Classification of Central Venous Catheter Tip Position on Chest X-ray Using Artificial Intelligence

Recent studies utilizing deep convolutional neural networks (CNN) have described the central venous catheter (CVC) on chest radiography images. However, there have been no studies for the classification of the CVC tip position with a definite criterion on the chest radiograph. This study aimed to develop an algorithm for the automatic classification of proper depth with the application of automatic segmentation of the trachea and the CVC on chest radiographs using a deep CNN. This was a retrospective study that used plain chest supine anteroposterior radiographs. The trachea and CVC were segmented on images and three labels (shallow, proper, and deep position) were assigned based on the vertical distance between the tracheal carina and CVC tip. We used a two-stage approach model for the automatic segmentation of the trachea and CVC with U-net⁺⁺ and automatic classification of CVC placement with EfficientNet B4. The primary outcome was a successful three-label classification through five-fold validations with segmented images and a test with segmentation-free images. Of a total of 808 images, 207 images were manually segmented and the overall accuracy of the five-fold validation for the classification of three-class labels (mean (SD)) of five-fold validation was 0.76 (0.03). In the test for classification with 601 segmentation-free images, the average accuracy, precision, recall, and F1-score were 0.82, 0.73, 0.73, and 0.73, respectively. We achieved the highest accuracy value of 0.91 in the shallow position label, while the highest F1-score was 0.82 in the deep position label. A deep CNN can achieve a comparative performance in the classification of the CVC position based on the distance from the carina to the CVC tip as well as automatic segmentation of the trachea and CVC on plain chest radiographs.

The impact of anatomical variables on haemodialysis tunnelled catheter replacement without fluoroscopy

AIM

Tunnelled haemodialysis (HD) catheters can be used instantly, but there are several anatomical variables that could impact it survival. This study aimed to examine the impact of different novel anatomic variables, with catheter replacement.

METHODS

In a single-centre a prospective cohort in chronic kidney disease G5 patients were conducted. The primary outcome was to determine the factors associated with catheter replacement during the first 6-month of follow-up. All procedures were performed without fluoroscopy. Three anatomic regions for catheter tip position were established: considered as superior vena cava (SVC), cavo-atrial junction (CAJ) and mid-to deep atrium (MDA). Many other anatomical variables were measured. Catheter-related bloodstream infection was also included.

RESULTS

Between January 2019 and January 2020 a total of 75 patients with tunnelled catheter insertion were analysed. Catheter replacement at 6-month occur in 10 (13.3%) patients. By multivariate analysis, the incorrect catheter tip position (SVC) (OR 1.23, 95% CI 1.07-1.42, p <.004), the presence of extrasystoles during the procedure (OR 0.88, 95% CI 0.78-0.98, p = .03), incorrect catheter tug (OR 1.31, 95% CI 1.10-1.55, p = .003), incorrect catheter top position (kinking; OR 1.40, 95% CI 1.04-1.88, p = .02) and catheter-related bloodstream infection (OR 2.60, 95% CI 2.09-3.25, p <.001) were the only variables associated with catheter replacement at 6-month follow-up.

CONCLUSION

The risk of catheter replacement at 6-month follow-up could be attenuated by avoiding incorrect catheter tug and top position, and by placing the vascular catheter tip in the CAJ and MDA.

Chest radiography for simplified evaluation of central venous catheter tip positioning for safe and accurate haemodynamic monitoring: a retrospective observational study

OBJECTIVES

The tip-to-carina (TC) distance on a simple chest X-ray (CXR) has proven value in the determination of correct central venous catheter (CVC) positioning. However, previous studies have mostly focused on preventing the atrial insertion of the CVC tip, and not on appropriate positioning for accurate haemodynamic monitoring. We aimed to assess whether the TC distance could detect the passage of the CVC tip into the superior vena cava (SVC) and the right atrium (RA), and to accordingly suggest cut-off reference values for these two aspects.

DESIGN

Retrospective observational cohort study.

SETTING

Single urban tertiary level academic hospital.

PARTICIPANTS

479 patients who underwent CXR and chest CT scan after the insertion of a CVC with a 24-hour interval during the study period.

