Reply to Stöllberger et al.

We thank Stöllberger and colleagues for their interest in our manuscript (Riekerk et al., 2021) which they consider ‘important’ (Stöllberger et al., 2022). The validity of our key findings is not being challenged, namely that higher spatial resolution improves the interpretation of the extent of ventricular trabeculation and that the prevalence of excessive trabeculation is much less prevalent in pathology assessments than in non‐invasive in vivo imaging. In addition, Stöllberger and colleagues recognize that our study addresses one of the gaps in our knowledge concerning comparisons ‘between pathoanatomic and cMRI findings of left ventricular (LV) trabeculations’. We are happy to provide a response to their additional questions.

We did not provide an explicit rationale for not including (ex vivo) echocardiography. The use of echocardiography, such as in Tsang et al. (2012), would have facilitated a comparison between our study and clinical ultrasound. Nonetheless, the key finding that higher spatial resolution improves the interpretation of the extent of ventricular trabeculation, will likely stand irrespective of the imaging modality (CMR, CT, or echocardiography).

Particular diagnostic criteria for excessive trabeculation were used (trabeculation/compact >2.3, Petersen et al., 2005; the volume of trabeculation >25% of LV myocardial volume), whereas other criteria (Stöllberger & Finsterer, 2021) were not. There is a plethora of diagnostic criteria (D’Silva & Jensen, 2020) some of which show considerable overlap in outcome (Ivanov et al., 2017). Therefore, the key finding that the prevalence of excessive trabeculation is much less prevalent in pathology assessments than in non‐invasive in vivo imaging will likely stand irrespective of the diagnostic criterion.

For the following reasons, we did not exclude papillary muscles, false tendons, and aberrant bands from the measurements of trabecular layer volume as they sometimes are (Stöllberger & Finsterer, 2021) but not always (Luu et al., 2022). It has long been emphasized that embryonic trabeculations are the common scaffold from which trabeculae carneae, papillary muscles, and Purkinje muscles develop (Benninghoff, 1933; Jensen et al., 2012; Miquerol et al., 2010; Moorman et al., 1998; Sedmera et al., 2003; van Weerd & Christoffels, 2016). Uncommon structures such as false tendons and aberrant bands could develop from embryonic trabeculations as well. The inclusion of the papillary muscles and other trabecular structures in diagnostic assessments is then justified from a biological perspective since they are derivatives of the trabecular layer that is considered aberrant in excessive trabeculation. Irrespective of biology, whether to include papillary muscles can be determined on the sensitivity and specificity of diagnosis, but we would a priori give precedence to measurements that are consistent with the state‐of‐the‐art knowledge. From the perspective of diagnosing hypertrabeculation on the basis of the number of prominent bands, we can appreciate the wish to disregard aberrant bands that may inflate the number of prominent trabeculations. But this issue is more of a problem in some diagnostic criteria and less in others. If aberrant bands are present, they will be small and few and some diagnostic criteria will not be much affected by their presence, for example layer volumes (e.g. Grothoff et al., 2012; Riekerk et al., 2021; Luu et al., 2022), the requirement of more than one segment with a high trabecular‐to‐compact width ratio (e.g. Kawel et al., 2012) or fractal dimension (Captur et al., 2017).

Stöllberger and colleagues ask whether we could add additional information on the distribution of fibrosis and the presence of symptoms of neuromuscular disorders in the five cases of excessive trabeculation we report on. Our text, figures, and Table 2 describe and document the fibrosis we found in the five cases (Riekerk et al., 2021), and at this point, we do not have additional detail we can offer. We revisited our patient archive. No descriptions of symptoms of neuromuscular disorders were found.

Our study was based on autopsy hearts. Stöllberger and colleagues wonder whether the duration of preservation affected the presentation of the ventricular anatomy. They cite a study that measured the effect of four fixatives on the relative position of pins inserted in porcine hearts (Hołda et al., 2016). In the cited study, it was found that most change occurs within the first day of fixation and that the relative positions hereafter are mostly stable in the total study period of 7 days. The hearts we studied were from 1998 or earlier. We cannot rule out that there has been an effect of preservation on the specimens of our investigation. We never noticed hearts that were desiccated or wrinkled from dehydration, however. Some hearts were very contracted and stiff, whereas others were dilated and pliable, the overall presentation not unlike that of more recently fixated hearts.

We characterized some hearts that did not fulfil the criteria for excessive trabeculation as ‘abnormally trabeculated’. Stöllberger and colleagues ask us to clarify what abnormal is meant to signify. We did not mean to imply that abnormally trabeculated should be considered a distinct setting, let alone a practical one since these hearts were very heterogeneous in presentation [the compact wall had clefts that could be mistaken as trabeculation (Figure 8), trabeculae carneae were unusually similar in size (Figure 9) or the trabeculae carneae very numerous locally without fulfilling diagnostic criteria for excessive trabeculation (Figure 10)]. In addition, on the population level, the trabecular‐to‐compact width ratio exhibits a skewed distribution with a long tail towards the high values (Anderson et al., 2017). Thus, some of the hearts we investigated were more trabeculated than the median, but not excessively trabeculated. Stöllberger and colleagues additionally ask whether some of the abnormally trabeculated hearts were dilated. In our experience, the left ventricle of autopsy hearts tends to be quite contracted. That suggests to us that if a much‐dilated ventricle is found in an autopsy heart, that ventricle was likely dilated in vivo as well. But we are not sure we can conclude that the absence of a dilated cavity means that there was not a dilated state in the patient when alive. It is also of note that the state of contraction affects the distinction of the trabecular and compact layer (e.g. Grothoff et al., 2012; Stöllberger et al.,2008), but there is no consensus on whether so‐called true cases (Towbin & Jefferies, 2017) are easier to identify in systole or diastole (D’Silva & Jensen, 2020).