Sage Journals: Discover world-class research

Abstract

BACKGROUND:

Amplified in breast cancer 1 (AIB1) is a candidate oncogene in human breast cancer, which has been identified to be amplified and overexpressed in several types of other human cancers. Abnormalities of AIB1 and its clinical/prognostic significance, however, in upper tract urothelial carcinoma (UTUC) remain unclear.

OBJECTIVE:

To explore what role AIB1 plays in upper tract urothelial carcinoma.

METHODS:

The expression of AIB1 was analyzed using immunohistochemical staining in 133 UTUC patients. Overall, cancer specific and recurrence-free survival rates (OS, CSS, and RFS) were estimated using the Kaplan-Meier method. Multivariable COX regression models containing relevant clinicopathological variables addressed the prediction of postoperative outcome.

RESULTS:

High AIB1 expression was observed to be associated with increased hazard ratios for 5-year CSS (80.6% vs. 55.8%, $p=$ 0.008) and OS (78.1% vs. 54.8%, $p=$ 0.006). Multivariable analysis revealed that elevated AIB1 expression was an independent prognostic predictor of OS, CSS and RFS. Additionally, pT, pN and hydronephrosis were independently associated with oncologic outcome of UTUC. Three proposed nomograms were proposed to provide an individualized risk estimate of postoperative outcome in patients with UTUC.

CONCLUSIONS:

AIB1 can be used as an independent molecular marker for the prognosis of clinical outcomes of UTUC.

Keywords

AIB1 immunohistochemistry prognosis upper tract urothelial carcinoma

1. Introduction

Table 1
Relationship between AIB1 expression and clinicopathological parameters

	Total	Low AIB1 expression	High AIB1 expression	$p$ -value
$n$	133
Median age, year (range)	66 (36–86)	66 (45–86)	66.5 (36–84)	0.775 ${}^{\rm a}$
Mean $\pm$ SD	64.7 $\pm$ 10.3	64.4 $\pm$ 9.5	64.9 $\pm$ 11.1
Gender, $n$ (%)
Male	83 (62.4)	48 (67.6)	35 (56.5)	0.185 ${}^{\rm b}$
Female	50 (37.6)	23 (32.4)	27 (43.5)
Tumor site, $n$ (%)
Rt	65 (48.9)	34 (47.9)	31 (50)	0.808 ${}^{\rm b}$
Lt	68 (51.1)	37 (52.1)	31 (50)
Tumor location, $n$ (%)
Calix or pelvis	69 (51.9)	39 (54.9)	30 (48.4)	0.753 ${}^{\rm b}$
Ureter	50 (37.6)	25 (35.2)	25 (40.3)
Multiple	14 (10.5)	7 (9.9)	7 (11.3)
Hydronephrosis, $n$ (%)
No	41 (30.8)	27 (38.0)	14 (22.6)	0.054 ${}^{\rm b}$
Yes	92 (69.2)	44 (62.0)	48 (77.4)
Tumor stage, $n$ (%)
Ta–T1	50 (37.6)	27 (38.0)	23 (37.1)	0.109 ${}^{\rm b}$
T2	55 (41.4)	34 (47.9)	21 (33.9)
T3	16 (12.0)	7 (9.9)	9 (14.5)
T4	12 (9.0)	3 (4.2)	9 (14.5)
pN stage, $n$ (%)
Negative	118 (88.7)	66 (93.0)	52 (83.9)	0.098 ${}^{\rm b}$
Positive	15 (11.3)	5 (7.0)	10 (16.1)
Tumor grade, $n$ (%)
Low	49 (36.8)	26 (36.6)	23 (37.1)	0.955 ${}^{\rm b}$
High	84 (63.2)	45 (63.4)	39 (62.9)

SD: standard deviation. ${}^{\rm a}$ $P$ value was determined using the $t$ -tests. ${}^{\rm b}$ $P$ values were determined using the $\chi^{2}$ test.

