Pilot Evaluation of a Vacuum-Assisted Biopsy Instrument for Percutaneous Renal Biopsy in Dogs

Introduction

Diagnosis of kidney failure in dogs and cats is based on the concurrent presence of persistently increased serum blood urea nitrogen and creatinine concentrations and an inappropriately low urine specific gravity after correction of dehydration.¹ However, because azotemia does not occur until approximately 75% of nephrons have substantial functional impairment, many dogs and cats with kidney disease are not diagnosed until acute renal damage is widespread or chronic damage has progressed beyond stage I in the International Renal Interest Society staging scheme.² Because early intervention is likely associated with delay in disease progression and increased time until development of uremia, earlier diagnostic evaluation of animals with suspected kidney disease, particularly those with proteinuria, is recommended.³

Although several relatively noninvasive diagnostic tests may be used for further investigation of animals with suspected kidney disease, renal biopsy may be required for definitive confirmation of presence and type of renal damage. The goals in obtaining a renal biopsy are to collect the smallest sample possible while still ensuring sufficient tissue is sampled to allow accurate diagnosis, minimizing damage to the remaining functioning nephrons, and avoiding biopsy-associated complications. The most common methods for obtaining percutaneous biopsies of the renal cortex are via 14–18 gauge trocar-tipped manual or spring-loaded (i.e., automated) biopsy needles using ultrasound guidance or during laparoscopic or surgical visualization of the kidneys.^4–6 Biopsy of the renal medulla is avoided because histopathologic examination of this portion of the kidney offers little additional diagnostic information, and because transection of the medullary arcuate blood vessels may result in life-threatening blood loss.⁶

In people, 14 versus 18 gauge instruments have resulted in no significant differences in renal biopsy sample quality, but fully automated instruments result in a significantly lower incidence and size of post-biopsy hematomas and less severe decreases in hematocrit.^7,8 The EZ Core biopsy needle,^a a trocar-tipped 14–18 gauge automated instrument designed for one-handed operation, is used for renal biopsies of dogs in many veterinary practices. A minimum of two biopsies are typically collected to ensure adequate tissue for histopathologic examination, including for light, immunofluorescent, and electron microscopy when indicated.^4,6 However, it has been the subjective impression at the authors’ institution that, although many biopsies collected using the EZ Core needle are of sufficient diagnostic quality, limited experience of the person collecting the biopsy, biopsies of renal tumors, and laparoscopy-obtained biopsies often result in significant crush artifact, nonrepresentative tissue samples, and the need for more than two biopsies, thus potentially increasing the frequency of associated complications.

The Suros Celero vacuum-assisted, spring-loaded, core biopsy device^b recently became available for obtaining 12 gauge biopsies of human mammary tissue. Vacuum-assisted biopsy devices have been used in human medicine, particularly for characterization of mammary lesions, since the late 1980s.⁹ Similar to the EZ Core instrument, the Celero instrument sample chamber advances slightly ahead of a cutting obturator during one-handed operation; however, a vacuum is applied to bring the biopsy into the sample chamber before it is severed from the surrounding tissue. Potential reported advantages of this instrument are the much larger size of biopsy samples (therefore allowing collection of a single vs multiple biopsy samples) and the lower complication rate reported during biopsy of human breast tissue.^9,10 The purpose of this study was to compare the quality of renal biopsy samples collected during ultrasound-guided percutaneous biopsy using the EZ Core versus the Celero biopsy instrument. The authors hypothesized that the larger-gauge Celero biopsy instrument would result in equivalent or higher quality biopsy samples when evaluated using criteria developed for evaluation of human renal biopsies.

Materials and Methods

Biopsy Instruments and Procedures

Kidney biopsies were obtained with the EZ Core Trocar Tip Biopsy Needle (14 gauge, 15 cm instrument length, 17 mm sample channel, with spring-loaded trocar)^a and the Suros Celero vacuum-assisted, spring-loaded core biopsy device (12 gauge, 22 mm sample channel, vacuum-assisted sample collection, spring loaded trocar)^b (Figure 1). Dogs were randomized into two groups of 10 animals for percutaneous ultrasound-guided renal biopsy. In each dog, the same instrument type was used to biopsy both kidneys. Renal biopsies were collected from each dog from one kidney by a board-certified veterinary radiologist and the contralateral kidney by a second-year diagnostic imaging resident; assignment of the radiologist and resident to the right versus left kidney in each dog was random, but the number of right versus left kidneys biopsied per individual was identical.

