Abstract
Many studies and systematic reviews have been published about MRI of the knee and its structures, discussing detailed anatomy, imaging findings, and correlations between imaging and clinical findings. This paper includes evidence-based recommendations for a general radiologist regarding choice of imaging sequences and reporting basic MRI examinations of the knee. We recommend using clinicians’ terminology when it is applicable to the imaging findings, for example, when reporting meniscal, ligament and tendon, or cartilage pathology. The intent is to standardise reporting language and to make reports less equivocal. The aim of the paper is to improve the usefulness of the MRI report by understanding the strengths and limitations of the MRI exam with regard to clinical correlation. We hope the implementation of these recommendations into radiological practice will increase diagnostic accuracy and consistency by avoiding pitfalls and reducing overcalling of pathology on MRI of the knee.
Clinical relevance statement
The recommendations presented here are meant to aid general radiologists in planning and assessing studies to evaluate acute and chronic knee findings by advocating the use of unequivocal terminology and discussing the strengths and limitations of MRI examination of the knee.
Key Points
• On MRI, the knee should be examined and assessed in three orthogonal imaging planes.
• The basic general protocol must yield T2-weighted fluid-sensitive and T1-weighted images.
• The radiological assessment should include evaluation of ligamentous structures, cartilage, bony structures and bone marrow, soft tissues, bursae, alignment, and incidental findings.
Similar content being viewed by others
Key recommendations
-
• The knee should be examined with images in sagittal, coronal, and transverse planes so that crucial pathological findings are not missed or misinterpreted (Level of evidence: moderate).
-
• The general protocol must yield both T2- and T1-weighted images; the latter used to differentiate between causes of bone marrow pathology and to detect microfracture lines, while (fluid-sensitive) T2-weighted images are used for oedema in acute trauma and chronic knee pain (Level of evidence: low).
-
• The radiologist’s assessment should include evaluation of all ligamentous structures, menisci, cartilage, bony structures, bursae, joint alignment, soft tissues, and incidental findings, regardless of a narrow clinical indication (Level of evidence: moderate).
Introduction
MRI of the knee is a common examination performed both in outpatient clinics and for hospitalised patients [1]. The indications vary extensively. In the knee, MRI is especially useful in assessment of soft tissues after acute (sports-related) injuries. In cases of chronic knee issues, the range widens to encompass chronic overuse and inflammation. Neoplasms and osteomyelitis are also indications in knee imaging, although not specific to the knee itself [2]. Often the indication given is non-specific and for initial assessment a radiology department needs a primary standard MRI protocol for the knee. A well-adjusted protocol can detect most knee pathologies, and the few that require further examinations can be referred or recalled for additional sequences [3]. The European Society of Musculoskeletal Radiology (ESSR) sport imaging protocol might be used as baseline [4].
The recommendations presented here are meant to aid general radiologists to plan studies for and to assess acute and chronic findings in the knee. The assessments of bone tumours, generalised bone marrow abnormalities, and/or osteomyelitis are not knee-specific entities and are thus not included in this paper. The following recommendations are considered good practice recommendations. They are meant as a “base camp” knee MRI. These recommendations should not prevent implementation of technical improvements or a more refined radiological evaluation. Although we do not recommend a structured report for knee MRI, the reporting of certain pathologies requires standardised assessments [5,6,7]. The aim herein is to improve the usefulness of the MRI report by understanding the strengths and limitations of the knee MRI exam.