Urothelial carcinomas are derived from the urothelium and have been classified into two different types, bladder cancer and upper tract urothelial carcinoma (UTUC), according to their tumor location. UTUC is a heterogeneous disease that constitutes approximately 5% of urothelial carcinomas [1]. UTUC incidence has increased in the last 30 years and has been shown to be prevalent in patients who are approximately 70 years old [2]. UTUC has a tendency to relapse and is associated with a generally poor clinical prognosis. Tumors with invasive stage, nodular involvement, metastasis, and residual tumor following surgery are associated with poor clinical outcomes [3]. Therefore, a substantial amount of research regarding UTUC has focused on the discovery of sensitive molecular markers present in UTUC cells that could function as reliable prognostic factors. Several studies have investigated the prognostic value of factors involved in the cell cycle, apoptotic pathways, cell adhesion, metastasis, and microsatellite instability (MSI) [4]. However, the reported specific molecular and/or genetic alterations in UTUC that have prognostic significance remain limited.

The amplified in breast cancer 1 (AIB1) gene, also referred to as SRC-3, p/CIP, RAC3, ACTR, and TRAM-1, is isolated from chromosome 20q and has been identified as a human oncogene [5]. AIB1 has been demonstrated to be frequently amplified and overexpressed in human breast [5], prostate [6], ovarian [7], esophageal [8], gastric [9], and hepatocellular cancers [10]. AIB1 amplification and expression in bladder cancer has been previously studied by our research group, and we have demonstrated that upregulation of AIB1 is associated with poor patient prognosis in patients with bladder cancer [11]. However, significant differences were identified regarding tumor location and behavior between the upper and lower urinary tract [4]. Thus, further research is required regarding the molecular state of the AIB1 gene in UTUC and the prognostic significance of this state. In this study, we examined AIB1 expression in UTUC using immunohistochemistry (IHC), and studied its correlation with clinicopathological features and prognosis. The primary goal of this study was to study whether AIB1 activity possesses prognostic value in UTUC patients.

2. Materials and methods

2.1 Study design and patient selection

Surgical specimens from 133 patients who underwent operation for UTUC in our hospital from 2000 to 2016 were used. Of the 133 UTUC patients selected, a total of 83 had invasive UTUC (pT2-4) and the remaining 50 UTUC patients had superficial UTUC (pTa-1). We identified 83 male and 50 female patients in this study. The median patient age was 66 years (mean 64.7 $\pm$ 10.3, range 36 to 86). Table 1 lists the patient characteristics. None of the UTUC patients had received preoperative chemotherapy or radiotherapy. Approval of this study was obtained from the Institutional Review Board of our hospital, and the requirement for informed consent was waived owing to the retrospective nature of this study.

All surgical specimens were processed according to standard pathological procedures and all slides were retrieved and re-reviewed by genitourinary pathologists to confirm the original diagnosis. Tumors were staged according to the American Joint Committee on Cancer-Union Internationale Contre le Cancer TNM classification. Tumor grading was assessed according to the 1998 WHO/International Society of Urologic Pathology consensus classification. The representative paraffin block was selected to freshly cut into 4- $\mu$ m-thick sections for IHC analysis. The regular follow-up included physical examination, routine blood and serum chemistry studies, urinary cytology, cystoscopy, and computed tomography scan of the chest and abdomen. Cystoscopy was suggested every 3 months for 2 years and every 6 months thereafter. Clinical data were obtained by review of all pertinent medical records and telephone interviews. Recurrent disease was defined as the initial recurrence at the surgical sites, lymph node metastasis, contralateral metastasis, and/or distant metastasis. Survival time was identified as the interval from the date of operation to final follow-up or death.

2.2 IHC

IHC study was performed using a standard strep- tavidin-biotin-peroxidase complex method [8]. Briefly, 4- $\mu$ m-thick sections were deparaffinized and rehydrated. Endogenous peroxidase activity was blocked with 0.3% hydrogen peroxide for 20 min. For antigen retrieval, tissue sections were microwave-treated and boiled in a 10 mmol/L citrate buffer (pH 6.0) for 8 min. Nonspecific binding was blocked with 10% normal rabbit serum for 20 min. The slides were then incubated with anti-AIB1 antibody at 4 ${}^{\circ}$ C overnight (BD Transduction Laboratories; Lexington, KY, USA; diluted 1: 100 in PBS). All incubations were carried out in a moist chamber. Subsequently, the slides were incubated with biotinylated rabbit anti-mouse immunoglobulin (1: 100 dilution) at 37 ${{}^{\circ}}$ C for 30 min. Slides were then incubated with a streptavidin-peroxidase conjugate for at 37 ${{}^{\circ}}$ C 30 min using 3’-3’ diaminobenzidine as a chromogen substrate. Nuclei were counterstained using Meyer’s hematoxylin. A negative control was obtained by replacing the primary antibody with normal murine IgG. Known immunostaining positive breast cancer slides were used as positive controls.