View Figure

• Download as Powerpoint
• Download Figure

Dogs were placed in dorsal recumbency for ultrasonography^d and renal biopsies. Free abdominal air present in those dogs that had undergone abdominal surgery was displaced as pressure from the transducer was applied to the exterior abdominal walls to locate the kidneys. After the ultrasonographer determined an appropriate site and angle for biopsy instrument insertion, an approximately 0.5–1 cm stab incision was made using a no. 10 scalpel blade. Biopsy instruments were inserted through a transducer biopsy guide^e and the skin incision, through the ventral body wall, and advanced until the instrument trocar tip was seen ventrocaudal to the kidney. The instrument tip was then inserted under ultrasound visualization into the ventrocaudal renal cortex and biopsies collected. Dogs assigned to the EZ Core group had two biopsies collected from each kidney after insertion of the instrument through the same skin stab incision but after entering the kidney at a slightly different angle; dogs assigned to the Celero group had one biopsy collected from each kidney. All biopsy samples were immediately placed in 10% neutral buffered formalin.

Results

Tissue samples were successfully collected from both sites in all 20 dogs. The ultrasonographers’ subjective impressions were that penetration of the renal capsule by the Celero was much easier than with the EZ Core; in many cases a slight thrust was required with the EZ Core to pass through the renal capsule (with a palpable “pop” as puncture occurred), whereas the Celero usually passed into the renal parenchyma without any resistance noted. Both instruments were equally hyperechoic and easily visualized during the biopsy procedures. However, when comparing ease of operation of the two instruments, the ultrasonographers felt that although the ease with which the Celero entered the renal parenchyma allowed for easier insertion into the kidney, it also resulted in an increased vigilance being required when the trocar tip was not visible on the ultrasound machine. In some instances, the Celero instrument was advanced farther than expected or unknowingly entered an abdominal organ other than the kidney once the trochar tip was located. Biopsy tracts within the renal cortex after use of the Celero instrument were subjectively larger than those created with the EZ Core (Figure 2).

View Figure

• Download as Powerpoint
• Download Figure

All 20 (100%) biopsy attempts with the EZ Core contained renal tissue, 19 (95%) of which contained renal cortex and 15 (75%) of which were of sufficient quality to be considered interpretable (Table 1). The five uninterpretable EZ Core kidney biopsies either had moderate, widespread compression artifact that prevented examination of glomeruli (four samples) or contained only renal medulla without cortex (one sample). Fifteen of 20 total biopsy attempts with the Celero contained kidney tissue, 15 of which (75% of all biopsy attempts; 100% of samples with renal tissue) contained renal cortex and 10 of which (50% of all biopsy attempts, 67% of samples with renal tissue) were of sufficient quality for interpretation. The five Celero biopsy samples that did not contain renal tissue contained either skeletal muscle (four samples) or colon (one sample). The five Celero biopsy attempts that successfully collected kidney tissue but which nevertheless were considered uninterpretable all contained renal cortex, but had an insufficient number of glomeruli; two of these samples also had moderate, widespread compression artifact. Two biopsy samples (one from each type of biopsy instrument) contained a small piece of pancreatic tissue in addition to renal cortex. There was no significant difference between instruments in the likelihood of successfully obtaining an interpretable kidney biopsy when comparing either the percent of interpretable biopsies per total biopsy attempts (P=0.102) or comparing the percent of interpretable biopsies per biopsies that successfully collected kidney tissue (P=0.589). There were also no significant associations between the individual performing the kidney biopsy or whether the left or right kidney was biopsied and the likelihood of successfully obtaining an interpretable kidney biopsy when comparing total biopsy attempts (P=0.102 or P=0.935, respectively) or only biopsies that successfully collected kidney tissue (P=0.109 or P=0.668, respectively).

TABLE 1 Histologic Assessment of Renal Biopsies Collected from Healthy Dogs with the EZ Core and Celero Instruments

See text for description of compression scoring system and criteria for determining whether biopsies were considered interpretable. ND, not determined; SD, standard deviation.

