Due to its complex nature, the assessment of TV morphology and disease mechanism requires multiple TTE and TEE windows [2].

The main TTE views are the left parasternal long-axis (LAX) view focused on the RV inflow, the left parasternal short-axis (SAX), the apical and the subcostal views (Table 1). The visualization of all leaflets in one two-dimensional (2D) view is only rarely achieved, mainly through modified subcostal or parasternal SAX windows. A precise TTE characterization of the leaflets is challenging but few anatomic landmarks may be helpful to guide the imager (Table 1) [2, 36, 37]. The aortic valve permits to localize the anterior leaflet, in particular the non-coronary cusp is adjacent to the antero-septal commissure. The entry point of coronary sinus into the right atrium (RA) locates the postero-septal commissure, instead.

A comprehensive TEE examination of the TV should include multiple windows from several depths and angles and the recurring use of biplane (or crossplane) modality [2, 38, 39].

The multilevel assessment of the TV classically begins at a mid-esophageal (ME) depth with two views (Table 2): ME 4-chamber view (at about 0${}^{\circ }$ degrees) which shows the septal leaflet and generally the anterior leaflet, though a retroflexion movement of the probe may reveal the posterior leaflet. The use of biplane mode is useful to clearly define which leaflet is visualized, also based on the above-mentioned anatomical markers (Table 2). The second ME view is the RV inflow-outflow view (at about 60${}^{\circ }$ degrees), which shows the anterior leaflet (adjacent to the aortic valve) and the posterior leaflet (attached to the posterolateral wall of the RV), while the septal leaflet lies behind on a different imaging plane. If imaged with biplane mode, this view allows to entirely span the coaptation line of the septal leaflet with both the anterior and the posterior leaflets.

At deep-esophageal (DE) level the probe becomes closer to the RA and the TV without the interposition of the left atrium. Due to this proximity, DE views are optimal for 3D acquisitions of the TV and for Doppler beam alignment with regurgitant jets. Like at ME level, main DE views are the 4-chamber and the RV inflow-outflow.

At trans-gastric (TG) level unlimited views of the TV may be imaged, with a meticulous manipulation of the probe (usually right and anterior flexion) and use of different angles and biplane mode. The SAX view of the TV is crucial to assess the valve anatomy (number and morphology of leaflets, presence of clefts/fold indentations), to identify the regurgitant orifice and to measure the coaptation gap size. This view is required for assessment of the TV anatomy according to the above-mentioned classification of Hahn et al. [11]. If a CIED-lead crosses the TV, this view allows to localize where the lead crosses the valve and to assess the presence and entity of a potential CIED-related interference with leaflet coaptation.

Finally, deep-transgastric (DT) views of the TV may be useful mainly for colour flow evaluation and optimal Doppler beam alignment. In case of intense shadowing of the TV at ME and DE levels due to the interposition of atrial septum, aortic and mitral valves, this window permits to solve this issue.

Given the limitations and challenges of 2D echocardiography for TV assessment, 3D imaging is nowadays recommended to provide an exhaustive evaluation of leaflets, annulus and subvalvular apparatus. 3D datasets may be acquired from any good-quality TTE or TEE view [28]. The TTE apical (RV-focused or foreshortened) and parasternal RV inflow views are generally the best approaches to achieve an optimal TTE 3D acquisition of TV, while the TEE DE view permits to reduce at minimum the distance between the TEE probe and the right heart and to acquire fulfilling 3D datasets [40, 41]. A good 3D acquisition always derives from a 2D view which has been adequately optimized in terms of gain, with a high tissue-blood contrast and low speckle noise. To achieve the best spatial resolution of 3D datasets, it is pivotal to maintain the whole TV within a small acquisition volume, while optimizing the acquisition volume size and shape, and the gain and temporal resolution settings. To comprehensively evaluate the anatomy of the TV, the valve is generally visualized “en face” from both the ventricular and the atrial perspectives. The atrial (or surgical) view is particularly useful when analysing a primary TR, as it allows a detailed assessment of the motion of the leaflets; the ventricular perspective provides information regarding the involvement of commissures or fold indentations into TR mechanism, the presence of a leaflet/chordal impingement due to a CIED or the entity/location of the regurgitant orifice. A multibeat acquisition is generally preferred, as it provides the highest resolution and frame rate. However, in presence of arrhythmias or marked respiratory variations, multibeat acquisitions are prone to stitching artifacts. Real-time or live 3D is particularly useful for the guidance of transcatheter procedures as it is less susceptible to motion artifacts and the use of real-time multiplanar reconstruction (MPR) allows rapid orientation within the 3D dataset. The above-mentioned anatomic landmarks (aortic valve, septum, coronary sinus) for leaflet identification need to be encompassed within the 3D dataset, in order to allow a correct interpretation of the 3D TV acquisition [41].

There is not a general agreement regarding the proper orientation of the en-face view of the 3D TV acquisition. Lang et al. [42] proposed to orient the TV with the septal leaflet at 6 o’clock regardless of the atrial or ventricular perspective. More recently, for interventional purposes, Muraru et al. [40] proposed an atrial perspective with the septal leaflet between 6 and 10 o’clock, whereas Agricola et al. [43] oriented the atrial view with superior vena cava (SVC) at 11 o’clock and inferior vena cava (IVC) at 7 o’clock.

