Adverse Pathologic Features Impact Survival Outcomes for Small Renal Masses Following Nephrectomy

Abstract

Purpose:

While most small renal masses (SRM) < 4 cm have an excellent prognosis following resection, the impact of adverse T3a pathologic features on oncologic outcomes of SRMs remains unclear. We sought to compare clinical outcomes for surgically resected pT3a versus pT1a SRMs at our institution.

Materials and Methods:

We retrospectively reviewed records of patients who underwent radical or partial nephrectomy (RN, PN) for renal tumors < 4 cm at our institution between 2010–2020. We compared features and outcomes of pT3a vs pT1a SRMs. Continuous and categorical variables were compared using Student’s T and Pearson’s Chi-squared tests, respectively. Post-operative outcomes of interest including overall, cancer-specific, and recurrence-free survival (OS, CSS, RFS) were analyzed using Kaplan-Meier method, Cox proportional hazard regression, and competing risk analysis. Analyses were performed using R statistical package (R Foundation, v4.0).

Results:

We identified 1,837 patients with malignant SRMs. Predictors of post-operative pT3a upstaging included higher renal score, larger tumor size, and presence of radiologic features concerning for T3a disease (OR = 5.45, CI 3.92 – 7.59, p < 0.001). On univariable modeling, pT3a SRMs had higher positive margin rates (9.6% vs 4.1%, p < 0.001), worse OS (HR = 2.9, 95% CI 1.6–5.3, p = 0.002), RFS (HR 9.32, 95% CI 2–40.1, p = 0.003), and CSS (HR = 3.6, 95% CI 1.5 – 8.2, p = 0.003). On multivariable modeling, pT3a status remained associated with worse RFS (HR = 2.7, 95% CI 1.04 – 7, p = 0.04), but not OS (HR 1.6, 95% CI = 0.83 – 3.1, p = 0.2); multivariable modeling was deferred for CSS due to low event rates.

Conclusions:

Adverse T3a pathologic features portend worse outcomes for SRMs, highlighting the crucial role of preoperative planning and case selection. These patients have relatively poor prognosis, and should be monitored more closely and counseled for consideration of adjuvant therapy or clinical trials.

Funding:

National Institutes of Health Cancer Center Support Grant (NIH, NCI; 2P30CA008748–48)

SUMMARY

Objectives:

We sought to compare clinical outcomes for patients with pT3a versus pT1a small renal masses (SRMs) < 4 cm that underwent resection at our institution, to determine the impact of adverse T3a pathologic features on oncologic outcomes in SRMs.

Methods and Materials:

We retrospectively reviewed records of patients who underwent radical or partial nephrectomy (RN, PN) for renal tumors < 4 cm at our institution between 2010–2020. We compared features and outcomes of pT3a vs pT1a SRMs. Continuous and categorical variables were compared using Student’s T and Pearson’s Chi-squared tests, respectively. Post-operative outcomes of interest including overall, cancer-specific, and recurrence-free survival (OS, CSS, RFS) were analyzed using Kaplan-Meier method, Cox proportional hazard regression, and competing risk analysis. Analyses were performed using R statistical package (R Foundation, v4.0).

Results:

We identified 1,837 patients with malignant SRMs. Predictors of post-operative pT3a upstaging included higher renal score, larger tumor size, and presence of radiologic features concerning for T3a disease (OR = 5.45, CI 3.92 – 7.59, p < 0.001). On univariable modeling, pT3a SRMs had higher positive margin rates (9.6% vs 4.1%, p < 0.001), worse OS (HR = 2.9, 95% CI 1.6–5.3, p = 0.002), RFS (HR 9.32, 95% CI 2–40.1, p = 0.003), and CSS (HR = 3.6, 95% CI 1.5 – 8.2, p = 0.003). On multivariable modeling, pT3a status remained associated with worse RFS, but worse RFS (HR = 2.7, 95% CI 1.04 – 7, p = 0.04), but not OS (HR 1.6, 95% CI = 0.83 – 3.1, p = 0.2); multivariable modeling was deferred for CSS due to low event rates.

Conclusions:

Adverse T3a pathologic features portend worse outcomes for SRMs, highlighting the crucial role of pre-operative planning and case selection. These patients have relatively poor prognosis, and should be monitored more closely and counseled for consideration of adjuvant therapy or clinical trials.

