Transcript ASCO_2009_files/Blanke GIST PD ASCO2009

Redefining our Knowledge of GIST:
From Old Chromosomes to New Drugs
Sarcoma Poster Discussion ASCO 2009
Charles D. Blanke, M.D., F.A.C.P.
Systemic Therapy Provincial Program Leader, B.C. Cancer Agency
Chief, Division of Medical Oncology, University of British Columbia
Abstracts for Discussion
• Spadaro: MDS developing after imatinib (#10532)
• Rink: Gene expression signatures and response in RTOG
0132 (#10533)
• Schoffski: HSP90 inhibitor IPI504 in mice with GIST xenografts
(#10534)
• Emile: PDGFR and KIT mutations in a French population
(#10535)
• Van Glabbeke: Type of progression on EORTC 62005
(#10536)
#10532-MDS developing after imatinib therapy
for GIST: Background and Importance
• Multiple other cancers described in GIST patients: colorectal, gastric,
small bowel, small-cell lung, prostate, renal, thyroid cancers,
angiosarcoma/ewing’s/PNET, melanoma, carcinoid, NHL, CLL, CML,
hepatic perivascular epithelioid tumour
• Potential etiologies:
– Coincidence (common cancers common, even in patients with rare cancers)
– Common underlying cause (genetic or environmental carcinogen)
– One resulting from treatment of other
• Question(s) raised: Do we need to alter monitoring of GIST patients
(or change therapy)?
MDS developing after imatinib therapy for
GIST: Methods
• 49 advanced GIST pts on imatinib underwent bone marrow
assessment
• Pathologic material reviewed to identify patients with MDS or
AML
MDS developing after imatinib therapy for
GIST: Results and Conclusion
• 8 patients acquired cytogenetic abnormalities
– Trisomy 8, developed in 7, occasionally reversibly
• 3 patients developed MDS
• 1 patient developed RAEB; subsequent transformation into
acute myelogenous leukemia
• The authors concluded we need to monitor GIST pts “closely”
MDS developing after imatinib therapy for
GIST: My Conclusions
• These numbers are scary and challenging; other large data
sets do not reflect this
– Counter-argument: we didn’t look hard
• TKIs should not be mutagenic
• No known common etiologies
• I do not think we need to bone marrow all pts
– Assess patients with persistant cytopenias
– ?Retrospective examination of existing data sets
#10533-Correlation of GIST gene expression
signatures and response on RTOG 0132: Background
and Importance
• 01321: Phase II neoadjuvant trial imatinib in KIT +
potentially resectable GIST (primary or metastatic)
• Primary objectives: feasibility and correlative projects
• Main trial results: Neoadjuvant (and post-operative)
imatinib safe and effective: 2-yr PFS/OS 81%/92%
• Question raised: Did trial identify a new prognostic or
predictive marker that will alter therapy for GIST patients?
1Eisenberg
J Surg Oncol 99:42, 2009
Gene expression signatures and response in
RTOG 0132: Methods
• Gene expression assessed pre/post imatinib therapy
(oligonucleotide microarrays)
• Differentially expressed genes identified (SAM)
• Genes’ ability to sensitize to imatinib assessed (custom
siRNA library with lethal screening approach)
Gene expression signatures and response in
RTOG 0132: Results
• 38 genes expressed lower levels pre-treatment
– 20 encoded Zinc Finger proteins (transcriptional repressors)
• Knocking down ZNF subset enhanced imatinib sensitivity
• Conclusions: genetic signature of response, functionally
associated
Gene expression signatures and response in
RTOG 0132: My Conclusions
• Prognostic versus predictive?
• Gene expression correlated with response, not survival (yet)
– However, ORR correlates strongly with OS in advanced GIST
• Importance:
– May be able to predict imatinib response in individual patient
– May have a new therapeutic target
• Further work needed (and ongoing)
– Not quite like mutant KRAS/cetuximab in metastatic CRC
– Won’t alter therapy as of now
#10534-Assessment of HSP90 inhibitor IPI-504
alone or with imatinib in mice with human GIST
xenografts: Background and Importance
• HSP90 is a “chaperone”, controlling proteins involved in cellular growth,
differentiation, and survival, especially after significant environmental
stresses
– Tumour cells are perpetually stressed and may be more dependent on HSP90 than
normal cells
• IPI-504 (retaspimycin) is a soluble form of 17-AAG, an HSP-inhibitor
• Mutant KIT and other important oncogenesis-related proteins (e.g. EGFR)
are HSP90 “clients”
• Question raised: Do we have a new GIST drug which is ready for primetime?