Positive AIB1 expression in UTUC exhibited a primarily nuclear pattern. AIB1 expression levels were assessed using a semiquantitative scoring system and the AIB1 score was calculated using the following formula: staining intensity $\times$ positive cell rate. Staining intensity was stratified from 0 to 3 (0, no staining; 1, slight staining; 2, medium staining; 3, strong staining). The scale of positive cell rate was identified as follows: 0, number of stained cells $<$ 10%; 1, 10 $\leqslant$ number of stained cells $<$ 25%; 2, 25% $\leqslant$ number of stained cells $<$ 50%; 3, 50% $\leqslant$ number of stained cells $<$ 80%; 4, number of stained cells $\geqslant$ 80%. We selected the median (score 4) as the cut-off value (score range: 0–12). An AIB1 expression score greater than 4 was considered high AIB1 expression.

2.3 Statistical analysis

All analyses were performed using SPSS version 20.0 (IBM Corp., Armonk, NY, USA), GraphPad Prism 5 (GraphPad Software, Inc., La Jolla, CA, USA), and R for Windows, version 3.3.1 (R Foundation for Statistical Computing, Vienna, Austria). The association between AIB1 protein expression and UTUC patient clinicopathological features was assessed by the $\chi^{2}$ and $t$ -tests. The Kaplan-Meier method was applied in order to calculate survival functions, and the log-rank test was used to compare different survival curves. Univariate analysis with the log rank test and multivariate analysis with Cox hazards regression were applied to evaluate the value of prognostic factors including age, gender, hydronephrosis, tumor site, tumor location, tumor stage, node, tumor grade, and AIB1 expression in predicting overall, cancer specific and recurrence-free survival. Schoenfeld residuals test was used to check proportional hazard assumption. In case of departure from the proportional hazards assumption, time-dependent covariates were considered. We investigated the predictive accuracy of AIB1 and patient clinicopathological features using receiver operating characteristic (ROC) analysis, and the area under the curve (AUC) was used to measure predictive accuracy. To predict overall, cancer specific and recurrence-free survival probability, multivariable Cox regression coefficients were then used to generate three nomograms and calibration curves. For validation of these nomograms, we assessed their discrimination abilities using Harrell’s concordance index (c-index). All $p$ values were two-sided and $p$ values $<$ 0.05 were considered to be significant.

Figure 1.

Immunohistochemical stainings of AIB1 in upper urinary tract urothelial carcinoma. (a) AIB1 score 0 points (negative staining), (b) Low expression of AIB1 (score 3 points), (c) High expression of AIB1 (score 9 points).

Figure 2.

Kaplan-Meier analysis of overall survival, cancer-specific survival and recurrence-free survival of patients following surgery for UTUC according to AIB1 expression (a-c) and according to the prognostic models with AIB1 expression, pT, pN, and hydronephrosis status that are stratified (g-i); ROC curves comparing the predictive accuracy by the combined AIB1, pT, pN, and hydronephrosis status model, the AIB1 alone model, the pT model, and the pN alone model for overall survival (d), cancer-specific (e) and recurrence-free survival (f).

3. Results

3.1 Patient characteristics

All specimens were fixed in 10% formalin and embedded in paraffin. IHC was performed on all slides using anti-AIB1. According to the cut-off value (score 4), patients were divided into low AIB1 ( $n=$ 71) and high AIB1 expression groups ( $n=$ 62) in order to analyze their oncologic outcomes. Representative images for AIB1 scores of 0, 3, and 9 are shown in Fig. 1a–c, respectively. Positive staining of AIB1 protein was mainly located in the nucleus of tumor cells in UTUC tissues. The relationship between AIB1 expression and tumor clinicopathological features in UTUCs was then further analyzed. Results demonstrated that no significant association exists between AIB1 expression and the clinicopathological features of the UTUC cohorts ( $p>$ 0.05, Table 1).