INTRODUCTION

Most renal masses, particularly small renal masses (SRMs), are now incidentally detected on imaging studies usually performed for nonspecific abdominal or musculoskeletal complaints¹. Following surgical resection, the management and follow-up of malignant SRMs are largely dependent on pathological staging^2,3. However, while pathologic staging for pT1-pT2 RCC is driven by tumor size (≤ 7 cm for pT1, > 7 cm for pT2), renal tumors of any size may be upstaged to pT3a if they exhibit specific adverse pathologic features, including tumor extension into the renal vein or its segmental branches, or peri-renal or renal sinus fat⁴.

Several retrospective studies have evaluated the influence of upstaging to pT3a disease on oncologic outcomes in cT1 renal masses (≤ 7 cm)^5–12, with a recent meta-analysis of these studies concluding that upstaging from cT1 to pT3a portends worse overall, cancer-specific, and recurrence-free survival (OS, CSS, RFS) for patients¹³. However, the incidence and influence of pathologic upstaging to T3a disease on oncologic outcomes in SRMs (< 4cm) remain unclear.

In this study, we retrospectively reviewed records of patients with SRMs who underwent surgery at our institution over a 10-year period. Our primary objectives were to compare the peri-operative features as well as oncologic outcomes of pT3a and pT1a SRMs, and identify the predictors of post-operative pathologic upstaging to pT3a disease in SRMs.

METHODS

Study Population and Data Collection

Following institutional review board approval, we retrospectively reviewed our institutional kidney cancer surgical database for patients who underwent RN or PN using open or minimally invasive surgical (MIS) techniques for suspected renal cortical tumors between 1/1/2010 and 05/10/2021. Exclusion criteria included masses > 4 cm, benign masses, known or suspected advanced (≥ pT3b and/or N1/M1) disease preoperatively and incomplete pathology or radiology reports. Following surgery, oncological follow-up schedule was in alignment with the AUA guideline on management of renal masses^2,14.

Pre-operative variables of interest included age, gender, BMI, Charlson comorbidity index (CCI), preoperative tumor size (maximum tumor diameter on imaging), enrollment in active surveillance, and adverse radiologic features (ARFs), defined as RFs concerning for cT3 disease noted on pre-operative cross-sectional imaging, such as tumor abutment or extension into the renal sinus, hilar vessels, or perirenal fat. RENAL nephrometry scores¹⁵ were also retrospectively calculated by the authors (SK, BD, PKL) for all pT3a masses and using a representative subset of pT1a masses (280, 16.8%) that were previously assessed using our institutional Arterial Based Complexity (ABC) nephrometry score¹⁶, a CT-based score developed and applied for all renal masses seen at our institution since August 2015 to estimate the morbidity of PN. The authors were not blinded to pT stage or the ABC score prior to RENAL nephrometric scoring. Our primary outcomes of interest included OS, CSS, and RFS, the latter including both local recurrence and distant metastases.

Statistical Analysis

Continuous and categorical variables were compared among SRM groups using Student’s T and Pearson’s Chi-squared tests, respectively. Potential predictors of post-operative upstaging to pT3a were examined using univariable and multivariable logistic regression modeling, with multivariable LRM utilizing predictors with p ≤ 0.1 on univariable LRM. OS was analyzed using Kaplan-Meier (K-M) and Cox-Proportional Hazard Regression (Cox-PHR) modeling, while CSS and RFS were assessed in terms of cumulative incidence of kidney cancer (KC)-related death and recurrence, respectively, using competing risk analysis (CR) models^17–19 to account for censoring by competing events that may have precluded detecting the event of interest, namely death unrelated to KC for CSS, and death due to any cause for RFS; CR subdistribution hazard ratios were used to assess potential predictors for CSS and RFS. Multivariable analyses were performed for OS, and RFS using clinically relevant variables, but deferred for CSS due to the low event rate. All analyses were performed using R statistical package (R Foundation, v4.0).

