Drug-targeted therapies and Predictive Prognosis: Changing Role for the Pathologist

Transcription

1 Drug-targeted therapies and Predictive Prognosis: Changing Role for the Pathologist Moderator: S. Terence Dunn, Ph.D. Associate Professor, Pathology Director, Molecular Pathology Laboratory University of Oklahoma Health Sciences Center What is Targeted Therapy? Drugs which target specific molecules involved in a disease process In context of targeted cancer therapies, these are specific molecules involved in growth and spread of tumor cells Could argue that all drug therapies are targeted therapies molecular basis to them all Targeted therapy should be measurable clinically Approaches to Drug Development Molecules for Targeted Therapy Traditional approach Identify drug to test against specific disease Evaluate its laboratory and clinical effects Targeted approach Identify key pathway in disease Develop drug to targeted molecule 4 types of molecules used Small molecule inhibitors e.g., Gleevec - brand name Imatinib generic name STI571 lab development name Monoclonal antibodies Identify the target molecule Evaluate its laboratory and clinical effects Human Murine Chimeric Humanized -umab -momab -ximab -zumab Fusion proteins Antisense oligonucleotides and PNAs Current Use of Targeted Therapies Monotherapy In combination with radiation In combination with chemotherapy In combination with other targeted therapies Signal transduction inhibitors -target protein kinases (esp. tyrosine kinases), their ligands or signal transducers Protein kinases Catalyze transfer of phosphate from ATP to protein substrate Regulate variety of cellular processes Receptor kinases extracellular receptor and intracellular portion Non-receptor kinases interact with receptor kinases, other proteins, lipids and DNA

2 Signal transduction inhibitors Small molecule inhibitors e.g., Gleevec (imatinib mesylate) - binds to ATP-binding site of ABL, PDGF, SCF and c-kit tyrosine kinases; GIST and CML e.g., Iressa (gefitinib) and Tarceva (erlotinib) targets EGFR; advanced NSCLC Monoclonal antibodies e.g., Herceptin (trastuzumab) targets HER2; breast cancer e.g., Erbitux (cetuximab) targets EGFR; advanced colorectal cancer with Irinotecan Antisense oligonucleotides e.g., Affinitak " (ISIS3521) targets PKC#; in clinical trials for treatment of breast and NSCLC Proteasome inhibitors Proteins Proteasome Proteins Proteasome Amino acids Apoptosis Normal cellular activity Proteasome inhibition Proteasome inhibitors Small molecule inhibitors e.g., Velcade (PS-341 or bortexomib) multiple myeloma that has failed other treatments Apoptosis-inducing drugs Antisense oligonucleotides e.g., Genesense " (oblimersen) targets BCL-2 mrna; in clinical trials for treatment of leukemia, NHL and solid tumors Angiogenesis inhibitors target angiogenic factors, angiogenic factor receptors, endothelial cell growth factors, or tissue metalloproteinases Small molecule inhibitors e.g., Neovastat " - in clinical trials for kidney cancer and others Monoclonal antibodies e.g., Avastin (bevacizumab; anti-vegf) first-line metastatic colorectal cancer in combination with 5FU Antisense oligonucleotides e.g., Angiozyme " (anti-angiogenesis ribozyme) in clinical trials for colorectal and breast cancer Immunotherapy Directed immunotargeting of tumor cells e.g., Rituxan (rituximab) targets CD20; relapsed or refractory NHL e.g., Campath (alemtuzumab) targets CD52; CLL Radioimmunotherapy agents e.g., Bexxar (tositumomab; I 131 -radiolabeled anti- CD20) and Zevalin (ibritumomab; Y 90 -radiolabeled anti-cd20) CD20-expressing relapsed or refractory NHL

3 Immunotherapy Cytotoxin-complexed immunotherapy agents e.g., IL13-PE38 " - fusion protein of Pseudomonas toxin conjugated to IL13; in clinical trials for gliomas e.g., Ontak - fusion protein of IL2 and diphtheria toxin; advanced or refractory CTCL e.g., Gemtuzumab (mylotarg) - calicheamicin conjugated anti-cd33; refractory AML Chemotherapy-complexed immunotherapy agents Others e.g., HLA class II molecules (e.g., anti-lym-1 and apolixumab) G-proteins (Ras) Histone deacetylase complex CD markers (e.g., CD19, CD22, CD25) Challenges for Targeted Therapy Multiple redundant pathways Alternative ligands for receptors Alternative signaling pathways Design of effective drugs Side-effects Defining appropriate patient groups Resistance through mutations Changing Role for the Pathologist Patients eligibility Immunohistochemistry Flow cytometry FISH Mutation analysis Microarray analysis Response to treatment Quantitative measurement methods as above Monitoring for resistance to treatment Mutation analysis

4 Association of Molecular Pathology USCAP Companion Meeting Sunday, February 12, :00 PM Dan Jones, MD, PhD Associate Professor Medical Director, Molecular Diagnostic Laboratory Division of Pathology and Laboratory Medicine U.T. M. D. Anderson Cancer Center Houston, Texas "Molecular Monitoring Strategies to Track Response to BCR-ABL Kinase Inhibitors in CML" Failure to achieve 3-log or 4-log fold reductions in bcr-abl transcript levels within 6 months, as assessed by quantitative reverse transcription PCR (RT-qPCR), provided additional risk stratification. 7 The bcr-abl RT-qPCR assay at M.D. Anderson Cancer Center: 8,9 10 ml PB or 3 ml BM aspirate as input for RNA extraction/cdna synthesis. Single tube TaqMan-based assay covering the b2a2, b3a2 and e1a2 transcripts. Normalization to total abl transcript levels (fusion + normal abl). 1;100,000 lower limit of detection. Post-PCR sizing to detect transcript type. Samples run in duplicate. Chronic myelogenous leukemia (CML) remains the success story in tumor molecular diagnostics because it fits the targeted therapy paradigm nearly perfectly: CML (and Ph+ ALL) are defined by the presence of the bcr-abl fusion transcript/protein. Imatinib (Gleevec/Glivec/STI571) is a widely used small molecule inhibitor that competes for the ATP-binding pocket of the abl kinase and thus (rather) selectively blocks proliferation of the CML clone. 