In this study, Group 2 demonstrated significantly higher manually calculated stone volumes and STONE scores. Correlation analysis showed that both manually and 3D software–calculated stone volumes had significant positive correlations with the longest stone diameter, surface area, GUY’s score, S.T.O.N.E score, operative time, total laser energy, and lasing time. Furthermore, S.T.O.N.E score was positively correlated with operative time and fluoroscopy time, while operative time showed significant positive correlations with laser energy and lasing time. While correlation analyses demonstrate associations between variables, they do not establish predictive relationships, and these findings should be interpreted accordingly.
According to the ROC analysis, the threshold value for predicting the presence of residual stones was determined to be 746.9 mm³ (AUC: 0.656, 95% CI: 0.528–0.783, sensitivity: 61.2%, specificity: 61.3%). Multivariate logistic regression identified manually calculated stone volume as the only factor significantly associated with stone-free rate. The discrepancy between manual and 3D-derived volumes may be attributed to methodological differences, including assumptions of geometric shape in manual calculations and segmentation variability in 3D analysis. Additionally, the absence of reproducibility assessment for 3D segmentation represents a limitation. Interestingly, manual volume showed stronger predictive value, which may be related to its consistency and closer alignment with simplified clinical measurements.
When the literature is reviewed, numerous studies can be found that examine the factors influencing stone-free rates in patients undergoing PCNL. Atalay et al. conducted a retrospective review of 164 patients treated with PCNL and reported that 53% achieved complete stone clearance. Unlike our study, their analysis included measurements of both stone volume and the volume of the renal collecting system. The predictive value of the ratio of these two volumes in terms of stone-free status was investigated. It was found that the stone volume/renal collecting system volume ratio was statistically superior in predicting stone-free status compared to stone surface area and stone volume measurements (14). The stronger predictive value of the stone volume to renal collecting system volume ratio may be explained by several practical points. Unlike conventional measurements based solely on stone size, this ratio indicates the extent to which the collecting system is occupied by the stone burden. This offers a more realistic understanding of the surgical challenge. The advantage becomes even clearer in patients with complex or staghorn stones, where irregular shapes make surface area or volume measurements less reliable. By accounting for individual anatomical differences and showing the true proportion of the system filled by the stone, this ratio provides a more consistent and meaningful indicator of stone-free outcomes. We think that these factors likely contributed to its superior performance in predicting postoperative results.
Another study designed by Tailly et al., the data of 313 patients who underwent PCNL were analysed. In this study, the stone-free rate was found to be 69.6%. In our study, this rate was 61.3%. Similar to our study, this publication also found that stone volume had a significant predictive value for surgical outcomes. However, in that study, only stone volumes measured using 3D software were used. Additionally, in the same study, stone diameter and surface area were also found to be significant predictors of stone-free rates (11).
In a prospective cohort of 142 patients undergoing retrograde intrarenal surgery, Treigny et al. found that 64% had no residual stones postoperatively. For stone analysis, stone diameters and manual stone volume calculations were performed using two different measurement methods (Ackermann’s and Sphere). The stones were also subdivided based on whether their diameter was below or above 20 mm and re-evaluated. While all three measurement methods accurately predicted the absence of residual stones for stones smaller than 20 mm, only volume-based measurements were effective for stones exceeding 20 mm (15).
In the study designed by Canat et al. using stone data from 27 patients who underwent PCNL, stone volumes were calculated using both the ellipsoid formula (also used in our study) manually and 3D software. These stones were then 3D printed, and their in vitro volumes were calculated using the water displacement method. The results of the study showed that volume calculations made using 3D software were more correlated with the volumes of the 3D-printed stones. A statistically significant difference was observed between the manually calculated volumes and those of the 3D-printed stones (8). Based on our study results, manually calculated stone volume was more predictive of stone-free status than volumes obtained using 3D software.
There are also similarly designed studies for other stone treatment modalities besides PCNL. In the study conducted by Bandi et al., which included 94 patients who underwent ESWL, 58 patients (62%) were found to be stone-free after the procedure, while 36 patients (38%) had residual stones. A statistically significant difference was found in stone volume between these two groups. Consistent with our study, stone volume was identified as the strongest predictor of stone-free status (16).
In the study by Ito et al., unlike our study, 314 patients who underwent ureteroscopy were evaluated to determine whether stone diameter or stone volume better predicted stone-free status. Stone volumes were manually calculated, similar to our study. In this study, the predictive value of diameter and volume was assessed by dividing the stones into subgroups based on size. For stones smaller than 20 mm, diameter and volume had a similar level of predictive value. However, for stones larger than 20 mm, stone volume was found to be a more successful predictor, consistent with the findings of our study (17).
