Added value of dedicated axillary hybrid 18F-FDG PET/MRI for improved axillary nodal staging in clinically node-positive breast cancer patients: a feasibility study

After approval of the study by the local medical ethical committee, a total of 12 patients provided signed informed consent to undergo dedicated axillary hybrid PET/MRI. Inclusion criteria were patients with histopathologically confirmed clinically node-positive breast cancer. Exclusion criteria were contraindications for either MRI, including the contrast agent gadobutrol, or PET, including radiotracer 18F–fluorodeoxyglucose (18F-FDG). Prior to inclusion, patients underwent axillary US. In addition, all patients were scheduled for MR imaging and, depending on the number of suspicious lymph nodes (i.e. 1–3 vs. 4+), for PET/CT as well [10].

Axillary US exams were performed by the radiologist on call, using an ultrasound system with a linear 2- to 17-MHz array transducer (iU22-xMATRIX, Philips Healthcare, Best, the Netherlands). Criteria for suspicious axillary lymph nodes included diffuse cortical thickening, focal cortical mass and/or thickening, and loss of fatty hilum [11]. During axillary US, tissue sampling of suspicious axillary lymph nodes was performed. In the case of multiple suspicious nodes, only the most suspicious node was sampled, and the number of suspicious axillary lymph nodes was reported.

In the Netherlands, PET/CT exams can be considered in the case of stage III/IV breast cancer [10, 12]. After patients had fasted for a period of at least 4 h and had a blood glucose level < 10 mmol/L, an intravenous injection of 18F-FDG was administered at a dose of 2 MBq/kg body weight, followed by a resting period of 60 ± 10 min. Patients were placed in supine position with elevated arms in the PET/CT scanner (Gemini TF, Philips Healthcare, Best, the Netherlands). A low-dose CT scan (120 kV, 30 mAs, slice thickness 4 mm) from head to thigh was performed, followed by the PET acquisition (3 min per bed position). CT images were reconstructed using filtered back-projection. PET images were reconstructed using the BLOB-OS-TF time-of-flight algorithm provided by the manufacturer, with a voxel size of 4 × 4 × 4 mm³.

Directly following whole-body PET/CT, state-of-the-art dedicated axillary hybrid PET/MRI images were acquired on a 3.0 T Biograph mMR integrated PET/MRI system (Siemens Healthineers, Erlangen, Germany). Dedicated axillary MRI was performed using a six-channel anterior body coil (Siemens Healthineers), which was placed on the ipsilateral shoulder. Patients were placed in prone position with both arms elevated.

In three cases, no whole-body PET/CT was performed. These patients underwent dedicated axillary hybrid PET/MRI after administration of 18F-FDG at a dose of 2 MBq/kg body weight and a resting period of 60 ± 10 min, with identical fasting time and a blood glucose level < 10 mmol/L as with PET/CT.

The dedicated axillary hybrid PET/MRI protocol consisted of an unenhanced T2-weighted (T2W) sequence, contrast-enhanced T1-weighted sequence (TW1), and a fusion sequence of PET images with the unenhanced T2W sequence. All sequences were performed in coronal view.

More specifically, a 2D standard T2W sequence without fat suppression was performed first using the following scan parameters: repetition time (TR) 2000 ms; echo time (TE) 118 ms; variable degree flip angle; turbo spin echo (TSE) factor 41; averages 1.4; voxel size 1.1 × 1.1 × 1.1 mm; field of view (FOV) 220 × 220 mm²; matrix 192 × 192, resulting in an in-plane resolution of 1.15 × 1.15 mm². Next, gadobutrol (Gadovist^®, Bayer HealthCare, Berlin, Germany) contrast at a volume of 15 cm³ was administered intravenously. After 6 min, a 3D T1W sequence was acquired using the following parameters: TR 7.27 ms; TE 2.46 ms; 15° flip angle; averages 2; voxel size 1.1 × 1.1 × 1.1 mm; FOV 220 × 220 mm²; matrix 192 × 192; resulting in an in-plane resolution of 1.15 × 1.15 mm². Total acquisition time of the MRI exam was 20 min, 21 s.

Three-dimensional iterative reconstruction was performed for all PET images of the hybrid PET/MRI system (number of iterations 3; subsets 21; voxel size 4.1 × 2.6 × 3.9 mm³; image matrix 172 × 172; zoom 1.0), with automatic attenuation correction through implementation of a four-compartment model attenuation map (Dixon-based μ-map). Images were acquired in a single bed position within 15 min after the start of MRI acquisition on the hybrid PET/MRI system.

All axillary lymph nodes on dedicated axillary MR, hybrid PET/MRI and PET/CT images were evaluated by a breast radiologist (M.B.I.L.) and nuclear medicine physician (S.V.). The breast radiologist has 7 years of dedicated experience in breast imaging, and the nuclear medicine physician has 9 years of experience in nuclear imaging.

