Keywords
contrast-enhanced ultrasound - guidelines - quantification - therapy monitoring -
tumor perfusion
Introduction
Dynamic contrast-enhanced ultrasound (DCE-US) is a technique to quantify tissue perfusion
down to the capillary level based on phase-specific enhancement after injection of
microbubble contrast agents for diagnostic ultrasound. In addition, the quantitative
analysis of the dynamics of contrast enhancement overcomes its subjective comparison
between normal and abnormal parenchyma, or between a focal lesion and the surrounding
tissue.
The guidelines of the European Federation of Societies for Ultrasound in Medicine
and Biology (EFSUMB) published in 2004 [1] and updated in 2008 [2], 2012 [3], 2013 [4]
[5], and 2020 [6]
[7] focused on the use of contrast-enhanced ultrasound (CEUS), including essential technical
requirements, training, investigational procedures and steps, guidance regarding image
interpretation, established and recommended clinical indications, and safety considerations.
However, the quantification of phase-specific tissue enhancement acquired with ultrasound
contrast agents (UCAs) is not discussed. The basis for the standardization of DCE-US
has been established and published by the EFSUMB introductory paper in 2012 [3]. It provided some recommendations and descriptions of the quantification of DCE-US
images, and technical explanations for the analysis of time-intensity curves (TICs).
As part of the development of professional standards for diagnostic ultrasound techniques
[8] and in accordance with the regulations for EFSUMB policy documents published in
2019 [9], the current update was prepared on the basis of an up-to-date literature search.
It includes clinical aspects for data collection, analysis, and interpretation in
the quantification of tumor perfusion, which are derived from recent studies. This
study focuses on the clinical assessment in oncology, but the basic considerations
are generally transferable to other DCE-US indications such as treatment monitoring
in inflammatory bowel disease or chronic kidney disease. The current study not only
aims to support future work in this research field but also to facilitate a transition
to clinical routine use of DCE-US.
Why do we need quantification?
Why do we need quantification?
Quantification of CEUS is needed to evaluate data objectively, to enable comparison
of imaging techniques, to evaluate new UCA applications, to quantify tissue and tumor
enhancement in order to characterize focal lesions, to evaluate therapeutic response,
and to limit variability in clinical diagnosis [3]. Tissue perfusion is a relevant functional imaging parameter with pathophysiological
and clinical relevance in different clinical settings and can be assessed with different
imaging techniques, e.g., brain perfusion in stroke imaging using magnetic resonance
imaging (diffusion) or dynamic contrast-enhanced computed tomography or myocardial
perfusion using dynamic contrast-enhanced echocardiography for the heart.
An objective and quantitative diagnosis of perfusion characteristics is of particular
relevance in the follow-up of cancer patients but can also be used for the diagnostic
assessment of other pathological changes associated with alterations in tissue perfusion.
This applies, for example, to the noninvasive diagnosis of the progression of parenchymal
liver disease, liver cirrhosis, and portal hypertension [10]
[11]
[12]
[13]
[14]
[15]
[16] and for the noninvasive evaluation of chronic kidney disease [17]
[18]
[19] and subclinical kidney transplant rejection [17]
[20]
[21]
[22]
[23]
[24]. There are partially contradictory data regarding the evaluation of inflammatory
activity and response to biologic therapy in inflammatory bowel disease [25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]. A relatively new field of research is the application of DCE-US for the differential
diagnosis, grading of the biological behavior, and outcome assessment of malignant
tumors [35]
[36]
[37]
[38]
[39]
[40]
[41]. This position paper is focused on the assessment of tumor perfusion.
DCE-US as a dynamic examination is based on relatively long video sequences that measure
changes in contrast signal over time from the bolus transit in the body. For precise
diagnostic evaluation, such data need to be analyzed to extract biomarkers and other
parameters that are related to relevant physiologic and patho-physiologic properties
and presented in a form that is compatible with the imaging process (e.g., color coded
maps). It may be anticipated that such quantitative measures may play a major role
in big data analysis and the development of machine learning, which itself may influence
diagnostic approaches. Thus DCE-US has the potential to strengthen the role of CEUS
in future diagnosis and follow-up [42]
[43].
