Combined 18F-fluorocholine and 18F-fluoride positron emission tomography/computed tomography imaging for staging of high-risk prostate cancer


Henrik Kjölhede, Department of Surgery, Växjö County Hospital, 351 85 Växjö, Sweden. e-mail:


Study Type – Diagnosis (cohort)

Level of Evidence 2a

What's known on the subject? and What does the study add?

Positron emission tomography/computed tomography (PET/CT) with choline and fluoride for the detection of metastases in patients with prostate cancer have each been evaluated, with mixed results. Choline PET/CT has been evaluated against pelvic lymphadenectomy, generally with a low sensitivity but a high specificity; however, the study populations have been heterogenous. Fluoride PET/CT has been evaluated against other imaging methods, such as bone scan, single photon emission CT and MRI, and has been shown to have high specificity as well as sensitivity for bone metastases, but there are no studies with biopsy verification.

This is the first study that evaluates the clinical use of both choline and fluoride PET/CT on the same patients in a well-defined population of patients with high-risk prostate cancer.


  • • To investigate how often positron emission tomography/computed tomography (PET/CT) scans, with both 18F-fluorocholine and 18F-fluoride as markers, add clinically relevant information for patients with prostate cancer who have high-risk tumours and a normal or inconclusive planar bone scan.


  • • Patients with prostate cancer with prostate specific antigen (PSA) levels between 20 and 99 ng/mL and/or Gleason score 8–10 tumours, planned for treatment with curative intent based on routine staging with a negative or inconclusive bone scan, were further investigated with a 18F-fluorocholine and a 18F-fluoride PET/CT.
  • • None of the patients received hormonal therapy before the staging procedures were completed.


  • • For 50 of the 90 included patients (56%) one or both PET/CT scans indicated metastases.
  • • 18F-fluorocholine PET/CT indicated lymph node metastases and/or bone metastases in 35 patients (39%).
  • • 18F-fluoride PET/CT was suggestive for bone metastases in 37 patients (41%).
  • • In 18 patients (20%) the PET/CT scans indicated widespread metastases, leading to a change in therapy intent from curative to non-curative.
  • • Of the patients with positive scans, 74% had Gleason score 8–10 tumours. Of the patients with Gleason score 8–10 tumours, 64% had positive scans.


  • • PET/CT scans with 18F-fluorocholine and 18F-fluoride commonly detect metastases in patients with high-risk prostate cancer and a negative or inconclusive bone scan.
  • • For 20% of the patients the results of the PET/CT scans changed the treatment plan.

positron emission tomography


extended pelvic lymph node dissection


The most common sites for metastases from prostate cancer are the pelvic and retroperitoneal lymph nodes and the axial skeleton. Radical local therapy is usually not considered for patients with skeletal metastases, widespread or large lymph node metastases. The standard of reference for detection of lymph node metastases is an extended pelvic lymph node dissection (ePLND) [1,2]. However, this procedure is invasive and associated with complications and morbidity [3]. Positron emission tomography (PET) with 11C- or 18F-labelled choline fused with CT is a promising non-invasive staging procedure for detecting lymph node metastases [4–10]. The reported sensitivity varies, but the specificity is consistently high, up to 100%. Choline is currently the most commonly used tracer for PET/CT for prostate cancer staging [11]. PET/CT with 18F-labelled fluoride has been shown to have a higher sensitivity and specificity than standard 99mTc-MDP bone scans for the detection of skeletal metastases [12–14]; however, it is not known whether the information from both 18F-fluorocholine and 18F-fluoride PET/CT scans identifies more patients with metastatic disease, than does 18F-fluorocholine PET/CT alone.

The hitherto reported studies of PET/CT for staging of prostate cancer are based on heterogeneous groups of patients whose tumour characteristics are highly variable. In some reports it is not specified how many of the patients were receiving endocrine therapy at the time of evaluation. It is therefore not yet possible to identify for which patients a PET/CT should be considered as a routine staging procedure or which tracer should be used [15]. Patients with poorly differentiated prostate cancer or high PSA values may have multiple metastases despite negative findings on standard non-invasive staging procedures, such as planar bone scan and CT [16,17]. When treatment with curative intent is planned for such patients, PET/CT may have the potential to add information of clinical value for a considerable proportion of the patients.

