Comparison of 11C-choline Positron Emission Tomography/Computed Tomography (PET/CT) and Conventional Imaging for Detection of Recurrent Prostate Cancer

We aimed to compare 11C-choline positron emission tomography/computed tomography (PET/CT) with conventional imaging, including pelvic magnetic resonance imaging (MRI), contrast-enhanced chest, abdomen, and pelvic computed tomography (CT), and bone scintigraphy, for prostate cancer restaging. Thirty patients (median prostate-specific antigen [PSA: 11.8 ng/mL]) with suspected recurrent prostate cancer following definitive treatment underwent 11C-choline PET/CT and conventional imaging, including pelvic MRI, contrast-enhanced chest, abdomen, and pelvic CT, and bone scintigraphy. The results were compared with regard to patient- and lesion-based diagnostic performance for local recurrence, and for lymph node and bony metastases using receiver operating characteristic (ROC) analysis and McNemar’s test. Documented local recurrence and node and bony metastases were present in 11 (36.7%), 10 (33.3%), and 17 (56.7%) cases, respectively, of the enrolled patients. Patient-based sensitivity / specificity / accuracy / area under the ROC curve for 11C-choline-PET/CT for diagnosing local recurrence were 90.9% / 94.7% / 93.3% / 0.975 and for conventional imaging were 90.9% / 100% / 96.7% / 1.0. Those who underwent 11C-choline-PET/CT for node metastasis were 90.0% / 95.0% / 93.3% / 0.925 and for conventional imaging were 70.0% / 95.0% / 86.7% / 0.905. Those who underwent 11C-choline-PET/CT for bone metastasis were 94.1% / 92.3% / 93.3% / 0.991 and who underwent conventional imaging were 94.1% / 84.6% / 90.0% / 0.982. No significant differences were observed among them. The lesion-based detection rate of 11C-choline PET/CT for local recurrences and node and bone metastases as compared to conventional imaging was 92.9% (13/14) vs. 92.9% (13/14); 87.1% (27/31) vs. 54.8% (17/31); and 96.9% (219/226) vs. 90.3% (204/226) respectively, with significant differences noted for detection of node and bone lesions (p=0.0044 and p=0.00030, respectively). 11C-choline-PET/CT is more accurate in the detection of recurrent prostate cancer nodes and bony metastatic lesions compared to conventional imaging and has the advantage of restaging the disease in a single step.


Introduction
In Western Europe and North America, prostate cancer is the most common tumor type in men and the second-most frequent cause of all deaths from cancer [1]. Although prostate cancer often develops slowly, it can later show an aggressive pattern and sometimes recurrence. For patients with recurrent prostate cancer, it is important to determine whether there is localized recurrent disease, metastasis to lymph nodes or bone, or a combination of localized recurrent and metastatic disease. This affects the subsequent management, such as consideration of salvage therapy for localized recurrence, systemic treatment for metastatic disease, or a combination of these. Computed tomography (CT), bone scintigraphy, pelvic magnetic resonance imaging (MRI), and trans-rectal ultrasound examinations have been traditionally used to determine the localization of recurrent or metastatic disease, though they lack the adequate sensitivity and accuracy. In recent years, positron emission tomography/computed tomography (PET/CT) using 11 C-or 18 F-choline has emerged as a promising molecular imaging tool and can provide full-body examination results in a single step [2][3][4].
No known study has been presented so far that directly compares the diagnostic capability of 11 C-choline PET/CT with that of conventional imaging, including pelvic MRI, contrastenhanced chest, abdomen, and pelvic CT, and bone scintigraphy, for restaging prostate cancer that has recurred following definitive treatment. Picchio et al. [5] compared 11 C-choline PET with conventional imaging, including pelvic MRI, CT, bone scintigraphy, and trans-rectal ultrasound, for restaging of prostate cancer, but not PET/CT, though PET using 11 C-choline was done while other modalities (pelvic MRI, CT, bone scintigraphy, trans-rectal ultrasound) were not performed in all of the enrolled patients. In addition, Beauregard et al. [6] compared 11 Ccholine PET/CT with conventional imaging, including pelvic CT and bone scintigraphy, for staging and restaging of prostate cancer, though pelvic MRI was not performed. Thus, the superiority of 11 C-choline PET/CT over conventional imaging modalities, including pelvic MRI, contrast-enhanced chest, abdomen, and pelvic CT, and bone scintigraphy, for restaging of prostate cancer in patients who have undergone definitive treatment, has yet to be confirmed. The purpose of the present study was to clarify which method is preferable for the detection of local recurrence, as well as lymph node and bone metastasis in patients suspected to have recurrent prostate cancer.

