Combining Mean and Standard Deviation of Hounsfield Unit Measurements from Preoperative CT Allows More Accurate Prediction of Urinary Stone Composition Than Mean Hounsfield Units Alone

Abstract

Introduction and Objectives:

The mineral composition of a urinary stone may influence its surgical and medical treatment. Previous attempts at identifying stone composition based on mean Hounsfield Units (HUm) have had varied success. We aimed to evaluate the additional use of standard deviation of HU (HUsd) to more accurately predict stone composition.

Methods:

We identified patients from two centers who had undergone urinary stone treatment between 2006 and 2013 and had mineral stone analysis and a computed tomography (CT) available. HUm and HUsd of the stones were compared with ANOVA. Receiver operative characteristic analysis with area under the curve (AUC), Youden index, and likelihood ratio calculations were performed.

Results:

Data were available for 466 patients. The major components were calcium oxalate monohydrate (COM), uric acid, hydroxyapatite, struvite, brushite, cystine, and CO dihydrate (COD) in 41.4%, 19.3%, 12.4%, 7.5%, 5.8%, 5.4%, and 4.7% of patients, respectively. The HUm of UA and Br was significantly lower and higher than the HUm of any other stone type, respectively. HUm and HUsd were most accurate in predicting uric acid with an AUC of 0.969 and 0.851, respectively. The combined use of HUm and HUsd resulted in increased positive predictive value and higher likelihood ratios for identifying a stone's mineral composition for all stone types but COM.

Conclusions:

To the best of our knowledge, this is the first report of CT data aiding in the prediction of brushite stone composition. Both HUm and HUsd can help predict stone composition and their combined use results in higher likelihood ratios influencing probability.

Introduction

The lifetime prevalence of urolithiasis is gradually increasing and it is currently reported to be approximately 8.8% for the population of the United States.¹ Urinary stone disease as a primary diagnosis was responsible for an annual medical expenditure in the Unites States of $2.1 billion in 2000, was estimated at $3.79 billion in 2007, and is expected to be well over $4 billion by 2030.^2,3 With increasing costs associated with urinary stone disease care, the cost-effectiveness of diagnosis, treatment, and prevention is gaining importance.^4
–6 Choosing the most useful diagnostic imaging modality and preventing unnecessary treatments based on preoperative knowledge are essential to this purpose. As stone composition may help guide surgical and medical treatment as well as preventive strategies,^7

–12 this knowledge is useful early on in the workup and treatment planning of a patient's urinary stone disease. The current imaging modality of choice for diagnosis of urinary stone disease in patients with a BMI <30 is a low-dose noncontrast computed tomography (CT).^4,13 Previous attempts at predicting urinary stone composition with noncontrast CT have shown that calcium containing stones and uric acid stones are easily distinguishable based on mean Hounsfield Units (HUm).^14

–18 However, differentiation of other stone types has not been consistently possible with only a few authors reporting identification of struvite^19
–21 or cystine stones.^21
–23 In addition, there is a lack of large cohort, in vivo studies that use commonly applied low-dose noncontrast scanning protocols and include the most commonly encountered stone compositions without excluding nonpure stones.

We aimed to identify whether or not the added value of standard deviation of HU (HUsd) could be used for more accurate prediction of stone composition.

Materials and Methods

Both centers involved in the study, the University of Western Ontario in London, Ontario, and Lenox Hill Hospital in New York, New York, obtained institutional review board approval. A retrospective chart review was performed to identify patients who had undergone a surgical procedure for urinary tract lithiasis between 2006 and 2013. Patients were eligible for analysis if the mineral composition of a stone fragment retrieved during the procedure was available and if the patient had undergone an abdominopelvic noncontrast CT no more than 6 months before the procedure. CT images were obtained with a GE Discovery HD 750 with a tube voltage of 120 kV and automated mA, in London and a Philips Brilliance with a tube voltage of 140 kV and 400 mA in New York. All images were obtained with a 5 mm slice thickness.

