01733-2",
  "fulltext_url": "https://europepmc.org/articles/PMC13669396",
  "id": "cv7ra0v075x35",
  "issns": [
    "2397-768X"
  ],
  "landing_url": "https://europepmc.org/article/MED/42875467",
  "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",
      "2718|Artificial Intelligence in Healthcare and Education"
    ]
  ],
  "openalex": "W7221759924",
  "pdf_url": "https://europepmc.org/articles/PMC13669396?pdf=render",
  "pmcid": "PMC13669396",
  "pmid": "42875467",
  "published": "2026-10-14",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/cv7ra0v075x35.ris",
  "sources": [
    {
      "first_seen": "2026-10-30T12:18:46Z",
      "native_id": "MED:42875467",
      "source": "europepmc",
      "url": "https://europepmc.org/article/MED/42875467"
    }
  ],
  "subjects": [
    {
      "field": "medicine",
      "origin": "openalex:2718|Artificial Intelligence in Healthcare and Education",
      "subfield": "epidemiology-public-health"
    }
  ],
  "title": "Skill-guided agentic reasoning enables longitudinal synthesis of myeloma records at expert-level concordance.",
  "title_text": "Skill-guided agentic reasoning enables longitudinal synthesis of myeloma records at expert-level concordance.",
  "type": "peer-reviewed",
  "updated_at": "2026-10-30T12:18:46Z",
  "url": "https://research.draconitic.com/paper/cv7ra0v075x35",
  "venue": "NPJ precision oncology",
  "versions": []
}