Is Computed Tomography Mandatory for the Detection of Residual Stone Fragments After Percutaneous Nephrolithotomy?

Clinical significance and fate of residual fragments (RF)

The increasing popularity of PCNL as the preferred treatment for patients with stones in the upper urinary tract has introduced the concept of RFs after PCNL. It is only logical that different definitions of treatment success are required, because we have moved from the era of open surgery for stones to less invasive treatment options such as PCNL and SWL. Since the implementation of PCNL for upper tract stones, intervention outcomes have been reported by both stone-free rates and success rates. And while the stone-free rate is more or less self-explanatory, the success rate includes patients who are completely stone free as well as those with so-called clinically insignificant residual fragments (CIRFs). The lack of consensus regarding the definition of CIRFs and the variety of imaging modalities used for assessing post-PCNL stone-free status (KUB radiography, nephrotomography, US, CT) have made the comparison of the outcome of endourologic stone removal procedures somewhat difficult.

The aim of PCNL is complete stone clearance; however, residual stone fragments can be left behind for several reasons, such as intraoperative migration of a stone fragment into an inaccessible calix, a complex pelvicaliceal system to be treated, early termination of the case because of complications, or because of difficulty visualizing the stone under fluoroscopy. Other reasons for failure to provide stone-free status include stone composition, poor visualization secondary to bleeding, and the subjective impression of prolonged operative time. Moreover, the clearance of stone fragments after PCNL can be neither complete nor immediate. The available studies on the rate of residual fragments post-PCNL have presented conflicting results, with rates ranging from 8% to 80% depending on the definition of “residual lithiasis” and the imaging modality used for stone detection ^{11 –14} (Table 1).

Although there is no agreement on the maximum size, following either SWL or PCNL, RFs are defined as stone fragments smaller than 4 mm or 5 mm in diameter associated with sterile urine, in an otherwise asymptomatic patient with no evidence of upper tract obstruction. ^{13,15 –17} Small residual stones, however, can sometimes become symptomatic. The clinical impact of RFs, therefore, centers around the risks of (a) recurrent stone episodes from remnant stone growth and (b) recurrent urinary tract infection (UTI), especially in cases of infection stones or struvite stones, where it is of outmost importance to remove all RFs regardless of size to prevent recurrent infection and further stone growth and (c) dislocation of fragments and subsequent obstruction. ^{18 –20}

The risk of recurrent stone formation is higher for patients with RFs of infection stones than for stones of different composition. The recurrence rate for patients with RFs after treatment of infection stones was estimated to be 78% at 3 months after SWL, corresponding to a stone-free rate of 20%. ²¹

In another study, 25% of patients treated with SWL for infection stones were found to have formed new stones 2 years after treatment, with those containing a high content of calcium phosphate stones being at higher risk for recurrent stone formation. ²²

The presence of asymptomatic residual stone fragments smaller than 4 to 5 mm in diameter, or so-called CIRFs, after SWL and PCNL were regarded as treatment success. Recent studies, however, have suggested that these CIRFs may after all be clinically significant, and thus CIRF is a misnomer. RFs also can act as nidi for formation of larger stones and hence become symptomatic. RFs could pose problems when they become acutely dislodged and obstructive causing pain or even infection, or when they act as sites of bacterial adherence and persistent UTIs, and chronic bacteriuria. Studies have demonstrated that by time, as the number and size of RFs increases, the risk for symptomatic stone episodes and complications necessitating intervention also increases. ²³

A plausible explanation is that fine debris, undetectable by KUB radiography, persists after PCNL and tends to settle in the most dependent calices, acting as nidi for formation of stones. It is for these reasons that small residual stones, although not considered of immediate clinical relevance, are likely to become bothersome for the patient in the long term. There is evidence that retained RFs during the first 6 months after intervention are the cause of an increased rate of complications. ²¹

One of the earliest studies by Streem and associates ¹⁸ showed that 43% of patients with RFs smaller than 4 mm post-SWL reported a symptomatic stone episode, many of whom needed intervention at a mean of 26 months postoperatively. Likewise, Khaitan and coworkers ¹⁷ reviewed 75 stone patients with 4 mm or smaller RFs and reported that only 18 patients passed fragments spontaneously while 44 (59%) eventually needed some type of intervention including repeated SWL, PCNL, or ureteroscopic lithotripsy.

