Abstract Body (Do not enter title and authors here): Background
Arrhythmogenic cardiomyopathy (ACM) is associated with high risk of ventricular arrhythmias (VA) and heart failure (HF). Different predictive models and diagnostic criteria include parameters related to myocardial structural abnormalities and right ventricular (RV) ejection fraction (EF). Nevertheless, the value of global longitudinal strain (GLS) ascertained from cardiac magnetic resonance (CMR) for risk-stratification in the context of ACM is unknown.

Aim
We aim to study the association of low CMR-derived GLS with adverse outcomes in ACM patients.

Methods
This was a retrospective cohort study of ACM patients seen at a single quaternary center. ACM diagnosis followed the modified 2010 ARVC modified task force criteria. Primary outcome was a composite of life-threatening VAs, heart transplant or HF hospitalization during follow up. We excluded patients who suffered events prior to diagnosis, and those with missing or inadequate CMR protocol. Imaging parameters were acquired by CMR including EF, GLS, and late gadolinium enhancement (LGE).

Results
A total of 82 patients with suspected ACM were included, with a median age of 40 years, and 43 (52.4%) were females. Most common reason of initial visit was family history of genetic cardiomyopathy in 46 (56%), while 16 (20%) patients visited our clinic due to history of syncope. Out of 60 patients who had genetic testing, 21 (35%) had mutations of DSP gene. Left ventricular involvement (non-ischemic LV-LGE by CMR) was present in 42 (56.8%) out of 74 patients who had CMR with gadolinium. Median LVEF was 62% (56, 66), and RVEF of 53% (48, 59), both by CMR. Median LVGLS was -17.01% (-14.71, -18.43), while median RVGLS was -20.5% (-18.3, -24.4).
A total of 15 (18%) patients suffered the composite endpoint during an average follow up period of 8 years. Patients who suffered the composite endpoint had significantly lower mean LVGLS (-15.04% vs -16.70%, p= 0.048), and lower average RVGLS (-18.45% vs -21.33%, p=0.035) compared to patients free from the adverse outcome. Meanwhile there was no significant difference in LVEF (60.13% vs 61.18%, p=0.67), nor in RVEF (49.13% vs 54.01%, p=0.065) between the two groups.

Conclusion
Lower GLS was significantly associated with VAs and HF hospitalizations in our patient population. CMR derived GLS can be used as an important prognostic parameter of adverse outcomes in patients with suspected ACM. Especially in those with LV involvement and high genetic predisposition as our cohort.