Introduction

Indeterminate renal masses are a common clinical problem with an estimated more than half of patients over the age of 50 years having at least one renal mass [1]. Renal masses are often found incidentally during an imaging examination [2]. Many renal masses are benign simple cysts, which can be diagnosed confidently with cross-sectional imaging alone [1]. Other than inflammatory masses, vascular abnormalities, and pseudotumors, most solid, contrast enhancing renal masses are malignant [3]. Complicated cystic lesions with enhancing components, excluding the actual cyst wall, are often malignant [4], however the new Bosniak classification considers thin < 2 mm septations as benign [5] and enhancement of fine septations without nodularity is a benign finding in CEUS [1]. These complicated cystic lesions are classified using the Bosniak classification system [3,4,5,6,7], originally described for computed tomography (CT) and often applied to other modalities. Because of the high spatial resolution of ultrasound, thin septae or small masses can be seen which may be missed in CT or MRI imaging [1, 8]. With subtle septations or nodular masses, determining enhancement with CT or magnetic resonance (MR) imaging may be difficult because of patient size, patient movement, mass size, technique used, image noise, partial-volume averaging, inability to use contrast agents and pseudo-enhancement. Therefore, classification of indeterminate renal masses at CT or MR imaging may be difficult. CT should not be the first line imaging modality to assess cystic lesions with most probably benign character, due to the related radiation exposure, as pointed out in a recent white paper by the FDA [9]. The current widespread use of CT for assessment of cystic renal lesions is due to its established role and the extensive amount of published literature. The aim of this paper is to review and summarize the literature for Contrast-enhanced ultrasound (CEUS) as a safe and radiation-free alternative.

Ultrasound contrast agents are not nephrotoxic or hepatotoxic. They are purely intravascular and do not enter the urinary compartment [8]. Elimination takes place by exhalation via the lungs, with metabolism of the lipoprotein shell by the reticuloendothelial system similar to other circulating phospholipids and proteins [10]. They have a short half-life (approximately 5 min), which allows for multiple injections in one session. The contrast-enhanced US (CEUS) software allows the operator to subtract the background soft tissue signal to visualize a contrast enhancement–only image, and a narrow US beam width allows for improved visualization of vascularity in small structures, such as septations or small mural nodules. The vascular enhancement patterns of contrast-enhanced US in the kidney are similar to those of CT and MR imaging in the arterial phase but may be different in delayed phases, as contrast-enhanced US agents do not extravasate [8]. CEUS has been used to evaluate renal abnormalities [1, 11,12,13,14,15,16,17,18,19,20,21]. The enhancement of renal neoplasms differs from that in the surrounding renal cortex, with a difference in degree of contrast enhancement distribution in at least one vascular phase. Pseudotumors enhance in a manner similar to normal renal parenchyma in all contrast enhancement phases and contain a medullary pyramid identified by enhancement less than that of normal renal cortex [1]. Solid benign tumors, such as oncocytomas, and malignant tumors, such as renal cell carcinoma (RCC), show similar enhancement patterns and may be differentiated, which is similar to findings at CT and MR imaging. Because of the excellent background tissue suppression, CEUS is more sensitive than CECT in demonstrating blood flow in hypovascular lesions [14]. It can also be used to distinguish between solid hypovascular masses and atypical cystic masses, and it is more sensitive than CT and MRI in showing the enhancement of the septations and solid components due to excellent background subtraction of a complex renal mass [1, 18, 19]. At this time renal CEUS is an off-label indication for the USA but does have reimbursement with a Category 1 CPT code. With a Negative Predictive Value (NPV of 100%, benign lesions can be diagnosed with high certainty [1]. The ability to use multiple injections is helpful in performing biopsies and procedures.

This systematic literature review and meta-analysis aims at assessing available evidence on the diagnostic performance of Lumason/Sonovue-enhanced ultrasonography of the kidney for characterization of renal masses.

Materials and methods

Data sources and search strategy

A systematic literature search was performed utilizing PubMed, a service of the US National Library of Medicine®, using the following search terms (SonoVue or Lumason or Sulphur hexafluoride or BR1 or “contrast-enhanced ultrasound”) and (renal or kidney). Limits for the literature search were human studies only and published up to February 29, 2020. In addition, 10 additional publications from other sources were identified that meet inclusion criteria.

Inclusion/exclusion criteria

Publications that met all the following inclusion criteria were evaluated for inclusion in the meta-analysis of Lumason/SonoVue use in kidney ultrasound for characterization of renal masses:

  1. 1.

    Original publication of a clinical study in human subjects with prospective or retrospective enrollment in the evaluation of cystic and/or solid renal masses;

  2. 2.

