The predictive value of diagnostic imaging for histological lesions of the upper poles in duplex systems with ureteroceles

Authors


  • Presented at the 13th Annual Meeting of European Society for Paediatric Urology, Budapest, Hungary, April 2002.

A.E. Khoury, Division of Urology 5, The Hospital for Sick Children, M299, 555 University Avenue, Toronto, Ontario, Canada M5G 1X8.
e-mail: tony.khoury@sickkids.ca

Abstract

OBJECTIVE

To compare the diagnostic imaging findings with the histological lesions in upper pole nephrectomy (UPN) specimens of duplex system ureteroceles, using renal ultrasonography (US) and nuclear renal scintigraphy.

PATIENTS AND METHODS

Between 1992 and 2000, 86 patients with a ureterocele in a duplex system underwent surgery. The results from US were reviewed in 84 patients by a radiologist, for echogenicity, parenchymal thinning and hydronephrosis; 77 nuclear renal scans describing the differential function of the upper poles were also reviewed. Fifty-five patients underwent UPN (25 antenatal, 30 postnatal; 18 intravesical, 37 extravesical) and the specimens were available for independent review by a pathologist, describing five histological categories, i.e. chronic interstitial inflammation, fibrosis, tubular atrophy, glomerulosclerosis and dysplasia. Histological lesions were categorized as severe (> 25%) or minimal (≤ 25%). Radiological features and the histology of UPN were then analysed, and correlated using Fisher's exact test and multivariate analysis.

RESULTS

Severe histological lesions were reported in 38 specimens (69%) and minimal lesions in 17 UPN specimens (31%). Severe parenchymal thinning was associated with a severe pathological lesion (P < 0.02). There was no correlation with the echogenicity or degree of hydronephrosis. Dysplasia, found in 64% of specimens, did not correlate with increased echogenicity. Minimally functioning upper poles on nuclear renal scan (≤ 4% overall function) were significantly associated with severe histological lesions (P = 0.01). There was no correlation between diagnostic imaging features and the mode of presentation or type of ureterocele.

CONCLUSIONS

Multivariate analysis, severe parenchymal thinning on renal US and minimal function on nuclear renal scan can be used to predict the severity of histological lesions of the upper pole. These results may be helpful in the making therapeutic decisions for patients with ureteroceles in duplex systems.

INTRODUCTION

Renal ultrasonography (US) and nuclear renal scintigraphy are complementary imaging techniques used to investigate the upper urinary tract of patients with ureteroceles in a duplex system. They provide information on morphology and function, which traditionally have been important in making therapeutic decisions. US is not invasive and provides morphological information. The degree of hydronephrosis, parenchymal thinning and increased echogenicity on US have often been associated with poor function and dysplastic changes [1–3]. Increased echogenicity is known to be present in 90% of medical renal disease, but it is not a specific finding and is associated with a heterogeneous group of genitourinary abnormalities that include obstructive uropathies, glomerular, tubular and interstitial infiltration, and vascular defects [4]. Furthermore, these assorted abnormalities differ in their potential for ultimate renal function. It is important to avoid confusion between increased echogenicity and accentuated corticomedullary differentiation; the latter is a normal finding in infant renal units. Conversely, isoechoic renal parenchyma in relation to the liver in premature infants and neonates may also be normal [5].

The pathogenesis of increased echogenicity is unclear but it has been associated with definite histological findings, e.g. the degree of interstitial inflammation, glomerular disease and vascular infiltration and sclerosis [6]. Increased echogenicity occurs because of the nonspecific response to many different changes in the structural echotexture of the renal tissue. Parenchymal thinning and the degree of hydronephrosis are often closely related and similarly, published data show a strong association between thinning and decreased renal function [7,8]. However, no objective analysis has been published to verify the association between renal histology and renal function.

Nuclear renal scans are widely used to provide functional information. The introduction of DMSA and MAG-3 renal scintigraphy to the available diagnostic methods has significantly improved the accuracy of measuring differential renal function compared with IVU.