3D echocardiographic imaging cannot always inform regarding the leaflet thickness and tissue features. In particular, the tricuspid leaflets are much thinner than the mitral ones, with a generally poorer echocardiographic definition, which is widely influenced by the leaflet orientation with respect to the ultrasonographic beam: an echocardiographic view with an annular plane perpendicular to the insonation beam usually provides better definition of the leaflets in systole (closed valve), while the reverse occurs for annular planes parallel to the beam [40]. Then, blurring artifacts alter the imager’s perception of leaflet thickness, while 3D traditional colour maps represent a depth map and do not directly mirror the tissue characteristics.

Unlimited useful information may be derived from high-quality 3D datasets thanks to offline MPR, which permit an in-depth exploration of the multiple 2D planes included within the acquired 3D volume at any point in the cardiac cycle (Fig. 2).

Echocardiography is the first-line and most widely used imaging tool for the assessment of dimensions and function of the right chambers.

The RV anatomy is notoriously complex, with a crescentic shape in the sagittal plane, a triangular profile in SAX view, and three different and indissoluble regions: inflow, apex and outflow [44]. Because of this unique complexity, a standardised approach is needed to achieve a reliable and reproducible echocardiographic assessment of this chamber. In particular, the RV-focused view should always be acquired and analysed as it has shown superior reproducibility for the evaluation of function and size parameters as compared with the standard apical view [45, 46]. Commonly used 2D measurements include basal, middle, and longitudinal dimensions for RV size assessment and tricuspid annular plane excursion (TAPSE), S’, fractional area change, and free-wall longitudinal strain for RV function evaluation. Even if each of these measures shows a significant prognostic meaning, their accuracy and reproducibility are suboptimal. Indeed, TAPSE and S’ assess only the longitudinal excursion of the RV base, whereas more comprehensive parameters such as fractional area change, and free-wall longitudinal strain are based only on a single tomographic plane. Thus, the use of 3D echocardiography is essential for a full and reliable assessment of the complex shape and systolic function of the RV. 3D echocardiography measures RV volumes and ejection fraction (EF) parameters otherwise measured only with CMR and CCT—with an excellent agreement as compared with CMR gold-standard, with a tiny underestimation of RV volumes by 3D echocardiography, but comparable values of RV ejection fraction (RVEF) between the two modalities [47, 48]. RV function assessment in the context of severe TR is even more challenging as all traditional parameters are closely dependent on loading conditions; in this perspective, 3D measurements are even more crucial. Moreover, 3D-derived parameters of RV size and function have been shown to provide additive and incremental prognostic information over other traditional echocardiographic parameters [49].

Differently from the left ventricle, the right one is a volume-loaded pump with a great compliance and a thin myocardial wall. In response to volume or pressure overload the right chambers exhibit different patterns of remodelling and consequently different mechanisms of secondary TR [50]. In patients without left-sided heart disease and without PH, the right atrial enlargement with consequent annular dilatation due to AF or HF with preserved EF is the primary driver of significant secondary TR, whereas the RV dimensions and function are normal (at least during the first stages of the TR pathophysiology) [50, 51]. RV pressure overload due to left-sided heart disease and PH causes various grades of RV dilation and dysfunction with consequent papillary muscle displacement and leaflet tethering-related secondary TR. Similarly, RV dilation and dysfunction due to primary RV disease (ischemic or cardiomyopathy) progressively causes a ventricular FTR [51]. Hence RV dilatation with TV tenting and right atrial dilatation with TA dilatation develop differently based on the specific underlying aetiology leading to FTR. 3D echocardiography, thanks to its volumetric accuracy, inter-operator reproducibility, and new tricuspid-specific quantification tools, is the best modality for an in-depth analysis of the interplay between the complex and long-neglected structures of the right heart [44, 50, 52, 53].

Moreover, due to the known sensitivity of the RV to pressure overload, the concept of RV-pulmonary artery coupling has been recently introduced to correctly assess the RV contractility response to increased afterload [54, 55]. The ratio between non-invasive measures of RV function (usually TAPSE) and echocardiographic-derived systolic pulmonary artery pressure (sPAP) has shown an independent prognostic value in several patient subsets, even those with severe TR [12]. As far as RV and pulmonary vasculature are coupled so that the increase of TAPSE (or any RV function non-invasive measure) and sPAP is comparable with a stable ratio, the RV is able to compensate the pressure overload. When the RV and pulmonary vasculature are not coupled anymore and the TAPSE/sPAP ratio decreases, that is the onset of RV failure. However, the diagnostic sensitivity for PH of echocardiography is limited (55%) and the patient cohort with discordant results from invasive and echocardiographic assessment showed the worst outcomes after transcatheter tricuspid valve repair (TTVR), probably because of a more severe TR degree [56]. Consequently, invasive measurements of pulmonary arterial pressures with right heart catheterization should always be performed in patients with severe TR, above all those screened for transcatheter therapies.

Proper analysis of TR mechanism is the first step for an adequate definition of the patient’s diagnostic and therapeutic pathways.