From a clinical perspective, the identified threshold of 746.9 mm³ may serve as a practical parameter in preoperative surgical planning. In our cohort, stone volumes exceeding this cutoff were associated with a markedly increased likelihood of requiring more complex intervention. Therefore, for patients with stone volumes above 746.9 mm³, surgeons may consider a more cautious approach, including planning for standard PCNL rather than less invasive alternatives, anticipating longer operative times, or the potential need for staged procedures. Conversely, patients with smaller stone volumes may be more suitable candidates for less invasive strategies. However, this threshold should not be interpreted as an absolute decision-making criterion but rather as a supportive tool to be integrated with other clinical factors such as stone location, anatomy, and surgeon experience. In the present study, the discriminative ability of manual stone volume was found to be modest (AUC: 0.656), with sensitivity and specificity values of approximately 61%. These findings suggest that stone volume alone may not be sufficiently robust as a standalone predictor for clinical decision-making. Rather, its clinical utility may be enhanced when used in combination with established nephrolithometry scoring systems such as S.T.O.N.E. score and Guy’s stone score, which incorporate additional anatomical and procedural factors. Integrating stone volume into such multifactorial models may improve risk stratification, surgical planning, and prediction of procedural outcomes. Therefore, we believe that future studies should focus on validating combined predictive models to better define the additive value of stone volume in clinical practice.
When discussing the limitations of the study, we must acknowledge the inherent constraints of retrospective studies. We believe that the design and execution of prospective, randomized studies would contribute more meaningfully to the literature. In such prospective studies, conducting a power analysis for sample size calculation would also help obtain statistically more robust results. The use of different imaging modalities with varying sensitivity for detecting residual fragments may have introduced detection bias. Another limitation is that the postoperative stone-free rate assessment was conducted solely at the first postoperative month. We believe that future studies should incorporate longer-term outcomes, such as evaluations at the third and sixth postoperative months, to provide a more comprehensive clinical perspective. Furthermore, the wide CI associated with the OR for manual stone volume indicates substantial uncertainty around the effect size, likely due to the limited sample size and potential sparsity in the data. Therefore, while the observed association is noteworthy, it should be interpreted cautiously. Additionally, we believe that not performing routinely non-contrast-enhanced CT for postoperative residual fragment evaluation may have led to the oversight of some small residual stones. However, since current guidelines recommend postoperative CT imaging only for symptomatic patients, it was not routinely performed. In addition, preoperative imaging was primarily based on non-contrast computed tomography rather than contrast-enhanced imaging, which may have limited detailed assessment of the collecting system anatomy in selected cases.
DISCUSSION
In this study, Group 2 demonstrated significantly higher manually calculated stone volumes and STONE scores. Correlation analysis showed that both manually and 3D software–calculated stone volumes had significant positive correlations with the longest stone diameter, surface area, GUY’s score, S.T.O.N.E score, operative time, total laser energy, and lasing time. Furthermore, S.T.O.N.E score was positively correlated with operative time and fluoroscopy time, while operative time showed significant positive correlations with laser energy and lasing time. While correlation analyses demonstrate associations between variables, they do not establish predictive relationships, and these findings should be interpreted accordingly.
According to the ROC analysis, the threshold value for predicting the presence of residual stones was determined to be 746.9 mm³ (AUC: 0.656, 95% CI: 0.528–0.783, sensitivity: 61.2%, specificity: 61.3%). Multivariate logistic regression identified manually calculated stone volume as the only factor significantly associated with stone-free rate. The discrepancy between manual and 3D-derived volumes may be attributed to methodological differences, including assumptions of geometric shape in manual calculations and segmentation variability in 3D analysis. Additionally, the absence of reproducibility assessment for 3D segmentation represents a limitation. Interestingly, manual volume showed stronger predictive value, which may be related to its consistency and closer alignment with simplified clinical measurements.
When the literature is reviewed, numerous studies can be found that examine the factors influencing stone-free rates in patients undergoing PCNL. Atalay et al. conducted a retrospective review of 164 patients treated with PCNL and reported that 53% achieved complete stone clearance. Unlike our study, their analysis included measurements of both stone volume and the volume of the renal collecting system. The predictive value of the ratio of these two volumes in terms of stone-free status was investigated. It was found that the stone volume/renal collecting system volume ratio was statistically superior in predicting stone-free status compared to stone surface area and stone volume measurements (14). The stronger predictive value of the stone volume to renal collecting system volume ratio may be explained by several practical points. Unlike conventional measurements based solely on stone size, this ratio indicates the extent to which the collecting system is occupied by the stone burden. This offers a more realistic understanding of the surgical challenge. The advantage becomes even clearer in patients with complex or staghorn stones, where irregular shapes make surface area or volume measurements less reliable. By accounting for individual anatomical differences and showing the true proportion of the system filled by the stone, this ratio provides a more consistent and meaningful indicator of stone-free outcomes. We think that these factors likely contributed to its superior performance in predicting postoperative results.
