Multivariate analysis of progression-free survival in the CLARINET study of lanreotide Autogel/Depot vs placebo identifies prognostic factors in neuroendocrine tumours

Konference: 2015 40th Congress ESMO a 18th ECCO - účast ČR

Kategorie: Gastrointestinální nádory

Téma: Postery

Číslo abstraktu: P336/2374

Autoři: MD Edward M. Wolin; Prof. M.D. Martyn E. Caplin, FRCP; Prof. Dr. Marianne E. Pavel, M.D.; Prof. M.D. Jaroslaw B. Cwikla; Prof. M.D. Alexandria T. Phan; M.D. Markus Raderer; MUDr. Eva Sedláčková, MBA; Prof. M.D. Guillaume Cadiot, Ph.D.; MD Jaume Capdevila; M.D. Lucy R. Wall ; M.D. Guido Rindi, Ph.D.; Alison Langley; Edda Gomez-Panzani; Prof. M.D. Philippe B. Ruszniewski

Background: Progression-free survival (PFS) in metastatic pancreatic and intestinal neuroendocrine tumours (NETs) was significantly increased with lanreotide Autogel (Depot in USA) 120mg vs placebo (hazard ratio [HR] for progressive disease [PD]/death 0.47 [95%CI: 0.30, 0.73]) in the CLARINET study. This was a preplanned exploratory covariate analysis to identify prognostic factors for PFS.

Methods: Patients with metastatic grade 1/2 (Ki-67 <10%) non-functioning pancreatic/intestinal NETs received lanreotide Autogel 120mg (n=101) or placebo (n=103) for 96 weeks or until PD/death (NCT00353496), with data analyses stratified according to PD (yes/no) and prior therapy for NET (yes/no) at baseline. These factors, alongside treatment, were tested here in separate Cox proportional hazards (PH) models for several baseline covariates (as listed in the table). Factors potentially significant in univariate analysis (p≤0.1, Wald chi-square test) were included in stepwise multivariate analysis to obtain a final model. Treatment interaction was investigated.

Results: With adjustment for covariates, treatment with lanreotide vs placebo reduced the risk of PD/death by 60% in the final model (Table). The adjusted effect of prior therapy was not significant, whereas PD at baseline was associated with higher risk of PD/death. Other baseline factors associated with an increased risk of PD/death were: hepatic tumor load (HTL) >25%; primary tumour type of pancreas (risk increased by 20–60% vs midgut, hindgut and other types); and below-median BMI (Table). No covariate-by-treatment interaction was significant.

Term (reference) HR [95%CI)* p-value*
Lanreotide (placebo) 0.40 [0.25,0.63] <0.0001
PD (no PD) 4.57 [1.67,12.54] 0.0032
Prior therapy (no prior therapy) 1.29 [0.72,2.31] 0.3914
HTL, % (0):   0.0005
 >0,≤10 0.81 [0.42,1.59]  
 >10,≤25 1.22 [0.59,2.52]  
 >25,≤50 2.82 [1.41,5.63]  
 >50 2.47 [1.21,5.03]  
Primary tumour type (pancreas):   0.0289
 Midgut 0.80 [0.33,1.94]  
 Hindgut 0.53 [0.32,0.88]  
 Other/Unknown 0.39 [0.17,0.86]  
BMI >median# (≤median) 0.64 [0.41,1.00] 0.0483


*From final multivariate Cox PH model; #26.2kg/m2; intent-to-treat population.

Conclusions: Lanreotide extends PFS across patient subgroups; patients without PD at baseline have the lowest risk of PD. Among other potential prognostic factors in this exploratory analysis, several covariates (including tumour grade, prior therapy and time since diagnosis) had no effect. Conversely, HTL and primary tumour type were identified as the most important prognostic factors for PFS.

Other baseline terms (sex, age, race, US/ex-US, region, time since diagnosis, Ki67, tumour grade, CgA, prior chemotherapy, prior surgery) had no prognostic value (p>0.1) either in individual models or in presence of other terms in final model.

Conflict of interest: Ownership: Authors Langley and Gomez-Panzani are employees of Ipsen.

neuroendocrine tumour
progression-free survival

Datum přednesení příspěvku: 28. 9. 2015