11C-Choline positron-emission tomography/computed tomography and transrectal ultrasonography for staging localized prostate cancer


Ludwig Rinnab, University of Ulm, Department of Urology, Prittwitzstrasse 43, 89075 Ulm, Germany.
e-mail: ludwig.rinnab@medizin.uni-ulm.de



To evaluate and compare the role of 11C-choline positron emission tomography (PET) and transrectal ultrasonography (TRUS) in the preoperative staging of clinically localized prostate cancer.


Fifty-five consecutive patients with biopsy-confirmed prostate cancer had TRUS and 11C-choline PET as a part of their clinical staging programme before radical retropubic prostatectomy (RP). The PET images were prospectively interpreted by a consensus decision of two nuclear medicine physicians and one radiologist with special expertise in the field. The TRUS was done by one experienced urologist. The criteria evaluated prospectively in each patient were extracapsular extension (ECE), seminal vesicle invasion (SVI) and bladder neck invasion (BNI). The results were compared with the histopathological findings after RP.


At pathology, 32 patients were classified pT2, 16 as pT3a and three had pT3b lesions. In four patients the histopathological examination showed pT4 with BNI. The overall accuracy of PET in defining local tumour stage (pT2 and pT3a−4) was 70%; the overall accuracy by TRUS was 26%. PET was more sensitive than TRUS for detecting ECE (pT3a) and SVI (pT3b) in advanced stages, and in pT4 stages. The sensitivity and positive predictive value (PPV) (95% confidence interval) in stages pT3a–pT4 for PET were 36 (17–59)% and 73 (39–89)%. The sensitivity and PPV in stages pT3a–pT4 for TRUS were 14 (3–35)% and 100 (29–100)%.


11C-choline PET and TRUS tended to understage prostate cancer. This series shows the current limited value of TRUS and PET for making treatment decisions in patients with clinically localized prostate cancer, especially if a nerve-sparing RP is considered. Treatment decisions should not be based on TRUS and 11C-choline PET findings alone. In future studies, the combination of metabolic and anatomical information of PET and endorectal magnetic resonance imaging should be evaluated, as this might optimize the preoperative staging in prostate cancer.


radical prostatectomy


extracapsular extension


seminal vesicle involvement


positron emission tomography


standardized uptake value


(positive) (negative) predictive value.


Accurate clinical staging based on a physical examination, TRUS and current radiographic techniques for detecting localized, curable prostate cancer is limited. Precise imaging techniques that are useful for devising tailored protocols in localized disease are needed. Of patients operated on for clinically localized tumours 40–50% show extraprostatic spread on pathological examination [1]. In patients treated by radical prostatectomy (RP), extracapsular extension (ECE) or seminal vesicle involvement (SVI) are associated with recurrent disease [2], and a nerve-sparing procedure might be contraindicated. TRUS is usually the first imaging method and its main role is primarily to guide prostatic needle biopsy.

Molecular imaging approaches using 11C- or 18F-labelled choline derivatives and positron emission tomography (PET) were used for detecting primary or relapsing prostate cancer. This technique makes use of the increased content of phosphorylcholine [3,4], up-regulated key enzymes of choline metabolism, increased phosphatidyl choline turnover and metabolic flux of radiolabelled choline through phospholipid biosynthesis and degradation in prostate cancer. 11C-choline was reported as an appropriate tracer for noninvasive imaging of prostate cancer [5–7].

However, with increasing options for organ-sparing therapy, e.g. interstitial brachytherapy, intensity-modulated radiotherapy or cryosurgery, the precise evaluation of intraprostatic extension of prostate cancer is gaining clinical importance. In addition, with current efforts to improve nerve-sparing surgical techniques and postoperative potency, the extent of ECE might change the surgical approach required for preserving the neurovascular bundle. If local ECE can be identified accurately before surgery, the ipsilateral neurovascular bundle can be excised widely to increase the likelihood of negative surgical margins.

To evaluate the staging capability of 11C-choline PET in clinically localized prostate cancer, we compared images obtained by TRUS and PET with the histopathological findings after RP. To our knowledge, this is the first study to directly compare these two imaging techniques in men with localized prostate cancer.