Materials And Methods
This retrospective study was performed in accordance with the principles of the declaration of Helsinki. The institutional review board (Hyogo University Hospital, Japan) approved this retrospective study (No 2019). Informed consent was obtained from each patient after the procedure details were fully explained.

Patients
A total of 30 males (mean age 71.3 years, range 47-90 years) with suspected recurrent prostate cancer after receiving definitive treatment were included in this study and underwent 11 Ccholine PET/CT as well as conventional imaging, including pelvic MRI, contrast-enhanced chest, abdomen, and pelvic CT, and bone scintigraphy, at our institution between October 2015 and July 2017, with a maximum interval of two weeks between examinations. The median serum prostate-specific antigen (PSA) level in our cohort was 11.8 ng/mL (range 0.23-946 ng/mL). At the time of the examination, 26 (86.7%) of the 30 patients had received androgen-deprivation therapy. Additional patient details are shown in Table 1.

Characteristics Value
Age (

C-choline PET/CT
11 C-choline was synthesized using a commercial module, as described by Hara [7], and a CYPRIS-325R cyclotron (SHI, Tokyo, Japan). Acquisition of emission scan images from the midthigh to the head was started six minutes after intravenous injection of 3.0 MBq/kg body weight of 11 C-choline. All PET/CT examinations were performed using a PET/CT scanner equipped with a 64-multidetector computed tomography device (Gemini TF64; Philips Medical Systems, Eindhoven, The Netherlands). The whole-body PET image acquisition in 3D mode was performed from the mid-thigh to the top of the head (1.5 minutes per bed position; 6-8 bed positions) and the obtained images were reconstructed using the ordered subsets expectation maximization reconstruction algorithm (33 subsets, three iterations, 4 mm per slice), with attenuation correction based on low-dose CT (120 kVp, 100 mA, slice thickness 2 mm, transverse field of view 600 mm), which was also used for anatomical correlations.

Contrast-enhanced CT
Pre-contrast and contrast-enhanced multi-detector CT (MDCT) images of the chest, abdomen, and pelvis were obtained using a 128-detector row CT device (SOMATOM Definition AS+, Siemens Medical Solutions, Erlangen, Germany) at 120 kV and effective mAs of 220 (CARE Dose4D), beam pitch of 0.6, collimation of 1.2 mm×32, and B31 + medium smooth + image reconstruction. Blood creatinine and the estimated glomerular filtration rate (eGFR) level were checked in all the patients before the examination. Iodinated contrast material (IopamironInj, Bayer Schering Pharma, Berlin, Germany) containing 300 mg of iodine per ml at a dose of 600 mg iodine per kilogram of body weight was intravenously administered using a power injector. The scan was started at 120 seconds after the injection to obtain venous phase images.

Bone scintigraphy
Whole-body bone scintigraphy was performed using a dual-head gamma camera (Forte, Hitachi, Tokyo, Japan) after intravenous administration of approximately 555 MBq of 99m Tcmethylene diphosphonate ( 99m MDP) (FujiFilm RI Pharma, Tokyo, Japan). Following the injection, the patient was orally hydrated and asked to void their bladder frequently as well as immediately prior to the scan. Total body images were obtained approximately three hours after administration of 99m MDP, with simultaneous anterior and posterior whole-body acquisition. Static additional acquisitions were acquired when needed. Single-photon emission computed tomography (SPECT) images were not acquired.

