Embed Size (px)
Citation preview
High-Intermediate Risk Endometrial Cancer: Can Gene Expression Predict Recurrence? Cindy Tawfik, BA1; Angelina I. Londoño, PhD2; Allison Montgomery, BS1; Alba Martinez, MS3; Katherine A. Shelton, BS1; Beomjy Kim, BS1; Ashwini A. Katre, MS3; Warner K. Huh, MD3;
Eddy Yang, MD, PhD2; Kerri S. Bevis, MD, MSPH3; J. Michael Straughn, Jr., MD3; Charles A. Leath III, MD3; Rebecca C. Arend, MD3 1 School of Medicine , 2Department of Radiation Oncology , 3Division of Gynecologic Oncology, Department of Obstetrics & Gynecology; University of Alabama at Birmingham; Birmingham, Alabama
Introduction • Endometrial cancer (EMCA) is the most common gynecologic cancer in the United States. • Most cases are diagnosed at Stage 1 and have an excellent prognosis after surgery alone. • A subset of early-stage endometrial cancer patients are at higher risk for recurrence. • Prognostic factors of recurrence of early stage EMCA include lymphovascular space invasion
(LVSI), histologic grade (2 or 3), depth of myometrial invasion (>50%), and patient age. • Results from GOG 99 and PORTEC-2 led to the designation of a subset of patients known as high-
intermediate risk (H-IR) EMCA based on reproducible pathologic risk factors. • H-IR EMCA patients are at 20-30% risk of recurrence even in the setting of early stage disease. • Studies have shown that adjuvant therapy increases progression free survival, but does not
affect overall survival in these patients Objective: To identify a gene signature that determines which H-IR EMCA patients are at the highest risk for recurrence and to identify RNA expression changes that occur in the recurrent tumor.
Results
Conclusions
Methods
• IRB-approved retrospective cohort study. • All patients who underwent surgery at UAB and met criteria for H-IR EMCA based on GOG 99
between 2000-2010. • Clinical data was collected on all patients, but only patients who did not receive adjuvant
treatment were included in this analysis. • FFPE slides were made from patients that recurred and an equal number of patients that did
not recur were matched on a case-by-case basis (controls). • Tissue was available for analysis from 5 patients at the time of recurrence. DNA- CARIS protocol- NextSeq, a custom-designed SureSelect XT assay (592 whole-gene targets) was performed on 30 archival FFPE tumors (15 that recurred, 15 controls). All variants were detected with >99% confidence based on allele frequency and average coverage of >500 and an analytic sensitivity of 5%. Tumor DNA was harvested by manual microdissection. Genetic variants identified were interpreted by board-certified molecular geneticists and categorized based on ACMG standards. RNA- Nanostring protocol- Gene expression data was collected for 770 genes using the Nanostring nCounter® PanCancer Pathways Panel on 26 primary archival FFPE tumors (13 that recurred, 13 control) and on tumor from 5 of the 13 pts that recurred at the time of recurrence. Molecular profiles and pathway analysis of the cohorts (recurred vs. did not; primary vs. recurrent) were compared using nSolver Advanced Analysis Software. Genes were evaluated using a fold change of ±2 and a p-value of >0.05.
RNA expression pathway analysis of patients that recurred vs. matched patients that did not recur. Two pathways were significantly up regulated in the patients that did recur: Cell Cycle-Apoptosis and DNA Damage-Repair.
RNA expression pathway analysis of recurrent patients at the time of diagnosis vs. recurrence. 9 of 13 pathways were significantly altered with Driver Genes, DNA Damage-Repair, and MAPK being the top 3 most significantly altered.
Similar numbers and types of DNA mutations were found in patients that recurred compared to matching patients that did not recur. MSI-high and mutations in JAK1, DICER1, BRD3, PMS2, PDE4DIP, KMT2D, and BCOR genes were
more commonly observed in patients that recurred.
