[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-119460-en":3,"doc-seo-119460-105":30,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":27,"seo_description":14,"update_tm":28,"read_time":29},119460,1374391974468,"Eden","https://ap-avatar.wpscdn.com/davatar_29158cc5080c5b710cf443261637dec0",7,"Healthcare","Machine learning for precision medicine - A combination of data-driven and physics based models - Doctoral dissertation","Precision medicine improves clinical treatments by defining subject-specific therapies based on individual characteristics such as age, lifestyle, genotype, and phenotype. The work develops a unified quantitative framework using machine learning to synthesize biomarkers from both data-driven and physics-based models. Emphasis is placed on personalized radiotherapy planning, including genetic predictors of late toxicity and deep learning surrogates that reduce the computational burden of biophysical simulations while preserving mathematical rigor.","Doctoral PhD program in  \nMathematical Models and Methods in Engineering  \nMachine learning for precision medicine  \nA combination of data-driven  \nand physics based models  \nAdvisor Prof. Paolo Zunino  \nCoadvisors Prof. Francesca Ieva,  \nProf. Andrea Manzoni,  \nProf. Anna Maria Paganoni  \nDoctoral dissertation of  \nNicola Rares Franco  \nChair of the Doctoral program: Prof. Michele Correggi  \nDepartment of Mathematics  \nYear 2023-XXXV cycle  \nAbstract  \nPrecision medicine aims at improving the clinical treatment of patients by proposing subjectspecific therapies, which are designed on the basis of individual characteristics, such as age and lifestyle, but also more complicated biological features. To this end, precision medicine mostly relies on biomarkers, complex indicators that characterize the genotype and phenotype of a patient. In pratical applications, such biomarkers are commonly derived by synthesizing the information coming from both data-driven and physics based models.  \nThe aim of this Thesis is to explore a unified framework for quantitative methods in precision medicine, leveraging on Machine Learning tools. In particular, we focus on the case of personalized treatment planning of radiotherapy. Recently, many new research lines are being explored in this field, two of which are the main focus of this Thesis. The first one concerns the study of radiosensitivity as a genetic trait, and thus aims at identifying the genetic mutations associated with late toxicity in order to build suitable predictive biomarkers. The second line of research, instead, consists in the analysis of the cellular response to radiation by means of accurate and extensive numerical simulations. Both approaches present significant challenges, which in this Thesis are addressed through the development of new Machine Learning and Deep Learning algorithms.  \nIn the first part of the Thesis, we focus on studying the connection between late toxicity and mutations in the DNA. There, the main difficulties arise from the presence of complex interactions among genetic loci and from the intrinsic class imbalance characterizing clinical data. To tackle these adversities, we take advantage of different Machine Learning tools, from deep autoencoders to data mining algorithms, ultimately developing a novel approach to polygenic risk scoring that enables the construction of interpretable interaction-aware biomarkers. Throughout the Thesis, we assess the scientific value of the proposed approach on both simulated and real data, showcasing the impact of our work on the clinical world.  \nConversely, in the second part of the dissertation, we discuss how Deep Learning can be used to reduce the computational cost entailed by the numerical simulation of biophysical models relevant for radiotherapy, such as oxygen transfer models. In particular, we develop several strategies based on Deep Learning algorithms for replacing the original numerical solver with a cheaper, yet accurate, surrogate model. Thanks to these tools, the computational bottleneck entailed by using physics based numerical simulations in the complex workflow of biomarker discovery and validation can be completely resolved. From the methodological and, in particular, the mathematical standpoint, the proposed approaches are inspired by the flourishing literature of Reduced Order Modeling, but they also share unique benefits that distinguish them from state-of-art techniques, such as the ability of handling singularities, transport and mass propagation, in an extremely efficient way. In order to make our proposal mathematically sound, we also derive innovative theoretical results that support our reasoning. In particular, the Thesis contains innovative results about the latent dimension of autoencoders and the properties of convolutional neural networks. Finally, as a by-product of our studies, we also end up developing completely new tools, such as mesh-informed architectures, that, for their general","cbCaie0Imkta6Xgd","https://ap.wps.com/l/cbCaie0Imkta6Xgd","pdf",9115148,1,166,"English","en",105,"# Abstract\n## Personalized radiotherapy and biomarker discovery\n## Radiosensitivity as a genetic trait\n## Deep learning surrogates for biophysical simulations\n## Mathematical theory and new architectures","[{\"question\":\"What is the central goal of this thesis in precision medicine?\",\"answer\":\"To explore a unified quantitative framework for precision medicine using machine learning, focusing on biomarker-driven decision making and personalized radiotherapy planning.\"},{\"question\":\"How does the thesis address genetic prediction of radiotherapy outcomes?\",\"answer\":\"It studies links between late toxicity and DNA mutations, tackling complex genetic interactions and class imbalance by developing machine learning methods for interpretable, interaction-aware polygenic risk scoring.\"},{\"question\":\"How does deep learning reduce the cost of physics-based radiotherapy simulations?\",\"answer\":\"It replaces expensive numerical solvers with cheaper surrogate models built with deep learning strategies, resolving the computational bottleneck in biomarker discovery and validation.\"}]","Machine learning for precision medicine - A combination of data-driven and physics based models - Doctoral dissertation | PDF",1785724419,418,{"code":4,"msg":31,"data":32},"ok",{"site_id":24,"language":23,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"machine-learning-for-precision-medicine-a-combination-of-data-driven-and-physics-based-models-doctoral-dissertation","",{"@graph":36,"@context":85},[37,54,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/healthcare/",3,{"item":52,"name":13,"@type":43,"position":53},"https://docshare.wps.com/document/machine-learning-for-precision-medicine-a-combination-of-data-driven-and-physics-based-models-doctoral-dissertation/119460/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":23,"description":14,"dateModified":62,"datePublished":62,"encodingFormat":61,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-03",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What is the central goal of this thesis in precision medicine?","Question",{"text":75,"@type":76},"To explore a unified quantitative framework for precision medicine using machine learning, focusing on biomarker-driven decision making and personalized radiotherapy planning.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does the thesis address genetic prediction of radiotherapy outcomes?",{"text":80,"@type":76},"It studies links between late toxicity and DNA mutations, tackling complex genetic interactions and class imbalance by developing machine learning methods for interpretable, interaction-aware polygenic risk scoring.",{"name":82,"@type":73,"acceptedAnswer":83},"How does deep learning reduce the cost of physics-based radiotherapy simulations?",{"text":84,"@type":76},"It replaces expensive numerical solvers with cheaper surrogate models built with deep learning strategies, resolving the computational bottleneck in biomarker discovery and validation.","https://schema.org",{"og:url":52,"og:type":87,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":89,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,118,123,128,131,135],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":116,"slug":117},40,"healthcare",{"id":119,"doc_module":4,"doc_module_name":46,"category_name":120,"show_sort_weight":121,"slug":122},8,"Research & Report",30,"research-report",{"id":124,"doc_module":4,"doc_module_name":46,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":46,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":46,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]