28In this review, the authors propose the selective use of multidetector computed tomography (MDCT) following transcatheter aortic valve implantation (TAVI) to detect subclinical leaflet thrombosis (SLT) and characterize the mechanical phenotype of the prosthesis. Their starting point is the changing patient population: TAVI is increasingly offered to younger patients at lower surgical risk, for whom long-term valve durability has become a central concern. In this regard, the authors note that in the Evolut Low Risk trial, valve reinterventions began to occur more frequently after TAVI than after surgery beyond the fifth year of follow-up, favoring surgical aortic valve replacement.

SLT encompasses three MDCT findings: (1) hypoattenuated leaflet thickening (HALT), defined as a meniscal-shaped lesion with an attenuation below 200 Hounsfield units (HU) and a thickness greater than 2 mm; (2) reduced leaflet motion (RELM); and (3) perivalvular thrombus, which may develop within the neo-sinus (the space between the native aortic sinus and the prosthetic frame) or beneath the leaflets along the inner skirt. Its reported prevalence ranges from 15% to 30%. SLT is a dynamic phenomenon, generally asymptomatic, and rarely associated with increased transvalvular gradients during its early stages.

The authors emphasize the dissociation between anatomical abnormalities and valve haemodynamics. Transthoracic echocardiography (TTE) assesses overall valve performance, but partial leaflet thickening may not significantly reduce the effective orifice area until a critical threshold is reached. Consequently, by the time the transvalvular gradient increases, leaflet pathology may already be advanced and potentially less reversible. They also suggest that late prosthetic dysfunction may present as regurgitation rather than stenosis, as observed in late reinterventions in the Evolut Low Risk trial. Furthermore, TTE cannot adequately characterize prosthesis expansion, ellipticity, canting, or commissural alignment.

The authors assign MDCT a dual role. First, it can detect leaflet and perivalvular thrombosis and characterize these abnormalities according to tissue attenuation. HALT typically exhibits attenuation values between 30 and 90 HU, clearly distinguishable from calcification and metallic artifacts. This may help differentiate soft, potentially reversible thrombus from more organized material. Second, MDCT defines the mechanical substrate, namely the interaction between the prosthesis and the patient’s anatomy. This includes underexpansion, ellipticity, canting, implantation depth, leaflet symmetry, and commissural alignment, all of which may influence neo-sinus washout and promote blood stasis. Rather than proposing MDCT as a universal screening or longitudinal surveillance tool, the authors regard it as a snapshot that defines the post-implant phenotype.

The proposed imaging window is within the first month after implantation. Early MDCT could identify high-risk mechanical configurations and potentially modifiable thrombosis, particularly in patients with a low bleeding risk, although this approach would not detect thrombosis developing later. The authors also highlight the dynamic nature of HALT. In the PARTNER 3 CT substudy, more than half of the HALT lesions present at 30 days had spontaneously regressed by 1 year, while new HALT developed in 21% of patients.

The review identifies three situations in which early MDCT might provide greater additional information: (a) suspected mechanical thrombotic substrate based on anatomy or procedural characteristics, including bicuspid valves, annular asymmetry, severe calcification, multiple resheathing maneuvers, aggressive post-dilatation, deep implantation, underexpansion, or commissural misalignment; (b) patients with a long life expectancy and low bleeding risk in whom antithrombotic treatment could potentially be modified; and (c) discordance between clinical findings and echocardiographic assessment.

Particular attention is given to perivalvular thrombosis. Neo-sinus thrombus could theoretically act as an embolic source, especially when located near a coronary ostium, and contribute to persistent leaflet thickening. Subvalvular thrombus, by contrast, may serve as a reservoir from which thrombotic material extends toward the leaflet base. Although TTE cannot directly identify this thrombus, it may reveal a large or asymmetric neo-sinus, a finding that, according to the authors, should prompt consideration of further assessment with MDCT.

Regarding patient management, the authors identify single antiplatelet therapy (SAPT) as the default strategy, although it does not prevent SLT. Anticoagulation, particularly with vitamin K antagonists, has been associated with HALT resolution. However, neither GALILEO nor the CT substudies evaluating direct oral anticoagulants (DOACs) demonstrated consistent clinical benefit. Nevertheless, the authors propose considering anticoagulation for 3 to 6 months, followed by repeat MDCT, in patients with documented thrombosis or a high-risk phenotype and low bleeding risk. In selected patients with an unfavorable mechanical substrate but no detectable thrombosis, they even consider anticoagulation for 6 to 12 weeks.