Another study designed by Tailly et al., the data of 313 patients who underwent PCNL were analysed. In this study, the stone-free rate was found to be 69.6%. In our study, this rate was 61.3%. Similar to our study, this publication also found that stone volume had a significant predictive value for surgical outcomes. However, in that study, only stone volumes measured using 3D software were used. Additionally, in the same study, stone diameter and surface area were also found to be significant predictors of stone-free rates (11).
In a prospective cohort of 142 patients undergoing retrograde intrarenal surgery, Treigny et al. found that 64% had no residual stones postoperatively. For stone analysis, stone diameters and manual stone volume calculations were performed using two different measurement methods (Ackermann’s and Sphere). The stones were also subdivided based on whether their diameter was below or above 20 mm and re-evaluated. While all three measurement methods accurately predicted the absence of residual stones for stones smaller than 20 mm, only volume-based measurements were effective for stones exceeding 20 mm (15).
In the study designed by Canat et al. using stone data from 27 patients who underwent PCNL, stone volumes were calculated using both the ellipsoid formula (also used in our study) manually and 3D software. These stones were then 3D printed, and their in vitro volumes were calculated using the water displacement method. The results of the study showed that volume calculations made using 3D software were more correlated with the volumes of the 3D-printed stones. A statistically significant difference was observed between the manually calculated volumes and those of the 3D-printed stones (8). Based on our study results, manually calculated stone volume was more predictive of stone-free status than volumes obtained using 3D software.
There are also similarly designed studies for other stone treatment modalities besides PCNL. In the study conducted by Bandi et al., which included 94 patients who underwent ESWL, 58 patients (62%) were found to be stone-free after the procedure, while 36 patients (38%) had residual stones. A statistically significant difference was found in stone volume between these two groups. Consistent with our study, stone volume was identified as the strongest predictor of stone-free status (16).
In the study by Ito et al., unlike our study, 314 patients who underwent ureteroscopy were evaluated to determine whether stone diameter or stone volume better predicted stone-free status. Stone volumes were manually calculated, similar to our study. In this study, the predictive value of diameter and volume was assessed by dividing the stones into subgroups based on size. For stones smaller than 20 mm, diameter and volume had a similar level of predictive value. However, for stones larger than 20 mm, stone volume was found to be a more successful predictor, consistent with the findings of our study (17).
From a clinical perspective, the identified threshold of 746.9 mm³ may serve as a practical parameter in preoperative surgical planning. In our cohort, stone volumes exceeding this cutoff were associated with a markedly increased likelihood of requiring more complex intervention. Therefore, for patients with stone volumes above 746.9 mm³, surgeons may consider a more cautious approach, including planning for standard PCNL rather than less invasive alternatives, anticipating longer operative times, or the potential need for staged procedures. Conversely, patients with smaller stone volumes may be more suitable candidates for less invasive strategies. However, this threshold should not be interpreted as an absolute decision-making criterion but rather as a supportive tool to be integrated with other clinical factors such as stone location, anatomy, and surgeon experience. In the present study, the discriminative ability of manual stone volume was found to be modest (AUC: 0.656), with sensitivity and specificity values of approximately 61%. These findings suggest that stone volume alone may not be sufficiently robust as a standalone predictor for clinical decision-making. Rather, its clinical utility may be enhanced when used in combination with established nephrolithometry scoring systems such as S.T.O.N.E. score and Guy’s stone score, which incorporate additional anatomical and procedural factors. Integrating stone volume into such multifactorial models may improve risk stratification, surgical planning, and prediction of procedural outcomes. Therefore, we believe that future studies should focus on validating combined predictive models to better define the additive value of stone volume in clinical practice.
When discussing the limitations of the study, we must acknowledge the inherent constraints of retrospective studies. We believe that the design and execution of prospective, randomized studies would contribute more meaningfully to the literature. In such prospective studies, conducting a power analysis for sample size calculation would also help obtain statistically more robust results. The use of different imaging modalities with varying sensitivity for detecting residual fragments may have introduced detection bias. Another limitation is that the postoperative stone-free rate assessment was conducted solely at the first postoperative month. We believe that future studies should incorporate longer-term outcomes, such as evaluations at the third and sixth postoperative months, to provide a more comprehensive clinical perspective. Furthermore, the wide CI associated with the OR for manual stone volume indicates substantial uncertainty around the effect size, likely due to the limited sample size and potential sparsity in the data. Therefore, while the observed association is noteworthy, it should be interpreted cautiously. Additionally, we believe that not performing routinely non-contrast-enhanced CT for postoperative residual fragment evaluation may have led to the oversight of some small residual stones. However, since current guidelines recommend postoperative CT imaging only for symptomatic patients, it was not routinely performed. In addition, preoperative imaging was primarily based on non-contrast computed tomography rather than contrast-enhanced imaging, which may have limited detailed assessment of the collecting system anatomy in selected cases.