ORIGINAL ARTICLE Selective improvement in Seattle Heart Failure Model risk stratification using iodine-123 meta-iodobenzylguanidine imaging Eric S. Ketchum, MD, MS, a Arnold F. Jacobson, MD, PhD, b James H. Caldwell, MD, a Roxy Senior, MD, c Manuel D. Cerqueira, MD, d Gregory S. Thomas, MD, MPH, e Denis Agostini, MD, PhD, f Jagat Narula, MD, PhD, g and Wayne C. Levy, MD a Background. The Seattle Heart Failure Model (SHFM) is a multivariable model that uses demographic and clinical markers to predict survival in patients with heart failure. Inappro- priate activation of the sympathetic nervous system, which contributes to the progression of heart failure and increased mortality, can be assessed using iodine-123 meta-iodobenzylgua- nidine (MIBG) cardiac imaging. This study investigated the incremental value of MIBG cardiac imaging when added to the SHFM for prediction of all-cause mortality. Methods. Survival data from 961 NYHA II-III subjects in the ADMIRE-HFX trial were included in this analysis. The predictive value of the SHFM alone and in combination with MIBG heart-to-mediastinum ratio (H/M) was compared for all-cause mortality (101 deaths during a median follow-up of 2 years). Results. The addition of H/M to the SHFM in a Cox model significantly improved risk prediction (P < .0001), with a greater utility in higher risk SHFM patients. The observed 2-year mortality in the highest-risk SHFM subjects (rounded SHFM score of 1) was 24%, but varied from 46% with H/M <1.2 to 0% with H/M >1.8. Net reclassification improvement was 22.7% (P < .001), with 14.9% of subjects who died reclassified into a higher risk category than sug- gested by SHFM score alone (P 5 .01) and 7.9% of subjects who survived reclassified into a lower risk category (P < .0001). The 2-year integrated discrimination improvement (14.14%, P < .0001) and the 1-year area under the receiver-operator characteristic curve (10.04, P 5 .026) both showed significant improvement for the combined model with H/M compared to the SHFM alone. Conclusion. The addition of MIBG imaging to the SHFM improves risk stratification, especially in higher risk patients. MIBG may have clinical utility in higher risk patients who are being considered for devices such as ICD, CRT-D, LVAD, and cardiac transplantation. (J Nucl Cardiol 2012;19:1007–16.) Key Words: Heart failure Æ Seattle Heart Failure Model Æ cardiac I-123 meta- iodobenzylguanidine imaging Æ prognosis Æ statistical modeling From the Division of Cardiology, a University of Washington, Seattle, WA; GE Healthcare, b Princeton, NJ; BRU, Royal Brompton Hos- pital, Imperial College, London and Northwick Park Hospital, c Harrow, United Kingdom; Imaging and Heart and Vascular Insti- tutes, d Cleveland Clinic, Cleveland, OH; Long Beach Memorial Medical Center, e Long Beach, CA; Centre Hospitalier Universitaire Cote de Nacre, f Caen, France; Division of Cardiology, g Mount Sinai School of Medicine, New York, NY. The study was supported financially by GE Healthcare. Received for publication Jun 1, 2012; final revision accepted Jul 2, 2012. Reprint requests: Eric S. Ketchum, MD, MS, Division of Cardiology, University of Washington, 1959 NE Pacific Street, P.O. Box 356422, Seattle, WA; eketchum@washington.edu. 1071-3581/$34.00 Copyright Ó 2012 American Society of Nuclear Cardiology. doi:10.1007/s12350-012-9603-0 1007