Local and Regional Recurrences of Clinically Localized Renal Cell Carcinoma after Nephrectomy: A 15 Year Institutional Experience with Prognostic Features and Oncologic Outcomes

Abstract

OBJECTIVE

To evaluate outcomes for patients with local recurrence (LR) of clinically localized renal cell carcinoma (RCC) without concurrent systemic metastasis from our institution, an event that occurs rarely (1%–3%) after surgery. LR may be a harbinger of poor outcomes, and the best management of these patients is unclear.

MATERIALS/METHODS

We retrospectively reviewed patients surgically treated for clinically localized RCC (cT1-2N0M0) with subsequent LR (in the partial or radical nephrectomy bed) and/or regional recurrence (RR; in the abdomen distant from the direct site of surgery) without concurrent metastasis from our institutional database (2004–2018). Comparative and survival analyses were performed.

RESULTS

Out of 3038 total patients, 1895 had clinically localized RCC, with 30 patients (1.6%) having isolated LR/RR. Median time to recurrence was 26.5 months (IQR:16–35). Of 26 patients treated with local therapy, 14 (53.8%) recurred over a median follow-up time of 29.5 months (IQR:12–45). The 1-year and 2-year secondary recurrence-free survival rates are 60.7% and 49.7%, respectively. Two or more sites of locoregional recurrence significantly predicted secondary recurrence/metastasis after local therapy for local recurrence (hazard ratio: 2.22, P= .04).

CONCLUSION

Our results suggest local therapy is appropriate for select patients with LR/RR, with almost 50% of patients undergoing a second local therapy remaining alive with “local cure” and no secondary recurrence. The number of sites of recurrence can be used to better select patients that will benefit from local therapy or systemic/combination therapy. This work provides a framework onto which further studies regarding local therapy and locoregional recurrence of RCC can be performed.

Isolated local recurrence (LR) without synchronous distant metastasis of clinically localized renal cell carcinoma (RCC) after initial nephrectomy is a rare occurrence, with an incidence of only 1%–3% in post-nephrectomy patients.^1–8 Isolated LR has been defined differently among studies and is typically defined as recurrent RCC within the renal fossa, but the definition may extend to a regional recurrence (RR) and include ipsilateral regional structures such as the retroperitoneal lymph nodes, adrenal gland, and psoas muscle.^1,2,4 Given its relative rarity, studies regarding natural history, management, and prognosis typically entail case series and retrospective reviews, highlighting the need for further studies to elucidate these details. The time to LR/RR after initial nephrectomy can be quite variable, with median times in different series ranging from 10 months- 83 months, and with case reports showing times to recurrence as low as 3 months and as high as 45 years.^2,7,9–11 Furthermore, predicting the patients most likely to have an isolated LR without systemic progression remains challenging; only coagulative tumor necrosis and pathologic stage are identified as risk factors.⁴

Without a standardized management strategy for isolated LR/RR of RCC, decisions regarding management can be challenging for physicians and patients. Repeat partial nephrectomy (PN), salvage radical nephrectomy (RN), radiofrequency or cryoablation, systemic therapy, radiation therapy, and surveillance are reported as potential management options for isolated LR. Some studies suggest a survival benefit to surgical therapy.^12–15 Itano et al first presented a retrospective study of 30 patients with LR/RR after initial RN, finding that patients managed expectantly for their LR had a 5 year survival rate of 13% compared with 51% for those managed surgically.¹ However, over a decade later, urologists still do not have a standardized approach to managing LR/RR of RCC. Recently, a novel classification system for LR/RR following resection of intermediate-high risk nonmetastatic RCC was developed from analysis of the ASSURE (Sunitinib Malate or Sorafenib Tosylate in Treating Patients with Kidney Cancer that was Removed by Surgery; ECOG-ACRIN E2805) Adjuvant Trial, suggesting a poorer prognosis for patients with multiple sites of LR.¹⁶ With improving systemic therapies and a growing body of literature minimizing the role of cytoreduction in metastatic RCC, establishing the role of local therapy (surgery, ablation) in locoregional recurrences, if any, is of temporal importance.^17,18 We hypothesize that select patients with LR and RR will have a survival benefit from local therapy and endeavor to define selection criteria based on institutional experience.

