Aortic imaging and biomechanics in Marfan syndrome: keep it simple but not too simple

This editorial refers to ‘Proximal aorta longitudinal strain predicts aortic root dilation rate and aortic events in Marfan syndrome’†, by A. Guala et al., on page 2047. While current international clinical guidelines still encourage the use of aortic root diameter as the key criterion for risk assessment and timing of surgery, the medical community has intuitively felt for some time that a purely dimensional threshold approach to recommend surgery may be too simplistic. Over the last decade, various studies have demonstrated that the ascending aorta most probably dissects at a diameter way below the threshold of 50 or 55 mm1,2 and, therefore, the majority of patients would not quality for preventive repair if only the diameter was used to select someone for surgery.3 Thus, there clearly is a need for a more comprehensive selection algorithm in non-syndromic conditions and particularly in syndromic conditions such as Marfan syndrome. The analysis by Guala et al.4 of 117 patients with Marfan syndrome in this issue of the European Heart Journal revealed that longitudinal strain of the ascending aorta (once adjusted to age, body surface area, heart rate, and root diameter) may be instrumental as a useful independent predictor of aortic root expansion, Z-score growth rate, and eventually aortic events (such as elective surgery and dissection). They also report that, interestingly, circumferential strain and distensibility of the aorta are (surprisingly) not predictive for aortic events. In essence, a multivariate model including aortic dimension at the aortic root level, longitudinal strain, and age is advocated to be best to predict events. Subsequently, they claim that longitudinal strain should be implemented in the work-up and risk assessment of patients with Marfan syndrome. The group at Hospital Universitari Vall d’Hebron in Barcelona should be congratulated on their extensive and meticulous work and their creative use of magnetic resonance imaging- (MRI) derived data to predict events and outcomes in a sensitive group of patients at risk for major aortic vascular events, e.g. patients with Marfan syndrome. They also demonstrated the benefit of a structured surveillance programme for patients with Marfan syndrome, which could be used as a model for other scenarios of syndromic or even non-syndromic aortopathies. How should the medical community actually use this ‘new’ information derived from MRI post-processing, and is longitudinal strain robust enough? Is such a sophisticated image analysis to derive longitudinal strain now required to adequately serve our patients with Marfan syndrome? Those are important questions that may have an impact on both healthcare delivery and costs. First, the observation of a link between changes in the ascending aorta such as elongation, which could be seen as a surrogate of longitudinal strain, and elevated risk of dissection is not entirely new and was previously shown to depend on age and hypertension.5 From the second to the eighth decade of life, an elongation from ∼60 to 75 mm measured in the centre line of a 3D-reconstructed aorta is to be expected as a normal age-related phenomenon with a parallel diameter change from 28 mm to ∼40 mm in healthy subjects [from the sinotubular junction (STJ) to the brachiocephalic trunk (BCT)]; in hypertensive patients, an extra 10% increase in length and diameter is also to be expected.5 The Tubingen Aortic Pathoanatomy project elegantly showed that the distance between the aortic valve and the BCT was ≥110 mm in patients that had eventually dissected or dissected later, regardless of the influence of age and hypertension. That distance measured on average 93 mm in normal controls (P ≤ 0.001), which led to the conclusion that ascending aortic elongation, measured by any c-curved multiplaner reformatting method on computed tomography (CT) or MRI, precedes dissection and should be used in combination with the diameter of the ascending aorta for a better risk stratification.6 Secondly, it is not clear whether this new parameter of longitudinal strain of the ascending aorta (which requires more sophisticated MRI post-processing) is adding incremental information to just simple assessment of elongation of the ascending aorta (which can be obtained from almost any imaging modality such as CT, MRI, or angiography). It would be very interesting to see whether (longitudinal) elongation alone has a similar predictive power to longitudinal strain in Marfan patients. Thirdly, while longitudinal elongation at least in non-syndromic patients has been repeatedly and independently promoted as a predictor of risk,5 , 7–9 the new suggested parameter of longitudinal strain in Marfan syndrome still awaits prospective validation. It may be possible that multicollinearity also plays a role in the proposed multivariate model. Considering the need for sophisticated statistics as shown here, only prospective validation of the performance of any predictive model both in Marfan and non-syndromic patients will lead to widespread acceptance by the medical community. Finally, dissection and aneurysmal evolution may also take place in the descending thoracic aorta (certainly not affected) by longitudinal strain of the ascending aorta. There is certainly no doubt that both better prediction of future events and better selection of patients for pre-emptive treatment are needed. Modern therapeutic options in particular in patients with Marfan syndrome range from targeted personalized medications, over prophylactic ‘exostent’ support using PEARS,10 to timely aortic root replacement using dedicated surgical approaches;11 but those options need to be offered early enough by expert teams and at the right time. New parameters derived from imaging are likely to be important in this context and may help to target patients better for specific therapies and personalized care. A potential link between longitudinal strain derived from MRI and structural or biomechanic properties of the aorta in vivo in Marfan patients and also in non-syndromic patients may, however, be more complex than it seems, with a variety of determinants and triggers for the evolution of an aortic dissection still unknown. The impressive variety of clinical features of Marfan syndrome, the varying penetrance of signs and symptoms, and the occasional presence of genetic variants of unknown significance may well reflect the wide variety and complexity of this issue. It seems that a common denominator of splitting of cohesive layers in the tunica media of the aortic wall is preceded by loss or dysfunction of structural elements in the media such as elastin, collagen, non-fibrous matrix, and smooth muscle cells. Ideally, any direct assessment or in vivo visualization of such biological phenomena would lead to the holy grail of predicting aortic outcomes in connective tissue disorders and even in patients with aortopathy related to hypertension or atherosclerosis without known genetic underpinnings (Figure 1). In order to explore such a pathobiological phenomenon affecting the aortic wall, ‘functional’ MRI methods have recently been introduced such as four-dimensional MRI looking at focal wall stress and aortic wall integrity, early aneurysm formation, changes with the chronification of a dissecting lamella, and flow phenomena.12–15 Other types of functional imaging such as multidetector CT in conjunction with positron emission tomography (PET) may also by instrumental to predict focal areas of risk of dissection rupture by virtue of PET being able to reveal molecular aspects of aortic dissection16–18 as integrated PET/CT imaging may simultaneously reveal anatomical, molecular, and functional information as well as focal inflammation. The holy grail is to use genetic information and image processing to predict aorto-vascular events originating from pathological tissues. The data presented by the imaging group from Hospital Universitari Vall d’Hebron surely add a useful piece to the puzzle in predicting outcomes in patients with Marfan syndrome, but for clinical decision-making and selection of patients for individualized therapies, important considerations including the classics of family history, surgical expertise, patient’s preference, and individual lifestyle, as well as good clinical judgement always need to be factored in. Conflict of interest: none declared. The opinions expressed in this article are not necessarily those of the Editors of the European Heart Journal or of the European Society of Cardiology.