TR can be divided into primary, secondary and CIED related [12]. The characterization of TR mechanism is based on a comprehensive imaging assessment of three elements: leaflet mobility, annular dimensions, type of RV and RA remodelling.

Table 3 shows a classification of TR mechanisms based on pathophysiology and leaflet mobility.

Primary or organic TR is caused by TV abnormalities and accounts for about 8% of TR [57].

Secondary or FTR is a consequence of right heart chambers remodelling, is the most common TR type and is differentiated into ventricular and atriogenic/atrial (or often defined as isolated TR). Ventricular secondary TR is caused by leaflet tethering and papillary muscle displacement, as a consequence of RV dysfunction and enlargement in conditions of volume overload or PH, generally secondary to left heart disease. Atrial secondary TR is characterized by annular dilatation and is related to AF, age, and HF with preserved EF [12, 50, 53]. Right ventricular and RA remodelling patterns in response to different pathological stimuli have been previously described.

Finally, implantation or extraction procedures of CIED in the RV leads may cause significant TR in 7–45% of cases with a plethora of mechanisms: leaflet impingement/adhesion/perforation/avulsion, chordal rupture/entanglement [12, 21]. When evaluating a CIED-related TR, a careful multi-view echocardiographic assessment of TR degree is pivotal, because an underestimation (more frequently with TTE) of the TR severity may occur with colour-Doppler due to acoustic impedance and reflectivity of the CIED leads [21]. 3D datasets of the TV are particularly useful to assess the transvalvular trajectory of the lead and its relationships with leaflets and the subvalvular apparatus. The transvalvular position of RV leads may be commissural, impinging on a leaflet, adherent to a leaflet or in the middle of the valve. Commissural and central trajectories are generally safe and not associated with significant TR. An adherent lead generally moves altogether with the leaflet and may cause TR only if it significantly interferes with the leaflet systolic closure. An impinging leaflet, instead, restricts the leaflet systolic closure and usually causes a significant TR [21].

In patients with TR the main goals of CMR are: assessment of right ventricular volume and function, TR grading, anatomical evaluation of venous systemic and pulmonary returns and, finally, tissue characterization. Detailed anatomical study of TV leaflets and mechanism of TR can be done in case of inconclusive echocardiographic studies.

Right chambers remodelling due to TV disease must be assessed with a dedicated CMR protocol that adds some important dedicated sequences to the traditional CMR protocol. The dedicated CMR protocol is summarized in Table 4.

Balanced steady-state free-precession (b-SSFP) acquisitions have good signal-to-noise ratios and high blood-to-myocardium contrast. Short- and long-axis views can be acquired using b-SSFP sequences. The SAX stuck must be acquired from RV base to apex, taking care to include the most basal part of RV that can be displaced in case of RV enlargement over the plane of LV base.

Additional RV long axis views can be obtained from SAX views, 4 chamber views, coronal and transaxial localizer. These specialized views are: the RV inflow view, the RV inflow/outflow view, and the RV outflow tract (RVOT) view (Table 4 and Fig. 3).

RV transaxial stack can be obtained in a transaxial plane from the level of the diaphragm to the pulmonary bifurcation or as a stack of 4 chamber cine views; it can be useful to identify WMA and TR jets in multiple plains that cannot be visualized in SAX and LAX views. Both RV transaxial stack and SAX stack can be used to calculate RV volumes and function by contouring endocardial borders of RV in systole and diastole from (Table 4). Most of available post-processing software performs automating tracing in few seconds, deriving RV stroke volume (RVSV) and RVEF from RV end-diastolic volume (RVEDV) and RV end-systolic volume (RVESV) calculated using the Simpson’s method without geometric assumption.

After acquisition of b-SSFP, direct flow measurements can be done through phase-contrast sequences. To measure the aortic and pulmonary flow, phase-contrast sequences must be prescribed perpendicular to the LAX of the vessel and the flow direction. Two type of images are than generated: magnitude images and phase velocity maps [59]. Anatomic delineation and contouring of the vessel are made mainly in the magnitude images, while the phase map is used to directly calculate the flow over the cardiac cycle, as it represents the velocities within each pixel of the vessel. TR-regurgitant volume (TR-RV) can be calculated in a direct or indirect way (Table 4). The indirect method is the most used and it integrates information from cine images and phase contrast images by subtracting the forward pulmonary flow obtained by phase velocity maps from the RVSV calculated from the SAX stack (TR-RV mL/beat = RVSV – Pulmonary flow). Once TR-RV is obtained, RF can be calculated as the ratio between TR-RV and RVSV as shown by the formula: RF = TR-RV/RVSV. It is also possible to directly calculate TR by prescribing a phase contrast sequence with a plane parallel to TV annular plane in systole. However, tricuspid annular excursion is usually very wide, TV annular plane is not planar and TR jets often eccentric causing suboptimal evaluation of TR using the direct method.