More recently, Raman and colleagues ¹⁴ followed 728 patients who underwent PCNL and did not undergo second look flexible ureteroscopy only to find that of 43% of patients with residual fragments of a median diameter of 2 mm experienced a stone-related event during surveillance. A stone-related event was defined as an increase in the diameter of a RF, a visit to the emergency depasrtment, hospitalization, or an auxilliary procedure to relieve pain or obstruction. Multivariate analysis showed that: (a) Residual fragments larger than 2 mm and (b) fragments located in the renal pelvis or ureter were independent predictors of a stone event. The authors reported that in 61% of cases of a stone-related event, a secondary surgical procedure was needed. ¹⁴

In a large study from India of 2469 patients who underwent PCNL, RFs were detected in 7.57% of cases. ¹³ In this study, approximately 45% of patients with RFs spontaneously passed their stones at a mean follow-up of 2 years, although the majority of RFs (65.47%) that passed spontaneously did so in the first 3 months after intervention. The authors also noticed that RFs <25 mm² in the renal pelvis were more likely to pass spontaneously, while none of the stones larger than 100 mm² passed spontaneously. Osman and associates ²³ evaluated 173 patients with RFs of 4 mm or smaller after SWL and noticed that in 21% of patients, some type of intervention was needed for those RFs.

A more recent study by Raman and colleagues ¹⁴ on the fate of RFs after PCNL reported a 8% rate of RFs (median diameter of largest RF 2 mm) detected by unenhanced helical CT (UHCT). Of the 42 patients with RFs, 18 (43%) patients experienced a stone-related event within 32 months. Of the 18 patients, 12 presented to the emergency department because of symptoms, and CT imaging confirmed residual stone growth or ureteral obstruction. Of the 18 patients who experienced a stone episode, 11 (61%) needed a secondary surgical procedure. The impact of RF size on secondary surgical procedures was also analyzed. The vast majority (92%) of the 25 patients with RFs of 2 mm or smaller did not need any intervention. On univariate analysis, location of the RFs in the renal pelvis or ureter, size of RFs larger than 2 mm and cumulative RF size larger than 2 mm were considered predictors of a stone-related event. ¹⁴

More recent data from the same study confirmed that 40% of patients with residual fragments of 4 mm or less will experience a stone episode, while 57% of those will need surgical intervention. ²⁴

A recent study assessed the natural history of CIRFs after PCNL. ²⁵ There were 430 patients who underwent PCNL for renal stones and had plain radiography and/or CT imaging within 48 hours of surgery. Based on imaging studies, 74.5% were considered stone free after the initial operation while 22% still had RFs <4 mm at 3 months follow-up. Thirty-eight patients with CIRFs were monitored with serial CT scans annually or when the patients reported a symptomatic stone episode. During a mean follow-up of 28 months, symptoms attributed to the presence of a stone (hematuria or renal colic) were reported by 26.3% of patients while 7.9% of patients spontaneously passed the fragments. ²⁵

The results of these studies have challenged the belief that leaving RFs less than 5 mm in diameter after PCNL is safe, because they were considered clinically insignificant. While this certainly holds true for fragments less than 2 mm, the decision to leave or chase fragments between 2 and 4 mm should be individualized.

Imaging of residual fragments after PCNL

The purpose of imaging after a surgical intervention for stones is to confirm stone-free status or identify the location and size of residual stone fragments and to exclude obstruction of the urinary tract. Additional advantages of CT imaging after urologic interventions include the ability to detect complications such as perirenal hematoma, urinoma, or adjacent organ injury. ²⁶ It is clear that complete stone removal after PCNL is crucial for preventing recurrence and regrowth of stones and further need for ancillary interventions, making postoperative imaging for residual fragments necessary.

