Introduction

Urolithiasis affects a significant proportion of the global population, with prevalence rates ranging between 1 and 13% worldwide [1]. Since 1990, there has been a rise in the overall number of urolithiasis cases, as well as the associated disability-adjusted life years and deaths [2, 3]. This underscores the importance of implementing diverse strategies aimed at preventing and treating urolithiasis.

To support urologists in making treatment decisions, identifying the characteristics of stones, deepening the understanding of lithogenesis, and predicting the risk of recurrence, various stone classification systems have been developed by guidelines committees [4, 5]. The aim of this manuscript is to provide an overview of different stone classification criteria to facilitate better understanding stone formation, guide treatment strategies, and prevent stone recurrence.

Methods

Two authors (V.D.C. and E.X.K.) conducted a review of literature using the Medline, Scopus, and Web of Science databases in July 2023. The search terms “classification” AND (“urolithiasis” OR “kidney stones”) AND (“anatomy” OR “size” OR “imaging” OR “recurrence” OR “etiology” OR “composition” OR “morphoconstitutional analysis”) were used, and the filters “English” and “humans” were applied. Only studies relating to the classification of urolithiasis were considered. Case reports, editorials, and letters were excluded. Additional articles identified through references lists were also included. A narrative synthesis for analysis of the studies was used. The content and structure of the manuscript were agreed upon by consensus between the two authors (V.D.C. and E.X.K.) (Table 1).

Table 1 Classification system
Full size table

Results

Anatomical position

Urinary stones can be situated in diverse anatomical locations within the urinary tract, such as the pyelocaliceal system, ureter, bladder, or urethra (Fig. 1). To further pinpoint the location, the renal pelvis is considered as an important location of its own, as are the upper, middle, and lower calyces. The ureter is segmented into proximal, middle, and distal portions. The proximal ureter ranges from the ureteropelvic junction until the superior crest of the sacroiliac joint. The mid-ureter ranges from the superior to inferior crest of the sacroiliac joint. From there, the distal ureter passes through the vesicoureteric junction into the bladder. [6].Such precision in localization plays a pivotal role in selecting appropriate treatment modalities, ranging from conservative approaches to surgical intervention [4, 5, 7, 8].

Fig. 1
figure 1

Anatomy of the urinary system

Full size image

Size

Stone size can be described using one, two, or three dimensions. The majority of literature reports stone burden based on the largest diameter, with stratification methods such as < 5 mm, 5–10 mm, 10–20 mm, and > 20 mm aiding in outcome comparison and treatment guidance [4, 5, 7,8,9]. This unidimensional measurement is arguably adequate for the description of small particles [10]. A minority of publications report stone size in surface area which can be calculated from both ultrasound (US) and kidney–ureter–bladder (KUB) radiography [9]. Finally, only a few publications report stone burden in terms of volume. This calculation method is on the rise for several reasons. Most patients receive non-contrast computerized tomography (CT). Reporting stone volume is undeniably the most accurate method for reporting stone burden [9]. It also allows the assessment of stone ablation efficiency during laser lithotripsy [11]. With the development of thulium fiber lasers and devices with integrated pressure-measuring and aspiration technology, it also pushes boundaries of retrograde intrarenal surgery for larger stones [12, 13]. To facilitate stone burden measurements, fully automated systems for kidney stone detection and volume quantification on CT have been developed recently [14].

Medical imaging characteristics

US, KUB, and CT stand out as the most frequently employed imaging techniques for the comprehensive evaluation of urolithiasis. Depending on the mineral composition, stones can be categorized based on their appearance on plain X-rays, resulting in classifications such as radiopaque (including calcium oxalate monohydrate (whewellite), calcium oxalate dihydrate (weddellite), and brushite), poorly opaque (comprising carbapatite, magnesium ammonium phosphate (struvite), and cystine), or radiolucent (encompassing uric acid, urate, and miscellaneous types). In both single-energy and dual-energy CT scans, the Hounsfield Unit (HU) values have demonstrated predictive capabilities for determining stone compositions (Fig. 2) [15,16,17,18,19,20,21]. The lowest HU values (approximately 400) have been notably associated with uric acid stones, showing a remarkably high success rate in oral chemolysis treatment [22]. Additionally, investigations have revealed that lower HU values are linked to heightened success rates in shockwave lithotripsy (SWL), while higher HU values have exhibited a strong correlation with SWL failure [23,24,25].