RESULTS

Overview of Cohort

We identified 2,130 patients with SRMs that underwent PN or RN between 1/1/2010 and 05/10/2021, of which 293 (13.76%) were benign and excluded from further analyses. The characteristics of the remaining 1,837 patients with malignant SRMs are summarized in Table 1. The median age at nephrectomy was 60 years (IQR 52–68), with most patients undergoing PN (94.7%). Median tumor size was 2.5 cm (IQR 1.9 – 2.6 cm), with a 4.6% positive surgical margin (PSM) rate. Clear-cell renal cell carcinoma (cc-RCC) was the most common histology (62.6%), followed by papillary RCC (15.8%), and chromophobe RCC (9.7%). The proportion of pT3a SRMs was 9.04% (166).

Table 1.

Peri-operative features of cohort.

Characteristic	Value*
Mean age at nephrectomy	59 ± 11
Male gender	1205 (65.6%)
BMI	30 ± 6.1
Charlson comorbidity Index	0.83 ± 1.3
Previously on active surveillance	119 (6.5%)
R.E.N.A.L nephrometry score	7.19 ± 1.8
Adverse radiologic features	341 (18.6%)
Underwent radical nephrectomy	97 (5.3%)
Surgery done via minimally invasive approach	771 (42%)
Tumor size (max diameter) on pathology	2.6 ± 0.9
Positive margin	85 (4.6%)
Tumor histology predominantly cc-RCC	1,149 (62.6%)
Nuclear grade
G1/2	827 (45%)
G3/4	672 (36.6%)
NA	338 (18.4%)

^*Mean ± SD; n (%)

Peri-operative Features of pT3a and pT1a SRMs

Compared to patients with pT1a SRMs, patients with pT3a SRMs were older and more likely to be male (median age at nephrectomy 62 vs 59, p < 0.001; 76% vs 58.7%, p = 0.003), but did not differ significantly in their BMI or Charlson Comorbidity Index (Table 2). Compared to pT1a SRMs, pT3a SRMs had higher mean RENAL nephrometry score¹⁵, mean tumor size (7.82 vs 6.83, p < 0.001; 3.0 vs 2.5 cm, p < 0.001, respectively), ISUP grade (G3/4) tumors (63 vs 34%, p < 0.001), and PSM rates (9.6% vs 4.1%, p < 0.001) (Table 2). pT3a SRMs had a higher incidence of ARFs on pre-operative imaging (49% vs 15.5%; OR 6.1, 95% CI 4.3 – 8.7, p < 0.001). Of these, the most commonly noted ARF was contact with the renal sinus (47.6% in pT3a vs 15.1% in pT1a, OR 5.1, p < 0.001; Supp Table 1). pT3a masses were more likely to undergo RN (20% vs 3.8%, p < 0.001), but with no significant difference in the odds of intra-op conversion from PN to RN relative to pT1a SRMs (1.8% vs 0.9%, respectively, p = 0.2).

Table 2.

Comparison of peri-operative features between pT1a and pT3a SRMs

Characteristic	pT1a* N = 1,671	pT3a* N = 166	p-value**
Mean age at nephrectomy	59 ± 11	62 ± 12	< 0.001
Male gender	1,079 (58.7%)	126 (76%)	0.003
BMI	30 ± 6.1	30.7 ± 6.2	0.2
Charlson comorbidity Index	0.82 ± 1.29	0.98 ± 1.4	0.1
Previously on active surveillance	106 (6.3%)	14 (8.4%)	0.3
R.E.N.A.L nephrometry score	6.83 ± 1.7	7.82 ± 1.9	< 0.001
Adverse radiologic features	259 (15.5%)	82 (49%)	< 0.001
Underwent radical nephrectomy	64 (3.8%)	35 (20%)	< 0.001
Surgery done via minimally invasive approach	703 (42.1%)	68 (41%)	0.8
Tumor size (max diameter) on pathology	2.5 ± 0.9	3.0 ± 0.9	< 0.001
Positive surgical margin	69 (4.1%)	16 (9.6%)	< 0.001
Tumor histology predominantly ccRCC	1,039 (62.2%)	110 (66%)	0.3
Nuclear grade			< 0.001
G1/2	786 (47%)	41 (25%)
G3/4	568 (34%)	104 (63%)
NA	317 (19%)	21 (13%)

^*Mean ± SD; n (%)

^**Student’s T-test; Pearson’s Chi-squared test

Predictors of Postoperative pT3 Upstaging

On univariable LRM (u-LRM), statistically significant predictors of post-operative pathologic upstaging of SRMs to pT3a included older age (OR = 1.03, CI = 1.02 – 1.05, p < 0.001), male gender (OR = 1.76, 1.23 – 2.57, p = 0.003), presence of ARFs (OR = 5.45, CI 3.92 – 7.59, p < 0.001), higher RENAL score (OR = 1.37, CI = 1.22 – 1.54, p < 0.001), and larger tumor size (OR = 1.92, CI 1.55 – 2.41, p < 0.001); these variables remained statistically significant on multivariable LRM (m-LRM). BMI and CCI were not significantly associated with upstaging to pT3a disease on u-LRM (p = 0.2 and 0.13, respectively) and were thus excluded from m-LRM (Table 3).

