In cats and dogs, EHPSs can be gradually occluded with either ameroid constrictors1–10 or cellophane bands8–15 placed around the anomalous vessel. There is residual shunting in up to 21% of cases with ameroid constrictor placement1 and up to 37.5% of cases after cellophane band placement because of acquired shunting or incomplete occlusion of the shunt caused by failure of cellophane banding.16 Residual shunting or recurrence of clinical signs following surgical attenuation of an EHPS can lead to considerable patient morbidity. Previous studies17–21 have used CTA, portovenography, transcolonic portal scintigraphy, ultrasonography, biochemical analyses, and assessment of hepatic biopsy specimens as diagnostic tests to determine whether residual or acquired shunting following attenuation of portosystemic shunts is present.17–21 Computed tomography angiography has been shown to be the most accurate diagnostic imaging method for detection of preoperative EHPSs and is the gold standard method for assessment of portal vasculature in humans.21–23 Postoperative CTA has also been used as a diagnostic tool for assessment of residual shunting in dogs that have undergone shunt occlusion with ameroid constrictors.24,25 These studies24,25 revealed marked streak artifact for metal-cased ameroid constrictors, which obscured verification of shunt closure. Thus, plastic ameroid constrictors have been developed as an alternative to metal-cased ameroid constrictors to avoid imaging artifact that obscures detail that would contribute to a diagnosis.24,26 Although few studies21,27 have used CTA to assess residual shunting in EHPSs occluded with cellophane bands in dogs, the results indicate that use of CTA as a diagnostic tool for assessment of residual shunting is valuable.
Cellophane banding is commonly performed for attenuation of EHPSs. The current technique employs use of metal vascular clipsa,b to secure the cellophane band after it is placed around the EHPS.21,27,28 Metal vascular clips have historically been proven to cause beam-hardening artifacts (ie, streaking and star-shaped imaging artifacts) in CT images.29,30 Beam-hardening artifacts occur when the low-energy x-rays of the polyenergetic spectrum of x-rays are attenuated by metal, which leads to artifacts that reduce image quality and obscure valuable details when trying to detect structures of interest.29,31 Imaging artifacts can be reduced through metal reduction algorithms; however, such algorithms do not eliminate these artifacts, and the usefulness of metal artifact reduction software is often limited because it can be time-consuming to apply, the loss of detail around the metal-tissue interface frequently remains, and blurring of structures with changes in contrast resolution can occur, all of which can affect clinical diagnoses.32–35
Because of the imaging artifacts caused by metal vascular clips in all surgical settings, a nonabsorbable locking polymer clipc has been developed and is widely used in human patients because it does not affect the quality of CT or MRI images. Locking polymer clips contain integrated teeth on the interior of the clip and may be an alternative to traditional metal vascular clips for use in securing cellophane bands during shunt attenuation. Although various studies have found locking polymer clips to be radio-lucent36,37 as well as radiopaque,38–41 these clips do not result in artifacts that adversely affect CT or MRI image quality and interpretation.
Postoperative CTA is not routinely performed in dogs after cellophane band placement because of expense, the need for general anesthesia with breath holding in portal CTA assessments of small dogs, and complications with evaluation owing to streaking artifact caused by a metallic object.23,24,29 However, dual-phase CTA was proven valuable in evaluation of cellophane band efficacy in a consecutive series of dogs that underwent cellophane band attenuation of an EHPS.21 Use of locking polymer clips may reduce the amount of imaging artifact with CTA, thereby facilitating evaluation of shunt occlusion and blood flow through the vasculature in close proximity to the cellophane band and clips.
The purpose of the study reported here was to compare the extent of CT imaging artifact and mechanical strength of cellophane bands secured with locking polymer clips on cadaveric splenic veins with findings for cellophane bands secured with metal vascular clips. We hypothesized that the use of locking polymer clips for cellophane banding would generate less CT imaging artifact, compared with that associated with use of metal vascular clips. We also hypothesized that the strength of locking polymer clips when applied to cellophane bands would be equivalent to that of similarly applied metal vascular clips.
Materials and Methods
Cadaveric specimens
Cadavers of 10 dogs euthanized for reasons unrelated to the study were used. The method of euthanasia was unknown, and cadavers were stored at −20°C until ready for use. Breeds represented included American Staffordshire Terrier (n = 3) and 1 each of the following: Boxer, Cattledog, Chihuahua, Golden Retriever mix, Greyhound, Mastiff mix, and Jack Russell Terrier mix. The mean weight of the cadavers was 20.1 kg (range, 5.2 to 34.2 kg).
Cellophane band placement in cadaveric specimens
The 10 canine cadavers were allocated to 1 of 2 groups (5 cadavers/group); 1 group was to be used to investigate cellophane bands secured with medium-large locking polymer clipsc and metal vascular clips,a and the other group was to be used to investigate cellophane bands secured with large locking polymer clipsc and metal vascular clips.a The choice of securing cellophane bands with locking polymer clips or metal vascular clips was initially randomized by coin toss for the first cadaver, then alternated between clip types for all successive cadavers.
In all 10 cadavers, a standard ventral midline laparotomy was performed. The portal and splenic veins were identified, and a cellophane band was placed around the splenic vein, adjacent to the entry to the portal vein. The cellophane band was secured with either 4 locking polymer clips or 4 metal vascular clips that were in a dorsal-to-ventral or cranial-to-caudal plane and were placed perpendicular to the long axis of the cellophane band in an alternating configuration from opposite sides of the band.28 After cellophane band placement, the abdominal incision was routinely closed, and CT imaging of the abdomen was performed. After acquisition of abdominal CT images, the abdominal incision was reopened, and the cellophane band and associated clips were removed. A new cellophane band was placed on the same region of the splenic vein and secured with 4 clips of the same size but of the alternate clip material. The abdominal incision was closed, and CT imaging of the abdomen was performed.