Imaging analysis
Two board-certified nuclear medicine physicians with no knowledge of the other imaging results or the final diagnosis interpreted the 11 C-choline PET/CT findings, then used a fivepoint scale (1: definitely absent, 2: probably absent, 3: indeterminate, 4: probably present, 5: definitely present) to grade lesions in each patient based on consensus. Discordant readings by the two observers were resolved by a subsequent consensus review.
Focal 11 C-choline activity in the prostate or a seminal vesicle greater than that in the adjacent background and not due to excreted radiotracer in urine was graded as 4 or 5 regardless of a corresponding structural abnormality. Lymph nodes were graded as 4 or 5 if distinct focal activity on the PET images was co-registered to a visible lymph node in CT images regardless of size. Focal skeletal sites of uptake above background marrow activity were graded as 4 or 5 unless inflammatory change, degenerative change, or fracture was evident. Semiquantitative analysis of abnormal radiotracer uptake for each lesion was also retrospectively performed using a maximum standardized uptake value (SUVmax), calculated as follows: SUV = volume of interest (VOI) radioactivity concentration (Bq/mL)/[injected dose (Bq)/patient weight (g)].
SUVmax, defined as the highest SUV value for pixels with the highest count within the VOI, was determined and recorded for the focal areas of uptake.
Two board-certified observers, a double board-certified nuclear medicine physician and radiologist, and a board-certified radiologist, with no knowledge of the other imaging results, or the final results, interpreted the conventional imaging findings and then used the same fivepoint scale to grade the lesions in each patient based on consensus. Discordant readings by the two observers for each modality were resolved by a subsequent consensus review.
The status of local recurrence was evaluated by pelvic MRI results. When T2WI findings were reviewed, a mass or soft-tissue area showing abnormal signal intensity was considered a positive finding (confidence score 4 or 5). In a review of DWI findings and apparent diffusion coefficient (ADC) maps, the presence of a lesion with high focal signal intensity on a DW image and low signal intensity on the ADC map relative to the background was considered to be a positive finding (score 4 or 5). Two positive interpretations of images in two sequences were given a confidence score of 5, while one positive interpretation in two sequences was given a score of 3 or 4. The number, location, and long-axis dimension of the suspected locally recurrent lesions were recorded.
The status of lymph node metastasis was evaluated by pelvic MRI as well as contrast-enhanced chest, abdomen, and pelvic CT findings. For scoring nodal status based on CT and MRI results, the following diagnostic criteria were used: short-axis diameter of node ≥10 mm, score 5; 8-9.9 mm, score 4; 5-7.9 mm, score 3; 1-4.9 mm, score 2; not seen, score 1. The number, location, and short-axis dimension of the suspected nodal metastatic lesions were recorded.
Bone metastasis status was also evaluated based on the combination of bone scintigraphy, contrast-enhanced chest, abdomen, and pelvic CT, and pelvic MRI findings. In bone scintigraphy images, a lesion was defined as malignant according to standard pathological criteria of intensity, localization, and the extension of 99m MDP uptake, with physiological uptake due to inflammatory change, degenerative change, or fracture excluded (score 4 or 5). In CT images, a lesion with osteoblastic or osteolytic change was graded as 4 or 5, unless inflammatory change, degenerative change, or fracture was evident. A lesion showing low signal intensity on T1WI and abnormal signal intensity on a DWI/ADC map is considered to be a positive finding (score 4 or 5). Inside the pelvis, three positive interpretations in three modalities were given a confidence score of five; two positive interpretations in three modalities a score of 4 or 5, and one positive interpretation a score of 3 or 4. Outside the pelvis, two positive interpretations of images in two sequences were given a confidence score of 5 and one positive interpretation in two sequences received a score of 3 or 4.

Gold standard
The final diagnosis was obtained based on histological confirmation, radiological imaging findings, or clinical follow-up results at least six months, including PSA level, MRI, CT, bone scintigraphy, and 11 C-choline PET/CT findings.

Statistical analysis
To determine the utility of each imaging modality, receiver-operating-characteristic (ROC) analysis was employed. A ROC contrast estimation was utilized to compare the diagnostic capability of the two types of modalities on a per-patient basis. To test whether the area under the ROC curve (AUC) values were different, correlations of the testing methods were accounted for in the analysis. Tests for differences in sensitivity, specificity, and accuracy between the types of modalities were conducted using McNemar's test. To calculate the sensitivity and specificity of each modality, scores of 4 and 5 were considered positive. A p-value of less than 0.05 was considered to indicate a statistically significant difference. All statistical analyses were performed using SAS software, version 9.3 (SAS Institute).

Local recurrence
Eleven (36.7%) of the 30 patients were found to have a total of 14 local recurrent lesions (12 in the prostate gland, two in the seminal vesicles). The SUVmax values and long-axis diameters for these locally recurrent lesions as shown by 11  shows that local recurrence in the prostate gland could be diagnosed by both 11 C-choline PET/CT (Figures 1-2) and pelvic MRI (Figure 2).     (d) Axial T2-weighted MR image shows hypointense area measuring 35×39 mm in the bilateral prostate gland with restricted diffusion on (e) DWI (arrows), with a score of 5 for diagnosing local recurrence in the prostate gland.

