Abstract
1307
Objectives This study was to see whether primary tumor SUVs can predict ALNM and to determine optimal SUV cutoff values for ALNM according to molecular subtypes.
Methods A total of 239 female patients (average age of 55.9±11.88 yr) who underwent 18 F-FDG PET/CT at staging and surgery for initial treatment were included. The presence of ALNM was based on pathological specimen. Molecular subtypes were divided into Luminal A (Lum A), Luminal B (Lum B), HER2, and Triple negative (TN). SUVmax of the primary tumors and axillary lymph nodes were measured. Receiver operating characteristic (ROC) analyses were performed to determine the best cutoff SUVmax values for ALNM in each subtype.
Results There were 86 patients with Lum A, 83 with Lum B, 34 with HER2, and 36 with TN in this study. Of the molecular subtypes, only HER2 group showed a statistically significant difference in primary tumor SUVmax between patients with ALNM (10.5±3.5) and patients without ALNM (7.9±3.7) (p=0.031). The mean SUVmax of ALNM was 1.2±0.6 in Lum A, 1.6±1.7 in Lum B, 5.0±4.9 in HER2, and 6.5±13.6 in TN. The mean SUVmax in ALNM was significantly lower in Lum A (p=0.006) or B (p=0.01) compared to that in HER2 group. Upon ROC analyses, in Lum A with a cutoff SUVmax of 0.6, the area under the curve (AUC), sensitivity, and specificity for ALNM were 0.845, 96.6%, and 57.8% (p<0.0001). In Lum B with a cutoff of 1.0, the values were 0.719, 50.0%, and 85.45%, respectively (p<0.0002). In HER2 group with a cutoff of 1.0, AUC was 0.856 with a sensitivity of 91.6% and a specificity of 72.7% (p<0.0001). In TN, there was no difference in SUVmax of axillary LNs between patients with metastases and patients without metastases.
Conclusions Only in HER2 group, the SUVmax of primary tumors was significant in predicting the presence of ALNM. There were significant differences in mean SUVmax of ALNM between HER2 and Luminal A or B group. Accordingly, different SUVmax cutoff values should be used for the diagnosis of ALNM based on molecular subtypes.