Radical versus partial nephrectomy for a small renal mass: does saving nephrons save lives?

Keywords::

• kidney cancer
• kidney function
• partial nephrectomy
• radical nephrectomy
• renal cell carcinoma

In the USA, more than 65,000 new cases of kidney cancer are diagnosed each year [1]. At least 90% of these are renal cell carcinomas (RCCs) arising from the renal cortex and most of the remaining cases are urothelial carcinomas arising from the renal pelvis. In addition to RCC, the differential diagnosis of a renal cortical mass includes rare malignant lesions, such as sarcomas, metastases from other primaries, and various benign tumors such as angiomyolipomas and oncocytomas. Benign tumors account for up to 20% of small renal masses. However, because many benign lesions cannot be distinguished from RCC based on imaging alone, most patients presenting with a renal cortical mass are assumed to have RCC until a definitive pathological diagnosis is made.

Management options for suspected RCC without known metastases include surgical resection, tumor ablation or radiographic surveillance [2]. Because RCC, even if it is initially detected as a small renal mass, has the potential to grow and metastasize, surveillance is mostly reserved for patients with limited life expectancy due to advanced age and comorbidities. Thermal ablation may be reasonable for small lesions, but it precludes definitive pathological staging and because it is a relatively new technique, its long-term oncologic outcomes are not well established. This leaves surgical resection as the current gold standard of therapy for a newly diagnosed renal mass.

Historically, most patients with suspected RCC were treated with radical nephrectomy (RN). However, in patients with bilateral tumors or those with significant renal impairment at baseline, partial nephrectomy (PN) was often performed to avoid or delay the need for dialysis. During the 1980s, data from several institutional series of PN suggested that oncologic outcomes of this treatment modality may be comparable with those seen with RN [3–5]. It was then hypothesized that in patients with a small renal mass (T1a and selected T1b tumors) and a normal contralateral kidney, PN may in fact offer the same or nearly the same oncologic control as RN, but with better preservation of kidney function, leading to improved overall survival (OS). To test this hypothesis, the European Organization for Research and Treatment of Cancer (EORTC) designed a randomized trial of PN versus RN in patients with a small (≤5 cm) renal mass and a normal contralateral kidney [6–8]. Accrual for EORTC 30904 started in 1992, and while the results of this trial were awaited, numerous observational studies were performed to examine the association of PN versus RN with the risk of significant renal dysfunction, cancer-specific mortality and OS [9–15].

Evidence from observational studies

Observational studies of PN versus RN were based on either institutional cohorts or the US surveillance epidemiology and end results (SEER) database [9–15]. Some of these studies restricted analysis to tumors ≤4 cm, while others looked at tumors 4–7 cm or imposed no restrictions on tumor size. A few studies looked only at benign lesions [9,11]. In a recent meta-analysis of 20 observational studies without restriction on tumor size or histologic type, OS was better after PN compared with RN, with a pooled hazard ratio of 0.80 (95% CI: 0.74–0.87; p < 0.001), although there was also substantial evidence of statistical heterogeneity, indicating that individual studies were not estimating the same treatment effect [9,10]. Indeed, in 8 of 20 studies, OS was better after RN compared with PN, although none of these differences were statistically significant [9]. In the other 12 studies, OS was better after PN, with a statistically significant difference in 8 studies [9]. After the completion of this meta-analysis, several additional observational studies were published, most reporting better OS with PN [11–14], although one large study found no significant difference in OS between PN and RN [15].

For the end point of cancer-specific mortality, there was likewise substantial disagreement between the studies in reported findings, but in general there was a tendency toward better cancer-specific survival after PN compared with RN, particularly in the SEER studies [9]. For the end point of severe chronic kidney disease (CKD), most observational studies reported substantial risk reduction in association with PN [9]. In a meta-analysis of 10 observational studies, the pooled estimate of the hazard ratio for this end point was 0.39 (95% CI: 0.33–0.47), although the definition of ‘severe’ CKD varied substantially between the studies, ranging from an estimated glomerular filtration rate (eGFR) <60 to the need for dialysis [9].