Table 1 and 2. Matched Paired Patient Demographics
Acknowledgments UAB Cancer Center, T32 5T32CA183926-02 Research Training Program in Basic and Translational Oncology, ABOG Early Career Grant, Norma Livingston Foundation, and Wilma Williams Education and Clinical Research for Endometrial Cancer Award (Foundation for Women’s Cancer), Caris Life Sciences.
• Similar number and type of DNA mutations were found in patients that recurred compared to matching patients that did not recur. • MSI-High and mutations in JAK1, DICER1, BRD3, PMS2, PDE4DIP, and BCOR genes were more comely observed in patients that
recurred. • RNA expression pathway analysis of EMCA patients that recurred vs matched patients that did not recur resulted in 2 pathways that
were significantly up regulated in the patients that recurred: Cell Cycle and DNA Damage. • Nine of 13 pathways were significantly altered in tissue samples from the time of recurrence compared to tissue from the time of
diagnosis. • The results of this study are hypothesis generating and will need to be validated in a much larger cohort. • These findings could potentially impact the decision to treat H-IR EMCA patients with adjuvant therapy.
Table 2: RNA Matches (n=26)
Did Not Recur (n=13) Recurred (n=13)
Age at diagnosis 69.85 69.77 p=0.95
Race
White 12 (92.3%) 9 (69.23%) p=0.135
African American
1 (7.69%) 4 (30.76%)
BMI
Range 19.37-43.79 23.82-39.62 p=0.19
Mean 27.17 30.98
Stage
1A 8 (61.54%) 8 (61.54%) p=1
1B 5 (38.46%) 5 (38.46%)
2 0 (0%) 0 (0%)
LVSI present 3 (23.08%) 3 (23.08%) p=1
Stromal Grade:1 13 (100%) 13 (100%)
Mean Follow Up 11.81 40.26 p<0.05
Table 1: DNA Matches (n=30)
Did Not Recur (n=15) Recurred (n=15)
Age at diagnosis 69.67 68.6 p=0.61
Race
White 14 (93.33%) 10 (66.67%) p=0.068
African American
1 (6.67%) 5 (33.33%)
BMI
Range 20.22-55.84 24.54-47.08
Mean 29.55 32.19 p=0.19
Stage
1A 6 (40%) 6 (40%) p=1
1B 7 (46.66%) 7 (46.66%)
2 2 (13.33%) 2 (13.33%)
LVSI present 6 (40.0%) 8 (53.3%) p=0.46
Stromal Grade
1 14 (93.33)% 9 (60%) p<0.05
2 1 (6.67%) 6 (40%)
Mean Follow Up 17.18 37.41 p<0.05
Gene Log2 fold change P-value Pathway
RAD50 -0.443 0.000907 DNA Damage - Repair
CCNA2 0.736 0.00216 Cell Cycle - Apoptosis
FGF18 1.33 0.00298 MAPK, PI3K, Ras
CDC25A 0.89 0.00684 Cell Cycle - Apoptosis
NOTCH1 0.436 0.00751 Driver Gene, Notch
Gene Log2 fold change P-value Pathway
LEFTY2 -3.26 2.64E-05 TGF-beta
PRKACB -1.58 9.26E-05 Cell Cycle - Apoptosis, Hedgehog, MAPK, Ras, Wnt
BCL2L1 1.01 0.000123 Cell Cycle - Apoptosis, JAK-STAT, PI3K, Ras, Transcriptional Misregulation
WT1 -2.77 0.000127 Driver Gene, Transcriptional Misregulation
CACNA1C -1.97 0.00022 MAPK
P=0.0233 P=0.051
Did not Recur Recurred
Primary Recurrent
Figure 2. A. Differences in 13 pathway scores of patients that did not recurred vs. patients that recurred. Box plots of individual patient pathways scores in B. Cell Cycle-Apoptosis and C. DNA Damage-Repair. D. Volcano plot of genes differentially expressed in patients that recurred / did not recur. E. Table of top five genes that are significantly different based on log fold change in those that recurred.