Fig. 2
figure 2

Stone composition in function of Hounsfield Units on CT in the range of 80 kV to 140 kV COM calcium oxalate monohydrate, COD calcium oxalate

Full size image

Artificial intelligence presents as a valuable tool in the field of urolithiasis imaging, a topic commonly known as radiomics. Several studies to date have unveiled the potential of radiomics, suggesting that radiomics has a high potential to help clinicians providing personalized medicine [26].

Risk of recurrence

Accurately estimating the risk of stone recurrence is a significant concern, as it plays a pivotal role in safeguarding patients from potential regrowth, episodes of renal colic, the development of chronic kidney failure, mineral and bone disorders, and the crucial decision of initiating pharmacological treatment. In the guidelines of the European Association of Urology (EAU) and the American Urological Association (AUA), patients are typically categorized into low- and high-risk stone formers. This classification takes into account a range of factors, including the timing of stone formation, stone composition, stone burden, the presence of (genetic) conditions associated with stone formation, medication history, anatomical anomalies, and environmental and occupational influences [27, 28].

To predict stone recurrence, Rule et al. developed a recurrence of kidney stones (ROKS) nomogram. This predictive tool relies on thirteen readily accessible features from a patient's clinical history, which are typically available at the time of the initial stone episode [29]. Performance of the ROKS nomogram has demonstrated a limited to modest ability in foreseeing stone-related events [30, 31]. Based on a survey on risk estimation of stone recurrence, clinicians seemed not be able to distinguish patients with high and low recurrence risk when compared with peers and the ROKS nomogram [32].

In conclusion, there exists a clear need for additional clinical tools within the healthcare workflow to optimize the precision of stratifying individuals as low or high-risk stone formers. Moreover, it is imperative to explore tailored approaches regarding the type and frequency of imaging, specifically aligned with individual patient risk profiles, to curtail the risks associated with costly and excessive radiation exposure [33, 34].

Etiology of formation

Another method of classifying stones involves categorizing them based on their etiology that can be stratified into four groups [27]. Stones associated with non-infectious origins, encompassing compositions like calcium oxalate, calcium phosphate, uric acid, and ammonium urate. Additionally, there are stones attributed to infections, which encompass magnesium ammonium phosphate and highly carbonated apatite, and various urates. Third, genetic defects can give rise to stones composed of cystine, xanthine, and 2,8-dihydroxyadenine. Finally, certain stones are linked to the use of specific medications.

Composition

Stones are generally stratified upon composition to distinguish the variations in structure among various stone constitutions and their associated lithogenic processes. In routine clinical practice, diverse techniques are used to analyze stone composition: microscopy (binocular, polarized light and scanning electron microscopy) and constituent analysis (Fourier transform infrared spectroscopy with or without attenuated total reflectance, Raman spectroscopy, X-ray diffraction, and thermal analysis) [35,36,37,38,39,40,41].

Microscopic examination is a valuable tool for understanding various aspects of crystalline components, such as their nature, shape, internal structure, and location. It also provides insights into crystalline conversions and the relationships between different crystalline species. This technique has been extensively used to explore the inner structure of stones and reveal the morphology of small crystals [42, 43]. It is worth mentioning that following laser lithotripsy, there has been an observed alteration in the crystalline organization of stone dust when compared to stone fragments [44]. Following Ho:YAG laser lithotripsy, these changes included the conversion of calcium oxalate dihydrate to calcium oxalate monohydrate, shifts in the carbapatite spectra toward an amorphous phase, alterations of magnesium ammonium phosphate toward different amorphous and crystalline phases, and the presence of hydroxyapatite on brushite fragments. On the other hand, after thulium fiber laser lithotripsy, most stone dust samples showed changes in their crystalline organization, with the exception of calcium oxalate monohydrate and carbapatite [45].