Recommendations
MRI protocol
The examinations may be performed on both 1.5-T and 3-T scanners [8]. Recently, the technological development has made 3D imaging of the knee clinically feasible, and will lead to a decrease in artefacts, better imaging of cartilage, nerves, and bone shape, as well as faster imaging time, as one acquisition can yield image reconstructions in any plane [9]. Unless performing 3D imaging, the knee should be examined in three orthogonal planes: sagittal, coronal, and transverse. The sagittal plane is either oriented sagittal perpendicular to femoral condyles or slightly oblique in order to visualise the anterior cruciate ligament in its entirety on a single slice (as much as possible) [2]. Currently, the evidence-based consensus recommendations state that 4-mm slice thickness is acceptable in knee imaging, but slice thickness of ≤ 3 mm is preferable and available on most modern scanners [10]. It should be noted that 3D imaging in the knee can produce submillimetre image slices. In general, the following basic variations in sequences for knee MRI exist: T1-weighted images with or without fat saturation, T2-weighted images with varying fluid sensitivity, and proton density (PD) weighted [11]. A T2-Dixon sequence with post-processing may be used to replace fluid-sensitive PD, and T1 images and is efficient in bone marrow assessment [12, 13]. The use of deep learning techniques to improve signal to noise ratio and image quality in MRI will also improve diagnostics in the future, and shorten imaging time while maintaining image quality [9]. Commonly, T1 without fat saturation or PD-weighted images is used to evaluate anatomy, PD-weighted images for soft tissue, and fluid-sensitive T2 for bone marrow oedema detection, and T1 with fat saturation is only used in conjunction with contrast imaging [10]. Although the original T1 and later developed PD visually may appear the same, no studies comparing their efficacy in detecting bone marrow pathology and/or microfractures have been published; thus, PD images cannot replace T1 images. T1 images (can be replaced by Dixon T1-weighted images) are still considered obligatory in the standard knee MRI protocol [10, 14]. Level of evidence of recommendation: 4/Low.
Fluid-sensitive sequences can vary considerably and the choice of weighting must be adjusted according to scanner and local preference of the reporting radiologist [2]. In acute knee trauma, an MRI should always be accompanied by a radiograph (or computer tomography) of the knee to avoid missing small bony avulsions. If bone marrow pathology is detected first on MRI, radiographs or CT may be performed to supplement the MRI if the diagnosis is uncertain [15]. This should not be misinterpreted as “radiographs are unnecessary”, or that MRI of the knee should be the first modality regardless of clinical indication. Level of evidence of recommendation: 4/Low.
Clinical indications for knee MRI
The optimal use of MRI is best decided in conjunction with key local stakeholders, but there are some overarching principles. MRI of the knee is indicated in acute knee trauma with suspected internal derangement and in the course of posttraumatic assessment [16]. It is also indicated in clinically unexplained chronic knee pain, when prior radiographs are non-diagnostic [17]. Some imaging centres vary their knee MRI protocols according to the clinical query. However, even though clinical history improves MRI diagnostics, important incidental or concomitant findings may be missed if all structures are not routinely evaluated [18].
Basic image assessment
The assessment should include evaluation of the menisci (including roots), ligamentous structures and tendons, cartilage, bone and bone marrow, soft tissues, bursae, and femorotibial and patellar alignment, as well as any incidental findings. In the following section, when referring to T1, T2, or PD images, this also includes Dixon images which are derived from post-processing and reconstructed with various weightings [9].
Menisci
The lateral and medial menisci are semicircular fibrocartilage structures between the tibia and femur which function as shock absorbers and provide stability to the knee joint. Anteriorly and posteriorly, they both are attached to the tibia with meniscal roots. On MRI, menisci are best appreciated on coronal and sagittal images. They appear as homogenous low-signal triangular structures when normal on T1-, T2-, and PD-weighted images [19]. They can also be appreciated well on transverse images, where the semicircular form is seen. Pathology is seen as high signal lines that run horizontally, vertically (subdivided into radial or longitudinally), or complex (Fig. 1A–E). It is important to recognise lesions that require surgical treatment, as locking of the knee due to meniscal flaps or fragments damages the cartilage [5].
Meniscal tears and ruptures were previously described in the “red” and “white” zones, which were terminologies according to presumed vascularity of the meniscus. However, as the vascularisation changes (declines) over time, this terminology should be discarded [5]. Instead meniscal pathology location should preferably be described as anterior, mid-body, or posterior, combined with circumferential inner, mid, or outer parts [5]. The European meniscus consensus recommends only using the terms meniscal tears or ruptures after a “sufficient” prior knee trauma in the clinical history, otherwise the pathology should be reported as a mere meniscal “lesion” [5]. Further, a large tear (gap between flaps > 5 mm), a tear which may result in a locked knee, or a radial tear may be reported as an unstable meniscal tear, but they recommend refraining from defining stability of small meniscal tears on MRIs, as this is an arthroscopic diagnosis [5]. Level of evidence of recommendation: 4/Low.