The objective of the present study was to investigate whether combining PET/CT with both 18F-fluorocholine and 18F-fluoride adds clinically relevant information for a well-defined group of patients with high-risk prostate cancer planned for treatment with curative intent after normal or inconclusive planar 99mTc bone scans.



In the study protocol, accepted by the research ethical review board at University of Lund (EPN Dnr 552/2007), urologists in the Southern healthcare region of Sweden referred patients with biopsy-verified high-risk prostate cancer considered for curative treatment for both a 18F-fluorocholine PET/CT and a 18F-fluoride PET/CT scan. Biopsies were performed with eight to 12 cores. High-risk prostate cancer was defined by a PSA ≥20 ng/mL and/or Gleason score of 8–10. The patients all had a recent 99mTc-MDP bone scan with normal or inconclusive findings. Patients already receiving hormonal therapy and patients with PSA ≥100 ng/mL were excluded from the study. The clinical management was determined by the urologist referring the patient, commonly after discussing positive findings on the PET/CT scans with a member of the study group.


18F-fluorocholine was synthesized by 18F-fluoroalkylation of N,N-dimethylaminoethanol on the TracerLab MXFDG module (GE Healthcare, Stockholm, Sweden) using the method described by Kryza et al.[18]. 18F-Fluorobromomethane was produced by reaction of dibromomethane with 18F-fluoride, assisted by tetrabutylammonium hydroxide instead of Kryptofix 2.2.2. Tracer purity was controlled by ion chromatography and thin layer chromatography. The specific activity was >74 GBq/µmol, with >99% radiochemical purity.

The PET/CT studies were acquired with an integrated PET/CT system (Philips Gemini TF, Philips Medical Systems, Cleveland, OH, USA) at the Centre for Medical Imaging and Physiology, Skåne University Hospital in Lund. Patients fasted for 4 h before 18F-fluorocholine injection. Whole-body PET (pelvis to neck) was acquired 1–1.5 h after i.v. injection of 4MBq/kg (max dose 400MBq) of 18F-fluorocholine and, on another occasion, 18F-fluoride, with 2 min per bed position. The two PET/CT studies were performed 1–24 days apart (median 4 days). The 18F-fluorocholine PET scan was acquired from proximal femur to the base of the skull. The 18F-fluoride PET scan also included the whole skull and the proximal half of the femur and humerus. CT scans were performed immediately before the PET scan with a multi-detector (16-slice) spiral CT scanner. For the 18F-fluorocholine PET scans, CT scans were acquired as diagnostic quality CT, with 5 mm reconstructed slice thickness, pitch factor 0.938, rotation speed 0.75 s, 120 kV and with high-beam tube current modulation (120–300 mA) based on the patient's total body mass. Oral contrast 1000 mL (50 mL Omnipaque [GE Healthcare], 240 mgI/mL and 30 mL 70% Sorbitol mixed with 1000 mL water) was given 60 min before the scan. An automatic injection pump (Medrad Stellant Dual Head Injector, Pittsburgh, PA, USA) was used to administer the i.v. contrast (Omnipaque 350 mgI/mL), 350 mgI/kgBW, with an injection speed of 2.5 mL/s. The contrast volume was calculated using a dedicated computer program (Omni-Ject, distributed in Nordic countries by GE Healthcare, Sweden). Three CT scans were obtained. The liver was scanned without i.v. contrast and then the thorax was scanned in the arterial phase – both scans during breath-hold. The third scan covered the body from the skull base to proximal thigh and was obtained in the portal phase during normal breathing. This CT dataset was used for the attenuation correction calculation and for image fusion. For the 18F-fluoride PET scans, low dose CT scans (50 mAs) were used for attenuation correction and image fusion, since the patients had already had a diagnostic quality CT scan.