Regarding constituent analysis, infrared spectroscopy is currently the reference method. It has a high sensitivity to detect scarce but clinically important components. It also identifies mineral and organic components, drugs, and foreign particles [42]. Until now, there has been no approved standard for conducting stone analysis. Several groups have noticed a strong variability in both qualitative and quantitative stone analysis results among different laboratories and a lack of standardization of used nomenclature [46, 47]. This can be explained by the fact that less than 10% of stones are pure or include only one chemical compound [48]. Since the constituent analyses do not identify stone structure and crystal phases, another technique called morphoconstitutional analysis is of interest to provide information of the lithogenic process of the entire stone.

Morphoconstitutional analysis

The initial endeavor to establish a morphoconstitutional classification of stones was undertaken by Ord and Shattock in 1895[49]. Afterward, multiple groups investigated relationships between stone characteristic and their composition [50,51,52]. Later, several authors attempted to correlate stone composition with etiopathogenesis [53,54,55,56]. In 1993, Daudon et al. provided a detailed understanding of the overall characteristics of urinary stones in their morphoconstitutional analysis [42]. It involves a combined approach to examine both the stone's morphologic appearance and its chemical composition examined with infrared spectroscopy [57]. On the one hand, their refined morphological and structural examination involves observing the stone's size, shape, surface features, and internal organization under optical microscopy including location of different components and crystalline conversion. Compositional analysis using infrared spectroscopy, on the other hand, helps identifying stone composition and to determine the ratio of different components within a single stone [42].

By employing morphoconstitutional analysis, Daudon et al. proposed a classification based on surface and section morphology with its corresponding pathophysiological factors aiding to a better understanding of stone-related diseases and their etiopathogenesis (Table 2, adapted from [58]). They classified stones into seven types and twenty-two subtypes. Each subtype exhibits correlations with specific pathophysiologic conditions, which provides valuable insights into the underlying causes of stone formation and may guide clinicians in the selection of treatment modalities [42, 57, 58]. A series of clinical applications derived from morphoconstitutional analysis are detailed hereunder.

Table 2 Morphoconstitutional analysis of urinary calculi (adapted from [42 and 58])
Full size table

Fast identification of crystalline phase principles

Various chemical components can crystallize in multiple forms. For instance, calcium oxalate can manifest as calcium oxalate monohydrate, calcium oxalate dihydrate, or calcium oxalate trihydrate. As early as four decades ago, research revealed that calcium oxalate monohydrate is influenced by oxalate concentration, while calcium oxalate dihydrate is more dependent on calcium concentration. These diverse crystalline forms have been linked to specific biochemical conditions causing hyperoxaluria and hypercalciuria, respectively [59,60,61].

Detection of crystalline conversion process

Calcium oxalate dihydrate stones are primarily linked to hypercalciuria, while calcium oxalate monohydrate stones are predominantly formed due to primary crystallization in hyperoxaluric conditions [57, 62]. Nevertheless, the latter can also arise from the conversion (dehydration) from calcium oxalate dihydrate toward calcium oxalate monohydrate stones [43]. It is crucial to detect this conversion process from calcium oxalate dihydrate to calcium oxalate monohydrate during morphoconstitutional analysis when establishing connections between stone morphology, crystalline composition, and potential etiopathogenic factors.