Current assessment of response to cancer treatment is still mainly based on interval
evaluation of the tumor size according to the Response Evaluation Criteria In Solid
Tumors (RECIST) [44]. Unfortunately, RECIST only reflects tumor size changes (which are often delayed,
if they occur at all) and is unable to identify non-responders at an early time-point,
when novel cytostatic biologic agents are employed [45]. A patient may be misclassified as a non-responder because the tumor size remains
unchanged, or even increases in the early stages of treatment due to hemorrhage, necrosis,
or edema, in spite of a decrease of the viable tumor. To add functional assessment,
new methods that also reflect tumor perfusion have been introduced in the form of
modified RECIST (mRECIST) criteria [46]. This has highlighted the need for alternative accurate and reproducible quantitative
techniques to assess changes in tumor vascularity, a question which is not addressed
satisfactorily by current standard diagnostic evaluation.
Clinical Applications
DCE-US quantification has been used to monitor changes induced by anti-angiogenic
[47]
[48] and anti-inflammatory [49]
[50]
[51]
[52]
[53] therapies, both as a potential marker of response and as a tool to enable dose optimization
of therapy in individual patients [54]. Early clinical trials assessing tumor response in gastrointestinal stromal tumor
(GIST) were based on the subjective and qualitative assessment of enhancement dynamics.
Subsequent studies assessed response in renal cell carcinoma, hepatocellular carcinoma
(HCC), breast cancer, pancreatic cancer, and colorectal metastases using semi-quantitative
techniques [55]
[56]
[57]
[58]
[59]
[60]
[61]
[62]
[63]
[64]
[65]. Additional studies [50]
[51]
[56] used quantitative techniques to derive parameters related to the time course of
contrast enhancement, in comparison to clinical endpoints such as Progression Free
Survival (PFS) and Overall Survival (OS) following anti-angiogenic treatment. Techniques
such as respiratory gating [59]
[66]
[67] and motion correction have been shown to improve the reproducibility of DCE-US measurements.
A number of clinical trials have since evaluated DCE-US in therapy monitoring or intervention
guidance, also demonstrating the potential of this technique in comparison to other
imaging techniques such as DCE-MRI [53]
[68]
[69]
[70]
[71]
[72]
[73]
[74]
[75]
[76], CT perfusion [77], or positron emission tomography [78]. Preliminary results have also been reported in children [79].
The number of clinical studies on DCE-US has increased since the initial publication
in 2012, as well as the variety of technical approaches used to acquire and analyze
contrast enhancement dynamics. The selection of these techniques may influence the
reliability of reported results and possibly explain contrasting observations between
studies. The following sections attempt to explain the available DCE-US techniques
and parameters, with the aim of establishing a more standard approach to DCE-US examinations.
General considerations
Clinical DCE-US is usually performed with pure blood pool agents, such as SonoVue/Lumason
[sulfur hexafluoride with a phospholipid shell, Bracco spa, Milan, Italy], or Definity
[Octafluoropropane with a phospholipid shell, Lantheus Medical Imaging, Billerica
MA, USA]. Quantitative contrast techniques can also be applied to agents, which are
targeted to accumulate through specific biological interactions or to be extracted
by a specific process (such as phagocytosis), but they require more complex multi-compartment
kinetic models and are beyond the scope of this paper [43].
DCE-US can be performed using two different administration methods, an intravenous
bolus injection or an infusion of UCA. The latter is followed by a disruption-replenishment
technique and is much less commonly used for the assessment of tumor perfusion than
the bolus injection.