Focal 18F-fluorocholine uptake above background, corresponding to an abdominal or pelvic lymph node, was reported as positive. Focal 18F-fluorocholine or 18F-fluoride uptake above background in bone not corresponding to other pathology, i.e. fracture, was also reported as positive. No specific uptake value was used as threshold. The PET/CT scans were interpreted by teams of two, consisting of a nuclear medicine physician and a radiologist.


Clinical stage, Gleason score and PSA levels at the time of referral for PET/CT were retrieved retrospectively from the medical records. Only descriptive statistics were applied.


Between March 2008 and June 2010, 90 men met the criteria of the study protocol and underwent both 18F-fluorocholine PET/CT and 18F-fluoride-PET/CT. The characteristics of these patients are shown in Table 1.

Table 1. Clinical characteristics of the 90 patients included in the study and the patients with either none, one or both positive PET/CT scans
 All patientsn= 90Patients with negative scansn= 40Patients with positive choline PET/CT scansn= 35Patients with positive fluoride PET/CT scansn= 37Patients with one or both positive scansn= 50
Age, years     
 Mean (sd)66.5 (5.6)66.8 (6.1)66.1 (5.8)66.1 (5.1)66.2 (5.3)
 Median (range)66.8 (49.9–7.2)66.8 (49.9–77.3)66.8 (53.4–76.9)67.1 (55.3–76.1)66.9 (55.4–76.9)
PSA (ng/mL)     
 Mean (sd)27.9 (20.2)26.7 (18.6)29.9 (23.4)28.2 (19.2)28.9 (21.6)
 Median (range)22.0 (2.4–95)22.0 (2.4–81)23.0 (5.2–95)23.0 (6.0–82)22.5 (5.2–95)
Biopsy Gleason score, n (%)     
 5–64 (4)4 (10)0 (0)0 (0)0 (0)
 728 (31)15 (38)7 (20)11 (30)13 (26)
 8–1058 (64)21 (53%)28 (80)15 (70)37 (74)
Local clincal tumour stage, n (%)     
 T1c14 (16)9 (23)2 (6)5 (14)5 (10)
 T230 (33)16 (40)9 (26)10 (27)14 (28)
 T346 (51)15 (38)24 (69)22 (59)31 (62)
99mTc-MDP bone scan     
 Negative73 (81)39 (98)23 (66)22 (59)34 (68)
 Inconclusive17 (19)1 (3)12 (34)15 (41)16 (32)

The 18F-fluorocholine PET/CT scans indicated metastatic disease in 35 (39%) patients and were normal, except for uptake in the primary tumour, in 55 (61%). The 18F-fluoride PET/CT scans indicated bone metastases in 37 (41%) patients and were normal in 53 (59%). Of the 35 positive 18F-fluorocholine PET/CT scans, 24 showed more than one positive site. For 18F-fluoride PET/CT, 19 of the scans showed multiple positive sites. Fifty (56%) patients had positive findings in either one or both of the scans, 30 of which (33% of all) had findings indicating more than one metastasis. 18F-fluoride PET/CT indicated skeletal metastases in 15 patients with a negative 18F-fluorocholine PET/CT scan, while 18F-fluorocholine PET/CT indicated lymph node metastases in 13 patients with a negative 18F-fluoride PET/CT scan. All skeletal metastases suggested by 18F-fluorocholine PET/CT were detected by 18F-fluoride PET/CT. The PET/CT scans indicated metastases in 0, 13/28 (46%) and 37/58 (64%) of the patients with Gleason score 6, 7 or 8–10 tumours, respectively. More than one metastatic site was indicated in 0, 3/28 (10%) and 27/58 (47%) of the patients with Gleason score 6, 7 or 8–10 tumours, respectively. Fifteen of the 17 patients with an inconclusive bone scan had positive results on the 18F-fluoride PET/CT scan and 22 (30%) with normal bone scans had positive results on 18F-fluoride PET/CT. In the latter group, 14 of the 22 patients had Gleason score 8–10 tumours.

