A retrospective comparative investigation evaluating Doppler ultrasound surveillance of transjugular intrahepatic portosystemic shunts has revealed that while modern handheld ultrasound devices can generate clinically interpretable images at the bedside, they exhibit systematic measurement deviations when compared against high-end ultrasound platforms. The clinical pilot study, conducted across twenty-three patients with established intrahepatic shunts, demonstrated that a pocket-sized wireless scanner successfully visualized the vascular prostheses and captured mid-shunt velocities across all examinations without rendering any non-evaluable scans. However, paired evaluations showed that the compact unit systematically compressed the hemodynamic spectrum, overestimating flow velocities in lower-velocity ranges while underestimating flow in higher-velocity states. Because transjugular intrahepatic portosystemic shunt surveillance relies heavily on precise velocity thresholds to detect early luminal compromise, stenosis, or hyperperfusion, these systematic discrepancies highlight both the emerging utility and current diagnostic boundaries of miniaturized point-of-care ultrasound systems.

Transjugular intrahepatic portosystemic shunt creation represents a well-established therapeutic intervention designed to alleviate severe portal hypertension and its life-threatening clinical sequelae, particularly refractory ascites and recurrent variceal hemorrhage. Color-coded Doppler sonography serves as the primary noninvasive imaging modality for long-term monitoring, allowing clinicians to quantify mid-shunt flow velocities, evaluate vessel patency, and detect developing stenoses before complete occlusion occurs. Historically, accurate hemodynamic assessment has required high-performance, stationary ultrasound platforms equipped with advanced acoustic transducers, sophisticated spatial resolution, sensitive Doppler architecture, and highly customizable acquisition presets. With recent engineering advancements yielding compact, battery-powered handheld scanners capable of performing vascular assessments, questions have emerged regarding whether these portable devices offer diagnostic equivalence to benchmark institutional systems during specialized abdominal examinations.

To rigorously investigate this comparative performance, researchers analyzed clinical data from twenty-three patients who underwent standardized Doppler surveillance at the Interdisciplinary Ultrasound Center of University Hospital Regensburg between August 2023 and March 2024. The study protocol was approved by the institutional ethics committee in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines, with all enrolled participants providing written informed consent. The evaluated patient cohort comprised thirteen men, representing sixty-five percent of the sample, seven women, representing thirty-five percent, and three individuals with unspecified gender data, presenting an overall mean age of 55.5 years with a standard deviation of 9.5 years. The underlying hepatic pathology was predominantly liver cirrhosis secondary to chronic alcohol toxicity, accounting for fourteen patients, while non-alcoholic steatohepatitis was documented in three patients and hepatocellular carcinoma in two. Individual diagnoses in the remaining participants included autoimmune hepatitis, Budd-Chiari syndrome, focal nodular hyperplasia, and primary biliary cholangitis.

All patient examinations were performed in a standardized supine position with the right arm elevated above the head, utilizing predominantly intercostal acoustic windows from the right lateral chest wall to optimize shunt visualization. The imaging protocol paired a high-end Mindray Resona R9 system operating with a curved multifrequency transducer spanning one to six megahertz against a General Electric Vscan Air CL handheld scanner. The handheld device incorporates a wireless, dual-headed transducer featuring a curved-array element with an operational frequency range of two to five megahertz and a center frequency of 3.3 megahertz. The probe connects wirelessly to a mobile iOS device running a dedicated control and visualization interface, maintaining encrypted local data storage in accordance with patient data protection standards. Identical sonographic protocols were executed across both platforms, extracting color-coded Doppler mid-shunt velocities, internal shunt diameters, total post-implantation dwell times, and anatomical skin-to-shunt acoustic distances.

Qualitative assessment of B-mode and color Doppler images was conducted using a standardized five-point scoring scale, categorized from score one for non-evaluable images and score two for scans hardly evaluable due to acoustic artifacts, up to score three for artifact-laden but usable images, score four for good diagnostic quality with minor limitations, and score five for excellent visualization without relevant limitations. Two experienced ultrasound examiners, blinded to each other’s assessments, evaluated the digitally archived examinations. Notably, neither modality produced images categorized in the non-evaluable or hardly evaluable grades. However, the high-end platform achieved significantly superior image quality scores compared to the handheld scanner in sixteen of the twenty-three cases, showing a statistically significant difference with a p-value below 0.001. While the high-end system clustered predominantly within ratings of four and five, the handheld unit exhibited a marked shift toward score three, reflecting increased image noise.