The dual blood supply of the liver complicates blood flow quantification. After a
bolus injection, the arterial blood supply is responsible for the initial enhancement
of the normal parenchyma and of focal lesions, as the microbubbles arriving through
the portal blood supply are delayed by 5 to 10 seconds. With the infusion technique,
the replenishment reflects a combination of arterial and portal flow inputs.
After a bolus injection of UCA with wash-in/wash-out (bolus-transit) analysis, single-plane
imaging at a low mechanical index (MI) is usually performed at about 10 frames per
second for the duration of the enhancement. Frame rates that are too high should be
avoided to prevent bubble destruction. Three-dimensional acquisition (corresponding
to a volume) rather than a single plane would be preferable to overcome some limitations
related to single plane analysis, but it is currently not feasible with the currently
available commercial hardware (in terms of transducers and computing speed of available
equipment). The average CEUS signal intensity within a region of interest (ROI) is
calculated in linear units and is displayed as a function of time, i.e., a time-intensity
curve (TIC), which describes the phases of progressive increase in enhancement of
the contrast agent in the ROI (also termed wash-in) and the subsequent phase of slow
decrease in contrast signal intensity (termed wash-out phase). Additional ROIs can
be placed in a reference tissue for comparison purposes or in different areas of the
lesion.
UCA administration
The approved doses are for bolus injection of SonoVue 2.4 mL for examinations of the
macro- or microvasculature, and 2 mL SonoVue or Definity (10 μL/kg) in echocardiography.
However, this dose may be reduced to 0.6–1.5 mL in most ultrasound systems, or increased
up to the dose of two bolus injections (4.8 mL of SonoVue) under certain conditions
depending on the sensitivity of the equipment, the transducer type and central frequency,
the degree of vascularity, and the depth of the target lesion [80]. With more recent and sensitive equipment, the lower doses are adequate and should
be preferred except when high-frequency transducers are used. For example, the dose
may be reduced to 1 mL when scanning the kidneys (and particularly in renal transplants),
while it can be increased to 4.8 mL in the case of a superficial lesion using a high-frequency
linear array or endoscopic transducers [81]
[82]
[83]. All microbubbles tend to go up in saline and should be shaken from time to time
– or a pump should be used.
The bolus injection in general and also for quantitative DCE-US using SonoVue should
be quick and be performed with a short angio-catheter typically 20G (never a smaller
diameter than 22G to avoid disruption of the contrast microbubbles when they cross
a too narrow catheter lumen), placed in an antecubital vein, without using a long
extension line. A 3-way stop valve may be used at the end of the catheter to allow
controlled access. In this case, it is preferable to connect the contrast syringe
to the lock directly in line with the intravenous tract (not the perpendicular one)
to avoid microbubble disruption that could occur when injecting contrast bolus against
the stop valve tube wall. A saline flush (5 mL) should immediately follow to further
sharpen the injected bolus and to limit the volume of UCA remaining in the angio-catheter
and stop valve.
For infusion studies up to 2 vials (9.6 mL) have been infused at a rate of about 1
mL/min (or less) depending on the enhancement level required [84]. Slow infusion requires either a drip bag that is gently shaken from time to time
or a pump that can be placed vertically or a specific rotating pump to continuously
agitate the microbubbles [84]. Analysis should be performed during a steady UCA concentration in the blood. An
acceptable steady state situation is usually achieved after about 2 minutes of infusion
depending on the infusion rate. An initial faster injection rate can be used to achieve
steady state earlier.
CEUS time intensity curve parameters
CEUS time intensity curve parameters
CEUS time intensity curve parameters have been summarized in the EFSUMB position paper
describing the bolus-transit of the contrast microbubbles in the ROI [85]
[86]. Time-related parameters can be differentiated from signal intensity-related parameters
[3]. Several derived TIC parameters are purely descriptive/empirical. Reliability and
potential sources of errors have been described [80].
