Author + information
- Published online April 1, 2019.
- Rungroj Krittayaphong, MD∗ (, )
- Shuo Zhang, PhD,
- Pairash Saiviroonporn, PhD,
- Vip Viprakasit, MD,
- Prajak Tanapibunpon, MSc,
- Boonying Rerkudom, BNS,
- Ahthit Yindeengam, BSc and
- John C. Wood, MD, PhD
- ↵∗Siriraj Hospital, Mahidol University, 2 Wanglang Road, Bangkoknoi, Bangkok 10700, Thailand
Cardiac iron overload (CIO) is a deadly complication in patients with thalassemia, particularly in Southeast Asian countries (1). Although cardiovascular magnetic resonance (CMR) T2* imaging using 1.5-T is the recommended gold standard for noninvasive CIO diagnosis, on 3.0-T, CMR has multiple technical challenges that lead to decreased accuracy (2,3).
The aim of this study was to investigate the relationships between the emerging techniques T1 and T2 (3), and T2* on 3.0-T and T2* on 1.5-T in an actual thalassemia cohort with a wide range of cardiac iron loading, to further determine their diagnostic performance, and to generate a model for CIO assessment in thalassemia patients on 3.0-T. The study was approved by the local institutional ethics committee with informed consent obtained before CMR examination.
Patients with thalassemia (aged 8 years and older) who were referred to CMR were recruited for this study. All subjects underwent CMR on 1.5- and 3.0-T systems (Achieva and Ingenia, Philips Healthcare, Best, the Netherlands) on the same day. A standard black-blood T2* sequence was performed in a single breathhold on both 1.5 and 3.0-T.
Myocardial T1 maps were obtained using the native modified Look-Locker inversion recovery sequence within 1 breathhold (4). T2 mapping was performed using a standard multiecho, spin-echo and a multiecho gradient spin-echo sequence (5). Both were a single-breathhold electrocardiographically-gated, black-blood technique with a double inversion recovery pre-pulse for nulling the blood signal to reduce flow sensitivity. All parametric maps were acquired in the left ventricular mid-cavity, short-axis level, and were analyzed either on the scanner or in MATLAB (The MathWorks, Inc., Natick, Massachusetts) offline.
The quality of the generated maps were graded from 0 (not assessable with strong artifacts) to 3 (excellent assessability with no artifacts). For quantitative assessment, a single septal region of interest was manually defined, avoiding vasculature and the blood pool. CIO was defined by T2* on 1.5-T if <20 ms, whereas severe CIO was considered to be <10 ms (3).
Of the 50 patients enrolled and assessed, most patients received regular blood transfusions and were prescribed iron chelation therapy. Different levels of myocardial iron loading were present covering the entire range of the disease, whereas 12 patients had CIO assessed by T2* on 1.5-T (6 severe, 6 mild to moderate) (Table 1). All 3.0-T parameters could well distinguish the CIO and non-CIO groups (all p < 0.001), and all had excellent sensitivity (91.7% to 100%) and specificity (92.1% to 100%) for CIO diagnosis based on the receiver-operating characteristic analysis, with no statistically significant differences among the different methods (p = 0.1922 to 0.8426 by the method of DeLong). Although T1 best differentiated mild-to-moderate CIO and severe CIO (p < 0.001), T2 and T2* were not significantly different between these 2 groups (p = 0.09 and p = 0.714, respectively).
For association between 3.0-T parameters and T2* on 1.5-T, strong linear correlation was found for all 3.0-T methods (all p < 0.001), whereas T1 (R2 = 0.922) and T2* (R2 = 0.957) presented the best prediction of T2* on 1.5-T in a linear regression model (Table 1). In addition, T1 on 3.0-T had the highest qualitative rating score (2.8 ± 0.4), which was significantly higher than all other maps (p < 0.001).
This study demonstrated that all CMR parameters on 3.0-T could accurately detect CIO in a thalassemia cohort, with high diagnostic performance based on the derived cutoff values from the receiver-operating characteristic analysis. Although T1, T2, and T2* declined inversely with respect to cardiac iron levels, they were linearly proportional to one another in the range of clinical iron overload. Native myocardial T1 and T2* were most closely correlated to T2* on 1.5-T and revealed the highest diagnostic specificity, whereas T1 better differentiated within the CIO groups and exhibited the highest image quality rating, which appears to be a promising technique. This study demonstrated and confirmed the clinical applicability using established clinical pulse sequences, foreshadowing an even wider adoption of the approach for investigation of potential changes in outcomes.
Please note: This study was supported by a Siriraj Grant for Research Development (R15433024) and by Novartis Pharmaceuticals Corporation through APIA (Asia Pacific Iron Academy) MRI Network, under the supervision by Faculty of Medicine, Siriraj Hospital, Mahidol University, Thailand. The authors gratefully acknowledge Supapon Nakyen and Wisanu Chitrotchanarak for technical assistance, and thank all relevant colleagues and teams in the Siriraj Hospital for preparing the methods during installment of the 3.0-T system and transition from the 1.5-T system in our department to support this study. Dr. Zhang is an employee of Philips. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- 2019 American College of Cardiology Foundation
- Carpenter J.P.,
- Roughton M.,
- Pennell D.J.
- Krittayaphong R.,
- Zhang S.,
- Saiviroonporn P.,
- et al.
- Wood J.C.
- Zhang S.,
- Le T.T.,
- Kabus S.,
- et al.