Achieving trifecta and pentafecta is the major goal of PN regardless of the surgical approach. Therefore, an effective and validated tool to evaluate tumor complexity and surgical difficulty is essential. However, the R.E.N.A.L. and PADUA systems are not without limitations[10, 14]. The SPARE system, a refined version of PADUA, includes tumor size, exophytic rate, sinus involvement and rim location (Fig. 1). Compared to R.E.N.A.L. and PADUA, the SPARE system had similar predictive ability in pentafecta achievement (Fig. 2). In other words, the fewer constituents of the SPARE system did not affect its efficacy while making it easier to calculate the score. Moreover, the interobserver concordance of the SPARE system was good in overall score and in most of the individual components (Table 6). As a result, the SPARE system appears to be a favorable choice when evaluating tumor complexity and predicting post-PN outcomes during clinical practice and patient counseling.
Most peri-operative outcomes in our study were similar to RECORd1 project, a 4-year prospective observational multicenter study. Major complication rate was 3.5%, positive surgical margin rate was 5.5%, and median ischemia time was 16 minutes in RECORd1 project[15, 16]. The longer median ischemia time (24 minutes) in our study may be due to larger tumor size and low volume center (less than 50 PN performed per year). Renal functional outcomes such as ACE at 3rd day and 30th day were similar between our study and RECORd1 project.
Current study revealed that surgical approach were correlated to complication rates (ρ= -0.23, p=0.001), ischemia time (ρ= -0.33, p<0.001) but not with positive surgical margin (ρ= -0.03, p=0.76), PCE (ρ= -0.06, p=0.36) nor with achievement of pentafecta (ρ= 0.08, p=0.23) (data not shown in tables). RECORd1 project mentioned that open surgical approach was a signiﬁcant predictive factor of complications. In contrast, Serni et al. showed that surgical approach was neither the predictor of trifecta outcome in patients with highly complex renal tumor underwent simple enucleation. The effect of open surgical approach on trifecta/ pentafecta outcomes varied between studies may be caused by different surgical technique and different complexity of renal tumor. Further studies are required to confirmed this hypothesis.
In the current study, SPARE nephrometry was correlated with peri-operative outcomes including ischemia time, operative time, and complication rate. RECORd1 project mentioned that modified PADUA is not an independent predictive factor of postoperative complication. In contrast to RECORd1 project, most patients in our cohort underwent standard PN by minimal invasive approach (76.8%). Since utilization rate of open partial nephrectomy constantly decreased in last decades. Therefore, SPARE would be a more suitable nephrometry in the era of minimal invasive surgery.
Although there was a trend toward greater functional loss in the higher risk group, Ficarra et al. found that the SPARE system was not associated with functional outcomes. In contrast, the SPARE system was correlated with PCE and pentafecta in our study. This may be due to the different approaches of PN between the two studies. PN was conducted using standard resection methods in our institute, whereas 25% of the patients in their cohort underwent PN by enucleation. Since resected renal volume plays an important role in functional loss, the predictive ability of the SPARE system in functional outcomes may be influenced by the volume of resected non-neoplastic renal parenchyma. In addition, tumor contact surface area has a greater ability to predict post-operative renal function than R.E.N.A.L. and PADUA[11, 19]. SPARE includes components such as radius (R) and exophytic rate (E), which is similar to tumor contact surface area. The other two components of sinus involvement and rim location are related to the vascular territory of the kidneys which affect renal function deterioration. As a result, the SPARE system may be correlated to functional outcomes to some extent. However, further well-designed studies are needed to confirm these hypotheses.
In our study both R.E.N.A.L. and PADUA had good predictive ability for pentafecta achievement. R.E.N.A.L. has been confirmed to be an independent predictive factor of pentafecta achievement with a negative association. Serni et al. showed that PADUA score was significantly associated with achievement of trifecta and with negative margin, but not with warm ischemia time. In contrast, Ubrig et al. and Harke et al. reported conflicting results about the predictive ability of PADUA for trifecta achievement[22, 23]. The difference regarding the predictive ability of PADUA in pentafecta achievement between studies may be explained by the following reasons. First, there were inconsistencies between studies in controlling for confounding factors such as comorbidities, and patient factors affect post-operative complication rates and functional change. Differences in the methods of multivariate analysis between studies may have resulted in conflicting results. In our study, we included possible factors including age, Charlson Comorbidity Index, BMI, and pre-operative renal function in order to reduce selection bias. Second, unimportant and non-concordant factors in PADUA and a lack of central image review may have led to the difference in results between studies. Third, different surgical approaches such as open/ laparoscopy/ robot or simple enucleation/ enucleoresection varied between studies may affect the pentafecta achievement.
Our study showed good interobserver concordance with the SPARE system, with a kappa value of 0.82. Hew et al. reported the limited reproducibility of the PADUA score, and they reported a Fleiss’ generalized kappa in their study cohort of 0.37 to 0.80 for the various components of the PADUA. Spaliviero et al. directly compared interobserver concordance among R.E.N.A.L., PADUA, and C-index, and found that agreement using the C-index method was higher than with PADUA or R.E.N.A.L.. However, limitations existed when scoring the constituents including location and involvement of the collecting system. Therefore, Ficarra et al. refined PADUA into the SPARE system which successfully improved interobserver agreement according to our results. In our cohort, the interobserver concordance of renal sinus involvement was lower and exophytic rate was higher compared with previous studies. This may be because exophytic rate is a semi-quantitative parameter while renal sinus involvement is a qualitative parameter.
To the best of our knowledge, the current study is the first to externally validate the SPARE system. We further confirmed that SPARE is not only a predictive factor in overall complication rate, but also in pentafecta achievement. Besides complication rate, we also found similar predictive abilities of pentafecta achievement between the SPARE and R.E.N.A.L./PADUA systems in ROC analysis. Another strength of the current study is that we provided evidence of the reproducibility of the SPARE system between urologists and radiologist. This result suggests that the SPARE system can be applied across different specialties. However, there are also limitations to this study. First, this is a single center retrospective study design with various confounding factors. However, we tried our best to reduce selection bias by including possible confounding factors which have previously been reported. Second, we lacked unified imaging protocols for CT and MRI because we are a tertiary referral center. Most constituents of the SPARE system are quantitative or semi-quantitative, so there may not have been significant inconsistencies in the scoring. Third, only a small proportion of the patients (6.3%) were classified as being at high risk, which may have limited the findings. Fourth, the PN technique used in the current study was standard resection, so the applicability of SPARE for PN with enucleation is still unclear, and further studies are needed to confirm the efficacy of the SPARE system in high-risk renal tumors and PN with enucleation. Finally, we did not evaluate renal function using radio-isotope scans, which has been proven to be a more precise tool than serum Cre or eGFR, because the aim of this study was to assess pentafecta as defined by a change in renal function as assessed by eGFR. This may not have limited the interpretation of the results.