Time-to-peak (TP), rise time (RT), mean transit time (MTT), peak intensity (PI), and
area under the curve (AUC) have been proposed as primary parameters and all others
are derived from those parameters [87]. In the EFSUMB position paper parameters such as time zero offset (T0), time-to-peak (TP), wash-in time (WIT), wash-out time (WOT), mean transit time (MTT),
full width half max (FWHM) are explained in detail [3]. Different from the other parameters, MTT can be calculated only in combination
with a fitted mathematical model, while the other parameters are curve-descriptive
parameters and thus can be derived also without a dedicated model. Since it is assumed
that the signal intensity in DCE-US is proportional to the number of microbubbles
(see below, linearized image data), and the microbubbles remain strictly intravascular,
the TIC parameters are related to the vascularization of the analyzed region. Some
signal-related parameters (peak intensity, area under the curve) are more correlated
to the local blood volume of the region (~ mL), while other time-related parameters
are more reflective of blood flow (TTP, WIT, AUC is also related to blood flow according
to the Steward-Hamilton relationship). All time and intensity values should be calculated
from a curve fitted to the linearized echo intensity values and not from image data.
Signal intensity-related parameters are given in arbitrary units [a.u.], with the
most important being peak intensity (PI) and area under the curve (AUC). Both are
described in detail in the already mentioned paper. The whole AUC describing the area
under the curve may be divided into two components: the AUC of the wash-in phase up
to peak intensity PI (WIAUC) and of the wash-out from peak intensity until the predefined
time of end (WOAUC). The total AUC is the sum of WIAUC + WOAUC.
Other parameters include the wash-in rate (WIR) [a.u./s], which describes the slope
of the TIC curve during wash-in [signal intensity/s]. The maximum slope of the TIC curve or the mean slope of a certain wash-in time interval (e.g., from 5% to 95% signal intensity)
is used for this empirical parameter that is related to the blood flow. In a similar
way, change during wash-out (WOR) can also be derived. In addition, combinations of
the above parameters exist, in particular ratios between a signal intensity and a
time-related parameter such as the wash-in perfusion index WIPI, which is the wash-in
AUC divided by the wash-in time (WIAUC/WIT).
Refilling kinetics describe the replenishment of microbubbles during the infusion
of UCA. UCA is first imaged without being disrupted at a low MI, then a few frames
are acquired at a high MI (often at the highest available) causing bubble disruption
in the image plane. Immediately thereafter, the MI is reverted to its low setting
and the arrival of fresh microbubbles is imaged. Refilling kinetics are described
by parameters that are different from those after bolus injection. T0 has an identical
definition as for the TIC curves after bolus injection. TP, WIT, and MTT can also
be calculated using a mathematical model that describes the refilling process. In
contrast to bolus injection TIC curves, the maximum signal here is no longer reached
at a peak but rather in the plateau phase. Ip is the maximum signal reached at the plateau (complete replenishment), often also called A, and is proportional
to the local blood volume. The rise of the replenishment curve [1/s], often called B or β (based on the model by Wei et al., see below), is a parameter that is proportional to the local blood flow velocity.
Since the replenishment curve usually has a sigmoidal shape, this parameter varies
with time, and its concrete definition depends on the model used. In principle, A
and B are parameters of the replenishment curves that are directly related to the
blood volume and flow velocity (and its product is directly related to the blood flow,
F=A*B). They can be extracted from the curves even without using a specific mathematical
model that requires a closed form analytical expression, and thus they can easily
be calculated and may be less prone to model-dependent limitations.
In 1998, Wei et al. [85] were the first to introduce the disruption-replenishment method and the development
of the mono-exponential model. Krix et al.
[88]
[89]
[90] used a similar approach as Wei et al. However, the modified formulas were no longer based on empiric assumptions and were
based on a multi-vessel model incorporating differences in the acoustic field properties
when using high- and low-MI imaging. This model was found to be at least equivalent
to the mono-exponential model, but it is nevertheless used much less frequently. Wei’s
model was improved by Arditi’s model [91], which was subsequently further improved by Hudson et al.