1 Monotherapy with imatinib is the standard therapy for CML worldwide. 2-4 However, other (curative) therapies are available, including IFN +/- cytarabine, marrow transplantation, other bcr-abl inhibitors, and combination therapies. 5 Primary or secondary resistance to imatinib requires dose escalation 6 or a change in therapy. The goals of molecular monitoring in CML thus include: Making the diagnosis (i.e. differentiating CML from other MPDs). Minimal residual disease monitoring. Monitoring for imatinib resistance. Predicting therapeutic response to new agents once resistance develops. Where is the evidence that molecular monitoring of CML response to imatinib is clinically important? The IRIS Trial ( ): 1,106 patients with newly diagnosed CML, randomized to 400 mg imatinib qd vs. IFNalpha and cytarabine (The International Randomized Study of Interferon and STI571 Study). 3 Imatinib was associated with higher response rates (74% predicted complete cytogenetic response (CCR) at 18 months) than IFNalpha/cytarabine (14.5% predicted CCR) and was more easily tolerated. Failure to achieve CCR for either group was strongly predictive of a shorter progressive-free survival. The molecular testing community is moving towards more calibrated and reproducible assays for bcr-abl RT-qPCR Assessment of accuracy requires development of universal standards o Favored approach is dissemination of a (readily available/commercial) bcr-abl standard that can be used to calibrate each lab s bcr-abl/control ratio at several dilutions (analogous to the INR in coagulation). Establishing the minimal analytical sensitivity required for clinical use o Minimum sensitivity would be in the range of 4-log reduction from baseline (i.e. the MMR established in the IRIS trial). o Sensitivity controls should be included in every run Monitoring precision by tracking run-to-run variability o Guidelines on when to reject or repeat testing (e.g. less than 2-4-fold variation in replicates down to the level of MMR). Establishing analytical specificity o How to assess low-level false-positives, including use of post-pcr sizing or a qualitative (nested) assay. Mechanisms of therapy resistance in CML Related to overcoming imatinib blockade (i.e. bcr-abl dependent): Point mutations in bcr-abl kinase domain (KD). 10,11 Amplification of the bcr-abl locus. 12 Complex rearrangements of bcr-abl transcript producing altered bcr-abl proteins. Altered gene regulation/ inhibitor feedback loops. Related to bypassing imatinib (i.e. bcr-abl independent mechanisms): Activation of others kinases besides bcr-abl (signal bypass). 13 Clonal evolution, particularly p53 loss secondary to isochromosome 17q 12 and acquisition of AML-associated translocations involving the core binding factors. 14 New bcr-abl kinase inhibitors have shown activity against imatinib-resistant CML 15 AMN-107 (Novartis) o Activation state-independent bcr-abl inhibitor, in contrast to imatinib. 16 o Similar kinase specificity to imatinib (including activity against PDGFR and kit/cd117 tyrosine kinases). 17 o In vitro, AMN-107 can block bcr-abl kinase activity in unmutated bcr-abl and in bcr-abl with a number of imatinib-associated KD mutations (but not T315I). 18,19 o In vivo. the range of response of AMN-107 is similar but not identical to predicted range. Dasatinib/BMS (Bristol-Myers Squibb) o Dual SRC/ABL kinase inhibitor with expanded kinase specificity (e.g. LYN). 20 o Might work by overcoming KD mutations in bcr-abl 18 or alternatively by blocking other growth regulatory kinases. 21 o bcr-abl KD mutations that arise following dasatinib treatment include novel sites not seen with imatinib. 22 o If 10-fold rise in bcr-abl transcripts levels as determined by qpcr (independent of cytogenetic findings). o Following shifts in therapy from imatinib to another agent at 6 weeks and again at 3 months. o Rescreen of KD mutation-negative cases at 1-year if still disease is still imatinib-resistant. Use of quantitative mutations screening methods (e.g. ASO-qPCR or Pyrosequencing using primers for mutational hotspots) may be useful in highintensity screening in the post-therapy-shift interval. Imatinib resistance mediated by bcr-abl amplification is assessed by karyotyping and FISH analysis at time of bone marrow aspirate (i.e. presence of extra Ph). Other approaches for monitoring for KD mutations: Screen for KD mutations for any case with >2X fold elevation in bcr-abl qpcr levels (Adelaide Australia group) 23 High-throughput screening of CML cases at regular intervals on imatinib using DHPLC/Wave methodology: confirm suspected KD mutations with definitive sequencing o May detect multiple preexisting, low-level KD mutations that will be difficult to interpret clinically. 24 Two major initiatives for worldwide harmonization in bcr-abl molecular testing CAP survey covering minimal residual monitoring (coming Spring 2006) International consensus conference (October th, 2005, NIH, Bethesda): Molecular monitoring of CML o Organized by John Goldman o 53 participants including representation from AMP, CAP, Pharma and biotech industry o Goals: consensus document on how to test (qpcr, KD mutations) and how often (clinical algorithms) Developing disease-specific algorithms that can applied over the clinical course of CML disease M.D. Anderson Cancer Center molecular monitoring algorithm for CML: RT-qPCR o PB used whenever possible o During routine therapy: q3 months. o During change in therapy: q6 weeks, or per trial design Monitoring for bcr-abl mutations is done by assessing the entire KD (codons ) by bidirectional Sanger sequencing o At time of initial visit for imatinib-resistant disease. o If complete cytogenetic response (CCR) is not obtained by 6-12 months. o If CCR is lost at any point during therapy (excluding inability to take imatinib).

