MRI scans do not accurately predict hamstring graft diameter for ACL reconstruction

Ricci Plastow, Clare Cullen, Herbert Imalingat, Jamie C. Sergeant, Neil Jain

Research output: Contribution to journalArticlepeer-review


Introduction: Identify whether magnetic resonance imaging (MRI) pre-operatively can help predict intra-operative hamstring graft size for anterior cruciate ligament (ACL) reconstruction. Previous literature has shown correlation but not accuracy or reproducibility of such a prediction. Methods: 60 pre-operative MR scans from patients who had undergone ACL reconstruction were studied retrospectively. We measured the cross sectional area of gracilis and semitendinosus tendons using MR axial views. The intraoperative graft diameter was recorded. If the diameter intra-operatively was below an 8 mm threshold the case was highlighted. We then correlated these results with the pre-operative MR measurements. Pearson correlation and Bland- Altman plots were used to assess correlation and agreement between the MR and intra-operative measurements. Results: The Pearson correlation co-efficient was 0.35. The Bland-Altman plots showed poor agreement and reproducibility between readers. We highlighted 6 cases that had a graft diameter below 8 mm. We were able to predict these 6 cases with the pre-operative MR scans producing 100% sensitivity, but low specificity of 19%. Conclusion: Pre-operative MR scan measurementof hamstring size is a poor predictor of graft size due to poor specificity and not being reproducible between readers. Therefore, we do not recommend pre-operative MR scan hamstring measurement to predict intra-operative graft size. Level of evidence: IV.

Original languageEnglish
Pages (from-to)482-487
Number of pages6
JournalMuscles, ligaments and tendons journal
Issue number4
Publication statusPublished - 1 Oct 2018


  • ACL
  • Knee
  • Magnetic resonance imaging


Dive into the research topics of 'MRI scans do not accurately predict hamstring graft diameter for ACL reconstruction'. Together they form a unique fingerprint.

Cite this