[92]. This model has 3 components that were not present in Wei’s model: accounts for
tissue perfusion through realistic microvascular geometry (Lognormal perfusion model),
considers the ultrasound field properties of the destruction beam, and also considers
the ultrasound imaging field. With the Arditi-Hudson model, it is possible to calculate
the relative mean flow rate.
For repeated DCE-US exams, identical contrast protocols, DCE-US parameters, and analysis
models have to be used in order to facilitate inter- or intra-patient comparison.
Standardization and harmonization of software-based solutions and the various solutions
integrated in the US platform are desirable but don’t exist yet. For a detailed description
of the equipment settings and patient-based factors, we refer to the published position
papers [3]. Most studies focusing on AUC and wash-out recorded 3-minute loops [68]. In studies using infusion of UCA and the destruction-replenishment protocol, a
shorter loop of the replenishment of the lesion or organ is sufficient (15–60 s) with
the option to repeat it.
Clinical aspects of a DCE-US protocol
Clinical aspects of a DCE-US protocol
Choice of DCE-US parameters
A key question is which DCE-US technique and parameter should be used and evaluated
in the various clinical settings. As described above, some parameters are more related
to the blood volume (like the peak intensity Ip or the plateau A in replenishment
kinetics) while others are more related to the dynamics of the blood supply, the blood
flow (like the MTT). This is a first relevant aspect when choosing a certain DCE-US
parameter. Furthermore, like with other imaging methods that analyze tissue vascularization
(e.g., CT or MRI perfusion imaging in stroke) also a set of parameters and the identification
of a potential mismatch between them may be useful to evaluate. In oncology treatment,
monitoring or even outcome prediction are key aspects for use of DCE-US. This means
parameters that could allow early assessment or prediction of treatment success or
failure are the candidates of choice. In theory, changes in vascular dynamics (blood
flow) would occur before a change in the vascular morphology (blood volume) becomes
evident, but this has not yet been clearly demonstrated with DCE-US. Still, a widely
used approach in research projects using DCE-US is to calculate more or less all feasible
parameters and then to analyze if there is a correlation between these parameters
and the specific clinical efficacy/outcome parameters. A few studies have suggested
a certain parameter to be preferable in a specific setting (e.g., the MTT in bevacizumab
therapy of metastases [53]) but a general broad consensus is lacking. AUC may be the most robust parameter
in terms of technical errors.
DCE-US study design
Future studies should report DCE-US results in a more specific manner, related to
certain parameters. Results should also be set in the context of the concrete tumor
and treatment being assessed. “DCE-US for chemotherapy monitoring” may be a too broad
and unspecific term. It should be clarified to which specific treatment or drug group
study results are reported. In general, confirmatory studies are still needed to determine
the crucial DCE-US parameters that should be focused on for the various clinical scenarios.
It means an a priori hypothesis is to be proven in a prospective multicenter approach–
such as “change of AUC tumor/AUC liver at time point x compared to baseline provides
decisive information for therapy management with drug xy”. So, a very narrow study hypothesis focusing on concrete parameters, time points,
etc. and using a valuable clinical endpoint should be applied. Several studies so
far have been explorative and have only used another biomarker for comparison such
as perfusion in MRI or microvessel density in pathology.
Here, a comparison between classic early RECIST and DCE-US results is per se of no
or low additional clinical value. Studies should rather focus on the potential additive
value of DCE-US compared to standard diagnostics, i.e., on the predictive value of
the method at early time points. The use of long-term outcome data as the standard
of reference should be preferred to demonstrate whether DCE-US performs better at
follow-up compared to RECIST.