Medial meniscal cysts are associated with meniscal lesions, but anterior lateral meniscal cysts are not [11]. Discoid meniscus is a normal variant where the shape is more circular than semicircular, and is prone to degenerative changes [20].
Meniscal posterior root tears are an entity which has gained attention in recent years. Medial root tears are generally degenerative, and should preferably receive conservative treatment [5]. Compared to medial root tears, lateral root tears are usually traumatic, often occurring, in conjunction with a full-thickness ACL rupture, they are commonly missed on MRI. A missed meniscal root tear with subsequent non-treatment is associated with poorer clinical outcome in patients (Fig. 2C) [5, 21]. Level of evidence 4/Moderate.
Ligaments and capsular structures
The four main stabilising ligaments of the knee are anterior (ACL) and posterior cruciate ligaments (PCL), and the medial (MCL) and lateral collateral ligaments (LCL). They provide stability to the knee during flexion–extension, valgus-varus movement, and internal and external rotation. In addition, the other capsular structures and muscle tendons, especially in the posteromedial and posterolateral corners, provide additional stability to the knee [22, 23]. Anterolateral capsular structures are considered important secondary rotational stabilisers in the knee [24]. All ligaments normally appear as dark signal structures on T1-, T2-, and PD-weighted images [19]. The cruciate ligaments are each two bundled which can be appreciated in all planes, and the ACL appears more striated than the PCL. This should not be misinterpreted as a partial thickness rupture of the ACL. Ligament pathology is seen as increased signal, thickening or thinning of the ligaments, or discontinuity of fibres (Fig. 2A–D). Bony avulsions at either the proximal or distal attachments of both the cruciate and collateral ligaments may also occur and should not be missed [24]. An associated avulsion is the Segond fracture, i.e. a small avulsion of the anterolateral tibial rim where the iliotibial band and anterolateral capsule attach. This is associated with an ACL rupture in adult patients. The arcuate complex is when the conjoined LCL and biceps femoris tendon attachment avulses the proximal fibular head, this findings is associated with an ACL rupture (Fig. 2D–E) [24]. The PCL differs from the ACL because it can appear to have intact fibres, but the knee may be PCL deficient. In addition to discontinuity of fibres, a complete PCL rupture is diagnosed when the short axis of the PCL is ≥ 8 mm. Both the ACL and PCL can show mucoid degeneration, which is seen as increased signal intensity in the centre of the ligament on T1-, T2-, and PD-weighted images. When the knee is extended, knee joint alignment is useful for assessing ACL or PCL deficiency in the sagittal plane. The tibia translates anteriorly in the lateral compartment in ACL deficiency; in PCL deficiency, the medial tibia moves posteriorly (> 10 mm) [11, 23]. A 3-mm anterior displacement of the medial meniscus in the mid-medial compartment on sagittal MRI can also be a sign of (chronic) PCL deficiency (Fig. 3A–C) [25]. Level of evidence 3b/Low.
We recommend reporting ligament pathology as full- or partial-thickness tears, bearing in mind the limitations of MRI in diagnosing partial tears [6, 26, 27]. Searching for indirect signs of ligament insufficiency or deficiency as well as evaluation of joint malalignment may increase confidence in diagnosis [23, 25]. Level of evidence 5/Low (Fig. 2A–D).
Cartilage and osteochondral lesions
The femoral, tibial, and retropatellar surfaces are capped with articular cartilage [19]. Pathology is seen on fluid-sensitive sequences as high signal abnormalities. Cartilage pathology should be reported according to the depth of the lesion, graded as laceration/fibrillation when < 50%, and full thickness when > 50%, extending into the subchondral bone [7, 26]. One should be aware that MRI underestimates cartilage lesions [28]. A cartilage lesion can progress and affect the underlying bone and bone marrow, to an osteochondral lesion. The term osteochondral lesion encompasses several different entities [7]. In osteochondritis dissecans, there is disruption of endochondral ossification in the epiphysis, which results in an unstable fragment [7, 29]. If the fragment is displaced, it becomes an osteochondral defect (Fig. 4A–D). Any focal defect visible in the cartilage, as well as the underlying bone, may be called an osteochondral defect. Traumatic or insufficiency subchondral fractures (Fig. 5A–C), osteochondritis dissecans, avascular necrosis, and degenerative disease, all of which have different underlying pathophysiology, might progress into an osteochondral defect—the end state [29]. If one is uncertain which subgroup the lesion belongs to, one can safely use “osteochondral lesion(s)”; though it is an unspecific term, it is very accurate. Finally, it is not recommended to use the term “osteoarthritis” on MRI as this does not correlate with clinical findings, compared to osteoarthritis on radiographs which does correlate with clinical findings [7]. Level of evidence 5/Low.
