CT Grading in Upper Urinary Tract Calculi with Kidney Infection

Background This study is to establish a CT imaging grading system and to explore its value in evaluating upper urinary tract calculi with kidney infection. Methods CT images of 126 patients with kidney infection caused by upper urinary tract calculi were retrospectively analyzed. CT grading system was developed based on CT images. General information, clinical symptoms and examination data of patients in different CT grades were analyzed. With the occurrence of systemic inammatory response syndrome (SIRS) as the end point, univariate and multivariate analysis was conducted to analyze the risk factors of SIRS. Results CT images were classied into 4 grades. For patients with different CT grades, the factors of fever, diabetes, blood white blood cell counts, urine leucocytes, CT1 (ratios of renal cortex CT values on the affected side and those on the healthy side), CT2 (ratios of renal pelvis CT values on the affected side and those on the healthy side), maximum body temperature, disease time, proportion of blood neutrophils, C-reactive protein, procalcitonin, and the size of stones had statistical signicance (P<0.05). Only CT grading was statistically signicant by multivariate analysis. The higher the CT grading, the greater the risk of SIRS, with 4.472 times higher of SIRS risk for each increasement of the CT grade. Conclusions CT images may reect the involvement extent and scope of upper urinary tract calculi with kidney infection. The higher the CT grade is, the more severe the clinical symptoms are and the greater the risk of SIRS is. with corresponding grading CT renal pelvis, 4 on the CT compared clinical laboratory tests,


Introduction
Upper urinary tract calculi are common in the urinary system. Obstruction caused by upper urinary tract calculi may lead to kidney infection, which may develop into renal abscess if not detected in time or poorly understood, and will lead to septicopyemia if severe [1]. There may be no obvious symptoms even for patients with severe hydronephrosis, which always results in delayed diagnosis and treatment [2]. Currently, the existing imaging examinations mainly focus on the obstruction location of upper urinary tract calculi and the size of the calculi [3,4]. However, kidney infection caused by obstruction has not received enough attention. Clinical evaluation of the severity of kidney infection is often based on the symptoms, signs, blood routine examination, in ammatory indicators and other examinations [5,6]. The direct and objective imaging diagnostic criteria are lacking.
In recent years, the application of multi-slice spiral CT in kidney diseases has been increasing. CT imaging has not only improved the image resolution, showing more three-dimensional images of kidney anatomy and morphological changes, but also has a marked effect on the detection rate and qualitative diagnosis of kidney diseases [7,8]. The main advantages of CT are signi cant improvement of soft tissue contrast, which may directly and accurately show stones and clearly show tiny stones [9,10]. Ureteral obstruction causes urine to accumulate in the kidney or above the ureteral obstruction section, thus causing kidney infections. CT scans can clearly show the extent and characteristics of local in ammatory lesions at the ureteral obstruction [11]. However, the clinical diagnosis of obstructive kidney infection is usually based on the patient's symptoms and laboratory tests, but not on CT imaging, which makes it impossible to effectively evaluate the degree and extent of kidney infection.
Systemic in ammatory response syndrome (SIRS) is a systemic in ammatory response caused by infectious or noninfectious factors. It is an immune disorder caused by the excessive release of various in ammatory mediators in the body.
If SIRS is not controlled in time, it can progress to sepsis. According to statistics, 30% -50% of SIRS have further developed into septic shock and multiple organ dysfunction syndrome. Therefore, the early diagnosis and treatment of SIRS is very important for the prevention and treatment of sepsis. This is also one of the important issues that urgently need to be solved in clinical urology. CT examination CT Scanning Apparatus Lightspeed VCT XT 64-slice spiral CT scanner and CT quality analysis specialized water phantom were all from GE Healthcare (USA). Before CT examination, the patients were requested to drink water for lling of the bladders. A 64-slice spiral CT scanner was used for plain scanning, in a scan range from diaphragm to pubic symphysis, and, with a thickness of 5mm and a reconstruction interval of 1.25mm. The scanning data was exported to the workstation for the three-dimensional reconstruction. The maximum intensity projection, curved planar reformation, volume reformation and multi-planar reconstruction were adopted to display the three-dimensional images and lesions of the urinary tracts. Meanwhile, the relation between lesions and surrounding anatomical structures was observed. The images were analyzed by 2~3 senior radiologists. According to the scope and extent of in ammation exudation involving the kidneys, CT evaluation criteria were developed and CT images were classi ed into different grades.

