FDG-PET Response Prediction in Pediatric Hodgkin’s Lymphoma: Impact of Metabolically Defined Tumor Volumes and Individualized SUV Measurements on the Positive Predictive Value

Cancers

Amr Elsayed M. Hussien⁴  Christian Furth⁵  Stefan Schönberger¹  Patrick Hundsdoerfer³  Ingo G. Steffen⁵  Holger Amthauer⁵  Hans-Wilhelm Müller⁴  Hubertus Hautzel⁴  Georg Lenz² 

[1] Department of Pediatric Oncology, Hematology and Clinical Immunology, University Children’s Hospital, Medical Faculty, Heinrich-Heine-University Düsseldorf, Düsseldorf, 40225, Germany; E-Mail:Department of Nuclear Medicine (KME), Forschungszentrum Jülich, Medical Faculty, Heinrich-Heine-University Düsseldorf, Jülich, 52426, Germany;;Department of Pediatric Oncology and Hematology, Charité Campus Virchow, Humboldt-University Berlin, Berlin, 13353, Germany; E-Mail:;Department of Nuclear Medicine (KME), Forschungszentrum Jülich, Medical Faculty, Heinrich-Heine-University Düsseldorf, Jülich, 52426, Germany; E-Mails:;Department of Radiology and Nuclear Medicine, Medical School, Otto-von-Guericke University Magdeburg, Magdeburg, 39120, Germany; E-Mails:

关键词: FDG-PET;    pediatric Hodgkin lymphoma;    response assessment;    SUV;   

DOI  :  10.3390/cancers7010287

来源: mdpi

【 摘 要 】

Background: In pediatric Hodgkin’s lymphoma (pHL) early response-to-therapy prediction is metabolically assessed by (18)F-FDG PET carrying an excellent negative predictive value (NPV) but an impaired positive predictive value (PPV). Aim of this study was to improve the PPV while keeping the optimal NPV. A comparison of different PET data analyses was performed applying individualized standardized uptake values (SUV), PET-derived metabolic tumor volume (MTV) and the product of both parameters, termed total lesion glycolysis (TLG); Methods: One-hundred-eight PET datasets (PET1, n = 54; PET2, n = 54) of 54 children were analysed by visual and semi-quantitative means. SUVmax, SUVmean, MTV and TLG were obtained the results of both PETs and the relative change from PET1 to PET2 (Δ in %) were compared for their capability of identifying responders and non-responders using receiver operating characteristics (ROC)-curves. In consideration of individual variations in noise and contrasts levels all parameters were additionally obtained after threshold correction to lean body mass and background; Results: All semi-quantitative SUV estimates obtained at PET2 were significantly superior to the visual PET2 analysis. However, ΔSUVmax revealed the best results (area under the curve, 0.92; p < 0.001; sensitivity 100%; specificity 85.4%; PPV 46.2%; NPV 100%; accuracy, 87.0%) but was not significantly superior to SUVmax-estimation at PET2 and ΔTLGmax. Likewise, the lean body mass and background individualization of the datasets did not impove the results of the ROC analyses; Conclusions: Sophisticated semi-quantitative PET measures in early response assessment of pHL patients do not perform significantly better than the previously proposed ΔSUVmax. All analytical strategies failed to improve the impaired PPV to a clinically acceptable level while preserving the excellent NPV.

【 授权许可】

CC BY   
© 2015 by the authors; licensee MDPI, Basel, Switzerland.