MRI has a valuable role in differentiating stable from unstable osteochondral lesions in adults where the risk of a fragment loosening is related to visible high-signal intensity line between the bony fragment and rest of the bone [30].
Other structures
The infrapatellar fat pad (also called Hoffa’s fat pad) acts as a “protective pillow” and stabilises the anterior knee joint structures during function [31]. Fat pad oedema or fibrosis may cause symptoms [31]. Patellar maltracking is one of the causes for anterior knee pain and may be associated with ACL injury [31]. If this is suspected, further imaging findings indicating chronic patellar pathology on MRI include lateralisation of the patella and an abnormal sulcus depth or sulcus angle, as these findings render the knee joint susceptible to patellar subluxations [31].
Rarely, soft tissue surrounding the knee may reveal neural ganglia or evidence of posttraumatic nerve injury and denervation of muscle compartments [32]. Finally, always look at all corners, as sometimes acute trauma of the tendons or fractures are not suspected clinically (Fig. 6A–C).
Summary statement
MRI of the knee is commonly performed. The recommended MRI protocol includes both T1-weighted and fluid-sensitive images, in three orthogonal imaging planes. It is important to correlate the clinical findings with MRI findings; however, all structures from menisci, tendons and ligaments, bones, and soft tissues should be evaluated. It is important to use correct terminology, such as for partial- and full-thickness tear when reporting ligament pathology, for meniscal tears and lesions, and regarding osteochondral defects. It is also important to be aware of which pathology is usually underreported on MRI Table 1.
Patient summary
MRI of the knee is a useful examination for assessing acute and chronic knee pain. Knee ligaments, menisci, bones, and soft tissue around the knee can be readily assessed. Radiographs serve as an adjunct to the MRI in the post-traumatic setting and are important so as to not miss subtle bony avulsions on MRI.
Abbreviations
- ACL:
-
Anterior cruciate ligaments
- LCL:
-
Lateral collateral ligaments
- MCL:
-
Medial cruciate ligaments
- PCL:
-
Posterior cruciate ligaments
- PD:
-
Proton density
References
Ahmed I, Moiz H, Carlos W et al (2021) The use of magnetic resonance imaging (MRI) of the knee in current clinical practice: a retrospective evaluation of the MRI reports within a large NHS trust. Knee 29:557–563. https://doi.org/10.1016/j.knee.2021.02.034
Chien A, Weaver JS, Kinne E, Omar I (2020) Magnetic resonance imaging of the knee. Polish J Radiol 85:509–531. https://doi.org/10.5114/pjr.2020.99415
Komarraju A, Maxwell C, Kung JW et al (2023) Causes and diagnostic utility of musculoskeletal MRI recall examinations. Clin Radiol 78:e221–e226. https://doi.org/10.1016/j.crad.2022.11.004
Kassarjian A, Fritz L, Afonso P et al (2016) Guidelines for MR imaging of sports injuries. Eur Soc Skelet Radiol. Available via https://essr.org/content-essr/uploads/2016/10/ESSR-MRI-Protocols-Knee.pdf. Accessed 15 Nov 2023
Kopf S, Beaufils P, Hirschmann MT et al (2020) Management of traumatic meniscus tears: the 2019 ESSKA meniscus consensus. Knee Surgery, Sport Traumatol Arthrosc 28:1177–1194. https://doi.org/10.1007/s00167-020-05847-3
Hall MM, Allen GM, Allison S et al (2022) Recommended musculoskeletal and sports ultrasound terminology: a Delphi-based consensus statement. J Ultrasound Med 18:2395–2412. https://doi.org/10.1002/jum.15947