CT grading
According to the scope and extent of in ammation exudation involving the kidney suggested by CT, we divided the lesions into four grades ( Figure 1). In Grade 1 ( Figure 1A), the CT signs included thickened ureter wall and vague edema, thickened wall of pelvis and renal calices, mild dilation of pelvis and renal calices, vague fat space of sinus renalis, but not involving renal parenchyma. In Grade 2 ( Figure 1B), the CT signs included rough in ammation exudation of kidney capsules, increased density of perirenal fat space, and thickened renal fascias. In Grade 3 ( Figure 1C), the CT signs included locally thickened renal parenchyma with decreased density, local wedge-shaped swelling and low-density focuses in renal parenchyma, with the scope less than or equal to 50%. In Grade 4 ( Figure 1D), the CT signs included the area of involved renal parenchyma greater than 50%, or with emphysema, or renal and perinephric abscesses.

Indicator measurement and calculation of CT values
CT images of the lateral axis plane of renal hilum were taken, and the region of interest (ROI) was manually selected to measure the double renal parenchyma value (Hu). ROI were all measured for 3 times and averaged, greater than 100mm 2 for each one. Meanwhile, CT values of bilateral renal pelvis hydrops were measured. The ratios of renal cortex CT values on the affected side and those on the healthy side (CT1), and the ratios of renal pelvis CT values on the affected side and those on the healthy side (CT2) were calculated, and the CT1 and CT2 of different CT grades were compared.
Evaluation Indicators: general data, clinical symptoms and examination data General data of patients, including age, gender, and stone location were collected. Clinical symptoms and examination data included fever (body temperature > 38℃), maximum body temperature, renal colic, diabetes, disease time, stone size, CT1, CT2, white blood cells, blood neutrophil percent, C-reactive protein (CRP), procalcitonin (PCT), midstream urine bacterial culture, blood culture and urine leukocytes. All the general data, clinical symptoms and examination data of patients with different CT grades were compared.

Analysis of SIRS risk factors
The occurrence of systemic in ammatory response syndrome (SIRS) was considered as the end point of observation. The diagnosis of SIRS required at least two criteria as follows: 1) body temperature higher than 38℃ or lower than 36℃. 2) heart rate over 90 times/min. 3) respiratory rate higher than 20 times /min or arterial blood partial pressure of CO 2 lower than 32mmHg or mechanical ventilation. 4) white blood cell counts higher than 12×10 9 /L or lower than 4×10 9 /L, or immature white blood cells higher than 10%. Univariate and multivariate analysis was conducted to analyze the risk factors of SIRS in patients with upper urinary tract calculi.
Statistical analysis SPSS Statistics 23.0 (IBM, American) was used for data analysis. Quantitative data conforming to normal distribution were shown as mean ± standard deviation (SD). The t-test was used for difference comparison between two groups of independent sample data, and one-way analysis of variance was conducted for data of more than two groups. The quantitative data that did not conform to normal distribution were described by medians (inter-quartile range) and compared by non-parametric test. Chi-square test was adopted for analyzing enumeration data. The correlation analysis between quantitative data and qualitative data was performed by Eta coe cient. Multivariate analysis of dichotomous data was performed by logistic regression analysis. A p value <0.05 was considered as statistically signi cant.

CT images
Ureter signs: Ureters above the obstruction plane of the affected side were dilated with hydrops. Ureter wall was slightly thickened. Ureter wall in stone plane were dropsical and vague, and adjacent fascias were thickened.
Signs of renal pelvis: There was thickening of the renal pelvis and calvarial wall. The renal pelvis and calices were dilated.
The density of dilated renal pelvis was higher than or equal to that of the healthy side.
Signs of renal parenchyma: Local wedge-shaped swelling diffused from the collecting system to the renal capsules, or in diffuse swelling. The boundaries of the renal cortex and medulla of the swollen part were vague. The swollen part showed slightly high density, and the density of the diseased part was decreased.
Signs of renal capsules and perirenal images: There were rough renal capsules, and thickened perirenal anadesmas and peri-renal space fat, with patchy and banded high density.

Comparison of general data of patients with different CT grades
The clinical data of patients were shown in Table 1. Based on the CT grading system, there were 48 cases of grade 1 (accounting for 38.1%), 30 cases of grade 2 (23.8%), 24 cases of grade 3 (19.0%) and 24 cases of grade 4 (19.0%). There were no signi cant differences in age, sex, side and stone position of patients with different CT grades (P > 0.05).
Comparison of clinical symptoms and examination results of patients with different CT grades Patients with different CT grades had no statistically signi cant differences in renal colic, urine bacteria culture positive rate or blood culture positive rate (P>0.05, Table 2), but they had statistically signi cant differences in fever, diabetes, blood WBC counts, urine leucocytes, CT1, CT2, maximum body temperature, disease time, proportion of blood neutrophils, CRP, PCT and the size of stones (P<0.05, Table 2). Thus, there was a difference of conditions among patients with different CT grades.