5 T References: 1. Druker, B. J. et al. Effects of a selective inhibitor of the Abl tyrosine kinase on the growth of Bcr- Abl positive cells. Nat Med 2, (1996). 2. Druker, B. J. et al. Efficacy and safety of a specific inhibitor of the BCR-ABL tyrosine kinase in chronic myeloid leukemia. N Engl J Med 344, (2001). 3. O'Brien, S. G. et al. Imatinib compared with interferon and low-dose cytarabine for newly diagnosed chronic-phase chronic myeloid leukemia. N Engl J Med 348, (2003). 4. Kantarjian, H. M. et al. Long-term survival benefit and improved complete cytogenetic and molecular response rates with imatinib mesylate in Philadelphia chromosome-positive chronicphase chronic myeloid leukemia after failure of interferon-alpha. Blood 104, (2004). 5. Hoover, R. R., Mahon, F. X., Melo, J. V. & Daley, G. Q. Overcoming STI571 resistance with the farnesyl transferase inhibitor SCH Blood 100, (2002). 6. Kantarjian, H. M. et al. Dose escalation of imatinib mesylate can overcome resistance to standarddose therapy in patients with chronic myelogenous leukemia. Blood 101, (2003). 7. Hughes, T. P. et al. Frequency of major molecular responses to imatinib or interferon alfa plus cytarabine in newly diagnosed chronic myeloid leukemia. N Engl J Med 349, (2003). 8. Cortes, J. et al. Molecular responses in patients with chronic myelogenous leukemia in chronic phase treated with imatinib mesylate. Clin Cancer Res 11, (2005). 9. Luthra, R., Sanchez-Vega, B. & Jeffrey Medeiros, L. TaqMan RT-PCR assay coupled with capillary electrophoresis for quantification and identification of bcr-abl transcript type. Mod Pathol 17, (2004). 10. Gorre, M. E. & Sawyers, C. L. Molecular mechanisms of resistance to STI571 in chronic myeloid leukemia. Curr Opin Hematol 9, (2002). 11. Roumiantsev, S. et al. Clinical resistance to the kinase inhibitor STI-571 in chronic myeloid leukemia by mutation of Tyr-253 in the Abl kinase domain P-loop. Proc Natl Acad Sci U S A 99, (2002). 12. Hochhaus, A. Cytogenetic and molecular mechanisms of resistance to imatinib. Semin Hematol 40, (2003). 13. Donato, N. J. et al. BCR-ABL independence and LYN kinase overexpression in chronic myelogenous leukemia cells selected for resistance to STI571. Blood 101, (2003). 14. Merzianu, M. et al. inv(16)(p13q22) in chronic myelogenous leukemia in blast phase: a clinicopathologic, cytogenetic, and molecular study of five cases. Am J Clin Pathol 124, (2005). 15. Walz, C. & Sattler, M. Novel targeted therapies to overcome imatinib mesylate resistance in chronic myeloid leukemia (CML). Crit Rev Oncol Hematol (2005). 16. Schindler, T. et al. Structural mechanism for STI-571 inhibition of abelson tyrosine kinase. Science 289, (2000). 17. Golemovic, M. et al. AMN107, a novel aminopyrimidine inhibitor of Bcr-Abl, has in vitro activity against imatinib-resistant chronic myeloid leukemia. Clin Cancer Res 11, (2005). 18. O'Hare, T. et al. In vitro activity of Bcr-Abl inhibitors AMN107 and BMS against clinically relevant imatinib-resistant Abl kinase domain mutants. Cancer Res 65, (2005). 19. Weisberg, E. et al. Characterization of AMN107, a selective inhibitor of native and mutant Bcr- Abl. Cancer Cell 7, (2005). 20. Shah, N. P. et al. Overriding imatinib resistance with a novel ABL kinase inhibitor. Science 305, (2004). 21. Martinelli, G., Soverini, S., Rosti, G. & Baccarani, M. Dual tyrosine kinase inhibitors in chronic myeloid leukemia. Leukemia 19, (2005). 22. Burgess, M. R., Skaggs, B. J., Shah, N. P., Lee, F. Y. & Sawyers, C. L. Comparative analysis of two clinically active BCR-ABL kinase inhibitors reveals the role of conformation-specific binding in resistance. Proc Natl Acad Sci U S A 102, (2005). 23. Branford, S. et al. Real-time quantitative PCR analysis can be used as a primary screen to identify patients with CML treated with imatinib who have BCR-ABL kinase domain mutations. Blood 104, (2004). 