The images were reviewed by a blinded urologist or radiologist in each center with standardized measurement guidelines. The size of the stones was measured in the coronal plane. The size was calculated using the formula of an ellipse with longest diameter x perpendicular diameter x π/4. The stone density was measured by drawing a region of interest (ROI) within the stone limits using the bone window on the axial CT slice depicting the largest surface area. In the case of multiple stones, the density of the largest stone was measured. The mean Hounsfield Units (HUm) and standard deviation of Hounsfield Units (HUsd) were automatically reported by the CT software as measurements within this ROI and were collected for analysis. As the density of stones smaller than 5 mm measured on NCCT images with 5 mm slice thickness is likely to be underestimated,²⁴ stones smaller than 5 mm in any dimension were excluded.

Mineral stone composition was analyzed with Fourier transform infrared spectroscopy by a single lab at each center. Both pure and mixed stones were included in this database. The stones were grouped per major contributing component (>50%) in 1 of 7 groups: calcium oxalate monohydrate (COM), calcium oxalate dihydrate (COD), cystine (Cy), struvite (Str), uric acid (UA), hydroxyapatite (HA) and brushite (Br). Stones that did not fit in any of those groups due to not having a mineral component contributing >50% of the total, were excluded from the data and not separately analyzed. In an effort to report AUCs, accuracies and likelihood ratios as accurate as possible, these stones were not left out of those analyses, as this would artificially improve our outcomes.

The HUm and HUsd were compared for each of the groups with Welch's ANOVA. Post hoc analysis was performed using Games–Howell's test.²⁵ Spearman correlation tests were performed to identify whether or not there is a relationship between the stone size and the stone's Hounsfield unit measurements. Predictive accuracy of the different HU measurements for each of the stone types was analyzed using receiver operative characteristic (ROC) curve analysis with area under the curve (AUC) calculation. AUC values between 0.7–0.8, 0.8–0.9, and >0.9 were considered to correspond to fair, good, and excellent predictive accuracy, respectively.²⁶ Cutoff values were calculated from the ROC curves using the Youden index. Likelihood ratios (LRs) were computed and applied to the pretest probabilities to obtain posttest probabilities. Data were statistically analyzed with SPSS version 22 (IBM Corp. Armonk, NY).

Results

Of the 466 patients with stones included in the study, 193 had stones that were mainly composed of COM, 22 of COD, 90 of UA, 58 of HA, 35 of Str, 27 of Br, and 25 of Cy. Seven patients had a 50/50 distribution of UA and COM. There was no component contributing >50% for nine stones with three contributing chemical substances. The mean and standard deviations of HUm, HUsd, and surface area for each of the stone types are presented in Table 1.

Table 1.

Mean and Standard Deviation of Hounsfield Units and Surface Area for All of the Stone Types in This Analysis

	COM	COD	UA	Ap	Str	Br	Cy	P-value
Hum	1073 ± 242	1043 ± 233	441 ± 86	1035 ± 325	732 ± 200	1370 ± 240	643 ± 65	<0.001
HUsd	115 ± 51	130 ± 50	53 ± 37	123 ± 52	97 ± 44	124 ± 54	59 ± 19	<0.001
Surface area (SA) (mm²)	398 ± 387	322 ± 214	645 ± 534	654 ± 591	1073 ± 862	488 ± 342	620 ± 569	<0.001

HUm = mean Hounsfield Units; HUsd = standard deviation of Hounsfield Units; COD = calcium oxalate dehydrate; COM = calcium oxalate monohydrate; UA = uric acid; Str = struvite; Br = brushite; Cy = cystine.

On post hoc analysis, the HUm of UA (441 ± 86) and Br (1370 ± 240) stones were significantly lower and significantly higher than every other stone type (P < 0.001), respectively. Although the HUm of Str (732 ± 200) and Cy (643 ± 65) stones are significantly different from that of all other stone types (P = < 0.001), they could not be distinguished from each other (P = 0.206). COM, COD, and HA could not be differentiated from each other (Table 2 and Fig. 1).

FIG. 1.

(A, B) Bar charts demonstrating the median (line in middle of bar), interquartile range (bar), 5–95% of data (whiskers), and outliers (dots) for mean Hounsfield Units and standard deviation of Hounsfield Units for all the stone types included in the analysis.

Table 2.

Between Group Comparison of Mean Hounsfield Units with P-Values from Post Hoc Analysis

HUm	COM	COD	UA	HA	Str	Br	Cy
COM	—
COD	0.997	—
UA	<0.001	<0.001	—
HA	0.982	1.000	<0.001	—
Str	<0.001	<0.001	<0.001	<0.001	—
Br	<0.001	<0.001	<0.001	<0.001	<0.001	—
Cy	<0.001	<0.001	<0.001	<0.001	0.206	<0.001	—

The bold values indicate statistically significant p-values.