An unselected population of 55 patients with biopsy-confirmed prostate cancer had TRUS and 11C-choline PET/CT as a primary staging, the latter ≥4 weeks after needle biopsy. Based on clinical criteria (age, performance status, comorbidity and chest X-ray) and standard preoperative staging (PSA level, DRE, TRUS, bone scan) 54 patients (median age 64 years, range 43–74) were thought to have clinically localized and locally advanced disease, and were offered RP (Table 1). Patients with a PSA level before RP of >10 ng/mL [8] had a bone scan; in no patient was the scan positive for bone metastasis. Of the 55 patients, one was suspected to have clinical T4 disease. No patients were denied RP on the basis of TRUS and PET findings. The median (range) PSA level was 9.04 (0.2–66.6) ng/mL. In the present study, Partin Tables were not used during the consultation with the patients. The study was approved by the local ethics committee of the University of Ulm and the national radiation protection authorities. All patients gave written informed consent.

Table 1.  The distribution of biopsy Gleason score, PSA level, clinical and pathological stage
VariableNumber of patients
  1. Clinical and pathological staging using the 2002 TNM classification.

PSA, ng/mL
 0.0–4.0 3
Gleason score
 2–4 4
 5 9
 8–10 4
Clinical stage
 cT1a 1
 cT1b 0
 cT2b 4
 cT2c 4
 cT3a 3
 cT3b 1
 cT4 1
Pathological stage
 pT2a 9
 pT2b 4
 pT3b 3
 pT4 4
 pN1 2

One urologist experienced in prostatic TRUS conducted the assessment using a 7.5-MHz transducer (Combison 530 D, ultrasound scanner A 00595, transducer 14843, General Electric, Milwaukee, WI, USA). The patient was scanned in the left lateral decubitus knee-chest position. Sagittal and transverse images of the prostate and periprostatic area were obtained from the level of the seminal vesicles to the apex. The diameter of the tumour, gland volume, tumour location, echo structure of the lesion, capsular penetration, SVI and periprostatic fat involvement were assessed. A TRUS-guided biopsy was taken according to a 10-core systematic biopsy scheme (standard sextant scheme plus transition and lateral zones). In all, 47 patients had results positive for cancer at the first biopsy, and eight at repeated biopsy; all patients had 11C-choline PET at a mean of 4 weeks after the positive biopsy.

11C-choline was synthesized according to the loop-methylation method described by Wilson et al.[9]. The patients were fasted for 5–8 h before 11C-choline PET/CT, using an integrated PET/CT scanner (GE Discovery LS, General Electric) after an i.v. injection with 1112 ± 131 MBq of 11C-choline. PET images were acquired 5–10 min after the injection, starting from the distal margin of the pelvic floor with a 3-min acquisition time per bed position [10]. Contrast-enhanced CT (140 kV, 160 mA, pitch 1.5 mm) was acquired with 120 mL of i.v. non-ionic contrast medium as a bolus (Ultravist, Schering, Germany) immediately before the PET acquisition. PET images were reconstructed with the iterative reconstruction ordered-subset expectation maximum likelihood algorithm of the manufacturer, after attenuation correction based on the CT data set. Consecutive transverse PET/CT slices of 4.25 mm thickness were generated.

All patients had a standard RP using a nerve-sparing technique or not, in conjunction with pelvic lymphadenectomy, according to current guidelines [11]. The resected prostate was weighed and the surface inked, followed by standard formalin fixation for 24 h. The urethral (apical) and the vesical (basal) margins (3–4 mm thick) were taken and sectioned para-sagittally at 3–4 mm intervals, perpendicular to the inked surface. The prostate was serially sectioned at 3–4 mm intervals perpendicular to the long axis from the apical to basal regions of the gland, until the junction of the seminal vesicles was reached. The transition zone from prostate gland to the seminal vesicles was embedded in separate cassettes, one for each side. The slices were further formalin-fixed and paraffin-wax embedded. Finally, microslices were placed on glass slides and stained with haematoxylin and eosin. The PET and TRUS findings were compared with the histological information from the step-sectioned RP specimens.

The PET/CT scans were visually assessed with no knowledge of the clinical data or results of previous imaging studies, as a consensus reading of two experienced nuclear physicians and two radiologists (respectively, F.M.M., 8 years in PET, 3 years in PET/CT experience; S.N.R. 13 years in PET, 4 years in PET/CT; N.M.B., 10 years in radiology; S.P. 8 years in radiology).