View Fullscreen

The mean number of glomeruli found in total biopsy attempts with the EZ Core (range, 0–30; mean±standard deviation [SD] 13.7± 8.4) or the Celero (range, 0–30; mean±SD, 10.6±10.0) were not significantly different (P=0.293). When comparing only those samples that contained kidney tissue, the mean number of glomeruli in EZ Core biopsies (range, 0–30; mean±SD, 13.7±8.4) versus Celero biopsies (range, 0–30; mean±SD, 14.1±9.0) was also not significantly different (P=0.888). Greater than or equal to 5 glomeruli were found in 18 of 20 (90%) of all biopsy attempts and in 18 of 20 (90%) successful renal tissue biopsies when using the EZ Core, and in 13 of 20 (65%) of all biopsy attempts and in 13 of 15 (87%) successful renal biopsies when using the Celero. Greater than or equal to 10 glomeruli were found in 13 of 20 (65%) of all biopsy attempts and in 13 of 20 (65%) successful renal tissue biopsies when using the EZ Core, and in 9 of 20 (45%) of all biopsy attempts and in 9 of 15 (60%) successful renal biopsies when using the Celero. There were no significant associations between the individual performing the kidney biopsy or whether the left or right kidney was biopsied and the number of glomeruli when comparing total biopsy attempts (P=0.221 or P=0.419, respectively) or only biopsies that successfully collected kidney tissue (P=0.255 or P=0.552, respectively).

Seven EZ Core biopsy samples (35% of total biopsy attempts; 35% of biopsy samples with renal tissue) contained renal medulla (six of which had both renal cortex and medulla, one of which was composed entirely of medulla). All six samples with both renal cortex and medulla contained a sufficient number of glomeruli to be considered interpretable. Four Celero renal biopsy samples (20% of total biopsy attempts; 27% of biopsy samples with renal tissue) contained renal medulla. Two of the four (50%) medulla-containing Celero renal biopsy samples had <10 glomeruli. There was no association between the presence of renal medulla within total biopsy attempts or only those biopsy samples that contained kidney tissue and the biopsy instrument used (P=0.288 and P=0.599, respectively), the radiologist performing the biopsy (P=0.288 and P=0.328, respectively), or the kidney being biopsied (P=0.873 and P=0.983, respectively).

The number of renal arterioles in total biopsy attempts with the EZ Core (range, 0–10; mean±SD, 3.9±2.6) versus the Celero (range, 0–9; mean±SD, 2.7±2.9) was not significantly different (P=0.099). When comparing only those samples that contained kidney tissue, the number of renal arterioles in EZ Core biopsies (range, 0–10; mean±SD, 3.9±2.6) versus Celero biopsies (range, 0–9; mean±SD, 3.6±2.9) was also not significantly different (P=0.662). There were no significant associations between the individual performing the kidney biopsy or whether the left or right kidney was biopsied and the number of renal arterioles when comparing total biopsy attempts (P=0.576 or P=0.827, respectively) or only biopsies that successfully collected kidney tissue (P=0.739 or P=0.934, respectively). However, there was a dependent association between the number of arterioles per sample and the kidney being sampled by the particular ultrasonographer (P=0.004): one ultrasonographer was more likely to have a larger number of arterioles present when the left kidney was biopsied, whereas the other ultrasonographer was more likely to have a larger number of arterioles present when the right kidney was biopsied; this effect was independent of the biopsy instrument being used.

Twenty-eight of 35 (80%) biopsies that contained renal tissue had some compression artifact noted by the pathologist, with 13 (34%) having greater than minimal compression (i.e., greater than a compression score of 1). There was no significant difference (P=0.094) in the degree of compression artifact noted in kidney tissue obtained with the EZ Core (mean compression score, 1.6±0.8) versus the Celero (mean compression score, 0.93±1.0). Presence or severity of compression in kidney tissue was not significantly associated with the individual obtaining the renal biopsy (P=0.869) or the kidney being biopsied (P=0.671).

Discussion

Based on the results of this pilot study, the Suros Celero and the EZ Core biopsy instruments were equally effective for collection of kidney tissue, renal cortex, and interpretable renal biopsy samples in healthy dogs. The prevalence of inadvertent biopsy of the renal medulla, presence and number of renal arterioles, and severity of compression artifact were the same in both groups. However, the number of renal biopsies that were considered interpretable (EZ Core, 75% of all biopsy attempts and of renal tissue samples; Celero, 50% of all biopsy attempts, 67% of renal tissue samples) was equal to or lower than that reported in previous studies with other instruments (50–92%).^5,9–11,13 The authors suspected that this lower-than-anticipated success rate was due to the radiologists’ limited experience obtaining ultrasound-guided renal biopsies due to the historical low frequency of performing this procedure at our institution. Consistent with this hypothesis, those biopsies obtained with the Celero biopsy instrument that did not contain renal tissue or were considered to be uninterpretable due to marked crush artifact were all collected in the first five instrument passes, as would be expected with a steep, early learning curve with a new device.¹³

