Proper scoring rules with general preferences: A dual characterization of optimal reports

2019 ◽  
Vol 117 ◽  
pp. 322-341
Author(s):  
Christopher P. Chambers ◽  
Paul J. Healy ◽  
Nicolas S. Lambert
Author(s):  
Bruno de Finetti ◽  
Maria Carla Galavotti ◽  
Hykel Hosni ◽  
Alberto Mura

2019 ◽  
Vol 180 ◽  
pp. 244-273 ◽  
Author(s):  
Piotr Skowron ◽  
Piotr Faliszewski ◽  
Arkadii Slinko

2018 ◽  
Vol 6 (3-4) ◽  
pp. 343-376 ◽  
Author(s):  
Arthur Carvalho ◽  
Stanko Dimitrov ◽  
Kate Larson

2015 ◽  
Vol 10 (2) ◽  
pp. 479-499 ◽  
Author(s):  
A. Philip Dawid ◽  
Monica Musio

2009 ◽  
Vol 76 (4) ◽  
pp. 1461-1489 ◽  
Author(s):  
THEO OFFERMAN ◽  
JOEP SONNEMANS ◽  
GIJS VAN DE KUILEN ◽  
PETER P. WAKKER

2013 ◽  
Vol 31 (4_suppl) ◽  
pp. 427-427 ◽  
Author(s):  
Ulrich Robert Mansmann ◽  
Ute Sartorius ◽  
Ruediger Paul Laubender ◽  
Clemens Albrecht Giessen ◽  
Regina Esser ◽  
...  

427 Background: The extent of tumor shrinkage in patients (pts) receiving chemotherapy +/- monoclonal antibodies has prognostic value for PFS and OS. "Deepness of response (DpR)" is a new efficacy outcome measure that could explain the impact of tumor shrinkage on long-term survival outcome. If shrinkage takes place DpR is the percentage of tumor shrinkage observed at the nadir compared to baseline. DpR is 0 for no change and negative if the tumor load increases. Longest diameter (LD) based on RECIST or a calculated tumor volume (ASCO GI 2012 #635) can quantify the tumor load at distinct time points. A joint model was presented (ASCO GI 2012 #580, ASCO 2012 #3603) which allows us to relate DpR to individual post-progression survival (PPS) time. Methods: Based on the data from 2 randomized trials (CRYSTAL, n=1198; OPUS, n=337), 4 treatment regimens (FOLFIRI +/- cetuximab and FOLFOX4 +/- cetuximab) were studied. A joint model was used to quantify individual changes in tumor size over time and to relate these changes to PFS and OS. Relationships between baseline tumor load and DpR and PPS were studied. Proper scoring rules were used to assess whether the LD-based or the volume-based approach allowed a better prediction of individual prognosis. Results: Results are reported for the CRYSTAL study using LD-based measures for 663 pts with KRAS wild-type tumors and imaging data. The 348 pts treated with FOLFIRI alone had a mean DpR of 35.52% (Interquartile range [IR]:12.09%, 59.86%), minimum DpR -80%. The 315 pts treated with FOLFIRI + cetuximab had a mean DpR of 50.07% (IR: 22.87%, 79.55%) and a minimum DpR of -49%. The DpR was significantly different between the 2 groups (p<0.00001). Individual DpR is a significant prognostic factor for PPS time in both the LD-based (p=0.0023) and volume-based (p=0.0003) models. Proper scoring rules provided evidence of a more precise estimation of individual PPS time based on volume algorithm-measured DpR. Results of the OPUS study will be presented. Conclusions: Our results emphasize the value of the variable DpR as a new efficacy outcome measure for clinical trials. The tumor-shrinking capacity of cetuximab was shown to be associated with its ability to prolong PPS.


Sign in / Sign up

Export Citation Format

Share Document