    Lumason/SonoVue was administered intravenously during renal ultrasound examinations;

  3. 3.

    True-positive, true-negative, false-positive and false-negative numbers are extractable from the published paper;

  4. 4.

    Publication in English language.

Publications that did not meet the inclusion criteria or met the following exclusion criteria were excluded:

  1. 1.

    Study was performed in non-human subjects (e.g. phantom, in vitro or animal studies);

  2. 2.

    Lumason/SonoVue was used in ultrasound examinations of the kidney for clinical indications other than characterization of renal lesions (such as follow-up evaluation after chemotherapy or local ablation treatment, viability of renal transplant, assessment of renal perfusion, inflammatory, traumatic or ischemic renal abnormalities);

  3. 3.

    Study assessed use of Lumason/SonoVue in ultrasound examinations of abdominal organs/structures other than kidneys;

  4. 4.

    Study did not include diagnostic performance results or no appropriate truth standard was included (histopathology, CECT or CEMR follow up);

  5. 5.

    Study included < 30 patients;

  6. 6.

    A contrast agent other than Lumason/SonoVue was used;

  7. 7.

    Publications were review articles, guidance/guideline articles, letters to the editor, case reports or commentary articles; meta-analyses of published data.

Quality assessment and data extraction

The assessment of study quality and data extraction was independently performed by one physician and one statistician. Inconsistencies were resolved by discussion and consensus.

Study quality and applicability were evaluated using the QUADAS-2 guideline with risk of bias and clinical applicability for patient selection, CEUS, and truth standard as well as risk of bias from flow and timing assessed.

Data extraction were performed for study characteristics and diagnostic performance related information. For study characteristics, we retrieved the first author, Journal, publication year, study design, mean age, gender distribution, volume of SonoVue administration, truth standard and the comparative modality, if any. The truth standard was histopathology, macroscopic fat on CT or MRI for diagnosis of angiomyolipoma, or a minimum two-year stability on follow-up studies. For diagnostic performance results, we extracted number of lesions, number of malignant lesions (oncocystoma was considered benign solid enhancing lesions), number of benign lesions, true positive (TP), true-negative (TN), false-positive (FP), false negative (FN); the lower diagnostic performance values were obtained when there were multiple sets of data due to multiple readers.

Statistical methods

All statistical analyses were performed using the statistical software SAS, version 9.4 (SAS Institute Inc., Cary, NC, USA).

A random effect model which formally considers heterogeneity among the studies was employed to estimate pooled sensitivity and specificity.

Forest plots for sensitivity/specificity were created to visually display sensitivity/specificity and their 95% CIs from the individual studies and the pooled results, with the relative weight of each study indicated by the relative size of the symbols in the forest plots.

Heterogeneity of the sensitivity/specificity across studies was assessed by Cochran’s Q-statistic and further quantified by degree of Inconsistency (I-square, I2) which describes the percentage of total variation attributable to heterogeneity.

Funnel plot was created to visually examine publication bias, plotting the log diagnostic odds ratio (DOR) against the standard error of the log DOR from each study. Publication bias was further tested for asymmetry using the method proposed by Egger et al. [22].

The analysis was performed separately for studies grouped by truth standard, i.e., separately for the group of studies with histopathology as truth standard and for the group of studies with a combination of histopathology and imaging follow-up as truth standard. Subgroup analyses were also performed to compare the diagnostic performance of CEUS versus CECT and to evaluate the diagnostic performance of CEUS in solid renal masses and complex cysts.

Results

Search results and study characteristics

A total of 473 publications were retrieved and were analyzed to identify relevant publications with use of Lumason/SonoVue in ultrasound of the kidney to characterize renal masses. Sixteen (16) publications met all inclusion criteria and were included in the meta-analysis[18, 19, 21, 23,24,25,26,27,28,29,30,31,32,33,34,35]. The selection procedure of publications is described in Fig. 1.

Fig. 1
figure 1

Flowchart for systemic review and study selection for meta-analysis

Full size image

A total of 1554 patients (1614 renal masses) were enrolled in the 16 studies included in the present meta-analysis. The characteristics of these studies are summarized in Error! Reference source not found below.

  • 6 studies enrolled patients prospectively and 10 studies used retrospective enrollment

  • 7 studies utilized histopathology as the sole reference standard (1088 patients; 1109 lesions) and 9 studies used a combination of pathological findings and imaging follow-up as truth standard (466 patients; 505 lesions)

  • 12 studies compared the diagnostic performance of Lumason/Sonovue-enhanced ultrasonography versus other imaging modalities (CECT – 7, CEMRI – 1, CECT and CEMRI – 1, CECT and conventional US – 1, unenhanced US – 1); 4 studies evaluated the diagnostic performance of Lumason/Sonovue-enhanced ultrasonography with no comparison versus another imaging modality;

  • 8 studies included patients with complex cystic renal masses, 3 studies included patients with solid renal masses, and 5 studies included both patients with cystic and solid renal masses.