MATERIALS AND METHODS

The IRB-approved Johns Hopkins Renal Mass Registry was queried for patients with clinically localized (cT1-2N0M0) and pathologically proven RCC who underwent RN or PN with curative intent and developed a LR or RR. Professional data abstractors prospectively maintain the registry with comprehensive demographic, clinical, and pathologic data collected and entered with multiple data reviews to ensure accurate and complete data. The diagnosis of LR or RR was made based on axial imaging. LR was defined as recurrent RCC in the ipsilateral kidney, nephrectomy bed, or renal fossa. RR was defined as recurrent RCC limited to the abdomen in the ipsilateral retroperitoneal lymph nodes, adrenal gland, abdominal wall, omentum, regional bowel, or pancreas. Inclusion of the pancreas as a LR/RR recurrence is supported by prior institutional experience and the ASSURE-based classification.^16,19 Distant metastatic recurrence was defined as RCC in the bone, CNS, thorax, skin, or liver.

Descriptive statistics and survival analyses were performed. χ2 and Fisher’s exact tests, Student’s t test, and Wilcoxon rank-sum test were used to evaluate differences between patients with and without secondary recurrence after their primary LR and/or RR. Recurrence-free and cancer-specific survival outcomes were examined using the Kaplan-Meier method and compared with the log-rank test. Cox regression was used to analyze potential predictors of secondary recurrence/metastasis and death following management of primary LR and/or RR. All tests were 2-sided, and a P-value <0.05 was considered statistically significant. Statistical analyses were performed using Stata/MP 15.0 (STATA Corp, College Station, TX, 2017).

RESULTS

Of 3038 patients in the Johns Hopkins Renal Mass Registry from 2004–2019, 1895 (62.4%) underwent RN or PN with curative intent for a clinically localized (cT1-2N0M0) and pathologically proven RCC. 75 out of 1895 (4%) patients had RCC LR and/or RR in general, while 30 out of 1895 (1.6%) had LR and/or RR without concurrent distant metastatic recurrence, as outlined in Supplementary Figure 1.

Patient, surgery, and tumor data for these 30 patients are detailed in Table 1. Of note, 18 patients (60%) had undergone RN and 12 patients (40%) had undergone PN as initial management of RCC. The locations of the primary LR and/or RR are detailed in Supplementary Table 1. 10 patients experienced LR, 13 RR, and 7 both LR and RR. Primary recurrence pathology, grade, and histologic subtype when available (20/30 patients) were consistent with initial RCC pathology, grade, and histologic subtype. The median time to primary LR and/or RR following initial nephrectomy for RCC was 26.5 months (IQR:16–35).

Table 1.

Demographics, RCC grade, pathologic stage, histology, and management of recurrence (n= 30).

	Number of Patients (n)
Gender
Male	22 (73%)
Female	8 (27%)
Race
Black	5 (17%)
White	25 (83%)
Age at Initial Surgery
Median: 57.7y, IQR (45.6–66)
35–44 y	6 (20%)
45–54 y	6 (20%)
55–64 y	8 (27%)
65–74 y	10 (33%)
Laterality of RCC
Left	14 (47%)
Right	16 (53%)
Initial Surgery for RCC
Radical Nephrectomy	18 (60%)
Open	5 (17%)
Laparoscopic	11 (37%)
Robotic	2 (7%)
Partial Nephrectomy	12 (40%)
Open	5 (17%)
Laparoscopic	5 (17%)
Robotic	2 (7%)
Location of Initial Surgery
Johns Hopkins Hospital	19 (63%)
Outside Hospital	11 (37%)
RCC Histologic Subtype
Clear Cell	21 (70%)
Chromophobe	2 (7%)
Mixed Papillary	1 (3%)
Papillary	4 (13%)
Translocation	1 (3%)
Unclassified	1 (3%)
RCC Grade *
Grade 1	1 (3%)
Grade 2	11 (37%)
Grade 3	11 (37%)
Grade 4	6 (20%)
RCC Pathologic T Stage *
pT1 or pT2	16 (53%)
pT3	13 (43%)
Margin Status *
Negative Margin	28 (97%)
Positive Margin	1 (3%)
Management of Recurrence
Local Therapy Alone	23 (77%)
Local + Systemic Therapy	3 (10%)
Systemic Therapy	2 (7%)
Radiotherapy	1 (3%)
Active Surveillance	1 (3%)

^*1 patient did not have an initial pathology report.

Of the 30 patients with primary LR and/or RR, 26 (86.7%) underwent local therapy (surgical resection: 24/26 (92.3%), thermal ablation: 2/26 (7.7%)) for their recurrence, as detailed in Supplementary Table 2. Of note, 3 of these 26 (11.5%) patients were concurrently managed with systemic therapy (IL-2: n = 1, pazopanib: n = 2). The remaining 4 patients (13.3%) were managed with active surveillance (n = 1, LR), systemic therapy (RR treated with sunitinib: n = 1; RR treated with IL-2: n = 1), or radiation therapy (LR: n = 1).

