EURURO Vol. 71 No. 4

Introduction

The tumor stage, size, grade, and necrosis (SSIGN) score was

reported in 2002 based on patients treated with radical

nephrectomy (RN) between 1970 and 1998 for clear cell

renal cell carcinoma (ccRCC)

[1] .

It was developed due to the

limited prognostic ability offered by the TNM staging

system to predict death from renal cell carcinoma (RCC)

following RN, specifically for ccRCC. As such, the SSIGN

score incorporates several pathologic features (tumor size,

grade, and presence of coagulative necrosis) beyond TNM

stage that are predictive of survival following nephrectomy

[2–4]

. Since its original description, the SSIGN score has

been externally validated

[5–8]

, compared favorably with

other prognostic models

[5]

, been included in guidelines

[9]

and is now being utilized to stratify patients for therapeutic

clinical trials and assess the role of biomarkers in predicting

survival for RCC patients

[10–15] .

However, the original description and subsequent valida-

tions may not reflect current clinical practice. Important

changes have influenced survival for RCC patients since the

initial publication of the SSIGN score including significant

stage migration

[16]

and the introduction of targeted agents

for metastatic RCC

[17] .

In addition, the utilization of partial

nephrectomy (PN) for localized RCC has dramatically

increased in the past decade

[18] ,

a procedure excluded

from the SSIGN score development. Finally, the original

description of the SSIGN score, and the subsequent valida-

tions, lacks long-term follow-up data and an assessment of

the competing risk of death from non-RCC causes.

Given the changes in the landscape of RCC management

since the inception of the SSIGN score and its ongoing

utilization in validating new biomarkers, we sought to

reassess the originally described cohort to evaluate the SSIGN

scorewith longer follow-up, investigate the competing risk of

non-RCC death, and evaluate the predictive ability of the

SSIGNscore in contemporary patientsmanagedwithbothRN

and PN.

Materials and methods

2.1.

Patient selection

Following institutional review board approval, we queried the Mayo

Clinic Nephrectomy Registry to identify the 1801 patients treated with

RN for sporadic unilateral ccRCC between 1970 and 1998 who were used

to develop the Mayo Clinic SSIGN score

[1] .

Of these patients, 6 declined

use of their medical records for research, leaving 1795 patients for

analysis. We also identified 1038 patients treated with RN and

767 patients treated with PN for sporadic unilateral ccRCC between

1999 and 2010 to serve as contemporary RN and PN cohorts.

2.2.

Clinicopathologic features

Clinical features assessed included age at surgery, sex, symptoms at

diagnosis, smoking history, preoperative estimated glomerular filtration

rate (eGFR; in ml/min per 1.73 m

), Eastern Cooperative Oncology Group

(ECOG) performance status, Charlson Comorbidity Index (CCI) score, and

body mass index (BMI; in kg/m

). Patients with a palpable flank or

abdominal mass, discomfort, gross hematuria, acute onset varicocele, or

constitutional symptoms (rash, sweats, weight loss, fatigue, early satiety,

and/or anorexia) were considered symptomatic at presentation.

All pathologic specimens were reviewed by one urologic pathologist

(J.C.C.) blinded to patient outcome for identification of histologic

subtype, tumor size, 2010 TNM classification

[19]

, 2016 World Health

Organization/International Society of Urological Pathology grade (iden-

tical to the nuclear grading system used for the Mayo Clinic

Nephrectomy Registry and the development of the SSIGN score

[1,20,21]

), coagulative necrosis, and sarcomatoid differentiation. The

original SSIGN score was developed using the 1997 TNM classification,

which has since been updated

[19]

; all patients have been restaged to

reflect the current system, and the SSIGN scores are reflective of this

change (Supplementary Table 1). For example, patients with level

0 tumor thrombi were originally classified as pT3b in the 1997 system

and are now classified as pT3a. Because 2 points are added to the SSIGN

score for all pT3 tumors, this did not result in a change in the calculated

SSIGN score. Similarly, 2 points are added to the SSIGN score for both

pN1 and pN2 tumors from the 1997 system, which are now both

classified as pN1 in the 2010 system

[1]

2.3.

Patient outcome

Vital status for patients in the Nephrectomy Registry is updated yearly,

with the most current follow-up utilized for analysis. For patients who

died within the previous year, the cause of death is determined by death

certificate review. If patients visited our institution for metastatic RCC

within 6 mo of death, they are considered to have died of RCC. If the

death certificate does not support this conclusion, the medical history is

reviewed by an attending urologist to determine cause of death, which

may include verification with the patient’s local physician.

2.4.

Statistical methods

Continuous features were summarized with medians and interquartile

ranges (IQRs); categorical features were summarized with frequencies

and percentages. Comparisons of features between patients in the

original and contemporary RN cohorts and between patients in the

contemporary RN and PN cohorts were evaluated using Wilcoxon rank

sum and chi-square tests. Cancer-specific survival (CSS) was estimated

using the Kaplan-Meier method, with duration of follow-up calculated

from the date of surgery to the date of death or last follow-up.

Associations of the SSIGN score with death from RCC were evaluated

using univariable Cox proportional hazards regression models and

summarized with hazard ratios (HRs) and 95% confidence intervals (CIs).

The predictive ability of the SSIGN score was summarized with a

bootstrap-corrected C-index. Additional multivariable models assessed

the effect of cohort assignment, M stage, and SSIGN score on outcome

within the RN cohorts. A separate analysis accounting for the competing

risk of non-RCC death was performed by calculating the adjusted

cumulative incidence of death from RCC

[22]

. Proportional subdistribu-

tion models were used to assess associations of the SSIGN score with the

adjusted cumulative incidence of death from RCC and were summarized

with HRs and 95% CIs. Statistical analyses were performed using

SAS v.9.3 (SAS Institute Inc., Cary, NC, USA) or R v.3.1.1 (R Foundation

for Statistical Computing, Vienna, Austria). All tests were two sided with

values

0.05 considered significant.

Results

3.1.

Cohort characterization

Table 1

summarizes the clinicopathologic features of the

3600 patients stratified by cohort. Comparing the original

E U R O P E A N U R O L O G Y 7 1 ( 2 0 1 7 ) 6 6 5 – 6 7 3

666