Markhardt BK, Huang BK, Spiker AM, Chang EY (2023) Interpretation of cartilage damage at routine clinical MRI: how to match arthroscopic findings. Radiographics 43:e239008. https://doi.org/10.1148/rg.239008
Cheng Q, Zhao FC (2018) Comparison of 1.5- and 3.0-T magnetic resonance imaging for evaluating lesions of the knee: a systematic review and meta-analysis (PRISMA-compliant article). Medicine (Baltimore) 97:1–9. https://doi.org/10.1097/MD.0000000000012401
Altahawi F, Pierce J, Aslan M et al (2021) 3D MRI of the knee. Semin Musculoskelet Radiol 25:455–467. https://doi.org/10.1055/s-0041-1730400
(2020) ACR–SPR–SSR practice parameter for the performance and interpretation of magnetic resonance imaging (MRI) of the knee. 1–16. Available via https://www.acr.org/-/media/ACR/Files/Practice-Parameters/mr-knee.pdf. Accessed 6 Nov 2023
Chien A, Weaver JS, Kinne E, Omar I (2020) Magnetic resonance imaging of the knee. Polish J Radiol 85:e509–e531. https://doi.org/10.5114/pjr.2020.99415
Lins CF, Salmon CEG, Nogueira-Barbosa MH (2021) Applications of the Dixon technique in the evaluation of the musculoskeletal system. Radiol Bras 54:33–42. https://doi.org/10.1590/0100-3984.2019.0086
Chiabai O, Van Nieuwenhove S, Vekemans MC et al (2023) Whole-body MRI in oncology: can a single anatomic T2-Dixon sequence replace the combination of T1- and STIR sequences to detect skeletal metastasis and myeloma? Eur Radiol 33:244–257. https://doi.org/10.1007/s00330-022-09007-8
Baumbach SF, Pfahler V, Pozza SBD et al (2020) How we manage bone marrow edema—an interdisciplinary approach. J Clin Med 9:551–568. https://doi.org/10.3390/jcm9020551
Kumahara R, Kimura Y, Sasaki S et al (2022) Prevalence of Segond fractures associated with anterior cruciate ligament injuries and their influence on knee joint stability; a case-control study. BMC Musculoskelet Disord 23:180. https://doi.org/10.1186/s12891-022-05127-w
Ridley U, Ridley L (2020) Imaging of the knee: common acute presentations to general practice. Aust J Gen Pract 49:344–349. https://doi.org/10.31128/AJGP-10-19-5120
Fox MG, Chang EY, Amini B et al (2018) ACR Appropriateness Criteria® chronic knee pain. J Am Coll Radiol 15:S302–S312. https://doi.org/10.1016/j.jacr.2018.09.016
Simpfendorfer C, Polster J (2014) MRI of the knee: what do we miss? Curr Radiol Rep 2:43–57. https://doi.org/10.1007/s40134-014-0043-2
Vohra S, Arnold G, Doshi S, Marcantonio D (2011) Normal MR imaging anatomy of the knee. Magn Reson Imaging Clin N Am 19:637–653. https://doi.org/10.1016/j.mric.2011.05.012
Kim JG, Han SW, Lee DH (2016) Diagnosis and treatment of discoid meniscus. Knee Surg Relat Res 28:255–262. https://doi.org/10.5792/ksrr.16.050
Krych AJ, Wu IT, Desai VS et al (2018) High rate of missed lateral meniscus posterior root tears on preoperative magnetic resonance imaging. Orthop J Sport Med 6:1–6. https://doi.org/10.1177/2325967118765722
Khodarahmi I, Alizai H, Alaia E, Gyftopoulos S (2022) MR imaging of the knee posterolateral and posteromedial corner injuries. Magn Reson Imaging Clin N Am 30:215–226. https://doi.org/10.1016/j.mric.2021.11.003
Naraghi AM, White LM (2016) Imaging of athletic injuries of knee ligaments and menisci: sports imaging series. Radiology 281:23–40. https://doi.org/10.1148/radiol.2016152320
Choi C, Lee SJ, Choo HJ et al (2021) Avulsion injuries: an update on radiologic findings. Yeungnam Univ J Med 38:289–307. https://doi.org/10.12701/yujm.2021.01102
Ahn DY, Park HJ, Kim MS et al (2022) Protruding anterior medial meniscus and posterior tibial translation as secondary signs of complete and partial posterior cruciate ligament tear. Br J Radiol 95:1–8. https://doi.org/10.1259/bjr.20210976
