Development and validation of a multicenter Cox regression model to predict all-cause mortality in patients with renal masses suspicious for renal cancer

The incidence of renal cell carcinoma (RCC) continues to rise, in large part based on the incidental detection of renal masses rather than symptomatic presentation with advanced RCC [1]. There have been significant advances in local and systemic therapies for RCC, but despite these, the mortality for advanced disease remains high. For those with apparently localized disease, however, cancer-specific survival is quite good, and much attention has been paid to the impact of various treatments on renal function and overall health [2,3]. Comparisons between radical nephrectomy (RN) and kidney-sparing interventions (KSI) have revealed functional benefits to KSI. Partial nephrectomy (PN) is the reference standard for localized RCC that is amenable to such an approach in guidelines from the AUA, NCCN, and EAU [4], [5], [6]. Nonsurgical KSI, including thermal ablation and stereotactic ablative radiotherapy, are considered alternatives in select patients [4,7,8]. Several studies, though, have questioned the magnitude of the benefit of treatment vs. surveillance [4,[9], [10], [11]].

Like active surveillance for prostate cancer, active surveillance of RCC has been gaining acceptance worldwide [4,5,8]. First, not all enhancing renal masses are malignant tumors [12]. Second, most small renal masses (SRM) pursue an indolent course, with metastases rarely observed (<1%–2%) in patients with SRM managed with surveillance [4,6]. Third, many patients with suspected RCC also suffer from comorbid conditions that pose a significant impact on their survival, making competing causes more important than cancer-related causes of mortality.

Several prior groups have investigated the role of comorbidities in all-cause mortality and have developed survival models accounting for competing causes of death (noncancer vs. cancer-specific mortality) [13], [14], [15]. Prior studies have relied upon comorbidity indices, such as the Charlson comorbidity index (CCI) and its subsequent iterations [16], [17], [18]. These prognostic tools are hindered though by the extensive amount of information required to estimate mortality risk; thereby, limiting their utility in clinical practice. In our prior work, we found that the simplified cardiovascular index (CVI) provides similar estimates of mortality to the CCI and its derivatives while requiring only seven elements of patient information [19]. In this study, we gathered pretreatment data and survival information from patients managed at 2 institutions to develop and validate a model to predict all-cause mortality in patients with localized renal masses suspicious for RCC.

留言 (0)

沒有登入
gif