NaF PET/CT for response assessment of prostate cancer bone metastases treated with single fraction stereotactic ablative body radiotherapy

Prostate cancer represents a major cancer burden in men, representing the most common male cancer diagnosis [1]. Prostate cancer staging determines appropriate treatment at initial presentation and during disease progression and makes use of various medical imaging techniques. The most probable site of distant metastases in prostate cancer is bone, thus imaging techniques used for visualization of prostate metastases must be able to accurately visualize sites of bone disease [2]. This is particularly so with increasing interest in metastasis-directed therapy (MDT) for oligometastatic prostate cancer [3‐5]. The standard imaging for determination of prostate cancer bone metastases has been whole body bone scan with 2D scintigraphy or single photon emission computed tomography (SPECT) approaches, using ^99mTc methylene diphosphonate [6, 7]. These tracers are taken up at sites of high osteoblastic activity representing bone turnover. The tumour burden as represented on bone scans can be quantified into a bone scan index (BSI), which has been shown as an independent prognostic marker for survival [8, 9]. Bone scans have many limitations however, such as poor anatomical correlation and low specificity and sensitivity [10, 11].

Prior to use of ^99mTc MDP, ¹⁸F sodium fluoride (¹⁸F-NaF) was used for planar scintigraphy [12]. In recent years however ¹⁸F-NaF has been used for PET/CT acquisition, which allows high spatial resolution 3D imaging of osteoblastic activity and blood flow [13‐15]. ¹⁸F-NaF has been shown to have improved sensitivity and specificity for prostate cancer metastases, compared with ^99mTc MDP [10], although improvements through use of quantitative SPECT have recently suggested consistent standardised uptake value (SUV) between the two modalities for prostate and breast bone metastases [16].

In the context of oligometastatic prostate cancer, ¹⁸F-NaF PET/CT imaging facilitates high quality detection and visualization of skeletal metastases which may be suitable for local MDT such as stereotactic ablative body radiotherapy (SABR). In this study we examine ¹⁸F-NaF uptake prior to and after single fraction SABR to bone metastases in patients enrolled in a prospective clinical trial. We investigate ¹⁸F-NaF uptake in tumour and non-tumour bone, with the hypothesis that tumour and normal tissue response to SABR can be assessed by ¹⁸F-NaF PET/CT.

This is a pre-specified exploratory analysis of a prospective clinical trial (POPSTAR, ‘Pilot Study of patients with Oligometastases from Prostate cancer treated with STereotactic Ablative Radiotherapy’, Universal Trial Number U1111–1140-7563) [17]. Between April 2013 and November 2014 33 patients with oligometastic prostate cancer were enrolled with written informed consent. They received a single fraction of 20 Gy to a total of 50 metastases. All lesions in a given patient were treated synchronously within in a single treatment course. All patients had a ¹⁸F-NaF PET/CT at screening, and 6 months post-treatment. Patients were excluded from the study if they had more than three metastases after PET/CT screening. The Quality of Life including pain scores, and disease progression for the whole cohort has previously been reported [17]. The current analysis is limited to those patients with demonstrable bone metastases who received the treatment protocol.

3 MBq/kg of ¹⁸F-NaF was administered by intravenous injection followed by a 60 min uptake period. A low-dose CT acquisition was obtained first followed by the PET acquisition. Patients were imaged from vertex to toes on a PET/CT scanner (Discovery 690 GE Healthcare, USA). No fasting was required. Patients were encouraged to void prior to imaging.

Radiotherapy simulation CT was performed less than 2 weeks prior to radiotherapy treatment. Scans were performed on a Philips Brilliance Big Bore CT scanner with a 2 mm slice thickness at 140 kV. The gross tumour volume (GTV) was contoured as visualised on the ¹⁸F-NaF PET and planning CT imaging, limited to bone. For non-verterbral metastases, an isotropic 5 mm planning target volume (PTV) margin was applied to the GTV to account for geometric uncertainties in the treatment. In the case of vertebral metastases, a clinical target volume (CTV) was applied according to the International Spine Radiosurgery Consortium consensus guidelines [18]. A 2–3 mm PTV margin was then applied to the CTV. Treatment planning was performed in the Eclipse treatment planning system (v11, Varian Medical Systems, Palo Alto, USA). Non-vertebral targets were treated with 3D conformal treatment plans which consisted of 7–9 beams typically including 1–2 non-coplanar beams. Dose was prescribed such that at least 99% of the PTV received 20 Gy, with a maximum dose between 125 and 140%. Vertebral targets were treated with a 9–12 coplanar IMRT fields, and prescribed such that at least 80% of the PTV received 18 Gy, with a maximum dose between 125 and 140%. Dose was calculated with the AAA algorithm at 2.5 mm resolution for non-vertebral targets and 1.5 mm resolution for vertebral targets. Radiotherapy was delivered on a Varian 21iX or Varian TrueBeam STx linear accelerator. Patients were immobilised in a vacuum immobilisation bag. Image guidance was performed using cone-beam CT (CBCT) and/or Exactrac planar x-ray imaging with a 0 mm tolerance for shifts. Mid-treatment CBCT was performed to ensure patient setup accuracy during treatment.