CT examinations
Each cadaveric specimen underwent CT with a 64-slice helical CT scanner.d Images (volume, 0.5mm voxels; rotation speed, 0.5 seconds; helical pitch, 0.83; and matrix, 512 × 512) were acquired. Kilovoltage and amperage were varied to optimize the images on the basis of dog size. The volume data were reconstructed with bone and soft tissue algorithms in isovolumetric transverse, sagittal, and dorsal planes at a slice thickness of 1.0 mm with a 0.5-mm reconstruction interval.
CT artifact evaluation
Examination of CT images with a slice thickness of 1.0 mm was conducted by a board-certified veterinary radiologist (SN) and by a small animal surgery resident (SLL) who were both unaware of whether locking polymer clips or metal vascular clips were being used for each evaluation. To evaluate beam-hardening artifact, objective and subjective data were collected from CT images created with a soft tissue algorithm.
To objectively quantify beam-hardening artifact on CT images, the transverse image within each scan with the most severe artifact for all 4 clips obtained by use of the soft tissue algorithm was chosen for each clip material and size type, and this image was used for all measurements of beam hardening. Artifact streak length was objectively measured and recorded by use of the electronic caliper tool; attenuation and heterogeneity were measured by use of the ROI tool of the commercially available software.e Mean attenuation (measured in HU) of each of 5 circular (0.5-cm2) ROIs drawn sequentially starting adjacent to the clip along the line used to measure the longest streak artifact was obtained (Figure 1). The ROI closest to the clip was labeled as area 1, with subsequent ROIs labeled as areas 2, 3, 4, and 5 on the basis of their increasing distance from the clip. The SD of the attenuation in each ROI was used as a quantification of heterogeneity.42 For each CT scan, attenuation of an internal control region was measured by placement of a 0.5-cm2 ROI within either the right or left epaxial musculature. Mean ± SD attenuation of each ROI was recorded.
Representative transverse CT image of 4 large metal vascular clips used to secure a triple-layer cellophane band placed on a splenic vein in a dog cadaver. The transverse image within each scan with the most severe artifact obtained by use of the soft tissue algorithm was chosen for each clip material and size type, and this image was used for all measurements of beam-hardening artifact. A line was drawn from the clip along the longest streak artifact, and a 0.5-cm2 ROI for each of 5 areas (area 1 is closest to the clip and area 5 is most distant from the clip [circles]) was drawn sequentially along the line. Mean attenuation (HU) of each region was measured, and mean values were used to quantify beam-hardening artifact, which was severe for this clip. L = Left. R = Right.
Citation: American Journal of Veterinary Research 78, 11; 10.2460/ajvr.78.11.1338
When beam-hardening artifact was not observed on CT images, the transverse image within each scan that best depicted all 4 clips of each material and size type by use of the soft tissue algorithm was chosen and used for all measurements. A tangential line to the clip was made within a region of the selected slice that had minimal gas- and mineral-attenuating areas associated with the line. Five 0.5-cm2 ROIs were drawn sequentially along the tangential line, and data were collected and recorded as previously described (Figure 2).
Representative transverse CT image of 1 of 4 large locking polymer clips used to secure a triple-layer cellophane band placed on a splenic vein in a dog cadaver. For this clip, there is minimal artifact and no beam-hardening artifact. When beam-hardening artifact was not observed on CT images, the transverse image within each scan that best depicted all 4 clips by use of the soft tissue algorithm was chosen and used for all measurements. Within a region of the selected slice that had minimal gas- and mineral-attenuating areas, a tangential line from the clip was drawn, and a 0.5-cm2 ROI for each of 5 areas (area 1 is closest to the clip and area 5 is most distant from the clip [circles]) was drawn sequentially along the tangential line. Mean attenuation (HU) of each region was measured, and mean values were used to quantify beam-hardening artifact. See Figure 1 for remainder of key.
Citation: American Journal of Veterinary Research 78, 11; 10.2460/ajvr.78.11.1338
Subjective scoring of CT artifacts was recorded. Artifact was defined as any discernible streaking or image distortion. All beam-hardening artifact sequences were scored on a scale of 1 to 3 as follows: 1 = clip caused minimal or no artifact and would not be expected to affect image interpretation, 2 = clip caused moderate artifact that could affect image interpretation, and 3 = clip caused severe artifact that would likely affect image interpretation.42
Cellophane band formation for mechanical testing
Commercially available cellophanef was cut into 12-mm-wide strips and folded twice longitudinally to produce 4-mm-wide triple-layer cellophane bands. A new cellophane band was used for each test. The same observer who applied the clips for CT imaging (SLL) also secured cellophane bands with either 4 locking polymer clips or 4 metal vascular clips applied perpendicular to the long axis of the cellophane band in an alternating configuration from opposite sides of the band as described by McAlinden et al.28 Mechanical testing was performed with medium-large and large locking polymer clips and medium-large and large metal vascular clips (n = 10 constructs [each with 4 clips]/clip group).
Mechanical testing
Each cellophane band was applied to a 10-mm-diameter split circular jaw (Figure 3) and tested by use of a computer-controlled servo-hydraulic universal materials testing machineg with a 50-N load transducerh that was calibrated prior to testing. Testing conditions including temperature and humidity were recorded.