Recently, volumetric 4D flow MRI has been developed to partially overcome these limitations [60]. 4D flow imaging has several advantages over conventional phase contrast cardiac MRI. The 4D flow can be prescribed as a single volume acquisition covering the entire heart, without the need to prescribe anatomical planes perpendicular to the flow of interest. MPR is easily used in post-processing to visualize regurgitant jets and measure TR-RV. The direct quantification of TR-RV can be done in a single measurement, taking care to measure at least 5 mm apart the valve plane, to avoid regions of velocity-aliasing and segments of severe signal dephasing. Quantification of TR by 4D flow MRI is highly reproducible and consistent across multiple methods of measurement [61]; it also showed excellent interobserver and intraobserver reliability for quantification of TR regurgitant fraction (TR-RF) and TR-RV [62].

Late gadolinium enhancement (LGE) sequences and native T1 mapping are both used for tissue characterization of the right ventricular walls. However, the normal RV wall thickness of 3–5 mm and the low spatial resolution of these sequences limit an adequate analysis of the RV walls. In case of right ventricular transmural infarction or RV walls hypertrophy, the presence of LGE can be easily appreciated, but TR is often accompanied by RV dilatation without hypertrophy due to chronic volume overload. Thus, the RV walls are usually thin and partial volume artifacts can affect RV wall evaluation with both LGE and T1 mapping sequences. Systolic acquisition of LGE sequences can improve RV walls visualization, while high resolution acquisition techniques have been developing and need definite validation [63].

The presence of RV LGE has never been studied in patients with isolated TR, while it has been described in different clinical scenarios. Location and extent of RV LGE may differ according to the underlying disease associated with the TR. Different types of RV LGE are described in Table 4. The most common pattern of RV LGE is the RV insertion point. The clinical relevance of isolated RV insertion point LGE in subjects without additional evidence of cardiac damage and in patients with non-ischemic dilated cardiomyopathy does not convey worse prognosis [64, 65]. There is some evidence that it can be associated with high left ventricular filling pressure and diastolic dysfunction in patients with hypertrophic cardiomyopathy [66]. PH is an important cause of RV walls fibrosis and LGE at the RV junctional insertion points and into the interventricular septum; different patterns have been described with involvement of both RV insertion points and interventricular septum, only one or neither of all [67]. The presence of LGE at RV seems to be more common in case of pre-capillary PH [68]. PH patients with LGE show significantly higher mean PA pressure (obtained with right heart catheterization) and lower RVEF than patients without LGE. In patients with Ebstein’s anomaly LGE can be localized in the RA, in the atrialized RV and in the RV and seems to be associated with the severity of TR and presence of supraventricular arrythmia [69]. In patients with Tetralogy of Fallot (ToF) the presence of LGE relates with higher right ventricular volumes, lower EF and a higher pulmonary regurgitant fraction. In ToF the most common locations of LGE are RVOT, the ventricular insertion points and around the ventricular septal defect patch [70]. In case of a prior RV infarction, transmural LGE is visualized into the RV free wall extending from the inferior LV myocardium or the inferior interventricular septum [71, 72]. In conclusion, the presence of RV LGE is often associated with concomitant LV pathology, PH or ischemic heart disease, however its role in patients undergoing TV intervention is still unknown.

In patients with TR the exam quality can be affected by the presence of arrythmia (atrial fibrillation is the most common) and/or PM leads or devices (like TriClip). Most common artefacts that occur during Cine bSSFP sequences are presented in Fig. 4. In case of extreme irregular RR intervals, prospective triggering or real-time free-breathing cine sequences can be acquired to avoid artefacts related to HR variability. Leads and other intracardiac devices are responsible of metallic artefacts than can compromise image quality, above all of bSSFP sequences; in this case the use of cine fast-spoiled gradient, instead of bSSFP, can reduce magnetic susceptibility artefacts.

Due to the complex shape and structure of RV, multiple CMR planes are required for a comprehensive analysis of right chambers and TV [73].

As previously mentioned, SSFP sequences in RV LAX planes must be part of the acquisition protocol. In the four-chamber view, the trabeculated apex, the moderator band, the inlet, the lateral wall and the interventricular septum are displayed. The inlet, the inlet/outlet and the outlet view are described together with the SAX in Fig. 3.

A change in RV shape and volume can be the first sign of RV dysfunction, pressure or volume overload. In severe TR both signs of pressure and volume overload can coexist according to the underlying disease and TR mechanism.

In primary TR, the RA or RV can be normal or mildly dilated during the first phases of the disease; however, as TR begets TR, in long-standing primary TR, signs of volume overload may develop, with RA, RV and TA enlargement. In secondary TR, the mechanism of regurgitation in mainly related to RA or RV dilatation without primary leaflet abnormalities (see Fig. 5).

Atrial and ventricular FTR have different CMR morphological features (Fig. 5). In atrial FTR, there is a severe TA, RA and basal RV dilation. A crucial feature of atrial FTR is the annular dilatation with increase of septum-to-lateral diameter, loss of TA saddle shape and presence of large central coaptation gap of the TV leaflets. This aspect of basal dilatation of RV, without elongation of the lateral wall and the trabeculated apex is called “conical deformation”. Ventricular FTR is usually associated with PH; the mechanism of TR is mainly due to papillary muscles displacement and leaflets tethering secondary to extreme right ventricular apex and lateral wall dilatation, while TA and RA dilation are less evident. In ventricular FTR the morphology of the RV becomes elliptical or spherical with less predominant basal enlargement as seen in atrial FTR [51].

