Oncology Evaluation of the Prognostic Significance of Altered Mammalian Target of Rapamycin Pathway Biomarkers in Upper Tract Urothelial Carcinoma Aditya Bagrodia, Laura-Maria Krabbe, Bishoy A. Gayed, Payal Kapur, Ira Bernstein, Xian-Jin Xie, Christopher G. Wood, Jose A. Karam, Alon Z. Weizer, Jay D. Raman, Mesut Remzi, Nathalie Rioux-Leclerq, Andrea Haitel, Marco Roscigno, Christian Bolenz, Karim Bensalah, Arthur I. Sagalowsky, Shahrokh F. Shariat, Yair Lotan, and Vitaly Margulis OBJECTIVE MATERIALS AND METHODS
RESULTS
CONCLUSION
To evaluate the prognostic value of altered mammalian target of rapamycin (mTOR) pathway biomarkers in upper tract urothelial carcinoma (UTUC). We performed a multi-institutional review of clinical and pathologic information on patients receiving extirpative surgery for UTUC from 1990 to 2008. Immunohistochemistry for phosphorylated-S6, mTOR, phosphorylated-mTOR, PI3K, phosphorylated-4EBP1, phosphorylated-AKT, PTEN, HIF-1a, raptor, and cyclin D was performed on tissue microarrays from radical nephroureterectomy (RNU) specimens. Prognostic markers were identified and the significance of altered markers was assessed with the Kaplan-Meier analysis and the Cox regression analysis. Six hundred twenty patients were included. Over a median follow-up of 27.3 months, 24.6% of patients recurred and 21.8% died of UTUC. On multivariate analysis, PI3K (odds ratio, 1.28; P ¼ .001) and cyclin D (odds ratio, 3.45; P ¼ .05) were significant predictors of clinical outcomes. Cumulative marker score was defined as low risk (no altered markers or 1 altered marker) or high risk (cyclin D and PI3K altered). Patients with high-risk marker score had a significantly higher proportion of high-grade disease (91% vs 71%; P 90%) but dramatically declined for patients with locally advanced or lymph node (LN)e positive disease (5%-10%).4
Aditya Bagrodia and Laura-Maria Krabbe contributed equally to this work. Financial Disclosure: The authors declare that they have no relevant financial interests. From the Department of Urology, University of Texas Southwestern Medical Center, Dallas, TX; the Department of Urology, University of Muenster Medical Center, Muenster, Germany; the Department of Pathology, University of Texas Southwestern Medical Center, Dallas, TX; Division of Biostatistics, Department of Clinical Sciences, University of Texas Southwestern Medical Center, Dallas, TX; the Department of Urology, MD Anderson Cancer Center, Houston, TX; the Department of Urology, University of Michigan Cancer Center, Ann Arbor, MI; the Department of Urology, Penn State Hershey Medical Center, Hershey, PA; the Department of Urology, Medical
University of Vienna, Vienna, Austria; the Department of Pathology, Centre Hospitalier Universitaire de Rennes, Rennes, France; the Department of Pathology, Medical University Vienna, Vienna, Austria; the Department of Urology, Ospedali Riuniti of Bergamo, Bergamo, Italy; the Department of Urology, Mannheim Medical Center, University of Heidelberg, Mannheim, Germany; and the Department of Urology, Centre Hospitalier Universitaire de Rennes, Rennes, France Address correspondence to: Vitaly Margulis, M.D., Department of Urology, University of Texas Southwestern Medical Center, 5323 Harry Hines Boulevard J8.130, Dallas, TX 75390-9110. E-mail: [email protected] Submitted: March 9, 2014, accepted (with revisions): July 29, 2014
1134
ª 2014 Elsevier Inc. All Rights Reserved
http://dx.doi.org/10.1016/j.urology.2014.07.050 0090-4295/14