For 18 (20%) of the patients, the treatment plan changed from curative to non-curative intent based on the findings of the PET/CT. Seventeen had multiple metastases, including bone metastases in nine of them, according to the 18F-fluorocholine PET/CT. Eleven of these patients had bone metastases according to the 18F-fluoride PET/CT. In one case where both scans showed only a single positive site, magnetic resonance tomography was performed that confirmed the suspicions of a bone metastasis, with changed treatment plan as a result (Table 2). The treatment plan was changed for 0, 7 and 26% of the patients with Gleason score 6, 7 or 8–10 tumours, respectively. These 18 patients would all have been exposed to either a radical prostatectomy with an ePLND or an ePLND alone, without the results from the PET/CT scans. All 18 patients later received primary endocrine therapy instead of treatment with, supposedly, unsuccessful curative intent.

Table 2. Findings on PET/CT scans for the patients whose treatment plan was changed from curative to non-curative intent
Patient no.Age, yearsPSA, ng/mLcT stageGleason scoreBone scan 18F-fluorocholine 18F-fluoridePositive lymph node sitesPositive skeletal sites
  • *

    Verified by MRI. + positive; – negative; +/− inconclusive.

 16756T24 + 3+/−++Multiple iliacRib, scapula
 26929T34 + 5+Paraaortal, multiple iliac, ventral of S2 
 35677T34 + 5++Multiple iliac, up to 4 cm sizeMultiple vertebrae
 46322T34 + 5+/−++Multiple iliacMultiple bone locations
 57125T24 + 4+/−++IliacRib, vertebra, pelvis
 66223T34 + 4++Multiple iliacVertebra
 76919T24 + 5+Multiple iliac, less than 1 cm in size 
 86417T34 + 5++IliacVertebra*
 96537T34 + 3+/−++Multiple paraaortal, multiple iliacMultiple bone locations
106811T34 + 5+/−++Multiple iliac, paraaortalMultiple bone locations
116916T34 + 5+Multiple paraaortal, multiple iliac 
127011T24 + 5+/−++IliacSkull, pelvis
137227T34 + 4+/−++IliacPelvis
145582T35 + 4++Multiple paraaortal, multiple iliacVertebra
157031T1c4 + 4+/−++Multiple iliacHumerus*
167431T34 + 4+Sacral, multiple iliac, pararectal 
176417T34 + 4+/−++NoneVertebra, rib, skull
186750T33 + 4++IliacVertebra*

Incidental findings on 18F-fluorocholine PET/CT scans were five cases of benign adrenal tumours, one of which underwent surgery, three benign lung tumours, a pituitary adenoma and a ureteric stone.


In the present study 18F-fluorocholine and 18F-fluoride PET/CT scans indicated metastatic disease in 56% of patients with high-risk prostate cancer without conclusive evidence of metastases on a previous 99mTc-MDP bone scan. For 20% of the patients, the metastatic spread was considered so extensive that the treating urologist changed the treatment plan from curative to non-curative. Patients with Gleason score 8–10 cancers had the greatest likelihood (64%) of positive findings on the PET/CT scans, which resulted in a change of treatment plan for 15 of 28 in the present study. Both 18F-fluorocholine and 18F-fluoride PET/CT each detected metastases that the other did not. The present study indicates that 18F-fluorocholine PET/CT and 18F-fluoride PET/CT are clinically valuable staging procedures for patients with high-risk prostate cancer when treatment with curative intent is considered, both separately and combined, especially for patients with Gleason score 8–10 cancers.