The data were interpreted with the information of the CT and PET alone and the fused images, the latter being the main method. If intraprostatic 11C-choline uptake was at least twice that of 11C-choline uptake in periprostatic soft tissue, pararectal fat tissue or in pelvic muscles [10,12] this was defined as prostate cancer in the visual analysis. ECE was suspected in cases in which there was accentuated 11C-choline accumulation in the transitional area between prostate and periprostatic tissue, and the corresponding CT slices showed an irregular boundary of the prostate. SVI was assumed when focal 11C-choline uptake was detected at the base of or within the seminal vesicles. Only 11C-choline uptake that was seen in two consecutive axial slices was interpreted as SVI. Care was taken to differentiate prostatic 11C-choline uptake from rectal activity. Capsular invasion cannot be differentiated in current PET/CT systems, due to the limited intrinsic spatial resolution of ≈4 mm. Potential movement artefacts [13] occurring between CT and PET acquisition furthermore obviate the possibility of evaluating capsular invasion in PET/CT images.

For the quantitative assessment, the mean and maximum 11C-choline standardized uptake value (SUVmean, max) of the focal uptake in suspected tumour lesions was determined using a circular 1 cm (diameter) region of interest. The SUV is defined as the measured activity concentration divided by the injected radioactivity, normalized to body weight [12].

To assess the tumour staging of PET and TRUS, the sensitivity and positive predictive value (PPV) were calculated. In this study sensitivity was the proportion of patients having a specific pathological tumour stage that was also detected by PET or TRUS. The PPV is the proportion of patients for whom PET or TRUS diagnosed a specific tumour stage and which was pathologically confirmed. Because there were few patients of each tumour stage the 95% CI was also calculated. First, the sensitivity and PPV were determined for each tumour stage, and second the tumour stages were classified into two groups, i.e. pT2a–pT2c and pT3a–pT4, and again the sensitivity and PPV were calculated. Furthermore, the overall accuracy for differentiating between these groups was computed. The McNemar test was used to investigate whether PET and TRUS equally often determined the correct tumour stage group. The correlation between SUVmean and SUVmax with PSA level, grading and Gleason score was examined using the nonparametric Spearman rank-order correlation coefficient.


In all, 53 of the 55 patients completed the study protocol and were suitable for evaluation; two were excluded because a precise T staging was not possible by TRUS. The histopathological assessment is listed in Table 1; two of the four men with pT4 were lymph node positive (Table 1). The primary prostate cancer could be visualized by PET as a focal accumulation within the prostate in 52 of 53 patients (SUVmean 3.89, sd 1.37; SUVmax 4.19, sd 1.49); in one patient PET failed to detect the tumour. TRUS also failed in this patient; in this case the histopathological result showed a pT3a tumour. Two patients had histologically confirmed lymph node metastasis; PET identified one (Fig. 1) and in the second there were micrometastases in a pelvic lymph node. The overall mean (95% CI) staging accuracy by TRUS was 26 (15–40)%, and for PET was 70 (56–82)%. The sensitivity and PPV for PET and TRUS for each stage is shown in Table 2, which also shows the diagnostic accuracies of the two imaging techniques: for ECE and SVI and pT4 PET was more sensitive than TRUS. In all cases TRUS and PET tended to understage the prostate cancer. The McNemar Test showed that PET significantly more often determined the correct tumour stage than TRUS (P < 0.001). There was a slight correlation (Fig. 2) between SUVmean and SUVmax and the PSA level (r = 0.58 and 0.61, respectively). There was no correlation between SUVmean and SUVmax with grading or Gleason score (SUVmean with grading, r = 0.14; SUVmean with Gleason score, r = 0.34; SUVmax with grading, r = 0.10; SUVmax with Gleason score, r = 0.31).

Figure 1.