HA = hydroxy apatite.

The HUsd of UA (53 ± 37) and Cy (59 ± 19) stones are significantly lower than the HUsd of any other stone type (P = < 0.001), but do not differ significantly from each other (Table 3 and Fig. 2). Spearman correlation showed only a moderate positive correlation of HUm with stone size for Str stones and a weak positive correlation for COM stones (Rho: 0.488 with P-value: 0.004 and Rho: 0.306 with P-value < 0.001 respectively).

FIG. 2.

(A–G) Probability of a stone being of a certain mineral composition. The values on the blue line (diamonds) demonstrate the probability of a stone being a certain composition (y-axis) when the HUm is below a certain value on the x-axis (Prob <HUm). The red line (squares) demonstrates that probability when above that certain HUm value (Prob >HUm). The other lines (triangles and Xs) show the probability of a stone being a certain mineral composition when below (<) or above (>) (depending on the stone type) a certain HUm and below (<) or above (>) a certain HUsd. (A) COM, (B) COD, (C) UA, (D) HA, (E) Str, (F) Cy, (G) Br. Color images available online at www.liebertpub.com/end

Table 3.

Between Group Comparison of Standard Deviation of Hounsfield Units with P-Values from Post Hoc Analysis

HUsd	COM	COD	UA	HA	Str	Br	Cy
COM	—
COD	0.786	—
UA	<0.001	<0.001	—
HA	0.916	0.997	<0.001	—
Str	0.362	0.160	<0.001	0.139	—
Br	0.975	0.999	<0.001	1.000	0.385	—
Cy	<0.001	<0.001	<0.939	<0.001	0.001	<0.001	—

The bold values indicate statistically significant p-values.

According to the AUC calculations, the HUm is a fair to excellent predictor for UA, Br, COM, and Cy stones (AUC = 0.975, 0.882, 0.743, 0.724, respectively) and the HUsd is a fair to good predictor for UA and Cy stones (AUC = 0.866 and 0.763). The cutoff values that where calculated for HUm and HUsd for each stone type with their accompanying sensitivity, specificity, calculated positive predictive value (PPV), negative predictive value (NPV), and test accuracy are shown in Tables 4 and 5. When combining these cutoff values, we identified an increased PPV for all stone types except COM (Table 6).

Table 4.

Area Under the Curve of Mean Hounsfield Units with Accompanying P-Values and 95% Confidence Intervals for All the Stone Types, with Calculated Cutoff Values and Their Accompanying Sensitivity, Specificity, Positive Predictive Value, Negative Predictive Value, and Accuracy

Type	AUC HUm	P-value	95% CI	Cutoff	Sensitivity	Specificity	PPV	NPV	Accuracy
COM	0.743	<0.001	0.698–0.788	>763	89.6%	59.3%	62.5	88.4	72.4
COD	0.624	0.049	0.539–0.709	>740	90.9%	38.7%	7.0	98.8	40.7
UA	0.975	<0.001	0.962–0.987	<588	97.8%	92.6%	78.6	99.4	94.2
HA	0.621	0.003	0.549–0.692	>644	93.1%	31.9%	16.7	96.8	39.1
Str	0.649	0.003	0.586–0.712	<1045	97.1%	40.1%	12.1	99.4	45.1
Cy	0.724	<0.001	0.681–0.767	<738	100%	66.6%	15.2	100	69.1
Br	0.882	<0.001	0.820–0.944	>1235	81.5%	81.3%	21.6	98.6	81.1

AUC = area under the curve; NPV = negative predictive value; PPV = positive predictive value.

Table 5.

Area Under the Curve of Standard Deviation of Hounsfield Units with Accompanying P-Values and 95% Confidence Intervals for All the Stone Types, with Calculated Cutoff Values and Their Accompanying Sensitivity, Specificity, Positive Predictive Value, Negative Predictive Value, and Accuracy

Type	AUC HUsd	P-value	95% CI	Cutoff	Sensitivity	Specificity	PPV	NPV	Accuracy
COM	0.649	<0.001	0.600–0.699	>73	80.8	44.7%	53.1	75.9	61.1
COD	0.683	0.004	0.586–0.780	>82.2	95.5%	42.3%	7.9	99.5	45.8
UA	0.866	<0.001	0.818–0.913	<68	81.1%	80.3%	50	94.4	80
HA	0.649	<0.001	0.579–0.718	>94	69%	54.7%	18.4	92.3	56.7
Str	0.506	0.905	0.416–0.596	>58	88.6%	25.1%	8.9	95.6	30.7
Cy	0.763	<0.001	0.706–0.821	<88	96%	56%	11.3	99.6	57.8
Br	0.646	0.011	0.569–0.723	>71	100%	35.1%	9.1	100	39.8

Table 6.