Using a standardized and validated method of reporting histological lesions in upper pole nephrectomy (UPN) specimens, we systematically analysed renal US and renal DMSA scans, comparing the results with histological findings. Our aim was to correlate these variables to determine if diagnostic imaging can predict the extent of the histological lesions in UPN.

PATIENTS AND METHODS

From a surgical database at The Hospital for Sick Children we identified 86 patients who underwent surgery for a duplex system with ureterocele between 1992 and 2000, and for whom diagnostic imaging was available for review.

Preoperative imaging studies were independently reviewed by one genitourinary radiologist unaware of the clinical history, surgical procedures and the histological lesions to the upper pole. To avoid bias, all results from US were reviewed first, independent of any from nuclear scans. This was followed by interpretation of the nuclear renal scans. All patients also underwent voiding cystography but their interpretation was not part of this study.

The results were available for US from 84 (98%) patients; the radiologist reviewed the following variables: echogenicity (normal vs increased), parenchymal thinning (mild to moderate vs severe) and degree of hydronephrosis (Society of Fetal Urology grading system [9]). The kidneys were evaluated for echogenicity relative to the adjacent liver or spleen.

In all, 77 (90%) 99mTc-DMSA or -MAG-3 nuclear renal scans describing differential function of involved upper poles were available for review. Three patients had no preoperative renal scans at our institution and six had DTPA scans with no description of differential function of the upper pole. A poorly functioning upper pole was arbitrarily defined as ≤ 4% of overall function; the upper pole was preserved in 31 of the 77 patients.

In all, 154 moieties (77 each of upper and lower poles) were available to compare US findings with the respective differential function on renal scan. Individual diagnostic imaging features were also compared for mode of presentation and type of ureterocele. The classification system proposed by the Urology Section of the American Academy of Pediatrics was used, defining intravesical ureteroceles (within the bladder but not into the bladder neck mechanism) and extravesical or ectopic (ureterocele into or beyond the bladder neck), based on imaging studies and/or cystoscopic findings [10].

Fifty-five patients had a partial or total nephrectomy as part of their treatment, either initially or as a secondary procedure. The genitourinary pathologist, categorizing and grading the histological lesions, independently reviewed the 55 UPN specimens. The Banff Working Classification of Renal Allograft Pathology 1997 [11] was used to describe the histology, an international and clinically validated method for reporting renal histopathology.

Histological lesions were classified into five categories of chronic interstitial inflammation, fibrosis, tubular atrophy, glomerulosclerosis and dysplasia. Involvement of the pathological specimen by> 25% in one or more of these categories was defined as ‘severe histological lesions’. Conversely, an involvement by all categories combined of ≤ 25% was defined as ‘minimal histological lesions’.

Indications for surgery included a nonfunctioning upper pole and/or recurrent UTI (in 34), persistent significant hydronephrosis after endoscopic decompression (in 12) and UPN as part of a total reconstruction (in nine).

The results were assessed statistically using Fisher's exact test and multivariate analysis to correlate the category and severity of histological lesions with US characteristics and upper pole function, based on nuclear renal scans, with statistical significance defined as P < 0.05.

RESULTS

The study included 86 patients (62 girls and 24 boys, mean age at initial surgery 20 months, range 1–180); the presentation was antenatal in 36 (42%) and postnatal in 50 (58%), and the ureterocele intravesical in 29 (34%) and extravesical in 57 (66%). Thirty-one (36%) patients had preservation of their upper poles and 55 (64%) underwent UPN.

The results from US available in 84 patients were analysed, controlling for mode of presentation and position of ureterocele (intra/extravesical). Echogenicity, thinning of parenchyma and degree of hydronephrosis were equally distributed among the four groups. There was increased echogenicity in 40 (48%), severe thinning in 27 (32%) and grade 4 hydronephrosis in 50 (60%). There was no correlation with the mode of presentation. There was an insignificant association of more severe parenchymal thinning and severe hydronephrosis for the extravesical (P = 0.052 and 0.10, respectively) when compared with intravesical ureteroceles. On follow-up US, three of six antenatally diagnosed patients who preserved their upper poles had increased echogenicity initially, but this improved after endoscopic decompression.

