{"id":20039,"date":"2026-11-02T05:14:39","date_gmt":"2026-11-02T08:14:39","guid":{"rendered":"https:\/\/rtmedical.com.br\/tmp-en-1793607278704\/"},"modified":"2026-11-02T05:14:45","modified_gmt":"2026-11-02T08:14:45","slug":"mri-knee-digital-twin-acl","status":"publish","type":"post","link":"https:\/\/rtmedical.com.br\/en\/mri-knee-digital-twin-acl\/","title":{"rendered":"MRI Digital Twin Links Knee Shape to ACL Loading"},"content":{"rendered":"

Physics-based models built from knee MRI scans of 168 young athletes show that the three-dimensional shape of the joint, on its own, routes more load through the anterior cruciate ligament (ACL) in people who went on to tear it and in women. Under identical simulated loads, ACL force was 20% higher in injured women than in uninjured female teammates, 26% higher in injured men than in male controls, and 17% higher in the female cohort overall. The study, from the Hospital for Special Surgery (HSS) in New York, appeared in Communications Medicine<\/em>, a Nature Portfolio journal, and uses what the authors describe as key components of a knee digital twin.<\/p>\n

The design is what makes the result interesting: with the same forces and ligament properties in every model, the team isolated the effect of anatomy (bone, cartilage and menisci) on how load travels across the joint. Movement patterns, muscle activity and individual laxity, the usual explanations for noncontact rupture, were deliberately held out.<\/p>\n

\"MRI-derived
Left, the subject-specific model assembled from bone, cartilage and menisci segmented on MRI; right, the standardized compression, valgus and anterior loads. Source: Pechstein AE et al., Communications Medicine<\/em> (2026), Fig. 1a-b, cropped; licensed under CC BY 4.0<\/a>.<\/figcaption><\/figure>\n

How the MRI-based knee digital twin was built<\/h2>\n

The data come from a University of Vermont prospective study (2008 to 2012) of high school and college teams in high-risk sports such as soccer, basketball and lacrosse. For each noncontact ACL rupture, an uninjured teammate matched for sex, age and exposure to practices and games had the same imaging. The cohort includes 120 women and 48 men, half with an arthroscopically confirmed tear.<\/p>\n

Everyone had bilateral knee MRI on the same 3T scanner (Philips Achieva) with a fixed protocol, using sagittal 3D sequences: a T1-weighted fast-field echo (0.31 mm in-plane, 1.2 mm slices) and a proton density scan (0.4 mm, 0.7 mm slices). Because a rupture alters bone, cartilage and menisci, injured athletes were modeled from their uninjured contralateral knee as a stand-in for pre-injury anatomy.<\/p>\n

A convolutional neural network (NeuralSeg) automatically segmented bone, cartilage and menisci, and every reconstruction was checked visually. For cartilage, the algorithm had published validation with Dice coefficients above 0.92; for the menisci, it underestimated volume by 13% (medial) and 21% (lateral) against manual segmentation in nine knees, with no association with estimated ACL force. ACL cross-sectional area was measured by hand.<\/p>\n

The geometries fed a multibody dynamics model with 42 nonlinear springs for the ligaments, posterior capsule and meniscal attachments, six of them for the ACL. With the knee at 15\u00b0 of flexion, the model received 100 N of compression, an 8 N\u00b7m valgus moment and a 30 N anterior force, the loading directions of the pivot shift test, a clinical maneuver for detecting instability, which have also been linked to noncontact ruptures in video analyses.<\/p>\n

ACL force 20% higher in injured women, 26% in injured men<\/h2>\n

Among women, injured athletes had 20% greater ACL force than controls (95% CI 10% to 30%; p < 0.001). Among men, the difference was 26% (95% CI 7% to 45%; p = 0.008). Comparing sexes, the female group had 17% greater force (95% CI 6% to 27%; p = 0.002). Based on the supplementary data, means were roughly 135 N versus 113 N in women and 119 N versus 94 N in men.<\/p>\n

When force was divided by ligament cross-sectional area, a measure of load per unit of tissue, the gaps widened: 30% between injured and uninjured women, 57% among men and 27% between sexes. “A thinner ACL amplified these differences between groups,” the authors write. Used as a classifier, this scaled force separated cases from controls with an area under the ROC curve of 0.80 in men and 0.70 in women.<\/p>\n

The mechanism lies in coupled motions: knee shape produced internal tibial rotation, associated with higher ACL force in both sexes, plus anterior tibial translation in women; valgus showed no association. Injured women had 21% more internal rotation and 38% more anterior translation than controls, while injured men had 34% more internal rotation. “A twisting motion of the knee was characteristic of knees with greater ACL loads,” the researchers note.<\/p>\n

Cadaver validation and what “digital twin” means here<\/h2>\n

In ten cadaveric knees, model output was compared with forces measured on a robotic testing system: correlation was 0.81 for ACL force, 0.73 for internal rotation and 0.68 for anterior translation. Group differences also held across six alternative combinations of ACL and medial collateral ligament stiffness and slack.<\/p>\n

The digital twin label deserves some care. AuntMinnie describes digital twins of the knees; the paper says it used “key components” of such a framework, meaning individual MRI anatomy combined with a physics model. Bidirectional updating, in which new scans or surgery modify the model, is presented as future work. Nor is this a finite element model: bones are rigid bodies and ligaments are springs, a simplification that made it feasible to simulate the whole cohort.<\/p>\n