Assessment of HSP90 inhibitor IPI504 alone or
with imatinib in mice with human GIST
xenografts: Methods
• Sensitive human GIST cells placed nude mice
• Mice were observed, or given imatinib or IPI-504, or both
• Assessed: tumour volume, mitotic + apoptotic activity, histologic
response, KIT expression, and activation of downstream effectors
Assessment of HSP90 inhibitor IPI-504 alone or with
imatinib in mice with human GIST xenografts: Results
• IPI-504 shrunk tumours, decreased mitotic and increased apopotic
activity, and downregulated KIT expression and phosporylation of
downstream markers
• The combination modestly to markedly improved all these effects,
except apoptosis
• No discussion of toxicity in these animals
• Authors concluded “strong rationale” for exploring IPI-504 alone and
with TKI in the clinic
Assessment of HSP90 inhibitor IPI504 alone or with
imatinib in mice with human GIST xenografts: My
conclusions
• These experiments supply strong rationale for clinical testing
• Too late! Recent phase III RING trial of IPI-504 closed early
– No discussion of efficacy
• What was responsible for failure of this study?
• This drug is obviously not ready for prime-time
#10535-Molecular epidemiology of GISTs: Incidence of
PDGFR + KIT exon 9 mutations in the large French
population-based MolecGIST: Background and importance
• GISTS demonstrate molecular heterogeneity
– 75-85% have activating KIT mutations and 5-7% have activating
PDGFR mutations
• Genotype is at least a predictive factor when using TKI
– Exon 11 mutants have best ORR, TTP, and OS with imatinib
• Data are largely derived from pts with gross mets
– However, Corless has demonstrated mutations in very early GISTs
• Questions arising: Are mutation rates different in the general
population?
– Does this affect how we treat patients systemically?
MolecGIST: Methods
• Variety of physician types indicate they have a GIST patient
• Pathologist involved asked for tissue sample and data
• Demographics collated
• Mutational assessment of KIT and PDGFR performed
MolecGIST: Results and Author Conclusions
• Registry received 843 cases from 236 pathologists
• Median age, KIT expression, location
– Incidence lower than other populations (but not clear all cases were captured)
• Rate of KIT exon 9 mutation lower (5%) and PDGFR exon 18 mutation
higher (11%)
• Authors concluded latter differences are because of referral patterns
MolecGIST: My Conclusions
• This work complements other registries (reGISTry)
– Much more data on mutations
• ?More accurate than data from study patients (more
representative of general GIST population) or less (referral
bias)
• Mutational data only currently used to choose imatinib dose
– We still need to genotype more patients
Type of progression in patients treated with imatinib for
advanced GIST: A study based on EORTC-ISG-AGITG
trial 62005: Background and importance
• 62005: Phase III study front-line imatinib 400 v. 800 mg/d
– Demonstrated no difference in PFS or OS, for all-comers
• Data contributed to meta-analysis (meta-GIST) showing statistical
but not clinically meaningful difference in PFS
• Though trial showed modest benefit of crossing-over from 400 to
800 mg, little data exist on survival after first progression
• Question arising: Does clinical and laboratory information garnered
before first progression help in decision-making?
Type of progression on EORTC-ISG-AGITG
trial 62005: Methods
• Utilized 516 patients progressing on phase III 62005 trial
• Using complicated initials, pts basically categorized as new lesions (CR+,
PR+, NC+, PD+) or size old lesions (PD-) or mixed (MXPD)
• Baseline co-factors (pt characteristics, imatinib dose, standard TTP) were
correlated with type of progression and survival after progression (SAP),
using standard COX modeling and univariate/multivariate analyses
Type of progression on EORTC-ISG-AGITG
trial 62005: Results
• Type of progression: 47% progressed without new lesions (PD-);
12.6% had mixed progression
• A number of factors associated with MXPD on univariate
analysis; on multivariate only TTP and prior chemorx significant
• SAP worse in pts with new lesions with PD old lesions, short
progression-free interval, prior RT (!), factors associated with
short PFI (high neutrophils, low albumin, no mutations)
Type of progression on EORTC-ISG-AGITG
trial 62005: My conclusions
• Interesting work contributing lots of new information
• Type of progression data interesting but not applicable to
many
• SAP data somewhat intuitive
– We would expect shorter PFI to be bad
– Makes RT question even murkier
• Bottom line: Fun stuff, but I would not monitor nor treat
differently
What have we learned or confirmed today?
• GIST patients get lots of other cancers
• Correlative science crucial in GIST
• New drugs are coming
• Registry data can supplement clinical trial information
• GISTs remain very, very cool