Predictive value of FDG PET/CT for pathologic axillary node involvement after neoadjuvant chemotherapy
- First Online:
- Cite this article as:
- Keam, B., Im, S., Koh, Y. et al. Breast Cancer (2013) 20: 167. doi:10.1007/s12282-011-0323-0
The purpose of this study was to determine the usefulness of sequential FDG PET/CTs for prediction of axillary lymph node (ALN) status after neoadjuvant chemotherapy (NAC).
Seventy-seven stage II or III breast cancer patients who received 3 cycles of neoadjuvant docetaxel/doxorubicin chemotherapy were enrolled in this prospective study. FDG PET/CTs were acquired before chemotherapy and after the first cycle of chemotherapy for early metabolic response prediction.
Patients with pN0 had significantly lower post-NAC ALN standard uptake value (SUV) than those who were pN+ (1.22 ± 1.46 in pN0 vs. 2.13 ± 1.99 in pN+, P = 0.017). Post-NAC ALN size on CT also differed according to pathologic ALN status (6.3 mm in pN0 vs. 11.1 mm in pN+, P = 0.014). When serial FDG PET/CT and chest CT were used, patients with an SUV > 1.5 and post-NAC ALN size ≥10 mm on CT did not achieve pN0 (specificity 100% and positive predictive value 100%).
The serial FDG PET/CT after NAC could predict the pathologic status of ALN before surgery in stage II/III breast cancer. Our findings suggest that the combined use of serial FDG PET/CTs and chest CT might provide better information regarding ALN before surgery.
Keywords18F-FDG PETBreast cancerNeoadjuvant chemotherapyAxillary node
It is important to evaluate axillary lymph node (ALN) accurately before treatment of primary breast cancer because ALN status has been repeatedly confirmed to be the single most important prognostic factor for breast cancer . In locally advanced breast cancer treated with neoadjuvant chemotherapy (NAC), pathologic ALN status differs from those of early breast cancer treated by conventional surgery . NAC might alter the yield of involvement of ALNs. However, the loss of prognostic value provided by tumor size and nodal status remains an important disadvantage of NAC [3, 4].
Many surgeons have felt difficulty in determining the optimum extent of ALN dissection in the neoadjuvant setting. In particular, routine performance of ALN dissection for patients with a low probability of ALN positivity could cause unnecessary side effects, for example lymph edema and numbness of the arm . Hence, a sentinel lymph node biopsy (SLNB), which can minimize the extent of ALN dissection and reduce the morbidity of axillary surgery for early breast cancer patients [6, 7], has been attempted even in locally advanced breast cancer, even though the role of SLNB in the neoadjuvant setting is still controversial [8–11].
Several studies have been conducted to identify the role of fluorine-18 fluorodeoxyglucose positron emission tomography (FDG PET) in predicting ALN status [12–19], and initial pre-operative axillary staging using FDG PET might give additional information for determining SLNB in early breast cancer [14–18]. The specificity of PET has been consistently high across studies, ranging from 85 to 100% . However, these data were obtained using single acquisition of FDG PET before surgery in patients with early breast cancer, and limited data using serial FDG PET for evaluating ALN status in a neoadjuvant setting are available. The purpose of this study was to determine the usefulness of FDG PET/CTs for prediction of pathologic ALN status after NAC.
Materials and methods
Study population and treatment
Seventy-eight stage II or III breast cancer patients were enrolled in this prospective clinical trial between July 2006 and September 2008 (ClinicalTrials.gov number, NCT01396655). For these 78 patients, we have reported the role of FDG PET/CT in predicting early metabolic response and differences by subtypes . Subsequently we re-analyzed the data for nodal evaluation. We excluded one patient who was cN0 stage, so 77 patients were analyzed. The detailed eligibility criteria were described in our previous reports [21–23]. In brief, the eligibility criteria were:
breast cancer pathologically-confirmed by core needle biopsy;
initial clinical stage II or III;
objective measurable lesion;
ECOG performance status 0–2; and
The patients received 3 cycles of neoadjuvant docetaxel/doxorubicin chemotherapy. The regimen consisted of docetaxel (75 mg/m2) and doxorubicin (50 mg/m2) by intravenous infusion every 3 weeks.