Images from a 61-year-old man with an adenocarcinoma of the right prostate lobe, confirmed by biopsy, T2a, Gleason 7 (3 + 4) with an initial PSA level of 8 ng/mL. In the PET/CT before RP, a cT2b cN1 carcinoma was suspected. After RP and histopathology examination a pT2b pN1, Gleason 7 (3 + 4) was confirmed. (a) Contrast-enhanced four-row helical CT of the prostate; the transaxial slice shows no carcinoma. (b) PET shows focal 11C-choline uptake in the pelvis (red arrow). (c) The combination of metabolic information by 11C-choline PET and the exact topography by CT using the dedicated integrated PET/CT machine visualized the precise location of the prostate carcinoma in the right lobe (red arrow). (d) Contrast-enhanced four-row helical CT shows a pelvic lymph node (see grid) (cN1?). (e) The focal 11C-choline uptake indicates a lymph node metastasis (red arrow). The black arrows show physiological 11C-choline uptake in the bowel. (f) focal PET visualized the metabolic information (cN1) of the pelvic lymph node metastasis and integrated PET/CT reflected the precise anatomical location and the necessary information for the surgeon by molecular imaging. Black arrows show unspecific bowel activity and physiological diffuse 11C-choline uptake in the gluteal muscles.

Table 2. 
The sensitivity and PPV of 11C-choline PET and TRUS for tumour stages, individually and for the combined groups
Tumour stagen/N, sensitivity (95% CI)n/N, PPV (95% CI)
T2a 0/9, 0 (0.00–0.34)  0/2, 0 (0.00–0.84)
T2b 2/4, 0.5 (0.07–0.93)  2/8, 0.25 (0.03–0.65)
T2c 11/18, 0.61 (0.36–0.83)  11/31, 0.35 (0.19–0.55)
T3a 4/15, 0.27 (0.08–0.55)  4/7, 0.57 (0.18–0.90)
T3b 2/3, 0.66 (0.09–0.99)  2/3, 0.66 (0.09–0.99)
T4 1/4, 0.25 (0.00–0.81)  1/1, 1.0 (0.00–1.00)
T2a 1/9, 0.11 (0.00–0.48)  1/10, 0.1 (0.00–0.45)
T2b 1/4, 0.25 (0.00–0.81)  1/5, 0.2 (0.00–0.72)
T2c 0/18, 0 (0.00–0.19)  0/4, 0 (0.00–0.60)
T3a 1/15, 0.07 (0.00–0.32)  1/2, 0.5 (0.01–0.99)
T3b 0/3, 0 (0.00–0.71) 
T4 1/4, 0.25 (0.00–0.81)  1/1, 1.0 (0.00–1.00)
Combined groups, %
pT290 (74–98) 68 (52–82)
pT3a−436 (17–59) 73 (39–89)
pT235 (19–55) 58 (34–80)
pT3a−414 (3–35)100 (29–100)
Figure 2.

The correlation of serum PSA level, PET and SUVmean and SUVmax in 55 positive PET cases. The PSA level was 0.2–66.6 ng/mL.


In the present study, PET was more sensitive for ECE and SVI, and for pT4 tumours, than TRUS (Table 2), but both methods understaged microscopic extension of disease. Since the introduction of PET/CT several authors have reported the usefulness of PET in staging and re-staging of prostate cancer [14–17]. However, the available data and the few patients assessed in previous reports do not permit a judgement of the real value of this imaging method in staging prostate cancer. TRUS is usually the first imaging method and its use in prostate cancer is based on the fact that most cancers are shown as hypoechoic lesions, but only 20% of the hypoechoic lesions are malignant, so its role is primarily to guide prostate needle biopsy. However, >30% of men with clinically organ-confined disease have evidence of ECE on pathological analysis [18]. Thus the exact extent of the disease (ECE, SVI, pT4, lymph node involvement) should be clarified before surgery as precisely as possible, to plan rational therapy related to the clinical state and the prognosis of the prostate cancer (Fig. 3). General statements about the pathological stage can be made by using nomograms (e.g. Partin Tables, Kattan nomogram), which use various clinical variables (PSA level, Gleason-score, clinical state). However, the use of these nomograms is limited, e.g. in intermediate-risk patients, as defined by Partin et al.[19] (PSA 0–4 ng/mL, Gleason score 7; PSA 4–10 ng/mL, Gleason score 5–7; and PSA 10–20 ng/mL, Gleason score 2–7) the incidence of extraprostatic disease can be 13–58% or 35–70% using the Partin and Roach equations, respectively [19,20]. Consequently, there is a need for more accurate diagnostic tools. To our knowledge, the present is the first study to compare TRUS and 11C-choline-PET/CT in the primary staging of prostate cancer.