Improved clinical outcomes may be expected with an earlier detection, enhanced risk stratification, and incorporation of effective systemic therapies. However, screening for UTUC is precluded by the low incidence; prognostic models currently have only approximately 75% accuracy for predicting clinicopathologic outcomes; and neoadjuvant and adjuvant chemotherapies are administered to 1300 patients in the original data set, pathologic specimens for tissue microarray (TMA) construction were obtained for 620 patients from RNU or PU specimens, and this subgroup comprised our study cohort. Surgical technique and indication for RNU vs PU was based on surgeon preference. The extent and completion of lymphadenectomy were based on surgeon discretion. In general, patients were seen at least every 3-4 months in the first postoperative year, semiannually for the second year, and then annually. Comprehensive clinical and pathologic data were maintained in an institutional review boardeapproved database. Multiple data reviews and quality assurance checks were performed throughout the data collection process to ensure accuracy and completeness of data points. All surgical specimens were processed according to the standard pathologic procedures. Pathologic variables were prospectively rereviewed by a genitourinary pathologist (P.K.), who was blinded to clinical outcomes, using defined uniform criteria. UROLOGY 84 (5), 2014
STATISTICAL ANALYSIS
1135
Table 1. Descriptive characteristics of 620 patients treated with radical nephroureterectomy and distal ureterectomy Variable Age, y Median (range), y Women Men Grade Low grade High grade Location Kidney Ureter Ureteroenteric anastomosis Pathologic stage Ta/CIS T1 T2 T3 T4 N stage Node/x Nodeþ Organ-confined disease Yes (T stage 2; N ¼ 0/x) No (T stage >2 and/or Nþ) Lymphovascular invasion No Yes Architecture Papillary Sessile Necrosis No Yes Concomitant CIS No Yes Unifocal Multifocal Perioperative chemotherapy No Yes Normal PI3K Overexpressed PI3K Normal cyclin D Overexpressed cyclin D Cumulative marker score Low risk (0-1 markers altered) High risk (2 markers altered) Recurrence of UTUC (only M0) No Yes Death due to UTUC No Yes
No. (%; Total Cohort, n ¼ 620) 68.8 69.5 (27-104) 271 (43.7) 349 (56.3) 163 (26.3) 457 (73.7) 450 (72.6) 158 (25.5) 12 (1.9) 103 185 108 187 37
(16.6) (29.8) (17.4) (30.2) (6.0)
569 (91.8) 51 (8.2) 388 (62.6) 232 (37.4) 483 (77.9) 137 (22.1) 528 (85.2) 92 (14.8) 532 (85.8) 88 (14.2) 509 111 495 125
(82.1) (17.9) (79.8) (20.2)
548 72 374 246 342 278
(88.4) (11.6) (60.3) (39.7) (55.2) (44.8)
528 (85.2) 92 (14.8) 466 (75.2) 154 (24.8) 486 (78.4) 134 (21.6)
CIS, carcinoma in situ; PI3K, phosphatidyl inositol 3-kinase; UTUC, upper tract urothelial carcinoma.
parameters. Differences in variables with a continuous distribution across dichotomous or ranked categories were assessed using the Mann-Whitney U test or the KruskalWallis nonparametric analysis of variance test, respectively. Significance of altered markers was assessed with 1136
the Kaplan-Meier analysis and the Cox regression analysis. Univariate and multivariate Cox regression models with known predictors of outcomes and incorporation of marker status type addressed CSM after surgery.4 To assess the potential additional clinical utility, biomarker profile was assessed within a context of prognostic nomograms using previously described independent predictors of recurrence and survival (T or N stage, grade, LVI, and tumor architecture).4 Potential augmentation of an existing preoperative nomogram that incorporates grade, architecture, and tumor location was also performed with an additional of biomarker score.19 Bootstrap analysis was performed for both nomograms and bias-corrected Cindices were compared. In a second comparison, we removed the strongest predictor variable from each nomogram and compared bias-corrected C-indices between a nomogram with and without markers status as a prognostic variable. All reported P values are 2-sided and significance is delineated at .05.