Calculated Cutoff Values for HUm and HUsd for All of the Stone Types, with the Accompanying PPV, NPV, Sensitivity, Specificity, and Accuracy for Their Combined Use

	Cutoff HUm	Cutoff HUsd	Sensitivity	Specificity	PPV	NPV	Accuracy
COM HUm + HUsd	>763	>73	71.0%	65.8%	60.9	75.1	68.0
COD HUm + HUsd	>740	>82	86.4%	54.7%	8.9	98.7	56.2
UA HUm + HUsd	<588	<68	81.1%	97.8%	90.1	95.4	94.4
HA HUm + HUsd	>644	>94	63.8%	62.2%	20.0	92.1	62.4
Str HUm + HUsd	<1045	>58	82.9%	61.7%	15.4	97.7	63.3
Cy HUm + HUsd	<738	<88	96.0%	76.9%	19.7	99.7	78.0
Br HUm + HUsd	>1235	>71	81.5%	87.0%	28.6	98.7	86.7

As PPV and NPV are dependent on prevalence and the current calculated values are based on the prevalence in our cohort, these values may not be applicable to each patient population. We therefore calculated LRs based on incremental cutoff values per 100 HUm, which allowed us to project posttest probabilities for all the stone types based on the cutoffs. Figure 2A–G clearly demonstrates the influence of adding a rounded cutoff value for HUsd on the posttest probabilities for each of the stone types. This effect is most apparent for UA and Cy stones, with stones that have a HUm below 500 and a HUsd below 70 having a probability of 91% of being a UA stone, whereas stones with a HUm below 500 and a HUsd above 70 have only a 42% probability of being a UA stone. Similarly, a stone with HUm below 800 and HUsd below 90 has a 17% probability of being a Cy stone, whereas that stone would have only a 1.5% probability of being a Cy stone if the HUsd was above 90. Adding HUsd as a parameter has the least influence on the probabilities of HA and Br stones.

Discussion

Many in vivo and in vitro studies have been performed to uncover the ability of preoperative imaging to predict the mineral composition of urinary stones.^{14

–23,27

–36} Similar to previous reports, the average HUm of UA was significantly lower than the average HUm of all other stone types.^{14
–16,18,23,36} Mostafavi et al. had likewise reported that Cy and Str could be differentiated from UA, COM, and Br, but not from each other after an in vitro evaluation of 102 pure stones.¹⁷ To the best of our knowledge, this is the first report of Br stones having a significantly higher HUm than any other stone type in an in vivo cohort. Although brushite is a distinct stone type, it has often been grouped with HA^18,21 or COM³⁰ or not included in the database at all,^{15,16,19,20,23,29,33} probably due to low numbers.

There are multiple reports indicating that shockwave lithotripsy (SWL) is less effective for stones with HUm more than 900–1000.^7,37
–39 Although the average HUm of cystine stones in this report is far below that threshold, we also know that cystine stones do not fragment well with SWL.^40,41 Correctly identifying a cystine stone before treatment could therefore help prevent unnecessary SWL.

In the current analysis, adding HUsd as a predictive variable increased the PPV and the probability of correctly identifying a stone's mineral composition for most stone types. Interestingly, Hillman et al. in 1984 described that including the standard deviation as a parameter increased the accuracy of correctly categorizing uric acid, struvite, and calcium oxalate stones in a small in vitro study.¹⁹ In an in vivo analysis of 100 patients, Patel et al., however, found no significant differences in HUsd.³⁶ To date, the influence of this parameter on probability ratio has not been previously reported.