Table 2
Univariate and multivariate Cox regression analysis of potential prognostic factors for prediction of overall, cancer specific and recurrence-free survival in patients treated radical nephroureterectomy for upper tract urothelial carcinoma

	Overall survival			Cancer specific survival			Recurrence-free survival
	Deaths per	Univariate	Multivariate	Cancer specific	Univariate	Multivariate	Recurrent	Univariate	Multivariate
	1000 person	HR (95%CI)	HR (95%CI)	deaths per 1000	HR (95%CI)	HR (95%CI)	events per 1000	HR (95%CI)	HR (95%CI)
	years			person years			person years
Age, year
$>$ 66 vs. $\leqslant$ 66	54.5 vs. 83.2	0.98 (0.54–1.78)	0.97 (0.50–1.90)	71.9 vs. 69.3	0.97 (0.51–1.86)	0.95 (0.46–1.99)	140.7 vs. 97.7	1.17 (0.68–2.03)	1.34 (0.73–2.44)
Gender
Female vs. male	78.8 vs. 86.6	0.91 (0.49–1.71)	0.61 (0.30–1.24)	57.8 vs. 77.7	0.75 (0.37–1.52)	0.41 (0.18–0.93)	80.2 vs. 139.1	0.57 (0.31–1.05)	0.29 (0.14–0.61)
Hydronephrosis
Yes vs. No	101.9 vs. 43.2	2.30 (1.02–5.18)	2.53 (0.94–6.85)	85.3 vs. 37.1	2.32 (0.97–5.59)	2.61 (0.87–7.84)	146.4 vs. 54.4	2.75 (1.29–5.86)	3.32 (1.28–8.61)
Tumor site
Rt vs. Lt	75.6 vs. 91.2	0.82 (0.45–1.49)	0.68 (0.36–1.31)	67.7 vs. 73.0	0.90 (0.47–1.72)	0.67 (0.33–1.38)	107.0 vs. 124.0	0.84 (0.48–1.45)	0.64 (0.35–1.18)
Tumor location
Ureter vs pelvis	93.3 vs. 73.8	1.23 (0.65–2.34)	0.97 (0.46–2.02)	72.5 vs. 62.7	1.13 (0.55–2.30)	0.93 (0.41–2.11)	146.7 vs. 92.3	1.58 (0.87–2.86)	1.05 (0.53–2.06)
Multiple vs pelvis	98.0 vs. 73.8	1.36 (0.54–3.42)	0.76 (0.28–2.09)	98.0 vs. 62.7	1.67 (0.65–4.29)	0.87 (0.30–2.47)	124.1 vs. 92.3	1.78 (0.76–4.20)	0.74 (0.28–1.97)
pT status
T2–4 vs ta-1	98.2 vs. 63.7	1.53 (0.82–2.949)	1.58 (0.65–3.82)	94.9 vs. 36.4	2.64 (1.20–5.80)	2.85 (0.98–8.26)	154.6 vs. 66.7	2.06 (1.10–3.87)	2.75 (1.19–6.36)
pN status
N $+$ vs N0	340.2 vs. 67.0	4.97 (2.44–10.10)	4.80 (2.10–10.97)	340.2 vs. 52.8	5.37 (2.62–11.02)	4.56 (1.96–10.59)	623.4 vs. 92.5	4.09 (2.12–7.89)	4.50 (2.06–9.83)
Tumor grade
High vs low	101.5 vs. 57.1	1.74 (0.89–3.39)	1.21 (0.52–2.83)	92.0 vs. 38.1	2.37 (1.08–5.20)	1.24 (0.45–3.42)	144.6 vs. 75.7	1.82 (0.99–3.38)	1.11 (0.50–2.45)
AIB1 expression
High vs low	128.9 vs. 53.8	2.30 (1.24–4.25)	2.47 (1.26–4.86)	109.8 vs. 44.3	2.40 (1.22–4.70)	3.06 (1.42–6.59)	162.9 vs. 85.6	1.74 (1.00–3.03)	1.97 (1.06–3.66)

HR: hazard ratio; 95%CI: 95% confidence interval.

3.2 Survival and recurrence

In the low AIB1 expression group, median survival was 46 months (range 6 to 140), whereas in the high AIB1 expression group the median survival was determined to be 25 months (range 1 to 163). Median recurrence-free survival in the low and high AIB1 expression groups was observed to be 41 months (range 2 to 140) and 22 months (range 1 to 163), respectively. Statistically significant differences were observed between the low and high AIB1 expression groups in the 5-year cancer specific survival rate (80.6% vs. 55.8%, $p=$ 0.008) and the overall survival rate (78.1% vs. 54.8%, $p=$ 0.006). While there was no statistically significant difference observed in the 5-year recurrence-free survival rate (67.8% vs. 48.4%, $p=$ 0.051), there did exist a trend of better recurrence-free survival in patients who were in the low AIB1 expression group (Fig. 2a–c).