Correlation analysis of clinical symptoms and examination data indicators with CT grades
Eta coe cient was used to analyze the correlation of clinical symptoms and examination data indicators (including CT1, CT2, CRP and PCT) with CT grades. Using CT grades as the dependent variable, the Eta coe cients of CT grades with CT1, CT2, CRP and PCT were 0.968, 0.782, 0.965 and 0.887, respectively (Table 3), indicating high correlation. When CT1, CT2, CRP and PCT were used as dependent variables, the Eta coe cients of CT grades with CT2 and PCT were less than 0.5, while Eta coe cients of those with CT1 and CRP were both higher than 0.7, also indicating high correlation.
Analysis of SIRS risk factors SIRS occurred in 36 of the enrolled patients. Univariate analysis was used to analyze the risk factors of SIRS. As shown Table 4, the CT1, CT2, CRP, PCT, proportion of neutrophils and CT grades between SIRS group and non-SIRS group was statistically signi cant (P < 0.05). CT1 in the non-SIRS group was signi cantly higher than that in the SIRS group, and CT2, CT grades, CRP, PCT and proportion of neutrophils in the SIRS group were signi cantly higher than those in the non-SIRS group. There was no statistical signi cance in stone size, disease time, blood culture, urine culture, diabetes, and renal colic between the SIRS group and the non-SIRS group (P > 0.05, Table 4).
In the study, dichotomous logistic regression was used for multivariate analysis. Indicators with signi cant P values in univariate analysis were included in logistic regression. To avoid missing signi cant in uencing factors, P values were set to be less than 0.1. Therefore, CT1, CT2, CRP, PCT, urine culture, neutrophil granulocyte percent and CT grades were used as independent variables in logistic regression analysis. The tolerances were all much greater than 0.1, and the variance in ation factors were all less than 10. Thus, there was no multicollinearity.
Finally, the logistic model was statistically signi cant (χ2 = 41.380, P <0.001). At the same time, the results showed that the model correctly distinguished 90.5% of the research subjects, with a sensitivity of 82.4% and a speci city of 93.5% Among the 7 independent variables included in the model, only CT grade was statistically signi cant, which was an independent factor in uencing SIRS. According to the values of partial regression coe cient (B), the higher the CT grade was, the greater the risk of SIRS was. The risk of SIRS was 4.472 times higher as each increasement of the grade (Table 5).