Directing etiopathogenesis

The refined morphologic examination proposed by Daudon et al. provides etiopathogenic orientations and guides metabolic evaluation [42, 57, 63]. Just to choose one are the carbapatite (calcium phosphate) stones, which account for 10% to 15% of all stones. Based on 1093 patients who had an available etiological diagnosis and stones containing at least 70% of calcium phosphate without magnesium ammonium phosphate, Dessombz et al. evaluated the stone morphology [64]. They found that 13% of carbapatite stones had a peculiar morphology termed type IVa2, characterized by a smooth aspect and a glazed brown yellow appearance with tiny cracks. Type IVa2 morphology was associated with inherited distal renal tubular acidosis (dRTA) in 53% of cases, medullary sponge kidney in 35%, Sjögren syndrome in 9%, carbonic anhydrase inhibitors intake in 1.5% and urinary tract infection in 1.5%. In contrast, these diseases were associated with IVa1 stones in 0.3%, 17%, 0.7%, 5%, and 40% of cases, respectively. Remarkably, 96% of patients with inherited distal renal tubular acidosis had stone morphology type IVa2, representing a sensitivity and specificity of 96% and 94%, respectively.

Identification of lithogenic process: papillary umbilication and Randall plaques

In 1936, the American urologist Alexander Randall described the presence of calcium phosphate deposits at the tip of the renal papilla. He also observed calcium oxalate renal stones connected to the plaques and put forward the hypothesis that these papillary calcifications could potentially be the initial trigger for stone formation in these patients [65, 66]. Recently, it was found that these plaques have a 83% prevalence in kidney stone formers [67]. The origin of Randall's plaque is thought to be the basement membranes of the thin limbs of the loop of Henle. This plaque formation may be linked to increased calcium reabsorption from the proximal tubule, leading to higher concentrations of calcium in the descending vasa recta and subsequently in the interstitium around thin limbs [68]. These plaques may be influenced by hypercalciuria, and the combined use of vitamin D and calcium could accelerate their formation [69]. Stones that grow over Randall plaques can dislodge and extract a piece of the papilla, creating a visible papillary pit or umbilication. This phenomenon results in a Randall plaque forming on the lower surface of the stone and a papillary pit corresponding to the area of attachment. In 2019, Almeras et al. introduced a classification system for renal papillary abnormalities visualized during flexible ureteroscopy with the aim of standardizing their description. This classification, denoted as Sx nPx Drx/i/px, comprises various elements: S designates the observed type of papillary stones, n denotes the count of abnormal papillae within the same kidney along with their specific type (P), and D encompasses deposits, which includes quantification of Randall's plaques (rx), as well as the presence of active intrapapillary papular deposits (i) and plugs (p) [67]. This observation has the potential to contribute to our understanding of nephrolithogenesis mechanisms and the risk of stone recurrence [70].

Assessing the stone surface can also be used to track other stones. Similar to beach pebbles, certain stones types (such as Id, IIc, IIIa, Vb) display distinct characteristics with smooth, rounded, and naturally polished surfaces. This appearance is a result of constant mechanical interaction, as the stones continuously toss and roll against each other, leading to abrasion and surface smoothing. This particular morphology indicates the presence of multiple stones and conditions of stasis, such as calyceal diverticulum, ureteropelvic junction stenosis, and benign prostatic obstruction [42, 58].

Assessment of crystal formation speed

Crystal formation speed and recurrence rates are important clinical parameters that may hint to the underlying metabolic disease. A typical example is primary hyperoxaluria, which is an uncommon genetic disorder that results in an excessive production of oxalate, mainly affecting the kidneys and resulting in fast growth of calcium oxalate monohydrate stones. To prevent complications and preserve kidney function, early diagnosis and intervention are essential [71]. A study conducted by Daudon et al. involved the stone analysis of 74 patients with primary hyperoxaluria. They discovered a distinct stone morphology within this group. The stones exhibited a whitish or pale-yellow surface and a loose, unorganized inner structure (type Ic), which was notably different from the dark-brown surface and well-organized, radiating inner structure (type Ia) commonly seen in most calcium oxalate monohydrate stones. Scanning electron microscopy confirmed that the crystalline structure found in the primary hyperoxaluria stones (type Ic) was different from the typical structure observed in the common whewellite stone (type Ia). The authors hypothesized that this dissimilarity in morphology is indicative of the rapid and permanent crystal formation induced by genetic hyperoxaluria. They concluded that this peculiar stone morphology is pathognomonic for primary hyperoxaluria and should prompt an early and comprehensive laboratory evaluation [72, 73]. If morphological analysis was omitted, patients with primary hyperoxaluria may suffer from delayed time point of diagnosis and impaired kidney function.