3.3 Independent prognostic factors of UTUC: Multivariate cox regression analysis

The proportional hazards assumption was accepted for OS ( $p=$ 0.087), CSS ( $p=$ 0.442) and RFS ( $p=$ 0.578), which allowed us to analyze these factors using Cox hazards regression analysis. According to univariate and multivariate analyses, AIB1 expression was found to be an independent prognostic factor for poor overall patient survival (HR: 2.47, 95% confidence interval (CI): 1.26–4.86, $p=$ 0.009), cancer specific survival (HR: 3.06, 95% CI: 1.42–6.59, $p=$ 0.004), and recurrence-free survival (HR: 1.97, 95% CI: 1.06–3.66, $p=$ 0.032), respectively. In regard to other parameters, the pN status and hydronephrosis were also found to be an independently unfavorable prognostic factor for overall, cancer specific and recurrence-free survival. Tumor stage and gender were found to be independent predictors for cancer specific survival and tumor recurrence (Table 2).

Table 3
The areas under ROC curves (AUC) of the combined AIB1, pT, pN and hydronephrosis status model, the AIB1 alone model, the pT model, and the pN alone model for overall survival, cancer-specific and recurrence-free survival

		AUC	95%CI	$p$ value ${}^{\rm a}$
Overall	AIB1	0.638	0.540–0.737	0.010
survival	pT	0.599	0.491–0.706	0.065
	pN	0.603	0.495–0.710	0.055
	Hydronephrosis	0.611	0.513–0.710	0.037
	Combination	0.733	0.644–0.823	$<$ 0.001
Cancer	AIB1	0.629	0.524–0.733	0.022
specific	pT	0.676	0.569–0.782	0.002
survival	pN	0.628	0.513–0.742	0.023
	Hydronephrosis	0.601	0.498–0.704	0.071
	Combination	0.755	0.660–0.850	$<$ 0.001
Recurrence	AIB1	0.615	0.517–0.713	0.026
-free	pT	0.657	0.560–0.755	0.002
survival	pN	0.599	0.497–0.702	0.055
	Hydronephrosis	0.623	0.527–0.718	0.017
	Combination	0.748	0.660–0.836	$<$ 0.001

${}^{\rm a}$ $P$ values are calculated using ROC curve analysis.

Figure 3.

Nomograms to predict overall survival probability (a), cancer-specific survival probability (b), and recurrence-free survival probability (c) at 2, and 5 years for UTUC patients. OS: Overall survival, CSS: Cancer-specific survival, RFS: Recurrence-free survival.

Figure 4.

Calibration plots demonstrate virtually ideal predictions for 2, and 5-year overall survival (a), cancer-specific survival (b), and recurrence-free survival (c), respectively.

3.4 ROC curve analysis

We applied the ROC curve analysis in order to explore the predictive value of AIB1 expression in regards to overall survival, recurrence-free survival, and cancer specific survival. A prognostic model was generated to combine AIB1 with three independent prognostic factors (tumor stage, pN status and hydronephrosis) using Cox proportional hazards regression (the model for overall survival $=$ 0.141 $\times$ AIB1 score $+$ 0.363 $\times$ pT stage $+$ 1.151 $\times$ pN status $+$ 0.803 $\times$ hydronephrosis status; for cancer specific survival $=$ 0.127 $\times$ AIB1 score $+$ 0.531 $\times$ pT stage $+$ 1.022 $\times$ pN status $+$ 0.798 $\times$ hydronephrosis status; for recurrence-free survival $=$ 0.104 $\times$ AIB1 score $+$ 0.480 $\times$ pT stage $+$ 0.968 $\times$ pN status $+$ 1.099 $\times$ hydronephrosis status). These results demonstrated a promising predictive value of AIB1 expression in regards to overall, cancer specific, and recurrence-free survival: AUC 0.638, 95% CI (0.540–0.737); AUC 0.629, 95% CI (0.524–0.733); AUC 0.615, 95% CI (0.517–0.713), respectively. The predictive value of AIB1 expression was prior to or approximate to that of the three clinicopathological factors. ROC curve was applied to compare the prognostic validity of the combined model with each factor alone. As depicted in Fig. 2d–f, the model combining AIB1, tumor stage, pN status, and hydronephrosis possessed a better prognostic value than that of the model with a single factor of these prognostic ones in overall survival (AUC 0.733, 95% CI: 0.644–0.823), cancer specific survival (AUC 0.755, 95% CI: 0.660–0.850), or recurrence-free survival (0.748, 95% CI: 0.660–0.836) (Table 3). Therefore, AIB1 expression could add prognostic value to the clinicopathological factors of UTUC.