24. Willis, S. G. et al. High-sensitivity detection of BCR-ABL kinase domain mutations in imatinibnaive patients: correlation with clonal cytogenetic evolution but not response to therapy. Blood 106, (2005). MJ\U.#+,&$.&J**($&' ]*&&$/*1^:Z2&(CJ:W: U3*5*3,&0U3$&$-@.2&0,($.& Epidermal growth factor receptor (EGFR) >?@A,&0$(%3$',&0%4*1*0$%-.5*1*0<6W1:H(,&3*6U.)*&C()*N.<*33,21*,(*P,&04,%/$1%(-3.&*0$&TSRV7T9:>?@A$%,3%.B&.4&,%-=>1<_=T,&0Z>AT:>?@A,&0()*-3.%*361*3,(*0/,#$36#*#<*1%Z>A"`&*27-=>1<_="9,&0>1<_=L,&0V,1*,33 -*33#*#<1,&*1*-*+(.1%(),(),5*$&(1$&%$-(61.%$&*B$&,%*,-($5$($*%:;)*/,#$36./()* >1<_%4,%.1$'$&,336&,#*0/.1()*%$#$3,1$(6(.()*(1,&%/.1#$&''*&*./()*,5$,& *16()1.<3,%(.%$%'*&*Cv-erb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`&*2:;)* 0$#*1$a,($.&+1.5$0*%%$'&,3,#+3$/$-,($.&7L9:D&3,<.1,(.16*K+*1$#*&(%C)*(*1.0$#*1 /.1#,($.&<*(4**&>?@A,&0Z>A"`&* *%,%(1.&'*1+1.3$/*1,($5*%$'&,3(),& >?@A).#.0$#*1%7V9: ;)*%$'&,3$&'-,%-,0*%(),(,1*%($#23,(*0<6>?@A,-($5,($.&,1*b2$(*-.#+3*K,&0$&5.35*#23($+3*0$//*1*&(+,()4,6%:^$',&0<$&0$&',&00$#*1$a,($.&-,2%*,2(.+).%+).163,($.&./()*$&(1,-6(.+3,%#$-0.#,$&%,&0,-($5,($.&./()*$&(1,-*3323,1 (61.%$&*B$&,%*:;)*,-($5,(*0(61.%$&*B$&,%*%+).%+).163,(*%(61.%$&*1*%$02*%$&(,1'*( +1.(*$&%: Therapies that involve EGFR >?@A),%<**&$0*&($/$*0,%<*$&',3(*1*0$&2+(.T`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b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he Pathologist s Role in Measuring Aberrant EGFR EGFR assessment at the protein expression level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not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c.1-3$&$-,3(1$,3%,1*#$&$#,3[,&6+,($*&(4$(),(2#.14$()dte./-*33%+.%$($5*/.1 >?@A$&,#*#<1,&.2%+,((*1&$%-.&%$0*1*0,-,&0$0,(*7Q9: EGFR assessment at the genomic level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a,($.&:;)*1*,1*,&($<.0$*%,5,$3,<3*/.1()* #2(,&(/.1#./()*1*-*+(.1C,&0()$%1*-.'&$a*%%+*-$/$-,336()*%).1(*1>?@A5DDD7TSP ""9: ;)*%*-.&0(6+*%./#2(,($.&%(),(,1*B&.4&(..--21$&>?@A.--21,3#.%( *K-32%$5*36$&32&'-,1-$&.#,%C,&0,1*#2-)#.1*+1*5,3*&($&-*1(,$&+.+23,($.&%./ +,($*&(%4$()%+*-$/$-(2#.1%[6.2&'4.#*&C&*5*1=%#.B*1%C,&0,0*&.-,1-$&.#,%7"LP "V9:W*%+$(*#2-)4.1B.&.()*1(6+*%./(2#.1%C()*1*),5*<**&5*16/*41*+.1(%./ >?@A#2(,($.&%.2(%$0*./32&'-,1-$&.#,%7"89:D&,3,1'*%*1$*%./1*%*-(*032&' -,&-*1%C#2(,($.&%4*1*/.2&0$&VFe./-,%*%C4$().&36.&*&.&=,0*&.-,1-$&.#, %).4$&',#2(,($.&7"O9:;)*%*>?@A#2(,($.&%,1*-32%(*1*0$&()*1*'$.&./*K.&%TR (."TC4)$-)$%()*,1*,(),(*&-.0*%()*M;\=<$&0$&'+.-B*(./()*(61.%$&*B$&,%* 0.#,$&./>?@A:@2&-($.&,336C()*#2(,($.&%,++*,1(.$&02-*1*02-*03*5*3%./ +).%+).(61.%$&*.&,-($5,($.&C,&01*b2$1*TF=/.303*%%./012'(.$&)$<$(+).%+).163,($.& 7"V9:;)*#.%(-.##.&#2(,($.&%$&5.35**K.&%TS,&0"TC4)$-),--.2&(%/.1#.1* (),&S8e./#2(,($.&%7"Q9:>K.&TS#2(,($.&%,1*0*3*($.&#2(,($.&%C2%2,336$&5.35$&' <*(4**&TF,&0"F<,%*+,$1%./()**K.&TS%*b2*&-*C2%2,336<*'$&&$&',1.2&0-.0.& QVOP.--,%$.&,3*K.&TS+.$&(#2(,($.&%,1*,3%.%**&7-.0.&QTS97"Q9:>K.&"T #2(,($.&%,1*,3#.%(,34,6%+.$&(#2(,($.&%C$&5.35$&'-.0.&R8R4$(),;(.? (1,&%5*1%$.&$&#.%(-,%*%7"O9: ;)*1*3,($.&%)$+<*(4**&%.#,($->?@A%*b2*&-*#2(,($.&%,&01*%+.&%*(.,&($= >?@A()*1,+$*%),%1*-*&(36<**&+2<3$%)*0$&%*5*1,3#,c.11*+.1(%:D(,++*,1%(),( #2(,($.&%$&()*>?@A'*&*.--21$&)$')*1/1*b2*&-6$&+,($*&(%(),(1*%+.&0$&%.#* #,&&*1(.,&($=>?@A()*1,+$*%7"OP"VP"RP"S9:Z.4*5*1C$#+.1(,&(36C,33./()*%* 1*+.1(%,3%.,-B&.43*0'*(),(%.#*+,($*&(%4$().2(#2(,($.&%,3%.1*%+.&0*0(.,&($= >?@A()*1,+6P/21()*1#.1*C&.%(206),%%).4&,%215$5,3,/(*1(1*,(#*&(4$(),&($= >?@A()*1,+$*%:

6 V M#+3$/$-,($.& (2#.1(6+*%:;)*#.%(4*33=%(20$*0,&0<*%(2&0*1%(..0,1*()*)$')'1,0*<1,$&(2#.1% novo9,&0 %*-.&0,167*5.35$&'/1.#3.4*1'1,0*3*%$.&9?_J%#,6<*)$%(.3.'$-,336 $&0$%($&'2$%),<3*C()*-3$&$-,3-.21%*,&0()*#2(,($.&,3+1./$3*,1*0$//*1*&(:\1$#,16?_J%4$33),5*,%).1(*1($#*-.21%*C+1*%*&(4$()/*,(21*%./?_JC,&04$33./(*&),5* <**&%).4&(.),5*,-($5$(6,',$&%('3$.#,%$&*K+*1$#*&(,3#.0*3%,&0$&%.#**,136-3$&$-,3(1$,3%7L89:Y()*1%(20$*%),5*%).4&#$&$#,3()*1,+*2($-,05,&(,'*%(. 7T9: 4$()$&-1*,%$&'-.+6&2#<*1%).4$&'$&-1*,%$&'DZU%(,$&$&'7LLPLT9P).4*5*1C()* +2<3$%)*0(),(5,3$0,(*()*2%*./-)1.#.'*&$-$&%$(2)6<1$0$a,($.&7UDHZ9: 8 ;,<3*Y&*[W$//*1*&((6+*%./(2#.1%(),(),5*B&.4&,3(*1,($.&%$&>?@A: Tumor Type J*().0/.1 #*,%21*#*&( ^2&',0*&.-,1-$&.#, U.3.&,0*&.-,1-$&.#, Somatic mutations: EGFR viii 7DZU,&0 \UA9 Somatic mutations: Exons \UAC %*b2*&-$&'9 Gene amplification Protein Expression 7@DHZ9 7DZU9 Clinical Testing = l l l`= \.%%$<36 A,1* = A,1* l i*%?3$.#,% l = l l i*% Z*,0,&0N*-B Hb2,#.2%-*33 -,1-$&.#, A,1* = A,1* l N.(-211*&(36 Practical Concerns for EGFR testing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`.11*%+.&%*(.()*1,+$*%7;,<3*T9:@21()*1#.1*C <*-,2%*()*>?@A%(.16$%*5.35$&'/.10$//*1*&((2#.1(6+*%C$($%3$B*36(),(()*%* +1.(.-.3%4$33),5*(.<*1*=1*5$*4*0+*1$.0$-,336(.1*#,$&2+=(.=0,(*4$()()*3$(*1,(21*: O A>@>A>NU>H T:N,()..NC?.3032%(HCf.'*3<,2#JM:>+$0*1#,3'1.4()/,-(.11*-*+(.1,&(,'.&$%(%[ &.5*3()*1,+6/.1()*(1*,(#*&(./)$')='1,0*'3$.#,%:N*21.%21'*16:"FFVC8V[TVRF= TVRR: ":>3=A,6*%_@C^.A2%%.\J:;,1'*($&'()**+$0*1#,3'1.4()/,-(.11*-*+(.1:_1$($%) ].21&,3./U,&-*1:"FFVCST[VTR=V"V: L:^,%B$&]]CH,&03*1M_:>+$0*1#,3'1.4()/,-(.11*-*+(.1[,+1.#$%$&'(,1'*($&%.3$0 (2#.21%:U,&-*1;1*,(#*&(A*5$*4%:"FFVCLF[T=TQ: V:H-,'3$.(($?fCH*35,''$?CN.5*33.HCZ$1%-)@A:;)*<$.3.'6./*+$0*1#,3'1.4() /,-(.11*-*+(.1$&32&'-,&-*1:U3$&$-,3U,&-*1A*%*,1-):"FFVCTF[T8: 8:W,&-*6]>C@1*$03$&_:;,1'*($&'*+$0*1#,3'1.4()/,-(.11*-*+(.1==,1*4*#$%%$&' ()*#,1Bg^,&-*(:"FFLCLO"[O"=OV: O:^,6/$*30^]C_*1&,10\HC?.30%(*$&NH:U.3.1#23($+3*K+.36#*1,%*-),$&1*,-($.&/.1 b2,&($(,($5*,&,36%$%./*+$0*1#,3'1.4()/,-(.11*-*+(.1'*&*%$&-.3.1*-(,3,0*&.-,1-$&.#,:].21&,3./h21'$-,3y&-.3.'6:"fflcrl[""q="lt: Q:H-,1(.aa$JC_*,1a$DC_*1,10$ACJ,&0.3*%$MC@,<1$%?CU,%-$&2H:>+$0*1#,3 '1.4()/,-(.11*-*+(.17>?@A9%(,(2%$&+1$#, *-(,3(2#.1%0.*%&.(-.11*3,(*4$() >?@A*K+1*%%$.&$&1*3,(*0#*(,%(,($-%$(*%[$#+3$-,($.&%/.1(1*,(#*&(4$()>?@A= (,1'*(*0#.&.-3.&,3,&($<.0$*%:].21&,3./U3$&$-,3Y&-.3.'6:"FFVC""[VQQ"=VQQR: R:U,+.&$'1.@:A,($.&,3*,&0-3$&$-,35,3$0,($.&./*+$0*1#,3'1.4()/,-(.11*-*+(.1,%,(,1'*($&()*(1*,(#*&(./)*,0,&0&*-B-,&-*1:M&($U,&-*1W12'%:"FFVCT8[LTT=L"F: S:?1,&0$%]ACJ*3)*#J@C?..0$&'h>CW,6ACZ.3%(fMCh,'*&*1JJ:^*5*3%./ ;?@=,3+),,&0>?@A+1.(*$&$&)*,0,&0&*-B%b2,#.2%-*33-,1-$&.#,,&0+,($*&( %215$5,3:].21&,3./()*N,($.&,3U,&-*1D&%($(2(*:TSSRCSF[R"V=RL": TF:A*%&$-BJ_CA.2()$*1]CX.&B$&;CH,<.>C\1$-.3.f>:>+$0*1#,3'1.4()/,-(.1 1*-*+(.1C-=J>;C<*(,=-,(*&$&C,&0+8L*K+1*%%$.&,%+1.'&.%($-$&0$-,(.1%$&%(,'*DD -.3.&-,&-*1[,($%%2*#$-1.,11,6%(206:U3$&$-,3U,&-*1A*%*,1-):"FFVCTF[LFOS=LFQ8: TT:?.30%(*$&NHCM1#$&J:>+$0*1#,3'1.4()/,-(.11*-*+(.1$##2&.)$%(.-)*#$-,3 1*,-($5$(6$&+,($*&(%4$()M#*1$-,&].$&(U.##$((**.&U,&-*1H(,'*Df-.3.&,0*&.-,1-$&.#,[$#+3$-,($.&%/.1,%(,&0,10$a*0%-.1$&'%6%(*#:U,&-*1:TLLTCS"[TLLT= TLVO: T":U2&&$&'),#WCZ2#<3*(iCH$*&,HCX),6,(WC_3*$<*1'ZCH,&(.1.MC_*(%WC J2*%*1JCZ,1%(1$-BMCf*1%36+*UCU),2DCf,&U2(%*#>:U*(2K$#,<#.&.()*1,+6,&0 -*(2K$#,<+32%$1$&.(*-,&$&$1$&.(*-,&=1*/1,-(.16#*(,%(,($--.3.1*-(,3-,&-*1:j%** -.##*&(k:n*4>&'3,&0].21&,3./j*0$-$&*:"ffvcl8t[llq=lv8: TL:H,3(a^_CJ*1.+.3N]C^.*)1*1\]CH1:CN**03*JNCX.+$(]CJ,6*1A]:\),%*DD(1$,3./-*(2K$#,<$&+,($*&(%4$()1*/1,-( *-(,3-,&-*1(),(*K+1*%%*%()**+$0*1#,3 '1.4()/,-(.11*-*+(.1:j%**-.##*&(k:].21&,3./U3$&$-,3Y&-.3.'6:"FFVC""[T"FT= T"FR: TV:;*<<2((NUCU,((*33>CJ$0'3*6ACU2&&$&'),#WCX*11W:H6%(*#$-(1*,(#*&(./ -.3.1*-(,3-,&-*1:>21.+*,&].21&,3./U,&-*1:"FF"CLR[TFFF=TFT8: T8:i.2&*%J:D%$##2&.)$%(.-)*#$%(16/.1*+$0*1#,3'1.4()/,-(.11*-*+(.1 *K+1*%%$.&,+..1+1*0$-(.1./1*%+.&%*(.*+$0*1#,3'1.4()/,-(.11*-*+(.1=(,1'*(*0 ()*1,+6g].21&,3./U3$&$-,3Y&-.3.'6:"FF8C"L[S"L=S"V: Q TO:h*33%MC_$%).+]JCZ*3#*%(*W:M#+3$/$*0'*&*/.1()**+$0*1#,3'1.4()/,-(.1 1*-*+(.1$&,)2#,&'3$.<3,%(.#,-*333$&**&-.0*%,&*&a6#,($-,336$&,-($5*+1.(*$&: J.3*-23,1mU*3323,1_$.3.'6:TSRRCR[V8OT=V8O8: TQ:N$%)$B,4,AC]$nWCZ,1#.&AUC^,a,1UHC?$33?NCU,5*&**hXCZ2,&'Z]:M #2(,&(*+$0*1#,3'1.4()/,-(.11*-*+(.1-.##.&$&)2#,&'3$.#,-.&/*1%*&),&-*0 (2#.1$'*&$-$(6:\1.-**0$&'%./()*N,($.&,3M-,0*#6./H-$*&-*%./()*G&$(*0H(,(*%./ M#*1$-,:TSSVCST[QQ"Q=QQLT: TR:H)$&.c$#,NC;,0,XCH)$1,$%)$HCX,#$16.;CX.