Localization of prostate cancer recurrence is a major challenge in patients with biochemical recurrence following curatively intended treatment. The currently applied imaging modalities such as bone scintigraphy and abdominal pelvic computed tomography (CT) are too insensitive, particularly at low serum prostate-specific antigen (PSA) values (1). These two modalities only reliably show the site of relapse in patients with very high PSA levels (>10 ng/ml) (2). This is of particular importance as salvage radiotherapy in these patients is most effective at serum PSA values <0.5 ng/ml (3,4).

Established positron emission tomography (PET) tracers such as ¹¹C- or ¹⁸F-choline or ¹¹C-acetate are able to directly detect tumor tissue, which is also true at the bone level, where these tracers show tumor tissue, as opposed to reactive bone as with bone scans and ¹⁸F-sodium fluoride (NaF) PET/CT. This result may convey higher sensitivity for both osteoblastic and osteolytic bone lesions at low PSA levels (5). However, the majority of large series have included patients with a PSA level >2 ng/ml (6,7). Most studies that included patients with low PSA levels have shown infrequent bone metastases, and little information is provided concerning tumor characteristics in individual patients (8–11). Generally, choline PET is only indicated in biochemical failure if the PSA level is >2 ng/ml or the PSA level is rising rapidly (12).

Gallium-68-labelled prostate-specific membrane antigen (⁶⁸Ga-PSMA) PET/CT has recently been introduced as a promising method for prostate cancer imaging, both for staging and restaging (13–17). In contrast to existing PET tracers, ⁶⁸Ga-PSMA PET has revealed pathological sites of uptake, even with low PSA levels, in several retrospective series (13–17).

In 2015, the present authors initiated two prospective trials with ⁶⁸Ga-PSMA PET/CT in prostate cancer, including a study of PSMA-11 (DKFZ-PSMA-11, also known as PSMA-HBED-CC) PET/CT in recurrent prostate cancer in comparison with magnetic resonance imaging (MRI) and NaF PET/CT (EudraCT no.; 2014-005073-37). The trial was approved by the Danish Health and Medicine Authority, The Danish Data Protection Agency, and the North Denmark Region Committee in Health Research Ethics. The patients received written and oral information and provided written informed consent, inclusive presentation of individual study results in a blinded fashion. The current study reports two cases of bone metastasis in patients with PSA levels <2 ng/ml and slow PSA kinetics.

Imaging in cases of biochemical recurrence of prostate cancer has been hampered by the lack of appropriately sensitive modalities, particularly for bone lesions. Bone scans and CT in general do not have adequate sensitivity, and choline PET is only indicated if the PSA level is >2 ng/ml or if the PSA has rapidly rising kinetics (e.g., PSAdt <6 months). The current study presents two cases of pathological ⁶⁸Ga-PSMA uptake on PET/CT scans with PSA levels <2 ng/ml and slow PSA kinetics. These data, along with a number of cases identified in the published literature (Table I, data captured from a search of 1,858 references in an ongoing systematic review with a cutoff date of August 2016), indicated that ⁶⁸Ga-PSMA PET/CT may be a valuable imaging technique to localize disease in very early biochemical recurrence.

⁶⁸Ga-PSMA PET/CT identified pathological uptake in the bones in the first patient and in both the lymph nodes and the bones in the other patient. One of the current cases, as well as a number of cases presented in Table I, were identified to have a solitary site relapse in the bone. From a clinical point of view, the majority of published cases have a minimum amount of clinical and laboratory data to indicate if relapse was expected to occur at the prostate level, in lymph nodes or in the bones. The missing information includes data about resection margins, the nodal status at staging and/or surgery, and the levels and duration of post-treatment PSA levels (or time since curative treatment).

The presented data with ⁶⁸Ga-PSMA PET/CT use in early recurrence are encouraging. Similar data with other PET tracers are scarce in the published literature. In several previous studies with choline PET/CT, no bone lesions were detected in early biochemical recurrence (10,11). Kjolhede et al (8) presented 5 patients with suspected bone lesions among 58 patients with biochemical recurrence and PSA levels of <2 ng/ml, but there were no data on PSA kinetics in patients with bone lesions. Castellucci et al (9) performed ¹¹C-choline PET/CT in 605 patients with biochemical recurrence and PSA levels of 0.2–2.0 ng/ml and a median PSAdt of 6 months, and showed pathological bone uptake in 51/172 patients with choline-positive scans. The PSAdts of these 51 patients were not reported. It is appropriate to say that lesions detected by choline PET/CT must be cautiously interpreted due to low to moderate specificity; a study with verification by histology showed a low predictive value of 24% at the node level (19).

A number of trials are emerging that directly compare ⁶⁸Ga-PSMA with other PET tracers and standard imaging in patients with biochemical recurrence. Although ⁶⁸Ga-PSMA PET/CT appeared to detect more lesions than choline PET/CT, both on the whole patient level and at the level of individual lesions, the majority of studies include patients with high PSA levels, and there are limited data on skeletal involvement and characteristics of individual patients (20–22).

⁶⁸Ga-PSMA PET/CT is developing rapidly (16,17). However, most studies are retrospective reports with inherent methodological deficiencies, including lack of compliance with the standards for reporting of diagnostic accuracy (STARD) criteria (23). One key feature of the STARD guidelines is the definition of the reference test. In the absence of a true reference, it remains unknown if PET uptake parallels tumor recurrence. This is well known with choline PET/CT (19), and it is shown with suspicious lesions without sclerosis on CT with NaF PET as in the current case no. 1. In the present cases, skeletal malignancy was confirmed in both the patients with positive findings with two functional methods (⁶⁸Ga-PSMA and NaF PET) and at least one anatomical method (low-dose CT and/or MRI). In addition, both patients responded biochemically to anti-cancer treatment.

In conclusion, ⁶⁸Ga-PSMA PET/CT has emerged as a very promising imaging technique for use in identifying tumor sites in patients with biochemical recurrence. In comparison with standard imaging modalities as well as existing PET tracers, ⁶⁸Ga-PSMA PET/CT appeared to be sensitive at very low PSA levels (and in patients with slow PSA kinetics), which is optimal in terms of salvage radiotherapy.