Table 3. Predictors of post-operative upstaging to pT3 on univariable and multivariable logistic regression modeling (LRM).

Variables with p ≤ 0.1 on univariable LRM were applied to multivariable LRM. OR = odds radio; CI = confidence interval

	Univariable LRM			Multivariable LRM
Characteristic	OR	95% CI	p-value	OR	95% CI*	p-value
Mean age at nephrectomy	1.03	1.02 - 1.05	< 0.001	1.02	1.00 – 1.04	0.016
Male gender	1.73	1.21 - 2.53	0.004	2.31	1.44 – 3.78	< 0.001
R.E.N.A.L nephrometry score	1.38	1.22 - 1.55	< 0.001	1.17	1.02 – 1.35	0.022
Adverse radiologic features	5.32	3.82 - 7.42	< 0.001	1.58	1.58 – 4.3	< 0.001
Tumor size (max diameter) on pre-op imaging	1.95	1.56 – 2.44	< 0.001	2.6	1.17 – 2.16	0.003
BMI	1.02	0.99 – 1.05	0.2	Excluded from multivariable model
Charlson comorbidity Index	1.09	0.97 – 1.21	0.13	Excluded from multivariable model

Post-operative outcomes of pT1a vs pT3a masses

Median follow-up was 3.5 years. The overall incidence of all and kidney cancer-related deaths were 72 patients (3.9%) and 7 patients (0.4%), respectively (Supp Table 2), with higher incidence in pT3a vs pT1a SRMs (overall mortality 7.2% vs 3.6%, p=0.02; cancer-specific mortality 1.8% vs 0.2%, p = 0.02). Similarly, the rate of disease recurrence was higher in pT3a compared to pT1a SRMs (4.2% vs 1.5%, p = 0.02).

Overall survival was worse in pT3a compared to pT1a SRMs (log-rank p = 0.001; Figure 1). On univariable Cox-PHR, pT3a features were associated with worse OS (HR = 2.9, 95% CI 1.6–5.3, p = 0.002), along with CCI, age at nephrectomy, and tumor size (Table 4). On multivariable Cox-PHR, pT3a features were not significantly associated with OS (HR 1.6, 95% CI = 0.83 – 3.1, p = 0.2), adjusting for ARFs, Charlson CI, tumor size, and age at nephrectomy (Supp Table 3).

Figure 1.

Kaplan-Meier curve analysis for overall survival comparing pT1a to pT3a SRMs

Table 4. Univariable survival analysis for predictors of OS, CSS, and RFS.

Cox proportional hazard modeling was used for OS, while competing risk subdistribution hazard modeling was used for CSS and RFS.

	OS		CSS		RFS
Predictor	HR & 95% CI	p-value	HR & 95% CI	p-value	HR & 95% CI	p-value
pT3a features	2.7 (1.5 - 5.1)	0.005	9.4 (2.14, 41.27)	0.003	3.58 (1.55, 8.26)	0.003
Adverse RFs	1.8 (1.04 – 3.02)	0.046	2.08 (0.4, 10.8)	0.38	1.43 (0.62, 3.3)	0.4
Radical nephrectomy	4.7 (2.4 – 9.22)	< 0.001	4.43 (0.55, 35.88)	0.17	2.59 (0.79, 8.51)	0.12
Age at nephrectomy	1.1 (1.06 - 1.11)	< 0.001	1.2 (1.05, 1.38)	0.01	1.05 (1.02, 1.09)	0.005
Tumor size on pathology (max diameter in cm)	1.45 (1.14 - 1.84)	0.003	1.95 (1.18, 3.22)	0.01	1.64 (1.13, 2.39)	0.009
ccRCC histology	1.32 (0.81 - 2.17)	0.26	0.81 (0.18, 3.64)	0.79	0.9 (0.45, 1.83)	0.78
Positive surgical margins	1.01 (0.37 - 2.78)	0.98	2.74 (0.31, 24.18)	0.36	1.15 (0.27, 4.82)	0.85
Charlson Comorbidity Index	1.39 (1.24 - 1.56)	< 0.001	1.69 (1.28, 2.23)	<0.001	1.53 (1.32, 1.78)	<0.001

