Abstract
Background/Aim: To integrate the results of studies on interventional metrics of cone-beam computed-tomography (CBCT) guidance and fluoroscopy guidance in transjugular intrahepatic portosystemic shunt (TIPS) placement in a meta-analysis.
Materials and Methods: A systemic literature search was conducted in PubMed, Ovid/Medline, Cumulative Index to Nursing and Allied Health Literature, Web of Science and Google Scholar. Inclusion criteria were original research articles in English, comparison of CBCT and fluoroscopy guided TIPS placements and reporting of outcome parameters (procedure time, fluorotime, radiation exposure). Study quality was assessed by modified Downs-and-Black checklist. Heterogeneity was evaluated using forest plots, I2, and considering study differences. A meta-analysis was conducted to combine the outcome effect using mean difference (MD) between CBCT and fluoroscopy guided TIPS placements.
Results: In this meta-analysis, five studies with 218 patients and TIPS placements were included. Study quality was limited with 12±0 points for procedure time, fluorotime and dose area product (DAP). Heterogeneity was indicated in procedure time (I2=48.7%), fluorotime (I2=44.2%) and DAP (I2=55.4%). By application of random-effects model, procedure time and fluorotime of TIPS placements were not significantly different between CBCT and fluoroscopy guidance with an overall MD of −14.43min [95% confidence interval (CI)=−38.82 to 9.97 min; p=0.16] and −6.23 min (95% CI=−22.27 to 9.80 min; p=0.24). DAP was not significant between TIPS placement under CBCT and fluoroscopy guidance with a MD of 27.44 Gy*cm2 (95% CI=−10.47 to 65.34 Gy*cm2; p=0.11).
Conclusion: CBCT guidance tends to accelerate fluoroscopy and procedure times with slightly elevated DAP compared to fluoroscopy guidance in TIPS placements. Evidence is limited by the small number of feasibility studies. Further investigations need standardized reporting, especially for procedural complications and shunt patency.
Introduction
Transjugular intrahepatic portosystemic shunt (TIPS) is an effective and well established treatment to reduce refractory ascites and variceal bleeding in patients with portal hypertension by reduction of the portal pressure (1, 2). The right hepatic vein is catheterized and the portal vein is punctured with a TIPS needle under fluoroscopy guidance followed by balloon dilatation of the intraparenchymal tract and implantation of the TIPS stent graft (3). The most challenging part of the intervention is the blind puncture of the portal vein from the hepatic vein, which is visualized under fluoroscopy (3). Thus, the fluorotime required to implant the TIPS stent graft is shorter for an experienced interventionalist than for interventionalists with limited experience (3). This is why, various guidance techniques have been developed and are applied in clinical practice to facilitate the portal vein puncture, e. g. percutaneous ultrasound or indirect mesentericoportography via contrast injection in the superior mesenteric artery (4).
Recently, employment of cone-beam computed-tomography (CBCT) during the intervention was reported to support TIPS guidance (5), as it was also useful in different treatments such as three-dimension puncture planning in nerve root block or dosimetry in image-guided radiation therapy (6, 7). Although this technique aims to visualize the portal vein, the complexity in displaying the portal venous anatomy via generation of a CBCT may prolong the intervention or escalate radiation exposure. For example, Chivot et al. (8) reported a significant decrease in the fluorotime under CBCT guidance in comparison to conventional fluoroscopy guidance (8). In contrast, Ketelsen et al. (9) could not detect a significant difference in the fluorotime and procedure time between CBCT and fluoroscopy guidance (9). Thus, the overall effectiveness of these novel CBCT guidance technique in TIPS placement is not clear.
To our knowledge, no meta-analysis has been published on CBCT guidance in TIPS interventions. Hence, the aim of this meta-analysis is to synthesize the reported results from published studies in examining the fluorotime, procedure time and radiation exposure of TIPS placements with CBCT and fluoroscopy guidance.