-)$JCN,B,#21,ZCJ,B$&.XC H,6,ZCZ$1,&.ZCX21,(%2]CYB,XCD%)$#,12iCG%)$.i:\1.'&.%($-5,32*./ *+$0*1#,3'1.4()/,-(.11*-*+(.1$&+,($*&(%4$()'3$.<3,%(.#,#23($/.1#*:U,&-*1 A*%*,1-):"FFLCOL[OSO"=OSQF: TS:@*30B,#+JJC^,3,\C^,2NCA.&-,1$^C?2),M:>K+1*%%$.&./,-($5,(*0 *+$0*1#,3'1.4()/,-(.11*-*+(.1%CA,%='2,&.%$&*(1$+).%+),(*C,&0#$(.'*&=,-($5,(*0 +1.(*$&B$&,%*$&)2#,&'3$.<3,%(.#,#23($/.1#*%+*-$#*&%:N*21.%21'*16:TVV"C V8[TVV"=TV8L: "F:N,()..NC?.3032%(HCf.'*3<,2#JM:>+$0*1#,3'1.4()/,-(.11*-*+(.1,&(,'.&$%(%[&.5*3()*1,+6/.1()*(1*,(#*&(./)$')='1,0*'3$.#,%:N*21.%21'*16:TVRFC 8V[TVRF=TVRR: "T:\*0*1%*&JhCJ*3(.1&JCW,#%(12+^C\.23%*&ZH:;)*(6+*DDD*+$0*1#,3'1.4() /,-(.11*-*+(.1#2(,($.&:_$.3.'$-,3%$'&$/$-,&-*,&0+.(*&($,3(,1'*(/.1,&($=-,&-*1 ()*1,+6:M&&,3%./Y&-.3.'6:"FFTCT"[QV8=QOF: "":\,&.2%$%UCA,6a#,&fJC].)&%;?CA*&&*1UC^$2oCU,1(41$')(?C^**@;Ch,&' WC?,&ZCU,.WCX6+1$0$%MCH#6()@>C_1*-)<$*3JhC_21'*%%MhCY30^]CH-.(( MJ:>&'$&**1$&',&0-),1,-(*1$%,($.&./-)$#*1$-#.&.-3.&,3,&($<.06RFO7-)RFO9/.1 (,1'*(*0$##2&.()*1,+6./(2#.21%*K+1*%%$&'0*"=Q>?@A.1,#+3$/$*0>?@A:_1$($%) ].21&,3./U,&-*1:"FF8CS"[TFOS=TFQQ: "L:^6&-);]C_*33WhCH.10*33,AC?212<),',5,(23,HCYB$#.(.AMC_1,&&$',&_hC Z,11$%\^CZ,%*13,(HJCH2+B.]?CZ,32%B,@?C^.2$%WNCU)1$%($,&$WUCH*((3*#,&]C Z,<*1WM:M-($5,($&'#2(,($.&%$&()**+$0*1#,3'1.4()/,-(.11*-*+(.12&0*136$&' 1*%+.&%$5*&*%%./&.&=%#,33=-*3332&'-,&-*1(.'*/$($&$<:N*4>&'3,&0].21&,3./ J*0$-$&*:"FFVCL8F["T"S="TLS: "V:\,.hCJ$33*1fCo,B.4%B$JCW.)*1(6]C\.3$($XCH,1B,1$,DCH$&')_CZ**3,&AC A2%-)fC@23(.&^CJ,10$%>CX2+/*1WCh$3%.&ACX1$%JCf,1#2%Z:>?@1*-*+(.1 '*&*#2(,($.&%,1*-.##.&$&32&'-,&-*1%/1.#p&*5*1%#.B*1%p,&0,1*,%%.-$,(*0 4$()%*&%$($5$(6./(2#.1%(.'*/$($&$<,&0*13.($&$<:\1.-**0$&'%./()*N,($.&,3M-,0*#6./H-$*&-*%./()*G&$(*0H(,(*%./M#*1$-,:"FFVCTFT[TLLFO=TLLTT: "8:^**]hCH.2&'iZCX$#HiC\,1BhHCN,#HhC^**]iCi..N]C^**HZ: M<%*&-*./>?@A#2(,($.&$&()*B$&,%*0.#,$&$&-.##.&)2#,&-,&-*1%<*%$0*%&.&= %#,33-*3332&'-,&-*1:D&(*1&,($.&,3].21&,3./U,&-*1:"FF8CTTL[8TF=8TT: "O:X.%,B,;Ci,(,<*iC>&0.)ZCX24,&.ZC;,B,),%)$;CJ$(%20.#$;:J2(,($.&%./ ()**+$0*1#,3'1.4()/,-(.11*-*+(.1'*&*$&32&'-,&-*1[<$.3.'$-,3,&0-3$&$-,3 $#+3$-,($.&%:U,&-*1A*%*,1-):"FFVCOV[RSTS=RS"L: "Q:J,1-)*(($MCJ,1(*33,UC@*3$-$.&$^C_,1,%%$@CH,35,(.1*HCU)*33,MCU,#+3*%*\\C D,12%%$;CJ2-$33$@CJ*aa*(($MCU @CH,--.AC_2(($((,@:>?@A#2(,($.&%$& &.&=%#,33=-*3332&'-,&-*1[,&,36%$%./,3,1'*%*1$*%./-,%*%,&00*5*3.+#*&(./,1,+$0

7 R,&0%*&%$($5*#*().0/.10$,'&.%($-%-1**&$&'4$()+.(*&($,3$#+3$-,($.&%.& +),1#,-.3.'$-(1*,(#*&(:].21&,3./U3$&$-,3Y&-.3.'6:"FF8C"L[R8Q=RO8: "R:M#,&&]CX,36,&B1$%)&,HCJ,%%$.&\\CY)#]>C?$1,10^CH)$'*#,(%2ZC\*6(.& JC]21.%B*WCZ2,&'iCH(2,1(H,3#.&]CX$#iZC\.33,-B]ACi,&,'$%,4,XC?,a0,1 MCJ$&&,]WCX21$*]JCU,1<.&*W\:M<*11,&(*+$0*1#,3'1.4()/,-(.11*-*+(.1 "FF8CO8[""O="L8: "S:X.<,6,%)$HC_.''.&;]CW,6,1,#;C],&&*\MCX.-)*1YCJ*6*1%.&JC].)&%.& -,&-*1(.'*/$($&$<:N*4>&'3,&0].21&,3./J*0$-$&*:"FF8CL8"[QRO=QS": LF:_$'&*1HZCh.&'M]CJ,1B]CJ2)3<,$*1^ZCX$&a3*1XhCf.'*3%(*$&_C_$'&*1 WW:A*3,($.&%)$+<*(4**&'*&*,#+3$/$-,($.&,&0-)1.#.%.#,30*5$,($.&%$&#,3$'&,&( )2#,&'3$.#,%:U,&-*1?*&*($-%mU6(.'*&*($-%:TSRQC"S[TO8=TQF: LT:J,1b2*aMCh2ACo),.]C;,.]CH)$o:>5,32,($.&./*+$0*1#,3'1.4()/,-(.1 $##2&.)$%(.-)*#$%(167DZU9$&,1-)$5,3'3$.#,%2%$&'<1$')(=/$*30#$-1.%-.+6: W$,'&.%($-J.3*-23,1\,().3.'6:"FFVCTL[T=R: _ZCJ,-)*33iCA.*<2-BJJCH,3$%<216]CH$<%.&WACW2\3*%%$%WC_1..#*]CA.%%$ J^:U),1,-(*1$%,($.&./#.3*-23,1,3(*1,($.&%$&#$-1.0$%%*-(*0,1-)$5,3'3$.#,%:M-(, N*21.+,().3.'$-,:"FFTCTFT[L"T=LLL: LL:M&0*1%%.&GC?2.WCJ,3#*1_C_*1'*&)*$#M;C_1,&&%(1.#;CZ*0#,&ZC #*&$&'$.#,%:M-(,N*21.+,().3.'$-,:"FFVCTFR[TL8=TV": LV:f,1*3,JCA,&2&-.3.HJCJ.1,&0MC^,%($1$]CW*X$*1].//*>_C\21$-*33$^DC $&+,($*&(%4$()3.4='1,0*,%(1.-6(.#,%:].21&,3./H21'$-,3Y&-.3.'6:"FFVCRO[LV=VF: L8:^$_CU),&'UJCi2,&JCJ-X*&&,h?CH)2ZX:A*%$%(,&-*(.%#,33#.3*-23* $&)$<$(.1%./*+$0*1#,3'1.4()/,-(.11*-*+(.1$&#,3$'&,&('3$.#,%:U,&-*1A*%*,1-): "FFLCOL[QVVL=QV8F: *K+1*%%$.&,&0'*&*,#+3$/$-,($.&$&)$')='1,0*&.&=<1,$&%(*#'3$.#,%./-)$30)..0: U3$&$-,3U,&-*1A*%*,1-):TSSSC8[TQRO=TQS":