The most common indication for renal biopsy in dogs is definitive diagnosis and histologic classification of glomerular disease; therefore, interpretability of renal biopsy samples often depends on the number of intact glomeruli.^4,5 The minimum number of glomeruli within a renal biopsy sample that are required to accurately determine the histologic subtype of glomerulopathies described in dogs is currently unknown. In people, some authors have recommended a minimum of 20 glomeruli for diagnosis of localized diseases, whereas others demonstrated that only five glomeruli were sufficient for diagnosis of more globally-distributed glomerulopathies.^7,8,12,14,15 Veterinary nephrologists have anecdotally recommended that a minimum of 10 glomeruli should be considered adequate to consider a renal biopsy sufficient for interpretation, which was found in the majority of the EZ Core renal biopsy attempts, but not with the Celero instrument. However, a minimum of five glomeruli was recommended in equivalent studies evaluating biopsies from humans with kidney diseases, which were successfully obtained in the majority of biopsy attempts using either instrument. ^11,12 The mean number of glomeruli per sample was not significantly different between the two instrument groups (13.7 versus 10.6 for all biopsy attempts, 13.7 versus 14.1 for biopsies that successfully collected renal tissue), and were comparable to the mean number of glomeruli reported in previous studies.^15–17 Additionally, although thin section evaluation (i.e., 3 μm) of glomeruli was recommended to assess mesangial cell cellularity and capillary basement membrane thickness in animals with protein-losing nephropathies, the use of standard sections (5 μm) in this pilot study should not have altered the results of those criteria (i.e., presence of cortex or medulla, number of glomeruli and arterioles, severity of crush artifact), which the authors used for initial assessment of biopsy adequacy.⁶

The number of renal biopsy samples that included renal medulla was similar in both groups (35% of all EZ Core biopsy attempts and of biopsies that successfully collected renal tissue; 20% of all Celero biopsy attempts, 27% of Celero biopsies that successfully collected renal tissue), but was greater than previously reported (20.5%) in a large study of renal biopsies in dogs that did not specify the instrument or instrument types used.⁵ Despite the inadvertent and undesirable sampling of medulla that occurred in some cases, most of these samples still contained sufficient renal cortex to allow adequate interpretation. The only renal biopsy sample that contained medulla without cortex was collected with the EZ Core instrument. Based on these results, despite the larger gauge and length of the Celero, a fear of sampling medulla with increased frequency using this larger instrument was unfounded.

Because microscopic evaluation of arterioles for vasculitis is required for assessment of rejection in human renal transplants, current guidelines stipulate that a minimum of two arterial/arteriolar profiles should be present within human renal transplant biopsies.¹¹ However, because vasculitides resulting in renal failure are very rare in dogs (and have not been reported as of yet in cats), presence of a large number of arterioles within renal biopsies is usually considered undesirable by veterinary nephrologists because of the presumptively increased risk of life-threatening blood loss.^4,6 There was no significant difference in the number of arterioles seen in biopsies obtained with the Celero instrument versus the EZ Core instrument. However, the authors did find an association between the number of arterioles and the person collecting the renal biopsy from a particular kidney; why this association occurred is unclear.

Major complications associated with renal biopsies are infrequent, with a 13.4% overall rate of major and minor complications in one retrospective study of 213 needle and 64 wedge biopsy procedures in dogs.⁵ Hemorrhage was the most common complication reported, and might lead to microscopic hematuria, intrarenal or perirenal hematomas, peritoneal hemorrhage due to vessel or organ laceration, and rarely, death. Because the dogs in this study were euthanized immediately for ethical reasons after renal biopsy collection, the authors were unable to evaluate the frequency of post-biopsy complications. Additionally, because all biopsies were performed after several hours of general anesthesia, and thus presumptively at a time of decreased renal blood flow and glomerular filtration rate, assessment of complications in this population would have been of questionable clinical applicability. Finally, absence of renal tissue or the content of attempted biopsies not being reported made up 16.2% of all biopsies in dogs in the previously mentioned retrospective study, which was less than the 25% of all biopsy attempts reported herein with the Celero instrument. What percent of these reported biopsies contained gastrointestinal tract tissues was unknown, but the frequency of sampling of these organs by the Celero instrument, particularly the colon and pancreas, also added to the possibility of septic peritonitis and pancreatitis as possible complications, which should be better assessed in the future.

