Ethics and institutional review board approval
This study was approved by the institutional ethics committee of hospital. It has been registered online at NIH ClinicalTrails.gov (NCT 02742168). Written informed consent was obtained from all participants. The authors had full control of the data and information submitted for publication.
Study participants
This prospective study recruited 46 women who were first diagnosed with locally advanced stage II-III breast cancer from October 2015 to October 2018. The ‘‘gold standard’’ for diagnosis of the primary tumors and axillary lymph nodes were based on histopathological findings of fine needle biopsy (FNA) before treatment. The stage of cancer was determined by 18F-FDG PET/CT, ultrasound, MR, and/or whole-body bone imaging according to the TNM classification.
Exclusion criteria were: 18 year’s old, pregnancy, and lactation period and the patients with distant metastatic lesions.
Five patients were excluded because two patients were diagnosed with distant metastases during NAC (a brain metastasis and a lumbar vertebrae metastasis diagnosed by MRI or PET/CT), and three patients missed followed up checks during NAC. Finally, forty-one patients were included in the final statistical analysis (mean age 61.50 ± 7.84 years, range from 25 to 65; mean weight 56.64 ± 7.53 kg, range from 40 to 69).
Histopathological analysis and treatment
Immunohistochemical (IHC) staining was used to determine the histological subtypes. ER or progesterone receptor (PR) status was defined as positive when the number of positive nuclei was > 1%. HER2 expressed on the membrane was scored from 0 to 3. Scores of 3 + or 2 + and fluorescence in situ hybridization positive detection was defined as HER2-positive. The breast cancer subtypes were categorized into luminal, HER2-positive and triple-negative.
All patients were treated with four cycles of NAC (epirubicin 90 mg/m2 on day 1 and cyclophosphamide 600 mg/m2 on day 1 every 14 days) followed by four cycles of T (taxane 175 mg/m2 on day 1 every 21 days). In HER2-positive patients, trastuzumab was given with the last fourth cycle of taxane.
Radiopharmaceutical preparation
A 3PRGD2 kit was supplied by the Medical Isotope Research Center of Peaking University. 99mTc-labled RGD peptide was a noninvasive imaging of integrin αvβ3 expression via SPECT and the preparation process of 99mTc-3PRGD2 have been fully characterized in previous studies [13, 26] The 18F-FDG was purchased from Advance Medical Systems Limited, Nanjing, China.
Imaging
All patients underwent 99mTc-3PRGD2 imaging and then 18F-FDG imaging within three days. Each imaging series included three single-photon-emission-computed-tomography/computed-tomography (SPECT/CT) scans followed by three PET/CT scans. Note that baseline SPECT/CT0 and PET/CT0 imaging (first scan) were performed before the start of NAC, and after the first (second scan) and fifth (third scan) NAC cycles.
99mTc-3PRGD2 Imaging
For patients at SPECT/CT: after intravenous injection of 11.1 MBq/kg 99mTc-3PRGD2, whole-body planar (10cm/min) and SPECT/CT (30s/frame/6) scans for thorax were performed using a double-headed γ camera equipped with low-energy high-resolution collimators (Infinia Hawkeye4, GE Healthcare). The matrix was 128 128 pixels, and the photo-peak was centered on 140 keV with a 20% energy window.
18F-FDG Imaging
Patients were prepared via a 6-hour fasting period with blood glucose levels <150mg/dl. An FDG dose of 170-230 MBq (0.1 mCi/kg) was given intravenously. PET/CT (Biography mCT (64), Siemens) scans were acquired at 60 minutes after 18F-FDG injection. A PET scan (2.00 min per bed position) was performed for each patient, and PET acquisition was followed by low-dose CT (3 mm slices). The standard supine PET/CT was from the skull to the middle femur region.
Imaging reading
The three experienced nuclear medicine reading physicians were blinded to the pathological results, and could refer to other imaging results. A consensus decision was required for any discordant initial assessments. The readers evaluated the images visually for the focal tracer uptake compared with the uptake in surrounding normal tissue and the maximum radioactive counts of the primary breast tumors and prominent ALNs were measured in the two imaging modalities. If the patient without visible lymph node trace uptake for both of the imaging modalities at baseline would not undergo the subsequent pathological response predicting analysis.
99mTc-3PRGD2imaging analysis
Analysis of 99mTc-3PRGD2 uptake into tumors was based on region of interest (ROI) analyses. Seeking to avoid nonspecific 99mTc-3PRGD2 uptake into benign lesions or normal breast tissue which were found in the previous studies [12], the blood pool of the aortic arch was used as a control for the primary tumor. For the controls in the for ALN image data, contralateral normal axillary areas were used. The mean counts for control regions of interest were measured and used as the background values for calculating the tumor-to-background (T/B) ratios. Note that if an apparent tumor remission yielded no obvious tracer uptake, we analyzed the same region of interest location as in SPECT/CT0. We calculated T/B ratios for three SPECT/CT scans: T/B0, T/B1, and T/B2, with each T/B ratio calculated as follow:
ΔT/B1= ((T/B0- T/B1) / T/B0) ×100%
ΔT/B2= ((T/B0- T/B2) / T/B0) ×100%
18F-FDG imaging analysis
Nuclear medicine physicians used a standard analytical pipeline identified the mostly focal 18F-FDG uptake of tumors and ALN. The ROIs were drawn manually in the primary tumors and ALNs and measured the SUVmax (SUVmax0, SUVmax1, SUVmax2) obtained by generating a 3D region of interest [21].
The relative changes of SUVmax were calculated as follows:
ΔSUVmax1= ((SUVmax0- SUVmax1) / SUVmax0) × 100%
ΔSUVmax2= ((SUVmax0- SUVmax2) / SUVmax0) × 100%
Response assessment
Surgery with ALN dissection was performed after 8 cycles of NAC routinely. We defined pathological complete response (pCR) as the simultaneous absence of any residual invasive tumor cells from all of the breast and axillary node specimens.
Statistical analysis
Statistical analyses were performed using SPSS 24.0 software. Student’s t-tests, Fisher’s exact test, and binary logistic regression were used for analysis of parameters comparisons. The ROC curve analysis was used to evaluate the predictive performance of pathological responders vs. pathological non-responders. Z tests were used to compare the areas under curve (AUCs). P < 0.05 was considered to be statistically significant.