The use of multiple parameters to aid in identifying a certain stone type has been attempted before. Nakada et al. have reported that the stone size may aid in identifying a certain stone type.¹⁴ They developed a formula for HU density (HUd) by dividing the HUm by the stone size in mm. A value below 80 would be indicative of a stone being UA. When applying this formula to our data, we see that not only UA but also Cy and Str stones have a HUd below 80. Interestingly enough, 75% of the Br stones also have a HUd <80. Williams et al. pointed out that this HUd is not linear and that the HUm plateaus with increasing stone size.⁴² This HUd as proposed by Nakada et al. may therefore adjust for lower HUm in small stones, but will be artificially low for larger stones, limiting its utility in a data set including large stones. Adding HA-supersaturation as a parameter did not improve predictive accuracy in distinguishing HA stones from COM/COD stones with dual-energy CT in a study by Liu et al.⁴³ Spettel et al., however, demonstrated that adding the patient's urinary pH with a cutoff of <5.5 to the HUm increased the PPV for UA stones. Similarly, it has been shown that a more alkaline urine predisposes for HA, Str, and Br stones.^44,45 Although these data were not available for analysis, we believe that incorporating these and other patient variables in an algorithm together with HUm and HUsd may help to increase the predictive accuracy of a stone's mineral composition.

There is a possibility that the stone composition reported for each stone is not fully representative of the entire stone's mineral composition as only fragments of the stone are sent for analysis. As the density measurements are only performed on one CT slice, it is equally possible that the HU measurements are not fully representative of the entire stone's CT density and thus the mineral composition. Despite this possible bias, this methodology reflects that clinical practice and significant differences were nonetheless identified between the different stone types. More elaborate measurements at different sections of the stone and analysis of the complete stone sample would be required for more rigorous data analysis.

Grosjean et al. have demonstrated that different CT scanners may procure different HU values for the same stone in an in vitro model.²⁸ Although these in vitro results cannot be extrapolated to in vivo due to differences in scanning methods and surrounding medium,^27,31 the use of two different scanners may be a possible limitation in our multicenter study. As our results show significant differences between the different stone types despite this possible limitation, we argue that this strengthens generalizability of our data. Additionally, it has been previously demonstrated that HU measurements would yield similar results on a low-dose and conventional non-contrast CT.^46,47

Conclusion

The combined use of HUm and HUsd with the proposed cutoff values increases the accuracy of predicting a urinary stone's composition. The likelihood ratios calculated in this study need to be prospectively validated and perhaps refined with a larger data set to empower these observations.

Footnotes

Author Disclosure Statement

No competing financial interests exist.

Abbreviations Used

References

Scales

, Smith

, Hanley

, Saigal

. Prevalence of kidney stones in the United States. Eur Urol, 2012; 62:160–165.

Pearle

, Calhoun

, Curhan

. Urologic diseases in America project: Urolithiasis. J Urol, 2005; 173:848–857.

Antonelli

, Maalouf

, Pearle

, Lotan

. Use of the National Health and Nutrition Examination Survey to calculate the impact of obesity and diabetes on cost and prevalence of urolithiasis in 2030. Eur Urol, 2014; 66:724–729.

Fulgham

, Assimos

, Pearle

, Preminger

. Clinical effectiveness protocols for imaging in the management of ureteral calculous disease: AUA technology assessment. J Urol, 2013; 189:1203–1213.

Hyams

, Matlaga

. Cost-effectiveness treatment strategies for stone disease for the practicing urologist. Urol Clin North Am, 2013; 40:129–133.

Lotan

, Buendia Jiménez

, Lenoir-Wijnkoop

, Daudon

, Molinier

, Tack

, et al. Primary prevention of nephrolithiasis is cost-effective for a national healthcare system. BJU Int, 2012; 110:E1060–E1067.

Perks

, Schuler

, Lee

, Ghiculete

, Chung

D-G

, D'A Honey

, et al. Stone attenuation and skin-to-stone distance on computed tomography predicts for stone fragmentation by shock wave lithotripsy. Urology, 2008; 72:765–769.

Kacker

, Meeks

, Zhao

, Nadler

. Decreased stone-free rates after percutaneous nephrolithotomy for high calcium phosphate composition kidney stones. J Urol, 2008; 180:958–960: discussion 960.

Xue

, Zhang

, Yang

, Chong

. The effect of stone composition on the efficacy of retrograde intrarenal surgery. Kidney stones 1–3 cm in diameter. J Endourol, 2015; 29:537–541.

10.