An exhaustive study of RA can be done by prescribing a stack of SAX images that extend over TV plane through the whole RA, as prolongation of SAX stack acquisition of left and RVs [74]. As alternative, RA can be visualized in standard 4 chamber, RV 2 chamber view (inlet view) and inlet/outlet view. RA volume can be calculated both from SAX views or LAX views. However, normal ranges for right atrial volumes vary significantly between methods. The area-length method is used from LAX views, it is faster, but less reproducible than Simpson’s method used with SAX stack.

Normal TA is oval and has a saddle-shaped structure less evident than that of mitral annulus. Given its non-planar structure it is difficult to visualize using CMR, but its maximum and minimum diameter and its motion can be assessed in SSFP cine sequences with good temporal and spatial resolution (Fig. 3). The two main movement of TA are the base-to apex contraction and the sphincteric contraction. During the sphincteric contraction the anterior and posterior part of the TA move toward the interventricular septum causing a 20–30% reduction in size (diameters, perimeter and area). Thus, TA has its maximum size in late diastole and its minimum in late systole. Measuring TA dimensions is possible both in RV LAX and SAX views (Fig. 3). As alternative, free-breathing whole heart b-SSFP sequences can be acquired both in systole and diastole. The dataset obtained can be analysed with multiplanar reformatting during post-processing. These sequences have the limit of long scan times (7–10 minutes) and the lack of dedicated post-processing software that allow 3D evaluation of TA.

CMR study of TV leaflets may be challenging due to several reasons. First of all, the TV morphology has high interindividual variability; leaflets number can vary from 2 to 5 according to Hahn classification, so that visualization and exact recognition of each leaflet in SAX is more complex and time-consuming than with 3D or 2D echocardiography. In RV LAX views only one or two leaflets can be identified and complex morphology (like type V according to Hahn classification) can be easily missed. Moreover, TV leaflets are usually thin, highly mobile and thus more prone to artifacts and limited characterization due to low temporal resolution. The exact mechanism of a primary TR cannot be fully understood with CMR, and integration with echocardiographic information is of crucial importance. Conversely, in FTR leaflets motion is often reduced because of tethering forces and a reliable measurement of leaflets length, tenting area, coaptation depth and regurgitant orifice area may be possible also with CMR. A reduction of slice thickness of SSFP cine sequences to 5–6 mm without interslice gap may allow a better visualization of both TA and TV leaflets.

Despite CMR is considered the gold standard for assessment of RV size and function, it has not been validated for the quantification of TR in clinical practice yet [75]. Nevertheless, in the last years the development and technical improvement of CMR have led to increased interest in quantification of valve diseases using non-echocardiographic methods. Many studies showed that CMR assessment of TR is feasible, but less established than that of other regurgitant valvular lesions. CMR quantification methods of TR may be qualitative or quantitative, direct or indirect (Table 5).

Qualitative assessment. Qualitative assessment of TR by CMR imaging can be done through bSSFP or gradient echo sequences. The regurgitant jet is visualized as an area of local signal void secondary to flow turbulence or acceleration. In a small study by Reddy et al. [76], twenty-nine TR were analysed and the visual assessment of cardiac regurgitant lesion resulted accurate and reproducible with good agreement with standard quantitative assessment. However, a lone visual evaluation is not additive as compared with a more feasible echocardiographic evaluation; moreover, the extent and the intensity of signal voids depend on the sequence used (b-SSFP or gradient echo) and the parameters of the sequence itself (flip angle, echo time, slice thickness, window level). Medvedofsky et al. [77] proposed a semi-quantitative approach to TR quantification based on the measure of the size and signal intensity (SI) of the cross-sectional TR jet area in the RA in SAX SSFP images and showed that the jet area significantly increased concurrently with TR severity assessed by echocardiography.

Quantitative assessment. The indirect method is the most common used for quantitative assessment of TR. TR-RV is calculated as the difference between RVSV calculated from RVSAX stack and Pulmonary flow derived by phase contrast images. In the 2017 JASE-JCMR guidelines cut-off values for TR severity were based on MR classification and severe TR was defined as having a TRF $>$48% [14]. Zhan et al. [78] validated this hypothesis in a population of 547 patients with FTR, using CMR to quantify regurgitant volume and RF. In this study TR-RV was measured as mentioned above, while TRF was calculated by dividing the TR-RV by the RV inflow, which, in the absence of pulmonary regurgitation is the RVSV. Both TRF and TR-RV were associated with increased mortality after adjustment for clinical and imaging covariates, including RVEF. They identified three different risk categories based on TRF and TR-RV: low risk (TR-RV 30 mL, TRF 30%), intermediate risk (TR-RV 30–44 mL, TRF 30–49%) and severe risk (TR-RV $\ge$45 mL, TRF $\ge$50%). The mortality from FTR at 1 year was 15% in those with a TRV of $\ge$45 mL and 14% in patients with a TRF $\ge$50% (HR 2,26 and 2,60 respectively) [78].

In a comparison study of 337 patients, echocardiographic parameters of TR severity had variable accuracy against TR-RV by CMR (AUC for VC of 0.65, AUC for EROA of 0.75 and AUC for TR-RV of 0.72). A multiparametric hierarchal approach resulted in 68% agreement with CMR and 100% agreement when a 1-grade difference in TR severity was considered acceptable [79].

