PURPOSE: Lateral patellofemoral ligament (LPFL) reconstruction addresses medial patellar instability, but uncertainty regarding the optimal femoral attachment site may affect isometry and increase complication rates. This study aimed to establish landmarks for the femoral attachment of the LPFL graft based on in vivo isometry during active knee extension. METHODS: Dynamic computed tomography scans of 104 knees from 58 healthy participants were employed to examine flexion-extension movements. Length changes were assessed in approximately 1335 virtual LPFL graft fibres, which extended from the proximal one third of the patellar height to attachments across the femoral condyle. Four methods were evaluated for achieving (near-)isometric LPFL graft behaviour: three radiographic methods (R1-R3) and one anatomic method (A). Method R1 positioned the femoral attachment at a lateral equivalent of Schöttle's point, Method R2 at the centre of the trochlear groove arc, Method R3 at the centre of the lateral trochlear ridge arc and Method A at a point relative to the lateral epicondyle. RESULTS: Median length changes during extension were 7.7 mm (Method R1), 3.4 mm (Method R2), 2.7 mm (Method R3) and 3.0 mm (Method A). Method R3 demonstrated significantly smaller length changes compared to Methods R1 (p <
0.001) and R2 (p <
0.01), while Method A yielded smaller changes than Method R1 (p <
0.001). Notably, Method R1 resulted in continuous LPFL graft tightening throughout knee motion, whereas Methods R2, R3 and A showed initial tightening until 20° flexion, followed by slackening and near-isometric behaviour. CONCLUSION: Femoral graft attachment is best determined at the centre of the lateral trochlear ridge arc (Method R3) or 15.1 mm anterior and 3.4 mm proximal to the lateral epicondyle (Method A). These guidelines help improve surgical precision and minimize complications in LPFL reconstructions. LEVEL OF EVIDENCE: Level III.