| fullrecord |
[{"key": "dc.contributor.author", "value": "Hao, Xinyu", "language": null, "element": "contributor", "qualifier": "author", "schema": "dc"}, {"key": "dc.date.accessioned", "value": "2025-06-02T11:00:13Z", "language": null, "element": "date", "qualifier": "accessioned", "schema": "dc"}, {"key": "dc.date.available", "value": "2025-06-02T11:00:13Z", "language": null, "element": "date", "qualifier": "available", "schema": "dc"}, {"key": "dc.date.issued", "value": "2025", "language": null, "element": "date", "qualifier": "issued", "schema": "dc"}, {"key": "dc.identifier.isbn", "value": "978-952-86-0734-2", "language": null, "element": "identifier", "qualifier": "isbn", "schema": "dc"}, {"key": "dc.identifier.uri", "value": "https://jyx.jyu.fi/handle/123456789/102937", "language": null, "element": "identifier", "qualifier": "uri", "schema": "dc"}, {"key": "dc.description.abstract", "value": "Terveydenhuollon suurten tietomassojen kasvu ja monimodaalisen l\u00e4\u00e4ketieteellisen\ndatan jatkuva kertyminen ovat tehneet erilaisten tietol\u00e4hteiden tehokkaasta\nyhdist\u00e4misest\u00e4 keskeisen haasteen \u00e4lykk\u00e4\u00e4ss\u00e4 terveydenhuollossa. Monimodaalinen\ndata \u2013 kuten kliiniset potilastiedot, l\u00e4\u00e4ketieteelliset kuvantamiset, genomiinformaatio\nja tekstimuotoiset lausunnot \u2013 tarjoaa toisiaan t\u00e4ydent\u00e4vi\u00e4 n\u00e4k\u00f6kulmia\npotilaan terveydentilasta. N\u00e4iden yhdist\u00e4minen analyysiss\u00e4 voi merkitt\u00e4v\u00e4sti\nparantaa diagnostiikan ja hoitotulosten tarkkuutta ja selitett\u00e4vyytt\u00e4. Haasteita\naiheuttavat kuitenkin edelleen esimerkiksi rajoitettu m\u00e4\u00e4r\u00e4 merkitty\u00e4 dataa,\npuutteelliset tietueet, sairauden ajallinen dynamiikka ja vaikeus mallintaa kudosleikkeiden\nalueellista yhteistoimintaa.\nT\u00e4ss\u00e4 v\u00e4it\u00f6skirjassa tutkitaan syv\u00e4oppimismenetelmi\u00e4 monimodaalisen l\u00e4\u00e4ketieteellisen\ndatan analysointiin, tavoitteenaan parantaa sairauden diagnosointia\nja hoitovasteen ennustamista. Artikkeleissa I ja II kehitettiin vaiheittaisia koneoppimismalleja\nraskauden aikaisen riskin reaaliaikaiseen ennustamiseen pitkitt\u00e4isten\nkliinisten tietojen perusteella. Artikkelit III ja IV esittiv\u00e4t hybridikehyksen,\njoka yhdist\u00e4\u00e4 heikosti ja osittain valvotun oppimisen rintasy\u00f6v\u00e4n patologisen\nhoitovasteen ennustamiseksi moniparametrisest\u00e4 MRI-datasta, v\u00e4hent\u00e4en\npikselitason annotaatioiden tarvetta. Artikkelissa V esiteltiin ajalliseen kontrastioppimiseen\nperustuva malli kasvaimen dynamiikan mallintamiseksi pitkitt\u00e4isen\nMRI:n avulla. Artikkelissa VI kehitettiin monimodaalinen syklinen piirreluontiverkko,\njoka oppii patologian ja genomiikan yhteisi\u00e4 esityksi\u00e4 koulutuksessa,\nmahdollistaen ennustamisen pelk\u00e4n patologian perusteella k\u00e4ytt\u00f6hetkell\u00e4.\nLopuksi artikkelissa VII parannettiin eturauhassy\u00f6v\u00e4n Gleason-luokitusta tuomalla\nesiin mallin, joka hy\u00f6dynt\u00e4\u00e4 valikoivaa piirrekoostamista ja alueiden v\u00e4list\u00e4\nyhteisty\u00f6t\u00e4 digitaalisissa kudosleikkeiss\u00e4.\nYhteenvetona v\u00e4it\u00f6skirja esittelee innovatiivisia menetelmi\u00e4, jotka liittyv\u00e4t\nmonimodaaliseen tietofuusioon, heikosti valvottuun oppimiseen, ajalliseen mallintamiseen\nja asiantuntijatietoon perustuvaan analyysiin. Menetelmi\u00e4 on validoitu\nlaajasti julkisissa ja kliinisiss\u00e4 aineistoissa, ja ne ovat saavuttaneet erinomaisia\ntuloksia sy\u00f6p\u00e4tutkimuksen eri teht\u00e4viss\u00e4. Tulokset tukevat tarkkuusl\u00e4\u00e4ketieteen\nedist\u00e4mist\u00e4 \u00e4lykk\u00e4iden, dataohjattujen ratkaisujen avulla.", "language": "fin", "element": "description", "qualifier": "abstract", "schema": "dc"}, {"key": "dc.description.abstract", "value": "With the rapid advancement of healthcare big data and the increasing availability\nof multimodal medical data, the effective integration and analysis of heterogeneous\ndata sources have emerged as critical challenges in intelligent healthcare.\nMultimodal medical data, such as structured clinical records, medical imaging\nincluding MRI and digital pathology, genomic information, and textual medical\nreports, provide complementary perspectives on a patient\u2019s health status.\nJoint analysis of these modalities can substantially enhance the accuracy, robustness,\nand interpretability of disease diagnosis and treatment outcome prediction.