Letter to the Editor “CT compared to MRI for functional evaluation of the right ventricle: a systematic review and meta-analysis”

by Jawdat Abdulla, MD PhD FESC (jawab@dadlnet.dk)

CT compared to MRI for functional evaluation of the right ventricle: a systematic review and meta-analysis

Dear Editor,

We have read with interest the paper by Fu et al, who have conducted a comprehensive meta-analysis of the published studies on computed tomography (CT) compared to magnetic resonance imaging (MRI) for functional evaluation of right ventricle [1]. Nineteen studies have been included. The results are reported as mean differences (MD) with confidence intervals (CI) of the left ventricular end diastolic and systolic volumes (LVEDV and LVESV) and stroke volumes (LVSV) measured by CT compared with MRI as standard reference. The conclusion is that CT is as accurate as MRI.

The included studies have provided data both as means ± standard deviations and bias (MD) with limits of agreements (LOA) for the above mentioned parameters. The authors have presented the results of their meta-analyses as forest plots (Fig.3). In these figures the results are shown as MD with CI, while in the section of statistics they report that “MD, LOA and r were weighted […] The pooled effects of EDV, ESV […].” First of all, the authors do not report exactly which method they used for data pooling, and secondly if they have pooled MD and LOA derived from Bland-Altman plots, then why they report it as MD and CI and not report as “bias with LOA” in the forest plots? This might be easily misinterpreted as if the authors have pooled the means ± standard deviations using weighted mean difference method for continuous variables. In case they have pooled MD (bias) and LOA, then they are required to explain which method they have used (inverse variance method?). Moreover, the authors do not comment the results (which seem to be composed of bias with LOA) according to the traditional terminology “good” or “poor” agreement. This has added even more confusion to the interpretation of the results. An important limitation of combination of MD and LOA occurs if the studies are not normally distributed.

These or other likely limitations have not been reported in the section of limitations. Besides, this paper lacks pooling of mean age, male sex %, mean interval between the two measurements by CT and MRI, mean LVEDV, LVESV, and LVSV in the two comparison groups. These variables are helpful for the reader to understand the general characteristics of the overall included populations. In order to avoid misleading results or misinterpretation by the readers, the authors are kindly required to explain their statistical approaches to clarify the ambiguities in the presented results.