Role of computed tomography with no contrast medium enhancement in predicting the outcome of extracorporeal shock wave lithotripsy for urinary calculi


M.S. Ansari, Department of Urology, All India Institute of Medical Sciences, New Delhi, India.



To evaluate the usefulness of urinary calculi attenuation values from non-contrast computed tomography (NCCT) in predicting the outcome of treatment by extracorporeal shock wave lithotripsy (ESWL).


We evaluated 112 patients with solitary renal and upper ureteric calculi of 0.5–2 cm undergoing ESWL. All patients had NCCT at 120 kV and 240 mA on a spiral CT scanner. During each ESWL session 3000 shock waves were given to a maximum of 3.0 kV. A final X-ray of the kidney, ureters and bladder was taken 12 weeks after the last ESWL session. Fragments of ≤ 5 mm were regarded as clinically insignificant residual fragments (CIRF). The calculi retrieved were analysed by X-ray diffraction and the results assessed by comparing the mean density (as measured in Hounsfield units, HU) with the number of ESWL sessions and clearance.


In all, 82 (76%) patients had complete clearance of stones and 26 (24%) had CIRF. There was a linear relationship between the calculus density and number of ESWL sessions required. Of patients with calculi of ≤ 750 HU, 41 (80%) needed three or fewer ESWL sessions and 45 (88%) had complete clearance. Of patients with calculi of > 750 HU, 41 (72%) required three or more ESWL sessions, and 37 (65%) had complete clearance. The best outcome was in patients with calculus diameters of < 1.1 cm and mean densities of ≤ 750 HU; 34 (83%) needed three or fewer ESWL sessions, and the clearance rate was 90%. The worst outcome was in patients with calculus densities of > 750 HU and diameters of > 1.1 cm; 23 (77%) needed three or more ESWL sessions and the clearance rate was only 60%. The calculus density was a stronger predictor of outcome than size alone.


The use of NCCT for determining the attenuation values of urinary calculi before ESWL might help to predict the treatment outcome, and so might help in planning alternative treatment in patients with a likelihood of a poor outcome from ESWL.


clinically insignificant residual fragments


non-contrast CT


Hounsfield units


Since its introduction by Chaussy et al.[1] in 1980, ESWL has become the preferred treatment for renal calculi of < 2 cm in diameter. The outcome of ESWL depends on many factors, including stone size, location, composition, fragility, the shock wave generator and the presence of obstruction or infection [2,3]. After the introduction by Dretler [4] of the concept of fragility, stone composition has emerged as the main factor influencing the efficacy of ESWL. Different techniques have been used to assist in determining the chemical composition of urinary calculi in vivo. Such tests include pH, identifying and characterizing urinary crystals, the presence of urea-splitting organisms, bone densitometry and radiographic studies [5–7]. CT with no enhancement by contrast medium (NCCT) has long been used clinically to evaluate causes of radiolucent filling defects using measurements of substance density in Hounsfield units (HU) to distinguish calculi from tumours or blood clots [8,9]. As it provides greater density discrimination than a conventional plain abdominal film it is now the preferred method to evaluate patients with renal colic [10]; its ability to detect density differences as low as 0.5% has been exploited to determine the composition and fragility of urinary stones [4,11]. The density of the stone varies with composition, and affects the fragility of a calculus, which ultimately governs the clinical outcome in ESWL. Hence it is vital to know the fragility of a calculus before ESWL, to increase the efficacy and reduce the number of hospital visits and thus cost. We evaluated the role of NCCT, using the attenuation value, in determining the fragility and clearance of calculi in patients treated with ESWL.


Between 1 March 2001 and 30 September 2002, we evaluated 112 patients (80 men and 32 women; mean age 33.6 years, range 19–54) undergoing ESWL for urinary calculi. Patients with solitary renal and ureteric calculi of 0.5–2 cm on a plain film and IVU were included in the study. Patients with ureteric and inferior calyceal calculi of > 1 cm, a solitary functioning kidney, congenital anomaly, those requiring a stent or developing steinstrasse during the therapy, or not giving consent, were excluded.

