Two-step-fusion 18F-fluorodeoxyglucose Positron Emission Tomography/Computed Tomography (FDG-PET/CT) based radiotherapy in locally advanced oropharyngeal cancer

Aims. To develop two-step-fusion 18F-fluorodeoxyglucose Positron Emission Tomography/Computed Tomography (FDG-PET/CT) based radiotherapy in locally advanced oropharyngeal cancer at the Beatson West of Scotland Cancer Centre and evaluate the technical and clinical aspect of this multi-modality imaging methodology. Methods. I conducted a radiotherapy service development project at the Beatson. Contrast enhanced radiotherapy simulation CT (CTsim) and FDG-PET/CT were acquired separately with the same set-up and fused using an automatic rigid fusion algorithm (Eclipse, Varian). The fusion accuracy was assessed with the spatial reproducibility index (R=intersection/union ratio) of bony structures. Radiotherapy target volumes for both primary (T) and nodal disease (N) were defined separately on CTsim and FDG-PET/CT using visual assessment (PET/CT-vis) and segmentation with 50% SUVmax (PET/CT-50%). Volumes (cc) and spatial reproducibility (R) were calculated for the various volumes. Changes in TNM staging definition due to FDG-PET/CT were evaluated and compared with the staging based on morphological imaging (CT±MRI) and clinical information (endoscopy). SUVmax was calculated for T and N and correlated with the HPV-status and the oropharyngeal prognostic groups (low risk: HPV+, ≤10 pack years smoking history; intermediate risk: HPV+, >10 pack years smoking history; high risk: HPV-). Patients were treated using the target volumes defined with PET/CT-vis. Volumetric Modulated Arc Therapy (VMAT) was used with 65Gy and 54Gy in 30 fractions to high and low risk volumes respectively. Tumour outcome and late toxicity were recorded and compared with an internal non-PET/CT-based oropharyngeal series. Data were analysed using Stata v14.2 (StataCorp LLC, Texas). Data were summarised using medians (with range or inter-quartile range IQR). P-values were calculated to test for differences. All tests were 2-sided and a p-value <0.05 was considered statistically significant. Results. A total of 30 patients were enrolled. The fusion accuracy of FDG-PET/CT and CTsim was calculated in 14 patients and resulted 0.89 (0.83-0.92). SUVmax was recorded for both primary and nodal disease in 27 patients. Among these patients median SUVmax was significantly higher in the primary tumour compared to the nodal disease (19.0 versus 14.0 g/ml, p=0.0001). Median SUVmax was higher in HPV- compared to HPV+ patients for both primary tumour (21.0 vs 16.9 g/ml) and nodal disease (17.0 vs 10.0 g/ml), however these differences were not statistically significant. Nodal SUVmax was higher in the high risk (i.e. HPV-) compared to the intermediate and low risk (i.e. HPV+) group (17.0 vs 8.8 vs 15.o g/ml) although again these differences were not statistically significant. FDG-PET/CT down-staged and up-staged T and N in 6/30 (20%) and 17/30 (57%) patients. Unsuspected distant metastases were not detected in any of the patients at baseline. The median volume of T and N defined with PET/CTvis and CTsim was 11.5cc vs 16.5cc (p=0.31) and 13.8cc vs 11.1cc (p=0.42), with reproducibility index R=0.49 and R=0.47 respectively. PET/CT50% identified hyper-metabolic sub-volumes inside PET/CTvis for both T and N: 4.6cc vs 11.5cc (p=0.001) and 3.5cc vs 13.8cc (p=0.04), DICE index 1. At median follow-up time of 16 (1-44) months, 74% of the patients had complete response, whilst 22% had progressive disease with median time to progression of 6.1 (3.1-15.9) months. The estimated overall survival (OS) at 2 years was 74% (95%CI, 49%-88%). In the sub-group analysis, the estimated OS at 2 years was 83% (95%CI 27-97%), 87% (39-98%) and 67% (19-90%) in the low, intermediate and high risk category respectively. Grade≥2 late xerostomia, dysphagia, dysgeusia and fatigue were recorded in 36%, 35%, 0% and 14% of the patients. Grade≥2 dysphagia was recorded in 38% of the patients who presented with bilateral and unilateral neck nodes (p=1.0). Conclusions. I developed a 2-step-fusion methodology between FDG-PET/CT and CTsim. PET/CT fusion has been introduced in the routine radiotherapy planning at the Beatson for selected oropharyngeal cancer patients. My data suggest that HPV- are more metabolically active than HPV+ oropharyngeal cancers. My results support the hypothesis of treatment intensification in the high-risk group because more biologically aggressive. Dose intensification to hypermetabolic tumour sub-volumes may improve the outcome especially in the high-risk sub-group. FDG-PET/CT modified tumour staging and radiotherapy target volumes. My outcome and late toxicity results are similar to an internal non-PET-based series and other published studies. A prospective randomised study stratified by risk group would clarify if a true difference exists in outcome and late toxicity between PET-based and non-PET-based radiotherapy.