Of the 77 renal scans reviewed, 23 (30%) showed residual function in the involved upper pole; the remaining 54 (70%) scans showed ≤ 4% function. There was no difference between the antenatally and postnatally diagnosed groups, with 23 of 31 (74%) and 31 of 46 (67%) having a poorly functioning upper pole, respectively. Renal scan function was also compared with the anatomical position; 17 of 24 (71%) and 37 of 53 (70%) had poor upper pole function for intravesical and extravesical ureteroceles, respectively. Follow-up nuclear renal scans were not routinely obtained in those patients who preserved their upper pole. Of these 31 patients, 13 had a poorly functioning upper pole before surgery and none of the six studies showed improvement afterward.

The histological lesions were no different when the antenatally and postnatally diagnosed groups were compared. Furthermore, there was no difference in the type of histological lesions between intravesical and extravesical ureteroceles. However, severe fibrosis and tubular atrophy were significantly more common for extravesical ureteroceles (P < 0.05), with glomerulosclerosis following this trend (P = 0.08).

We identified chronic interstitial inflammation, fibrosis, tubular atrophy and glomerulosclerosis in almost every specimen (96–100%). These categories were graded as severe (> 25% involved) in 55%, 67%, 66% and 53%, respectively. Dysplasia was severe in 12 (22%), minimal in 23 (42%) and absent in 20 (36%) specimens. There was no correlation with mode of presentation or type of ureterocele.

The US variables of the 53 patients who underwent UPN were initially compared with the histology by univariate analysis. The echogenicity was described as normal or increased and this feature was compared with the severity of the histological lesions of UPN specimens. Severe lesions tended to be associated with increased echogenicity (P = 0.10). Dysplasia, often associated with increased echogenicity on US, was not statistically significant (P = 0.15). The second US variable, parenchymal thinning, was described as mild-to-moderate or severe. Severe thinning was associated with a severe histological lesion (P = 0.01). The third variable, degree of hydronephrosis, did not correlate with more severe lesions. By multivariate analysis, only parenchymal thinning was significant (P = 0.01) but echogenicity still provided power if related to thinning.

The histology was also compared with the differential function of the affected upper pole on the 46 nuclear renal scans available. Poorly functioning upper poles (≤ 4%) were significantly associated with severe histological lesions (P = 0.01) but not to a specific pathological category of lesion. Of the five patients who had upper poles removed for indications of clinical reflux/pyelonephritis (reflux to the upper pole after a primary endoscopic decompression and recurrent UTI), despite the finding of some upper pole function (7–20%) on DMSA renal scan, four had minimal histological lesions and none had dysplasia.

In all, 154 moieties (77 upper and 77 lower poles) were available to compare US variables with their respective differential function on renal scans. Each US variable independently correlated with poor function of the moiety, but the coexistence of increased echogenicity, severe parenchymal thinning and grade 4 hydronephrosis were highly predictive of poor function on nuclear renal scan, with a specificity of 91% and a positive predictive value of 94% (P < 0.001) (Figs 1 and 2). Furthermore, the association of poor US values and poor differential function of the upper pole was a strong predictor of severe histological lesions (P < 0.001).

Figure 1.

Diagnostic imaging of a 12-month-old boy with an extravesical ureterocele in a left duplex system, with preserved function of the affected upper pole: a, renal ultrasonogram showing normal echogenicity, grade 3 hydronephrosis and mild-moderate cortical thinning; b, DMSA renal scan showing function in the upper pole (20%) accounting for 10% of the overall function.

Figure 2.

Diagnostic imaging of a 3-month-old girl with an extravesical ureterocele in a right duplex system with a nonfunctioning upper pole: a, renal ultrasonogram showing increased echogenicity, grade 3 hydronephrosis and moderate cortical thinning; b, DMSA renal scan showing amputation of the upper pole.

