Institutional members access full text with Ovid®

Lymph Node to Primary Tumor SUV Ratio by 18F-FDG PET/CT and the Prediction of Axillary Lymph Node Metastases in Breast Cancer

Park, Jihyun MD*; Byun, Byung Hyun MD*; Noh, Woo Chul MD; Lee, Seung Sook MD; Kim, Hyun-Ah MD; Kim, Eun-Kyu MD; Choi, Chang Woon MD*; Lim, Sang Moo MD, PhD*

doi: 10.1097/RLU.0b013e3182a75477
Original Articles

Purpose: The authors evaluated the usefulness of axillary lymph node (ALN) to primary breast tumor SUV ratio (determined by 18F-FDG PET/CT) for predicting the presence of ALN metastasis in breast cancer.

Methods: One hundred thirty-six consecutive female patients with breast cancer were enrolled in this retrospective study between January 2009 and November 2012. All patients underwent surgical resection without neoadjuvant chemotherapy, and ALN metastases were histologically confirmed by ALN dissection (n = 75) or sentinel lymph node (LN) biopsy (n = 61). The maximum SUVs of FDG-avid ALNs (SUVLN) and of primary breast tumors were measured on preoperative 18F-FDG PET/CT images, and ALN to primary breast tumor SUV ratios (LN/T ratios) were calculated. In a subgroup of patients with FDG-avid ALNs, optimal cutoff values for SUVLN and LN/T ratio were determined by receiver operating characteristic curve analysis for predicting the presence of ALN metastasis. Subsequently, the diagnostic performances of visual analysis (presence of FDG-avidity), SUVLN, and LN/T ratio for the prediction of ALN metastasis were determined.

Results: In a subgroup of patients with FDG-avid ALNs (n = 65), the area under the curve and the optimal criteria of SUVLN for detecting ALN metastasis were 0.655 and greater than 2.1, and those of LN/T ratio were 0.739 and greater than 0.2, respectively. For these criteria, the sensitivity, specificity, and diagnostic accuracy of detecting ALN metastasis were 71.4%, 77.3%, and 74.3%, respectively, for visual analysis; 47.1%, 93.9%, and 69.9%, respectively, for SUVLN; and 62.9%, 92.4%, and 77.2%, respectively, for LN/T ratio in all patients. The specificity of LN/T ratio was significantly higher than that of visual analysis (P = 0.0259). Although the sensitivity of LN/T ratio was higher than that of SUVLN, it did not reach a statistical significance (P = 0.0874).

Conclusions: The LN/T ratio better predicts the presence of ALN metastasis than visual analysis or SUVLN in breast cancer.

From the Departments of *Nuclear Medicine, †Surgery, and ‡Pathology, Seoul Korea Cancer Center Hospital, Seoul, Republic of Korea.

Received for publication March 25, 2013; revision accepted June 4, 2013.

Conflicts of interest and sources of funding: This work was supported by Establishment of Center for PET Application Technology Development, Korea Institute of Radiological and Medical Sciences (KIRAMS), and by grants from the Ministry of Education, Science and Technology (50441-2013).

Reprints: Sang Moo Lim, MD, PhD, Department of Nuclear Medicine, Korea Cancer Center Hospital, 75 Nowongil, Nowon-Gu, Seoul, 139-706, Republic of Korea. E-mail address:

© 2014 by Lippincott Williams & Wilkins