First, the radiologist scored each axillary lymph node on dedicated axillary T2W images, according to the criteria previously defined by Baltzer er al., using a confidence scale from 0 (no lymph nodes) to 4 (definitely malignant) [13]. Lymph nodes with a score of 0–2 were considered benign, and those with a score of 3–4, malignant. Suspicious characteristics included irregular margins, inhomogeneous cortex, perifocal edema, absence of fatty hilum, asymmetry, and absence of chemical shift artifacts [13‐15]. Next, contrast-enhanced dedicated axillary T1W images were shown, which allowed the radiologist to adjust his score for each lymph node, if deemed necessary. Suspicious characteristics included absence of contrast hyperintensity of the lymph node and absence of an intact nodal border [15].

Second, the breast radiologist and nuclear medicine physician evaluated in consensus each axillary lymph node on dedicated axillary hybrid PET/MRI using a four-point system as previously described by Aukema et al. for the assessment of lymph nodes with 18F-FDG PET/CT: 0 = similar to other surrounding lymph nodes, 1 = slightly more intense than other lymph nodes, 2 = moderately intense, 3 = very intense [16]. Lymph nodes with a score of 0–1 were considered benign, while a score of 2–3 was considered malignant. PET information was considered to be decisive regarding the number of suspicious axillary lymph nodes in the evaluation of dedicated axillary hybrid PET/MRI. The nuclear medicine physician further measured maximum standardized uptake values (SUVmax) of the primary breast tumor and the most FDG-avid axillary lymph node on each dedicated axillary hybrid PET/MRI exam. Finally, the nuclear medicine physician evaluated each axillary lymph node on 18F-FDG PET/CT, using the same four-point system [16], and measured SUVmax of the primary breast tumor and the most FDG-avid axillary lymph node on each PET/CT exam. Both PET/CT and PET/MRI exams were shown anonymously and in randomized order to the nuclear physician for evaluation of the number of suspicious axillary lymph nodes.

In this study, clinically node-positive disease was based on US-guided cytological and/or histological confirmation of axillary lymph node biopsy. In cases where US was performed in the referring hospital, the pathology of the axillary lymph node biopsy was revised and metastases were verified by the pathologist in our hospital (L.K.).

Statistical analyses were performed using SPSS software (version 22, IBM Corp., Armonk, NY, USA). Axillary nodal status was based on the number of suspicious axillary lymph nodes, as follows: none (N0), 1–3 (N1), 4–9 (N2), or 10 or more (N3), according to axillary US (including results of tissue sampling), dedicated axillary (PET)MRI and/or PET/CT, respectively. When the result of dedicated axillary hybrid PET/MRI led to either up- or down-staging of clinical nodal status compared to standard imaging modalities (US, MRI or PET/CT), it was considered a change in clinical nodal status according to dedicated axillary hybrid PET/MRI findings.

A total of 12 consecutive patients with clinically node-positive breast cancer over a period from April 2016 to March 2017 were included in this feasibility study. The mean age was 49 years (range, 28–70 years). In all patients, clinically positive nodal status was histopathologically confirmed with US-guided biopsy. Eleven of 12 patients had invasive carcinoma of no special type; one patient had mixed lobular carcinoma with carcinoma of no special type. In addition, 9 of 12 patients underwent PET/CT. A more detailed overview of the general characteristics is shown in Table 1.

When compared to standard imaging modalities, dedicated axillary hybrid PET/MRI resulted in changes to clinical nodal status as follows: 40% according to US findings (4 of 10 patients), 75% according to T2W MRI findings (9 of 12 patients), 40% according to CE-T1W MRI findings (4 of 10 patients) and 22% according to PET/CT findings (2 of 9 patients), respectively. Differences between PET/CT and PET/MRI findings were due to better delineation of the lymph nodes on the MRI component. A more detailed overview of all cases, including up- or down-staging of nodal status according to dedicated axillary hybrid PET/MRI findings, is provided in Table 2.

In two patients, axillary US findings regarding the number of suspicious axillary lymph nodes was missing, because US was performed in the referring hospital. However, in addition, neither patient underwent contrast-enhanced T1W MRI, as this was refused by both patients.

Figures 1 and 2 illustrate examples of differences in the number of suspicious axillary lymph nodes between PET/CT (Fig. 1) and T2W MRI (Fig. 2) when compared to dedicated axillary hybrid PET/MRI. The Appendix provides a complete overview of all axillary lymph nodes evaluated on dedicated axillary (PET/)MRI and PET/CT images.

The mean SUVmax of the primary tumor and the most FDG-avid axillary lymph node on PET/CT were 8.6 (± 5.7) and 7.7 (± 4.3), respectively. Comparable values were achieved on hybrid PET/MRI, with mean SUVmax of 9.0 (± 5.0; p = 0.678) and 7.8 (± 5.3; p = 0.767) for the primary tumor and the most FDG-avid lymph node, respectively. Table 3 provides an overview of all SUVmax measurements on PET/CT and hybrid PET/MRI for each patient.