Figure 3. A. Changes in 13 pathway scores based on RNA expression at the time of diagnosis and recurrence. Box plots of individual patient pathway scores in B. Cell Cycle-Apoptosis, C. DNA Damage-Repair, D. MAPK, and E. Driver Genes. F. Volcano plot of genes differently expressed in tissue at time of recurrence / diagnosis. G. Table of top five genes that are significantly altered at recurrence based on log fold change.
29 35
Non Recurrent Recurrent
Nu
mb
er o
f M
uta
tio
ns
A. Average Number of Mutations
0 5 10 15 20 25 30 35 40
CTNNB1
MSI
TPR
PIK3R1
KMT2D
PIK3CA
RNF213
ARID1A
T.M.L.
PTEN
B. Top 10 Genes in All Patients
Gen
es
Number of mutations
3
3
3
3
4
4
4
4
4
4
5
5
5
5
6
9
9
12
15
18
0 5 10 15 20
RPL22
TCF3
TPR
TRIP11
CTCF
DDX10
IL6ST
KMT2C
KRAS
TP53
BCORL1
CTNNB1
LRP1B
MN1
RNF213
PIK3CA
PIK3R1
ARID1A
T.M.L.
PTEN
C. Top 20 Genes in Patients that did NOT Recur
Gen
es
Number of mutations
5
5
5
5
5
5
5
6
6
6
6
6
7
8
9
12
13
13
15
19
0 5 10 15 20
BCOR
CTNNB1
PDE4DIP
PIK3R1
PMS2
TP53
TRIP11
BRD3
DICER1
JAK1
RPL22
SPEN
MSI
TPR
PIK3CA
KMT2D
ARID1A
RNF213
T.M.L.
PTEN
D. Top 20 Genes in Patients that Recurred
Number of mutations
71%
16%
4%
3% 2%
2% 1% 1% <1%
<1%
E. Types of Mutations in All Patients (N= 30)
Missense
Frameshift
CODON_DELETION
NA
Nonsense
Splicing
Noncoding
CODON_INSERTION
CODON_CHANGE_PLUS_CODON_DELETION
STOP_LOST
71%
14%
5%
4% 3%
2% 1%
<1%
<1%
F. Types of Mutations in Patients that did NOT Recur (N= 15)
Missense
Frameshift
CODON_DELETION
NA
Nonsense
Splicing
CODON_INSERTION
Noncoding
CODON_CHANGE_PLUS_CODON_DELETION
71%
18%
3%
2% 3%
2% 1% <1%
<1%
G. Types of Mutations in Patients that Recurred (N= 15)
Missense
Frameshift
NA
Nonsense
CODON_DELETION
Splicing
Noncoding
CODON_INSERTION
STOP_LOST
Figure 1. Characterization of DNA mutations in EMCA patients that Recurred and did NOT Recur. A. Average number of total mutations found in matched groups of H-IR EMCA patients. B Top 10 genes in all patients. * (Mutations in the top 10 identified by Atlas Research Network Study in all EMCA pts). Top 20 Genes mutated in C. Patients that did not recur and D. Patients that recurred. Analysis of types of mutations observed in E. All patients, F. Patients that did not recur, and G. Patients that recurred.
DNA mutations and pathways changes in 4 cases that recurred
Figure 4. List of DNA mutations in primary tumor at the time of diagnosis in 4 H-IR EMCA patients that recurred and RNA expression pathway analysis in primary tumor compared to recurrent tumor.
N=15
13
15
9
7
7
3
2
3
5
4
2
4
3
4
4
4
2
3
3
3
N=15
11
15
8
9
7
8
7
7
5
6
5
4
4
4
3
5
4
4
4
5
Gen
es
A.
Did not Recur Recurred
B.
Did not Recur Recurred
C.
D.
E.
A.
P=0.049 B.
G.
F.