Materials and Methods
Systematic literature research. A systematic literature research was performed in the electronic databases PubMed, Ovid/Medline (Ovid), Cumulative Index to Nursing and Allied Health Literature (CINAHL), Web of Science (WOS) and in the search engine Google Scholar (GS) with the following term: “(transjugular intrahepatic portosystemic shunt) AND (guidance) AND ((fluoroscopy) OR (conventional) OR (standard)) AND ((cone beam computed tomography) OR (CBCT) OR (C-arm computed tomography) OR (CACT)) AND ((procedure time) OR (fluorotime) OR (dose area product))”. In addition, the Research Assistant of WOS (WOSRA) was used with the following prompt: “Find studies on transjugular intrahepatic portosystemic shunt with comparison of procedure time or fluorotime or dose area product between fluoroscopy/conventional/standard guidance and cone beam computed tomography/CBCT/C-arm computed tomography/CACT guidance!”. The time interval was not limited.
Study inclusion criteria. Inclusion criteria for the meta-analysis were original articles in the English language, the comparison between CBCT and fluoroscopy guidance, the clinical application and the reporting of the defined outcome parameter [procedure time, fluorotime and radiation exposure using dose area product (DAP)]. Fluorotime was defined as the duration fluoroscopic images were generated and the procedure time was defined as the duration of the TIPS procedure. DAP was defined as the total amount of radiation exposure during the intervention. Exclusion criteria were abstracts or conference presentations, case or case series, reviews, commentaries, letters, editorials, pictorial essays, book chapters, records in different languages, records without comparison or control cohort, records with different interventional techniques in the study or control cohort, experimental or animal investigation and records with the reporting of different parameters. The reference manager Zotero was applied (https://www.zotero.org/).
Data extraction. The information on first author, year of publication, country, sex, age, liver disease, indication and the defined outcome parameters (procedure time, fluorotime and DAP) were extracted from the included studies. For the outcome parameter, the cohort size, mean and standard deviation (SD) were extracted and in case of different reported values (median, quartiles, minimum and maximum) transformation in mean±standard deviation (SD) was performed via the methods by Hozo et al. (SD=range/4) and Wan et al. (10, 11). Both investigators (A.M. and T.C.M.) reviewed each study for inclusion and extracted data independently; in case of disagreements, a group consensus was reached. Data were converted by one investigator (T.C.M.) and given for verification to another (H.J.M.).
Study quality. The study quality of the included studies was assessed with the modified Downs-and-Black checklist (12, 13). Items 1 to 10 refer to reporting, 11 to 13 refer to external validity, 14 to 26 refer to internal validity. Item 27 was modified by Nascimento et al. (13) with a maximum score of 1 (power analysis was conducted) and minimum score of 0 (power analysis was not conducted) (13). Thus, the highest score for the modified checklist was 28 (13). Study quality was reported in the following categories: excellent (26-28), good (20-25), fair (15-19) and poor (≤14) (13). The investigators (D.E.D. and T.C.M.) evaluated the study quality independently and a group consensus was reached when disagreements occurred.
Statistical analysis. Firstly, study heterogeneity was assessed with forest plots, I2, and evaluation of differences in study design. It was rated using I2 as limited (I2: 0-40%), moderate (I2: 40-60%), substantial (I2: 60-80%) or considerable (I2: 80-100%) (14). When I2 was >40% and/or differences in study design were present, meta-analysis was conducted using a random-effects model, Sidik-Jonkman with Knapp-Hartung error adjustment, to combine the outcome effect, mean difference (MD), between CBCT guided and fluoroscopy guided TIPS placements. Finally, leave-one-out sensitivity analysis was performed to evaluate the impact of individual studies on the overall effect. MD, 95% confidence interval (CI) and p-values were given. The level of statistical significance was set to <0.05. All analyses were conducted using SPSS (SPSS Statistics, Version 29, IBM, Armonk, NY, USA).