Assessment of recurrence rates

Guidelines of EAU and AUA recommend close follow-up and metabolic testing in recurrent stone formers [27, 28]. This assessment considers particular stone compositions, including brushite, uric acid, urate, and cystine. Recently, Daudon et al. assessed the risk of recurrency based on the morphoconstitutional analysis of 38,274 stones [74]. The age of first stone occurrence varied, with dihydroxyadenine occurring at the youngest age (15.7 years) and anhydrous uric acid at the oldest (62.5 years). Calcium oxalate stones fell in between (40.7 years for calcium oxalate dihydrate and 48.4 years for calcium oxalate monohydrate). When considering composition alone, calcium oxalate monohydrate stones had a lower recurrence rate, with only 38.0% coming from patients with previous episodes. However, when examining different calcium oxalate monohydrate morphologies, type Ic had a significantly higher recurrence rate (82.4%). Similarly, among stones composed of apatite, type IVa2 had a higher recurrence rate compared to other apatite morphologies (78.8% vs. 39–42%). They concluded that stone composition alone is insufficient to predict recurrence risk, and considering stone morphology can aid in identifying highly recurrent stones.

Guidance of treatment modalities

Initially, it was thought that resistance to SWL was related to crystalline compositions [75]. It was only after later research that stone fragility was found to be associated with stone mineral content measured by dual photon absorptiometry [76, 77]. Williams et al. found that the variability in stone fragility to shock waves was large within groups of the same mineral composition. They hypothesized that the variation in stone structure could underlie the variation in stone fragility within the same stone composition [78]. This was further analyzed by Daudon et al. on several morphologies of calcium oxalate monohydrate stones [42]. To gain a deeper understanding of the crystalline structure, a single-crystal neutron study was conducted using a four-circle automated diffractometer. Additionally, scanning electron microscopy was employed to investigate the mesoscopic scale, while powder neutron diffraction was used to explore the nanometric scale. All types of whewellite stones showed a similar structure at the nanometric scale. However, significant differences were observed at the mesoscopic scale. The study revealed a close relationship between stone morphology and the organization of the crystals at the mesoscopic level, which also had an impact on the effectiveness of SWL. From an analysis involving 1270 SWL procedures performed on calcium oxalate monohydrate stones (including type Ia, Ib, and Id), the failure rates were found to be 80% for type Ib stones; whereas for type Ia and Id stones, the failure rates were less than 10% [79]. This research provided valuable insights into the morphological and structural characteristics of different types of whewellite stones, shedding light on the association between their formation and response to SWL. These findings were in line with prior studies showing higher success rates for smooth surfaced stones compared to rough surfaced stones [80, 81]. In conclusion, the resistance to fragmentation by shock waves appears to be more related to the organization of the stones, particularly for poorly organized structures (e.g., type Ib and Va), rather than their stone composition.

Discussion

Stones can be classified according to various criteria such as anatomical position, size, medical imaging features, risk of recurrence, etiology, composition, and morphoconstitutional analysis (Table 1). Categorizing stones by anatomical position, size, and medical imaging characteristics is a straightforward process that can arguably directly impact treatment recommendations. Given that CT imaging is typically administered prior to treatment for most patients, determining the precise anatomical location, stone dimensions, and HU can be easily interpreted and should be recommended (expert opinion). In the process of offering treatment modalities such as SWL, RIRS, or PNCL, these criteria must be considered in conjunction with patient preferences. Radiomics, the utilization of artificial intelligence capabilities within the domain of medical imaging, has the potential to enhance this personalized patient approach [26].