Lymph node metastasis
Ten (33.3%) of the 30 patients had a total of 31 metastatic lymph nodes (six external iliac, five internal iliac, two obturator, four common iliac, eight abdominal para-aortic, three mediastinal, and three supraclavicular). The SUVmax values and short-axis diameters of those metastatic nodes as shown by 11  Case 1 shows that a tiny pelvic lymph node metastasis could be diagnosed by 11   (h) Contrast-enhanced CT shows mild swelling of the left external iliac node measuring 7 mm in short diameter (arrow), with a score of 3 for diagnosing lymph node metastasis.
Four false-negative cases were determined by 11 C-choline PET/CT and consisted of two tiny para-aortic metastatic nodes measuring 4 and 5 mm, respectively (one tiny internal iliac node measuring 6 mm and one tiny external iliac node measuring 6 mm). Those four nodal lesions showed no or faint 11 C-choline uptake (SUVmax: 0.89, 1.21, 1.49, and 1.77). Fourteen falsenegative cases were determined by conventional imaging and were all tiny lesions whose shortaxis diameter measured less than 8 mm. One false-positive case determined by 11 C-choline PET/CT showed 11 C-choline nonspecific uptake (SUV max: 3.61) in the inguinal node. One false-positive case was also determined by conventional imaging and had non-specific enlarged hilar and mediastinal nodes measuring larger than 10 mm.   (Figures 4-5), pelvic MRI ( Figure 5), and bone scintigraphy ( Figure 6). Table 4 presents the distribution and visualization rates of bone metastases shown by 11 C-choline PET/CT and conventional imaging.    Eight false-negative cases shown by 11 C-choline PET/CT were composed of one cervical spine, two thoracic spine, two lumbar spine, one sacrum, and one ischium case, while 22 falsenegative cases shown by conventional imaging were composed of one cervical spine, two thoracic spine, one lumbar spine, one sacrum, one iliac bone, one ischium, one sternum, one clavicle, five ribs, three scapula, two upper extremities, two lower extremities, and one skull case. One false-positive case shown by 11 C-choline PET/CT had moderate 11 C-choline uptake (SUVmax: 4.13) observed in a normal thoracic vertebra, and false-positive cases shown by conventional imaging were degenerative changes in the cervical and lumbar vertebra, which showed strong 99m MDP uptake and osteoblastic changes in CT images.

All lesions
There was a total of 271 recurrent or metastatic lesions in 25 (83.3%) of the 30 patients.