INTERVENTION

The TC distance was measured on CXR, and the position of the CVC tip was assessed on the chest CT images. The TC distance was described as a negative or positive number if the CVC tip was above or below the carina, respectively. Receiver-operating characteristics curve analyses were conducted to ascertain the TC distance to detect SVC entrance and RA insertion of CVC tip.

RESULTS

The TC distance could significantly detect both SVC entrance and RA insertion (p<0.001 for both; area under curve 0.987 and 0.965, respectively), with a reference range of -6.69 to 15.61 mm.

CONCLUSION

The TC distance in CXR is a simple and precise method to confirm not only the safe placement of the CVC tip but also its optimal positioning for accurate haemodynamic monitoring.

Practical anatomical landmark for optimal positioning of left-sided long-term central venous catheter (a pilot study)

Background:

Long-term central venous catheter (CVC) insertion in dialysis patients is an accepted method of hemodialysis. The appropriate CVC tip placement may reduce both early and late complications related to catheter and increase patency rate. This study aimed to evaluate a new, simple, and feasible method based on surface anatomy for the proper placement of tunneled CVC in the left internal jugular vein for hemodialysis or chemotherapy.

Materials and Methods:

The study was carried out as a quasi-experimental model at Saint Al-Zahra Education Hospital in 2016. A total of forty patients with an indication of left-sided (upper) long-term CVC insertion were enrolled. The length of catheter to be inserted in the left internal jugular vein was considered as the sum of distance from the insertion point of the needle up to sternal notch plus the total distance between the left and right sternoclavicular joint and half-length of the sternum. The right atrium (RA) or superior vena cava-RA junction was the correct region for inserting the catheter tip. The collected data were analyzed using Fisher's exact test and t-test using SPSS (version 22).

Results:

The patients were 63.75 ± 17.96 years of age, weighed 67.33 ± 13.20 kg, and height of 166.92 ± 8.99 cm. Catheters were inserted successfully in 95% of patients (n = 38). No significant relationship was found between the success of new method and age, gender, height, weight, body mass index, and sternum half-length plus the distance between the right and left sternoclavicular joint.

Conclusion:

“The mid – sternal length plus sternoclavicular joints spacing” as a new formula (based on anatomical landmarks) was found practical and safe and could easily be used among adult patients who undergo tunneled CVC in the left internal jugular vein.

Comparing the conventional 15 cm and the C-length approaches for central venous catheter placement

Introduction: The present guidelines recommend placing the catheter tip in the superior vena cava (SVC) above the pericardial cephalic reflection. The aim of this study was to compare the accuracy of two different approaches in locating the tip of the Central venous catheter (CVC) at the suggested vascular zone. Methods: This was an interventional study on two hundred patients undergoing Coronary artery bypass surgery (CABG) operation who required a central venous cannulation. They were randomly assigned into two groups. In the first group catheter placement was applied through using the conventional 15 cm method. In the second group a C-length method was applied for measuring the depth of catheter tip placement from the preoperative chest radiographs. For statistical analysis Chi-square test and T-test were used. Results: In the first group (15 cm) 100% of the patients had their catheters placed below the C-line (Carina line) and the average distance between the catheter tip and the C-line was +4.22±2.10 cm. In the second (C-Length) group 52% of the catheters were below C-line with an average distance of +0.77±0.5 cm. There was a meaningful difference between the two groups in respect to the catheter location depth and zone of placement (P<0.001). Conclusion: The C-Length approach in comparison to the conventional 15 cm approach resulted in a considerable higher number of catheters above the recommended C-line, thus it can provide a more reliable and safe mode for CVC placement in the SVC.

Evaluation of a central venous catheter tip placement for superior vena cava-subclavian central venous catheterization using a premeasured length: A retrospective study

Subclavian central venous catheterization is a common procedure for which misplacement of the central venous catheter (CVC) is a frequent complication that can potentially be fatal. The carina is located in the mid-zone of the superior vena cava (SVC) and is considered a reliable landmark for CVC placement in chest radiographs. The C-length, defined as the distance from the edge of the right transverse process of the first thoracic spine to the carina, can be measured in posteroanterior chest radiographs using a picture archiving and communication system. To evaluate the placement of the tip of the CVC in subclavian central venous catheterizations using the C-length, we reviewed the medical records and chest radiographs of 122 adult patients in whom CVC catheterization was performed (from January 2012 to December 2014) via the right subclavian vein using the C-length. The tips of all subclavian CVCs were placed in the SVC using the C-length. No subclavian CVC entered the right atrium. Tip placement was not affected by demographic characteristics such as age, sex, height, weight, and body mass index. The evidence indicates that the C-length on chest radiographs can be used to determine the available insertion length and place the right subclavian CVC tip into the SVC.

