Introduction
According to the European Association of Urology (EAU) guidelines, bone scintigraphy (BS) is the recommended imaging modality for the detection of bone metastases in patients with newly diagnosed unfavorable intermediate- to high-risk prostate cancer (PCa) [
1]. The presence of bone metastases as well as the number of bone metastases are of great significance to tailor the treatment [
1‐
3] and to determine patients’ prognosis [
4].
Positron emission tomography/computed tomography with ligands of the prostate-specific membrane antigen (PSMA PET/CT), especially with the
68Ga-labelled ligand PSMA-11, has been used widely to assess PCa metastases, particularly at the time of biochemical recurrence [
5,
6], but
68Ga-PSMA PET/CT has also been shown to possess premium diagnostic accuracy for the detection of bone metastases at the time of primary staging [
7,
8].
The purpose of this study was twofold. First, we aimed to evaluate the added value of 68Ga-PSMA-11 PET/CT in patients with newly diagnosed PCa versus the value of the initial BS and to assess the diagnostic accuracy of 68Ga-PSMA-11 PET/CT for skeletal metastasis patients with PCa at the time of primary staging.
Materials and methods
Patients
From May 2015 to October 2018, all patients undergoing PSMA PET/CT at our department were screened for inclusion in the present retrospective study. The eligibility criteria were as follows: (1) newly diagnosed with PCa, (2) no prior treatment for PCa, (3)
99mTc BS performed within 3 months of the
68Ga-PSMA-11 PET/CT, and (4) no treatment for PCa initiated between BS and PSMA PET/CT. All patients had BS and a CT scan of the thorax, abdomen, and pelvis conducted as part of the routine staging procedure according to the recommendations of the EAU [
1].
68PSMA-PET/CT
In short, the
68Ga-PSMA-11 ligand was used in the present study and was synthesized as previously described [
9]. The
68Ga-PSMA-11 PET/CT was performed in accordance with the guidelines of the European Association of Nuclear Medicine/Society of Nuclear Medicine (EANM) on
68Ga-PSMA-11 PET/CT [
10]. Images were acquired approximately 60 min after an intravenous injection of 2 MBq/kg body weight (minimum 100 MBq, maximum 200 MBq). Patients were examined using either a VCT Discovery True 64 PET/CT system (GE Healthcare, USA) or a Siemens Biograph mCT Flow 64 PET/CT system (Siemens, Erlangen Germany). The patients were scanned from the base of the skull to the upper thigh, and the PET images were acquired in 3D mode. The PET images were reconstructed using attenuation correction using an ordered subset expectation-maximization algorithm. For the Siemens Biograph, mCT Flow 64 PET/CT time-of-flight and point-spread-function were applied. A low-dose CT was performed immediately after the PET scan and used for attenuation correction and anatomical co-registration.
99mTc bone scintigraphy
BS was conducted in accordance with the EANM guidelines for BS [
11]. A planar whole body BS scan was acquired on a two-headed gamma camera (Symbia T16, Siemens Medical Solutions, Erlangen, Germany) 2–3 h after the intravenous injection of 9.4 MBq
99mTc-labelled methylene bisphosphonate per kilogram body weight (minimum 750 MBq). A supplemental single-photon emission computed tomography (SPECT)/CT, covering one- or two-bed positions (each 40 cm field of view), was conducted at the discretion of the physician in charge. SPECT/CT was acquired using the following parameters: 16 views with 10 s per view, as previously described [
12]. The images were reconstructed using iterative reconstruction with scatter correction. A low-dose CT was performed for attenuation correction and anatomical co-registration. Planar bone scan and any supplementary SPECT/low-dose CT was considered the standard bone evaluation per clinical guideline recommendation. According to institutional practice, a contrast-enhanced CT was performed after the bone scan for the assessment of lymph node and soft tissue metastasis. The information from the “bone window” was not included in the comparison with PSMA PET. In retrospect, no patients had any findings in the contrast-enhanced bone window, which changes the overall BS classification of metastasis.
