Two tendon measurements that work when TT-TG does not

Design

Cross-sectional MRI case-control study, level of evidence 3

Population

284 adolescent knees - 142 with patellofemoral instability, 142 controls

Primary outcome

Diagnostic accuracy of patellar tendon-lateral trochlear ridge distance and patellar tendon inclination angle

Effect

PT-LTR area under the curve 0.903 (cutoff 4 mm); inclination angle 0.866 (cutoff 16 degrees); both remained informative with normal TT-TG

Patellofemoral instability has been assessed for years on tibial tubercle-trochlear groove distance, patellar height and trochlear shape. This case-control study of 284 adolescent knees - 142 with instability, 142 controls - tests two measurements of the patellar tendon itself: patellar tendon-lateral trochlear ridge distance, already proposed as a marker of lateralisation, and a newly defined patellar tendon inclination angle quantifying the tendon's obliquity relative to the femoral condyles.

Both were higher in unstable knees (PT-LTR mean 9.76 mm, inclination angle 22.76 degrees, both P < .001) and both correlated with patellar tilt, Caton-Deschamps index and sulcus angle. Patella alta significantly raised PT-LTR. On receiver operating characteristic analysis, PT-LTR reached an area under the curve of 0.903 with an optimal cutoff of 4 mm, and the inclination angle 0.866 with a cutoff of 16 degrees.

The finding that earns attention is the last one in the conclusion: both measurements remained informative in knees with a normal TT-TG distance. That is the group that causes trouble in clinic - an adolescent with recurrent dislocation, a reassuring TT-TG, and no obvious structural explanation for why the patella keeps leaving the groove. If these tendon-based parameters hold up in prospective work, they describe part of what the established measurements are missing. For now they are worth measuring alongside the standard set rather than instead of it.

• Consider adding PT-LTR and tendon inclination when TT-TG is normal but instability is recurrent
• Cutoffs from this cohort: 4 mm for PT-LTR and 16 degrees for the inclination angle
• Both measurements move with patella alta and trochlear dysplasia - they are not independent of them
• Case-control design means these are diagnostic markers, not validated predictors of recurrence
• Record which MRI sequence and slice you measured on; reproducibility depends on it

Why it matters

It offers an explanation for the unstable knee whose standard measurements come back normal.

Don't overread it

Case-control diagnostic accuracy overstates real-world performance, and the cutoffs were derived in the same dataset.

The statistics, in plain English

An area under the curve of 0.903 means that if you picked one unstable knee and one normal knee at random, PT-LTR would rank them correctly about 90% of the time - good discrimination for a single measurement. But this is a case-control design, where cases and controls were chosen in equal numbers and the cases already have the diagnosis. Diagnostic accuracy in that setting is almost always better than it will be in a clinic population where instability is uncommon and borderline knees are the norm. The 'optimal' cutoffs were derived from the same dataset that generated the curves, so they will perform less well the first time they are applied elsewhere.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.