8 Molecular Markers for Targeted Therapy and Prediction of Prognosis in Breast Cancer Thomas M. Williams, MD University of New Mexico USCAP AMP Companion Meeting February 12, 2006 Prognostic Problems in Patients with Breast Cancer Approximately 1/3 of women with lymph node negative cancers develop recurrent disease. About 1/3 of women with positive lymph nodes at diagnosis are free of recurrence at 10 years. Conventional prognostic markers do not accurately predict which patients will develop recurrent disease and which patients will not, with or without adjuvant therapy. More accurate genomic and other markers of prognosis are needed. HER2 Amplification as a Genomic Marker for Targeted Therapy with Trastuzumab Searching for Prognosis Signatures via Microarray Gene Expression Studies HERA Trial: 1694 women with node positive or negative breast cancer who received 1 year of trastuzumab had a hazard ratio 0.54 for recurrence, death, or second primary in comparison to a control group. B-31and N9831 Trials: Fraction of patients with any first adverse event = 171/872 and 90/807 without trastuzumab and 83/864 and 50/808 with trastuzumab. Piccart-Gebhart, et al. NEJM 353: , 2005 Romond, et al. NEJM 353: , 2005 Isolate mrna from tumor cells (generally from fresh or frozen tissue and prepare fluorescently labeled cdna. Hybridize tumor cdnas and reference cdnas to an immobilized oligonucleotide library representing >20,000 genes. Analyze hybridization signals with a variety of algorithms to detect signatures of interest. Pitfalls: Quality control of arrays, analytical methods chosen, risks of over fitting data. Expression Signature Predictive of Recurrence in Women with N0 Breast Cancer Treated with Tamoxifen Identified 250 candidate genes based the literature. Focused on 16 within this group based on three clinical studies. Expression of 16 prospectively selected cancer related genes and 5 control genes assessed by RT-PCR on paraffin embedded tissue. 675 patients enrolled in NSABP trial. RT-PCR profiles obtained for 668 patients to segregate them into low, mid, and high risk groups via an algorithm. Ki67 STK15 Survivin CCNB1 MYBL2 GRB1 HER2 ER PGR BCL2 SCUBE2 MMP11 CTSL2 GSTM1 CD68 BAG1 Expression Signature Predictive of Recurrence in Women with Node Negative Breast Cancer Treated with Tamoxifen Recurrence scores could be calculated that were predictive of distant recurrence independent of patient age and tumor size. Low Risk Intermediate Risk High Risk Patients (%) Kaplan-Meier Estimate of 10 yr recurrence risk (%) (95% confidence limits) 6.8 ( ) 14.3 ( ) 30.5 ( ) P<0.001 Paik, et al. NEJM 351: , 2004

9 A 70-Gene Expression Signature Predictive of Survival in Women with Breast Cancer A 70-Gene Expression Signature Predictive of Survival in Women with Breast Cancer Employed previously devised 70-gene expression signature segregate 295 women with primary breast carcinomas into good and poor prognosis groups. Patients had node negative (151) or positive (144) disease and all were Stage I-II and < 53 years old. Algorithm Derived Risk Good Poor Year Survival 94.5% (+/-2.6) 54.6% (+/-4.4) Hazard Ratio 5.1 ( ) P< Van de Vijver, et al. NEJM 347: , Van t Veer, et al. Nature 415: , Examining Published Datasets with a Concurrent Data Mining Approach Genomic Targets Selected for Analysis Concurrently analyze public datasets linking genomic information with outcomes in women with breast cancer: Microarray-based gene expression data Van t Veer, et al. Nature 415: , Comparative genomic hybridization data Pollack, et al. PNAS, USA 99: , Sorlie, et al. PNAS USA 98: , Employ a computational platform to select 17 genomic targets for subsequent analysis CYP24 EXT1 NR1D1 MLN64 FANCA BIRC5 ZNF144 RAD21 GRB7 HEPSIN ZNF207 STK3 IMPA1 AL SMARCE1 ANXA11 PDCD6IP A collaboration of University of New Mexico and Exagen Diagnostics Figure 1. Map of 17 Genomic Regions PDCD6IP Study Population 229 women with primary invasive breast cancer treated at UNM Hospital At least 5 years follow-up with known outcomes (8.9 year mean) IMPA1 AL080059, STK3 RAD21, EXT1 ANXA11 NR* vs R* Age Hormone Overall Tumor Size Node Nuclear Receptor Stage Status Grade 4 N/A* 0 N/A 1 N/A 0 N/A 12 N/A FANCA ZNF207 ZNF144 MLN64 GRB7 NR1D1 SMARCE1 BIRC5 No Recurrence Avg Age < 50 Age > 50 n Age yrs yrs (yrs) N=92 N= II 2 & I & 3 HR+ HR- I III T1 T>1 N0 N> Recurred HEPSIN CYP24 *Key: NR= Non-recurrent; R=Recurrent; N/A=Not available