Observers and procedure for image assessment
The evaluation of BS was a part of the daily clinical practice. The BS results were evaluated by at least two observers, mostly one physician in training and re-read by at least one experienced board-certified specialist in nuclear medicine. The consensus result was categorized as either (1) no bone metastases, (2) equivocal for bone metastases in which case it was an institutional practice to conduct a 68Ga-PSMA-11 for confirming or ruling out bone metastases, or (3) bone metastases.
The
68Ga-PSMA-11 PET/CT images were evaluated by three board-certified nuclear medicine physicians with experience in the evaluation of
68Ga-PSMA-11 PET/CT [
13]. The evaluation of
68Ga-PSMA-11 PET/CT was conducted according to Rauscher et al. [
14]. The PSMA PET/CT images were categorized on a three-point-scale: bone metastases (M1), no bone metastases (M0), or equivocal in patients with bone lesions not typical for bone metastases (Me). Furthermore, the exact number and location of all equivocal or metastatic bone lesions were described in patients with 10 or fewer lesions. The observers had no access to any clinical information except that the patients were newly diagnosed with prostate cancer and were instructed to evaluate the PSMA PET using the corresponding CT-images as they would do in their daily clinical practice. Likewise, the observers were blinded to the evaluation of the BS. Cases of disagreement were resolved by consensus.
Best valuable comparator
In most cases, a histopathologic reference standard was not available for ethical and practical reasons. For the present study, we defined a “best valuable comparator” (BVC) for the presence or absence of bone metastases at the patient level similar to previous diagnostic studies of bone metastases [
6,
8,
15]. The BVC was based on all available imaging results available at the time of staging as well as supplementary and/or follow-up imaging (all modalities). A minimum of 12 months of clinical and laboratory follow-up was required. Patients with PSA < 0.1 ng/mL after radical prostatectomy without any systemic treatment were categorized as having no bone metastases at the time of staging [
16].
Statistics
Diagnostic accuracy was calculated among patients with a valid BVC. Sensitivity analyses were performed for each component of the diagnostic accuracy endpoints (sensitivity, specificity, positive, and negative predictive values), where patients with equivocal findings were first considered positive for metastases (pessimistic analyses) and then calculated as negative for metastases (optimistic analysis). The results are summarized by the mean or median values, standard deviations and ranges. Statistical analyses were performed using STATA®11 (StataCorp LP, College Station, TX, USA). All results are reported with 95% confidence intervals (CIs).
Ethics
Due to the retrospective nature of the study, no formal approval from the ethics committee was required according to our national legislation. The study was approved by the Danish Data Protection Agency (approval number 2008-58-0028).
Discussion
The present study investigated the added value of 68Ga-PSMA-11 PET/CT in patients with newly diagnosed PCa who recently underwent BS. 68Ga-PSMA-11 PET/CT diagnosed bone metastases in 10% of patients with negative BS results, provided a firm diagnosis in 20 of 22 patients referred for 68Ga-PSMA-11 PET/CT due to equivocal BS results, and confirmed bone metastatic disease in all patients with positive BS results, but 68Ga-PSMA-11 PET/CT also identified a notable proportion of patients in whom PSMA-avid lesions in the ribs were false positive.
Pyka et al. published the first comparison of
68Ga-PSMA-11 PET/CT with BS in patients with PCa, including 37 patients at primary staging [
8]. They found a sensitivity and specificity of
68Ga-PSMA-11 PET/CT of 100% at primary staging. Likewise, Lengana et al. reported a sensitivity of 96% and specificity of 100% of
68Ga-PSMA PET/CT for the detection of bone metastases at primary staging [
7]. In the present study, the sensitivity was comparable to that found in prior studies, whereas the specificity was slightly lower than the previously reported specificity, which might be explained by the four patients with false-positive PSMA-avid uptake in the ribs.
68Ga-PSMA-11 PET/CT has previously shown an ability to detect bone metastases in a proportion of patients without bone metastases apparent on BS. Lengana et al. reported that PSMA PET/CT revealed bone metastases in 8.4% of patients with negative BS results [
7], which is in line with the present findings in which 10% of the patients without bone metastases on BS were considered metastatic by PSMA. Among patients with bone metastases on BS, PSMA PET/CT confirmed M1 bone disease in all patients and showed more metastasis-suspected lesions than those detected by BS. These findings are in line with recent observations [
7].