The largest of the previously reported choline PET/CT series are those by Beheshti et al.[4] and de Jong et al.[5]. The present results are consistent with those reported by Beheshti et al.[4]. In their study, 18F-fluorocholine PET/CT detected lymph node metastases in 34 (41%) and bone metastases in 12 (14%) of 83 patients with Gleason score 8–10 tumours and PSA-values above 20 ng/mL. However, their patients were not pre-screened with a bone scan. A smaller proportion of patients with 11C-choline PET/CT detected metastases was reported by de Jong et al.[5], 22%, but their study population was more heterogeneous, with only 15% of patients with Gleason score 8–10 tumours. Also, 40% of their patients with positive PET/CT had PSA-values >100 ng/mL, a patient group for which treatment with curative intent is usually not considered.

No previous study has reported the outcome of 18F-fluoride PET/CT for a large group of patients with high-risk prostate cancer considered for treatment with curative intent. Some smaller studies of 18F-fluoride PET/CT in more heterogeneous groups of patients with prostate cancer have been performed, all showing a higher sensitivity and specificity for skeletal metastases than standard 99mTc-MDP bone scans [12,14,19]. These other reports are consistent with the results of the present study. We saw no patient with a negative 18F-fluoride PET/CT scan having a 18F-fluorocholine PET/CT scan suggesting bone metastases, adding to previous evidence that 18F-fluoride is the tracer of choice for detecting metastases to the skeleton. However, data on specificity and, especially, sensitivity for methods of detection of bone metastases should be interpreted with caution because no study providing histopathological verification of suspected bone metastases has been reported. Depending on local resources, fluoride PET/CT could be considered instead of a standard bone scan or as an additional procedure for patients with a normal or inconclusive bone scan.

We believe that the greatest benefit of PET/CT for patients with high-risk prostate cancer is for the group identified with extensive metastatic disease. Those patients would probably not benefit from, and could hence be spared the adverse effects of, radical local therapy. Since the optimum treatment of patients with prostate cancer with metastases limited to the pelvic lymph nodes is not known, the clinical benefit of PET/CT is difficult to define for the patients with findings indicating such limited metastases only.

The main limitation of the present study is the lack of histopathological verification of the metastases detected by PET/CT. The positive findings for the patients whose treatment plan was changed were extensive and obviously pathological, and not unexpected considering the tumour characteristics (Table 2, Fig. 1). Also taking into account the high positive predictive value previously reported, we consider it highly unlikely that these represented false-positive findings. It could be argued that this should have been evident in those cases where the bone scan was considered inconclusive. However, it is often difficult to make the decision to abstain from further curative treatment when there is only a single uptake on the bone scan. Furthermore, it could also have been desirable to compare the two PET/CT scans with just the diagnostic CT scan. This was not part of the original study protocol, and has therefore not been analysed.

Figure 1.

PET/CT images from patient number three in Table 2. A, Full-body overview of the bone scan with uptake at joints assessed to be arthrosis and normal accumulation in bladder. B, Full-body overview of choline PET scan with normal uptake in bladder, liver, kidneys and salivary glands. Multiple metastatic lesions in vertebrae and pelvis. Multiple suspected large metastases in iliac lymph nodes are not visible on PET owing to extensive necrosis, but are visible on CT. C, Full-body overview of fluoride PET scan. Multiple metastatic lesions in vertebrae and pelvis. D, Transverse slice of fused choline PET/CT showing metastatic lesion in the seventh thoracic vertebra. E, Transverse slice of fused fluoride PET/CT with metastatic lesions in the seventh thoracic vertebra and a rib.

In conclusion, 18F-fluorocholine PET/CT and 18F-fluoride PET/CT provided clinically valuable staging information for a substantial proportion (20%) of patients with high-risk prostate cancer and a previous normal or inconclusive bone scan. Combined, they detected more metastases than either method did alone. Patients with Gleason score 8–10 tumours were much more likely to have extensive metastatic disease according to PET/CT than patients with more differentiated cancers. 18F-fluoride PET/CT is the better choice when some other staging method suggests, but not is conclusive for, bone metastases.


Grants for the study were received from FoU Kronoberg, Philips, Region Skånes FoU-enhet, and The Swedish Cancer Foundation.


None declared.