Discussion
The present is the first study to directly compare the diagnostic capability of 11 C-choline PET/CT with that of conventional imaging, including pelvic MRI, contrast-enhanced chest, abdomen, and pelvic CT, and bone scintigraphy, for restaging prostate cancer in patients who have undergone definitive treatment. We found that the lesion-based sensitivity of 11 C-choline PET/CT for detection of node and bone metastases in patients with recurrent prostate cancer was significantly higher as compared to conventional imaging modalities, with 11 C-choline PET/CT and conventional imaging exhibiting equal sensitivity for the diagnosis of local recurrent disease. Although patient-based analysis did not reveal a significant difference between the two types of imaging, our results confirm those of a previous study, in which Picchio et al. [5] demonstrated that the patient-based sensitivity of 11 C-choline PET and conventional imaging for restaging prostate cancer was 47% and 49%, respectively.
In Western countries, 11 C-and 18 F-choline PET/CT have been successfully used for restaging prostate cancer in patients with biochemical disease recurrence after undergoing definitive therapy, especially in those with a PSA level >1.0 ng/mL [2,4,8]. Another systematic review and meta-analysis of 19 selected studies with a total of 1,555 patients reported a pooled sensitivity of 85.6% and a pooled specificity of 92.6% [9]. Picchio et al. concluded that the routine use of 11 C-choline PET/CT for restaging prostate cancer after a radical prostatectomy cannot be recommended for patients with a PSA value <1 ng/mL [8]. Furthermore, they noted that in addition to the PSA value, PSA doubling time and other clinical and pathological features, such as a locally advanced tumor (pT3a-T4) or nodal involvement at the initial staging, should be considered when considering patients for 11 C-choline PET/CT.
The current gold standard for the diagnosis of a locally recurrent prostate cancer in patients who have undergone a radical prostatectomy procedure and radiation therapy is a dynamic gadolinium contrast-enhanced MRI [10][11]. Dynamic contrast-enhanced MRI was not performed in the present cases, thus the sensitivity of conventional imaging for local recurrence may have been underestimated. Kitajima et al. [3] compared multi-parametric MRI modalities, including T2WI, DWI, and dynamic contrast-enhanced imaging, with 11 C-choline PET/CT for detection of local recurrence after a radical prostatectomy in 87 patients and reported that patient-based sensitivity, specificity, and accuracy of multi-parametric MRI were 88.5%, 84.6%, and 87.4%, respectively, whereas they were 54.1%, 92.3%, and 65.5%, respectively, for 11 C-choline PET/CT.
Evaluation of lymph node metastasis is crucial for restaging patients with PSA failure after treatment. In a study of 25 prostatectomy patients with PSA failure using pelvic and retroperitoneal lymphadenectomy histopathology results as a reference [12], the values for sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of 11 Ccholine PET/CT for depicting nodal metastasis were 100%, 66%, 90%, 100%, and 92% respectively, in patient-based analysis and 64%, 90%, 86%, 72%, and 77% respectively, in lesion-based analysis. The low negative predictive value for lesion-based analysis in that the study indicated the limited performance of 11 C-choline PET/CT in detecting microscopic nodal metastasis, whereas the high positive predictive value is an important result for facilitation of appropriate treatment. Kitajima et al. [3] also compared pelvic MRI and 11 C-choline PET/CT results for detecting pelvic nodal metastasis following a radical prostatectomy in 70 patients, and they reported that the patient-based sensitivity, specificity, and accuracy values for pelvic MRI were 64.0%, 85.0%, and 70.0%, respectively, whereas those of 11 C-choline PET/CT were 90.0%, 100%, and 92.9%, respectively. In the present study, we also noted the superiority of 11 C-choline PET/CT as compared to MRI results.
In the present patient series, the ability to diagnose metastasis to the bone outside of the pelvis was significantly different between 11 C-choline PET/CT and conventional imaging (bone scintigraphy and CT), whereas with regard to metastasis to bone inside the pelvis, it was the same between the imaging types because of the features of MRI. Bone scintigraphy is a low-cost whole-body clinical examination method commonly used to detect skeletal involvement in patients with prostate cancer, though it is known to be inferior to 11 [13]. Similar to the findings of the present cohort, Kitajima et al. [3] reported equally excellent accuracy for the diagnosis of pelvic bone metastases after undergoing a radical prostatectomy in 95 patients, with values for patient-based sensitivity, specificity, and accuracy for pelvic MRI found to be 87.5%, 96.2%, and 94.7%, respectively, whereas those for 11 C-choline PET/CT were 81.3%, 98.7%, and 95.8%, respectively. Although the superiority of 11 C-choline PET/CT as compared to the whole-body MRI is controversial [13][14] and further analysis is needed, we consider that it is necessary to use 11 C-choline PET/CT or a whole-body MRI for a careful examination to check for wholebody bone metastasis. In contrasting reports, a group reported the superiority of the wholebody MRI [13], while another found that 11 C-choline PET/CT was superior [14].
Our study has some limitations, including the small number of patients enrolled from a single institution. A prospective multicenter trial with a larger cohort would help to clarify the exact role of 11 C-choline PET/CT in clinical decision-making and long-term outcomes in clinical settings. Also, the enrolled population was heterogeneous, as it included treatment-naive patients, as well as those who did or did not undergo hormonal treatment. Such heterogeneity likely introduced confounding factors into the analysis. The follow-up period was relatively short. In addition, the ideal gold standard for any analysis would be a histological confirmation of obtained findings. On the other hand, clinical follow-up results are considered to be valid for evaluations of diagnostic accuracy and response to therapy, and it would have been unethical to investigate all PET/CT-detected lesions with the use of an invasive procedure. Although the positive findings are easy to confirm, negative findings only indicate that it was not possible to acquire positive findings during a follow-up examination, making it uncertain whether the findings are truly negative. Furthermore, new and more sensitive PET tracers for prostate cancer, such as 18 F-FACBC and 68 Ga-PSMA, have been recently introduced for clinical use in Western countries [4], though they are not yet available in Japan. One study demonstrated that 68 Ga-PSMA PET/CT showed a higher detection rate than 11 C-choline PET for lymph nodes as well as bone lesions in patients with prostate cancer [15]. Finally, SPECT/CT findings were not available in the present study. A hybrid imaging method (SPECT/CT) may increase the sensitivity and specificity of bone scintigraphy by identifying benign bone conditions with increases in bone turnover and whenever equivocal findings of planar bone imaging are noted [16].