1 Background

The lifetime prevalence of urinary stones is about 10–14% [1]. Stone composition significantly affects both treatment outcomes and preventive measures [2]. Nearly no currently available stone analysis methods can facilitate preoperative assessment. The attenuation value in some radiological modalities can predict the stone composition. There are two ways to measure the attenuation value, density by non-contrast CT (NCCT), and radio-opacity by plain X-ray KUB (PKUB) [3].

The Hounsfield unit (HU) determined by the NCCT was suggested as a useful predictor of the urinary stone compositions [4]. HU is a normalized index of X-ray attenuation, named after Sir Godfrey Newbold Hounsfield, based on a scale of − 1000 (air) to + 1000 (bone), with water being zero [5]. A previous study classified the stone's radio-opacity in PKUB in comparison with the opacity of the last rib into 4 categories: dense opaque, opaque, faint opaque, and radiolucent [6].

There are many techniques for stone analysis; chemical analysis, thermogravimetric analysis, polarizing light microscopy, scanning electron microscopy, infrared spectroscopy, X-ray powder diffraction analysis (XRD), and energy-dispersive X-ray analysis [7].

Our study aims to see if urinary stone composition could be predicted by PKUB and/or NCCT characteristics validated by in vitro XRD analysis.

2 Methods

This is a prospective hospital-based cross-sectional study. The study included 100 cases with urinary tract stones attending our department between April 2014 and March 2016. The sample size was calculated using Epi-Info™ 7.1 for Statistical Calculation using the following parameters: power of the study 80%, population size 20,000, and expected frequency 50%, the margin of error 9%, and confidence level of 97%. This yielded a sample size of 99. The sample size in our study is 100 stone samples. The study was submitted to and approved by our institutional ethics committee.

Stone radio-opacity in PKUB was determined in comparison with last rib opacity into; more opaque than the last rib (dense opaque), nearly equal to last rib opacity (opaque), less opaque than the last rib (faint opaque) or radiolucent determined by two investigators (M.G., A.E.). The patients scanned with the use of a second-generation dual-energy multi-detector CT scanner, using our clinical urinary stone protocol: 80 and 120 kV with tin filter, 240 quality reference mAs, and 1 mm collimation, 0.5 s rotation time, 64 slides, and stone density (HU) reported.

2.1 The technique of XRD analysis

Stones are retrieved either after lithotripsy, extracorporeal shockwave lithotripsy (ESWL), or open surgery. XRD is a tool used for identifying the atomic and molecular structure of a crystal, in which the crystalline atoms cause a beam of incident X-rays to diffract in many specific directions. The distance between similar atomic planes in a mineral, the inter-atomic spacing, called the d-spacings, is measured in angstroms. Sixty-Nine Files of d-spacings for hundreds of thousands of inorganic compounds are available from the International Centre for Diffraction Data (ICDD). Many other sites contain d-spacings of minerals such as the American Mineralogist Crystal Structure Database (AMCSD) [8].

The urinary stone sample was pulverized and analyzed by XRD using Philips PW 1710 X-Ray Diffractometer (Japan) with Beta filtering using a graphite monochromator. The crystalline components were identified using the ICDD database and the semiquantitative composition determined using the relative intensity of the different bands.

A few grams of the stone powdered and mounted uniformly on a glass slide assuring a flat upper surface, and then rotated in the collimated X-ray beam of the XRD unit. The diffraction pattern is recorded using a cylindrical camera. Once all d-spacings of a certain substance have been determined, manual or automated search–match routines compare the d-spacings of the unknown to those of known materials derived from the ICDD cards.

2.2 Statistical analysis

We encoded all data in the Microsoft Excel spreadsheet and performed statistical analysis using IBM SPSS program version 20. The data are tested for normality using the Anderson–Darling test and for homogeneity variances before further statistical analysis. Categorical variables are described by number and percent, while continuous variables are described by mean, standard deviation, and median. Chi-square and Fisher exact tests were used to compare categorical variables. Comparison between continuous data was performed using t test and ANOVA (for normal distribution variables). Mann–Whitney U and Kruskal–Wallis H were used for the nonparametric variables. The receiver operating characteristic (ROC) curve was constructed to determine the best cutoff value.

3 Results

The study included 100 stones from 100 patients (74 males and 26 females). The median age was 32 years (range 2–70). The mean BMI was 27.3 kg/m2. The following interventions were performed, PCNL for 46 patients, open renal surgery for 23, ureteroscopy for 21, and 10 cases were post-ESWL. The mean stone size was 2.9 ± 1.3 cm.

XRD revealed only 3 (3%) stones were of pure component (two calcium oxalate monohydrate COM stones and one cystine stone). The remaining 97 (97%) stones were of the mixed component. There were no brushite or calcium carbonate stones. Table 1 shows the distribution of the major components of the mixed stones.

Table 1 Distribution of mixed stones
Full size table

In PKUB, the degree radio-opacity was as follows; 30 stones were dense opaque, 44 opaque, 21 faint opaque, and 5 radiolucent. There was no significant correlation between the BMI and the stone opacity by PKUB (P = 0.59). Most of the calcium oxalate monohydrate (COM) stones (60%) were dense opaque, all urate stones were radiolucent, and most of the struvite stones (75%) were faint opaque.