3.5 Survival and recurrence analysis for risk stratification

Based on the median scores of the three prognostic models described above, patients were stratified into low- and high-risk groups. Results from these studies demonstrated that patients in low-risk group exhibited a much better overall survival rate, 5-year cancer specific, and recurrence-free survival rate compared with those patients in the high-risk group (81.7% vs. 52.7%; 84.2% vs. 43.9%; 76.1% vs. 40.2%, respectively, Fig. 2g–i). These results demonstrate that these three prognostic models with AIB1 could practically identify patients with UTUC who had poor clinical outcome and needed further adjuvant chemotherapy.

3.6 Construction of the prognostic nomograms

We developed three competing-risk nomograms to predict the probability of overall survival, cancer specific survival, and intravesical recurrence at 2, and 5 years. The models were based on AIB1, hydronephrosis, tumor stage, and pN status. Calibration plots demonstrated that the nomograms compared well with an ideal model (Figs 3 and 4), and the c-index of the nomograms for the predictions of overall survival, cancer specific survival, and recurrence-free survival was 0.723, 0.747, and 0.716, respectively. According to these nomograms, a patient with a T2 UTUC, high AIB1 expression, no hydronephrosis, and no lymphatic metastasis received a total of 8.9 points, corresponding to a greater than 85% overall survival rate for 2 years following surgery. A patient with T3 UTUC, high AIB1 expression, hydronephrosis, and no lymphatic metastasis accumulated 15.9 points with a 63% of 2-year cancer specific survival rate, and about 38% of 5-year cancer specific survival rate. This same patient also accumulated 16.4 points with a 60% risk of tumor recurrence in 2 years following surgery, and a greater than 70% risk for 5-year recurrence.

4. Discussion

At present, current pTNM staging and histopathological grading systems are established for prognosis prediction in patients with UTUC [12]. However, these parameters may have reached their limits in providing critical information that influences patient prognosis and treatment strategies. Patients who stay at the same clinical stage and/or histological grade often present considerable variability in terms of disease recurrence and survival [13, 14]. Therefore, there exists an urgent need for the development of novel variables that can function to distinguish between patients with an unfavorable prognosis and other patients with a better prognosis. As is known, patients’ survival might be prolonged by more effective adjuvant therapy, if individuals with different prognosis could be identified by molecular biomarkers at the time of surgery [15, 16]. Therefore, further studies are required for the development of appropriate UTUC biomarkers.

The AIB1 gene is located at 20q12 and has been reported to promote proliferation in hormone-dependent cancers and some hormone-independent cancers [5, 8, 9, 17]. Though our previous studies have observed that in bladder cancer, AIB1 overexpression contributes to a shortened survival by promoting bladder cancer cell proliferation through the AKT pathway and acting as coactivator of E2F1 in non-muscle-invasive bladder cancer [11, 18]. However, recent studies have reported different genetic presentations between upper and lower urinary tract urothelial carcinoma [19], and the role of AIB1 in UTUC remains unclear. In order to study the potential oncogenic effects of AIB1 in UTUC, we initially studied AIB1 expression using IHC in a cohort of UTUC tissue samples. The key findings in this study were that AIB1 was much more frequently overexpressed in UTUC and was strongly linked with overall survival, cancer specific survival, and recurrence free survival in patients with UTUC. These findings present a potentially critical role of AIB1 as an underlying biological mechanism related to UTUC recurrence and/or progression. Therefore, AIB1 expression analysis can be applied in order to optimize individual UTUC therapy management.