Evidence from EORTC 30904

From March 1992 to January 2003, a total of 541 patients with a small (≤5 cm) renal mass and a normal contralateral kidney were enrolled and randomized to RN (n = 273) or PN (n = 268). Perioperative complications requiring re-operation, although fairly rare in both arms, were slightly more common after PN (4.4%) compared with RN (2.4%) [6]. The most surprising finding of this trial was better OS after RN compared with PN. With a median follow-up of 9.3 years, 18% of RN and 25% of PN subjects had died, with a hazard ratio of 1.50 (95% CI: 1.03–2.16; p = 0.03) [7]. Kidney cancer mortality occurred in 3.0% of PN and 1.5% of RN subjects (p = 0.23) [7].

In the recently reported analysis of kidney function, with a median time to last eGFR measurement of 6.7 years, percentages of subjects reaching eGFR <60 were RN: 85.7%, PN: 64.7%, difference: 21.0% (95% CI: 13.8–28.3%); eGFR <45, RN: 49.0%, PN: 27.1%, difference: 21.9% (95% CI: 13.8–30.2%); eGFR <30, RN: 10.0%, PN: 6.3%, difference: 3.7% (95% CI: −1.0–8.5%); eGFR <15, RN: 1.5%, PN: 1.6%, difference: −0.1% (95% CI: −2.2–2.1%) [8]. Hence, compared with RN, PN substantially reduced the incidence of at least moderate renal dysfunction stage A (eGFR <60) and B (eGFR <45), although the incidence of advanced kidney disease (eGFR <30) was relatively similar in the two treatment arms and the incidence of kidney failure (eGFR <15) was nearly identical.

Discussion

The failure of PN to improve OS relative to RN in EORTC 30904 may seem surprising, particularly because the incidence of stage III CKD was substantially lower in the PN arm. In the general population, even a modest decrease in eGFR in the range of 30–59 ml/min/1.73 m² is associated with decline in OS [16]. To understand this apparent discrepancy, it is important to recognize that in the general population, low eGFR usually results from systemic illnesses like diabetes or hypertension, which in addition to their effects on the kidney have multiple direct adverse effects on other vital organs. The latter does not seem to occur with RN, which just targets the kidney. In other words, moderate renal dysfunction resulting from surgery may not have the same negative implications for OS as moderate CKD resulting from systemic medical illness. This is strongly suggested by EORTC 30904 and is in agreement with findings reported from large cohorts of renal donors, where no excess risk of kidney failure or all-cause mortality is seen relative to the matched set of the general population even after 30 years of follow-up [17].

Disagreement between observational studies reporting improved OS with PN relative to RN and findings from EORTC 30904 is likely explained by uncontrolled confounding in the observational studies. In fact, cancer-specific survival in these studies also tended to be better after PN, which can only be explained by selection bias. In contrast, randomized treatment assignment in EORTC 30904 assured that for the purpose of causal inference, all systematic error was eliminated. Results from EORTC 30904 definitively refuted the hypothesis that PN confers an OS advantage relative to RN, at least in patients similar to those enrolled in the trial, since not only the point estimate of the hazard ratio but also the entire CI was on the opposite side of the null. Whether PN actually makes survival worse is not as certain. Although the p-value of 0.03 for the end point of OS was below the conventional level of significance (so that the null hypothesis of equal OS was formally rejected), this is not very strong evidence against the null [7]. In fact, if PN decreases OS relative to RN in this patient population, the biological mechanism of this is unclear although some very speculative explanations have been proposed [8].

Given the significant uncertainty about the true impact of PN versus RN on OS and considering the widespread use of PN, we believe that another randomized trial of RN versus PN is needed to confirm or refute the findings of EORTC 30904. Meanwhile, RN should be considered a viable treatment option for patients with a renal mass and a normal functioning contralateral kidney at baseline. It should definitely be considered a reasonable choice when PN appears to be technically difficult or potentially tenuous with respect to oncologic control. Although many patients may develop moderate renal dysfunction after RN, it does not tend to progress to kidney failure and apparently has minimal adverse consequences for OS. Patients should be counseled accordingly to reduce anxiety that may be associated with the diagnosis of CKD.

Financial & competing interests disclosure

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.