Detection of RFs after PCNL can be a challenging task, however. Plain KUB radiography and linear tomography have been used in earlier studies to estimate the outcome of PCNL in relation to the presence of RFs. While the majority of urinary calculi are radiopaque, sometimes they might be difficult to see on plain abdominal radiographs because of their size, location, and also to the presence of stents and tubes and bowel loops. It is true that the superimposition of bowel gas or feces, the presence of extraurinary calcifications, nephrostomy tubes or stents make plain abdominal radiographs difficult to interpret resulting in uncertainty about the presence or absence of residual stone fragments. ^10,27,28

Although it has been suggested that tomography alone or plain abdominal radiography plus tomography is superior to plain KUB radiography alone in evaluating stone-free status, according to Denstedt and coworkers, ²⁹ in patients with large renal calculi managed by PCNL and SWL, plain radiography overestimated their stone-free status by 35% and 17%, respectively, compared with flexible nephroscopy. A recent study also revealed that the mean size of stones missed by plain abdominal radiography was 7.4 mm with 45.5% of stones being >4 mm in diameter. ¹¹

US is able to directly visualize residual fragments in the upper collecting system down to a diameter of 2 mm or provide indirect signs of their present (dilation of the collecting system). ¹⁷ Because of its noninvasive nature and the lack of radiation exposure, US has been suggested for the routine evaluation of asymptomatic patients after SWL. ^9,30

Routine follow-up with only US, however, for the detection of residual stones after PCNL is usually not advised, because it is considered unreliable in detecting residual stones, especially in the presence of a nephrostomy tube; it can be difficult for the operator to differentiate properly between the acoustic shadow of the tube and shadows representing residual stone fragments. ¹⁰

Intravenous urography (IVU) is currently of limited use as an imaging modality for detection of residual stones after PCNL. Nonenhanced, low dose CT provides more information than IVU at a comparable cost and a relatively lower radiation exposure for the patient. ³¹ A prospective study on the relative efficacy of abdominal radiography plus renal US compared with excretory urography for the evaluation of asymptomatic patients 1 month after SWL revealed that the combination performed equally well and even better than IVU in identifying residual stone fragments. ³²

On the other hand, some urologists rely on antegrade pyelography (or nephrostography) via a nephrostomy tube in evaluating treatment outcome and deciding whether to remove the nephrostomy tube after PCNL. The accuracy of antegrade pyelography for the detection of stones has been recently compared with unenhanced multidetector CT (UMDCT) in 49 patients having nephrostomy tubes placed for upper urinary tract stones. ³³ In 64.5% of patients, these stones were diagnosed only by UMDCT; in other words, the sensitivity of antegrade pyelography was limited to 35.5%. Antegrade pyelography missed renal and ureteral stones of significant size (mean size 5.1×6.2 mm, and 6×5.3 mm, respectively). This study showed the clear advantage of UMDCT in the evaluation of residual stones after endoscopic management compared with antegrade pyelography. The main disadvantage of UMDCT remains the higher radiation delivered to the patient compared with IVU and nephrostomography.

Flexible nephroscopy under local anesthesia and intravenous sedation had been considered the “gold standard” for assessment of residual stones after PCNL. ²⁹ Its routine use, however, was challenged by Pearle and colleagues ¹² in view of the high sensitivity of CT in detecting residual stones after PCNL. Pearle and coworkers prospectively compared KUB radiography, noncontrast CT, and flexible nephroscopy in the detection of residual stones in a study of 36 patients with 41 renal units undergoing PCNL. With the assumption that flexible nephroscopy was the gold standard, CT demonstrated a sensitivity and specificity of 100% and 62%, compared with 46% and 82%, respectively, for KUB radiography. The authors concluded that the selective, and not routine, use of second-look flexible nephroscopy based on positive CT findings would have spared 20% of the patients an unnecessary procedure resulting in substantial cost savings. The sensitivity of CT was even better in cases of faint, lucent, and small stones compared with the sensitivity of the other imaging modalities.

Recent data derived from the same group confirmed that besides being perhaps unnecessary, second-look flexible nephroscopy is not cost-effective compared with expectant management for residual fragment smaller than 4 mm. ²⁴ On the other hand, Waldmann and associates ³⁴ in a retrospective study of 121 patients who underwent CT after PCNL (including 59% stone-free patients and 16% patients with fragments of 1 to 3 mm), reported that routine nephroscopy could not have been omitted in three of four cases. Arguments against the routine use of flexible nephroscopy include the increased fluoroscopic exposure, its frequent inability to reach all renal calices, as well as the risks of complications and resultant prolonged hospital stay and increased cost.