Photograph of a cellophane band secured with 4 medium-large locking polymer clips and applied to a split circular jaw mounted on a computer-controlled servo-hydraulic materials testing machine with a 50-N load transducer to illustrate the mechanical testing of the constructs intended for use in the treatment of EHPSs in dogs. Commercially available cellophane was cut into 12-mm-wide strips and folded twice longitudinally to produce 4-mm-wide triple-layer cellophane bands. A new cellophane band was used for each test. Four medium-large or large locking polymer clips or 4 medium-large or large metal vascular clips were applied perpendicular to the long axis of the cellophane band in an alternating configuration from opposite sides of the band.
Citation: American Journal of Veterinary Research 78, 11; 10.2460/ajvr.78.11.1338
Cellophane bands were loaded to failure in tension at a distraction rate of 0.1 mm/s as described by McAlinden et al.28 Force displacement curves were generated and recorded for each test by means of standard materials testing machine software.i After linear loading was established, the yield load was determined for each cellophane band tested as the force where a deviation of 5% from linear loading first occurred. Failure method was identified and recorded.
Statistical analysis
For the CT data, all statistical analyses with the exception of the computation of the Lin CCC were performed by means of a statistical program.j The Lin CCC was determined by means of an epidemiological statistical program.k
The Lin CCC was calculated for all 5 subjective and objective variables (ie, subjective CT artifact score, mean ROI attenuation, SD of the ROI attenuation, artifact streak length, and beam-hardening impairment of CT interpretation) to quantify concordance between the 2 observers. A descriptive scale as outlined by McBride43 was used to determine strength of agreement, where a score of > 0.99 is almost perfect, 0.95 to 0.99 is substantial, 0.90 to 0.95 is moderate, and < 0.90 is poor. A weighted κ statistic was also calculated to quantify agreement among the subjective CT artifact scores, and 95% CIs were obtained. For each variable, the mean for the 2 observers was calculated for analysis. For each clip size and material, means of the 5 sequential ROI values (means, SDs, and ranges) were calculated prior to further analysis.
A Mann-Whitney U test was used to test for differences in subjective CT artifact scores among clip material and size groups. A Student t test was used to compare ROI attenuation means, ROI attenuation SDs, and artifact length (in centimeters) between clip material and size types. The folded form F statistic was used to determine whether variances were equal among clip material and size groups. If unequal, then a Satterwaithe approximation for degrees of freedom for the Student t test was used. In all tests of CT data performed, a value of P < 0.05 was considered significant.
For the mechanical testing data, statistical analyses were performed by means of statistical software.l Yield loads for each cellophane band test were reported as mean ± SD. A 1-way ANOVA was used to compare differences in yield loads for the 4 different clip groups. To determine pairwise differences, a post hoc analysis was performed with a Tukey test. In all comparisons of mechanical testing data performed, a value of P < 0.05 was considered significant.
Results
CT artifact evaluation
The Lin CCCs and 95% CIs for subjective CT artifact score, mean ROI attenuation (measured in HU), SD of the ROI attenuation (measured in HU), artifact streak length (measured in centimeters), and beam-hardening impairment of CT interpretation (measured in centimeters) determined by the 2 observers were calculated; a weighted κ statistic with 95% CI associated with subjective CT artifact scoring was also calculated. The Lin CCCs were as follows: subjective CT artifact score, 0.974 (95% CI, 0.937 to 0.990); mean ROI attenuation, 0.998 (95% CI, 0.994 to 0.999); SD of the ROI attenuation, 0.945 (95% CI, 0.882 to 0.975); artifact streak length, 0.990 (95% CI, 0.975 to 0.996); and beam-hardening impairment of CT interpretation, 0.991 (95% CI, 0.979 to 0.996). The weighted κ statistic associated with subjective CT artifact scoring was 0.95 (95% CI, 0.855 to 1.000). Thus, for the findings of the 2 observers, mean ROI attenuation and beam-hardening impairment of CT interpretation had almost perfect concordance; the agreement for subjective CT artifact score and artifact streak length were each substantial, and that for the SD of the ROI attenuation was moderate.
Mean ± SD and range were calculated for subjective CT artifact score, mean ROI attenuation, SD of the ROI attenuation, artifact streak length, and beam-hardening impairment of CT interpretation for clips of each material (Table 1). These same variables were also calculated for clips of each material according to size (medium-large [Table 2] or large [Table 3]). The subjective CT artifact score for metal vascular clips was significantly (P < 0.001) higher than that for locking polymer clips. The SD of the ROI attenuation, artifact streak length, and beam-hardening impairment of CT interpretation for metal vascular clips were also significantly (P < 0.001) higher than those for locking polymer clips. Mean ROI attenuation was significantly (P < 0.001) lower for the metal vascular clips, compared with findings for the locking polymer clips.
Characterization of CT image artifact severity (as determined by 2 observers) associated with metal vascular clips and locking polymer clips used to secure cellophane bands on splenic veins in 10 canine cadavers.
