Better Regimens & Predictive Markers
For Advanced Esophagogastric Cancer?
Peter C. Enzinger , M.D.Dana-Farber Cancer Institute/Harvard Medical School
Posters DiscussedKishimoto et al. Does Paclitaxel or Irinotecan improve a Japanese
Standard – S1?-----------------------------------------------------------------Shah et al. Is “Docetaxel, Cisplatin + mod de Gramont” better than
DCF?
Al-Batran et al.Is “Docetaxel + mFOLFOX” better than DCF?-----------------------------------------------------------------
Atmaca et al. What do we do with a new prognostic molecular factor?
Evolution of S1 Therapy in Gastric CA
FLAGS**
Cisplatin + S1
Cisplatin + 5-FU
29% 8.6 mos
32% 7.9 mos
P = NS
SPIRITS*
S1
Cisplatin + S1
*Koizumi. Lancet Oncol 2008**Ajani. J Clin Oncol 2010
31% 11 mos
54% 13 mos
P = 0.04Japan
“West”
RR: OS:Less toxic
Random Phase II: S1+ CPT or S1 + Pacli
•Arm B: 3 weeks / course
Day1
80 mg/m2/day x 2 wks, p.o.
Paclitaxel(50 mg/m2, iv)
S-1 1-week rest
•Arm A: 5 weeks / course
Day8 Day14 Day21
Day1
80 mg/m2/day x 3 wks, p.o.
CPT-11(80 mg/m2, iv)
S-1 2-weeks rest
Day15 Day21 Day35
•Unresectable, measurable advanced or recurrent gastric cancer
•No prior chemotherapy except adjuvant CTX
H0: 30% RR vs. HA: 50% RRpower 80%; 2-sided α 5%
Phase III:START
0 200 400 600 800 1000 12000
20
40
60
80
100
Time
Su
rviv
al r
ate
S-1+CPT
S-1+paclitaxel
S-1+CPT
S-1+paclitaxel
(%)
(days)
Progression-Free Survival Overall Survival
S-1+CPT
S-1+paclitaxel
S-1+CPT
S-1+paclitaxel
0 200 400 600 800 1000 1200 14000
20
40
60
80
100
Surv
ival
rat
e
(%)
(days)
HR=1.18; p=0.42 HR=0.99; p=0.96
Regimen Response PFS OS
S1 alone*/# 28-31% 4.2 mos 11.0-11.4 mos
S1 + Pacli 31% 4.6 mos 12.0 mos
S1 + Doce** 88% 6.5 mos 12.0 mos
S1 + CPT 33% 5.7 mos 12.5 mos
S1 + CPT*** 42% N/A 12.8 mos (IRIS)
S1 + CDDP* 54% 6.0 mos 13.0 mos*Koizumi. Lancet Oncol 2008#Boku. Lancet Oncol 2009**Ikeda. Proc ASCO 2009***Imamura. Proc ASCO 2008
ConclusionsThe authors conclude that neither S1+ paclitaxel or S1 +
irinotecan should be developed further – I agree!
S1 seems to be happiest by itself … or perhaps with cisplatin.
Where are the biologics – there seem to be no combinations with S1?
DCF: What do you do with an active regimen that causes 82% G3-4 ANC and 81% G3-4 non-heme toxicity?
Dose and Schedule
Author Regimen Site Pts. RR ANC Non-heme OS
Van Cutsem.
J Clin Oncol ‘06
DCF
original
Gastric 221 37% 82% Stomatitis
21%
9.2mo.
Park. Am J Clin Oncol’05
Low dose D + C F
Gastric 47 40% 68% Stomatitis 21%
9.7mo.
Lorenzen. Ann Oncol 2007
DC q 2wks
+ AIO
GEJ
Gastric
60 47% 22% Diarrhea
20%
15.1mo
Overman. Cancer 2010
Weekly
DCF
retrospective
Esoph
Gastric
95 34% 4% Hospital
23%
8.9mo
Tebbutt. Br J Cancer 2010
Weekly D
q 3 wk CF
Esoph
Gastric
50 47% 10% Diarrhea
22%
11.2mo
DCF: What do you do with an active regimen that causes 82% G3-4 ANC and 81% G3-4 non-heme toxicity?
Substitutions
Author Regimen Site Pts. RR ANC Non-heme OS
van Cutsem.
J Clin Oncol ‘06
DCF
original
Gastric 221 37% 82% Stomatitis
21%
9.2mo.
Shankaran.