Twenty-one patients from the patient population with bone metastases were included in this analysis. A total of 33 bone lesions were irradiated (Additional file 1: Table S1). The baseline SUV characteristics of the lesions is shown in Additional file 2: Table S2. The mean (± 1 st. dev.) time of post-therapy PET was 7.4 ± 1.0 months. The SUV_mean of normal, un-irradiated bone was consistent between pre and post-treatment scans, with a mean ratio, post/pre of 0.98 ± 0.08. In this cohort, ¹⁸F-NaF had detected an additional 14 metastases, over that detected with CT and bone scan.

The POPSTAR trial was a prospective evaluation of SABR for oligometastases from prostate cancer and involved high, single fraction doses to bone and lymph node metastases [17]. This is the first study to demonstrate ¹⁸F-NaF as a response assessment tool in the context of SABR to bone metastases. ¹⁸F-NaF provides high spatial resolution and high sensitivity/specificity measurement of bone metastases and is representative of osteoblastic activity and blood flow. In the current study we have demonstrated that a single high dose fraction of external beam radiotherapy reduces ¹⁸F-NaF uptake in bone at 6 months, thus can be considered to reduce osteoblastic activity in bone metastases in the majority of patients. We have also shown a reduction of osteoblastic activity in regions of non-tumour bone treated to high doses per fraction. Bone fracture remains a side effect in SABR to bone lesions, with two Grade 2 and one Grade 3 fractures observed in the POPSTAR trial. In all the patients that had bone fracture post-SABR, there was high residual NaF uptake within the treated field, or there was new uptake adjacent to the treated volume as visualised on the follow up ¹⁸F-NaF PET scan. As reported previously for this cohort, for the bone metastasis-specific Quality of Life module (EORTC QLQ-BM22), painful sites, pain characteristics, and functional interference increased from baseline only at the 24-mo timepoint [17].

In the POPSTAR trial, bone metastasis was defined using a combination of CT and ¹⁸F-NaF scan information. The primary tumour was delineated by an experienced radiation oncologist and a 5 mm margin was applied to account for geometric uncertainty in the treatment delivery (planning target volume, [PTV]). In the subset of vertebral tumours, a CTV was applied according to international consensus guidelines, with a 2 mm CTV-PTV expansion. The use of ¹⁸F-NaF PET in the current study however has shown the potential inadequacy of direct expansion to PTV in non-vertebral bone without a CTV margin. Three patients had regions of increased uptake immediately adjacent to the treated volume, suggesting marginal failure. This may be mitigated by the use of a CTV margin for all bone metastases, similar to the international consensus guidelines for spine metastases. Despite the inclusion of a CTV in vertebral targets, there are still limitations with full coverage of the GTV due to the close proximity of dose limiting structures such as the spinal cord.

In more recent years, 18F-Fluoromethylcholine has been used for directing SABR to prostate cancer metastase, with some prognostic value [19]. Prostate specific membrane antigen (PSMA) PET scanning has however become the scanning modality of choice for prostate cancer in the metastatic setting [20]. PSMA PET has been shown to have comparable sensitivity and specificity as ¹⁸F-NaF PET for bone metastases in one study [21], however Uprimny et al. [22] found ¹⁸F-NaF PET detected more skeletal metastases and had a higher tumour to background ratio than PSMA PET. PSMA PET has the additional advantage of visualising soft-tissue metastases, however is limited to prostate cancer and renal cell carcinoma [23]. ¹⁸F-NaF PET thus has a role in response assessment for skeletal metastases in both prostate and non-prostate cancer.

In this study we have shown the response to a high dose single fraction SABR treatment to prostate bone oligometastases as visualised on ¹⁸F-NaF PET/CT. In the majority of patients, SABR reduces ¹⁸F-NaF PET/CT uptake in the tumour, and in high dose regions reduces uptake in non-tumour bone. Regions of increased uptake immediately adjacent to treated volumes suggest that increased clinical target volumes are required in the treatment of bone metastases with SABR.