Given the large body of evidence concerning echocardiographic quantification and prognostication of TR, the recent CMR data, despite promising, should be used with caution and need to be validated yet in the field of percutaneous treatment of TR.

Direct quantification with phase contrast sequences of atrio-ventricular valves is less validated than PC imaging of semilunar aortic and pulmonic valves. Semilunar valves are more fixed and planar than atrioventricular valves, so that quantification of anterograde and retrograde flow is more reproducible [80, 81]. In a small study, Jun et al. [82] reported good agreement of tricuspid flow quantification using direct phase contrast method; however, the published data are very few and this technique is not widespread in clinical practice.

New direct quantification of TR using 4D flow techniques showed high concordance with both direct or indirect methods of quantifying regurgitation [83]. The 4D flow CMR consist of a volumetric, isotropic, time-resolved cine sequence that enables three-directional velocity encoding; post-processing analysis allows calculation of forward flow, reverse flow, regurgitation fraction, and peak velocity. Interestingly, in both children and adult studies comparing indirect quantification to 4D flow, the quantification of TR with 4D flow was more accurate than 4D quantification of MR. Severe eccentric jets are more common in MR, increasing the technical difficulty of maintaining an orthogonal plane to the mitral regurgitant jet, while TR jets are more often central and larger, making easier to find the right cut-plane.

Notwithstanding direct interrogation of the regurgitant jet with 4D flow showed high intra and inter-observer consistency and reproducibility, no data are currently available regarding the prognostic implications of 4D flow derived TR-RV and the specific cut-off values.

Patient preparation: no specific medication is recommended; however, AF is common in this population and thus rhythm irregularity and high heart rate can affect scan quality. Beta-blockers must be used i.v. at the time of the scan or started orally days before. Reasonably an HR 100 bpm should be reached, but an optimal rate of 60–70 bpm can allow contemporary evaluation or coronary arteries provided that nitrates are administered, and an arteriosus phase is acquired. High quality ECG monitoring is essential for a good scan, thus ECG tracing needs to be checked during breath-hold to identify possible artefacts and low tracing signal and eventually improved with lead position change or skin brushing.

Scanner: scanner type affects acquisition protocol; if locally available, last generation scanner (at least 64 multidetectors, or dual-source) shall be used to obtain the maximal temporal resolution and to reduce the scan time.

Coverage: the acquisition should cover at least a 12–14 cm of the chest, while in case of extreme right/left chamber enlargement even a 16–18 cm coverage is required. Additional non-gated scan can be acquired to cover the venous system and in particular the position of SVC and IVC.

Slice thickness: $\le$0.75 mm.

Acquisition: a gated acquisition from the tracheal bifurcation to just below the diaphragm needs to be planned from traditional scout acquisition of the thorax. The entire cardiac cycle (0–100% of R-R interval) or at least a 10–90% acquisition to cover end-systolic and end-diastolic phases are the preferred acquisition windows. In case of TTVI planning, the scan mode depends mainly on the scanner characteristics and the coverage required. Prospective or retrospective triggering can be used. In case of multidetectors (256–320 detectors rows) a prospective full-volume-1 beat acquisition can be performed. If a less advanced multidetector scanner is available, only a prospective multibeat or a retrospective acquisition are possible; in this case motion and misregistration artifacts are expected. Dual-source scanners allow for high pitch helical scanning with excellent temporal resolution. Non-gated acquisition of venous accesses (from neck to pelvis) can be obtained right after the gated acquisition of the heart and usually do not need further administrations if adequate volume of contrast has been previously used. Tube voltage and current settings should be adapted to patient’s body habitus; in general, 80–120 kV and 400–550 mA are used.

Contrast administration: iodinated contrast agents with at least 350 mgI/mL are suggested. Test bolus or bolus tracking techniques are in general chosen based on local expertise. In case of test bolus the region of interest (ROI) must be positioned in the centre of the RV in order to obtain adequate contrast in both RV and venous system. The bolus tracking strategy is a simple and reproducible alternative to test-bolus, ideal in case of step-an-shoot approach without the need for extra-contrast dose administration. Type of contrast injection protocol are summarized in Table 5. A single bolus strategy may not be sufficient for adequate image quality in TTVR planning, while biphasic or split (triphasic) bolus are preferred. Single bolus requires more contrast and lower injection velocities and is associated with frequent beam hardening and mixing artifacts in SVC. When a dual flow injection system is available, triphasic mut be preferred over the biphasic protocol. In triphasic bolus, the first injection is contrast, the second a mixture of contrast/saline solution (30/70 in general), followed by a final saline injection. Flow velocity is set at 4 or 5 mL/sec. This protocol is ideal for bi-ventricular opacification and is associated with lower right heart/SVC streak artifacts compared to both single and dual bolus strategy [84, 85, 86, 87, 88].

Reconstruction: we suggest one reconstruction at each 5% of the R-R interval to cover the entire cardiac cycle (0–100% at each 5% of R-R interval).