Discussion
Obstruction of upper urinary tract by calculi can cause urine accumulation, lead to bacterial proliferation and kidney infections. This may further lead to rapid destruction of kidney tissue structure. In recent years, the diagnostic accuracy of CT examination for calculi has reached 95% ~ 100% [12], and it has been used as a routine preoperative examination of calculi patients. However, the diagnosis for kidney infection by CT has not been paid much attention in clinical practice. Therefore, this study aims to at establish a CT imaging grading system and analyze its clinical value in evaluating upper urinary tract calculi with kidney infection.
The normal intrapelvic pressure is 4-10cm H2O. In the early stage of acute ureteral obstruction, the renal pelvis and calves are dilated and stagnant, manifested by the thickening of the renal pelvis and calves and the slight expansion of the renal pelvis and calves on CT images [13]. When the pressure continues to rise, the effusion will enter the renal sinus through the calyx. In plain CT scan, the fat in renal sinus is di cult to be distinguished from renal parenchyma [14], especially during edema, both manifesting as thickening of renal parenchyma. Only when uid accumulated within the renal sinus at the renal hilum can the separation of renal sinus edema and vague fat space be found in CT scan. In this study, patients with dilation and hydrops of the renal pelvis and calyces, thickened walls of renal pelvis and calyces, and visible obese fat space were classi ed as CT grade 1.
When pressure in the renal pelvis is signi cantly increased, urine containing pathogenic bacteria and in ammatory cells may in ltrate from the renal interstitium or lymphatic vessels into the subrenal capsule or perirenal space, causing thickened and coarse kidney capsules, increased density of the perirenal space and thickened renal fascias [15]. According to studies from Raptopoulos and other scholars [16], there are multi-layers of fat space in the perirenal space, which are called bridge septa. The function of bridge septa is mainly to buffer the kidney pressure, and to drain uid, pus and urine beneath the renal capsule and the renal sinus. Under normal circumstances, it is di cult for CT scan to identify the bridge septa, but there is thickening on CT if the bridge septa drainage results in increased hydrops during perirenal space edema [17]. Therefore, we de ned patients with CT manifestations of in ammatory exudation of the renal capsule, increased density of the interrenal fat space, and thickened renal fascia as CT grade 2. The reasons are considered as follows: pathogenic bacteria and in ammatory cells may spread to the perirenal fascia through renal interstitial and lymphatic vessels. Renal parenchyma may be not yet involved, with less blood supply to the perirenal fascia. Thus, pathogenic bacteria and in ammatory cells are less likely to enter other body parts. The results of this study showed that the clinical symptoms and rise of in ammatory indicators of patients in CT grade 2 did not increase signi cantly.
Hydrops in the renal pelvis and calyces increases, and the pressure increases accordingly, therefore, urine within the renal sinus further enters the renal tubules through the corresponding renal papilla of the renal calyces, thus reversely owing to the renal parenchyma [18]. Acute in ammatory reactions occur in the renal parenchyma, showing aggregation and in ltration of in ammatory cells, telangiectasia, edema of localized or patchy tissues, presence of edema and slight ischemic areas and presence of diffuse or punctate distribution [19]. On plain CT images, these features are manifested as kidney enlargement, thickened renal parenchyma, decreased density, and unclear decomposition of renal cortex and medulla [19,20]. Multiple abscesses or infarct focuses are visible, characterized by one or more wedge-shaped banded lowdensity focuses extending from the renal papilla to the renal cortex, in striated signs [19,20]. Thus, we classi ed the patients with CT manifestations of local thickening of the renal parenchyma, decreased density, unclear boundary of the renal cortex medulla, and local wedge swelling of the renal parenchyma and low-density foci as CT grade 3.
Emphysematous pyelonephritis or renal abscess may occur with the aggravation of pathological changes or because of the highly toxic pathogenic bacteria [21]. Emphysematous pyelonephritis is a rare urological emergency, presenting with in ammatory necrosis of acute renal parenchyma and perirenal tissues, and characteristic emphysematous of renal parenchyma, perirenal tissues, and collecting systems [22]. Emphysematous pyelonephritis is a fatal disease, and if not diagnosed and treated in a timely manner, it may rapidly develop to septic shock, with a mortality rate of over 50% [23].
Therefore, in this study, renal abscess and emphysema pyelonephritis were classi ed as CT grade 4.
Paick SH et al [24] classi ed CT images into 4 grades according to renal parenchymal involvement, including no renal parenchymal involvement (grade1), less than 25% involvement (grade 2), 25% or greater involvement but less than 50% involvement (grade 3), and 50% or greater involvement (grade 4). The higher the grade was, the more severe the clinical symptoms were. However, the grading system was established based on the CT enhancement scan. Numerous patients with upper urinary tract calculi are complicated with fever and are not suitable for injection of contrast agents. Thus, CT enhancement scan was not possible in these patients. In addition, pathogenic bacteria and in ammatory cells could in ltrate into subrenal capsules or perirenal space through renal interstitium or lymphatic vessels, which were not included in their grading system. In this study, plain CT scan was adopted, which was easy to be carried out in clinical practice. Plain CT images of upper urinary tract calculi with kidney infection were collected and analyzed, and corresponding grading system was established. According to CT images of renal pelvis, renal parenchyma and perirenal involvement, CT was classi ed into 4 grades. Finally, based on the CT grades, we compared the clinical manifestations, laboratory tests, and prognostic outcomes of patients at different levels.
