Article Text

Download PDFPDF
DI-026 Overall survival benefit with pomalidomide: area under curves-based reanalysis
  1. J Diaz-Navarro,
  2. EJ Alegre-DelRey,
  3. S Fenix-Caballero,
  4. MA Blanco-Castaño,
  5. MJ Gandara-Ladron de Guevera,
  6. C Palomo-Palomo,
  7. JC Garciade Paredes-Esteban,
  8. E Rios-Sanchez,
  9. C Martinez-Diaz,
  10. JM Borrero-Rubio
  1. Hospital Universitario Puerto Real, Hospital Pharmacy, Puerto Real (Cádiz), Spain


Background The pomalidomide pivotal clinical trial was recently published in Lancet Oncology (PubMed PMID:24007748). Pomalidomide shows a significant benefit on overall survival (OS), with a difference between medians of 4.6 months (12.7 vs. 8.1). However, difference in median survival (DMS) is erratic and sometimes doesn’t provide a good assessment of survival benefit.

Purpose To reanalyse pomalidomide OS benefit from pivotal clinical trial using an area under the curve (AUC)-based method.

Materials and methods Kaplan-Meier OS curves were extracted from the pomalidomide pivotal clinical trial. A graphical AUC method was applied to pomalidomide + low-dose dexamethasone vs. high-dose dexamethasone curves and compared to DMS.

  • Using the AUC method, three lines were defined for pomalidomide OS graph: V, H and T. A vertical cutting line (V) intersects the abscissa at the longest time (t) with at least 10 patients at risk in each group or 30 in total. A horizontal cutting line (H) intersects the point where V and the upper curve cross. AUC was defined between the ordinate axis, the curve and H. A reference area (RA) was defined as the rectangular area between the ordinate axis, V, H and T. It represents the survival time in the event that no patients died (t). The AUC method quantifies the difference between areas, relates them to a RA and the results are expressed in units of time.

  • Survival reanalysis was calculated as (AUC/RA)*t for each curve. Photoshop CS6 was used for graphical AUC calculation.

Results The V-line intersects the abscissa with 31 patients at risk for a “t” equal to 16 months. With 61.5% patients included in this AUC-based analysis (the H line), pomalidomide + low-dose dexamethasone AUC was 91602 pixels, high-dose dexamethasone AUC, 59240 pixels and RA, 196968 pixels. OS reanalysis and comparison to DMS are shown in the table below.

Abstract DI-026 Table 1

Conclusions The pomalidomide OS AUC-based method gives a value 2 months less than DMS and could have relevant implications for pomalidomide evaluation, positioning and cost utility in clinical practice.

No conflict of interest.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.