DCE-US exam time points
This is related to the question at which clinical time point(s) a DCE-US examination
should be performed. Treatment monitoring requires follow-up examinations while predictive
messages or data for guidance of interventions can be derived from a single, early
exam. Clinical trials using DCE-US in monitoring have often focused on standard time
points, i.e., before the start of treatment and follow-up exams performed at standard
time points in parallel to established imaging (e.g., for RECIST). Early time points
for follow-up sometimes have been added, in particular a DCE-US exam after a first
cycle of chemotherapy. Currently, further studies are still needed to determine the
optimum monitoring regime for a specific treatment. Not only the duration after the
general start of treatment can be relevant, but also the duration after a certain
cycle of chemotherapy can have a considerable impact. Anti-angiogenic effects may
be observable already within a short period of time, maybe the optimum only within
a specific period of time after administration. DCE-US should clearly report how the
used exam time points have been chosen and further studies can increase knowledge
for optimization of monitoring schemes.
How to perform DCE-US, how to interpret the results, technical advice
How to perform DCE-US, how to interpret the results, technical advice
To optimize the machine settings for DCE-CEUS, the following issues are important.
One single focus position should be set in a deeper region of the scanning plane,
which must include by large all the regions of interest. The lowest but still reliable
mechanical index (MI) should be used to avoid any unnecessary bubble disruption. The
most convenient MI value varies depending on the specific equipment. The receive gain
should be set so that it is usually aligned in the middle position. The persistence
mode should be turned off and the dynamic range should be kept tendentially high despite
the fact that these two adjustments may not provide the best ultrasound images.
The conditions of the patient and surrounding factors (including posture, resting
time, heart rate, blood pressure) and also of the scanning plane (acoustic window,
probe position) at each acquisition, to help explain discrepancies in unexpected findings
taken at different sessions during follow-up should be standardized and recorded.
For adequate reproducibility, the follow-up examinations require the scanning plane
to be exactly the same. This is often very difficult to achieve, even for expert users.
A clear description of the probe position for examining the lesion, with landmarks
in relation to the skin surface and documentation of representative anatomical structures,
e.g., liver segment(s), major vessels, as well as the CEUS acquisition parameters,
such as the depth of the lesion, mechanical index, etc., are essential to ensure standardization
of these subsequent studies. It is important to keep all imaging (machine) parameters
unchanged after the baseline scan to allow the comparison of the effects of therapy
in subsequent scans.
Find a tumor in conventional B-mode and choose a tumor plane to study. Inject the
appropriate dosage of microbubble contrast agents and scan in contrast mode (side-by-side).
In order to keep the probe stationary, be aware of and compensate for any motion.
Some examiners have also used an articulated arm to stay on the same plane. Scan continuously
for 2 up to 5 minutes (depending on the clinical application) avoiding bubble destruction.
Commercially available software (e.g., Vuebox) also allows the merging of smaller
videos into one video, which can be helpful in the case of motion but also to reduce
bubble destruction. Save the DICOM video loop in a format that allows data linearization.
If more than one TIC curve may be recorded, then rotate the probe to select a different
tumor plane (to evaluate tumor heterogeneity) and repeat the steps above for both
infusion and replenishment.
The data analysis involves the use of a software package that allows forming of the
TIC from linearized data from ROIs in the lesion and one in the normal parenchyma.
One ROI should cover the whole tumor, and the placement of optional additional ROI(s)
should follow representative areas of the “whole tumor” guided by highly vascularized
parts of the tumor. For early relapse prediction, focusing on highly vascularized
ROIs may be useful. In partially necrotic tumors, this guidance can make an important
difference. For some of the mentioned recommendations, no consensus has been reached
so far.
Next, a curve is fitted to the TIC data and the important perfusion parameters (rise
time RT, mean transit time MTT, peak intensity PI, and area under the curve AUC) are
extracted. Interpretation of the results involves statistically correlating the perfusion
parameters with physiological data and clinical outcomes.
Further technical and methodological aspects
Technical considerations also contribute to the choice of an optimum DCE-US parameter.