Urinary leukocyte is an important indicator of the in ammatory response after bacterial infection in urine [25]. The results of our study showed that urine-leukocyte-positive patients had higher CT grade. The reason may be that the more in ammatory mediators and in ammatory cells caused by pathogenic bacteria in the urine, the more they ow back into the renal parenchyma, and the more severe the damage to the renal parenchymal structure. However, some of the patients with CT grade 4 did not have leukocytes in the urine. We suppose that this may because of the complete obstruction of urethra by stones. Nevertheless, such patients often were in more serious conditions. The positive urine culture results among different grades showed no statistical signi cance, which may be due to small sample size and the small number of urine-positive samples of various grades. Koh et al [26] found that diabetes was a high-risk factor for incidence of patients with urinary sepsis, and pathogenic bacteria were more likely to reproduce in the environment of high blood glucose. The results of the study showed that patients with diabetes were in higher CT grade.
The CT value of the substance re ects the density of the substance, that is, the higher the CT value, the higher the density. When pathogenic bacteria, in ammatory mediators and in ammatory cells involve the kidney, the density of the involved parenchymal kidney decreases, and the CT value decreases correspondingly [27]. To avoid individual difference, the ratio of CT values of the affected side and that of the healthy side was calculated in this study to compare the CT value differences of patients with different CT grades. The results showed that the lower the ratio of CT values of the affected kidney cortex to that of the healthy kidney cortex was, the higher the CT grade was, and the difference was statistically signi cant. The Eta correlation analysis showed that CT1 was correlated with CT grading with Eta coe cients all higher than 0.7, that is, the ratio of CT values of the affected renal cortex to that of the healthy renal cortex was negatively correlated with the CT grading. With high degree of correlation, the lower the ratio was, the higher the CT grades were. Urine containing pathogenic bacteria accumulates in the renal pelvis and it produces an in ammatory response, resulting in an increase of pathogenic bacteria, in ammatory mediators and in ammatory cells, as well as an increase in density, and a corresponding increase in CT values of the renal pelvis [28]. The results of this study showed that the ratio of CT values of the affected renal pelvises with those of the healthy renal pelvises were different in patients in different CT grades.
When there are pathogenic bacteria, the kidneys produce an in ammatory response under the stimulation of pathogenic bacteria and their toxic products. In addition, pathogens and their toxic products may directly enter the circulatory system through the renal parenchyma, thereby causing a systemic in ammatory response and increased production of in ammation indicators [29,30]. It is found that CT images accurately show the degree of kidney infection, and that patients with more severe kidney infections have more severe clinical symptoms [31]. Our results showed that the blood white blood cell count, blood neutrophil ratio, CRP and PCT of patients with different CT grades were signi cantly different. The correlation analysis results of Eta coe cient also showed that the Eta coe cients of CRP and CT grades were higher than 0.7, indicating a high degree of correlation. The higher the CT grade, the more severe the kidney infection, and the higher the CRP index of systemic in ammation. The univariate analysis results showed that CRP was a risk factor for SIRS. CRP may be used as a reference index for predicting SIRS at early stage.
With the presence or absence of SIRS as the end point of observation, univariate and multivariate analysis was conducted on the risk factors in uencing the occurrence of SIRS in patients with upper urinary tract calculi. Univariate analysis results showed signi cant differences in CT1, CT2, CRP, PCT, neutrophil ratio and CT grading between the SIRS group and the non-SIRS group. Multivariate analysis results showed that only CT grading was statistically signi cant, which was an independent factor in uencing SIRS. According to the values of partial regression coe cient (B), the higher the CT grades were, the greater the risk of SIRS was. The risk of SIRS was 4.472 times higher as each increasement of the grade.
Therefore, results of the study showed that the higher the CT grade was, the greater the risk of SIRS in patients with upper urinary tract calculi were, which provides clinical guidance for early clinical diagnosis and treatment of SIRS. Kim et al. showed that CT image grading could be used to predict the severity and course of disease [32]. The higher the CT grades were, the worse the clinical symptoms and the prognosis were [32]. By CT grading, we evaluated the degree of kidney infection and our results may better guide the treatment. For severe kidney infection caused by upper urinary tract calculi, priority should be given to the removal of obstruction and effusion drainage before the second stage of stone treatment until infection control.
This study has some limitations. For example, the retrospective study and small sample size may affect the true level of CT grading and disease severity changes. Thus, results of the study need further veri cation by multi-center and largesample clinical data.
In conclusion, CT images may re ect the involvement and scope of upper urinary tract calculi with kidney infection. The

Consent for publication
Written informed consents were obtained from every patient.

Availability of data and materials
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests.  Notes: CRP, C-reactive protein; PCT, procalcitonin; CT1, the ratios of renal cortex CT values on the affected side and those on the healthy side; CT2, the ratios of renal pelvis CT values on the affected side and those on the healthy side. Table 4. Univariate analysis of SIRS in patients with upper urinary calculi.
Notes: * P<0.05. CRP, C-reactive protein; PCT, procalcitonin; CT1, the ratios of renal cortex CT values on the affected side and those on the healthy side; CT2, the ratios of renal pelvis CT values on the affected side and those on the healthy side; SIRS, systemic inflammatory response syndrome.