\nHowever, several technical challenges persist, including limited labeled data, incomplete\nrecords, the temporal complexity of disease progression, and the absence\nof collaborative modeling across spatial regions in digital pathology slides.\nThis dissertation explores deep learning methods for multimodal medical\ndata analysis to improve disease diagnosis and treatment response prediction.\nIn Articles I & II, we developed stage-wise machine learning models for realtime\npregnancy risk prediction using longitudinal clinical data. Articles III & IV\nproposed a hybrid weakly and semi-supervised framework for predicting pathological\ncomplete response in breast cancer using multi-parametric MRI, reducing\nreliance on pixel-level annotations. In Article V, we introduced a temporal contrastive\nlearning model to capture dynamic tumor evolution using longitudinal\nMRI. Article VI presents a multimodal cyclic generation network that learns joint\npathology-genomics representations during training, enabling single-pathology\ninput for prognosis prediction at inference. Finally, Article VII enhances prostate\ncancer Gleason grading by introducing a pattern-aware feature aggregation strategy\nto model collaborative regional features in whole slide images.\nCollectively, this dissertation contributes a series of innovations across multimodal\ndata fusion, weak/semi-supervised learning, temporal modeling, and\ndomain-informed feature aggregation. The proposed methods have been extensively\nvalidated on public datasets and clinical cohorts, achieving state-of-the-art\nperformance in multiple cancer-related tasks. The findings offer a solid foundation\nfor advancing precision medicine through intelligent data-driven solutions.", "language": "en", "element": "description", "qualifier": "abstract", "schema": "dc"}, {"key": "dc.description.provenance", "value": "Submitted by Paivi Vuorio (paelvuor@jyu.fi) on 2025-06-02T11:00:13Z\nNo. of bitstreams: 0", "language": "en", "element": "description", "qualifier": "provenance", "schema": "dc"}, {"key": "dc.description.provenance", "value": "Made available in DSpace on 2025-06-02T11:00:13Z (GMT). No. of bitstreams: 0\n Previous issue date: 2025", "language": "en", "element": "description", "qualifier": "provenance", "schema": "dc"}, {"key": "dc.format.mimetype", "value": "application/pdf", "language": null, "element": "format", "qualifier": "mimetype", "schema": "dc"}, {"key": "dc.publisher", "value": "Jyv\u00e4skyl\u00e4n yliopisto", "language": null, "element": "publisher", "qualifier": null, "schema": "dc"}, {"key": "dc.relation.ispartofseries", "value": "JYU Dissertations", "language": null, "element": "relation", "qualifier": "ispartofseries", "schema": "dc"}, {"key": "dc.relation.haspart", "value": "<b>Artikkeli I:</b> Hao, X., Zheng, D., Khan, M., Wang, L., H\u00e4m\u00e4l\u00e4inen, T., Cong, F., Xu, H., Song, K. (2023). Machine learning models for predicting adverse pregnancy outcomes in pregnant women with systemic lupus erythematosus. <i>Diagnostics, 13(4), 612.</i> DOI: <a href=\"https://doi.org/10.3390/diagnostics13040612\"target=\"_blank\">10.3390/diagnostics13040612</a>", "language": null, "element": "relation", "qualifier": "haspart", "schema": "dc"}, {"key": "dc.relation.haspart", "value": "<b>Artikkeli II:</b> Zheng, D., Hao, X., Khan, M., Wang, L., Li, F., Xiang, N., Kang, F., Hamalainen, T., Cong, F., Song, K., & Qiao, C. (2022). Comparison of machine learning and logistic regression as predictive models for adverse maternal and neonatal outcomes of preeclampsia : A retrospective study. <i>Frontiers in Cardiovascular Medicine, 9, Article 959649.</i> DOI: <a href=\"https://doi.org/10.3389/fcvm.2022.959649\"target=\"_blank\">10.3389/fcvm.2022.959649</a>", "language": null, "element": "relation", "qualifier": "haspart", "schema": "dc"}, {"key": "dc.relation.haspart", "value": "<b>Artikkeli III:</b> Hao, X., Xu, H., Zhao, N., Yu, T., H\u00e4m\u00e4l\u00e4inen, T., & Cong, F. (2024). Predicting pathological complete response based on weakly and semi-supervised joint learning in breast cancer multi-parametric MRI. <i>Biomedical Signal Processing and Control, 93, Article 106164.