The FDG PET/CTs of all patients were carried out using the same scanner (Gemini PET/CT system; Philips, Milpitas, CA, USA). Patients fasted for at least 6 h before intravenous injection of fluorine-18 FDG (5.18 MBq/kg). All patients were studied in the supine position with both arms raised above the head to pull their breast away. The CT scan settings were: tube voltage, 120 kV; current intensity, 50 mAs; scan time, 43.2 s; effective radiation dose, 4.8 mSv; table speed, 0.75 s/rotation; transaxial FOV, 56.5 cm; axial FOV, 18 cm. Patients were administered a weight-adjusted dose of fluorine-18 FDG (5.18 MBq/kg), and images were acquired approximately 60 min (range 50–75 min) after intravenous injection of fluorine-18 FDG. Whole-body emission scans were obtained for 2 min per bed position. PET/CT scanners automatically calculated the decay-corrected injected activity. We flushed the syringe and venous catheter three times, and residual activity in the syringe was less than 20 μCi, and therefore negligible. Attenuation correction was based on the CT data, and PET data reconstruction was achieved by use of a three-dimensional row action maximum likelihood algorithm. For quantitative assessment of tumor FDG uptake, regions of interest were manually drawn on the slice of the primary breast tumor with the highest radioactivity concentration and on the adjacent slices. The region of interest was manually drawn on the slice with the highest radioactivity concentration.
The standard uptake values (SUVs) were calculated from the amount of FDG injected, body weight, and soft tissue uptake in the attenuation-corrected regional images, as follows: SUV = (activity/unit volume)/(injected dose/body weight). We measured the maximum SUV of the ALN found on FDG PET/CT images. The pre-treatment maximum SUV of the ALN (pre-ALN SUV) and the post-treatment SUV of the ALN (post-ALN SUV) were compared with pathologic ALN status. The metabolic response of FDG PET/CT was evaluated from the relative change in SUVs before and after chemotherapy. Calculation of the uptake index was as follows: ΔSUV% = 100 × (pre-ALN SUV − post-ALN SUV)/pre-ALN SUV.
Radiologic and pathologic assessment
For precise evaluation of the ALNs, we acquired chest CTs twice, before and after three cycles of NAC (Fig. 1). The initial clinical stage and the post-NAC pathologic stage were evaluated on the basis of the AJCC (6th edition) . Radiologic response was evaluated using both breast MRI and chest CT, using Response Evaluation Criteria In Solid Tumors (RECIST) criteria . At the baseline chest CT, ALNs with a long axis ≥10 mm were measured and summed according to RECIST criteria , because CT slices were 5 mm thick. We performed immunohistochemistry using tissues obtained before treatment; details of pathologic evaluation have been described in our previous reports [23, 28].
Dichotomous variables were compared by use of a chi-squared test or Fisher’s exact test, where appropriate. The Mann–Whitney U test was used to compare SUVs between different groups. Receiver-operating characteristics (ROC) analysis was performed to determine a cut-off for the hold for prediction of pathologic nodal status. All statistical tests were two-sided, with the level of significance established at P < 0.05. SPSS software (SPSS, Chicago, IL, USA) was used for all statistical analysis.
Baseline characteristics for the 77 patients
No. of patients (%)
Median age (range)
45 (range 31–69)
Invasive ductal carcinoma
Initial clinical stage
Inflammatory breast cancer
Type of surgery
Level of axillary node dissection
Adjuvant hormonal therapy
Adjuvant radiation therapy
Pathologic complete response
Nodal status of the patients
No. of patients (%)
Pre-NAC initial clinical nodal staging
Post-NAC pathologic nodal status
Number of involved nodes
Axillary SUV on PET/CT
Pre-NAC axillary SUV
Median 3.60 (range 0.00–23.20)
Post-NAC axillary SUV
Median 1.50 (range 0.00–8.90)
Correlation between pathologic nodal status and serial FDG PET/CTs
pN+ (no PNCR)
(N = 21)
(N = 56)
4.15 ± 3.91
4.54 ± 3.83
1.22 ± 1.46
2.13 ± 1.99
60.7 ± 34.5
46.6 ± 44.4
Pre-node size on CT (mm)
18.1 ± 7.9
22.1 ± 11.3
Post-node size on CT (mm)
6.3 ± 5.8
11.1 ± 7.5
65.5 ± 28.7
49.0 ± 28.6
The correlation between pathologic nodal status and SUV with different cut-off levels
pN0 (n = 21)
pN+ (n = 56)
Post-ALN SUV cut-off
SUV = 0
SUV > 0
SUV ≤ 0.5
SUV > 0.5
SUV ≤ 1.0
SUV > 1.0
SUV ≤ 1.5
SUV > 1.5
SUV ≤ 2.0
SUV > 2.0
Post- node size on CT
Combined use of CT and PET
<10 mm or SUV ≤ 1.5
≥10 mm and SUV > 1.5
In this study, we found that serial FDG PET/CT after NAC could enable prediction of pathologic status of ALN before surgery for patients with stage II/III breast cancer. Patients with a post-NAC ALN SUV > 1.5 and a post-NAC ALN size ≥10 mm on CT did not achieve pN0; this finding can provide additional information enabling surgeons to determine the extent of ALN dissection before surgery.