Results
Systematic literature research and study inclusion. The systematic literature research was conducted in March 2025, and 283 records were identified in the databases and search engine. After duplicate removal, the remaining 250 records were screened for the relevant topic and potential eligibility. In 30 records on the relevant topic, the abstracts or full texts were examined for study inclusion. Finally, 5 studies fulfilled the inclusion criteria and were analyzed (8, 9, 15-17). The diagram in Figure 1 illustrates the study inclusion and exclusion, adapted to the PRISMA statement (Preferred Reporting Items for Systematic Reviews and Meta-Analysis) (https://www.prisma-statement.org/).
Study inclusion The study inclusion process is illustrated adapted to the PRISMA statement. CINAHL: Cumulative Index to Nursing and Allied Health Literature; GS: Google Scholar; n: number; Ovid: Ovid/Medline; PRISMA: Preferred Reporting Items for Systematic Reviews and Meta-Analysis; WOS: Web of Science; WOSRA: Research Assistant of WOS.
Data extraction. A total number of 218 patients suffering from therapy refractory portal hypertension were involved in five studies with four studies (80%) from Europe and one study (20%) from Asia (8, 9, 15, 16). Overall, patients included in the studies were predominantly male and around 60 years of age with common indications for TIPS placement. Of note, the CBCT guidance technique differed slightly between the included studies. For example, Chivot et al. (8) and Leger et al. (16) generated a CBCT, but it was performed after catheterization of the hepatic vein and in wedged portography to identify the adjacent portal vein branch (8, 16). Ketelsen et al. (9) acquired a contrast enhanced CBCT to display the portal vein using intravenous application contrast media (9). In accordance, part of the TIPS placements reported by Boening et al. (15) were performed with intravenous contrast injection and generation of a CBCT to guide the portal vein puncture (15). In addition, the remaining part of the included TIPS placements by Boening et al. (15) were performed with the generation of a native CBCT and co-registration with pre-interventional computed tomography (15). In the preprint by Huang et al. (17), CBCT was generated after portal vein puncture to evaluate portal vein entry (17). If portal vein entry was inappropriate, re-puncturing was performed by Huang et al. (17). The studies by Chivot et al. (8), Ketelsen et al. (9) and Leger et al. (16) reported all outcome parameters (fluorotime, procedure time, DAP), whereas Boening et al. (15) did not report the fluorotime and Huang et al. (17) only reported on DAP (8, 9, 15, 16). Data are given in detail in Table I and Table II.
Characteristics of the studies.
Interventional metrics of the studies.
Study quality. The included studies were retrospective studies with feasibility design (8, 9, 15-17). Hence, the overall study quality was poor with a mean±SD of 12±0 points for procedure time, fluorotime and DAP in the modified Downs-and-Black checklist (13). Data are given in Table III.
Assessment of the quality of the studies.
Statistical analysis of the procedure time, fluorotime and dose area product and leave-one-out sensitivity analysis. Four studies reported procedure times of the intervention (8, 9, 15, 16). The analysis of the forest plot and I2 (I2=48.7%) revealed moderate heterogeneity among the included studies on the procedure time. In addition, subtle differences in the CBCT technique could be identified as described before. Thus, the random-effects model was applied. Overall MD for the four studies for the procedure time between CBCT guidance and fluoroscopy-guidance was −14.43 min (95% CI=−38.82 min to 9.97 min; p=0.16). Although this effect on the procedure time was not significant, TIPS placements under CBCT guidance tend to be faster as compared to fluoroscopy guidance. The forest plot is given in Figure 2.
Forest plot of procedure time. The effect sizes are represented by blue boxes and the corresponding confidence intervals are shown with black bars. Estimated overall effect size is symbolized with the violet rhombus and estimated overall confidence interval is displayed with whiskers. The overall effect size value is marked with a dashed red line, and the reference is labelled by a grey line. The numbers on the x-axis represent the unit minutes. The dashed red line is on the left side of the grey line, which symbolizes an overall effect with lower procedure time for CBCT guidance than for fluoroscopy guidance. Lower: Lower confidence interval border; Std. Error: standard error; Upper: upper confidence interval border.