UTUC in our series presented aggressive behavior, as 63.2% of our patients exhibited a pejorative clinical outcome. Patients with invasive tumors typically have a poor survival rate and have a distinct tendency to relapse, with work suggesting that the tumor stage is a primary prognostic factor [20]. In a study carried out by Ng et al. [21] preoperative clinical features were controlled and it was confirmed that hydronephrosis was independently associated with cancer metastasis and cancer specific survival by preoperative multivariable analysis. Chung et al. [22] graded hydronephrosis as none/mild vs. moderate/severe, and demonstrated degree of hydronephrosis in patients with advanced UTUC was associated with local/systemic recurrence-free survival (HR $=$ 2.5; 95% CI: 1.07–5.64; $p=$ 0.04). We also found an association between tumor recurrence and hydronephrosis ( $p=$ 0.037). In the current study, the combination of AIB1 and these significant clinicopathological factors was carried out in an effort to improve the predictive sensitivity and specificity for the diagnosis of recurrence and the prognosis of cancer specific survival. Furthermore, survival analysis based on the risk subset of the predictive combination model showed a more significant difference than the one based on AIB1 expression alone. Although some studies have demonstrated that clinicopathologic factors, including pathologic stage and grade, were recognized to have a poor reproducibility, the combination of molecular biomarkers and clinicopathological factors was considered to provide a more stable way to predict postoperative outcome [15, 23].

The accurate estimation of treatment success, recurrence, and long-term morbidity is essential for both physicians and their patients to make informed medical decisions regarding the management of their disease [14]. To obtain the most accurate and reliable predictions of outcomes, we designed nomograms with AIB1 to predict overall survival, distant or metastatic recurrence, and cancer-specific mortality following surgery, respectively. We observed that our nomograms were well calibrated, as the slopes of the calibration curves were close to 1. With our accurate, well-calibrated predictive models, it is helpful to improve clinicians’ abilities to become familiar with the strengths and weaknesses of the predicted probabilities according to their distribution. These results enrich the present findings from the current literatures [14, 24].

The main limitation of our study centers on the retrospective analysis of data which was collected from a single center. Therefore, the number of cases studied is small. External validation of the current nomograms is required before widespread application. However, we still believe that our postoperative nomograms of tumor recurrence and cancer specific survival are useful for the consideration of the management of postoperative chemotherapy and can provide meaningful counseling for the patients. In addition, lymphovascular invasion, surgical margins and tumor necrosis have been proved significant postoperative factors, [25, 26, 27] and it might be more effective to add these factors to the nomograms. However, in this retrospective study, these pathological factors were not available for all patients.

5. Conclusion

In the current study, we have identified certain factors that predict poor clinical outcome for UTUC. High AIB1 expression was found to be an important predictor for shortened cancer specific and recurrence-free survival in patients with UTUC. We identified a mixed panel of four markers (AIB1, hydronephrosis, tumor stage, and pN status) that proved to be a powerful predictor of UTUC recurrence. Further prediction models can be used for patient counseling and selecting postoperative systemic therapy.

Footnotes

Acknowledgments

This study was funded by National Key Research and Development Program of China (Grant/Award Number: 2016YFC0902600), Guangdong Provincial Science and Technology Foundation (Grant/Award Number: 2014B020212015, 2017B020227004, 2017A 030313538), National Natural Science Foundation of China (Grant/Award Number: 81372357, 81725016, 81572905, 81602219), Guangdong Medical Science and Technology Research Foundation (Grant/Award Number: A2018040), and the Guangzhou Science and Technology Foundation (201607010238, 2017040 20174).

Conflict of interest

The authors declare that they have no conflict of interest.

References

Raman

J.D.

et al., Incidence and survival of patients with carcinoma of the ureter and renal pelvis in the USA, 1973–2005, BJU Int 107 (2011), 1059–1064.

Munoz

J.J.

and Ellison

L.M.

, Upper tract urothelial neoplasms: Incidence and survival during the last 2 decades, J Urol 164 (2000), 1523–1525.

Balaji

K.C.

et al., Upper tract recurrences following radical cystectomy: An analysis of prognostic factors, recurrence pattern and stage at presentation, J Urol 162 (1999), 1603–1606.

Lughezzani

et al., Prognostic factors in upper urinary tract urothelial carcinomas: A comprehensive review of the current literature, Eur Urol 62 (2012), 100–114.

Anzick

S.L.

et al., AIB1, a steroid receptor coactivator amplified in breast and ovarian cancer, Science 277 (1997), 965–968.

