Predictive ability of renal cortex enhancement in dynamic computed tomography for residual renal function after nephroureterectomy: Comparison with 99m Tc‐diethylenetriaminopentacetic acid renography and validation study

Development and Validation of a Multivariable Nomogram Predictive of Post-Nephroureterectomy Renal Function

BACKGROUND AND OBJECTIVE

The timing of perioperative nephrotoxic chemotherapy for upper tract urothelial carcinoma (UTUC) remains controversial and strongly depends on predicted platinum eligibility after radical nephroureterectomy (RNU). The study objective was to develop and validate a multivariable nomogram to predict estimated glomerular filtration rate (eGFR) following RNU.

METHODS

This was a multi-institutional retrospective study of patients with UTUC treated with RNU from 2000 to 2020 at seven high-volume referral centers. Use of adjuvant chemotherapy was risk-stratified. Patients were retrospectively randomly allocated 2:1 to discovery and validation cohorts. Discovery data were used to identify independent factors associated with GFR at 1-3 mo after RNU on linear regression, and backward selection was applied for model construction. Accuracy was defined as the percentage of predicted eGFR results within 30% of the corresponding observed eGFR.

KEY FINDINGS AND LIMITATIONS

We included 1100 patients, of whom 733 were in the discovery and 367 were in the validation cohort. Multivariable predictors of postoperative eGFR decline included advanced age (odds ratio [OR] -0.18, 95% confidence interval [CI] -0.28 to -0.08), diabetes (OR -2.38, 95% CI -4.64 to -0.11), and hypertension (OR -2.24, 95% CI -4.16 to -0.32). Factors associated with favorable postoperative eGFR included larger tumor size (OR 10.57, 95% CI 7.4-13.74 for tumors >5 cm vs ≤2 cm) and preoperative eGFR (OR 0.44, 95% CI 0.39-0.49). A composite nomogram predicted postoperative eGFR with good accuracy in both the discovery (80.5%) and validation (78.6%) cohorts. Limitations include exclusion of patients who received neoadjuvant chemotherapy.

CONCLUSIONS

A nomogram that incorporates ubiquitous preoperative clinical variables can predict post-RNU eGFR and was validated with an independent cohort.

PATIENT SUMMARY

We developed a tool that uses patient data to predict eligibility for chemotherapy after surgery to remove the kidney and ureter in patients with cancer in the upper urinary tract.

Changes in kidney function according to ischemia type during partial nephrectomy for T1a kidney cancer

To compare the postoperative estimated-glomerular-filtration-rate (eGFR) and parenchymal changes between cold ischemia and zero/selective ischemia for a T1a mass. We analyzed 104 patients who underwent open partial nephrectomy with cold ischemia (53) or zero/selective ischemia (51) for T1a between 2008 and 2018 to determine postoperative renal function changes and associated factors. Postoperative renal function was expressed as (postoperative-eGFR - preoperative-eGFR)/preoperative-eGFR × 100%. Parenchymal enhancement and thicknesses of the ipsilateral kidney as tissue changes were measured on postoperative CT to identify the correlation with the renal function change. Patients with 10% or 25% decrease in eGFR were significantly more in the cold ischemia group (p = 0.032, p = 0.006). On multivariable analysis, preoperative eGFR, ischemic type, and percent change of parenchymal thickness were identified to be significantly associated with postoperative 12 months renal function (B = - 0.367, p = 0.020; B = 6.788, p = 0.042; B = 0.797, p = 0.029). Change in parenchymal thickness was negatively correlated with changes in postoperative renal function (r = - 0.277, p = 0.012). Changes in eGFR were associated with a decrease in parenchymal thickness and the type of ischemic technique. Zero/selective ischemia during partial nephrectomy may have an advantage in preserving postoperative renal function compared to cold ischemia.

Split renal function in patients with renal masses: utility of parenchymal volume analysis vs nuclear renal scans

OBJECTIVES

To evaluate the utility of parenchymal volume analysis (PVA) for estimation of split renal function (SRF) in patients with renal masses. SRF is important for deciding about partial vs radical nephrectomy (PN/RN) and assessing risk for developing severe chronic kidney disease after surgery. For renal donors PVA is routinely used to estimate SRF, but the utility of PVA for the more complex renal mass population remains undefined.

PATIENTS AND METHODS

All patients (n = 374) with renal tumours and a normal contralateral kidney managed with PN (2010-2018), with preoperative/postoperative nuclear renal scans (NRS) and cross-sectional imaging were analysed. Parenchymal volumes were measured by free-hand scripting or software analysis. Concordance between ipsilateral estimated glomerular filtration rate (eGFR) values based on SRF from NRS vs PVA were evaluated by Pearson correlation and Bland-Altman plots. Parallel analysis of all 155 patients managed with RN at our centre (2006-2016) with preoperative NRS and imaging was also performed.

RESULTS

For PN, the median age and tumour size were 62 years and 3.4 cm, respectively. The median preoperative ipsilateral parenchymal volume and eGFR were 181 cm³ and 36.9 mL/min/1.73 m² , respectively. Parenchymal volumes estimated by free-hand scripting vs software analyses correlated strongly (r = 0.98, P < 0.001). Preoperative ipsilateral eGFR based on SRF from PVA vs NRS also correlated strongly (r = 0.94, P < 0.001). Ipsilateral eGFR saved after PN correlated strongly with parenchymal volume preserved (all r >0.60); however, the correlation was much stronger when ipsilateral eGFRs were based on SRF from PVA rather than NRS (z-statistic = 3.15, P = 0.002). For RN patients, preoperative eGFR in the contralateral kidney based on SRF from PVA vs NRS also correlated strongly (r = 0.87, P < 0.001).

CONCLUSION

PVA has utility for estimation of SRF in patients with renal masses, even though this population is older and more comorbid than renal donors and the tumour can complicate the analysis. PVA can be obtained by software analysis from preoperative cross-sectional imaging and thus readily incorporated into routine clinical practice.