Randomized controlled trials in central vascular access devices: A scoping review

BACKGROUND

Randomized controlled trials evaluate the effectiveness of interventions for central venous access devices, however, high complication rates remain. Scoping reviews map the available evidence and demonstrate evidence deficiencies to focus ongoing research priorities.

METHOD

A scoping review (January 2006-December 2015) of randomized controlled trials evaluating the effectiveness of interventions to improve central venous access device outcomes; including peripherally inserted central catheters, non-tunneled, tunneled and totally implanted venous access catheters. MeSH terms were used to undertake a systematic search with data extracted by two independent researchers, using a standardized data extraction form.

RESULTS

In total, 178 trials were included (78 non-tunneled [44%]; 40 peripherally inserted central catheters [22%]; 20 totally implanted [11%]; 12 tunneled [6%]; 6 non-specified [3%]; and 22 combined device trials [12%]). There were 119 trials (68%) involving adult participants only, with 18 (9%) pediatric and 20 (11%) neonatal trials. Insertion-related themes existed in 38% of trials (67 RCTs), 35 RCTs (20%) related to post-insertion patency, with fewer trials on infection prevention (15 RCTs, 8%), education (14RCTs, 8%), and dressing and securement (12 RCTs, 7%). There were 46 different study outcomes reported, with the most common being infection outcomes (161 outcomes; 37%), with divergent definitions used for catheter-related bloodstream and other infections.

CONCLUSION

More high quality randomized trials across central venous access device management are necessary, especially in dressing and securement and patency. These can be encouraged by having more studies with multidisciplinary team involvement and consumer engagement. Additionally, there were extensive gaps within population sub-groups, particularly in tunneled devices, and in pediatrics and neonates. Finally, outcome definitions need to be unified for results to be meaningful and comparable across studies.

The Different Impacts on the Long-Term Survival of Tunneled Internal Jugular Hemodialysis Catheters Based on Tip Position and Laterality

Background/Aims: The aim of this study was to examine the impact of different catheter tip positions on the life of the catheter, dysfunction, infection, and quality of hemodialysis and possible differences between the access site laterality in jugular-tunneled hemodialysis catheters. Methods: Catheters were evaluated for the following parameters: place of insertion, time of insertion, duration of use, and reason for removal. In all patients, the catheter tip position was checked using an X-ray. Results: The mean duration of implanted catheters with the tip placed in the cavo-atrial junction and right atrium was significantly longer. There were no differences in catheter functionality at follow-up or complications based on catheter laterality for each catheter tip position. Conclusion: According to our results, the localization of the catheter tip in superior vena cava still remains the least preferable method. Our results showed that the main factor responsible for better catheter functionality was not laterality but the depth to which the catheter tip is inserted into the body.

Simple Formula to Place Central Venous Catheter Tip at T6 After Surgical Cutdown in Neonates

The objective of this paper was to develop a generally applicable formula to estimate correct catheter length after surgical cutdown in right internal jugular vein (RIJV) in neonates. The carina has been utilized as an anatomic landmark indicating superior vena cava-right atrium junction (SVC-RA) for the optimal placement of the central venous catheter (CVC) tip position. However, this landmark may not be accurate in neonates. Recent researches noted that the sixth vertebral body (T6) could better serve as a new landmark of SVC-RA in neonates and smaller children. We prospectively performed RIJV cutdown. For a controlled and reproducible surgical procedure, the venous entry site was consistently taken as the point where the omohyoid muscle crosses the RIJV. On intraoperative infantogram, the vertical distance between the venous entry site and T6 was measured and the catheter was inserted to this length. A linear regression model was investigated using the following variables to elicit the best prediction model for catheter length: gestational age, postconceptional age, birth weight, and weight at operation. Weight at operation best correlated with the measured CVC length (R2 = 0.916, P = 0.00), and the following linear equation was derived: estimated CVC length (mm) = 9 × [weight at operation (Kg)] + 30. There was no statistically significant difference between measured and estimated CVC length. With this formula, the optimal catheter length could easily be estimated when considering RIJV cutdown.