| Variable | Metal vascular clip | Locking polymer clip | P value |
|---|---|---|---|
| Subjective CT artifact score | 2.95 ± 0.16 (2.50 to 3.00) | 1.00 ± 0.00 (1.00 to 1.00) | < 0.001 |
| ROI attenuation (HU) | −83.91 ± 54.63 (−176.77 to 2.79) | 55.21 ± 13.81 (26.15 to 71.67) | < 0.001 |
| SD of the ROI attenuation (HU) | 99.35 ± 34.16 (55.90 to 158.70) | 21.41 ± 12.53 (3.93 to 38.82) | < 0.001 |
| Artifact streak length (cm) | 5.11 ± 1.57 (3.30 to 8.25) | 0.00 ± 0.00 (0.00 to 0.00) | < 0.001 |
| Beam-hardening impairment of CT interpretation (cm) | 2.87 ± 1.12 (1.45 to 5.10) | 0.00 ± 0.00 (0.00 to 0.00) | < 0.001 |
Data are reported as mean ± SD (range). A crossover study was performed in which metal vascular clips and locking polymer clips were placed in all 10 cadavers. For each cadaver, clips of 1 size were used (medium-large or large; 5 cadavers/clip size of each material). A standard ventral midline laparotomy was performed, and a cellophane band was placed around the splenic vein, adjacent to the entry to the portal vein. The cellophane band was secured with either 4 locking polymer clips or 4 metal vascular clips that were in a dorsal-to-ventral or cranial-to-caudal plane and were placed perpendicular to the long axis of the cellophane band in an alternating configuration from opposite sides of the band. After cellophane band placement, the abdominal incision was routinely closed, and CT imaging of the abdomen was performed. After acquisition of abdominal CT images, the abdominal incision was reopened, and the cellophane band and associated clips were removed. A new cellophane band was placed on the same region of the splenic vein and secured with 4 clips of the same size but of the alternate clip material. The abdominal incision was closed, and CT imaging of the abdomen was performed. Examination of CT images was conducted by 2 observers who were both unaware of whether locking polymer clips or metal vascular clips were being used for each evaluation. To objectively quantify beam-hardening artifact on CT images, the transverse image within each scan with the most severe artifact obtained by use of the soft tissue algorithm was chosen for all 4 clips, and this image was used for all measurements of beam hardening. Artifact streak length was objectively measured and recorded by use of the electronic caliper tool; attenuation and heterogeneity were measured by use of the ROI tool of the commercially available software.e Mean attenuation (measured in HU) of each of 5 circular (0.5-cm2) ROIs drawn sequentially starting adjacent to the clip along the line used to measure the longest streak artifact was measured. The SD of the attenuation in each ROI was used as a quantification of heterogeneity. For each CT scan, attenuation of an internal control region was measured by means of placing a 0.5-cm2 ROI within either the right or left epaxial musculature. When beam-hardening artifact was not observed on CT images, the transverse image within each scan that best depicted all 4 clips by use of the soft tissue algorithm was chosen and used for all measurements. A tangential line to the clip was made within a region of the selected slice that had minimal gas- and mineral-attenuating areas associated with the line. Five 0.5-cm2 ROIs were drawn sequentially along the tangential line, and data were collected and recorded as previously described. The CT image artifacts (defined as any discernible streaking or image distortion) were subjectively scored. All beam-hardening artifact sequences were subjectively scored on a scale of 1 to 3 as follows: 1 = clip caused minimal or no artifact and would not be expected to affect image interpretation, 2 = clip caused moderate artifact that could affect image interpretation, and 3 = clip caused severe artifact that would likely affect image interpretation.
Characterization of CT image artifact severity (as determined by 2 observers) associated with medium-large metal vascular clips and locking polymer clips used to secure cellophane bands on splenic veins in 5 canine cadavers.
| Variable | Metal vascular clip | Locking polymer clip | P value |
|---|---|---|---|
| Subjective CT artifact score | 3.00 ± 0.00 (3.00 to 3.00) | 1.00 ± 0.00 (1.00 to 1.00) | 0.018 |
| ROI attenuation (HU) | −79.03 ± 37.03 (−118.03 to −31.61) | 50.09 ± 15.40 (26.15 to 62.40) | < 0.001 |
| SD of the ROI attenuation (HU) | 89.65 ± 27.77 (55.90 to 124.64) | 21.96 ± 13.07 (3.95 to 36.02) | 0.001 |
| Artifact streak length (cm) | 4.95 ± 1.62 (3.30 to 7.35) | 0.00 ± 0.00 (0.00 to 0.00) | 0.002 |
| Beam-hardening impairment of CT interpretation (cm) | 3.10 ± 1.20 (2.05 to 5.10) | 0.00 ± 0.00 (0.00 to 0.00) | 0.005 |
See Table 1 for key.
Characterization of CT image artifact severity (as determined by 2 observers) associated with large metal vascular clips and locking polymer clips used to secure cellophane bands on splenic veins in 5 canine cadavers.
| Variable | Metal vascular clip | Locking polymer clip | P value |
|---|---|---|---|
| Subjective CT artifact score | 2.90 ± 0.22 (2.50 to 3.00) | 1.00 ± 0.00 (1.00 to 1.00) | 0.022 |
| ROI attenuation (HU) | −88.52 ± 72.73 (−176.77 to 2.79) | 60.32 ± 11.23 (46.46 to 71.67) | 0.010 |
| SD of the ROI attenuation (HU) | 109.05 ± 40.23 (64.55 to 158.70) | 20.87 ± 13.48 (6.37 to 38.82) | 0.002 |
| Artifact streak length (cm) | 5.27 ± 1.69 (4.25 to 8.25) | 0.00 ± 0.00 (0.00 to 0.00) | 0.002 |
| Beam-hardening impairment of CT interpretation (cm) | 2.64 ± 1.12 (1.45 to 4.00) | 0.00 ± 0.00 (0.00 to 0.00) | 0.006 |
See Table 1 for key.