RESULTS Clinicopathologic Outcomes Clinicopathologic information and immunohistochemical staining data were collected on 620 patients. Patient, tumor, and immunohistochemical data are presented in Table 1. Overall, 72 patients (11.6%) received perioperative chemotherapy (11 [1.8%] neoadjuvant and 61 [10%] adjuvant). Over a median follow-up of 27.3 months, 24.6% of patients recurred, and 21.8% died of UTUC. Biomarker Assessment On univariate marker analysis for event at 2 years, PI3K (P 1 was considered altered and cyclin D H score >2 was considered altered based on natural integral proportion cutpoints. Forty percent of UTUC tumors had PI3K overexpression and 44.8% had cyclin D overexpression. Cumulative marker score was defined as either low risk (no altered markers or 1 altered marker) or high risk (cyclin D and PI3K altered). High-risk marker profile was present in 92 patients (14.8%). High-risk marker score was uniformly associated with adverse pathologic and clinical outcomes (Table 2). The Kaplan-Meier analysis demonstrated a significant difference in recurrence-free survival (RFS) and CSM based on the biomarker score (Fig. 1A, B). Median RFS estimate for patients with a low-risk score was significantly longer than for patients with a high-risk score (142 months; 95% confidence interval [CI], 133-151 months vs 68.4 months; 95% CI, 54-83 months; P 2 and/or Nþ) Lymphovascular invasion No Yes Architecture Papillary Sessile Necrosis No Yes Concomitant CIS No Yes Focality Unifocal Multifocal Perioperative chemotherapy No Yes Recurrence of UTUC (only M0) No Yes Death due to UTUC No Yes
Entire Cohort (%)
Low Risk (%; 0-1 Marker Altered)
High Risk (%; 2 Markers Altered)
620
528 (85.2)
92 (14.8)
163 (26.3) 457 (73.7)
155 (29.4) 373 (70.6)
8 (8.7) 84 (91.3)
103 185 108 187 37
97 173 88 145 25
6 12 20 42 12
(16.6) (29.8) (17.4) (30.2) (6.0)
(18.4) (32.8) (16.7) (27.5) (4.6)
(6.5) (13.0) (21.8) (45.7) (13.0)
569 (91.8) 51 (8.2)
497 (94.1) 31 (5.9)
72 (78.3) 20 (21.7)
388 (62.6) 232 (37.4)
352 (66.7) 176 (33.3)
36 (39.1) 56 (60.9)
483 (77.9) 137 (22.1)
423 (80.1) 105 (19.9)
60 (65.2) 32 (34.8)
528 (85.2) 92 (14.8)
457 (86.6) 71 (13.4)
71 (77.2) 21 (22.8)
532 (85.8) 88 (14.2)
456 (86.4) 72 (13.6)
76 (82.6) 16 (17.4)
509 (82.1) 111 (17.9)
442 (83.7) 86 (16.3)
67 (72.8) 25 (27.2)
495 (79.8) 125 (20.2)
431 (81.6) 97 (18.4)
64 (69.6) 28 (30.4)
548 (88.4) 72 (11.6)
475 (90.0) 53 (10.0)
73 (79.3) 19 (20.7)
466 (75.2) 154 (24.8)
415 (78.6) 113 (21.4)
51 (55.4) 41 (44.6)
486 (78.4) 134 (21.6)
431 (81.6) 97 (18.4)
55 (59.8) 37 (40.2)
P Value
RESULTS
CONCLUSION
To evaluate the prognostic value of altered mammalian target of rapamycin (mTOR) pathway biomarkers in upper tract urothelial carcinoma (UTUC). We performed a multi-institutional review of clinical and pathologic information on patients receiving extirpative surgery for UTUC from 1990 to 2008. Immunohistochemistry for phosphorylated-S6, mTOR, phosphorylated-mTOR, PI3K, phosphorylated-4EBP1, phosphorylated-AKT, PTEN, HIF-1a, raptor, and cyclin D was performed on tissue microarrays from radical nephroureterectomy (RNU) specimens. Prognostic markers were identified and the significance of altered markers was assessed with the Kaplan-Meier analysis and the Cox regression analysis. Six hundred twenty patients were included. Over a median follow-up of 27.3 months, 24.6% of patients recurred and 21.8% died of UTUC. On multivariate analysis, PI3K (odds ratio, 1.28; P ¼ .001) and cyclin D (odds ratio, 3.45; P ¼ .05) were significant predictors of clinical outcomes. Cumulative marker score was defined as low risk (no altered markers or 1 altered marker) or high risk (cyclin D and PI3K altered). Patients with high-risk marker score had a significantly higher proportion of high-grade disease (91% vs 71%; P 90%) but dramatically declined for patients with locally advanced or lymph node (LN)e positive disease (5%-10%).4