Although PSMA PET/CT was indeterminate in some cases, it provided a definitive imaging diagnosis among 96.4% of the patients in this population at the primary staging. These findings are comparable to findings in patients with biochemical recurrence after curatively intended treatment, where PSMA has been shown to provide a definite diagnosis in 99% of patients [
17].
The fact that four patients obtained false-positive results based on PSMA PET/CT was unexpected and has not been reported in prior studies [
8,
17]. In each of these four patients presenting with a total of five PSMA-avid lesions of non-prostatic origin, morphologic changes were observed in the corresponding CT images. However, biopsy ruled out bone metastases in one patient. In addition, the incorporation of a PSA < 0.1 ng/mL 12 months after radical prostatectomy served as a relatively reliable verification of non-metastatic diseases in three patients. In accordance with the BVC, PSA-negative metastatic PCa in the mentioned patients is very unlikely. The false-positive findings were likely not due to reader inexperience; two of the observers (HDZ and AAO) were highly experienced and evaluated PSMA PET/CT in numerous trials [
5,
6,
17]. After we performed the blinded evaluation of the
68Ga-PSMA-11 PET/CT in the present study, a number of cases with PSMA-avid bone uptake in benign skeletal conditions were published [
18‐
21], revealing that fibrous dysplasia in the ribs or rib fractures may be PSMA-avid. In summary, these findings emphasize the need for careful interpretation of PSMA PET/CT in rib lesions.
The strength of the present study is the consecutive inclusion of patients. Due to the unique national security number in Denmark, it was possible to perform a thorough follow-up on patients included in the study, even if the patients moved across regions within the country. In the present study, detailed follow-up data were available in most patients. A shortcoming is that histologic confirmation is rarely available in imaging studies of the bone. However, performing biopsy routinely is not ethically reasonable, and consequently, a composite endpoint (BVC) was applied in the present study. A BVC has previously been used in diagnostic studies of bone metastases, including studies comparing
68Ga-PSMA-11 PET/CT to BS [
6‐
8,
15]. One of the limitations of the BVC is that the index test itself often plays a key role in the definition of the BVC, as the interpretation of a new and promising method might be unreasonably trusted in the final conclusion based on the BVC. In addition, the extent of clinical, imaging, and biochemical data available for the BVC have seldom been reported. Here, we included detailed information for a minimum of 12 months for all patients with at least one lesion according to either BS or PSMA. A minority of patients did not have data allowing for a definitive conclusion of BVC. For these reasons, it is imperative to consider the intrinsic verification bias when using a BVC. For the present study, the limitations are reflected in the sensitivity of 1.00, which is likely to be overestimated.
Patients were excluded if they had received any kind of treatment for PCa prior to BS and
68Ga-PSMA-11 PET/CT, which indicates that the scans were not influenced by the negative effects of ADT on PSMA PET/CT as described previously [
22].
Despite the consecutive inclusion of patients, the present population was biased because more than 20% of the patients were referred to
68Ga-PSMA-11 PET/CT due to equivocal findings in BS. The study was not a head-to-head comparative diagnostic test accuracy study. Therefore, no comparative analysis of diagnostic accuracy for the detection of bone metastases by BS was conducted. Likewise, the high proportion of patients with equivocal BS does not reflect the true frequency of equivocal findings in BS when SPECT/CT is applied, which in unselected populations has been shown to be approximately 10% at the time of initial staging [
23]. In the present setting, approximately 80% of the patients had a SPECT/CT performed as an add-on to the planar bone scintigraphy which is a limitation of the study. However, the use of SPECT/CT in patients with planar BS with possible benign or equivocal lesions reflects everyday clinical practice at our institution, where patients with a normal bone scintigraphy or harboring several obvious malignant lesions are not succumbed to additional SPECT/CT.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.