FDG PET/CT as a prognostic biomarker in the era of molecular-targeting therapies: max SUVmax predicts survival of patients with advanced renal cell carcinoma

This was a prospective study that followed enrolled patients slated to undergo systemic therapies for pathologically proven advanced RCC between June 2008 and January 2014. The patients were initially assessed by conventional imaging techniques (computed tomography, magnetic resonance imaging, or bone scintigraphy) and diagnosed as stage IV or recurrent RCC. Patients with uncontrolled diabetes mellitus (fasting blood sugar > 150 mg/dL), other known malignancies, and patients who had received treatment within 2 weeks prior to enrollment were excluded. The study protocol was approved by the Yokohama City University Institutional Review Board. Written informed consent was obtained from all patients.

Initially, 110 patients were enrolled in the study. Nine were eventually eliminated: four whose pathology could not be confirmed conclusively, three who decided against treatment after evaluation by FDG PET/CT, one patient had a fasting blood sugar over 150 mg/dL, and one with contralateral kidney metastases for which accurate SUV could not be measured owing to the urinary excretion of the radiotracer. This left a total of 101 patients for the analysis, including 24 who had been analyzed in the preliminary report [12]. The first therapeutic interventions after enrollment in this study were decided before the evaluation by FDG PET/CT.

Patients fasted for at least 6 h prior to intravenous injection of ¹⁸F FDG. PET/CT images were acquired (Aquiduo 16^®; Toshiba Medical Systems, Tokyo, Japan). One hour after injection of 2.5 MBq/kg of ¹⁸F FDG, PET/CT images were acquired from the top of the head to the mid-thigh. A low-dose, non-contrast CT scan was acquired first and used for attenuation correction. Emission images were acquired in three-dimensional mode for 2 min per bed position. After PET acquisition, contrast-enhanced CT was performed with a 2-mm section thickness, 120 kV, 400 mA, 0.5 s/tube rotation, from the top of the head to the mid-thigh, with breath holding. A total of 100 mL contrast medium (iopamidol) was administered intravenously at a rate of 1.0 mL/s. The scan delay was set at 120 s after the start of the injection of contrast material. Patients with serum creatinine levels > 1.5 mg/dL were examined without contrast material. The all cases with origin RCC were evaluated by contrast enhanced CT scan. Images were reconstructed by attenuation-weighted, ordered-subset expectation maximization (four iterations, 14 subsets, 128 × 128 matrix, with 5-mm Gaussian smoothing). The SUV was determined according to the standard formula, with activity in the volume of interest (VOI) recorded as Bq per mL/injected dose in Bq per total body weight (kg). VOIs were positioned to encompass targets within areas of increased uptake and measured on each slice by two experienced physicians (DU and KM), who were blinded to clinical data. Discrepancies were resolved by consensus reading. Analysis of FDG uptake in the primary tumor was made with reference to contrast-enhanced CT images to differentiate tumor from physiologic parenchymal and urinary tract activity. The maximum activity of all VOIs of each patient was defined as the max SUVmax.

Survival time was calculated from the date of evaluation by ¹⁸F-FDG PET/CT to the date of death. A Cox proportional hazards model was used to assess the effects of max SUVmax on survival. OS curves were estimated by the Kaplan-Meier method, and the resulting curves were compared using the log-rank test. The impacts on overall survival (OS) of max SUVmax and other standard clinicopathologic factors (performance status, the interval from diagnosis to start of treatment, LDH, corrected calcium, age, sex, and pathology) were analyzed by a univariate Cox hazard model, and the factors with p < 0.05 were analyzed by a multivariate Cox hazard model. All statistical analyses were carried out with commercial software (SPSS^®, SPSS Inc., Chicago, IL, USA).

The characteristics of the 101 patients are shown in Table 1. Of 40 patients with Stage IV disease, 24 had not undergone prior nephrectomy. The FDG PET/CT evaluation of the 17 patients who had received prior therapy was performed more than 2 weeks after the end of any previous treatment.

After evaluation by PET/CT, 40 patients were treated with sorafenib, 38 with sunitinib, 12 with interferon-α (IFN-α), eight with temsirolimus, one with axitinib, one with pazopanib, and one with chemotherapy. Comprehensive decisions regarding first interventions were made from pathological and clinical information before the FDG PET/CT evaluation.

The median observation period was 18 months (range 1–70). During the observation period, 44 patients were treated with a single intervention (20 sorafenib, 19 sunitinib, four IFN-α, and one temsirolimus), 22 with two interventions (10 TKI to mTORI, 9 TKI to TKI, and 3 mTORI to TKI), 20 with three interventions, and 15 with more than three interventions. Ninety-six patients were treated with TKI, 43 with mTORI, and 16 with IFN-α. Six patients underwent metastasectomy, and five, nephrectomy (Table 2). There were 57 cases of death due to cancer; we confirmed that the other 44 patients were still alive at the time of this writing. There were no cases of death due to other causes.

The max SUVmax of all patients ranged from undetectable to 23.0 (median 6.9). When max SUVmax was analyzed as a continuous variable, high max SUVmax was associated with shorter OS, (Fig. 1) (p < 0.001, hazard ratio 1.257, 95 % confidence interval [CI] 1.177–1.342).

The impact of max SUVmax on OS was compared with that of numerous standard risk factors. The multivariate analysis of max SUVmax with performance status, LDH, corrected calcium, interval between diagnosis and entry, and pathology (p < 0.05 in univariate analysis) revealed that max SUVmax was a significant independent predictor of survival (Table 3).

At first, we validated the application of a cutoff of max SUVmax of 8.8, which was the same cutoff point for OS prediction used in our previous report [13], focusing on the 77 patients who were enrolled after the preliminary analysis. The median OS of the 52 patients with RCC having a max SUVmax < 8.8 was 57.3 months, and that of the 25 patients with RCC having the max SUVmax ≥ 8.8 was 13.2 months (95 % CI 5.89–20.51) (p < 0.0001) (Fig. 2).

We then divided the 101 patients into three subgroups by max SUVmax. Because the existence of the subgroup of patients with RCC showing very high max SUVmax whose survival time were less than 1 year became apparent in Fig. 1. The max SUVmax of 51 patients (50 %) was < 7.0 and the median OS of this subgroup was 41.9 months (95 % CI 34.12–49.68). The max SUVmax of 32 patients (32 %) were ≥ 7.0 and < 12.0, and median OS was 20.6 months (95 % CI 12.4–28.8). The max SUVmax of 18 patients (18 %) was ≥ 12.0, and median OS was 4.2 months (95 % CI 0.7–7.7). Differences in OS for these patient subgroups were statistically significant (< 7.0 vs. ≥ 7.0 and < 12.0: p=0.0001, ≥ 7.0 and < 12.0 vs. ≥ 12.0: p=0.0004) (Fig. 3).

Figure 4 presents the features of the FDG PET/CT scans. Regardless of the tumor size and the organs where the metastasis was located, the patients with lower max SUVmax exhibited better OS than the patients with higher max SUVmax.