Out of five studies, three studies reported fluorotimes (8, 9, 16). In analogy to the procedure time, forest plot, I2 (I2=44.2%) and study design showed heterogeneity among the studies reporting on the fluorotime. By application of the random-effects model, the MD for the fluorotime between CBCT guided and fluoroscopy guided TIPS placements was not significant with −6.23 min (95% CI=−22.27 to 9.80 min; p=0.24). This also indicates a tendency for an improved or accelerated fluorotime of TIPS placements under CBCT guidance compared to fluoroscopy guidance. The forest plot is displayed in Figure 3.
Forest plot of fluorotime. The effect sizes are depicted with blue boxes and the corresponding confidence intervals are displayed with black bars. Estimated overall effect size is represented by the violet rhombus and estimated overall confidence interval is shown with whiskers. The overall effect size value is labelled with a dashed red line, and the reference is symbolized by a grey line. The numbers on the x-axis represent minutes. The dashed red line is on the left side of the grey line, which represents an overall effect with lower fluorotime for CBCT guidance than for fluoroscopy guidance. Lower: Lower confidence interval border; Std. Error: standard error; Upper: upper confidence interval border.
All studies reported on DAP of the intervention (8, 9, 15-17). Heterogeneity was also present for DAP according to the forest plot, I2 (I2 =55.4%) and study differences. The random-effects model revealed an overall effect between TIPS placement under CBCT guidance and fluoroscopy guidance with a MD of 27.44 Gy*cm2 (95% CI=−10.47 to 65.34 Gy*cm2; p=0.11). This effect was also not significant with slightly higher DAP in TIPS placements under CBCT guidance than in TIPS placements under fluoroscopy guidance. The forest plot is shown in Figure 4.
Forest plot of the dose area product. The effect sizes are shown by blue boxes and the corresponding confidence intervals are depicted with black bars. Estimated overall effect size is labelled with the violet rhombus and estimated overall confidence interval is displayed with whiskers. The overall effect size value is marked with a dashed red line, and the reference is symbolized by a grey line. The numbering of the x-axis is in the units of minutes. The dashed red line is on the right side of the grey line, which reflects an overall effect with higher dose area product for CBCT guidance than for fluoroscopy guidance. Lower: lower confidence interval border; Std. Error: standard error; Upper: upper confidence interval border.
When a single study was excluded, the overall effect in procedure time, fluorotime and DAP was not different to the results of the meta-analysis including all studies. The data are given in Table IV.
Leave-one-out sensitivity analysis.
Discussion
TIPS placement is the standard treatment for therapy refractory ascites and variceal bleeding (1, 2), but it remains one of the most challenging interventions due to the invisible target: the portal vein. Several 3D guidance techniques with generation of a CBCT during the intervention or co-registration of pre-interventional imaging have been reported to improve the visualization and puncture of the portal vein (5, 8, 9, 16, 18-21). However, the reported results of the procedural characteristics of TIPS interventions performed with CBCT guidance techniques have not been systematically analyzed previously. In the present meta-analysis, we harmonized the reported results of the studies comparing CBCT guidance and fluoroscopy guidance in TIPS placements (8, 9, 15, 16). In short, there is no significant difference between CBCT and fluoroscopy guidance in the interventional metrics in TIPS placements in our meta-analysis. However, procedure and fluorotimes tend to be shorter, while radiation exposure can be slightly increased. These results are consistently shown among the groups and the outcome parameter in the leave-one-out sensitivity analysis. This implies that no study showed a disproportionate impact on the results, although subtle differences in the CBCT guidance technique were noted.
First, patient cohorts of the included studies are predominantly male patients with an age of around 60 years and refractory ascites or bleeding as indications for TIPS placement. Of note, technical success rate for CBCT guided TIPS placements is reported with 90% in the study by Boening et al. (15) and 100% in the other included studies (8, 9, 15-17). In addition, no immediate or major complications have been detected in relation to TIPS placements under CBCT guidance in the studies by Boening et al. (15), Chivot et al. (8), Ketelsen et al. (9) and Leger et al. (16) while Huang et al. (17) have not reported on this data (8, 9, 15-17). However, no standardized reporting of complications is performed and the selection of the patients in these feasibility studies might have affected the technical success rate. Nevertheless, the included studies have reported overall on typical patient cohorts which indicates a general transferability of the results on patients referred for TIPS placements.