Mechanical testing
Tests were performed in conditions of laboratory air with a mean ± SD ambient temperature of 21.9 ± 1.4°C and mean relative humidity of 39 ± 2%. Mean yield load varied significantly (P < 0.001) by clip type and size as determined by ANOVA. The mean ± SD yield loads (n = 10 constructs [each with 4 clips/clip group]) were as follows: medium-large metal vascular clips, 6.0 ± 1.9 N; large metal vascular clips, 8.4 ± 2.7 N; medium-large locking polymer clips, 1.9 ± 0.6 N; and large locking polymer clips, 2.8 ± 1.3 N. All constructs underwent failure by clip slippage. Most constructs failed during loading as a result of each clip slipping, with 18 of 20 metal vascular clip constructs and 17 of 20 locking polymer clip constructs failing in this manner with each of the 4 clips failing in series from closest to farthest from the band (Figure 4). Few constructs failed because of single clip slippage at peak load, with 2 of 20 metal vascular clip constructs and 3 of 20 locking polymer clip constructs failing in this manner. Tukey tests revealed significant differences in load between medium-large locking polymer clips and medium-large metal vascular clips, between medium-large locking polymer clips and large metal vascular clips, and between large locking polymer clips and medium-large metal vascular clips.
Force displacement curves representative of cellophane band-clip constructs during mechanical testing. A—Most constructs failed because of clip slippage occurring multiple times, with initial clip slippage as indicated by deviation of 5% from linear loading. Each of the 4 applied clips failed in series from closest to furthest from the band during loading and up to peak load. B—Few constructs failed because of single clip slippage at peak load as indicated by deviation of 5% from linear loading. See Figure 3 for remainder of key.
Citation: American Journal of Veterinary Research 78, 11; 10.2460/ajvr.78.11.1338
Discussion
Results of the present study indicated that use of metal vascular clips to secure cellophane bands is associated with CT imaging artifact that could compromise an accurate postattenuation diagnosis, similar to the artifacts associated with the metal casing in ameroid constrictors.9 Locking polymer clips used to secure cellophane bands were still visible on CT images; however, minimal to no imaging artifact was observed, making them a potentially appealing substitute for metal vascular clips to secure cellophane bands around EHPSs. Mechanically, the locking polymer clips were less resistant to slipping, compared with the metal vascular clips, and although supra-physiologic forces have not been previously determined, the locking polymer clips should still be considered a viable alternative to metal vascular clips for securing cellophane bands in the treatment of EHPSs.
There are various causes of persistent shunting after surgery, including failure of the cellophane band to promote complete occlusion, suboptimal band placement, recanalization, presence of other hepatic abnormalities, and development of acquired shunts.16,21,24 Although postoperative scintigraphy can be used to quantitatively assess persistent portal bypass of the liver, the residual shunt fraction obtained is not indicative of the underlying cause. In a study16 conducted by Landon et al, 6 of 16 dogs had increased scintigraphic shunt fractions at 10 weeks after cellophane banding. Among those 6 dogs, mesenteric portovenography was used to determine the underlying cause of increased shunt fractions; 3 dogs had incomplete closure of the shunt, and 3 dogs had multiple acquired shunts.16 Additionally, residual flow through the shunt does not always lead to increased scintigraphic shunt fractions, as previously determined.21,24 Therefore, given that use of locking polymer clips for cellophane banding was associated with minimal to no CT imaging artifact, use of those clips instead of metal vascular clips for cellophane banding of EHPSs would likely facilitate more accurate evaluation of shunt closure, blood flow through the vasculature in close proximity to the clip, pre- and postocclusion liver size, and characteristics of the liver parenchyma by means of CTA and its detailed 3-D anatomic imaging.
In human medicine, it is known that use of CT in patients with metal vascular clips results in beam-hardening artifact.29–35,44 When low-energy x-rays of the polyenergetic spectrum are attenuated by metal, beam-hardening artifact occurs, characterized by areas of streaking or dark band formation around the metal object.29,44,45 The size and density of the metal object changes the number and energy range of the x-rays that were lost, and would be expected to influence the severity of beam-hardening artifact. More severe beam hardening results in longer or more numerous dark bands.42 With regard to CT, regions of the images that are darker have lower attenuation when measured in HU. Thus, having a dark streak occupying part of an ROI would be expected to decrease the mean HU value of the ROI while increasing the SD of the HU value within the ROI.45,46
Whereas several studies have used < 4 metal vascular clips to secure cellophane banding, McAlinden et al28 determined that banding with 4 metal vascular clips provided optimal resistance to tensile testing. On the basis of that finding, we elected to test cellophane bands secured with a total of 4 clips, regardless of clip material, as this is the currently suggested method for cellophane band securement and the standard surgical technique at our institution. Use of 4 metal vascular clips would likely cause more severe CT imaging artifact than if fewer metal vascular clips were used. To our knowledge, physiologic forces of cellophane banding have not been evaluated; therefore, we realize that in a clinical setting, fewer than 4 clips to secure cellophane bands may be sufficient, which would likely result in less CT imaging artifact. In the present study, beam-hardening artifact was assessed by artifact length, attenuation, and a subjective grading scale ranging from 1 through 3 for mild to severe imaging artifacts. Results indicated substantial to near perfect agreement between observers with regard to mean ROI attenuation, beam-hardening impairment of CT interpretation, subjective CT artifact score, artifact streak length, and SD of the ROI attenuation.
Imaging artifacts are clinically relevant if they impede the ability to correctly interpret the imaging findings and make a diagnosis. After all variables between observers were averaged, the metal vascular clips were found to be significantly correlated with beam-hardening artifact impairment of CT image interpretation and with artifact streak length. Although locking polymer clips were mineral attenuating47 on CT images, these clips did not generate beam-hardening artifact that impaired CT interpretation or artifact streak length, as both of these values were 0 cm. Perhaps the most clinically relevant of the study findings was that the metal vascular clips were found to significantly impair CT image interpretation, whereas the locking polymer clips did not. Although this was a subjective assessment, the CT scores assigned by a board-certified veterinary radiologist and small animal surgery resident mirrored objective CT measurements of artifact associated with the metal vascular clips and with the locking polymer clips. Nearly all the metal vascular clips caused severe artifact that would likely affect image interpretation, and all locking polymer clips caused minimal to no artifact and would not be expected to affect image interpretation.
