Variable parameters:

Remove patients with 0% LGE on MRI 1 from survival analysis: FALSE
Imported file name: ici_data_with_outcomes_updated_8_5_24.csv
Definition of ACE: MI, arrhythmia, cardiomyopathy, heart failure and death
End of study (follow up) date: 9/20/2023

For this analysis I only include patients with true symptomatic ICI (removed controls for now).

ROC Curve to establish optimal LGE % level threshold for predicting outcome (ACE)

It establishes that the best LGE threshold for predicting ACE is 2.9% with specificity of 0.44 % and sensitivity of 0.86%.

Table with different thresholds on ROC curve, and their corresponding sensitivity and specificity

threshold	specificity	sensitivity
Inf	1.00	0.00
32.55	0.96	0.00
29.90	0.96	0.05
27.55	0.96	0.10
24.50	0.96	0.14
23.50	0.92	0.14
22.65	0.88	0.14
21.60	0.84	0.14
19.65	0.84	0.19
17.85	0.80	0.19
17.45	0.76	0.19
16.80	0.76	0.24
14.15	0.72	0.24
10.70	0.68	0.24
9.35	0.64	0.24
8.35	0.64	0.33
7.20	0.64	0.38
6.75	0.60	0.38
6.40	0.60	0.43
6.20	0.60	0.48
5.90	0.60	0.52
5.55	0.60	0.57
5.35	0.60	0.62
5.15	0.56	0.62
4.75	0.52	0.62
4.25	0.52	0.67
3.95	0.48	0.67
3.80	0.48	0.71
3.60	0.48	0.81
3.25	0.44	0.81
2.90	0.44	0.86
2.65	0.40	0.86
2.40	0.36	0.86
2.25	0.32	0.86
2.05	0.32	0.95
1.80	0.28	0.95
1.60	0.24	0.95
0.75	0.24	1.00
-Inf	0.00	1.00

Survival analysis of patients using LGE threshold of 2.9%

Median follow up in days (IQR): 435 (190.25 - 725.75)