Quality assessment using QUADAS-2

Each of the 16 publications was assessed for study quality and applicability for inclusion in the meta-analysis using the checklist based on the QUADAS-2 guidelines. For the QUADAS-2 domain of patient selection, all studies had low risk of bias except 3 [23, 25, 34]. For the domain of index test (i.e. CEUS) all studies had low risk of bias except 2 [24, 33] that did not report whether the CEUS assessment was blinded or not.

For the domain of reference standard, all 16 studies had high risk since none of the studies mentioned if the reference standard interpretation was independent of CEUS results. For the domain of patient flow, all the 7 studies that had histopathology as truth standard had low risk of bias while the rest of 9 studies had high risk due to combination of histopathology and imaging follow-up as truth standard (Table 1).

Table 1 Characteristics of included studies
Full size table

None of the studies showed concerns regarding applicability for patient selection, index test and reference standard. The quality and applicability assessment for all studies are presented in Table 2.

Table 2 Bias and applicability assessment based on QUADAS-2
Full size table

Diagnostic performance of CEUS

Individual study results for the diagnostic performance of Lumason/SonoVue-enhanced ultrasonography versus truth standard are provided in Error! Reference source not found.

The pooled sensitivity and specificity of Lumason/SonoVue-enhanced ultrasonography for characterization of renal masses were calculated separately for studies with histopathology as truth standard (n = 7) and studies with imaging follow-up and/or histopathology as truth standard (n = 9).

Based on the random effect model, the pooled sensitivity and specificity for the 7 studies with histopathology as truth standard were 97% (95% CI: 96%, 99%) and 79% (95% CI: 74%, 84%), respectively (Fig. 2), with no significant heterogeneity indicated by Cochran’s Q statistic (p = 0.0517 and 0.6565, respectively) or I-square (52% and 0%, respectively) (Table 3).

Fig. 2
figure 2

Forest plot of Pooled Sensitivity and Specificity of Studies with Histopathology as Truth Standard

Full size image
Table 3 Individual Study Results for CEUS diagnostic performance
Full size table

Among the 9 studies with histopathology or imaging follow-up as truth standard, the pooled sensitivity and specificity were 96% (95% CI: 91%, 100%) and 88% (95% CI: 82%, 94%), respectively (Fig. 3). Significant heterogeneity for specificity was observed based on both Q Statistic (p = 0.0081) and I-square (61.3%); while no significant heterogeneity for sensitivity was suggested by either Q statistic (p = 0.9199) or I-square (0%).

Fig. 3
figure 3

Forest plot of Pooled Sensitivity and Specificity, Studies with Histopathology or Imaging Follow-up as Truth Standard

Full size image

The publication bias was visually investigated by funnel plots (Fig. 4), with each circle representing data from one study. The asymmetry test of the funnel plots suggested no significant publication bias for both groups of studies (p = 0.1368 and 0.1502, respectively).

Subgroup analysis: diagnostic performance of CEUS vs CECT

There were 9 studies (845 patients; 861 lesions) with diagnostic performance results from CECT as the comparative modality (Table 1). For those studies, the pooled sensitivity and specificity of CEUS were 95% (95% CI: 93%, 97%) and 82% (95% CI: 75%, 89%), respectively (Fig. 5); the pooled sensitivity and specificity of CECT were 83% (95% CI: 75%, 91%) and 75% (95% CI: 67%, 83%), respectively (Fig. 6). The results indicated that CEUS had higher sensitivity and specificity than CECT.

Fig. 4
figure 4

Funnel Plot of Publication Bias

Full size image
Fig. 5
figure 5

Forest plot of Pooled Sensitivity and Specificity of Studies with CECT as Comparative Modality—CEUS Results

Full size image

Subgroup analysis: CEUS in characterization of complex cystic and solid renal masses

There were 11 publications with data of diagnostic performance on CEUS available based on lesion characteristics (8 studies for renal complex cystic lesions and 3 studies for renal solid lesions) (Fig. 6).

Fig. 6
figure 6

Forest plot of Pooled Sensitivity and Specificity of Studies with CECT as Comparative Modality—CECT Results

Full size image

Among the 8 studies with data available for patients with complex cystic renal lesions (419 patients; 436 lesions), the pooled sensitivity and specificity of CEUS were 95% (95% CI: 91%, 99%) and 84% (95% CI: 77%, 90%), respectively (Fig. 7).