Yield load was significantly higher for cellophane bands secured with metal vascular clips, compared with that for cellophane bands secured with locking polymer clips, regardless of size of clips used. When clips are placed onto cellophane bands, they are secured by means of friction, dependent on surface area. After biomechanical evaluation of cellophane bands secured with metal vascular clips,b McAlinden et al28 determined that four 11.5-mm-open-length metal vascular clips arranged in an alternating configuration provided optimal resistance to tensile testing at 7.46 ± 0.32 N and that, in a clinical setting, placement of > 4 metal vascular clips would likely result in a marginal increase in resistance to slippage. In the present study, results of tensile testing of medium-large and large metal vascular clips were comparable to those of 11.5-mm-open-length metal vascular clips used in the study by McAlinden et al.8 Although a significantly greater yield load was noted for cellophane bands secured with metal vascular clips, compared with bands secured with locking polymer clips, resistance to tensile forces as indicated by the results of this study are likely supraphysiological. Portal pressures can be measured at a distance from the shunt vessel via a jejunal or splenic vein catheter,47,48 but the actual pressure and forces on the vessel wall within the shunt are unknown.28 It is unlikely that portal pressures following cellophane band attenuation would result in forces that could lead to clip slippage,47 regardless of clip type or size. Therefore, even though the yield loads for bands secured with locking polymer clips in the present study were significantly lower than those for bands secured with metal vascular clips, these yield loads likely exceeded loads that cellophane bands are subjected to in vivo.
In the present study, most constructs failed with multiple slips relating to each of the 4 clips during loading, from the clip closest to the band up to the clip furthest from the band. Because yield loads were identified as a deviation of 5% from linear loading, this generally corresponded to slippage of the first clip but not absolute failure. We chose a deviation of 5% from linear loading because we thought that this would be clinically relevant, as it would indicate slippage of a clip securing the cellophane band that caused loosening of the cellophane band. However, this was not absolute failure, and if absolute failure was chosen as the experimental end point, then the yield loads across clip groups would be higher. Clinically, the authors recommend that a portion of the cellophane band should be left distal to the clips to prevent absolute failure at the loads exerted in the present study. In clinical cases, clip slippage is more likely to occur with continued surgical manipulation after cellophane band placement or postoperative patient movement28 and should ideally be avoided.
In analyzing the clinical relevance of the mechanical data obtained in the present study, it is important to consider the forces placed on cellophane bands secured with clips in vivo. Although typical portal pressures after cellophane band attenuation are unlikely to result in forces that could cause clip slippage, these pressures are measured at a distance from the shunt vessel47; thus, the actual pressure and forces on the shunt vessel wall are unknown. Therefore, our testing protocol was designed to be similar to that used previously to test the number and size of cellophane band clips that provided most resistance to tensile forces in a reproducible manner.28
There are known limitations of use of cadaveric specimens. Use of cadavers in the present study resulted in evaluation of clip artifact by means of CT instead of CTA, which has been proven to be the most accurate diagnostic imaging method for evaluation of portal vasculature, detection of preoperative EHPSs, and confirmation of residual postattenuation shunting.16,21–24 Therefore, we propose that postoperative CTA would be the best diagnostic imaging technique. Another limitation of the present study was the in vitro nature of the cellophane band tensile strength testing. Although the use of a split cylinder to emulate the vessel expansion on the cellophane band has been used in a previous study,26 the tensile testing configuration does not uniformly apply circumferential force on the band as would be expected from a radially expanding vessel. Furthermore, the split cylinder had a stiff metal composition, unlike the more compliant wall of a vessel. Likewise, forces exerted on the cellophane bands in the present study were likely supraphysiological and may not be directly correlated with in vivo findings.
From the results of the present study, we concluded that cellophane banding with locking polymer clips causes minimal CT imaging artifact, compared with the effects of cellophane banding with metal vascular clips, and therefore, we accepted our first hypothesis. Because locking polymer clips were found to be less resistant to slipping than were the metal vascular clips, we rejected our second hypothesis. However, the pressures required for any of the clips to slip are unlikely to occur with clips placed on cellophane bands surrounding the portal vasculature of a dog in vivo. With regard to cellophane banding of EHPSs, the use of locking polymer clips may be considered as an alternative to the use of metal vascular clips, especially for cases where postoperative imaging may be necessary. In vivo clinical trials involving locking polymer clips are needed to further assess their safety and security before routine application for cellophane banding of EHPSs in dogs.
Acknowledgments
Supported by an Oregon State University Department of Clinical Sciences Intramural Resident Training grant.
None of the authors have a conflict of interest in the manuscript.
The authors thank Jason Wiest and Cynthia Viramontes for assistance with acquisition of images and Dr. Deborah Keys for assistance with statistical analysis of data.
ABBREVIATIONS
| CCC | Concordance correlation coefficient |
| CI | Confidence interval |
| CTA | Computed tomography angiography |
| EHPS | Extrahepatic portosystemic shunt |
| HU | Hounsfield units |
| ROI | Region of interest |
Footnotes
Hemoclip, Weck Closure Systems, Research Triangle Park, NC.
Ligaclip, Ethicon US LLC, Cincinnati, Ohio.