The patients were examined by haematology, biochemical and urine tests. Before ESWL all patients had NCCT (with no oral or intravenous contrast medium) with 3 mm contiguous sections through the renal calculus, using a soft-tissue setting of a window width and level of 300 and 40 HU, respectively, at 120 kV and 240 mA on a spiral CT scanner (Somatom Plus; Siemens, Germany). The longitudinal calculus dimension was calculated using collimation thickness, the reconstruction interval and the number of images in which the calculus could be visualized. The image showing the calculus in the largest width was selected. The maximum diameter and the mean density of the stone were calculated by drawing a region of interest over the calculus. The maximum dimension of the calculus included in the study was either the longitudinal or the transverse diameter, whichever was the largest.

All patients underwent ESWL (Siemens Lithostar Shock Wave System C; Erlangen, Germany) under analgesia and sedation. The fragmentation of the calculus during the therapy was monitored by fluoroscopy. A maximum of 3.0 kV was given to each patient, starting at 0.1 kV and increasing gradually stepwise after every 200 shock waves. During each ESWL session 3000 shock waves were given, and an interval of 14 days maintained between ESWL sessions. A plain film was taken after each ESWL session to document fragmentation and before the next session to ascertain position and clearance. Clearance was documented by a plain film 12 weeks after the last ESWL session, and defined as complete disappearance of the renal calculus; fragments of ≤ 5 mm were defined as clinically insignificant residual fragments (CIRF) and patients with CIRF were subsequently managed conservatively. All calculus fragments retrieved were analysed by X-ray diffraction.

The CT values were analysed with the outcome of ESWL (the number of sessions required and clearance of the calculi), and were further correlated with stone composition. Results were analysed using Student's t-test and chi-square test, multivariate analysis and one-way anova, followed by a multiple-range test.


Of the 112 patients enrolled, four developed steinstrasse after ESWL and were excluded from the analysis. Of the 108 patients evaluated (77 men and 31 women) 57 had a calculus on the right and 51 on the left, with a mean (range) calculus size of 1.32 (0.7–2.0) cm. The calculus was in the pelvis in 52 patients (48%), in the inferior calyx in 22 (20%), in the superior calyx in five (5%), in the middle calyx in two (2%), in the pelvis and either calyx in eight (7%) and in the upper ureter in 19 (18%). Eighty-two patients (76%) had complete calculi clearance and 26 (24%) had CIRF. Fourteen (13%) patients completed therapy in one ESWL session, 25 (23%) in two, 18 (17%) in three, 32 (30%) in four, 13 (12%) in five and three each (3%) in six and seven. The mean calculus density and number of ESWL sessions needed had a linear correlation (Fig. 1), which was maintained at all levels, except that the three patients who needed six sessions had a mean calculus density of 817 HU, which was less than in those who needed five sessions (1039 HU). On a one-way anova the differences in distribution were significant (P < 0.001), but the conclusions require caution, given the few patients needing six and seven sessions.

Figure 1.

Relationship between renal calculus density (mean HU) and number of ESWL sessions.

Of 51 patients with a calculus density of ≤ 750 HU, 41 (80%) needed three or fewer ESWL sessions and 45 (88%) had complete clearance. The stone was pulverised in 24 patients at the first session, in 17 at the second and in 10 at the third. Of 57 patients with a calculus density of > 750 HU, 41 (72%) needed three or more ESWL sessions (P < 0.001 vs patients with a calculus of ≤ 750 HU) and only 37 (65%) had complete clearance (P < 0.01 vs patients with a calculus ≤ 750 HU). Only six patients (12%) with a calculus of ≤ 750 HU had CIRF, vs 20 (35%) with a calculus of > 750 HU (P < 0.01).

Treatment outcome in relation to calculus diameter is shown in Table 1. The calculus density was a stronger predictor of outcome than size alone. This comparative study used univariate analysis; for multivariate analysis we applied stepwise logistic regression taking calculus density (>750 or ≤750 HU) and diameter (≤1.1 or >1.1 cm) as independent variables, and the number of sessions needed (≤3 or >3) as a dependant variable. Only calculus density was a significant factor, with an odds ratio (range) of 10.5 (4.27–25.86), indicating that patients with a calculus of >750 HU had a 10.5 times greater chance of needing three or more ESWL sessions than those with a calculus of ≤750 HU.