In contrast, the classification of stones according to risk of recurrence and etiology is more complex since it depends on multiple variables, including stone composition and morphology. Importantly, the composition of most kidney stones is typically mixed [82,83,84]. This can arise from the presence of concurrent lithogenic factors on one hand, or from subsequent conditions triggering new lithogenic processes, such as dietary hyperoxaluria, dietary hypercalciuria, gout, diabetes mellitus, chronic diarrhea, abuse of laxatives, primary hyperparathyroidism, intake of carbonic anhydrase inhibitors, or urinary tract infections. As stones are infrequently extracted in one piece since the advent of minimally invasive techniques, it becomes crucial to identify the structural characteristics of stone surfaces and sections (including judgement of stone core localization, e.g., Randall plaque) during surgery [42, 57, 58, 63]. This is particularly important since stone composition can alter during laser lithotripsy [44, 45, 85]. Consequently, the stone dust generated might not adequately represent the crystalline organization of the stones before laser lithotripsy. To address this, Estrade et al. evaluated the feasibility of intra-operative stone composition recognition by a single expert urologist using Daudon et al. morphoconstitutional classification [42, 86]. They discovered an 82% agreement between endoscopic and microscopic characterizations of urinary stones. In contrast, Henderickx et al. also assessed the diagnostic accuracy and intra-observer agreement of endoscopic stone recognition compared to formal stone analysis by fifteen urologists. They concluded that diagnostic accuracy was limited, and intra-observer agreement fell below an acceptable threshold [87]. One possible solution lies in utilizing automated computer-assisted in situ recognition of morphological characteristics of both pure and mixed urinary stones, as demonstrated in a preliminary study with promising discrimination results [88]. Incorporating specialized artificial intelligence into ureteroscopes may represent a significant advancement toward perioperative morphoconstitutional analysis [89,90,91].

Currently, the guidelines of the EAU and the AUA primarily propose treatment algorithms based on the stone's anatomical position and size as a unidimensional measurement. However, considering all three dimensions would offer a more accurate assessment of stone burden and could assist urologists in selecting treatment modalities based on their individual stone ablation efficiency. When implementing this approach, the incorporation of other aforementioned classifications, including medical imaging characteristics and patients' medical history, could further aid in the appropriate selection of treatments and the assessment of success rates in everyday practice [92]. Moreover, the inclusion of morphoconstitutional analysis could significantly enhance the management of urolithiasis patients. This analysis allows for the rapid identification of crystalline phase principles, the detection of crystalline conversion processes, the determination of etiopathogenesis, the recognition of lithogenic processes, the assessment of crystal formation speed, and the evaluation of recurrence rates, and the guidance for selecting appropriate treatment modalities. Therefore, as no single classification system can cover all aspects comprehensively, the integration of all classification approaches is essential for tailoring a patient-specific treatment approach.

Conclusion

Classification of urinary stones involves several criteria, each offering unique insights into stone characteristics and formation. Anatomical position, stone size, and medical imaging characteristics provide essential information for treatment decisions. Classification based on risk of recurrence and etiology is more complex, demanding a comprehensive understanding of stone composition and morphology. The use of advanced techniques like morphoconstitutional analysis, which is rapid and low cost by combining morphological appearance and chemical composition, has proven invaluable in identifying stone-related diseases and guiding treatment strategies. This holistic approach has enabled the identification of distinct crystalline phases, determination of lithogenic processes, and even tracking stone history. Becoming familiar with and combining all these classification systems will assist urologists in refining personalized clinical management approaches. Furthermore, as technology continues to evolve, the integration of automated stone recognition and artificial intelligence holds promising potential for enhancing stone recognition accuracy and for integrating all the described stone classifications, further aiding urologists in adopting a tailored approach.