After reviewing the relevant literature, there is little to no doubt that currently noncontrast helical CT scan with 1 to 2 mm axial slices is the imaging modality of choice for evaluation of stone-free status, detection, and localization of residual lithiasis after PCNL. ^{10,11,13,14,25,35}

CT can detect, with sensitivity and specificity exceeding 90%, all types of stones, including otherwise radiolucent uric acid concrements with the exception of indinavir stones, which are radiolucent on plain KUB and CT. ³⁶

Recently, the development of helical technology with the elimination of respiratory artifact by rapid image acquisition and the availability of image multiplanar reconstruction has enhanced the sensitivity of CT for detecting urinary tract calculi. Moreover, given the high sensitivity and availability of noncontrast enhanced CT, there is little doubt that flexible nephroscopy should no longer be considered the gold standard for the detection of residual calculi after PCNL. ³⁷

Post-PCNL imaging with MDCT is helpful and particularly beneficial when complete stone clearance is the goal—for instance, in cases of infectious stones, or stones that are faint or lucent at conventional radiography. CT is also helpful in identifying the exact location of residual fragments in relation to the PCNL tract, facilitating retrieval of residual fragments. The findings of post-PCNL CT can help guide treatment decisions—for instance, the need for placement of a stent or planning of a repeated PCNL using the same or a different tract. ^26,34

Another significant advantage of the panoramic view provided by CT lies in its ability to detect extraurinary complications related to PCNL, such as splenic or hepatic injuries. ^38,39

The superiority of unenhanced MDCT in assessing both the location (intrarenal vs extrarenal) and the size of RFs was clearly shown in a prospective study comparing antegrade pyelography, plain KUB radiography, and unenhanced CT after PCNL. ¹¹ This study showed that estimated stone-free rates were 73.6% for pyelography, 62.3% for plain KUB radiography, and only 20.8% for CT. In other words, there is a three-fold reduction in completely stone-free patients if CT was used to evaluate residual stone status at 1 month. It is interesting to note that of the patients in whom residual stones were detected by CT but missed by KUB radiography, almost half of them (45.5%) had stones of significant size (>4 mm in diameter on CT), with a mean estimated size of 7.4 mm.

A different perspective was introduced in a recent study in which noncontrast MDCT was compared with KUB and linear tomography with regard to their sensitivity in the detection of residual stones after PCNL. ¹⁰ The study showed that regarding radiopaque stones, the sensitivity of CT for the overall detection was 100%, 62.9% for plain radiography, 74.3% for linear tomography, and 48.6% for US (P<0.05). The sensitivity for detecting significant residual stones, however, was 100% for CT, 95.2% for linear tomography, and 85.7% for plain radiography. The differences in sensitivity for opaque RFs greater than 5 mm were not considered as statistically significant. This led the authors to introduce the concept of the CT CIRFs including one or two gravels less than 5 mm, as measured by CT. The authors' conclusion challenged the suggested common practice of routine post-PCNL CT for the evaluation of residual lithiasis. Although the authors admitted that currently CT is the most sensitive tool for detecting residual stones after PCNL, they argued that “CT should not be routinely performed in patients with opaque stones because it yields no statistically valuable increase in the diagnosis of significant residual stones compared with that of plain x-ray and linear tomography.”

Currently, there are no available guideline recommendations for the follow-up of residual fragments after PCNL. For cases of residual fragments after SWL, the 2008 European Association of Urology (EAU) guidelines were advocating follow-up with good quality KUB radiography and reserving CT for cases of uric acid stones (EAU guidelines 2008). In the updated 2012 EAU guidelines, however, it is suggested that patients with residual fragments are to be followed up in regular intervals to monitor for recurrent disease without any further recommendations or suggestions on the optimal imaging modality.

At the end of the day, however, it is important that criteria be implemented as to which patients may be most reasonably imaged with sensitive as well as such costly imaging techniques as MDCT. For example, patients in whom a stone-free outcome is mandatory—for instance those with infection calculi and increased risk of recurrent stone formation—should routinely undergo post-PCNL CT.