Figure 3.

a–d; Images from a 66-year-old man with adenocarcinoma of the prostate confirmed by biopsy (T2c cNx), Gleason 8 (4 + 4) with an initial PSA level of 10.7 ng/mL. Before RP, PET/CT detected a cT3a carcinoma (cN0) in the left prostate lobe in the apex, verified by the RP and histopathological result (pT3a pN0, 0/14, Gleason 8, 4 + 4). (a) Four-row helical contrast-enhanced CT could not detect the primary prostate carcinoma. (b) PET visualized the prostate carcinoma; the focal 11C-choline uptake indicated carcinoma (arrow). (c) Clear visualization and precise location by PET/CT; focal 11C-choline uptake indicates carcinoma (arrow). (d) TRUS shows cT3a carcinoma in the left prostate lobe of the apex (arrow).

May et al.[1] evaluated, in a similar study, the role of endorectal MRI and TRUS in staging clinically localized prostate cancer; 54 patients with biopsy-confirmed prostate cancer had TRUS and endorectal MRI (eMRI) before RP. The eMRI images were prospectively interpreted by two radiologists, and the findings compared with the histopathological results. The overall accuracy of eMRI in defining local tumour stage was 93% by radiologist A and 56% by radiologist B, and had a poor correlation with eMRI studies. The overall accuracy of TRUS was 63%, and higher than in the present study. In that study, eMRI was more sensitive than TRUS for detecting capsular penetration and SVI. Mullerad et al.[21] compared in a retrospective study of 106 patients with prostate cancer, the accuracy of eMRI, TRUS-guided biopsy and DRE for detecting the location of cancer in the prostate and seminal vesicle. They concluded that eMRI adds significantly to DRE or TRUS biopsy findings for prostate cancer location. The value of eMRI for detecting ECE and SVI was investigated in several studies of patients with localized prostate cancer. It was reported that the accuracy, sensitivity, specificity, PPV and negative PV (NPV) of eMRI was 77–91%, 23–75%, 84–97%, 36–83%, and 80–93%, respectively, for detecting SVI [22–26] and 64–91%, 13–66%, 82–97%, 40–66%, and 70–85%, respectively, for detecting ECE [22–25]. The accuracy, sensitivity and specificity for predicting the pathological local stage was 68–88%, 51–89% and 68–87%, respectively [24–26]. In agreement with these results, Nakashima et al.[27] recently reported a study of 95 patients with localized prostate cancer; they found an accuracy, sensitivity, specificity, PPV and NPV of 94.7%, 33.3%, 98.9%, 66.6% and 95.7%, respectively, for predicting SVI, and 74.7%, 57.1%, 82.1%, 57.1% and 82.1%, respectively, for detecting ECE. Recent studies showed that the MRI spectroscopic evaluation of citrate and choline metabolism improves the specificity of this diagnostic tool [28,29]. This technology is still developing and continued advances in accuracy and use are to be expected. Compared with the earlier results, 11C-choline PET in the present study had a sensitivity of 66% and a PPV of 66% for predicting SVI, and 27% and 57% for detecting ECE. For detecting a pT4 tumour the sensitivity was 25% and the PPV 100% (Table 2).

One of the major limitations of the present study was that there were few patients in the subgroups. Based on these results, multicentre studies are warranted, including measures of interobserver variability across different centres. However, to our knowledge this is the first study using a multimodal approach including TRUS, PET/CT and histopathological confirmation.

Based on the present results using 11C-choline PET and TRUS, it is possible that eMRI, if the published data are valid, might be preferable for predicting tumour site, size and local extension of prostate cancer. In future studies, the combined metabolic and anatomical information provided by eMRI and 11C-choline PET should be further investigated. This integrated approach might be promising in accurately assessing the presence, location and extent of prostate cancer.

In conclusion, the present results on the accuracy, sensitivity and PPV of 11C-choline PET and TRUS before RP had limited value, and subsequent treatment should not be altered based on TRUS or 11C-choline PET findings alone, especially if a nerve-sparing RP is planned. 11C-choline PET and TRUS might only assist in deciding the appropriate management for each patient. Macroscopic disease beyond the prostatic capsule and into the periprostatic fat or in seminal vesicle is often accurately detected, but both imaging methods understage the microscopic extension of prostate cancer.


We thank Dr Sandra Pauls, Department of Radiology University Hospital Ulm for participating in the consensus reading sessions of the PET/CT.


None declared.