*
*
*
*
DNA
Patient 1
ARAF
ARID1A
CTNNB1
CTNNB1
EP300
ERBB3
ERC1
FCRL4
HGF
HOXA9
-5
-4
-3
-2
-1
0
1
2
3
4
Path
way
Sco
re
Patient 1 RNA Pathway Analysis Cell Cycle - Apoptosis
Chromatin Modification
DNA Damage - Repair
Driver Genes
Hedgehog
JAK-STAT
MAPK
Notch
PI3K
Ras
TGF-beta
Transcriptional Regulation
WntPrimary Recurrent
P=0.019 D.
Primary Recurrent
Primary Recurrent
P=0.004 C.
Primary Recurrent
P=0.003 E.
Primary Recurrent
P= 0.02 P= 0.05
H-IR EMCA n=292
Did not recur n=185
Recurred n=37
DNA n= 15
RNA n= 13
DNA n= 15
RNA n= 13
Primary tumor n= 5
Recurrent tumor n= 5
Selected matched patients based on:
Age Race Stage LVSI
Grade
DNA n= 30
RNA n= 26
Both n= 10
Radiation n=45
Observation n=222
Chemotherapy n=21
Radiation/Chemotherapy n=4
MSI High
Did not Recur Recurred
-6
-4
-2
0
2
4
6
Path
way
Sco
re
Patient 8 RNA Pathway Analysis Cell Cycle - Apoptosis
Chromatin Modification
DNA Damage - Repair
Driver Genes
Hedgehog
JAK-STAT
MAPK
Notch
PI3K
Ras
TGF-beta
Transcriptional Regulation
Wnt
Primary Recurrent
DNA Patient 8
AFF1 ALK
ASPSCR1 ATP2B3
C15orf65 CASC5 CASP8 CD274 CNTRL
CTNNB1 DICER1 DICER1
ELN ELN
EP300 EPHB1 EXT2
KMT2D
MAML2 NCOA2
NOTCH2 NUP214 NUP214 PIK3CA PLAG1 PTPRC
RNF213 RPN1 SPEN T.M.L. TET1
TNFRSF14 TNFRSF14
TRIP11 TSC2 USP6 WAS
-8
-6
-4
-2
0
2
4
6 Cell Cycle - Apoptosis
Chromatin Modification
DNA Damage - Repair
Driver Genes
Hedgehog
JAK-STAT
MAPK
Notch
PI3K
Ras
TGF-beta
Transcriptional Regulation
Wnt
Primary Recurrent
DNA Patient 4
ABL1 AKAP9 ASXL1 BCOR
BCORL1 BCR
BRD3 BRD3
CCDC6 cMET
CNOT3 CRTC3 CSF1R DOT1L ELF4
EPHB1 ERCC5 ERCC5 FANCA
FLT1 FSTL3 IRS2 IRS2 JAK1
KAT6A
-4
-2
0
2
4
6
8
10 Cell Cycle - Apoptosis
Chromatin Modification
DNA Damage - Repair
Driver Genes
Hedgehog
JAK-STAT
MAPK
Notch
PI3K
Ras
TGF-beta
Transcriptional Regulation
WntPrimary Recurrent
DNA Patient 7
AFF1
ARID1A BRAF
CDH1
CREB3L1
CTCF
CTNNA1 DDR2
EXT2
FANCC
FLT1
MSI
NFE2L2
PIK3CA
KDM5C KIAA1549
KRAS LCP1
MEN1 MLLT10
MSI NSD1
NT5C2 PIK3CA PMS2 PRCC PTEN PTEN
RECQL4 RECQL4 RNF43 RPL22 RUNX1
RUNx1T1 SMARCE1
SPECC1 T.M.L. TCEA1
TPR ZMYM2
IGF1R
NOTCH1
PIK3R1
PMS2
PTEN
PTEN
STIL
T.M.L.
TRIM27
TRRAP
ZBTB16
Patient 4 RNA Pathway Analysis Patient 7 RNA Pathway Analysis
Path
way
sco
re
Path
way
sco
re
PIK3CA
PTEN
ROS1
RPL22
RPN1
RPTOR
SETD2
SUZ12 T.M.L.
TCEA1
TET2 TPR
U2AF1
WHSC1