Currently, high accuracy has been reported for ALN staging with FDG PET. The specificity of FDG PET has been consistently high across studies, ranging from 85 to 100%; however, the sensitivity was lower and broader, ranging from 20 to 94% [12, 13, 15, 20, 29–31]. In this study, given an axillary SUV cut-off level of 1.5, the sensitivity, specificity, NNP, and PPV were 51.8, 85.7, 40.0, and 91.3%, respectively; the high specificity and PPV are in agreement with other studies. However, there were more false negative cases than expected. Twenty-seven patients (37.0%) who were PET-negative but pathology-positive were classified as false negatives. This phenomenon might be because of less aggressive tumor biology with a corresponding decrease in glycolytic activity. Microscopic minimum residual cancer cells in ALN could not express detectable level of glycolytic activity by current FDG PET/CT resolution. The timing of obtaining the 2nd FDG PET/CT would also be ascribed to high false negative rate.
Given the high specificity and PPV, sequential FDG PET/CT may aid identification of patients likely to have positive pathologic nodes before surgery. With the combined use of conventional CT and FDG PET/CT, specificity and PPV increased and both reached 100%. This finding suggests that patients with a post-NAC ALN SUV > 1.5 and a post-NAC ALN size ≥10 mm on CT could avoid SLNB or minimum ALN dissection, and these patients would benefit from adequate ALN dissection. Given low sensitivity and NPV, FDG PET/CT seems inappropriate for identification of patients who would benefit from SLNB, and this is a shortcoming of our study. Certainly, there is abundant controversy regarding the use of SLNB after NAC, because of the lack of a large randomized prospective trial. It is not yet certain whether or not we can omit axillary surgery after NAC. It is, however, important to select inappropriate or appropriate candidates for SLNB or minimum axillary dissection.
FDG PET-guided SLNB or ALN dissection for breast cancer patients is still controversial [8, 14, 18, 19, 32]. Veronesi et al.  reported that given the high specificity of FDG PET, patients who have PET-positive axilla should have an ALN dissection rather than an SLNB. Kim et al.  reported the clinical usefulness of pre-operative FDG PET/CT as a guide for ALD or SLNB in a prospective study. This study suggested that identification of patients from the results of pre-operative FDG PET/CT avoids an unnecessary SLNB and positive axillary basin. However, Ueda et al.  reported the limited value of FDG PET/CT for detecting axillary metastasis and selecting candidates for SLNB. They investigated baseline FDG PET/CT results for 183 operable breast cancer patients and the diagnostic accuracy of FDG PET/CT was shown to be nearly equal to that of ultrasound. The general opinion is that the technique is not sufficiently accurate for use in place of axillary node sampling for routine staging of axillary involvement . To date, information about the utility of SNLB and FDG PET in the neoadjuvant setting is limited. A large prospective study involving SLNB and FDG PET, especially in the neoadjuvant setting, is warranted.
This study had some limitations. First, because of the relatively short follow-up duration (median, 18.7 months), we could not evaluate the correlation between survival and axillary nodal status provided by FDG PET/CT. If the duration of follow-up is prolonged, the correlation with survival may give us additional information regarding the prognostic value of FDG PET/CT and enhance the clinical interpretation of FDG PET/CT. Second, we performed only three cycles of NAC, which led to a lower pCR rate than other studies which used conventional 6 or 8 cycles of NAC. This low pCR rate is probably associated with a low pathologic axillary nodal complete response in our study. If pCR rate is increased by more cycles of NAC, the predictive value of SUV might be changed. Third, we did not routinely perform fine needle aspiration for axillary node before NAC; this could be limitation of our study. Fourth, reactive lymphadenopathy provoked by a general inflammatory response or recent breast biopsy would increase FDG uptake of the ALNs [30, 33, 34]. Optimum cut-off of SUV for determining positivity on PET should be standardized.
However, our study had many strengths. The study was a well-designed, prospective study, and the patient population was highly homogenous in terms of treatment, with a large sample size (N = 77). SUVs were obtained by the same FDG PET/CT machine, with the same protocol, in a single institution. Furthermore, our study is the first report that attempted to predict ALN status using serial FDG PET/CT scans in a neoadjuvant setting.
In conclusion, we found that serial FDG PET/CT scans after NAC added incremental diagnostic confidence in the pathologic status of ALN before surgery for patients with stage II/III breast cancer. Our findings suggest that the combined use of serial FDG PET/CT and chest CT scans would provide better information regarding ALNs before surgery.
This research was supported by grants from the Priority Research Centers Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (2009-0093820) and the Seoul National University Hospital research fund (03-2010-019-0).
This study was presented at the 46th American Society of Clinical Oncology Annual Meeting in Chicago, IL, USA (June 7, 2010).
Conflict of interest