The ultrasonographic medullary “rim sign” versus medullary “band sign” in cats and their association with renal disease

Abstract Background Medullary rim sign (MRS) refers to a hyperechoic line in the renal medulla, reported on ultrasound examination (US) in both dogs and cats with and without kidney disease (KD). Objective To describe the different aspects of MRS in cats and to assess its association with KD. Animals Cats that underwent US examination, with MRS (study group) with and without KD and without MRS with and without KD (control groups). Methods Retrospective case‐control study: cats with MRS, with or without KD (rim sign groups) and cats without MRS, with or without KD (control groups). Ultrasonographic images were blindly reviewed with attention given to the thickness and margins of the MRS recorded. Results Eighty‐four cats with MRS were included and 60 cats recruited for each control group. The MRS had 2 distinct aspects: a thin hyperechoic line with well‐defined margins (MRS‐line) in 50/84 cats (59%) and a thick hyperechoic band with ill‐defined margins (MRS‐band) in 34/84 cats (41%). Twenty of 50 (40%) cats with MRS‐line and 25/34 (74%) of cats with MRS‐band had KD. The frequency of MRS‐line was higher in cats without KD, whereas the presence of MRS‐band was more frequent in cats with KD (P = .003). Conclusions and Clinical Importance A thick hyperechoic ill‐defined band (for which the term medullary band sign is proposed) was more frequently associated with KD, whereas a thin hyperechoic well‐defined line (true MRS) may be seen in cats with or without KD.


| INTRODUCTION
On ultrasound (US) examination, the medullary rim sign (MRS) is defined as a distinct hyperechoic line in the renal medulla, parallel to the corticomedullary junction. 1,2 It has been described in dogs with both acute and chronic kidney disease (KD) such as hypercalcemic nephropathy, chronic interstitial nephritis, acute tubular necrosis, 1,3,4 and in dogs with no signs of renal dysfunction. 2 A hyperechoic area between the cortex and the medulla, corresponding to the outer medulla is considered normal in dogs, especially in small breeds, and should not be interpreted as a rim sign. 5 Additionally, a hypoechoic band in the corticomedullary junction in kidneys with a hyperechoic cortex and medulla, referred to as a "halo sign," also has been reported in some dogs and cats with ethylene glycol toxicity. 4,6 This sign has been associated with a poor prognosis. 4,6 Medullary rim sign is a common US finding in healthy cats of all breeds. 7 Furthermore, a hyperechoic line in the outer medulla has been associated with a band of mineral deposits in patients without KD. 8 However, the MRS and the halo sign also have been described in cats with KD such as pyogranulomatous vasculitis associated with feline infectious peritonitis and chronic interstitial nephritis. 1 In a recent study, MRS was identified in both azotemic and nonazotemic cats. 9 Another recent study reported the prevalence and clinical relevance of the MRS in cats. 10 Its clinical relevance however still remains unclear. In another study, both presence of MRS and visualization of a thick MRS were associated with KD. 10 In addition, some confusion regarding the definition of MRS persists. We hypothesized 2 forms of MRS appearance may exist: a physiologic thin line and a pathologic thick band. Our aim was to describe the US appearance of MRS in cats and to assess its association with the presence of KD.

| Case selection criteria
For this retrospective case-control study, the electronic medical records of all cats examined at the Veterinary Teaching Hospital of the University of Bologna between June 2008 and May 2017 were reviewed. Cats were included if the keyword "rim sign" appeared in the US report, if US images of the urinary tract were available for review, and if clinical and laboratory data (CBC, serum biochemistry and urinalysis) at the time of US examination were available. Cats were excluded if they underwent partial US examination or had incomplete images of the urinary tract or renal parenchymal changes preventing visualization of the MRS (eg, lymphoma, polycystic kidney disease) or urinary tract obstruction.
The MRS patients selected were divided into 2 groups: cats with KD (group MRS-KD) and cats without KD (group MRS-N).
Cats were diagnosed with KD based on the presence of compatible history, clinical, laboratory and US findings. In particular, included cats must have had either persistent azotemia (serum creatinine concentration >1.6 mg/dL) or persistently low urine specific gravity (USG <1.035), both assessed and confirmed over 1-month as well as an US report consistent with KD. 11 In this way, cats with International Renal Interest Society (IRIS) stage 2, 3 and 4 were included in the KD group.
Patients without these abnormalities were included in the cats without KD disease (N) group.
The control group was selected by reviewing medical records of cats that underwent US examination during the same time period (2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015)(2016)(2017). Cats were included in the cats without medullary rim sign with kidney disease (NoMRS-KD) group if they met the criteria for KD (as defined above) without visualization of a MRS and included in the cats without medullary rim sign without kidney disease (NoMRS-N) if they had neither evidence of KD nor MRS. Division of the control and study groups is summarized in Table 1.

| Ultrasound image review
All US images of the kidneys were randomly and blindly reviewed by a  if present, with dimensions in mm) was evaluated in the transverse plane as previously described, 13 and echogenicity of the perirenal tissue (normal or abnormal) also was recorded.