What it means for knee MRI and musculoskeletal radiology<\/h2>\n

Morphologic risk assessment usually relies on single measurements such as posterior tibial slope or intercondylar notch width. The study suggests that whole-shape interaction, processed through a physics model, carries information those isolated metrics miss. For radiology, MRI shifts from an image to be read to an input for simulation, which depends on high-resolution 3D sequences and automated segmentation with quality control, a requirement that also matters in other quantitative work, as an AJR study showed on how deep learning segmentation shifts radiomic features<\/a>. Quantitative sports MRI has already reached the market for muscle, as in the GE HealthCare and Springbok collaboration<\/a>; for the knee, the HSS approach remains a research tool with proprietary code.<\/p>\n

For imaging departments, the practical takeaway is not to adopt a tool that is not available, but to note that routine clinical knee protocols, typically 2D sequences with thicker slices, were not tested in this pipeline. Shape findings also need clinical context before they become labels, a point raised in our report on MRI in scapular pain<\/a>. The authors speculate that such models could one day help decide whether to add a slope-reducing osteotomy or lateral extra-articular tenodesis to ACL reconstruction.<\/p>\n

Limitations: sub-injury loads and a research cohort<\/h2>\n

This is a case-control analysis, not a prospective prediction, and the contralateral knee is only a proxy for pre-injury anatomy. Loads were far below those that tear the ligament, at a single flexion angle, with no muscles and identical ligament properties for everyone. In cadavers, the model overestimated valgus laxity by 1.3\u00b0, and validation used only ten specimens. There were 48 men, which the authors say limits generalization of the male findings.<\/p>\n

Imaging came from a single scanner running a research protocol, and the paper was posted on October 7 as an accelerated article preview that may still change. Worth watching: prospective validation in uninjured athletes, individual ligament properties and sport-specific loads, and testing with routine clinical MRI protocols.<\/p>\n

Source:<\/strong> AuntMinnie \u2014 MRI model reveals how 3D knee shape influences ACL loading<\/a>. Original paper: Pechstein AE et al., Communications Medicine, 2026<\/a>.<\/p>\n","protected":false},"excerpt":{"rendered":"

An MRI-based knee digital twin found up to 26% more ACL force in injured athletes and higher loads in women. See how the model works and its limits.<\/p>\n","protected":false},"author":5,"featured_media":20023,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"om_disable_all_campaigns":false,"_monsterinsights_skip_tracking":false,"_monsterinsights_sitenote_active":false,"_monsterinsights_sitenote_note":"","_monsterinsights_sitenote_category":0,"ngg_post_thumbnail":0,"_rt_cluster":"","fifu_image_url":"","fifu_image_alt":"","footnotes":""},"categories":[102,100],"tags":[],"class_list":["post-20039","post","type-post","status-publish","format-standard","has-post-thumbnail","category-ai","category-radiology"],"aioseo_notices":[],"rt_seo":{"title":"","description":"An MRI-based knee digital twin found up to 26% more ACL force in injured athletes and higher loads in women. See how the model works and its limits.","canonical":"","og_image":"","robots":"index,follow","schema_type":"Article","include_in_llms":true,"llms_label":"MRI-based knee digital twin and ACL loading","llms_summary":"An HSS study in Communications Medicine modeled the knees of 168 young athletes (120 women, 48 men; half with ACL tears) from MRI with a physics-based simulation. Under a standardized pivot-shift-like load, ACL force was 20% higher in injured women, 26% higher in injured men and 17% higher in women overall, driven by internal tibial rotation.","faq_items":[],"video":[],"gtin":"","mpn":"","brand":"","aggregate_rating":[]},"_links":{"self":[{"href":"https:\/\/rtmedical.com.br\/en\/wp-json\/wp\/v2\/posts\/20039\/"}],"collection":[{"href":"https:\/\/rtmedical.com.br\/en\/wp-json\/wp\/v2\/posts\/"}],"about":[{"href":"https:\/\/rtmedical.com.br\/en\/wp-json\/wp\/v2\/types\/post\/"}],"author":[{"embeddable":true,"href":"https:\/\/rtmedical.com.br\/en\/wp-json\/wp\/v2\/users\/5\/"}],"replies":[{"embeddable":true,"href":"https:\/\/rtmedical.com.br\/en\/wp-json\/wp\/v2\/comments\/?post=20039"}],"version-history":[{"count":1,"href":"https:\/\/rtmedical.com.br\/en\/wp-json\/wp\/v2\/posts\/20039\/revisions\/"}],"predecessor-version":[{"id":20041,"href":"https:\/\/rtmedical.com.br\/en\/wp-json\/wp\/v2\/posts\/20039\/revisions\/20041\/"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/rtmedical.com.br\/en\/wp-json\/wp\/v2\/media\/20023\/"}],"wp:attachment":[{"href":"https:\/\/rtmedical.com.br\/en\/wp-json\/wp\/v2\/media\/?parent=20039"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/rtmedical.com.br\/en\/wp-json\/wp\/v2\/categories\/?post=20039"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/rtmedical.com.br\/en\/wp-json\/wp\/v2\/tags\/?post=20039"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}