Aditya Bagrodia and Laura-Maria Krabbe contributed equally to this work. Financial Disclosure: The authors declare that they have no relevant financial interests. From the Department of Urology, University of Texas Southwestern Medical Center, Dallas, TX; the Department of Urology, University of Muenster Medical Center, Muenster, Germany; the Department of Pathology, University of Texas Southwestern Medical Center, Dallas, TX; Division of Biostatistics, Department of Clinical Sciences, University of Texas Southwestern Medical Center, Dallas, TX; the Department of Urology, MD Anderson Cancer Center, Houston, TX; the Department of Urology, University of Michigan Cancer Center, Ann Arbor, MI; the Department of Urology, Penn State Hershey Medical Center, Hershey, PA; the Department of Urology, Medical
University of Vienna, Vienna, Austria; the Department of Pathology, Centre Hospitalier Universitaire de Rennes, Rennes, France; the Department of Pathology, Medical University Vienna, Vienna, Austria; the Department of Urology, Ospedali Riuniti of Bergamo, Bergamo, Italy; the Department of Urology, Mannheim Medical Center, University of Heidelberg, Mannheim, Germany; and the Department of Urology, Centre Hospitalier Universitaire de Rennes, Rennes, France Address correspondence to: Vitaly Margulis, M.D., Department of Urology, University of Texas Southwestern Medical Center, 5323 Harry Hines Boulevard J8.130, Dallas, TX 75390-9110. E-mail: [email protected] Submitted: March 9, 2014, accepted (with revisions): July 29, 2014
1134
ª 2014 Elsevier Inc. All Rights Reserved
http://dx.doi.org/10.1016/j.urology.2014.07.050 0090-4295/14
Improved clinical outcomes may be expected with an earlier detection, enhanced risk stratification, and incorporation of effective systemic therapies. However, screening for UTUC is precluded by the low incidence; prognostic models currently have only approximately 75% accuracy for predicting clinicopathologic outcomes; and neoadjuvant and adjuvant chemotherapies are administered to 1300 patients in the original data set, pathologic specimens for tissue microarray (TMA) construction were obtained for 620 patients from RNU or PU specimens, and this subgroup comprised our study cohort. Surgical technique and indication for RNU vs PU was based on surgeon preference. The extent and completion of lymphadenectomy were based on surgeon discretion. In general, patients were seen at least every 3-4 months in the first postoperative year, semiannually for the second year, and then annually. Comprehensive clinical and pathologic data were maintained in an institutional review boardeapproved database. Multiple data reviews and quality assurance checks were performed throughout the data collection process to ensure accuracy and completeness of data points. All surgical specimens were processed according to the standard pathologic procedures. Pathologic variables were prospectively rereviewed by a genitourinary pathologist (P.K.), who was blinded to clinical outcomes, using defined uniform criteria. UROLOGY 84 (5), 2014
STATISTICAL ANALYSIS
1135
Table 1. Descriptive characteristics of 620 patients treated with radical nephroureterectomy and distal ureterectomy Variable Age, y Median (range), y Women Men Grade Low grade High grade Location Kidney Ureter Ureteroenteric anastomosis Pathologic stage Ta/CIS T1 T2 T3 T4 N stage Node/x Nodeþ Organ-confined disease Yes (T stage 2; N ¼ 0/x) No (T stage >2 and/or Nþ) Lymphovascular invasion No Yes Architecture Papillary Sessile Necrosis No Yes Concomitant CIS No Yes Unifocal Multifocal Perioperative chemotherapy No Yes Normal PI3K Overexpressed PI3K Normal cyclin D Overexpressed cyclin D Cumulative marker score Low risk (0-1 markers altered) High risk (2 markers altered) Recurrence of UTUC (only M0) No Yes Death due to UTUC No Yes
No. (%; Total Cohort, n ¼ 620) 68.8 69.5 (27-104) 271 (43.7) 349 (56.3) 163 (26.3) 457 (73.7) 450 (72.6) 158 (25.5) 12 (1.9) 103 185 108 187 37
(16.6) (29.8) (17.4) (30.2) (6.0)
569 (91.8) 51 (8.2) 388 (62.6) 232 (37.4) 483 (77.9) 137 (22.1) 528 (85.2) 92 (14.8) 532 (85.8) 88 (14.2) 509 111 495 125
(82.1) (17.9) (79.8) (20.2)
548 72 374 246 342 278
(88.4) (11.6) (60.3) (39.7) (55.2) (44.8)
528 (85.2) 92 (14.8) 466 (75.2) 154 (24.8) 486 (78.4) 134 (21.6)
CIS, carcinoma in situ; PI3K, phosphatidyl inositol 3-kinase; UTUC, upper tract urothelial carcinoma.