10 Fluorescent In Situ Hybridization Analysis Review charts to select patients that meet study criteria Review pathology reports and biopsy slides to select paraffin blocks Hybridize BAC clones to tissue sections Assess gene copy number with Metasystems image analysis software FISH personnel blinded to outcome and probe identity Computational scientists blinded to probe identity Probe Chromosome Location P (Wilcoxon) CYP24 20q EXT1 8q NR1D1 17q MLN64 17q FANCA 16q BIRC5 17q ZNF144 17q RAD21 8q GRB7 17q HEPSIN 19q ZNF207 17q STK3 8q IMPA1 8q AL q SMARCE1 17q ANXA11 10q SMARCE1 (dup) 1q PDCD6IP 3p Predictive Gene Subsets and Prognostic Index for HR+ Tumors BIRC5, CYP24, PDCD6IP Copy Number Recurrence Risk Prediction in Women with HR+ Tumors Amplification pattern of the BIRC5 CYP24 PDCD6IP trio Test Set N=67 Overall Risk Low Risk Moderate Risk High Risk Is the best predictor of recurrence PI = [log(cn BIRC5 )] [log(CN CYP24 )] [log(CN PDCD6IP )] Prognostic Index < >0.328 Recurrence 29.9% 9.1% 26.3% 50% PI = (CN BIRC5 ) X (CN CYP24 ) / (CN PDCD6IP ) 2 Fraction of Individuals at Risk 100% 33% 28% 40% Where CN equals copy number Odds Ratio of high: low risk = 9.52, CI = 2.12, P =

11 Hormone Receptor Positive Patients Hormone Receptor Positive, Node Negative Patients Recurrence Risk Recurrence Free Survival Low Risk Moderate Risk High Risk Avg. risk = 29.9% Risk = 9.1% Risk = 26.3% Risk=50.0% NPV=90.9% PPV=50.0% 32.8% 28.4% 40.6% Percentage of patients Independent test sample results (N=67) Recurrence Risk Recurrence Free Survival Avg. risk = 17.9% Low Risk Moderate Risk High Risk Risk = 6.3% Risk = 15.4% Risk=40.0% NPV=93.8% PPV=40.0% 41.0% 33.3% 25.6% Percentage of patients Time (years) Time (years) Independent test sample results (N=39) Predictive Gene Subsets and Prognostic Index for HR- Tumors NR1D1, SMARCE1, BIRC5 Copy Number Recurrence Risk Prediction in Women with HR- Tumors Amplification pattern of the NR1D1 SMARCE1 BIRC5 trio best predictor of recurrence Test Set N=35 Prognostic Index Overall Risk Low Risk Moderate Risk High Risk < >0.329 Recurrence 34.3% 9.1% 33.3% 58.3% PI = [log(cn NR1D1 )] [log(CN SMARCE1 )] [log(CN BIRC5 )] PI = (CN NR1D1 ) 2 / (CN SMARCE1 ) X (CN BIRC5 ) Fraction of Individuals at Risk 100% 31% 34% 34% Where CN equals copy number Odds Ratio of high: low risk = 12.3, CI = 1.45, P = Hormone Receptor Negative Patients Hormone Receptor Negative, Node Negative Patients Recurrence Risk Recurrence Free Survival Low Risk Moderate Risk High Risk Avg. risk =34.3% Risk = 9.1% Risk = 33.3% Risk=58.3% NPV=90.9% PPV=58.3% 31.4% 34.3% 34.3% Percentage of patients Independent test sample results (N=35) Recurrence Risk Recurrence Free Survival Low Risk Moderate Risk High Risk Avg. risk = 10.0% Risk = 0.0% Risk = 14.3% Risk= 20.0% PPV= 20.0% NPV=100.0% 40.0% 35.0% 25.0% Percentage of patients Independent test sample results (N=20) Time (years) Time (years)

12 Conclusions Acknowledgements Genomic markers that powerfully predict prognosis in breast cancer and allow targeted therapy have been identified. RNA expression markers DNA copy number markers The diversity of the genome and the multifactorial nature of breast cancer makes it likely that different studies will identify a variety of non-overlapping useful markers. Markers that can be assessed in typical surgical pathology assays may be of value. University of New Mexico John Hozier Therese Bocklage Katie Doeden Brian Hall Ian Rabinowitz Tom Williams Exagen Diagnostics Cole Harris Lisa Davis Lei Tang Patti Doherty Peter Hraber Yumiko Sakai