Worldwide, over 200,000 new cases of kidney cancer are diagnosed and approximately 100,000 deaths occur from this disease each year. Renal cell carcinoma (RCC) constitutes up to 85% of renal malignancies in adults. Despite the established role of radical or partial nephrectomy as a standard of treatment, a fair number of patients with localized tumors, ranging from 20% to 40%, will experience disease relapse [1]. In patients with recurrent RCC, the clinical course can vary, and survival can be stratified by an objective parameter called the memorial sloan-kettering cancer center risk score, which includes time to recurrence, lactate dehydrogenase, hemoglobin, corrected calcium, and performance status. However, limited information is available on clinical characteristics, prognostic factors, and outcomes in patients with late-recurring RCC [2,3]. In this study, we evaluated patients with disease recurrence after radical nephrectomy with respect to clinicopathological characteristics and focused on determining the predictive factors affecting different cancer-free intervals.

Disease recurrence in patients with localized RCC after curatively intended radical nephrectomy can occur at any time. However, late recurrence after radical nephrectomy is not common. The definition of 'late recurrence' in RCC is not clearly established. The reason we determined 5 years as a cutoff value was because surveillance patterns change at 5 years after curative treatment for RCC [4]. Also, some have suggested that surveillance after 5 years is no longer necessary for cost-effectiveness in low-risk patients [3,4]. Nevertheless, about 10% to 20% of patients with disease recurrence develop late recurrence more than 5 years after nephrectomy [2]. The final objective of our study was to determine the risk factors predictive of late recurrence, at the point of radical nephrectomy, which could thus be incorporated into the postoperative surveillance guideline. In our studies, of a total of 104 patients with follow-up for more than 5 years after radical nephrectomy for localized RCC, 9 patients (9.1%) developed late recurrence. To date, several studies have been conducted to determine the differential characteristics of late recurrence of localized RCC after radical nephrectomy. However, on the basis of the current medical literature, no consensus has been reached about the parameters predicting late recurrence of localized RCC because of the small patient numbers [2-16]. According to the study by Adamy et al. [3], which was conducted with 44 patients with late recurrence (beyond 5 years after nephrectomy), patients with late recurrence tended to have fewer initial symptoms, smaller tumor size, and less aggressive disease (pT1) compared with patients with early recurrence. Adamy et al. [3] also suggested that patients with late recurrence tend to be in an MSKCC favorable risk group. Park et al. [5] evaluated 41 patients with late recurrence (beyond 5 years after nephrectomy) and suggested that old age and high high-sensitivity C-reactive protein levels at the time of operation were independent predictive factors for late recurrence. Brookman-May et al. [2] studied a total of 310 patients with cancer recurrence more than 5 years after radical nephrectomy and compared the characteristics of these patients with recurrence-free patients. They proved that lymphovascular invasion, Fuhrman grade 3-4, and pT stage>pT1 were significantly associated with late recurrence. Ha et al. [6] evaluated 14 patients with disease recurrence more than 5 years after radical or partial nephrectomy among 423 patients with pathologically confirmed stage T1 clear cell RCC and showed that symptoms at diagnosis and pT stage were independent predictive factors for late recurrence. In the present study, unlike in previous studies, we tried to determine the factors that affected time to recurrence after radical nephrectomy and to discover the differential characteristics of patients in the early and late recurrence groups, respectively. We found that tumor necrosis was associated with late disease recurrence and initial clinical symptoms were associated with early recurrence. Large tumor size, advanced pathologic stage, and advanced Fuhrman nuclear grade were associated with both early and late recurrence. We also evaluated the correlation of tumor size and tumor necrosis in all the patients involved in this study. According to the logistic regression analysis, tumor size was shown to be a risk factor of tumor necrosis (OR, 1.261; 95% CI, 1.086 to 1.463; p=0.002).

Tumor necrosis was initially recognized in the 1970s as a predictor of aggressive RCC behavior [17]. Minervini et al. [18] confirmed that histological tumor necrosis is a statistically significant prognostic factor in patients with nonmetastatic clear cell RCC. Kim et al. [7] suggested that the survival rate of patients with tumor necrosis was significantly lower than that of patients without tumor necrosis. According to the previously published medical literature, the definitions for tumor necrosis are diverse. However, most investigators have defined histological necrosis, as a prognostic factor, as the presence of any microscopic coagulative tumor necrosis, without consideration of degenerative changes such as hyalinization, hemorrhage, and fibrosis [17,18]. We also determined tumor necrosis as the presence of microscopic coagulative tumor necrosis. As we showed in Tables 2, 3, tumor necrosis was statistically significant only in the late recurrence group. In consideration of the hypothesis that rapid tumor cell growth outgrows its own blood supply and subsequently creates a hypoxic condition and resultant tumor necrosis, this result might be contradictory. However, in several recent reports, a high proliferation index or insufficient oxygen supply (hypoxia inducible factor-1, or HIF-1a) and high Ki-67 (a proliferation marker) expression were not interchangeable with tumor necrosis. Some medical literature suggests that host immunologic factors, such as differential expression of chemokines, may be involved in tumor necrosis [19]. Although the exact nature of immunologic triggers and tumor necrosis remain to be clarified, there might be a kind of 'time-related function' between 'tumor necrosis' and a 'tumor dormancy state'.

This study also showed the diversity of the recurrence site in the late recurrence group. The lung, liver, and bone are known as the usual recurrence organs of malignant neoplasms, including RCC [20], and we found such a tendency in the early recurrence group. However, there was a predominance of cancer recurrence at other sites in the late recurrence group, such as the lymph node, brain, pancreas, gall bladder, stomach, duodenum, colon, muscle, contralateral kidney, and pleura. The question of differences in metastasis site according to recurrence-free interval remains to be answered. Bruin et al. [21] suggested that organ-specific metastasis localization can be predicted by specific genomic aberrations in primary colorectal cancer [5]. Yerushalmi et al. [22] showed that cancer antigen 125 levels varied among the different sites of metastasis in breast cancer [5]. Koo et al. [23] reported that metastatic breast cancer showed different phenotypes of estrogen receptor, progesterone receptor, and human epidermal growth factor receptor 2 according to the different recurrence sites [5]. Therefore, understanding the molecular biological mechanisms underlying RCC might solve these questions.

The potential limitations of our study lie in its retrospective design. There was no standardized postoperative follow-up protocol and the quality of imaging modality has improved substantially during the past 30 years. No incorporation of molecular markers is another limitation of this study. Furthermore, the small number of patients who were evaluated in a single institution might have affected our study data. For this reason, we are planning to conduct a multicenter study with a large patient pool to minimize these biases.

This study suggested that the presence of microscopic coagulative necrosis in the resected specimen may be a predictive factor of late recurrence after radical nephrectomy for localized RCC. Also, clinical symptoms such as hematuria, flank pain, and a palpable mass at diagnosis may predict disease recurrence in a short period after radical nephrectomy. Large tumor size, advanced pathologic stage, and advanced Fuhrman nuclear grade may be risk factors for both early and late recurrence. Therefore, we suggest that patients with tumor necrosis may need to undergo long-term, thorough surveillance after radical nephrectomy for localized RCC.