After a median follow-up time of 29.5 months (IQR: 12–45) from the time of management of their primary recurrence, 14 (53.8%) of the 26 patients managed with local therapy for their primary recurrences went on to develop secondary recurrences; the median time to secondary recurrence for these 14 patients was 11.5 months (IQR: 7–29). These patients with secondary recurrences were more likely to have had initial high-grade (Grade 3–4) RCC compared to those who did not experience secondary recurrences (85.7% vs 18.2%, P= .001). No other significant differences in terms of demographics, initial tumor pathology, and type and approach of initial nephrectomy were found for patients with and without eventual secondary recurrences. The 1-year and 2-year secondary recurrence-free survival rates were 60.7% and 49.7%, respectively. Notably, upon univariate Cox regression analysis, the number of primary recurrence sites (≥2) was highly predictive of secondary recurrence/metastasis (hazard ratio: 2.22; 95% CI: 1.21–9.42, P= .04; Fig. 1A). Type of primary recurrence (LR vs. RR; Fig. 1B), management approach (surgical resection vs thermal ablation), and primary recurrence grade and histology did not significantly predict subsequent recurrence in our cohort (Table 2). No meaningful multivariable analysis could be performed due to the small sample size and number of events.

Figure 1.

(A) Kaplan Meier survival analysis of the effect of the number of different sites of primary recurrence of RCC on secondary recurrence of RCC/metastasis. (B) Kaplan Meier survival analysis of the effect of the type of primary recurrence of RCC on secondary recurrence of RCC/metastasis. (Color version available online.)

Table 2.

Univariate Cox regression analysis^* for predicting the hazard of secondary recurrence/metastasis in cT1–2 RCC patients who underwent local therapy (surgical resection or thermal ablation) for a primary local/regional/locoregional recurrence.

Variable		HR (95% CI)	P-value
Type of Primary Recurrence	Local	REF	–
	Locoregional	1.83 (0.61–5.53)	.3
Management of Primary Recurrence	Thermal ablation	REF	–
	Surgical resection	0.55 (0.12–2.54)	.4
Number of Primary Recurrence Sites	1	REF	–
	≥2	2.22 (1.21–9.42)	.04
Primary Recurrence Grade	Low	REF	–
	High	5.20 (0.57–17.61)	.1
Primary Recurrence Histology	Clear cell	REF	–
	Non-clear cell	0.52 (0.11–2.53)	.4

HR, hazard ratio; 95% CI, 95% confidence interval

^*Analysis performed in 23 of 26 patients with available follow-up after management of primary recurrence.

Figure 2 demonstrates detailed outcomes for the 26 patients who underwent local therapy for their primary LR and/or RR. Of the 26 patients, 15 (71.4%) had only 1 site of primary recurrence, while 11 (28.6%) had more than 1 site at initial recurrence. Of 15 patients with 1 site of primary recurrence treated with local therapy, 5 (33.3%) had subsequently no evidence of disease after a median follow-up time of 40 months (IQR: 26–78 months). Of the 11 patients with >1 site of primary recurrence treated with local therapy, 7 (63.6%) had subsequently no evidence of disease after a median follow-up time of 3 months (IQR: 1–12 months); of note, 3 patients had no available follow-up beyond 1 month, and excluding them results in 4 remaining patients with no evidence of disease after a median follow-up time of 12 months (IQR: 3–30 months). Overall, 12 (46.2%) of the 26 patients who underwent local therapy for primary LR and/or RR remained disease-free and alive (2 received pazopanib). Of the remaining 14 (53.8% of 26) patients who developed secondary recurrences, 2 (7.7% of 26%) remained alive without evidence of distant metastatic disease after management of the secondary recurrences. In total, 14 (53.8%) of the 26 patients managed with local therapy for primary recurrence remained distant metastasis-free and alive without any initial or subsequent systemic therapy. The 5-year cancer-specific survival rate after local management of primary LR and/or RR was 67.5%. The number of primary recurrence sites (≥2) was associated with higher cancer-specific mortality, but this association did not meet traditional levels of statistical significance (hazard ratio: 5.83; 95% CI: 0.80–22.44, P= .08).

Figure 2.