Gnanapragasam

V.J.

et al., Expression of RAC 3, a steroid hormone receptor co-activator in prostate cancer, Br J Cancer 85 (2001), 1928–1936.

Tanner

M.M.

et al., Frequent amplification of chromosomal region 20q12-q13 in ovarian cancer, Clin Cancer Res 6 (2000), 1833–1839.

F.P.

et al., SRC-3/AIB1 protein and gene amplification levels in human esophageal squamous cell carcinomas, Cancer Lett 245 (2007), 69–74.

Sakakura

et al., Amplification and over-expression of the AIB1 nuclear receptor co-activator gene in primary gastric cancers, Int J Cancer 89 (2000), 217–223.

10.

Wang

et al., Prognostic significance of c-myc and AIB1 amplification in hepatocellular carcinoma, A broad survey using high-throughput tissue microarray, Cancer-Am Cancer Soc 95 (2002), 2346–2352.

11.

Luo

J.H.

et al., Protein expression and amplification of AIB1 in human urothelial carcinoma of the bladder and overexpression of AIB1 is a new independent prognostic marker of patient survival, Int J Cancer 122 (2008), 2554–2561.

12.

Cha

E.K.

et al., Predicting clinical outcomes after radical nephroureterectomy for upper tract urothelial carcinoma, Eur Urol 61 (2012), 818–825.

13.

Margulis

et al., Preoperative multivariable prognostic model for prediction of nonorgan confined urothelial carcinoma of the upper urinary tract, J Urol 184 (2010), 453–458.

14.

Xylinas

et al., Prediction of intravesical recurrence after radical nephroureterectomy: Development of a clinical decision-making tool, Eur Urol 65 (2014), 650–658.

15.

van Rhijn

B.W.

et al., Molecular grading of urothelial cell carcinoma with fibroblast growth factor receptor 3 and MIB-1 is superior to pathologic grade for the prediction of clinical outcome, J Clin Oncol 21 (2003), 1912–1921.

16.

Kosaka

et al., Expression of snail in upper urinary tract urothelial carcinoma: Prognostic significance and implications for tumor invasion, Clin Cancer Res 16 (2010), 5814–5823.

17.

Xie

et al., Correlation of AIB1 overexpression with advanced clinical stage of human colorectal carcinoma, Hum Pathol 36 (2005), 777–783.

18.

Tong

Z.T.

et al., AIB1 predicts bladder cancer outcome and promotes bladder cancer cell proliferation through AKT and E2F1, Br J Cancer 108 (2013), 1470–1479.

19.

Yates

D.R.

et al., Genetic polymorphisms on 8q24.1 and 4p16.3 are not linked with urothelial carcinoma of the bladder in contrast to their association with aggressive upper urinary tract tumours, World J Urol 31 (2013), 53–59.

20.

Hall

M.C.

al.,

, Urology 52 (1998), 594–601.

21.

C.K.

et al., Does the presence of hydronephrosis on preoperative axial CT imaging predict worse outcomes for patients undergoing nephroureterectomy for upper-tract urothelial carcinoma? Urol Oncol 29 (2011), 27–32.

22.

Chung

P.H.

et al., Degree of hydronephrosis predicts adverse pathological features and worse oncologic outcomes in patients with high-grade urothelial carcinoma of the upper urinary tract, Urol Oncol 32 (2014), 981–988.

23.

Tosoni

et al., Clinical significance of interobserver differences in the staging and grading of superficial bladder cancer, BJU Int 85 (2000), 48–53.

24.

Xiong

et al., Prognostic and predictive value of epigenetic biomarkers and clinical factors in upper tract urothelial carcinoma, Epigenomics-Uk 7 (2015), 733–744.

25.

Novara

et al., Prognostic role of lymphovascular invasion in patients with urothelial carcinoma of the upper urinary tract: An international validation study, Eur Urol 57 (2010), 1064–1071.

26.

Colin

et al., Influence of positive surgical margin status after radical nephroureterectomy on upper urinary tract urothelial carcinoma survival, Ann Surg Oncol 19 (2012), 3613–3620.

27.

Zigeuner

et al., Tumour necrosis is an indicator of aggressive biology in patients with urothelial carcinoma of the upper urinary tract, Eur Urol 57 (2010), 575–581.

Impact of AIB1 expression on the prognosis of upper tract urothelial carcinoma after radical nephroureterectomy