DISCUSSION

The clinical management of ureteroceles in duplex systems remains a challenge, as no two patients appear to have identical combinations of age, ureterocele position, size, total and differential function, and reflux. Furthermore, the precision and reliability of the diagnostic imaging used in their evaluation is often doubted by investigators, especially when assessing differential function of the putatively obstructed upper pole [1,3,12–15]. Earlier series relied on visualizing the upper pole on IVU and reported improved function after surgery, based on the same test [12,15,16]. The precision of IVU in assessing function is not as reliable as nuclear renal scintigraphy. Since the advent of nuclear renal scans (DMSA and MAG-3) differential function can be measured with some confidence, even for an obstructed system, and many have encouraged its use [2,3,17–19].

A principal goal in treating duplex system ureteroceles is preserving renal function. Most authors prefer reconstruction to ablative surgery when the affected upper pole provides clear and significant function [18,20,21]. The main dilemma occurs when faced with a poorly or nonfunctioning upper pole. There is published evidence that these upper poles have severe irreversible histological lesions. Dysplasia is reported to be present in 40–70% of specimens, in addition to chronic interstitial inflammation, fibrosis and other changes [13–16,22]. We confirmed these observations with an exhaustive standardized review of 55 specimens, showing that 69% had severe histological lesions and 64% had dysplasia.

In the present study, diagnostic imaging was compared with the UPN specimen for histology, confirming that a poorly or nonfunctioning upper pole on DMSA scan has severe irreversible histological lesions. Five patients had residual function in the affected upper pole and most had only minimal lesions. This result reinforces the findings of Coplen and Duckett [1], that a preoperative renal scan is important before a staged reconstruction. A poorly functioning upper pole seems to harbour more severe irreversible lesions, with little or no potential for recovery. From our review, the DMSA renal scan is a reliable and accurate preoperative tool to predict histological lesions.

The other imaging study used to investigate the upper tract is US, which has replaced IVU; US has become the predominant study to antenatally diagnose genitourinary anomalies. In the present review, parenchymal thinning on US significantly correlated with severe histological lesions, although echogenicity did not. The assumption has been that increased echogenicity is associated with dysplasia [1]. The present study showed a tendency to this association but larger series are warranted to further delineate it. The imaging studies analysed in this review were undertaken before any manipulation (i.e. incision) of the ureterocele itself. As histopathology is the ultimate arbiter of function, those cases where imaging correlates with poor histology reflect, according to the present data, an untouched ureterocele/duplex system with no biological potential for recovery of upper pole function. Thus, from a management perspective, such imaging findings should provide grounds for avoiding routine ureterocele incision unless it is being used to treat a dilated upper pole ureter complicated by infection.

By comparing US with differential function on renal scan for 154 moieties, we confirmed an important role for US; it can be used to predict histological lesions and differential function. It is probably the most valuable imaging study for the follow-up of upper tracts, being noninvasive and readily available. However, isoechoic or hyperechoic renal parenchyma in the neonatal period can be normal; with no associated severe parenchymal thinning, it has a chance to improve during the follow-up.

In conclusion, the histology of UPN specimens can be documented using a standardized validated method. We confirmed that severe and irreversible lesions occur in 69% of such cases. By having one radiologist independently review preoperative renal US and DMSA renal scans, we confirmed the predictive value of diagnostic imaging. Renal US documenting severe parenchymal thinning, and nuclear renal scans confirming a poorly or nonfunctioning upper pole, are reliable tests for predicting severe histological lesions in upper poles. Furthermore, renal US can also be used to predict differential function on nuclear renal scans. As histopathology is the final arbiter of functional potential in kidney tissue, these data provide the first direct evidence that renal imaging reflects renal histology. Such knowledge should provide renewed interpretative potential for renal imaging in children, and may be of help in making therapeutic decisions about duplex system ureteroceles. A long-term follow up and further studies are warranted to determine if the poorly or nonfunctioning upper pole remaining in situ, presumably with severe lesions, is of any risk for the child.

Abbreviations
US

ultrasonography

UPN

upper pole nephrectomy.

Ancillary