Combined PET/CT-perfusion in patients with head and neck cancers might predict failure after radio-chemotherapy: a proof of concept study

Overall 41 patients were prospectvely included in a study with PET/CT and integrated CT-perfusion. Of those, thirteen patients (mean age: 63 years; range: 44–78 years; one female, 12 male) received a PET/CT with CT-perfusion before and after therapy. The excluded patients hat either 1. only PET/CT without CT-perfusion or 2. were directly operated after staging or 3. received (salvage) surgery pre-or post radiochemotherapy during their cause of disease. Those patients were prospectively enrolled between 09/2008 and 02/2010. All patients had clinical suspicion for head and neck cancer and all were referred for routine primary staging and post-therapy follow up staging with FDG-PET/CT. The CTP-scan was integrated into the standard contrast-enhanced PET/CT-procedure. Patients were included consecutively, no further selection was applied. All patients had initially a biopsy proven squamous cell carcinoma: Nine carcinomas of the oropharynx (including three cancers of base of tongue and three of the tonsils), one oral cavity (floor of the mouth), one larynx carcinoma and two hypopharynx carcinomas. Eight of these patients had suspected lymph node metastases at the time of study enrollment. All patients received radio- and/or chemotherapy. Two patients additionally received salvage surgery (after radio- chemotherapy), nine patients received combined radio/chemotherapy (70Gy), one patient only chemotherapy and one patient only radiotherapy (70Gy). Different scientific aspects of the patients included in this study are already published in a different study [11]. Outcome of patients was assessed regularly based on local guidelines in our hospital. Patients with advanced head and neck cancer are seen clinically and with imaging after completion of therapy at 3, 6, 12 and 18 month or at occurrence of symptoms. If no recurrence is detected, patients are seen clinically every 2 years.

The study was performed in accordance with the regulations of the local institutional review board and ethics committee (Zurich Cantonal Ethics Committee). Written informed consent was obtained from all patients before the examination and enrolment into the study.

All data were acquired on a combined PET/CT in-line system (Discovery VCT, GE Healthcare, Milwaukee, WI, USA). These dedicated systems integrate a full-ring PET with a multislice helical 64-slice CT and permit the acquisition of co-registered CT and PET images in the same procedure. The patients fasted for at least 4 h prior to the examination. PET/CT imaging with integrated CTP was started 60 min after the injection of a standard dose of 300–340 MBq FDG. All patients received a non-enhanced CT examination which was acquired with the following parameters: 80 mA, 140 kV, 0.5-s tube rotation, 4.25-mm section thickness. The CT examinations were acquired during shallow breathing in the head and neck area, the thorax and the lower abdomen and during non-forced expiration in the upper abdomen. The CT examinations included the area from the head to the upper thighs. The PET emission examination was acquired with an acquisition time of 2 min per bed position after the low-dose CT examination. Images (15 cm axial field-of-view (FOV)/bed position) were reconstructed by using a standard fully 3D iterative algorithm (ordered subset expectation maximisation, OSEM: subsets, 28; iterations, 2; recon matrix, 128 × 128; overlap, nine slices). The selection of target lesions in the primary study was based on focal FDG uptake and the same region was selected in follow-up studies to CTP scan [11].

After target lesion definition, intravenous contrast injection was started by injecting a total dose of 70 mL contrast media (Ultravist 370, Bayer Schering Pharma, Germany). First, 40 mL of contrast media was applied at a flow of 5 mL/s via a cubital vein, and after a 5-s delay, the perfusion data were acquired for 50 s in the area of interest (1 s rotations time with one image/s, cine duration 50 s, eight slices, 5 mm slice thickness, 80 mA, 80 kV). Thus, an anatomical craniocaudal coverage of 4 cm was achieved for the perfusion examination of the target lesion. The first contrast media bolus injection was followed by a saline flush (20 mL, 5 mL/s). Directly after the perfusion imaging acquistion, a second bolus of another 30 mL of contrast media was applied at 4 ml/s followed by a saline flush of 30 mL to ensure full diagnostic (contrast-enhanced) ceCT data of the neck and thorax because all CTP examinations were fully integrated into the routine, clinical cePET/CT staging and re-staging examination [11].

The pre- and post-therapeutic values for CTP as well as pre- and post-therapeutic PET parameters were tested for differences using Spearman’s rho test and Pearson’s correlation. A p-value of p <0.05 was considered statistically significant. The correlation between all CT-perfusion data, the PET-data were and the Δ-values respectively were calculated using Pearson’s correlation. Secondary Spearman’s rho test was done for all data mentioned above. A p-value of p <0.05 was considered statistically significant. In addition the Bootstrap confidence intervals for both correlations were calculated to test for additive confidence intervals. All statistical analyses were performed using SPSS statistical software (Version 16.0.1, SPSS Inc., Chicago, Ill). To assess the prediction of local tumor recurrence by means of BVpre and SUVmaxpre, logistic regression models with stepwise forward entry were built for BVpre and SUVmax separately and using a combination of both with bootstrapping (1000 replications). The goodness-of-fit and the accuracy of the model were determined using the Hosmer-Lemeshow test and c-statistics (area under curve of the model).