These recommendations are incorporated into the algorithm presented in Figure 8. In selected patients with thrombotic risk factors, long life expectancy, acceptable renal function, and no previous indication for oral anticoagulation, MDCT is performed between the predischarge period and 30 days after implantation. If neither thrombosis nor an unfavorable mechanical substrate is identified, SAPT is continued. If an unfavorable mechanical substrate is present without thrombosis, individualized anticoagulation or closer surveillance may be considered. When thrombosis is detected, anticoagulation is initiated and MDCT repeated after 6 to 12 weeks. If thrombosis has resolved, treatment is stepped down to SAPT; if it persists, anticoagulation is continued until resolution. In patients with thrombus attenuation exceeding 90 HU, considered the most unfavorable phenotype, further prolongation of anticoagulation may be considered.

In their discussion, the authors emphasize that they are not advocating systematic or serial MDCT surveillance, but rather selective imaging when its findings are likely to influence treatment. They acknowledge the additional costs, radiation exposure, contrast administration, and potential downstream investigations associated with this approach. They also recognize that the proportion of patients who would meet the proposed imaging criteria cannot currently be reliably estimated. Accordingly, they present their strategy as a hypothesis-generating conceptual model and call for prospective trials comparing conventional follow-up with MDCT-guided strategies, using clinical endpoints defined according to Valve Academic Research Consortium-3 (VARC-3) criteria.

COMMENTARY:

The question raised by this review is both relevant and timely. The 2025 ESC guidelines for valvular heart disease lowered the age threshold for recommending TAVI to 70 years. However, long-term trial results suggest that caution is warranted when extending TAVI indications to younger patients. For example, at 7 years, reinterventions in the Evolut Low Risk trial were more frequent after TAVI than after surgery (9.8% vs 6%), whereas bioprosthetic valve failure in PARTNER 3 was similar between groups (6.9% vs 7.5%).

The central concept of the review, integrating post-implant prosthesis geometry into the pathophysiology of thrombosis, is supported by observational evidence. However, the proposed causal pathway linking early HALT to structural valve degeneration and subsequent reintervention remains unproven. In the Evolut Low Risk CT substudy, HALT did not affect valve haemodynamics at 1 year. In PARTNER 3, only persistent HALT, rather than early HALT, was associated with a modest increase in transvalvular gradient (17.8 vs 12.7 mmHg). Furthermore, in ADAPT-TAVR, valvular or perivalvular thrombus, detected in 43% of patients, was not associated with new cerebral lesions or clinical events.

The trials underlying this review also reveal an interesting paradox. In Evolut Low Risk, HALT occurred with similar frequency after both procedures (30.9% vs 28.4% at 1 year), yet reintervention rates diverged later, driven by regurgitation (5.6% vs 1.6%) rather than stenosis. In PARTNER 3, early HALT was more frequent after TAVI (13% vs 5%), without an accompanying increase in bioprosthetic valve failure. Therefore, HALT did not predict which treatment group would subsequently experience greater valve deterioration.

Regarding antithrombotic treatment, the authors’ proposals are supported by limited evidence, since the available data suggest that anticoagulation improves imaging findings without demonstrating a corresponding improvement in clinical outcomes. In GALILEO, rivaroxaban reduced HALT in the CT substudy, but this was accompanied by an increased risk of death or thromboembolic events (HR = 1.35). Neither apixaban nor edoxaban demonstrated a clinical benefit. For this reason, the 2025 guidelines do not recommend systematic screening for HALT and suggest anticoagulation only when leaflet-motion restriction or a significant increase in transvalvular gradient is present. Nevertheless, the algorithm proposed by the authors goes further: it advocates treatment of HALT without haemodynamic consequences and even considers anticoagulation for unfavorable prosthetic geometry in the complete absence of detectable thrombus.

In summary, from my perspective, MDCT after TAVI is fully justified when there is clinical or echocardiographic suspicion of prosthetic valve dysfunction, such as an increasing transvalvular gradient, new regurgitation, or an embolic event, in line with current European guidelines. However, in asymptomatic patients with normal TTE findings, I would reserve its use for research purposes. Its main value may ultimately be prognostic rather than therapeutic, at least regarding antithrombotic therapy, by helping anticipate future management options such as TAVI-in-TAVI. In any event, prospective clinical trials evaluating these emerging MDCT-guided strategies are needed before their strengths, limitations, and true clinical value can be adequately established.

REFERENCE:

Moscarelli M, Pibarot P, Van Mieghem N, Lancellotti P, Modine T, Piazza N, et al. The role of selective multidetector computed tomography after a transcatheter aortic valve implant. Eur J Cardiothorac Surg. 2026;68(8):ezag200. doi:10.1093/ejcts/ezag200.