Table 1.  Outcome of ESWL, as n (%)
Groups of patients withCalculus densityP
≤750 HU>750 HU
Calculi ≤ 1.1 cm in diameter
≤3 sessions34 (83) 9 (33)<0.001
>3 sessions 7 (17)18 (67)<0.001
Clearance37 (90)19 (70)>0.05
Residual fragments 4 (10) 8 (30)>0.05
Calculi >1.1 cm in diameter
≤3 sessions 7 7 (23)<0.05
>3 sessions 323 (77)<0.05
Clearance 818 (60)>0.05
Residual fragments 212 (40)>0.05

Calculus fragments were collected in only 72 patients; these were analysed by X-ray diffraction crystallography, but we were able to establish a clear correlation between composition and density. A comparison of calculus densities (mean HU) calculated for various types of calculi in the present study and those reported in previous series are given in Table 2[11–13].

Table 2.  Densities (HU) of urinary calculi
Calculus compositionPresent studyPrevious studies, mean (sd) or range
Nrangemean (sd)[11][12][13]
Calcium oxalate monohydrate48507–16391008 (254)1645 (238)1077–13451690
Calcium oxalate dihydrate 5324–1015 748 (281)1417 (234) 865–10391690
Carbonate apatite 2321–462 4441400
Uric acid 2347–542 391 (99) 409 (118) 426 (51) 480
Struvite 3548–869 662 (179) 666 (87) 725 (118)1285 (284)
Calcium mono and dihydrate12540–14781036 (296)1555 (193)


The outcome of ESWL is measured in terms of fragmentation and clearance of the calculus fragments. Fragmentation of a calculus largely depends on its size and composition [3,4], and the ability to predict stone composition would help to increase the efficiency of ESWL. A plain abdominal film is the initial test for calculus disease, and the ability of a plain film to indicate stone fragility was proposed by Dretler and Polykoff [4,6]. Using X-ray patterns to predict the fragility, smooth-edged calculi with a homogeneous structure required more shock waves to be completely fragmented than calculi with round, radially reticulated, spiculated edges and irregular margins [14]. Bon et al.[3] found that smooth, uniform, bulging calculi that appeared denser than bone (12th rib or transverse process) responded poorly to ESWL; the stone-free rate for smooth, radiologically dense calculi and calculi with irregular outlines was 34% and 79%, respectively. In a prospective study [15] the overall accuracy of predicting calculus composition from plain radiographs was reported to be only 39%, which is at present insufficient for clinical use.

NCCT is noninvasive and provides better density discrimination than conventional radiography; it can be used to detect a density difference of 0.5%, whereas plain radiography requires a density difference of ≈ 5%[16]. The same principle was exploited to analyse the composition of urinary calculi in vitro. Segal et al.[17] measured the density of urinary calculi and reported values of 50–120 ‘CT units’, roughly equivalent to 100–240 HU, for three uric acid calculi. In a study by Federle et al.[8] the calculus was identified as a high-density object (370–586 HU) and calcium oxalate and cystine calculi had higher attenuation values than uric acid or xanthine calculi. Newhouse et al.[18] used NCCT to determine CT attenuation values to allow an accurate analysis of stone composition; uric acid and cystine calculi could be identified, but calcium-containing calculi such as struvite, brushite and oxalate could not be distinguished reliably from each other. Similarly Hillman et al.[12] reported that uric acid calculi could be differentiated clearly from struvite and calcium oxalate calculi, but the last two were less easily differentiated.

