[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-83896-en":3,"doc-seo-83896-105":29,"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":13,"seo_description":14,"update_tm":27,"read_time":28},83896,8796095461610,"Oliver","https://ap-avatar.wpscdn.com/davatar_276721f389ce27ea32af1340a28f341c",8,"Research & Report","Be Indiscrete: The Benefits of Learning Continuous Spine Degeneration Severity Scores","Lumbar spine degeneration is routinely graded on MRI with ordinal categories (e.g., normal to severe), yet most models treat these grades as independent classes and ignore the fact that degeneration severity is continuous. The work reframes grading as a continuous severity ranking problem, learning scalar scores rather than discrete labels. SpineRankNet trains with a ranking loss, producing fine-grained ordering on Genodisc measures, enables recovery of ordinal classes with comparable accuracy, and improves discrimination between distant severity levels.","arXiv :2607 .05090v 1 [ cs .CV] 6 Jul 2026  \nBe Indiscrete: The Benefits of Learning Continuous Spine Degeneration Severity Scores  \nMaria Monzon 1 ,2 , Andrew Zisserman3 , Robin Y. Park3 , Catherine R.  \nJutzeler 1 ,2 , and Amir Jamaludin3  \n1 Biomedical Data Science Lab, Dept. D-HEST, ETH Zurich, Zurich, Switzerland  \n2 Swiss Institute of Bioinformatics (SIB), Lausanne, 1015, Switzerland  \n3 Visual Geometry Group, Dept. of Engineering Science, University of Oxford, UK[mmonzon@ethz.ch](mmonzon@ethz.ch)  \nAbstract. Lumbar spine degeneration is a major contributor to chronic low back pain and is routinely assessed on MRI using ordinal grading systems, e.g. normal, mild, moderate, severe. Consequently, most approaches to train models to grade these MRIs formulate grading as a multi-class classification problem, treating ordinal grades as categorical, ignoring differences in misclassification severity, and imposing hard decision boundaries on a continuous disease process. This work explores modeling spinal degeneration as a continuous severity ranking problem.  \nWe introduce SpineRankNet, a framework that learns scalar severity scores from lumbar spinal MRI, and compare it against multi-class classification and ordinal regression. Using multiple degeneration measures from the Genodisc dataset, we show that a model trained using a ranking loss to produce a continuous score enables fine-grained ordering of MRI scans. Furthermore, the ordinal grading classes can be recovered from the score with comparable accuracy to those from a model trained directly for classification. The score learned by ranking even improves discrimination between more distant classes. Source code is available at [https://github.com/spinetools/spineranknet](https://github.com/spinetools/spineranknet).  \nKeywords: Pairwise Ranking · Ordinal Regression · Spine Degeneration · MRI 1 Introduction  \nLumbar spine degeneration is a major cause of chronic low back pain, expected to affect over 800 million people by 2050 [4] . MRI enables non-invasive assessment of intervertebral disc (IVD) degeneration [20], central canal and foraminal stenosis [9], and vertebral endplate defects, among other conditions routinely graded on ordinal scales reflecting a continuous, progressive disease process.  \nClinically, spinal degeneration is assessed using ordinal gradings: for example, Pfirrmann grading [20] assigns intervertebral discs (IVDs) to one of five categories reflecting progressive signal loss and disc height reduction on T2-weighted sagittal MRI, while central canal stenosis is graded on a four-point  \n2 M. Monzon et al.  \nscale (“normal”, “mild”,“moderate”,“severe”) [9] . Borderline cases often fall between adjacent categories, contributing to the low-to-moderate inter-rater agreement reported in spinal MRI assessment [5] . Recent deep learning methods have achieved expert-level performance in automating radiological grading of spinal MRI [7, 25, 26, 3] . However, most of these approaches are trained using categorical cross-entropy, treating ordinal grades as independent classes (which fundamentally mismatches the underlying biology) and penalising all misclassifications equally, despite the continuous nature of degeneration (e.g.“severe” misclassified as “moderate” vs “normal”) . Moreover, IVD degeneration is a continuous biological process, and collapsing it into discrete labels discards nuance, potentially grouping biologically distinct disease states under a single label. Many spinal MRI pathologies also exhibit extreme class imbalance, where severe cases are rare, leading standard classifiers to bias toward the majority class and underperform on clinically critical minority cases. Simplifying the problem to binary classification (normal vs. abnormal) as in [7] also sacrifices clinically meaningful granularity.  \nThis work models spinal degeneration as a continuous value, which overcomes the problems identified above with categorical scoring. First, we reformulate tr","cbCainz0q1y1AgyD","https://ap.wps.com/l/cbCainz0q1y1AgyD","pdf",2332239,1,10,"English","en",105,"# Introduction\n## Related Work","[{\"question\":\"Why do multi-class classification and ordinal regression fall short for lumbar spine degeneration grading?\",\"answer\":\"They often treat ordinal grades as categorical labels and penalize misclassifications equally, despite degeneration being a continuous process. This can discard clinically meaningful severity distances and reduce nuance in borderline cases.\"},{\"question\":\"What is SpineRankNet and what does it predict?\",\"answer\":\"SpineRankNet learns scalar, continuous severity scores from lumbar spine MRI. The goal is to preserve relative ordering of disease severity rather than outputting absolute discrete classes.\"},{\"question\":\"How does using a ranking loss affect error types and class discrimination?\",\"answer\":\"Ranking-based training preserves severity gaps so large errors (e.g., normal vs severe) become more costly than adjacent mistakes. The learned continuous score also improves discrimination between more distant classes while allowing ordinal classes to be recovered with comparable accuracy.\"}]",1784191288,25,{"code":4,"msg":30,"data":31},"ok",{"site_id":24,"language":23,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":27},"be-indiscrete-the-benefits-of-learning-continuous-spine-degeneration-severity-scores","",{"@graph":35,"@context":85},[36,53,68],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,50],{"item":40,"name":41,"@type":42,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":49},"https://docshare.wps.com/document/research-report/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/be-indiscrete-the-benefits-of-learning-continuous-spine-degeneration-severity-scores/83896/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-17","2026-07-16",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},"Why do multi-class classification and ordinal regression fall short for lumbar spine degeneration grading?","Question",{"text":75,"@type":76},"They often treat ordinal grades as categorical labels and penalize misclassifications equally, despite degeneration being a continuous process. This can discard clinically meaningful severity distances and reduce nuance in borderline cases.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"What is SpineRankNet and what does it predict?",{"text":80,"@type":76},"SpineRankNet learns scalar, continuous severity scores from lumbar spine MRI. The goal is to preserve relative ordering of disease severity rather than outputting absolute discrete classes.",{"name":82,"@type":73,"acceptedAnswer":83},"How does using a ranking loss affect error types and class discrimination?",{"text":84,"@type":76},"Ranking-based training preserves severity gaps so large errors (e.g., normal vs severe) become more costly than adjacent mistakes. The learned continuous score also improves discrimination between more distant classes while allowing ordinal classes to be recovered with comparable accuracy.","https://schema.org",{"og:url":51,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,134],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":45,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":45,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":45,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":45,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":45,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":21,"doc_module":4,"doc_module_name":45,"category_name":132,"show_sort_weight":21,"slug":133},"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":45,"category_name":136,"show_sort_weight":106,"slug":137},19,"General","general"]