Focused on procedure time, the procedure time in our analysis tends to be reduced by about 14 minutes when CBCT guidance is applied in comparison to fluoroscopy guidance in TIPS procedures. An explanation for a potential reduction of the procedure time could rely on specific details of the CBCT guidance technique. In case of the generation of a CBCT after the catheterization of any hepatic vein as described by Chivot et al. (8), the virtual needle path is the ideal route for the TIPS placement (8). This can save the time to remove the catheter from the middle or left hepatic vein in order to catheterize the right hepatic vein, because it clearly visualizes the unfamiliar route from the middle or even left hepatic vein to the appropriate portal vein branch in any individual case (8). Thus, the visualization of the entry and target for the TIPS tract is likely to support optimal positioning of the TIPS stent graft, which could reduce procedure time. Since there is no standardized definition of the procedure time, which might vary among the studies, subtle differences might be difficult to detect.
In analogy to the procedure time, the fluorotime tends to be shorter in the group with CBCT guidance than in the group with fluoroscopy guidance by about 6 minutes. A reduction of the fluorotime under CBCT guidance could be explained with improved guidance for the puncturing of the portal vein. The fluorotime increases with the time required to propagate or adjust the orientation of the TIPS needle until the portal vein is successfully punctured. Thus, the fluorotime of an experienced interventionalist will be lower than of an interventionalist with limited experience (3). Additionally, certain conditions, e.g., portal vein thrombosis, Budd-Chiari syndrome, postoperative status after liver surgery or orthotopic liver transplantation and anatomic anomalies of the portal vein, will be more complex and even experienced interventionalists are likely to have an increased fluorotime (3). Therefore, CBCT guidance could reduce the fluorotime by potential acceleration of successful puncturing of the portal vein and reduction of off-target punctures. This could be an advantage in complex cases for experienced interventionalists due to the visualization of the puncture target. In accordance, Ketelsen et al. (9) and Huang et al. (17) showed a significant reduction of the puncture time using CBCT guidance in comparison to the puncture time using fluoroscopy guidance (9, 17). On the other side, Chivot et al. (8) and Leger et al. (16) did not detect a significant difference in the puncturing time for portal vein entry and Boening et al. (15) did not compare this time between CBCT and fluoroscopy guidance (8, 15, 16). An explanation might be the definition of this puncture time or portal vein entry time, which differed among the studies and limits the comparability. However, an improved needle navigation and accelerated portal vein puncture under CBCT guidance could also reduce the overall procedure time.
The radiation exposure was slightly increased in TIPS placements under CBCT guidance than in TIPS placements under fluoroscopy guidance by an overall DAP excess of 27.44 Gy*cm2. Of course, this additional amount of radiation exposure can be attributed to the generation of the CBCT. In particular, Huang et al. (17) reported a significant DAP increase in CBCT group compared to control group. Since CBCT was generated after portal vein puncture to evaluate the portal vein entry in the study by Huang et al. (17), the application of the CBCT could not facilitate puncture guidance and reduce the time for initial portal vein puncture in this study. Thus, the increased DAP in the CBCT group is conceivable in that study. If inappropriate portal vein entry detected on CBCT and re-puncturing the portal vein can optimize the TIPS position and shunt patency, it might be still an interesting question. In the included studies, the mean DAP for TIPS placements varies between 135 Gy*cm2 and 563 Gy*cm2 under CBCT guidance and between 134 Gy*cm2 and 469 Gy*cm2 under fluoroscopy guidance. Thus, the DAP has a wide range among studies. In general, radiation exposure is affected by several factors, e.g., the level of experience of the interventionalist, the complexity of the intervention, the patient’s height and weight as well as the generation and settings of the angiography system. Moreover, appropriate devices are also essential to facilitate TIPS placement, e.g., the development of steerable TIPS-cannula might be promising in analogy to steerable sheaths and microcatheters (22-24). However, given these multiple factors and the large variation in DAP in the studies, the additional radiation exposure of a CBCT might not be a major contribution to the total DAP of the procedure. Of course, optimization of the CBCT protocol should always be performed to minimize radiation exposure as low as reasonably achievable for patients and healthcare workers. When healthcare workers could stay outside the angiography room during CBCT generation, the occupational radiation exposure could be decreased by potentially reduced fluorotime using CBCT guidance.