Thus identifying uric acid calculi became feasible by NCCT, but overlapping CT attenuation values posed problems in accurately determining different calcium calculi. Two modifications were of great help in identifying different calcium calculi [11,13,19]. Mostafavi et al.[11] proposed that the chemical composition of urinary calculi could be determined from the absolute CT values measured at 120 kV and the dual-voltage CT values measured at 80 and 120 kV (HU at 80 kV minus HU at 120 kV); the absolute CT value at 120 kV could identify the chemical composition of uric acid, struvite and calcium oxalate calculi, and the use of the dual-voltage CT value was able to differentiate calcium oxalate from brushite and struvite from cystine. Later, in an in vitro study, Saw et al.[20] reported that the number of shock waves required to fragment the stone correlated with size (volume, weight, diameter) and helical CT attenuation values, and concluded that for calcium calculi, the number of shock waves to comminution was generally less than half the calculus CT attenuation value. This ‘half-attenuation rule’ predicted the number of shock waves needed for complete fragmentation of 95% of calcium calculi.

In an interesting in vivo study, Nakada et al.[21] compared the attenuation and the attenuation size ratio (peak attenuation/calculus size) with the results of calculus analysis; there was a significant difference between uric acid calculi, with a mean (sd) of 344 (152) HU, and calcium oxalate calculi, at 652 (490) HU, and by using an attenuation/size ratio threshold of > 80, the negative predictive value was 99% that a calculus would be predominantly calcium oxalate. In another in vivo study, Motley et al. [19] found no significant difference between the density values of calcium oxalate and calcium phosphate calculi, and so they were analysed collectively as calcium calculi. There was less overlap in the densities of the calculi studied and no non-calcium calculi had a density of > 6 HU/mm2.

To date, few clinical studies have compared the density of calculi with the outcome of ESWL in vivo. Joseph et al.[22], in a study of 30 patients, found that patients with calculi of < 500 HU had complete clearance and required 2500 shock waves (median), while patients with calculi of 500–1000 HU had a clearance rate of 86% and required a median of 3390 shock waves, and patients with calculi of ≥ 1000 HU had a clearance rate of 55%, requiring a median of 7300 shock waves These authors used an electromagnetic lithotripter (Lithostar Multiline, Siemens) and recommended that if the attenuation value of the calculus was > 950 HU and 7500 shock waves had not achieved adequate fragmentation, percutaneous nephrolithotomy should be considered. More recently, Pareek et al.[23] correlated calculus density with clearance in a study of 50 patients who were treated with a second-generation electrohydraulic lithotripter. They concluded that 36% of patients with residual calculi had a mean calculus density of ≥ 900 HU, compared with mean of 500 HU in 74% of patients who had clearance. However, they did not correlate the calculus density with fragmentation.

In the present study, when patients were categorised by calculus density, 80% with calculi of ≤ 750 HU needed three or fewer ESWL sessions and 88% had complete clearance. Conversely, of patients with calculi of > 750 HU, 72% required three or more sessions for complete clearance. The best outcome was in patients with stone diameters of ≤ 1.1 cm and a density of ≤ 750 HU; 35% needed three or fewer sessions and the clearance rate was 90%. The worst outcome was in patients with a stone of > 750 HU and diameter of > 1.1 cm; 23 (77%) of these patients needed three or more sessions and the clearance rate was only 60%. Analysis indicated that the attenuation value (calculus density) had a greater impact on outcome than the calculus size.

This is the largest in vivo study comparing the attenuation values of urinary calculi and stone size with treatment outcome. Patients with a mean stone density of > 750 HU had 10.5 times more chance of requiring three or more sessions than patients with mean stone densities of ≤ 750 HU. Hence we recommend that, for calculi of ≤ 750 HU, irrespective of size (<2 cm), ESWL should be the preferred treatment. Calculi of > 750 HU are associated with a poor outcome when treated with ESWL and an alternative treatment like percutaneous nephrolithotomy and/or ureteroscopy should be considered, depending on the location of the stone. Last, this study also opens the possibility of a better outcome of ESWL for calculi of > 2.0 cm and ≤ 750 HU.

NCCT is noninvasive and its use before ESWL can help in determining the mean stone density, which can predict its fragility and hence treatment outcome. This might help in planning alternative treatments in patients with a probable poor outcome, and to increase the efficiency of ESWL, thus decreasing the cost of treatment.


None declared.