Similarly, pT3a stage was associated with worse CSS and RFS compared to pT1a on univariable CR analysis (HR 9.4, 95% CI 2.1 – 41.3, p = 0.003: HR = 3.6, 95% CI 1.6 – 8.3, p = 0.003, respectively; Table 4 and Figure 2). On multivariable modeling for predictors of RFS, pT3a status was significantly associated with worse RFS (HR 2.7, 95% CI 1.04 – 7, p = 0.041), adjusting for histology (ccRCC or non-ccRCC), tumor size, PSM, and nephrectomy type (RN or PN) (Supp Table 4). Multivariable modeling was deferred for CSS due to the low event rate.

Figure 2.

Comparison of (a) cancer specific and (b) recurrence-free survival between pT3a and pT1a SRMs, assessed in terms of cumulative incidence of kidney cancer deaths and disease recurrences, respectively, using competing risk model analyses.

DISCUSSION

The management of SRMs has evolved over the past two decades due to several factors, including the increasing incidental detection of SRMs on radiologic workup for unrelated complaints rather than the historic delayed diagnosis of kidney tumors when patients presented with classic symptoms of hematuria, flank pain, and flank mass ¹. The resultant size and stage migration lead to the expansion of kidney sparing operations and evidence for equivalent oncological outcomes between PN and RN in the management of SRMs^20–23. The important added advantage of PN was renal functional preservation and avoidance or delay in the onset of chronic kidney disease and associated cardiovascular morbidity and mortality^24,25.

Furthermore, as advances in histopathologic diagnosis revealed large variation in the oncologic implications of different tumor histologies, they also noted that certain adverse pathologic features predicted worse survival, even in masses that would have been staged pT1/pT2 based on size alone. Bonsib et al²⁶ noted tumor invasion into the renal sinus, branched muscular renal veins and perinephric fat portended a worse prognosis regardless of tumor size^27,28. These findings questioned the AJCC size-based staging system for pT1/pT2 masses, prompting the 6^th and following editions of the AJCC staging system to redefine pT3a stage based on the presence of these features, independent of tumor size^4,29.

In this study, we demonstrate that pT3a status remains an independent poor prognostic factor even in SRMs. pT3a SRMs were twice as likely to have positive margins (9.6 vs 4.1%, p < 0.001), and pT3a status was significantly associated with worse OS, RFS, and CSS on univariable models. On multivariable modeling, we found pT3a status to remain associated with worse RFS (HR = 2.7, 95% CI 1.04 – 7, p = 0.041, respectively) but not OS (HR 1.6, 95% CI = 0.83 – 3.1, p = 0.2). Comparing preoperative features of pT1a vs pT3a SRMs, pT1a were smaller (mean size 2.5 ± 0.9 cm vs 3.1 ± 0.9 cm; p < 0.001) and had lower RENAL nephrometry scores (mean score 6.83 ± 1.7 vs 7.82 ± 1.9), but did not differ from their pT3a counterparts in proportion of ccRCC (66% vs 62.2%, p = 0.3), although pT3a tumors had a higher ratio of ISUP grade G3/4 tumors (63% vs 34%, p < 0.001).

Limitations of our study include its retrospective nature and the relatively short median follow-up time (3.5 years), and the relatively low incidence of our outcomes of interest, which may be related to our short follow-up, thereby limiting our statistical assessment of oncologic outcomes. RENAL scores were also calculated retrospectively for all pT3a masses and only a representative subset of pT1a masses that were previously assessed using our institutional ABC nephrometry score¹⁶, and the lack of blinding of authors to final pathology and ABC score may have biased their RENAL score assessments. However, our findings are in agreement with prior findings by Chevinsky et al³⁰, which found pT3 stage to predict fourfold risk of disease recurrence compared to pT1/pT2 masses, even after adjusting for tumor size, with similar association noted when examining the SRM subsets of pT3 vs pT1/T2 masses in the cohort (HR 4.4, 95% CI 1.4 – 13.6, p = 0.01)³⁰.