Hem-o-lok, Weck Closure Systems, Research Triangle Park, NC.
Toshiba Aquilion, Tochigi, Japan.
eFilm, Merge HealthCare, Milwaukee, Wis.
Darice, Strongsville, Ohio.
Instron 8501, Instron Corp, Norwood, Mass.
SM-50N, Interface, Scottsdale, Ariz.
WaveMaker, Instron Corp, Norwood, Mass.
SAS, version 9.3, SAS Institute Inc, Cary, NC.
epiR: Tools for the Analysis of Epidemiological Data, version 0.9–69, R package, Mark Stevenson, Telmo Nunes, Cord Heuer, et al, Melbourne, VIC, Australia. Available at: CRAN.R-project.org/package=epiR. Accessed Dec 12, 2015.
StatGraphics Centurion, Statpoint Technologies Inc, Warrenton, Va.
References
1. Mehl ML, Kyles AE, Hardie EM, et al. Evaluation of ameroid ring constrictors for treatment for single extrahepatic portosystemic shunts in dogs: 168 cases (1995–2001). J Am Vet Med Assoc 2005; 226: 2020–2030.
2. Havig M, Tobias KM. Outcome of ameroid constrictor occlusion of single congenital extrahepatic portosystemic shunts in cats: 12 cases (1993–2000). J Am Vet Med Assoc 2002; 220: 337–341.
3. Adin CA, Gregory CR, Kyles AE, et al. Effect of petrolatum coating on the rate of occlusion of ameroid constrictors in the peritoneal cavity. Vet Surg 2004; 33: 11–16.
4. Kyles AE, Hardie EM, Mehl M, et al. Evaluation of ameroid ring constrictors for the management of single extrahepatic portosystemic shunts in cats: 23 cases (1996–2001). J Am Vet Med Assoc 2002; 220: 1341–1347.
5. Kyles AE, Gregory CR, Jackson J, et al. Evaluation of a portocaval venograft and ameroid ring for the occlusion of intrahepatic portocaval shunts in dogs. Vet Surg 2001; 30: 161–169.
6. Murphy ST, Ellison GW, Long M, et al. A comparison of the ameroid constrictor versus ligation in the surgical management of single extrahepatic portosystemic shunts. J Am Anim Hosp Assoc 2001; 37: 390–396.
7. Vogt JC, Krahwinkel DJ Jr, Bright RM, et al. Gradual occlusion of extrahepatic portosystemic shunts in dogs and cats using the ameroid constrictor. Vet Surg 1996; 25: 495–502.
8. Youmans KR, Hunt GB. Experimental evaluation of four methods of progressive venous attenuation in dogs. Vet Surg 1999; 28: 38–47.
9. Sereda CW, Adin CA. Methods of gradual vascular occlusion and their applications in treatment of congenital portosystemic shunts in dogs: a review. Vet Surg 2005; 34: 83–91.
10. Greenhalgh SN, Reeve JA, Johnstone T, et al. Long-term survival and quality of life in dogs with clinical signs associated with a congenital portosystemic shunt after surgical or medical treatment. J Am Vet Med Assoc 2014; 245: 527–533.
11. Harari J. Lateral thoracotomy and cellophane banding of a congenital portoazygous shunt in a dog. J Small Anim Pract 1990; 31: 571–573.
12. Lawrence D. Results of surgical management of portosystemic shunts in dogs: 20 cases (1985–1990). J Am Vet Med Assoc 1992; 201: 1750–1753.
13. Frankel D, Seim H, MacPhail C, et al. Evaluation of cellophane banding with and without intraoperative attenuation for treatment of congenital extrahepatic portosystemic shunts in dogs. J Am Vet Med Assoc 2006; 228: 1355–1360.
14. Hunt GB, Kummeling A, Tisdall PL, et al. Outcomes of cellophane banding for congenital portosystemic shunts in 106 dogs and 5 cats. Vet Surg 2004; 33: 25–31.
15. Youmans KR, Hunt GB. Cellophane banding for the gradual attenuation of single extrahepatic portosystemic shunts in eleven dogs. Aust Vet J 1998; 76: 531–537.
16. Landon BP, Abraham LA, Charles JA. Use of transcolonic portal scintigraphy to evaluate efficacy of cellophane banding of congenital extrahepatic portosystemic shunts in 16 dogs. Aust Vet J 2008; 86: 169–179.
17. Burton CA, White RN. Portovenogram findings in cases of elevated bile acid concentrations following correction of portosystemic shunts. J Small Anim Pract 2001; 42: 536–540.
18. Daniel GB, Bright R, Ollis P, et al. Per rectal portal scintigraphy using 99mtechnetium pertechnetate to diagnose portosystemic shunts in dogs and cats. J Vet Intern Med 1991; 5: 23–27.
19. Szatmári V, Rothuizen J, van Sluijs FJ, et al. Ultrasonographic evaluation of partially attenuated congenital extrahepatic portosystemic shunts in 14 dogs. Vet Rec 2004; 155: 448–456.
20. Faverzani S, Trombetta R, Grieco V, et al. Clinical, laboratory, ultrasonographic and histopathological findings in dogs affected by portosystemic shunts, following surgery or medical treatment. Vet Res Commun 2003; 27(suppl 1): 755–758.
21. Nelson NC, Nelson LL. Imaging and clinical outcomes in 20 dogs treated with thin film banding for extrahepatic portosystemic shunts. Vet Surg 2016; 45: 736–745.
22. Kim SE, Giglio RF, Reese DJ, et al. Comparison of computed tomographic angiography and ultrasonography for the detection and characterization of portosystemic shunts in dogs. Vet Radiol Ultrasound 2013; 54: 569–574.