In NCCT, the mean (SD) HU was 1195.3 (146.8) for COM stones, 443.8 (99.3) for urate stones, and 527.8 (138.5) for struvite stones. Calcium oxalate dihydrate (COD) and calcium phosphate (CaPo4) stones are misidentified by PKUB and NCCT (no significant relation).

Table 2 shows that there was a significant relation between COM, struvite, and urate stones and their densities in HU by NCCT. The more the HU density of the stone, the more liability to have COM as a major component, while the less HU density of the stone, the more liability to have urate and/or struvite as the major components.

Table 2 Relation between density of the stone in HU, size of the stone, and its PKUB findings
Full size table

Also, there was a significant relation between COM, struvite, and urate stones with their degrees of opacity in the PKUB. Most of the COM stones were dense opaque, while all the urate stones were radiolucent in the PKUB. Most of the struvite stones were faint opaque in the PKUB. However, about 25% of struvite stones included in this study were radiolucent and overlapped with urate stones as shown in Table 3.

Table 3 Relation between type of the stone and its KUB findings
Full size table

To determine the NCCT sonogram HU attenuation level with both the highest specificity and sensitivity, the study applied the receiver operating characteristic (ROC) curve to set the best cutoff value of HU. The cutoff value of HU density by NCCT to the dense opaque stones in the PKUB was more than 1020 (Fig. 1).

Fig. 1
figure 1

ROC curve of dense opacity in PKUB about density in Hounsfield units

Full size image

Stones of HU density more than 590 were mostly radiopaque stones and those of HU density less than 590 were mostly radiolucent stones (Fig. 2). Figure 3 shows that most of the stones of HU density between 590 and 680 were faint opaque stones in PKUB. Figure 4 shows that most of the stones of HU density between ≤ 590 were radiolucent stones in PKUB.

Fig. 2
figure 2

ROC curve of opaque stone in PKUB regarding its density in Hounsfield units

Full size image
Fig. 3
figure 3

ROC curve of faint opacity in plain PKUB about the density of the stones in Hounsfield units

Full size image
Fig. 4
figure 4

ROC curve of radiolucency in the PKUB about its density in Hounsfield units

Full size image

Table 4 represents multiple linear regression analysis performed to assess the most significant predictors of the stone composition. It revealed a significant correlation between COM, struvite, and urate stones with their degrees of radio-opacity by PKUB, and their HU densities by NCCT were found as shown in Table 4. In other words, the presence of radiolucent stones on PKUB could predict a urate composition in 40% of cases. Besides, the presence of dense opaque stone on PKUB could predict a COM stone composition in 75.9%. The presence of faint opaque stone on PKUB can predict a struvite stone composition in 15%.

Table 4 Multiple linear regression analysis to assess the most significant predictors of stone composition
Full size table

4 Discussion

Preoperative prediction of the stone density and composition is of utmost importance affecting both the treatment option and the prevention measures [9]. Correlation of the CT stone attenuation level with the stone composition is studied extensively in the recent literature; however, no data are available in the literature about the correlation with PKUB findings [10].

An in vivo study evaluated NCCT in predicting stone composition and found a significant difference between the HU measurements of uric acid and calcium oxalate stones [11]. Another study suggested that ‘attenuation/stone size ratio’ was an important predictor in differentiating uric acid and calcium oxalate stones (3). Others stated that for calcium stones, the ability of NCCT (density in HU) to predict stone composition was limited, likely due to the mixed stone composition [12]. An ex vivo study stated that the dual-source CT simultaneously operated in the dual-energy mode allows for the accurate differentiation between uric acid (UA)-containing and non-UA-containing urinary calculi using XRD as the gold standard [13].

In our study, the stone size was not a significant factor in the prediction of the stone composition by the NCCT or PKUB, but there was a significant relation between stone type (COM, struvite, and urate stones) and their densities in HU by NCCT.

The correlation between the stone compositions by infrared spectrometry and stone morphology, location, and size by plain PKUB studied and identified five different patterns of radiographic appearances of the stones. They concluded that the prediction of stone composition from stone morphology on plain PKUB may not be accurate enough [14].

We found that the cutoff point of HU density by NCCT as regarding radio-opacity in PKUB is 590 HU. Stones of HU attenuation values more than this level are radio-opaque and vice versa. Some authors stated that redoing the XRD phase analysis after heat treatment considerably enhances the capabilities of the method [15].

As we get some stones after lithotripsy, parts of the stone for analysis will be missing, and stone analysis may not reflect the complete picture; this is a limitation of our study. Another limitation that needs further investigation is the relationship between bone density, BMI, and skin-to-stone distance and the categorization of stone opacity in PKUB.

A recent study revealed that PKUB has reasonable accuracy for urolithiasis > 5 mm composed of pure calcium salts, particularly those of the upper renal tract [16]. Another study stated that the use of HU stone density by NCCT revealed statistically significant differences among all pure and most of the mixed urinary stones; however, differentiation of pure COM and mixed COM was possible only in the tissue window [17].

5 Conclusion

Stone radio-opacity by PKUB and its attenuation value by NCCT could successfully predict its calcium oxalate monohydrate, struvite, and urate composition and hence its treatment option. However, the chemical stone analysis is still required as most stones are mixed.