Similarly to CMR, CCT may be used to assess RV heart; its main weaknesses are the radiation exposure, the iodinated contrast medium administration and the lower temporal resolution. The great advantages of CCT are represented by the low scan time, the excellent spatial resolution, and the feasibility in patients with implantable devices. Studies comparing CCT and CMR showed great agreement in terms of volumes and function [89, 90, 91], hence the same threshold values for right chambers quantification could be considered in case of TV disease, together with their prognostic implications. Normal CCT values of right chambers have been derived from studies on healthy subjects [91, 92].

Most of commercially available softwares allow semi-automated contouring of endocardial borders, similarly to CMR. In presence of isolated TR, CCT can be used to calculate TR-RV by subtracting LVSV to RVSV; however, if a concomitant valve regurgitation is present, CCT loses its theoretical utility in valve disease quantification.

The assessment of the elliptical shape of the TA is challenging both with echocardiography (lower spatial resolution than CT, often poor acoustic windows) and with CMR (axial images or a 3D-whole heart acquisition, but with lower spatial resolution than CT).

On the other side, CT is able to provide high quality images of TA with adequate temporal resolution and excellent spatial resolution throughout the whole cardiac cycle [93]. A SAX plane at the TA plane can be recreated using MPR with a reliable assessment of anteroposterior and septolateral diameters, perimeter and area both in systole and in diastole. TA can be traced manually or using semi-automated softwares. In a comparison study of Praz et al. [94], agreement between TEE and CT for TA sizing resulted superior using semi-automated methods than with direct manual measurements. TA has two main contraction patterns: the sphincteric contraction and the excursion towards the right ventricular apex. Excursion toward the apex can be easily assessed by CMR or echocardiography in 4-chamber view, while sphincteric contraction is best appreciated with CT or 3D TEE echocardiography. A 20 to 30% reduction in size of TA can be seen in systole, so that maximal dimensions are obtained in late-diastole [64]. Normal values of TA are derived by echocardiographic studies using 3D acquisition (end-diastolic maximal circumference and area: 105 $\pm$ 12 mm and 860 $\pm$ 200 mm${}^2$, respectively) [95] and a MPR post-processing of 3D dataset. Patients with moderate-to severe and severe TR showed a mean CT-perimeter of 148 $\pm$ 16 mm and mean CT-area of 1612 $\pm$ 295 mm${}^2$ in a study of 250 patients by Rosendael et al. [96]. In severe TR mean antero-posterior TA diameter was 50.3 $\pm$ 5.2 mm, and 41.2 $\pm$ 5.6 mm, significantly higher than in patients with mild or moderate TR. In another study, even mild TR showed larger TA dimensions than patients without TR [97]. Given the prevalence of TA dilatation in severe TR, accurate TA sizing is pivotal before TV annuloplasty, valve replacement and transcatheter treatment (Fig. 6).

Optimal visualization of TV leaflets in CT requires homogeneous contrast medium opacification around the leaflets with a dedicated contrast medium administration protocol (preferably, a triphasic protocol). Even if with a good spatial resolution, CT may show motion artifacts due to high and/or irregular heart rate and excessive leaflets motion, like in case of primary TR (flail, endocarditis). In ventricular FTR the leaflets are tethered and less mobile than in primary TR, thus motion artifacts are generally reduced. Four-chamber and 2-chamber views of the RV may be used to assess the leaflet tethering and accordingly measure leaflet length, tethering angle, tenting eight and tenting area. Compared to patients with moderate (3+) TR, patients with moderate to severe or severe TR show higher degree of tethering of the anterior and septal tricuspid leaflets, with no differences in terms of tethering angle of posterior leaflet [93]. Interestingly, the tethering height ($>$7.2 mm) of the TV seems to be associated with the recurrence of TR after TV annuloplasty [98]. Tricuspid anatomical regurgitation orifice area (AROA) can be measured with MPR by contouring the tips of TV leaflets in systole in a similar way to mitral valve regurgitation [99]. The AROA was recently calculated using Dual Source CT in 60 patients with symptomatic TR and compared to TR severity and VCA assessed at TEE. The AROA showed good intra and interobserver reliability and excellent linear correlation with 3D VCA and TR severity [100].

Coaptation devices are based on edge-to-edge and edge-spacer-edge valve repair [33, 117]. Patient’s selection for the most commonly available coaptation devices, Triclip (Abbott Vascular, Santa Clara, CA, USA) and Pascal (Edwards Lifesciences, Irvine, CA, USA), rely exclusively on echocardiographic evaluation and CCT is not necessary. However, lead position, leaflets calcifications and the presence of unfavourable angles between IVC and TA can be easily assessed by CCT images during pre-procedural planning. Moreover, in case of inconclusive echocardiographic data, coaptation gap and leaflets length can be measured and TTVI strategy can be changed in case of large coaptation gap or extreme tethering unsuitable for TEER. The Forma spacer device (Edwards Lifesciences, Irvine, CA, USA) is another kind of edge-to spacer device based on a foam-filled polymer balloon that fills the leaflets coaptation gap. The anchoring system is placed in the right ventricular apex, the spacer is placed at the annular level into the coaptation gap, and then the system is then locked in subclavian region. Device sizing is based on coaptation gap and TA dimensions, while the distance between TA and RV apex, as well as between TA and papillary muscles are assessed for device suitability and to avoid anchoring problems (Table 7A). The Forma system requires a large 24F sheath, hence CT scan is useful also to assess adequate dimensions and patency of left subclavian and axillary veins [118].