23. Nelson NC, Nelson LL. Anatomy of extrahepatic portosystemic shunts in dogs as determined by computed tomography angiography. Vet Radiol Ultrasound 2011; 52: 498–506.
24. Hunt GB, Culp WT, Mayhew KN, et al. Evaluation of in vivo behavior of ameroid ring constrictors in dogs with congenital extrahepatic portosystemic shunts using computed tomography. Vet Surg 2014; 43: 834–842.
25. Zwingenberger AL, Daniel L, Steffey MA, et al. Correlation between liver volume, portal vascular anatomy, and hepatic perfusion in dogs with congenital portosystemic shunt before and after placement of ameroid constrictors. Vet Surg 2014; 43: 926–934.
26. Inou T, Tomoike H, Watanabe K, et al. A newly developed X-ray transparent ameroid constrictor for study on progression of gradual coronary stenosis. Basic Res Cardiol 1980; 75: 537–543.
27. Yoon H, Choi Y, Han H, et al. Contrast-enhanced computed tomography angiography and volume-rendered imaging for evaluation of cellophane banding in a dog with extrahepatic portosystemic shunt. J S Afr Vet Assoc 2011; 82: 125–128.
28. McAlinden AB, Buckley CT, Kirby BM. Biomechanical evaluation of different numbers, sizes and placement configurations of ligaclips required to secure cellophane bands. Vet Surg 2010; 39: 59–64.
29. Barrett JF, Keat N. Artifacts in CT: recognition and avoidance. Radiographics 2004; 24: 1679–1691.
30. Yu H, Zeng K, Bharkhada DK, et al. A segmentation-based method for metal artifact reduction. Acad Radiol 2007; 14: 495–504.
31. Nawaz S, Fu J, Fan D. Metal artifacts reduction in x-ray CT based on segmentation and forward-projection. Biomed Mater Eng 2014; 24: 3287–3293.
32. Psychogios MN, Scholz B, Rohkohl C, et al. Impact of a new metal artefact reduction algorithm in the noninvasive follow-up of intracranial clips, coils, and stents with flat-panel angiographic CTA: initial results. Neuroradiology 2013; 55: 813–818.
33. van der Bom IM, Hou SY, Puri AS, et al. Reduction of coil mass artifacts in high-resolution flat detector conebeam CT of cerebral stent-assisted coiling. AJNR Am J Neuroradiol 2013; 34: 2163–2170.
34. Prell D, Kyriakou Y, Struffert T, et al. Metal artifact reduction for clipping and coiling in interventional C-arm CT. AJNR Am J Neuroradiol 2010; 31: 634–639.
35. Joemai RM, de Bruin PW, Veldkamp WJ, et al. Metal artifact reduction for CT: development, implementation, and clinical comparison of a generic and a scanner-specific technique. Med Phys 2012; 39: 1125–1132.
36. Abreu AL, Gill IS, Desai MM. Zero-ischaemia robotic partial nephrectomy (RPN) for hilar tumours. BJU Int 2011; 108: 948–954.
37. Ganpule A, Dhawan DR, Desai MR. Hem-o-lok clip eroding into the urethra following laparoscopic radical prostatectomy: a case report and review of literature. Indian J Urol 2010; 26: 580–581.
38. Lucioni A, Valentin C, Gong EM, et al. Computed tomography appearance of the Lapra-Ty and Weck Hem-o-lok clips in patients who recently underwent laparoscopic urologic surgery. J Comput Assist Tomogr 2006; 30: 784–786.
39. Izaki H, Fukumori T, Takahashi M, et al. Clinical research of renal vein control using Hem-o-lok clips in laparoscopic nephrectomy. Int J Urol 2006; 13: 1147–1149.
40. Matsushita K, Matsubara S, Tsumura K, et al. Computed tomography appearance of Hem-o-lok clips in patients who have undergone laparoscopic nephrectomy or nephroureterectomy. JSLS 2011; 15: 517–519.
41. Yip SK, Tan YH, Cheng C, et al. Routine vascular control using the Hem-o-lok clip in laparoscopic nephrectomy: animal study and clinical application. J Endourol 2004; 18: 77–81.
42. Terry JL, Milovancev M, Nemanic S. In vitro evaluation of a novel fiducial marker for computed tomography and magnetic resonance imaging of soft tissues in small animals. Am J Vet Res 2014; 75: 974–981.
43. McBride GB. A proposal for strength-of-agreement criteria for Lin's concordance correlation coefficient. NIWA client report: HAM2005-062. Hamilton, New Zealand: National Institute of Water and Atmospheric Research Ltd, 2005.
44. Boas FE, Fleischmann D. CT artifacts: causes and reduction techniques. Imaging Med 2012; 4: 229–240.
45. Bushberg JT, Seibert JA, Leidholt EM, et al. Computed tomography. In: The essential physics of medical imaging. 3rd ed. Philadelphia: Lippincott Williams & Wilkins, 2012; 312–374.
46. Thrall DE. Principles of computed tomography and magnetic resonance imaging. In: Textbook of veterinary diagnostic radiology. 5th ed. St Louis: Elsevier, 2007; 50–77.
47. Swalec KM, Smeak DD, Brown J. Effects of mechanical and pharmacologic manipulations on portal pressure, central venous pressure, and heart rate in dogs. Am J Vet Res 1991; 52: 1327–1335.
48. Tobias K, Johnston S. Hepatic vascular anomalies. In: Veterinary surgery: small animal. St Louis: Elsevier/Saunders, 2012; 1642.