Fig. 7
figure 7

Forest plot of Pooled Sensitivity and Specificity of Studies that Enrolled Patients with Complex Cystic Lesions

Full size image

For the 3 studies with data available for patients with solid lesions (331 patients; 341 lesions), the pooled sensitivity and specificity of CEUS were 98% (95% CI: 95%, 100%) and 78% (95% CI: 68%, 88%), respectively (Fig. 8).

Fig. 8
figure 8

Forest plot of Pooled Sensitivity and Specificity of Studies that Enrolled Patients with Solid Renal Lesions

Full size image

Limitations

This study is limited as it does not have a comparison to either CT or MRI or grouped classifications based on Bosniak score.

Discussion

This meta-analysis found that CEUS can characterize both cystic and solid renal masses with a pooled sensitivity and specificity of CEUS were 95% (95% CI: 91%, 99%) and 84% (95% CI: 77%, 90%) for complex cystic lesions. and for solid lesions a pooled sensitivity and specificity of CEUS were 98% (95% CI: 95%, 100%) and 78% (95% CI: 68%, 88%), respectively. The results suggest there may be significant heterogeneity for specificity (and not for sensitivity) in lesion characterization. This is most likely due to a lack of a standardized classification system. The development of a Bosniak CEUS classification should greatly improve the characterization and reporting of indeterminate masses in CEUS.

Renal masses are very common and are often followed, biopsied, or removed at surgery. To prevent unnecessary interventions a reliable method of determining the benignity or malignancy of a renal lesion is needed. The Bosniak classification provides guidance for characterization of cystic renal masses when using a defined protocol. However, many lesions are identified on suboptimal renal mass characterization CT or MRI protocols and a follow-up study is needed. In cystic lesions thin septa or small intracystic masses may be missed in CT or MRI, due to the overlapping background signal. CEUS can depict even single microbubbles and thus detect even minor blood flow in these intracystic structures with high sensitivity.

Also, some patients cannot receive an appropriate CT or MRI due to contrast allergy, renal insufficiency or other contraindications. CEUS is a non-radiation technique that can be used in essential all patients. With a negative predictive value of 100% it can classify lesions as benign with high confidence.

Guidelines for the use of microbubble contrast agents in renal abnormalities have been published [8, 11]. A Bosniak classification for CEUS has not been developed. However, the new Bosniak classification for CT and MRI [5] can be modified to incorporate CEUS. The definitions of septa number, wall and septa thickness, wall and septa irregularity from nodules, and calcifications can all be applied to US although US may detect more septations and nodularity. With a much thinner slice thickness, measurements of septations may be slightly different than CT or MRI but this should not substantially affect the classification. In general, CEUS is more sensitive to demonstrate small amounts of blood flow, resulting in a higher probability of malignancy for some classifications, especially for Bosniak III and IV lesions. Consequently, CEUS has been shown to be more sensitive and specific in characterizing these lesions [1, 36]. For Bosniak classification 3 lesions, approximately 50% are benign and 50% are malignant on CT or MRI [5]. However, in a large single center study, using CEUS to characterize indeterminate renal masses, CEUS was able to characterize the lesion as benign or malignant with a > 95% accuracy [1, 36] appropriately triaging patients to follow-up or intervention, thereby eliminating the need for many biopsies or surgeries.

For those cases where CT or MRI contrast is contraindicated CEUS is the examination of choice. For renal lesions which do not fulfill the criteria for the use of the Bosniak classification, including solid lesions, CEUS has been shown to have high sensitivity and specificity for characterization of these masses and should be considered as a first line study in their evaluation. For cases where a renal lesion is detected on a CT or MRI where a renal mass protocol has not been used and the lesion needs further evaluation, CEUS can characterize these lesions with high accuracy without the need for additional radiation or high cost. CEUS is not included in the original or the recently revised Bosniak classification but until a CEUS Bosniak classification is devised CEUS requires no such classification for excellent differentiation of malignant tumors, and diagnosis of benign and malignant renal masses.

This meta-analysis taken together with other large renal mass characterization studies not included in this meta-analysis [1] confirm that CEUS has an important role in characterization of renal masses, both cystic and solid. Further studies comparing CT, MRI and CEUS are needed to better define the precise criteria for benign and malignant lesions and the strengths and weakness of each modality.

The addition of CEUS to a standard grayscale ultrasound when an indeterminate renal mass is identified can expedite definitive diagnosis at the time of the initial detection.