Simple formulas to determine optimal subclavian central venous catheter tip placement in infants and children

BACKGROUND/PURPOSE

Optimal central venous catheter (CVC) tip location is necessary to decrease the incidence of complications related to their use. We sought to create a practical method to reliably predict the length of catheter to insert into the subclavian vein during CVC placement in children.

METHODS

We performed a retrospective review of 727 chest radiographs of children who underwent either left or right subclavian CVC placement. We measured the distance from the subclavian entry site to the to the right atrium/superior vena cava (RA/SVC) junction, following the catheter's course. We analyzed the relationship between that length and patient characteristics, including: age, gender, height, weight and body surface area (BSA).

RESULTS

Two derived formulas using the BSA best correlated with the optimal subclavian CVC length. For the left subclavian vein approach, the optimal catheter length was 6.5 BSA+7 cm, and for the right subclavian vein approach it was 5 BSA+6. The use of these formulas correlated in CVC tip placement in a clinically proper location in 92.9% of smaller children and in 95.7% of larger children.

CONCLUSION

The optimal length of central venous catheter to insert into the subclavian vein may be determined through the use of a simple formula using the BSA.

Central vascular catheter placement evaluation using saline flush and bedside echocardiography

OBJECTIVES

Central venous catheter (CVC) placement is a common procedure in critical care management. The authors set out to determine echocardiographic features during a saline flush of any type of CVC. The hypothesis was that the presence of a rapid saline swirl in the right atrium on bedside echocardiography would confirm correct placement of the CVC tip, similar to the accuracy of the postplacement chest radiograph (CXR).

METHODS

This was a prospective convenience sample of emergency department (ED) and intensive care unit (ICU) patients who had CVCs placed. Investigators used subcostal or apical four-chamber echocardiography windows to evaluate the onset and appearance of turbulent flow in the right atrium when the distal port of the CVC was flushed with 10 mL of saline. Onset was rated as "immediate" (within 2 seconds), "delayed" (2 to 6 seconds), or "absent" (did not appear within 6 seconds). Appearance was rated as "prominent," "speckling," or "absent." Digital video review was used later to objectively determine precise timing of turbulence onset. The rapid atrial swirl sign (RASS) was defined as the echo appearance of turbulence entering the right atrium immediately (within 2 seconds) after the saline flush of the CVC distal port. The observance of RASS ("positive") was considered "negative" for CVC malposition. Echocardiographic results were compared to CVC tip locations within predetermined zones on the CXR. Superior vena cava (SVC) region was considered the optimal CVC tip position for subclavian and internal jugular CVC. Left CVC tips within the mid left innominate vein were also considered appropriately placed.

RESULTS

A total of 142 patients enrolled, yielding 152 CVCs. Two CVCs were excluded from analysis due to incomplete data. Both CXR and echocardiographic images for 107 internal jugular CVCs and 28 subclavian CVCs were available for analysis. Saline flush echo evaluations were also performed on 15 femoral CVCs. Either 16-cm triple-lumen or 20-cm PreSep CVCs were used. CVC malposition was discovered on CXR in four of 135 (3.0%) of the subclavian and internal jugular CVCs. RASS for subclavian and internal jugular CVC evaluations versus CXR results for CVC tip malposition yielded 75% sensitivity, 100% specificity, positive predictive value (PPV) 100% (95% confidence interval [CI] = 29.24% to 100%), and negative predictive value (NPV) 99.24% (95% CI = 95.85% to 99.98%). Mean (±SD) time for onset of saline flush turbulence was 1.1 (±0.3) seconds for subclavian and internal jugular CVC tips within the target CXR zone.

CONCLUSIONS

The rapid appearance of prominent turbulence in the right atrium on echocardiography after CVC saline flush serves as a precise bedside screening test of optimal CVC tip position.