Two potential advantages of the Celero instrument over the EZ Core instrument were noted during this study. First, both ultrasonographers were impressed with the ease with which the Celero instrument was able to penetrate the kidney. Because the kidneys are partially mobile and are surrounded by a tough, fibrous capsule, smaller gauge instruments, such as the EZ Core, often bend or are deflected off the kidney surface, or displace the kidney without penetrating it during percutaneous biopsy, particularly when the angle of approach is oblique to the organ surface. Both ultrasonographers commented that the slight thrusting motion occasionally required with the EZ Core to enter the kidney was not needed with the Celero instrument. In contrast, in some cases, the latter instrument entered the dog kidney being biopsied without any change in resistance being felt. Unfortunately, however, the ease with which this instrument entered abdominal organs might have also contributed to the occasional biopsying of nonkidney tissue.

The second potential advantage of the Celero instrument over the EZ Core instrument was that only a single instrument pass was required to obtain sufficient tissue for microscopic examination. Kidney biopsies from dogs with suspected glomerular disease are frequently processed for possible light, immunofluorescence, and electron microscopy.^4,6 Because these three modalities require that tissue be placed in different fixatives, when biopsies are obtained with 14–18 gauge biopsy needles such as the EZ Core, typically one biopsy is placed in formalin for light microscopy, whereas a second is split between electron and immunofluorescence microscopy-appropriate media.^4,6,12 The larger biopsies obtained in this study with the Celero instrument, although not objectively measured, were more than sufficient for all three modalities as judged by a veterinary nephrologist (BMP) who routinely collects kidney biopsies from patients with glomerular disease. Admittedly, because only 50% of all biopsy attempts with the Celero instrument resulted in collection of interpretable kidney biopsies, this advantage of only a single biopsy being required was negated if the low success rate reflected inherent problems with the instrument rather than the operator inexperience, which the authors suspect. In addition, examination of biopsy tissue by low-power microscopy (such as with a dissecting microscope) has been advocated to ensure adequacy of tissues, and in the case of the Celero instrument, would prevent collection of a second biopsy if not truly needed.^14,18

The potential disadvantages of the results reported here included the small power of the study, the possible effect that operator inexperience had on success of obtaining renal biopsies (interpretable or not), and the previously discussed inability to evaluate dogs for post-biopsy complications. Assuming that the success rates the authors reported in this study for obtaining interpretable renal biopsies out of all biopsy attempts were accurate (i.e., 75% when using the EZ Core instrument; 50% when using the Celero instrument), slightly more than three times the number of biopsies, or 132 attempts, would have been required to confirm that these success rates were significantly (i.e., P<0.05) different. If only those samples that successfully collected renal tissue were considered, and again the assumption was made that the authors’ reported success rates (75% when using the EZ Core instrument; 67% when using the Celero instrument) were accurate, then approximately 24 times the number of biopsy attempts would have been required to confirm that performance of these instruments was truly significantly different. In other words, the authors’ results suggested that perhaps the performance of these instruments was sufficiently similar, such that any true difference in performance would only be evident in a much larger cohort of dogs, particularly if the ultrasonographer was sufficiently skilled in that kidney tissue was obtained in all biopsy attempts. Unfortunately, in this study, the authors did not record the order in which the final biopsies were collected, and thus were unable to perform a post hoc analysis to determine if elimination of early biopsy samples (when the ultrasonographers were still mastering the Celeros instrument) improved performance.

In conclusion, the Celero biopsy instrument facilitated collection of larger renal biopsy samples that were considered adequate for interpretation with only a single instrument pass. However, success rate of this device for obtaining interpretable renal biopsy samples was equal to that of the EZ Core instrument. Although differences in the renal biopsy procedure were noted between the two instruments (i.e., the need for one instrument pass vs two, and subjectively easier penetration of the renal capsule), whether these differences were clinically relevant advantages remains unknown. Further studies are required to evaluate the prevalence and severity of preacute and delayed biopsy-associated complications with this instrument in laboratory animals before use in clinical patients. However, the authors hypothesized based on results and their experience with the vacuum-assisted instrument that the Celero might be useful in those patients where larger biopsies are desirable, such as those with renal masses, or in patients where only one instrument pass (i.e., one biopsy) is possible. Smaller gauge versions of the Celero instrument that still allow a single instrument pass to obtain sufficient interpretable renal tissue should be evaluated as well.