parameters. Differences in variables with a continuous distribution across dichotomous or ranked categories were assessed using the Mann-Whitney U test or the KruskalWallis nonparametric analysis of variance test, respectively. Significance of altered markers was assessed with 1136
the Kaplan-Meier analysis and the Cox regression analysis. Univariate and multivariate Cox regression models with known predictors of outcomes and incorporation of marker status type addressed CSM after surgery.4 To assess the potential additional clinical utility, biomarker profile was assessed within a context of prognostic nomograms using previously described independent predictors of recurrence and survival (T or N stage, grade, LVI, and tumor architecture).4 Potential augmentation of an existing preoperative nomogram that incorporates grade, architecture, and tumor location was also performed with an additional of biomarker score.19 Bootstrap analysis was performed for both nomograms and bias-corrected Cindices were compared. In a second comparison, we removed the strongest predictor variable from each nomogram and compared bias-corrected C-indices between a nomogram with and without markers status as a prognostic variable. All reported P values are 2-sided and significance is delineated at .05.
RESULTS Clinicopathologic Outcomes Clinicopathologic information and immunohistochemical staining data were collected on 620 patients. Patient, tumor, and immunohistochemical data are presented in Table 1. Overall, 72 patients (11.6%) received perioperative chemotherapy (11 [1.8%] neoadjuvant and 61 [10%] adjuvant). Over a median follow-up of 27.3 months, 24.6% of patients recurred, and 21.8% died of UTUC. Biomarker Assessment On univariate marker analysis for event at 2 years, PI3K (P 1 was considered altered and cyclin D H score >2 was considered altered based on natural integral proportion cutpoints. Forty percent of UTUC tumors had PI3K overexpression and 44.8% had cyclin D overexpression. Cumulative marker score was defined as either low risk (no altered markers or 1 altered marker) or high risk (cyclin D and PI3K altered). High-risk marker profile was present in 92 patients (14.8%). High-risk marker score was uniformly associated with adverse pathologic and clinical outcomes (Table 2). The Kaplan-Meier analysis demonstrated a significant difference in recurrence-free survival (RFS) and CSM based on the biomarker score (Fig. 1A, B). Median RFS estimate for patients with a low-risk score was significantly longer than for patients with a high-risk score (142 months; 95% confidence interval [CI], 133-151 months vs 68.4 months; 95% CI, 54-83 months; P 2 and/or Nþ) Lymphovascular invasion No Yes Architecture Papillary Sessile Necrosis No Yes Concomitant CIS No Yes Focality Unifocal Multifocal Perioperative chemotherapy No Yes Recurrence of UTUC (only M0) No Yes Death due to UTUC No Yes
Entire Cohort (%)
Low Risk (%; 0-1 Marker Altered)
High Risk (%; 2 Markers Altered)
620
528 (85.2)
92 (14.8)
163 (26.3) 457 (73.7)
155 (29.4) 373 (70.6)
8 (8.7) 84 (91.3)
103 185 108 187 37
97 173 88 145 25
6 12 20 42 12
(16.6) (29.8) (17.4) (30.2) (6.0)
(18.4) (32.8) (16.7) (27.5) (4.6)
(6.5) (13.0) (21.8) (45.7) (13.0)
569 (91.8) 51 (8.2)
497 (94.1) 31 (5.9)
72 (78.3) 20 (21.7)
388 (62.6) 232 (37.4)
352 (66.7) 176 (33.3)
36 (39.1) 56 (60.9)
483 (77.9) 137 (22.1)
423 (80.1) 105 (19.9)
60 (65.2) 32 (34.8)
528 (85.2) 92 (14.8)
457 (86.6) 71 (13.4)
71 (77.2) 21 (22.8)
532 (85.8) 88 (14.2)
456 (86.4) 72 (13.6)
76 (82.6) 16 (17.4)
509 (82.1) 111 (17.9)
442 (83.7) 86 (16.3)
67 (72.8) 25 (27.2)
495 (79.8) 125 (20.2)
431 (81.6) 97 (18.4)
64 (69.6) 28 (30.4)
548 (88.4) 72 (11.6)
475 (90.0) 53 (10.0)
73 (79.3) 19 (20.7)
466 (75.2) 154 (24.8)
415 (78.6) 113 (21.4)
51 (55.4) 41 (44.6)
486 (78.4) 134 (21.6)
431 (81.6) 97 (18.4)
55 (59.8) 37 (40.2)
P Value
![[Prognostic factors of upper tract urothelial carcinoma].](/assets/img/hold.png)
