Baseline 18F-FDG PET/CT parameters as imaging biomarkers of overall survival in castrate-resistant metastatic prostate cancer

Abstract

The aim of this prospective investigation was to assess the association of parameters derived from baseline (18)F-FDG PET/CT with overall survival (OS) in men with castrate-resistant metastatic prostate cancer.

Methods: Eighty-seven men with castrate-resistant metastatic prostate cancer underwent (18)F-FDG PET/CT and were followed prospectively for OS. Median follow-up in patients who were alive was 22.2 mo (range, 1.6-62.5 mo). OS was defined as the time between the PET/CT imaging or the start of chemotherapy, whichever was later, and death, with patients who were alive censored at the last follow-up date. PET parameters included maximum standardized uptake value (SUV(max)) of the most active lesion, sum of SUV(max), and average SUV(max) of all metabolically active lesions, after subtraction of patient-specific background-liver average SUV. Comparison of OS was based on univariate and multivariable Cox regression analyses of continuous PET parameters adjusted for standard clinical parameters (age, serum prostate-specific antigen level, alkaline phosphatase, use of pain medication, prior chemotherapy, and Gleason score at initial diagnosis). Survival curves based on Kaplan-Meier estimates are presented.

Results: Among the 87 patients, 61 were dead at the time of last follow-up. Median OS was 16.5 mo (95% confidence interval [CI], 12.1-23.4 mo), and the OS probability at 24 mo was 39% ± 6%. For the univariate analysis, the hazard ratios associated with each unit increase were 1.01 (95% CI, 1.006-1.02) for sum of SUV(max) (P = 0.002), 1.11 (95% CI, 1.03-1.18) for maximum SUV(max) (P = 0.010), and 1.13 (95% CI, 0.99-1.30) for average SUV(max) (P = 0.095). For the multivariable analysis adjusting for relevant clinical parameters, the continuous parameter sum of SUV(max) remained significant (P = 0.053), with a hazard ratio of 1.01 (95% CI, 1.001-1.02). When sum of SUV(max) was grouped into quartile ranges, there was poorer survival probability for the patients in the fourth-quartile range than for those in the first-quartile range, with a univariate hazard ratio of 3.8 (95% CI, 1.8-7.9).

Conclusion: Sum of SUV(max) derived from (18)F-FDG PET/CT contributes independent prognostic information on OS in men with castrate-resistant metastatic prostate cancer, and this information may be useful in assessing the comparative effectiveness of various conventional and emerging treatment strategies.

Trial registration: ClinicalTrials.gov NCT00282906.

Keywords: 18F-FDG; cancer; castrate-resistant; prostate; survival.

Figure 1

OS probability for all patients.

Figure 2

Kaplan–Meier plot of OS probability against sum of SUVmax grouped into quartiles. Patients in fourth-quartile group (blue line) have significantly poorer survival probability than reference first-quartile group (green line). First-quartile range = 0–4.6, second-quartile range = 4.7–13.9; third-quartile range = 14–28.6; fourth-quartile range = 28.7–217.5.

Figure 3

OS of low- and high-risk patients further grouped by sum of SUVmax. Low risk and high risk were based on linear prediction scores from Cox model that included all considered standard clinical variables with prediction scores dichotomized at median values (below median as low risk, above median as high risk). Within each risk group, patients were further dichotomized on basis of median sum of SUVmax. For low-risk group, patients with higher sum of SUVmax had significantly worse outcome than patients with lower sum of SUVmax (P = 0.006). By contrast, for high-risk group, association between sum of SUVmax and OS was not significant (P = 0.21).

Figure 4

Moving hazards of death (blue line) in relation to sum of SUVmax interpreted as chance of death per person per month. Cubic spline smoothed line (in red) is superimposed. Marked upward shift of curve is seen for sum of SUVmax greater than 20.

Figure 5

A 61-y-old man with castrate-resistant metastatic prostate cancer that developed 8.2 y after primary treatment with radical prostatectomy (Gleason 5 1 4) followed by hormonal therapy. ¹⁸F-FDG PET/CT (left: abdominal CT only; middle: abdominal PET/CT, right: coronal PET) demonstrated only enlarged metaboli-cally active retroperitoneal lymph nodes (arrowhead). Patient was started on docetaxel chemotherapy. AVG 5 average SUVmax; MAX 5 maximum SUVmax; SUM 5 sum of SUVmax.

Figure 6

A 67-y-old man with dedifferentiated castrate-resistant metastatic prostate cancer who was initially treated with radical prostatectomy followed by hormonal therapy 11.8 y before imaging study. ¹⁸F-FDG PET/CT (left: chest PET/CT; middle: brain PET/CT; right: maximum-intensity-projection PET) demonstrated widespread metastatic disease involving bone (lesions in rib and thoracic vertebra [short arrows]), lungs [arrowheads], and brain [long arrow]). AVG = average SUVmax; MAX = maximum SUVmax; SUM = sum of SUVmax.