Outcomes of patients with primary LR and/or RR without concurrent distant metastasis. Of note, secondary recurrence and metastasis are not grouped unlike in Table 2. (A) The median overall follow-up time for patients who did not undergo local therapy for primary recurrence (n = 4) is 44 months (IQR 40–48 months).(B) For patients with 1 site of LR and/or RR, the median overall follow-up times for each grouping are as follows: NED after local therapy (n = 5): 40 months (IQR 26–78 months); secondary recurrence after initial local therapy with NEMD after subsequent management (n = 1): 42 months; secondary recurrence after initial local therapy with distant metastasis/death after subsequent local therapy (n = 7): 43 months (IQR 33–51 months); secondary recurrence after initial local therapy with distant metastasis/death after subsequent medical management (sunitinib) (n = 2): 62 months (IQR 28–96 months).(C) For patients with >1 site of LR and/or RR, the median overall follow-up times for each grouping are as follows: NED after local therapy (n = 7): 3 months (IQR 1–12 months); secondary recurrence after initial local therapy with distant metastasis/death after subsequent local therapy (n = 1): 18 months; secondary recurrence after initial local therapy with NEMD after subsequent medical management (nivolumab and ipilimumab) (n = 1): 23 months; secondary recurrence after initial local therapy with distant metastasis/death after subsequent medical management (vorinostat and bevacizumab: n = 1, sunitinib: n = 1) (n = 2): 35 months (IQR 19–51 months). NED, no evidence of disease; NEMD, no evidence of meta-static disease. *3 patients, all with >1 site of LR and/or RR, concurrently underwent systemic therapy (IL-2, n = 1; pazopanib, n = 2). **2 patients with 1 site of LR underwent ablation for their LR. Both patients had secondary recurrence, 1 is alive with NEMD, the other developed distant metastases. (Color version available online.)

DISCUSSION

Given the low incidence of isolated locoregional recurrence of RCC, the literature is predominantly comprised of case series and small retrospective studies. Urologists currently lack a standardized management strategy for these isolated LRs of RCC, but to date, surgical management in select patients has been shown to be preferable compared to non–surgical interventions alone.^6,8,13–15 Additional studies report 5-year survival rates of 30%–62% for patients with isolated LR of RCC treated surgically.^6,8 In 2009, Margulis et al conducted a retrospective analysis of 54 patients, one of the largest LR of RCC cohorts to date, who had undergone initial RN for RCC with curative intent.² All of these patients were managed surgically with 69% of patients receiving perioperative systemic therapy as an adjunct.² 35 (64.8%) patients had secondarily recurrent disease at the time of publication, with median recurrence-free and cancer-specific survival of 11 and 61 months respectively.² Notably, the authors found multiple adverse risk factors for increased risk of cancer-specific death including positive surgical margin after resection of LR, size of recurrent tumor, sarcomatoid histologic features, abnormal serum alkaline phosphatase, and increased lactate dehydrogenase at time of recurrence.² However, the field continues to lack guidelines as to which patients are best managed surgically, highlighting the need for data to assist with patient selection for different types of management for LR of RCC. While a rare occurrence (1.6% of our cohort), 38.1%–52.1% of patients with isolated LR or RR managed with local therapy remain disease-free over an extended time-period. In this study, patients with 1 site of locoregional recurrence had the greatest rate of durable response with local therapy, although some patients with multiple sites of recurrence or multiple recurrences experienced a durable disease-free interval without systemic therapy.

Our data support the concept that recurrence of RCC after definitive surgery most often manifests as a systemic disease. Very few patients have isolated locoregional recurrence, with most recurring with metastatic disease.²⁰ As many of our patients with LR were referred in from outside institutions, our data may overestimate isolated locoregional recurrence. Our patients had a median time to primary recurrence following surgical management of RCC of 26.5 months, consistent with prior studies.¹² While rare, LR and/or RR of RCC portends a poor prognosis.¹ With such a potentially long-time to recurrence, appropriate long-term follow-up after initial nephrectomy is important in identifying patients who need intervention as 50%–80% of these LR and/or RR are asymptomatic.^1,2,5–8 Furthermore, since a subset of patients with isolated locoregional recurrence benefit from local therapy, identifying these patients is essential.