Timing of imaging after PCNL

Given that unenhanced CT is the imaging study of choice for the detection and evaluation of RFs, another issue that arises is the optimal timing for such imaging. The appropriate time to perform post-PCNL UHCT is a subject of debate. Although in many centers an UHCT is performed routinely at postoperative day 1, the end of the first month after surgery is considered optimal, taking into account that an imaging study performed earlier than that might produce false-positive results from stone dust resulting in poor image resolution, and may also detect RFs that will ultimately pass spontaneously during the immediate postoperative course. ^11,40,41

Another advantage of delayed imaging over more immediate imaging is the presence of tubes or stents immediate postoperatively, which may obscure or hide RFs resulting in a high rate of false-negative scans. Under these circumstances, the use of bone window during UHCT may be necessary to differentiate between an actual residual stone (<1600 Hounsfield units [HU]) and stents or tubes (>1600 HU). ⁴²

CT scan protocols and radiation exposure concerns

Patient exposure to ionizing radiation is a significant concern related to the routine use of CT imaging in stone disease, especially because, at present, there is no strong evidence for routine CT imaging after PCNL. This concern is particularly worrisome for younger patients who are bound to undergo repeated CT examinations because of recurrent stones and are therefore likely to be at a higher risk for radiation-induced secondary malignancies. ^43,44 To understand the magnitude of the problem, one should take into account that the amount of radiation delivered during UMDCT (effective dose: 12.7–13.1 mSv for men and 17.5–18.0 mSv for women—Table 2) is 7 to 10 times higher when compared with antegrade pyelography (depending on the number of images taken during antegrade pyelography). ^45,46

With the use of new-generation multidetector CT and low-dose imaging protocols, however, the radiation dose delivered to the patient may be significantly reduced while maintaining high efficacy for stone detection. ^47,48 The mean effective dose delivered with UMDCT at 30 mAs (1.2 mSv in men and 1.9 mSv in women for ∼30 cm scan region) is comparable to that of a three-film excretory urography examination (1.5 mSv). ^49,50 Sustained efficacy in detecting stones with low-dose MDCT, however, is a concern that has been evaluated. Spielmann and associates ⁵¹ showed that renal calculi as small as 2 mm (smaller than the CIRFs after PCNL) could be depicted at CT exposures obtained at 60 mAs. ⁵¹

Spiral scanners are able to maintain a good balance between the radiation dose delivered to the patient and spatial resolution (sensitivity of 97%, specificity of 96% of low-dose CT for detecting renal stones). ⁵² This is accomplished using a slice thickness of 3 mm and pitch of 1.6, at 30–60 mA and 120 kV with a reconstruction thickness of 1.5 mm based on the patient's body habitus, with automatic tube current modulation to limit radiation exposure. As shown by Rimondini and coworkers, ⁵³ 3 mm is the most appropriate slice thickness for imaging patients with suspected stones in the upper urinary tract. Tack and colleagues ⁴⁹ acquired 2.5 mm thick images and reconstructed 3-mm sections at 2-mm increments, 120 kV, and 30 effective mA, because they considered this the ideal slice thickness for MDCT. Compared with other single detector CT (SDCT) investigations, the authors suggested that the use of a 2.5-mm slice thickness combined with multiplanar reconstructions provides higher accuracy for the detection of renal lithiasis. The radiation dose delivered is the 25% of that reported by Sourtzis and coworkers ⁵⁴ using a SDCT technique with 5-mm collimation, 160 mA, and a pitch of 1.5 and similar to that reported by Hamm and associates, ⁵⁵ using a SDCT technique with 5-mm collimation, 70 mAs, and a pitch of 2 (97% accuracy rate).

In an effort to further minimize radiation exposure, Katz and colleagues ^56,57 used 5-mm–thick images on their multidetector scanners, when imaging patients with suspected renal stones. They support that although thicker images worsen the quality of the multiplanar reconstructions, those are not so important for the detection of suspected ureteral or renal stones.

Finally, the issue of imaging obese patients with MDCT has also been addressed in the previous studies. According to Hamm and associates, ⁵⁵ who performed low-dose SDCT using an equivalent dose level, patients with a body mass index exceeding 30 kg/m² should be excluded. On the other hand, Poletti and coworkers ⁵² suggested an intermediate dose acquisition at 60 mAs for these patients. Performing these examinations with low-dose protocols has, in fact, been established as common practice and is facilitated by the usually high density of stones, so that even noisy images can be diagnostic.