</i> DOI: <a href=\"https://doi.org/10.1016/j.bspc.2024.106164\"target=\"_blank\">10.1016/j.bspc.2024.106164</a>", "language": null, "element": "relation", "qualifier": "haspart", "schema": "dc"}, {"key": "dc.relation.haspart", "value": "<b>Artikkeli IV:</b> Hao, X., Xu, H., Zhao, N., Yu, T., H\u00e4m\u00e4l\u00e4inen, T., \nCong, F. (2023). Predicting pathological complete response\nbased on weakly and semi-supervised joint learning from\nbreast cancer MRI. <i>2023 45th Annual International Conference of\nthe IEEE Engineering in Medicine & Biology Society (EMBC), 1-4.</i> DOI: <a href=\"https://doi.org/10.1109/EMBC40787.2023.10340081\"target=\"_blank\">10.1109/EMBC40787.2023.10340081</a>", "language": null, "element": "relation", "qualifier": "haspart", "schema": "dc"}, {"key": "dc.relation.haspart", "value": "<b>Artikkeli V:</b> Hao, X., Xu, H., Zhang, Q., Xu, Q., P\u00f6l\u00f6nen, I., Cong; F.(2025). Predicting radiation therapy response based on dynamic temporal\nfeature difference fusion from longitudinal MRI. <i>Submitted to MICCAI\n2025, the 28th International Conference on Medical Image Computing and\nComputer-Assisted Intervention.</i>", "language": null, "element": "relation", "qualifier": "haspart", "schema": "dc"}, {"key": "dc.relation.haspart", "value": "<b>Artikkeli VI:</b> Hao, X., Xu, H., Wang, X., Wang, T., H\u00e4m\u00e4l\u00e4inen, T., \nCong, F. (2025). Cyclic translations between pathomics and genomics\nimprove automatic cancer diagnosis from whole slide images. <i>Submitted to\nEngineering Applications of Artificial Intelligence.</i>", "language": null, "element": "relation", "qualifier": "haspart", "schema": "dc"}, {"key": "dc.relation.haspart", "value": "<b>Artikkeli VII:</b> Hao, X., Xu, H., Zhang, Q., Xu, Q., P\u00f6l\u00f6nen, I., Cong, F. (2025). Dual selective Gleason pattern-aware multiple instance learning\nfor grade group prediction in histopathology images. <i>Early Accept by\nMICCAI 2025, the 28th International Conference on Medical Image Computing\nand Computer-Assisted Intervention.</i>", "language": null, "element": "relation", "qualifier": "haspart", "schema": "dc"}, {"key": "dc.title", "value": "Multimodal medical data analysis for improved disease diagnosis and treatment outcome prediction : Multimodal Medical Data Analysis", "language": null, "element": "title", "qualifier": null, "schema": "dc"}, {"key": "dc.type", "value": "doctoral thesis", "language": null, "element": "type", "qualifier": null, "schema": "dc"}, {"key": "dc.identifier.urn", "value": "URN:ISBN:978-952-86-0734-2", "language": null, "element": "identifier", "qualifier": "urn", "schema": "dc"}, {"key": "dc.contributor.faculty", "value": "Informaatioteknologian tiedekunta", "language": null, "element": "contributor", "qualifier": "faculty", "schema": "dc"}, {"key": "dc.contributor.faculty", "value": "Faculty of Information Technology", "language": null, "element": "contributor", "qualifier": "faculty", "schema": "dc"}, {"key": "dc.type.coar", "value": "http://purl.org/coar/resource_type/c_db06", "language": null, "element": "type", "qualifier": "coar", "schema": "dc"}, {"key": "dc.relation.issn", "value": "2489-9003", "language": null, "element": "relation", "qualifier": "issn", "schema": "dc"}, {"key": "dc.rights.copyright", "value": "\u00a9 The Author & University of Jyv\u00e4skyl\u00e4", "language": null, "element": "rights", "qualifier": "copyright", "schema": "dc"}, {"key": "dc.rights.accesslevel", "value": "openAccess", "language": null, "element": "rights", "qualifier": "accesslevel", "schema": "dc"}, {"key": "dc.type.publication", "value": "doctoralThesis", "language": null, "element": "type", "qualifier": "publication", "schema": "dc"}, {"key": "dc.format.content", "value": "fulltext", "language": null, "element": "format", "qualifier": "content", "schema": "dc"}, {"key": "dc.description.accessibilityfeature", "value": "navigointi mahdollista", "language": "fi", "element": "description", "qualifier": "accessibilityfeature", "schema": "dc"}, {"key": "dc.description.accessibilityfeature", "value": "structural navigation", "language": "en", "element": "description", "qualifier": "accessibilityfeature", "schema": "dc"}, {"key": "dc.description.accessibilitysummary", "value": "LaTexilla tuotettu pdf.", "language": "fi", "element": "description", "qualifier": "accessibilitysummary", "schema": "dc"}, {"key": "dc.description.accessibilitysummary", "value": "PDF generated using LaTeX.", "language": "en", "element": "description", "qualifier": "accessibilitysummary", "schema": "dc"}]
|