Of the 26 patients treated with local therapy (surgical resection or ablation) for their primary LR and/or RR, 46.1% (n = 12) were rendered disease-free by 1 local therapy, with an additional 3.8% (n = 1) rendered disease-free with a second local therapy. None of these patients developed distant metastasis, with only 3 patients getting adjuvant systemic therapy (rationale for therapy unavailable in chart review), showing that durable metastasis-free survival is indeed possible by local therapy alone in a large proportion of well-selected patients (eg surgical candidates with surgically resectable disease). This supplement the prior literature suggesting that patients managed locally for their LR and/or RR can have 5-year survival of 30%–62% vs 18% and 13% for medical and expectant management respectively.^1,2 These patients had excellent 1- and 2-year secondary recurrence-free survival rates, suggesting that aggressive surgical intervention can remove the need for and/or delay the administration of systemic therapy. A median time to secondary recurrence of 11.5 months suggests that patients could potentially delay onset of systemic therapy a year after their initial diagnosis. The avoidance of systemic therapy for such a period could be a benefit for these patients, similar to conclusions made regarding cytoreductive nephrectomy for select patients with meta-static RCC in the modern era of systemic therapy.²¹ Of note, 4 patients in this series were managed with robotic surgery, helping supplement the literature regarding the feasibility of this approach.²²

Patients with a single site of locoregional recurrence had significantly improved secondary recurrence-free survival, as well as higher cancer-specific survival that did not meet traditional levels of statistical significance, compared to patients with multiple sites of locoregional recurrence. This is consistent with the findings of the analysis of the ASSURE trial, where patients with type IV recurrences (> 1 site of locoregional recurrence) were associated with worse survival outcomes vs other recurrence pattern types. Tumor grade and histology did not predict subsequent recurrence, suggesting that the underlying genetic changes facilitating locoregional recurrence may not correlate with these 2 histologic descriptors. The number of sites of LR and/or RR, along with the adverse risk factors detailed earlier by Margulis et al, could be used to help identify which patients are most amenable to local management for their LR and/or RR of RCC. These adverse risk factors together could signal more aggressive RCC recurrence or occult distant meta-static RCC recurrence. While prospective and larger studies are ideally needed to identify exactly which patients with LR and/or RR of RCC are amenable to local therapy, these studies provide a foundation upon which recommendations can be designed.

Our study included only 3 patients treated with local and systemic therapy (IL-2, n = 1; pazopanib, n = 2) for their primary recurrence. Combination therapy with surgery and systemic or radiation therapy is of unclear benefit. Minimal benefit was shown with addition of radiotherapy or systemic therapy to surgical management in several small studies.^8,23,24 With new and more effective types of systemic therapy including tyrosine kinase and immune checkpoint inhibitors, further studies using combination therapy are needed.^25,26

While a variety of factors may affect the likelihood of LR for patients, more granular data is necessary to further delve into these potential relationships. Potential factors include surgeon experience, differences in minimally invasive techniques, and surgical approach. With 11 of 30 (37%) patients being referred in for their disease, it is likely that surgeon experience plays a role in LR. In terms of surgical approach, no clear patterns were noticed regarding potential port site hernia recurrences, with the majority of patients in our study having LR in the renal fossa or nephrectomy bed. Biology also likely plays a role in the LR of RCC, as some patients with multiple sites of locoregional recurrence still went on to have durable survival after local therapy, speaking to a possible attenuated progression of disease for patients with isolated locoregional recurrence rather than a rapid progression to distant metastases.²⁷

Limitations of this study include the retrospective study design (with potential for selection bias given tumor size, complexity, patient comorbidities in locally vs non–locally managed recurrences), small sample size, and inherent selection bias to a large referral center for complex RCC care. To date, no prospective trials have been performed regarding locoregional recurrence of RCC nor are they likely to be completed in the future given the rarity of isolated LR and/or RR of RCC. This study could not assess the natural history of untreated recurrences or analyze other forms of locoregional control (eg stereotactic body radiation therapy) given small sample size and lack of standardized follow up. Patients were referred to our institution and selected for surgical management based on the volume of recurrent disease and suitability for surgery. While this limit the scope of the conclusions, it does not detract from the substantiation that select patients with locoregional recurrence of RCC benefit from local therapy.

CONCLUSION

A minority of patients will experience an isolated local or regional recurrence of RCC after definitive surgery, supporting the notion that most recurrences of RCC manifest as systematic disease. However, almost 50% of patients with LR or RR in this select surgical/ablative series demonstrated a disease-free interval without the need for systemic therapy. Patients with a solitary site of recurrence may be most appropriate for local therapy, although durable disease-free survival was also observed in patients with multiple sites of recurrence. This work provides a framework onto which further studies regarding surgery/ablation and locoregional recurrence of clinically localized RCC can be performed.