GI Symp’09
Docetaxel
Oxali /5-FU
GEJ
Gastric
35 74% 18% Neuro
21%
10.3mo
Moehler.
ASCO’09
Docetaxel
Oxali/Cape
Gastric 25 36% 3% Diarrhea
23%
N/A
Sato.
ASCO’08
D C S1 Gastric 31 87% 52% N/V 27% 18.9mo.
Enzinger.
Ann Oncol ‘09
D C Cpt
weekly
Esoph
Gastric
56 54% 21% Diarrhea 26%
11.9mo.
Conclusions
DCF is too toxic – worst way of giving either cisplatin or fluorouracil
Smaller, more frequent doses or substitutions are better but which one? or both?
Randomized trials are needed to move this forward:
DCF: What do you do with an active regimen that causes 82% G3-4 ANC and 81% G3-4 non-heme toxicity?
Dose and Schedule Substitutions(Shah) (Al-Batran)
FLO (mFOLFOX)5-FU 2600 mg/m² Folinic acid 200 mg/m² Oxaliplatin 85 mg/m²Every two weeks
FLOTDocetaxel 50mg/m2
5-FU 2600 mg/m² Folinic acid 200 mg/m² Oxaliplatin 85 mg/m²Every two weeks
RANDOMIZATION
DCF (original)Docetaxel 75mg/m2 Cisplatin 75mg/m2
5-FU 750mg/m2/d CI x 5dEvery three weeks
mDCFDocetaxel 40mg/m2
Cisplatin 40 mg/m², day 3Leucovorin 400 mg/m²5-FU 400 IVP5-FU 1000 mg/m²/d CI x 2d Every two weeks
RANDOMIZATION
DCF: What do you do with an active regimen that causes 82% G3-4 ANC and 81% G3-4 non-heme toxicity?
Dose and Schedule Substitutions(mDCF) (FLOT)
Disease site GEJ, Gastric GEJ, Gastric
Stage inop, metastatic
inop, metast, loc adv, recurrent
Response Rate
52% 49% (incl loc adv)
6-month PFS
63% 60% (inop/metastatic)
Overall Survival
15.1 mo (9.7 – 20.4)
est. 17mo (incl loc adv)
Neutropenia 54% 52%
Nausea 0% 21%
Diarrhea 2% 9%
Neuropathy 2% 20%
Fatigue 10% 11%
ConclusionsHard to choose between the two; mDCF perhaps
less toxic but more complicated
mDCF arm is still in progress and must wait for 6 mo PFS primary endpoint
FLOT– higher RR and toxicity more suited for adjuvant comparison
Randomization against ECF would require large pt numbers for significant difference
Matrix Metalloproteinase-9mRNA expression
What do we do with a new prognostic molecular factor in esophagogastric ca?
Other Molecular Prognostic Markers
“Hundreds” of prognostic molecular markers for gastric cancer in the literature …
HER2/neu – prognostic and antibody is effective
VEGF – prognostic – AVAGAST on Monday
EGFR – prognostic REAL 3 and EXPAND
Impressive Results!
months
0
0,2
0,4
0,6
0,8
1
0 5 10 15 20 25 30
pro
babilit
y
<best cut-offbest cut-off
P=0,00765
p=0,0193
0
0,2
0,4
0,6
0,8
1
0 10 20 30 40OS months
Quartile 1Quartile 2Quartile 3Quartile 4
p=0,0193
0
0,2
0,4
0,6
0,8
1
0 10 20 30 40OS months
Quartile 1Quartile 2Quartile 3Quartile 4
Median OS: 5,5 mo vs. 11,9 mo1 yr Survival: 25% vs. 50%2 yr Survival: 0% vs. 26%
Validation Cohort Pooled OS by Quartile
MMP-9 effect only in highest quartile
Independant by multivariate analysis
Questions?Training & Validation Cohorts in “cisplatin-based” therapy – same results
in non-platinum regimens or Asian patients?
Are there agents that target MMP-9? - ATRA*? Epigallocatechine-3-gallate**?
If interacts with angiogenesis or EGFR pathway, what would be the effect in bevacizumab or anti-EFGR studies?
Should one change treatment based on MMP9 status? More aggressive for high? What about adjuvant?
*Dutta. Cell Adh Migr. 2010**Farabegoli. Biosci Rep. 2010
Next steps …
1st step is for another group to run a validation cohort (West and East Asia)
Ideal cohorts to test interaction with MMP-9 would be the AVAGAST, REAL 3, and EXPAND studies
Too early for treatment changes – everybody gets max tolerated chemo for metastatic dz and in the adjuvant setting anyway.