EEWOWW is an online personal knowledge management tool. Click to find out more!
 
Updated in 2018/12/11 下午 06:47:08      Viewed: 121 times      (Journal Article)
European journal of haematology 86 (4): 289-98 (2011)

Combining FDG-PET/CT with laboratory data yields superior results for prediction of relapse in multiple myeloma.

Brian M Elliott , Steven Peti , Keren Osman , Eileen Scigliano , David Lee , Luis Isola , Lale Kostakoglu
ABSTRACT
  The precise role of positron emission tomography (PET/CT) for predicting relapse/progression in multiple myeloma remains uncertain. We compared the predictive values of PET/CT, concurrent laboratory testing (labs), and their combination in prediction of 12-month progression, as determined by current International Myeloma Working Group (IMWG) criteria.  PET/CT and labs (serum chemistry, β2-microglobulin, immunofixation, bone marrow biopsy, serum free light chains) were reviewed, and date of relapse/progression was determined by IMWG criteria.  The median time from therapy to PET/CT imaging was 12.0 months (1.0-110) and median time to progression (TTP) was 29.8 months (1.6-130+). Overall survival and survival-without-progression at last follow-up were 84% and 49%, respectively. Sensitivity of PET/CT for predicting relapse/progression was lower than that of labs (0.67 vs. 0.89, ns), but PET/CT was more specific (0.89 vs. 0.79, ns). When labs and PET/CT data were combined, a positive result for either test was 89% sensitive and a positive result for both tests was 100% specific for predicting 12-month progression of disease. Kaplan-Meier analysis showed significantly greater TTP for those with a negative vs. positive PET/CT (P = 0.0005), negative vs. positive labs (P < 0.0001), and both tests negative vs. both tests positive (P < 0.0001).  Combining PET/CT with laboratory data improves the accuracy of prediction of relapse/progression within 12 months compared with each test alone. Thus, integration of PET/CT into myeloma follow-up is recommended, and the impact of this approach on management should be explored.
DOI: 10.1111/j.1600-0609.2010.01575.x      ISSN: 0902-4441