Clinical outcomes for patients with recurrent prostate cancer could improve further if salvage treatments were individualized based on accurate delineation of cancer recurrence. Unfortunately, the localization of recurrent prostate cancer remains a significant diagnostic challenge, particularly at low PSA levels. In this study, it proved possible to identify local, lymphatic, and hematogenous routes of metastatic disease progression on the basis of whole-body FCH PET/CT with tumor detection rates directly correlated to PSA levels at the time of imaging.
Such a relationship between PSA level and the detection of prostate cancer by PET was first reported in studies using the tracer 11
C-labeled choline. In one study, 11
C-choline PET correctly detected tumor recurrence in 78 % of patients after external beam RT and 38 % of patients after RP, although there were no positive scan results at a PSA level below 5 ng/mL [12
]. Another study which focused on post-RP patients with biopsy-confirmed tumor recurrences reported a more favorable diagnostic sensitivity of 89 % and positive predictive value of 72 % for 11
C-choline PET/CT detection of recurrent prostate cancer at PSA <2.5 ng/mL [19
]. A direct relationship between positive findings on 11
C-choline PET/CT and PSA level in patients with recurrent prostate cancer was suggested in one study, with detection rates of 36, 43, and 62 % noted at PSA values of <1, 1 to <2, and 2 to <3 ng/mL, respectively [13
]. Higher detection rates with 11
C-choline PET/CT have also been associated with higher PSA velocities or shorter doubling times [20
The potential relationship between PSA level and tumor detection using fluorine-18 labeled choline analogs has also been investigated. Despite potential advantages to imaging with a fluorine-18 labeled PET tracer, findings from two studies also suggest that FCH PET/CT sensitivity diminishes as the PSA level approaches the 1–4 ng/mL range [9
]. Another recent study of FCH PET/CT estimated the diagnostic sensitivity in post-prostatectomy patients to be 20, 44, and 81.8 % at PSA levels of ≤1, >1 and ≤5, and >5 ng/mL, respectively [22
]. These previous studies are in accord with the current study. Notably, the current study was drawn from a predominantly Asian population and 32/50 of the patients enrolled were of Asian ancestry. In contrast, the previous studies were conducted in Europe at institutions where the clinical use of FCH was pioneered. The current study therefore serves to generalize some of the performance characteristics of FCH PET/CT to a population of prostate cancer patients that may differ from Western patients, particularly with regard to responsiveness to hormonal therapy and rates of disease progression [23
Sub-optimal detection rates at lower PSA values could potentially limit the value of FCH PET/CT in guiding treatment decisions for prostate cancer salvage therapy. This is because disease control rates for salvage treatments such as external beam radiation for recurrent prostate cancer following RP are also known to vary as a function of PSA level. For example, 6-year progression-free survival for salvage radiotherapy in patients with PSA levels of 0.5, 0.51–1.0, 1.01–1.5, and >1.5 ng/mL have been estimated at of 48, 40, 28, and 18 %, respectively [25
]. Thus, the PSA range associated with ideal detection rates for choline-based PET/CT may fall above the critical range of PSA values, where salvage therapy would provide the greatest clinical benefit. However, since detection failures with PET could be related to low tumor volume, there should be some prognostic relevance associated with a negative result on FCH PET/CT. Prognostic information gleaned from FCH PET/CT could therefore potentially aid in decisions on the timing of aggressive treatments. Further research is needed to determine the value of FCH PET/CT as a prognostic tool in patients with recurrent prostate cancer.
As a potential limitation, this study employed more than one method for confirming abnormal PET/CT findings, since biopsy confirmation was not always feasible to pursue in the clinical setting. Longitudinal follow-up and imaging were therefore necessary when biopsy data were unavailable for correlation with FCH PET/CT findings. The use of several methods as the standard of reference is not unusual for studies evaluating diagnostic imaging tests for recurrent prostate cancer due to the limited availability of tissue biopsy confirmations of metastatic disease in these patients.
In this study, visual analyses served to classify PET scans as either positive or negative for tumor recurrence, introducing potential subjectivity in PET/CT interpretations. An alternative approach would have been to apply a semi-quantitative threshold such as SUV as a discriminator of malignant involvement. Although a tumor-to-background ratio >2 correctly classified all abnormal lesions in this study, diagnostic thresholds based on SUV have not yet been validated for FCH PET/CT, and such criteria could vary depending on the imaging protocol being used [26
]. Fortunately, many anatomic regions predisposed to metastatic involvement by prostate cancer, such as the bone marrow and pelvic lymphatics, exhibit relatively low FCH uptake under normal conditions, making these areas amenable to visual inspection. Furthermore, since all patients underwent primary treatments for prostate cancer, focal increases in activity within the prostatic region could be considered abnormal for the purpose of this study. A decade of mutual experience in FCH PET and PET/CT likely underlies the strong concordance between the two image interpreters in this study. However, as with any PET tracer, all readers must be aware of the normal tracer biological distribution and normal variants in uptake for image interpretations to be reproducible. Although no false-positive results were encountered in this study, there are several issues known to limit the diagnostic specificity of FCH PET/CT [27
]. First, is non-specific increased uptake by benign lymph nodes, particularly those in the axillary, mediastinal/hilar, and inguinal regions [27
]. Another issue which may lead to interpretive pitfalls is the excretion of intact radiopharmaceutical by the renal system. As illustrated in Fig. , careful study of retroperitoneal and pelvic anatomy with anatomical CT correspondence is essential for accurately assessing pathologic findings in prostate cancer.
Treatments for recurrent prostate cancer could benefit from further individualization based on the mode of tumor recurrence depicted by imaging. For example, localized RT may be guided to lesions that may lead to pathologic fracture or urinary obstruction, thus reducing the risk of metastatic complications. By differentiating the sites of tumor recurrence, imaging may also help in selecting patients who may benefit from novel drugs that target a specific mechanism of metastatic progression. For example, a patient found to have bone-dominant metastatic disease may be steered toward treatments that can specifically interfere with the osteomimetic properties of prostate cancer. Estimations of tumor burden in the later stages of prostate cancer may also prove useful for planning palliative options. Thus, there are multiple opportunities for integrating FCH PET/CT into the clinical management of recurrent prostate cancer.