Skolarikos

, Straub

, Knoll

, Sarica

, Seitz

, Petrik

, et al. Metabolic Evaluation and Recurrence Prevention for Urinary Stone Patients: EAU Guidelines. Eur Urol, 2015; 67:750–763.

11.

Bres-Niewada

, Dybowski

, Radziszewski

. Predicting stone composition before treatment - can it really drive clinical decisions?. Cent Eur J Urol, 2014; 67:392–396.

12.

Pak

, Poindexter

, Adams-Huet

, Pearle

. Predictive value of kidney stone composition in the detection of metabolic abnormalities. Am J Med, 2003; 115:26–32.

13.

Drake

, Jain

, Bryant

, Wilson

, Somani

. Should low-dose computed tomography kidneys, ureter and bladder be the new investigation of choice in suspected renal colic?: A systematic review. Indian J Urol, 2014; 30:137–143.

14.

Nakada

, Hoff

, Attai

, Heisey

, Blankenbaker

, Pozniak

. Determination of stone composition by noncontrast spiral computed tomography in the clinical setting. Urology, 2000; 55:816–819.

15.

Motley

, Dalrymple

, Keesling

, Fischer

, Harmon

. Hounsfield unit density in the determination of urinary stone composition. Urology, 2001; 58:170–173.

16.

Spettel

, Shah

, Sekhar

, Herr

, White

. Using Hounsfield unit measurement and urine parameters to predict uric acid stones. Urology, 2013; 82:22–26.

17.

Mostafavi

, Ernst

, Saltzman

. Accurate determination of chemical composition of urinary calculi by spiral computerized tomography. J Urol, 1998; 159:673–675.

18.

Nakasato

, Morita

, Ogawa

. Evaluation of Hounsfield Units as a predictive factor for the outcome of extracorporeal shock wave lithotripsy and stone composition. Urolithiasis, 2015; 43:69–75.

19.

Hillman

, Drach

, Tracey

, Gaines

. Computed tomographic analysis of renal calculi. AJR Am J Roentgenol, 1984; 142:549–552.

20.

Demirel

, Suma

. The efficacy of non-contrast helical computed tomography in the prediction of urinary stone composition in vivo . J Int Med Res, 2003; 31:1–5.

21.

Deveci

, Coşkun

, Tekin

, Peşkircioglu

, Tarhan

, Ozkardeş

. Spiral computed tomography: Role in determination of chemical compositions of pure and mixed urinary stones—an in vitro study. Urology, 2004; 64:237–240.

22.

Newhouse

, Prien

, Amis

, Dretler

, Pfister

. Computed tomographic analysis of urinary calculi. AJR Am J Roentgenol, 1984; 142:545–548.

23.

Torricelli

FCM

, Marchini

, De

, Yamaçake

KGR

, Mazzucchi

, Monga

. Predicting urinary stone composition based on single-energy noncontrast computed tomography: The challenge of cystine. Urology, 2014; 83:1258–1263.

24.

Stewart

, Johnson

, Ganesh

, Davenport

, Smelser

, Crispen

, et al. Stone size limits the use of hounsfield units for prediction of calcium oxalate stone composition. Urology, 2015; 85:292–295.

25.

Games

, Howell

. Pairwise multiple comparison procedures with unequal N's and/or variances: A Monte Carlo Study. J Educ Behav Stat, 1976; 1:113–125.

26.

Swets

JA.

Measuring the accuracy of diagnostic systems. Science, 1988; 240:1285–1293.

27.

Saw

, McAteer

, Monga

, Chua

, Lingeman

, Williams

. Helical CT of urinary calculi: Effect of stone composition, stone size, and scan collimation. AJR Am J Roentgenol, 2000; 175:329–332.

28.

Grosjean

, Daudon

, Chammas

, Claudon

, Eschwege

, Felblinger

, et al. Pitfalls in urinary stone identification using CT attenuation values: Are we getting the same information on different scanner models?. Eur J Radiol, 2013; 82:1201–1206.

29.

Sheir

, Mansour

, Madbouly

, Elsobky

, Abdel-Khalek

. Determination of the chemical composition of urinary calculi by noncontrast spiral computerized tomography. Urol Res, 2005; 33:99–104.

30.