Our findings emphasize careful pre-operative planning and vigilance in the management of SRMs, through careful examination of pre-operative imaging for adverse RFs associated with pT3a features, such as tumor abutment of the renal sinus and perinephric fat³¹, as well as muscular branch vein invasion^32,33. Despite marked improvements in imaging technology, the sensitivity of CT and MRI modalities for pT3a features remains somewhat limited with high levels of inter-reader variability and discrepancy^34–36, with one series noting radiologic staging based on axial imaging (CT or MRI) to have the lowest sensitivity for stage III RCC - 27.7%, versus 90.7%, 67.3%, and 64.2% for stages I, II, and IV, respectively³⁴. Similarly, while pT3a SRMs had a significantly higher incidence of adverse RFs compared to pT1a masses in our series (49% vs 15.5%; OR 6.1, 95% CI 4.3 – 8.7, p < 0.001), 51% of pT3a masses did not exhibit any concerning radiologic features preoperatively. This issue of “unexpected pathologic upstaging” of seemingly cT1 masses to pT3a post-operatively has been shown to portend worse OS, CSS, and RFS compared to cT1/pT1a masses^5–13. Such issues may again be mitigated by the use of additional imaging modalities such as pre- and intra-op renal ultrasound (RUS), which are routinely performed at our institution in addition to pre-operative cross-sectional imaging with CT or MRI, to detect otherwise unrecognized branched vein or renal sinus fat invasion³⁷. In our series, we found the presence of adverse RFs concerning for T3a disease on pre-operative imaging to be a significant predictor of upstaging to pT3a disease on both univariable and multivariable LRM, along with higher RENAL scores (univariable LRM OR = 5.32 and 1.38, respectively; multivariable LRM OR = 1.58 and 1.17; all with p < 0.001). pT3a tumors were more likely to undergo RN than their pT1a counterparts (21% vs 3.8% rates of RN) with no significant difference in the odds of intra-op conversion from PN to RN relative to pT1a SRMs (p = 0.2), suggesting that the operating surgeons were aware of these adverse features as they designed their operations. Interestingly, male gender was also a predictor of upstaging to pT3a disease (univariable LRM OR 1.73, p = 0.004; multivariable LRM OR 2.3, p < 0.001), a finding noted on previous analyses of surgically resected tumors in SEER³⁸, NCDBI³⁹, and the Dutch PALGA⁴⁰ registries, which found female gender to be associated with smaller size and lower stage and grade on presentation^38–40, although the etiology underlying this difference remains unclear.

In summary, our findings emphasize that pT3a stage portends worse survival outcomes in SRMs, a subgroup of renal masses where current guidelines on renal cell carcinoma such as the EAU⁴¹, NCCN³, and AUA⁴² guidelines list active surveillance as a management option. It is critical to note that this recommendation of active surveillance as a management option is restricted to cT1a SRMs in the latter two guidelines (NCCN³ and AUA)⁴², i.e., SRMs with no adverse cT3 RFs. Therefore, as with surgical management of SRMs, the decision to proceed with active surveillance cannot be solely made based on tumor size alone, and requires a similar level of vigilance with regards to thorough examination of imaging for features suggestive of T3a stage, which would warrant an interventional rather than surveillance approach. However, given the variable and limited sensitivity of imaging modalities for pT3a features, which allow for discordance between radiologic and pathologic staging ³⁴, and the equivalence in oncologic outcomes for PN and RN in appropriate patients^43,44, the presence of adverse RFs may not necessitate a more aggressive surgical approach upfront. Post-operatively, our findings are in alignment with current guideline recommendations for closer surveillance, and potentially referral to medical oncology to discuss adjuvant therapy or clinical trials in patients with pT3 SRMs given their worse outcomes compared to pT1a^45–48.

CONCLUSIONS

Adverse T3a pathologic features portend worse cancer-specific and recurrence-free survival outcomes for patients with SRMs. Our findings emphasize the crucial role of pre-operative planning, case selection, and a multidisciplinary approach to post-surgical management for T3a masses, regardless of their size, given their poorer prognosis compared to lower stage masses.