When evaluating broad quantitative metrics across the entire patient group, aggregate flow parameters demonstrated no statistically significant differences between the two modalities. The median mid-shunt flow velocity recorded by the high-end platform was 110 centimeters per second, compared to a median of 105 centimeters per second measured with the handheld scanner, yielding a nonsignificant p-value of 0.393. Similarly, measurements of luminal shunt diameter demonstrated broad visual and quantitative comparability between platforms, with paired Brunner-Munzel statistical testing confirming no significant inter-device divergence at a p-value of 0.122. These macroscopic similarities indicated that at a gross population level, both ultrasound systems captured the general architectural features and patency of the deployed intrahepatic metal stents, affirming the fundamental capability of modern portable hardware to conduct basic vascular visualization.

Despite the lack of significant divergence in median values, granular analysis of paired measurements uncovered pronounced systematic velocity deviations across individual examinations. Bland-Altman agreement analysis revealed an overall mean difference, calculated as high-end velocity minus handheld velocity, of 6.5 centimeters per second, bounded by ninety-five percent limits of agreement extending from minus 34.3 to positive 47.3 centimeters per second. Crucially, the distribution of paired differences exhibited proportional bias, demonstrating a downward trend across the measurement spectrum. The handheld scanner consistently overestimated true flow velocities at lower velocities, while increasingly underestimating velocities as hemodynamic flows reached higher ranges. Kendall rank correlation confirmed a highly significant relationship between the inter-device flow difference and the arithmetic average flow across devices, displaying a tau coefficient of 0.656 and a p-value below 0.001, confirming measurement range compression.

Statistical modeling further characterized the physiological and anatomical factors driving this measurement divergence. Shunt dwell time also positively correlated with high-end velocity measurements, exhibiting a tau of 0.38 at a p-value of 0.021. In multivariable linear regression analysis, the average mid-shunt flow velocity remained a robust independent predictor of velocity discrepancy, while skin-to-shunt distance maintained statistical significance with an unstandardized beta coefficient of 5.6 and a p-value of 0.009. Conversely, shunt dwell time lost statistical significance upon multivariable adjustment, yielding an estimated coefficient of 0.00 and a p-value of 0.531.

The pronounced influence of tissue depth and baseline flow velocity highlights key physical and technological constraints governing portable Doppler instrumentation. Insonating intrahepatic vascular structures situated several centimeters beneath the costal margin presents substantial acoustic challenges, as biological tissue causes severe ultrasound attenuation and wave distortion. While premium ultrasound consoles deploy substantial computing hardware to maintain Doppler sensitivity, suppress acoustic noise, and permit fine adjustments to wall filters, pulse repetition frequencies, and angle correction, handheld devices operate within constrained processing architectures and simplified preset menus. As the skin-to-shunt distance increases, degraded Doppler signal-to-noise ratios in compact transducers amplify measurement instability, leading to underestimation of flow velocities.

These systematic biases bear vital clinical implications for everyday hepatology practice. Overestimating velocities during low-flow states creates the risk of falsely categorizing an incipient shunt stenosis as adequately patent, potentially delaying timely endovascular revision until symptomatic portal hypertension recurs. Because this single-center pilot investigation lacked invasive pressure gradient measurements to determine whether observed velocity errors would have altered concrete therapeutic decisions, researchers emphasize that handheld scanners should currently serve as complementary triage tools rather than wholesale replacements for high-end systems. Technical advancements, including artificial intelligence-assisted angle correction and enhanced software-based Doppler processing, combined with prospective multicenter validation cohorts, will be essential before pocket-sized devices can independently guide definitive shunt management.

Subject of Research: Diagnostic performance and hemodynamic measurement agreement of handheld versus high-end Doppler ultrasound systems in transjugular intrahepatic portosystemic shunt surveillance

Article Title: Comparative evaluation of handheld vs. high-end ultrasound systems for TIPSS hemodynamic assessment: first impressions

Article References: Dropco, I., Vehling-Kaiser, U., Kaiser, F., Kaiser, E., Thurn, S., Kück, F., Herr, W., Schlitt, H. J., Stroszczynski, C., Jung, E. M., & Kaiser, U. (2026). Comparative evaluation of handheld vs. high-end ultrasound systems for TIPSS hemodynamic assessment: first impressions. BMC Medical Imaging, 26(1), Article 606. https://doi.org/10.1186/s12880-026-03011-w

Image Credits: AI Generated

DOI: 10.1186/s12880-026-03011-w

Keywords: transjugular intrahepatic portosystemic shunt, TIPSS surveillance, handheld ultrasound, Doppler ultrasonography, point-of-care ultrasound, hemodynamic monitoring, portal hypertension, proportional bias, Bland-Altman analysis, color-coded Doppler sonography, blood flow velocity, skin-to-TIPSS distance, liver cirrhosis, Vscan Air CL, Resona R9, acoustic attenuation, shunt stenosis