Study limitations. First, the literature is very limited on CBCT guidance in TIPS placements. The published studies and the preprint by Huang et al. (17) were designed to show the feasibility of these novel navigation tools with a relatively small number of TIPS placements. Hence, the study quality is relatively poor, a sampling error could not be excluded, and heterogeneity could not be further evaluated for potential sources, e.g. with a subgroup analysis. Second, publication bias could not be excluded, because the analysis (e.g., funnel plot or Egger’s test) cannot be reliably analyzed due to the limited number of studies. Third, there is no standardized technique for the CBCT guidance and for the reporting of all interventional metrics (e.g., procedure time) that makes it difficult to compare and synthesize the results of different studies. Fourth, no standardized reporting system, e.g. the classification system recommended by the Cardiovascular and Interventional Radiological Society of Europe (25), have been applied and only immediate or major complications related to the procedure have been presented in a part of the included studies. Fifth, mid- and long-term complications or shunt patency or overall survival of the patients have not been published yet. This limits the analysis of the full potential and overall clinical benefit of the CBCT guidance technique. Finally, the comparison of CBCT guidance to co-registration or image fusion techniques and the application of intravascular ultrasound, especially in complex cases, would also support the generation of an overall conclusion on the image guidance techniques in TIPS placements.
Conclusion
Fluoroscopy and procedure times tend to be improved while radiation exposure appears slightly increased in TIPS placement under CBCT guidance compared to fluoroscopy guidance. This indicates that CBCT guidance has the potential to accelerate TIPS placements. However, the current evidence is limited due to the feasibility design of the studies. Further investigations need standardized reporting, especially for procedural complications and shunt patency, to elucidate the benefit of CBCT guidance.
Acknowledgements
We thank Soeren Sievers for preparatory work which supported us in the development of the project outline of this study. We thank Johannes Uhlig for advice and preparatory work which supported us in the development of the project outline of this study. Supported by PRACTIS–Clinician Scientist Program of Hannover Medical School, funded by the German Research Foundation (DFG, ME 3696/3).
Footnotes
Authors’ Contributions
A.M.: Data curation, Formal analysis, Investigation, Software, Writing – review & editing. D.E.D.: Data curation, Formal analysis, Investigation, Project administration, Writing – review & editing. T.C.M.: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Roles/writing – original draft, Writing – review & editing. H.J.M.: Conceptualization, Formal analysis, Investigation, Supervision, Validation, Visualization, Roles/writing – original draft, Writing – review & editing.
Data Availability Statement
Initial unpublished data were part of a review by A.M. (medical student) and T.C.M. (medical doctor) for the scientific module in medical school. All data relevant to the study are included in the article.
Conflicts of Interest
The Authors declare that they have no conflicts of interests that are directly relevant to the content of the study.
Funding
T.C.M. was supported by PRACTIS.
Artificial Intelligence (AI) Disclosure
During the preparation of this work authors used the AI-powered Web of Science Research Assistant (WOSRA) for literature search, but not to produce any content.
- Received August 27, 2025.
- Revision received October 8, 2025.
- Accepted October 14, 2025.
- Copyright © 2026 The Author(s). Published by the International Institute of Anticancer Research.
This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC-ND) 4.0 international license (https://creativecommons.org/licenses/by-nc-nd/4.0).










