Prognostic Value of Pretherapeutic Tumor-to-Blood Standardized Uptake Ratio in Patients with Esophageal Carcinoma

Patient Characteristics

In the present study, 130 consecutive patients with ¹⁸F-FDG PET/CT–staged esophageal carcinoma were included retrospectively. Evaluation of the data was approved by the Institutional Ethics Committee, and all subjects signed a written informed consent form. All cases were discussed in an interdisciplinary tumor board, and patients—because of either locally advanced disease or comorbidities/functional inoperability—received curative definitive radiochemotherapy between September 2005 and October 2013. Inclusion criteria were age more than 18 y, histologically confirmed esophageal carcinoma, ¹⁸F-FDG PET/CT staging, no distant metastases, curative treatment intention, and a minimum follow-up of 12 mo. All patients had a clinical staging of Union for International Cancer Control (UICC) stage based on PET imaging and endoscopy/endosonography. The median age of all included patients was 65 y (range, 42–85 y), and most were male (87%). A summary of patient and tumor characteristics is given in Table 1.

Treatment

All patients were treated with 3-dimensional CT-planned conformal radiotherapy. Total radiation doses of 60, 66, or 70 Gy (cervical tumor localization) were applied in 2 Gy per fraction. The gross tumor volume was separately delineated in the PET and CT images. Up to 50 Gy an elective nodal irradiation of the mediastinum was additionally performed. For boost irradiation, margins of 1.5 cm in axial and 3.5 cm in craniocaudal directions were used to create the clinical target volume. The chemotherapy schedule consisted of cisplatin (70 mg/m²) and 5-fluorouracil (3,000 mg/m² as an infusion over 96 h) in week 1 and 4 of the treatment.

¹⁸F-FDG PET/CT Protocol

All patients underwent a hybrid ¹⁸F-FDG PET/CT scan before treatment. Scans (3-dimensional PET acquisition, 3-min emission per bed position) were obtained with a Biograph 16 (Siemens Medical Solutions Inc.). Data acquisition started 77 ± 16 min after injection of 212–417 MBq of ¹⁸F-FDG. All patients had fasted for at least 6 h before ¹⁸F-FDG injection. The serum glucose concentration measured before injection was 5.8 μmol/mL on average (range, 3.3–9.8 μmol/mL). Tomographic images were reconstructed using attenuation-weighted ordered-subset expectation maximization reconstruction (4 iterations, 8 subsets, gaussian filter of 5 mm in full width at half maximum). The resulting image data had a voxel size of 4.1 × 4.1 × 5 mm.

Data Analysis

Region-of-interest (ROI) definition and ROI analyses were performed using ROVER software (version 2.1.20; ABX). Here and in the following sections, ROI is used synonymously for VOI, for denoting a 3-dimensional volume of interest.

The metabolically active part of the primary tumor was delineated in the PET data by an automatic algorithm based on adaptive thresholding considering the local background (15). The resulting delineation was inspected visually by 2 experienced observers in consensus. Lesions were delineated manually where this was deemed necessary (observers were masked to patient outcome). Manual delineation happened in 5 of 130 cases exhibiting only low diffuse tracer accumulation in the respective lesion. In 6 further cases, the primary tumor was not visible in the PET data, and a small ROI (<1 mL) was manually placed in the esophagus.

For the delineated ROIs, SUV_max, SUV_mean, MTV, and TLG (MTV × SUV_mean) were computed. The arterial blood SUV needed for computation of SUR values was determined by defining a roughly cylindric aorta ROI in the attenuation CT data, which then was transferred to the PET data (Fig. 1). To exclude partial-volume effects in this evaluation, a concentric safety margin was used in the transaxial planes, centering the ROI in the aorta. Planes showing high tracer uptake close to the aorta (pathologic or otherwise) were excluded. The minimum volume of the resulting aorta ROI was 5 mL. Blood SUV was computed as the SUV_mean in the aorta ROI. Figure 1 shows an example of a delineated aorta ROI. SUR of the lesions was computed as the ratio of lesion SUV and blood SUV.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Illustration of delineation of aorta. Actual delineation was performed on attenuation CT (B, marked in red). Resulting ROI was then transferred to PET image (A). High tracer uptake of lesion close to aorta will cause spillover into neighboring aorta, which may adversely affect blood SUV determination. Therefore, this part of aorta is excluded from ROI delineation.

Additionally, lesion SUV and SUR values were scan-time-corrected as described van den Hoff et al. (14):Eq. 1where T is the actual scan time after injection, and T₀ is the chosen standard scan time to which the SUV and SUR values are normalized (60 min in the present work). V_r is a fixed estimate of the apparent volume of distribution, corresponding to the y-axis intercept of a Patlak plot in dynamic investigations (we used V_r = 0.53 mL/mL), and b is a parameter describing the shape and decrease of the AIF over time (we used b = 0.313 (14)). Corresponding to the 4 possibilities to quantitatively assess the tracer uptake (conventional SUV; scan-time-corrected SUV, SUV^tc; tumor-to-blood uptake ratio SUR; scan-time-corrected tumor-to-blood SUR, SUR^tc), 4 different measures of maximum tracer uptake, mean tracer uptake, and TLG, respectively, were computed and further analyzed ([Table 2]; subscripts mean and max refer to the mean and maximum value of the respective parameter in the lesion ROI; superscript tc indicates that scan-time correction of SUV and SUR was performed during computation of the respective parameter; and for TLG, the subscript SUR indicates that SUR instead of SUV was used for computation of the TLG product).

Clinical Endpoints and Statistical Analysis

The 3 clinical endpoints of this study were OS, locoregional tumor control (LRC), and distant metastases-free survival (DM) measured from the start of radiotherapy to death or event. Patients who did not keep follow-up appointments and for whom information on survival or tumor status was thus unavailable were censored with the date of last follow-up.

The association of OS, LRC, and DM with clinically relevant parameters (sex, age, smoking state, histology, grading, T stage, and N stage) as well as quantitative PET parameters was analyzed using univariate Cox proportional hazards regression in which the PET parameters were included as binarized parameters. The cutoffs used for binarization were calculated by performing an univariate Cox regression for each measured value. The value leading to the hazard ratio (HR) with the highest significance was used as cutoff. To avoid too-small group sizes, only values within the interquartile range were considered as potential cutoff. The cutoff values were separately computed for OS, LRC, and DM. The probability of survival was computed and rendered as Kaplan–Meier curves.

Independence of PET parameters from clinically relevant parameters was analyzed in multivariate Cox regression. Those clinical parameters with at least a trend for significance according to univariate Cox regression (P < 0.1) were included.

Correlation was tested by the Spearman rank correlation method. Statistical significance was assumed at a P value of less than 0.05. Statistical analysis was performed with the R Language and Environment for Statistical Computing (version 3.1.2; R Foundation for Statistical Computing) (16).