Koch JEJ, Ben-Elyahu R, Khateeb B et al (2021) Accuracy measures of 1.5-tesla MRI for the diagnosis of ACL, meniscus and articular knee cartilage damage and characteristics of false negative lesions: a level III prognostic study. BMC Musculoskelet Disord 22:1–9. https://doi.org/10.1186/s12891-021-04011-3
Sanchez-Munoz E, Lozano Hernanz B, Zijl JAC et al (2023) Accuracy of magnetic resonance imaging in the diagnosis of multiple ligament knee injuries: a multicenter study of 178 patients. Am J Sports Med 51:429–436. https://doi.org/10.1177/03635465221145697
Krakowski P, Karpiński R, Jojczuk M et al (2021) Knee MRI underestimates the grade of cartilage lesions. Appl Sci 11:1–16. https://doi.org/10.3390/app11041552
Gorbachova T, Melenevsky Y, Cohen M, Cerniglia BW (2018) Osteochondral lesions of the knee: differentiating the most common entities at MRI. Radiographics 38:1478–1495. https://doi.org/10.1148/rg.2018180044
Hu H, Zhang C, Chen J et al (2019) Clinical value of MRI in assessing the stability of osteochondritis dissecans lesions: a systematic review and meta-analysis. AJR Am J Roentgenol 213:147–154. https://doi.org/10.2214/AJR.18.20710
Kim JH, Lee SK (2020) Superolateral hoffa fat pad edema and patellofemoral maltracking: systematic review and meta-analysis. AJR Am J Roentgenol 215:545–558. https://doi.org/10.2214/AJR.19.22263
Dalili D, Isaac A, Fayad LM, Ahlawat S (2021) Routine knee MRI: how common are peripheral nerve abnormalities, and why does it matter? Skeletal Radiol 50:321–332. https://doi.org/10.1007/s00256-020-03559-w
Sardanelli F, Hunink MG, Gilbert FJ et al (2010) Evidence-based radiology: why and how? Eur Radiol 20:1–15. https://doi.org/10.1007/s00330-009-1574-4
Martí-Bonmatí L (2021) Evidence levels in radiology: the insights into imaging approach. Insights Imaging 12:10–11. https://doi.org/10.1186/s13244-021-00995-7
Acknowledgements
This paper was endorsed by the Executive Council of the European Society of Radiology (ESR) and the Executive Committee of European Society of Musculoskeletal Radiology (ESSR) in February 2024.
A great thank you to Prof. Mario Maas, ESSR president, and Prof. Antoine Feydy, ESSR President-elect, for their valuable input regarding these recommendations.
The study was performed at Haraldsplass Deaconess Hospital, Postboks 6165 Posterminalen, 5892 Bergen, Norway.
Funding
Open access funding provided by University of Bergen (incl Haukeland University Hospital) The authors state that this work has not received any funding.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Guarantor
The scientific guarantor of this publication is Anagha P Parkar.
Conflict of interest
The authors of this manuscript declare no relationships with any companies, whose products or services may be related to the subject matter of the article.
Statistics and biometry
No complex statistical methods were necessary for this paper.
Informed consent
Written informed consent was not required.
Ethical approval
Institutional Review Board approval was not required.
Study subjects or cohorts overlap
Not applicable.
Methodology
• practice recommendations
Additional information
Publisher's note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
This article belongs to the ESR Essentials series guest edited by Mark Dewey (Berlin/Germany).
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Parkar, A.P., Adriaensen, M.E.A.P.M. ESR essentials: MRI of the knee—practice recommendations by ESSR. Eur Radiol (2024). https://doi.org/10.1007/s00330-024-10706-7
Received:
Revised:
Accepted:
Published:
DOI: https://doi.org/10.1007/s00330-024-10706-7