The goal of annuloplasty devices is to reduce TA area in FTR. Trialing, TriCinch, Cardioband and Traipta systems are some of the annuloplasty devices currently accessible [119, 120, 121, 122] (Table 7A). Right coronary artery (RCA) complications occur in 15% of patients treated with Cardioband [123] and are considered a major worry in transcatheter annuloplasty. Cardioband anchors are placed in the periannular tissue, and the risk of coronary injury depends on the course and the distance of the RCA around the TA (Fig. 7). The use of a triphasic contrast medium administration protocol may allow to achieve a good visualization of right heart and coronary arteries at the same time. In a study of 250 patients with TR who underwent CT evaluation, the course of RCA had 3 main configurations: along the TA (65% of patients), superior to TA (10%) and crossing the TV (25%). Distance between RCA and TA was measured in mid-diastole, using either a SAX view of the TA in case of RCA running at the same level or a long axis view in case of superior or inferior course. The authors suggest that a maximal distance between the anterior or posterior part of the TA (at the level of the anterior or posterior leaflet insertion) and the proximal or distal RCA of 2 mm, may be associated with high risk of RCA impingement [111]. In the TriCinch device, a corkscrew delivery system is advanced through femoral access to the target region of TA, where the coil is implanted. The target zone is the antero-posterior commissure of TA (at 9–10 o’clock with a ventricular en-face view of TV); at this level, the distance between TA and RCA must be assessed. Among the first 18 patients implanted, RCA injury occurred in one patient. Instead, the goal of the Trialign system is the bicuspidalization of TV, obtained by positioning two pledgets in the septo-posterior and antero-posterior commissures; the two pledgets are then approximated to reduce annular size and regurgitant orifice area. In the first trial, RCA damage requiring stenting occurred in about 6% of cases, with the proximal RCA near to the antero-posterior commissure being the area at higher risk. Finally, the Transatrial intrapericardial tricuspid annuloplasty (TRAIPTA) consists of an indirect annuloplasty repair thorough a loop delivered along the atrioventricular groove within the pericardial space, which is reached by puncturing of right atrial appendage (RAA). Relationship between RAA and the surrounding structures must be evaluated during pre-operative CT planning.

An alternative approach to direct intervention to TV is the heterotopic implantation of valves in the caval veins. The aim of this approach is to reduce the systemic congestion that characterize patients with severe untreatable TR and right HF. After a first attempt to adapt transcatheter aortic valve prostheses to cava veins, specific devices have been developed.

The TricValve system (P&F Products and Features, Vienna, Austria) consists of two dedicated self-expanding valves, implanted in the superior and inferior cava veins; the Tricento device (NVT AG, Muri, Switzerland) consist of a bicavally anchored covered stent with a lateral bicuspid porcine valve, implanted from top (SVC) to IVC. Role of CT is crucial to assess IVC and SVC dimensions. The SVC must be measured at the level of the innominate vein confluence, of the pulmonary artery and of SVC-RA junction; the IVC must be measured at the level of IVC-RA junction, at the confluence of the hepatic veins and at 5 cm below IVC-RA shift. To avoid hepatic vein obstruction, a distance $>$10 mm from the RA-IVC junction must be ensured [124, 125] (Table 7B).

The complex anatomy of TV and the great variety of mechanisms leading to TR make TTVI challenging, and many patients result non suitable for TEER of the TV. Theoretically a transcatheter tricuspid valve prosthesis should be ideal for this non-operable subset of patients; however, many technical issues still remain to be solved. A first limitation is based on the large TA dimensions that require larger prostheses than usual, causing problems in venous access routes and delivery systems management. Secondly, the TV is usually a non-calcified valve with thin and degenerated leaflets, that reduce prosthetic valve stability. Third, TA is very mobile, has very wide systolic excursions and it has extremely variable size during the cardiac cycle, complicating accurate prosthetic sizing. Moreover, the septal leaflet is part of the Koch triangle that is in close relationship with the atrioventricular node, thus a prosthesis implantation may lead to permanent conduction disturbance and the need for a pacemaker implantation.

A great variety of prostheses has been developed for the TV, with a great variability of pre-procedural planning CT protocols. In most cases each device has a specific acquisition protocol, and the images are then analysed by the manufacturer’s Core-lab.

Assessment of TA dimensions, valvular and subvalvular morphology and vascular access routes, are only some of the aspects of the pre-procedural CT imaging and prosthesis sizing. The MPR of CT 3D dataset allows a careful measurement of distance between the TA and the subvalvular apparatus (moderator band, papillary muscles, trabeculae) and between the TA and the apex, and the TA-RVOT angle and distance (Table 7B). Given the large variety in shape of the new TV prostheses, all these parameters must to be taken into account to choose the optimal implantation depth and to reduce the risk of procedural complications (for example RVOT obstruction) [126, 127, 128].

Strength and weakness of each imaging modality are summarized in Table 8.