Manual long axis strain compared to automated feature-tracking deformation imaging for identification of masked left heart involvement in pulmonary hypertension.

BACKGROUND

Left heart disease (LHD) may cause or coexist in pulmonary hypertension (PH). Left atrial (LA) deformation identifies LHD in PH, however, feature-tracking (FT) deformation imaging remains underused due to limited inter-vendor agreement. Manual long axis strain (LAS) has been introduced as a software independent approach for deformation assessment. Consequently, we sought to compare their diagnostic accuracies for identification of LHD in PH.

METHODS

Patients referred to both right heart catheterisation (RHC) and cardiovascular magnetic resonance (CMR) imaging were enrolled in this monocentric registry. Patients were classified by RHC according to current guideline recommendations. CMR assessment included FT-deformation imaging for LA total strain (Es) and left ventricular (LV) global longitudinal strain (GLS) as well as LV/LA LAS.

RESULTS

In total n = 209 PH patients were included with n = 126 undergoing exercise-stress testing (n = 55 normal, n = 72 pre-capillary, n = 27 combined post-/precapillary, n = 15 isolated postcapillary, n = 34 exercise and n = 6 unclassified PH). LAS showed strong correlation with FT for both LA (r = 0.78, p < 0.001) and LV (r = 0.83, p < 0.001) deformation. LA (AUC 0.81 vs. 0.81, p = 0.89) and LV LAS (AUC 0.75 vs. 0.77, p = 0.55) equally identified LHD at rest compared to their respective FT counterpart. LA LAS showed good diagnostic performance for identification of LHD unmasked during exercise stress for pulmonary capillary wedge pressure (PCWP) ≥25mmHg (AUC 0.73 vs. 0.79, p = 0.042) and PCWP/Cardiac Output (CO) Slope > 2 (AUC = 0.66 vs. 0.73, p = 0.045) but significantly inferior to LA Es. LV LAS showed similar performance compared to LV GLS, but worse compared to LA LAS.

CONCLUSION

Software-independent LAS shows strong correlation to its FT counterparts and good diagnostic accuracy for detection of LHD in PH. However, for detection of masked LHD detected during exercise-stress only, FT LA Es remains the most accurate parameter.