Reproducibility is an important factor, and DCE-US exams can be influenced by various
aspects. Thus, the most robust parameters can be preferable. Time-related parameters
(rise time) are robust since they do not depend on the acoustic signal level – if
the bolus arrival time is subtracted to avoid circulation time dependencies. Integrals
are per se more robust than single values, thus the AUC or also parameters based on
a mathematical integral and a closed form analytical expression (MTT) can be beneficial
in the clinical routine, but this has been a topic of controversial discussion between
the authors. However, the quality of the fit must be recorded to avoid misinterpretation.
Even then, all parameters related to the CEUS signal intensity can crucially be influenced
and biased by the various acoustic and patient conditions which may drastically limit
inter- or longitudinal intra-patient comparison. This is mainly related to signal-related
parameters. Normalization is the key to reducing this variability. Instead of using
parameters of a single ROI (in oncology usually this is the tumor), values obtained
from this ROI should be normalized, usually placing a second ROI in normal tissue
adjacent to the relevant tissue (for instance the liver) and calculating the ratio
of the parameters in these ROIs. Such normalized parameters are less prone to external
bias. Time-related parameters are less influenced but also these parameters of a tumor
can be compared with the surrounding tissue, e.g., as the difference in the rise time
in the tumor compared to the liver.
The choice of the US plane may be affected by the visibility of representative tissue.
The second ROI should be placed at the same depth. If no normal (healthy) parenchyma
is present, other normal organs visible in the US plane could be used as an exception.
Due to signal linearization, large vessels should not be used as the standard ROI
for comparison. The focus should be positioned just at the level of the target lesion
for most ultrasound scanners. Deeper focal zones might be used to achieve a more uniform
acoustic field, which improves sensitivity to the agents and lessens the risk of bubble
disruption. Detailed general technical recommendations have been published elsewhere
in a consensus paper [93].
Perspectives
Modern oncologic therapies not only aim at a decrease in tumor size but may also focus
on a “return to normal” situation, i.e., a tumor then may still have a high but relatively
normal blood volume or perfusion. Thus, even more sophisticated DCE-US parameters
beyond those related to blood volume or perfusion could be needed to describe DCE-US
patterns correlated with the vascular architecture. Existing models are able to derive
such additional information, but they are not used in clinical practice. Finally,
the described DCE-US parameters provide a temporal analysis of DCE-US exams, not a
spatial analysis. Vessel architecture analysis, however, also requires a spatial component.
Placing more than one ROI in a tumor, e.g., in the periphery and the center, is the
simplest approach to add a spatial analysis. When color-coded parameter maps are generated
with suitable software, more complex approaches are feasible, up to a pixel-wise comparison
and correlation of DCE-US parameters. A simple spatial approach is also to use the
size of the colored area in a tumor as an additional parameter – it is not pure DCE-US
but DCE-US is used here to create such spatial parameters. For instance, the size
of the AUC above a certain threshold/size of the whole tumor can reflect the vascularization
– similar to “% of vascularized tumor”. A combination of both a spatiotemporal analysis
and the use of 3D-US for DCE-US may provide a more complete description of the UCA
transport process and better characterization of perfusion, contrast dispersion, and
vascular architecture [43]. In brain perfusion studies using MRI, such parameters are relevant – e.g., to identify
a mismatch between perfusion parameters and to see if there is viable tissue at risk
that justifies treatment after stroke. In regard to 3D DCE-US, CEUS techniques are
limited and publications are lacking [94], and these topics are beyond the scope of this document. Variability studies using
phantoms and models across multiple scanners and quantification software have been
described in detail; refer to [80]
[87].
Although promising, all studies identified so far on AI in CEUS are single-center,
retrospective studies, or studies on limited, selected case series using different
algorithms for machine learning and with various clinical aims, even if characterization
of liver lesions is the most frequent. Most often the algorithms are run in post-processing,
making them less useful in a clinical workflow. There is a need to perform prospective,
multi-center studies with clinically useful endpoints, preferably using open-access
software in order to find the place for AI in the evaluation of CEUS cine loops.