| Statistical analyses
Normal distribution of data was assessed by means of the    Signalment of the cats included in each group are summarized in Table 2. Selected laboratory data comparing cats with a MRS with and without KD are presented in Table 3.
In the MRS-N group, 30/39 (77%) cats had an MRS-line and 9/39 (23%) cats had an MRS-band ( Figure 2). An MRS-line was significantly more frequent in cats without KD, whereas presence of an MRS-band was significantly more frequent in cats with KD (P = .003).
When comparing the group with an MRS-line and the group with an MRS-band, no statistically significant difference was seen regarding the age of the cats (P = .31).
In total, 105 cats had KD (KD group) and 99 cats did not have KD (N group).
No significant difference was found in renal length for either kidney among the 4 groups. All other findings (ie, renal contour, corticomedullary distinction, mineral foci in the peridiverticular recesses, nephroliths, pelvic distension, and perirenal tissue) were more frequent in the groups of cats with KD, when considering all groups combined ( Figure 3).
Poor corticomedullary distinction differed within the N group, being significantly more frequent in cats from the MRS-N group than in cats from the NoMRS-N group (P < .001).
Other US signs for the 4 groups are summarized in Table 5.

| DISCUSSION
In this retrospective case-control study, we confirmed the presence in cats there were some substantial differences between the studies. In our study, all US images of the kidneys were randomly and blindly reviewed by a board-certified radiologist, unaware of the clinical diagnosis and US findings, which was not the case in the previous study, where selection bias may have been introduced. 10 Despite these differences, the 2 studies agreed on the most T A B L E 6 Results of the receiver operating characteristic (ROC) curve analysis for the discrimination between 105 cats with kidney disease (KD) and 99 cats without kidney disease (N) relevant observation and concluded that the presence of a MRSband was associated with KD.
The underlying cause of the appearance of an MRS is currently under investigation and not fully understood. The MRSline may be the result of an area of intraluminal mineral deposits within renal tubules in patients without KD as described previously. 8 Unfortunately, no histological examination of the kidneys was available in our study, making the confirmation of mineral deposits impossible.
Ours is the second report describing a thick band with ill-defined margins, which may lead to confusion between its appearance and that of a thin medullary rim line. An MRS-line was significantly more frequent in cats with KD, making it an important variable to consider when evaluating cats with MRS. A different histopathological mechanism for formation may be involved other than benign mineral deposits. 8 We hypothesize that a main mechanism involved may be vascular in origin, because the area in which the band is located corresponds to the outer medulla, a substantially more hypoxic region, even in normal kidneys. 15,16 Intrarenal oxygen availability is the balance between supply, mainly dependent on renal blood flow and demand, determined by metabolic needs. Renal blood flow is carefully maintained to ensure stable glomerular filtration, and therefore increased intrarenal oxygen consumption can lead to tissue hypoxia. 15 Tubulointerstitial hypoxia stimulates production of collagen and smooth muscle actin resulting in increased fibrogenesis. Furthermore, the hypoxic environment induces epithelial-mesenchymal transdifferentiation thus worsening fibrosis, and resulting in decreased peritubular perfusion and oxygen delivery because of capillary rarefaction. 15 On US examination, fibrosis, collagen, and smooth muscles fibers appear generally hyperechoic. For these reasons, we hypothesize that the presence of a thick hyperechoic band in the outer medulla may be the consequence of increased fibrogenesis because of chronic tubulointerstitial hypoxia, possibly enhanced by the age. Renal histopathology would have been necessary to confirm this hypothesis. Cats with KD were older than cats without KD, which was expected considering that the prevalence of the KD increases with age and is higher in geriatric patients. 7,9,17,18 Surprisingly, the MRS-KD group showed a higher median age with respect to the remaining groups; no relation is known between age and the presence of MRS. Although no statistical difference was found between the age of the cats with MRS and MBS, it is possible that age plays a role in the appearance of MRS in cats, and additional studies would be necessary to rule out or confirm this hypothesis.
In almost all cases except for 1 cat, the MRS was bilateral. This cat showed a marked difference in renal size: the right kidney was substantially smaller than the left kidney and had a completely altered US appearance because of atrophy, which may have prevented visualization of the MRS.  10 It was not possible to further characterize the type of KD and correlate MRS with the chronicity of KD. In addition, cats with IRIS stage 1 KD were not identified in our study, because of the inclusion criteria required for the diagnosis, which may have led to the inclusion of some of these patients in the N groups. In addition, both convex  and linear  transducers were used.
Type of transducer employed also can affect renal echogenicity, 8,19,22 and this factor can be considered a potential limitation of our study.
Furthermore, a single radiologist reviewed the images and categorized the type of MRS according to qualitative and semiquantitative criteria.
Another limitation is lack of renal histopathology to further define the origin of the MRS.
In conclusion, MRS was observed in both cats with and without KD. A thin hyperechoic well-defined line (MRS-line) was more frequent in cats without KD, whereas a thick hyperechoic ill-defined band (MRSband or so-called MBS) frequently was associated with KD. Presence of an MBS in association with poor corticomedullary distinction, irregular contours, and pelvic distension is US evidence of KD in cats.

ACKNOWLEDGMENTS
The results of the study were presented at the EVDI conference in

CONFLICT OF INTEREST DECLARATION
Authors declare no conflict of interest.

OFF-LABEL ANTIMICROBIAL DECLARATION
Authors declare no off-label use of antimicrobials.

INSTITUTIONAL ANIMAL CARE AND USE COMMITTEE (IACUC) OR OTHER APPROVAL DECLARATION
Authors declare no IACUC or other approval was needed.

HUMAN ETHICS APPROVAL DECLARATION
Authors declare human ethics approval was not needed for this study.