Abstract
1535
Objectives The aim was to investigate the predictive value of Metabolic Tumor Volume (MTV) measured by 18F-FDG PET/CT for global (gLN) and para-aortic (PA LN) lymph node involvement in patients with cervix cancer (CC), as a clue of visual PET analyses. The second objective was to determine a predictive MTV threshold for gLN involvement and to analyze the risk factors associated with global lymph node metastasis (FIGO, histology, visual PET analysis, SUVmax of primary).
Methods We performed a retrospective study of 73 patients (with FIGO IB1 n= 6; IB2 n= 7; IIA n= 2; IIB n= 26; III n= 31; IV n= 1) included between April 2008 and July 2014 for CC in a single center, who underwent positron emission tomography (PET) imaging and laparoscopic para-aortic (PA) and/or pelvic lymphadenectomy. The standard uptake value (SUVmax) and volumetric parameters (MTV and TLG) of cervical tumor were measured in addition to the visual PET analysis of nodal area. All clinical, histological and imaging data were compared to the lymphadenectomy histopathologic findings. Univariate and multivariate analyses were used to analyze the associations between the pathologic gLN and PA LN status, FIGO stage, SUVmax, MTV drawn with a threshold of 40% (MTV40%) and TLG40%.
Results Forty-three patients (59%) had positive LN status (gLN+), including PA LN and/or pelvic LN. Visual PET performance analysis for gLN and PA LN were Se, 70%; Sp, 63% and Se, 56%; Sp, 91%, respectively. No statistically different values were found for SUVmax of tumor (p=0.357). The gLN+ group showed statistically significant higher values of MTV40% (p=0.005) and TLG40% (p=0.021) than those of gLN- group. The PA LN+ group showed statistically significant higher values of MTV40% (p<0,001) (median: 40mL for pathologically LN and 16mL for non pathological LN), than those of PA LN- group. MTV 蠅 25mL was highly predictive of gLN involvement. In multivariate analysis, MTV40% 蠅 25 mL combined with visual PET positive for PA LN is associated with higher risk of gLN (AUC = 0,78 (IC95% 0,67 - 0,88)) and PA LN (AUC = 0,87 (IC95% 0,77-0,96)) involvement. However, advanced FIGO stage was not associated with LN metastasis (p=0,101).
Conclusions Quantification of the primary tumor with volumetric parameters (MTV, TLG) can contribute to improve nodal involvement prediction for patients with cervix cancer.