Bellin

M-F

, Renard-Penna

, Conort

, Bissery

, Meric

J-B

, Daudon

, et al. Helical CT evaluation of the chemical composition of urinary tract calculi with a discriminant analysis of CT-attenuation values and density. Eur Radiol, 2004; 14:2134–2140.

31.

Grosjean

, Sauer

, Guerra

, Kermarrec

, Ponvianne

, Winninger

, et al. Determination of the chemical composition of human renal stones with MDCT: Influence of the surrounding media. In: Manduca

, Hu

, eds. Med. Imaging, International Society for Optics and Photonics, San Diego, CA. 2007, pp. 65112O–65112O–9.

32.

Mitcheson

, Zamenhof

, Bankoff

, Prien

. Determination of the chemical composition of urinary calculi by computerized tomography. J Urol, 1983; 130:814–819.

33.

Al-Ali

, Patzak

, Lutfi

, Pummer

, Augustin

. Impact of urinary stone volume on computed tomography stone attenuations measured in Hounsfield units in a large group of Austrian patients with urolithiasis. Cent Eur J Urol, 2014; 67:289–295.

34.

Marchini

, Remer

, Gebreselassie

, Liu

, Pynadath

, Snyder

, et al. Stone characteristics on noncontrast computed tomography: Establishing definitive patterns to discriminate calcium and uric acid compositions. Urology, 2013; 82:539–546.

35.

Gupta

, Ansari

, Kesarvani

, Kapoor

, Mukhopadhyay

. Role of computed tomography with no contrast medium enhancement in predicting the outcome of extracorporeal shock wave lithotripsy for urinary calculi. BJU Int, 2005; 95:1285–1288.

36.

Patel

, Haleblian

, Zabbo

, Pareek

. Hounsfield units on computed tomography predict calcium stone subtype composition. Urol Int, 2009; 83:175–180.

37.

Wiesenthal

, Ghiculete

, D'A Honey

, Pace

. Evaluating the importance of mean stone density and skin-to-stone distance in predicting successful shock wave lithotripsy of renal and ureteric calculi. Urol Res, 2010; 38:307–313.

38.

Ouzaid

, Al-qahtani

, Dominique

, Hupertan

, Fernandez

, Hermieu

J-F

, et al. A 970 Hounsfield units (HU) threshold of kidney stone density on non-contrast computed tomography (NCCT) improves patients' selection for extracorporeal shockwave lithotripsy (ESWL): Evidence from a prospective study. BJU Int, 2012; 110:E438–E442.

39.

El-Nahas

, El-Assmy

, Mansour

, Sheir

. A prospective multivariate analysis of factors predicting stone disintegration by extracorporeal shock wave lithotripsy: The value of high-resolution noncontrast computed tomography. Eur Urol, 2007; 51:1688–1693: discussion 1693–1694.

40.

Williams

, Saw

, Paterson

, Hatt

, McAteer

, Lingeman

. Variability of renal stone fragility in shock wave lithotripsy. Urology, 2003; 61:1092–1096.

41.

Ringdén

, Tiselius

H-G

. Composition and clinically determined hardness of urinary tract stones. Scand J Urol Nephrol, 2007; 41:316–323.

42.

Williams

, Saw

, Monga

, Chua

, Lingeman

, McAteer

. Correction of helical CT attenuation values with wide beam collimation: In vitro test with urinary calculi. Acad Radiol, 2001; 8:478–483.

43.

Liu

, Qu

, Carter

, Leng

, Ramirez-Giraldo

, Jaramillo

, et al. Differentiating calcium oxalate and hydroxyapatite stones in vivo using dual-energy CT and urine supersaturation and pH values. Acad Radiol, 2013; 20:1521–1525.

44.

Mobley

, Hausinger

. Microbial ureases: Significance, regulation, and molecular characterization. Microbiol Rev, 1989; 53:85–108.

45.

Hesse

, Heimbach

. Causes of phosphate stone formation and the importance of metaphylaxis by urinary acidification: A review. World J Urol, 1999; 17:308–315.

46.

Sohn

, Clayman

, Lee

, Cohen

, Mucksavage

. Low-dose and standard computed tomography scans yield equivalent stone measurements. Urology, 2013; 81;(2):231–234.

47.

Alsyouf

, Smith

, Olgin

, Heldt

, Lightfoot

, Li

, et al. Comparing stone attenuation in low- and conventional-dose noncontrast computed tomography. J Endourol, 2014; 28;(6):704–707.