Open questions
TIC curve analysis of CEUS bolus injections provides several parameters that reflect
local blood flow. None of the parameters alone represent clear-cut tissue characterizing
abilities, although differences are observed for e.g., neoplastic and non-neoplastic
tissue [37]
[95]
[96]
[97]
[98]
[99]. The ability to combine several parameters simultaneously using AI may provide improved
characterization, but this must be shown in prospective multi-center studies using
standardized technology. The dynamic contrast assessment methods need to be integrated
in the clinical workflow, not requiring too much time, and the results should provide
information influencing the clinical management of patients.
Concluding remarks
Results of recent monocentric and multicentric clinical trials propose that quantitative
DCE-US may be useful in oncology, in particular in the assessment of response to targeted
therapies beyond classic RECIST assessment. The current article provides general information
about the technique and parameters utilized in DCE-US quantification and recommendations
on its use to provide a standardized approach, which may improve clinical management.
Statements
Compared to a purely subjective comparison of the phase-specific enhancement of different
tissues or of the same tissue under different pathological or therapeutic conditions,
DCE-US allows a more objective assessment when used in a standardized way.
Using only tumor diameter changes (i.e., RECIST) is a suboptimal method for tumor
response assessment. Treatment monitoring assessment of vascularization/perfusion
adds relevant information both in the early and later phases after initiation of pharmacological
treatments.
Further research is recommended to investigate the potential of DCE-US to noninvasively
improve the differential diagnosis of focal lesions in parenchymal organs, to graduate
the biological aggressiveness of various malignant tumors, and to predict their outcome,
as well as to record the temporal dynamics of pathological processes in parenchymal
organs associated with changes in perfusion characteristics.
DCE-US provides quantitative information about local blood flow and can be carried
out with two main DCE-US modalities, which provide different information and parameters:
the bolus technique and the infusion technique (using the disruption-replenishment
method).
The bolus technique quantifies the entire course of contrast kinetics, from wash-in
to wash-out. The analysis is carried out along one single plane for each injection
and a cineloop of at least one minute in duration is recommended. The disruption-replenishment
method is carried out at a steady-state high signal enhancement level. The analysis
requires a shorter cineloop (usually 10–25 seconds), so that multiple planes can be
assessed. Parameters and information obtained with the two methods differ from each
other.
Relative quantification of perfusion using a reference area at the same depth should
be preferred to absolute evaluation of contrast enhancement.
In order to optimize machine settings for DCE-CEUS, the following recommendations
are important: a) use a single focus position to be set in a deeper region of the
scanning plane that must include all regions of interest; b) use a low mechanical
index (MI); c) set the receive gain high with TGC usually aligned in the middle position;
d) turn off the persistence mode and keep the dynamic range tendentially high despite
the fact that these two adjustments may not provide the best B-mode ultrasound images.
The MI should be set as low as possible, with the goal of avoiding any unnecessary
bubble disruption. The most convenient MI value varies depending on the specific equipment
and the contrast agent being used.
To assess tumor response in a patient, the same machine settings should be used for
consecutive DCE-US examinations as for the baseline examination. It is recommended
to keep a detailed record of patient conditions and surrounding factors (including
posture, resting time, heart rate, blood pressure) and also the scanning plane (acoustic
window, probe position) for each acquisition, to help explain discrepancies in unexpected
findings taken in different sessions during follow-up.
Suitable planning and choice of a representative imaging plane is crucial to avoid
respiratory motion of the ROI which is a major source of error in the quantification
of DCE-US. Especially out-of-plane motion cannot be corrected, and out-of-plane acquisitions
must be excluded from the DCE-US analysis, which is a time-consuming and demanding
process.
Quantification software may be embedded in ultrasound equipment or may be work off-line
on separate hardware. It is necessary to perform calculations on linearized data to
maintain the linear relationship between microbubble concentration and signal intensity.
