PET/MRI in Oropharyngeal Squamous Cell Carcinoma Evaluation

Introduction and aim of study: Head and neck cancer (HNC) continues to be a significant health care problem in Taiwan and oropharyngeal squamous cell carcinoma (SCC) is the common subtype. With the concern of organ preservation in recent years, concurrent chemoradiation is the major treatment modality for oropharyngeal SCC. Hence, accurate tumor anatomic and biologic evaluations are warranted for pretreatment planning, posttreatment monitoring and prognosis determination. Endoscopy with biopsy serve as the main diagnostic tools in patients with oropharyngeal SCC. While computed tomography (CT) and magnetic resonance imaging (MRI) are commonly used to evaluate the tumor extent of HNC, MRI is more preferred in the oropharyngeal area by virtue of its high contrast resolution, particularly in detecting perineural invasion. Whole body MRI is now feasible and thus distant site status of HNC can be evaluated in the single examination session. FDG- positron emission tomography/CT (PET /CT) is another common imaging modality to evaluate HNC. According to the investigators' experience in patients with head and neck cancer, MRI and PET/CT have different advantages over each other and may complement to each other. In evaluating primary tumor status, the high spatial and contrast resolution of MRI can delineate the tumor extent from surrounding normal tissue in the complex head and neck anatomical region, while delineation of FDG-positive tissue by PET may help differentiation tumor growth from surrounding noncancerous tissue, particularly in irradiated bed. In evaluating regional nodal status, MRI is superior to PET/CT in detecting retropharyngeal or cystic necrotic metastatic adenopathy, whereas PET/CT excels MRI in detecting metastasis in sub-centimeter nodes. The combined use of MRI and PET can clearly demonstrate the patterns of nodal spread.\[7\]. Regards to the evaluation of the distant sites, whole body MRI is better in detecting metastatic lesion in high metabolic organs such as brain, liver or spleen, while PET/CT is superior in detecting those metastatic lesions in the curved, flat bones, and second primary bowel tumor such as esophagus cancer or colon cancer. Currently, diffusion-weighted MR imaging (DWI) and dynamic contrast-enhanced perfusion weighted imaging (DCE-PWI) become clinically feasible to assess the functional aspects of HNC. The quantified parameter of DWI sequence is apparent diffusion coefficient (ADC) that relates to cellularity. DCE-PWI provides the volume transfer rate constant (Ktrans), relative extravascular extracellular space (Ve) and relative vascular plasma volume (Vp) as well as the efflux rate constant (Kep). These functional MRI techniques may provide biologic information, such as cellularity, angiogenesis and permeability. On the other hand, PET can provide its metabolic quantified parameters: standardized uptake value (SUV) that reflects glucose metabolism, metabolic tumor volume (MTV) that reflects tumor burden, and total lesion glycolysis (TLG) that integrates both glucose metabolism and tumor burden. These MRI and PET functional parameters appear promising for predicting chemoradiation response and prognosis of HNC. In our previous studies of oropharyngeal or hypopharyngeal SCC patients, Ktrans of the primary tumor was the only imaging parameter associated with local control, while ADC and Ve values of the neck metastatic nodes were independent prognostic factors for neck control. The maximal SUV of the regional lymph nodes was significantly associated with the occurrence of distant metastasis. TLG were significantly associated with overall survival. Integrated PET/MRI is a novel imaging technology introduced into the clinical practice. The combination of PET and MRI in one single, hybrid scanner is promising. Initial data convey that PET/MRI performed favorably for HNC. It showed good diagnostic capability similar to PET/CT and can serve as a legitimate alternative to PET/CT in the clinical workup of patients with HNC. The metabolic ratios measured PET/MRI showed excellent agreement with those on standard PET/CT. PET/MRI incorporating the Dixon sequence for attenuation correction purposes yielded similar SUV values when compared to PET/CT.Nevertheless, most studies were limited in number of cases or no particular tumor entity was specifically analyzed, which somewhat weakened their power in terms of evidence-based medicine. In addition, the predictive value of PET/MRI in treatment outcome of HNC has not been reported. A prospective study of PET/MRI in a large cohort of patients with specific tumor origin and uniform treatment protocol would complement these pioneering efforts to fully validate the clinical usefulness of the novel integrated system. Recently, integrated PET/MRI scanner (Biograph mMR, Siemens) has been installed in our hospital. In this scanner, the PET detectors are fully integrated into the 3 Tesla MR system with one single gantry, enabling simultaneous acquisition of both PET and MR data. Before its wide practical use, its clinical usefulness should be explicitly defined. In this research project, the investigators will investigate our 150 patients with oropharyngeal SCC subjected to chemoradiation with FDG-PET/MRI. Non-contrast chest CT will also be performed on the same day. The investigators have the following aims: Aim 1: to determine whole-body staging/restaging MRI and PET/CT have different advantages over each other, and visual correlation of these separate imaging modalities can yield a slightly higher diagnostic capability. However, some lesions may not be matched well for each other because of different examination date and positioning. Therefore, the investigators believe that tumor involvement should be made more accurately with simultaneously acquired PET/MRI imaging, leading to more accurate tumor staging/restaging, and in turn, to more precise treatment planning and better treatment outcome. However, the data of integrated PET/MRI for whole whole-body staging/restaging on oropharyngeal SCC are still unavailable currently. In this prospective project, the investigators aim to determine the feasibility and clinical impact of integrated PET/MRI in tumor staging and restaging of oropharyngeal SCC Aim 2. to predict treatment response and prognosis Functional MRI parameters and FDG-PET/CT parameters can be quantified and have been used for predicting response to chemoradiation in HNC. However, variable results had been obtained, mainly due to different tumor origins, sample sizes, methodologies, and treatment protocols. Another important confounding factor is that MRI and PET/CT were done at different time intervals. Because integrated PET/MRI can provide simultaneous PET and MR functional data that allow for direct comparison in the selected regions of interest, and thus should be more accurate and more reproducible. However, the predictive values of simultaneous functional PET/MRI techniques in oropharyngeal SCC have not been reported till now. On the other hand, DWI with intravoxel incoherent motion (IVIM ) technique can quantify both molecular diffusion and microcirculation in the capillary network that may be useful for predicting tumor chemoradiosensitivity, while DWI with kurtosis (biexponential or non-Gaussian fitting) modeling has been documented to yield a better fit of the in vivo water molecule diffusion than does that with monoexponential modeling\[32\]. In this research project, the investigators added functional MRI sequences as parts of PET/MRI procedures, including DCE-PWI as well as dedicated DWI that can yield monoexponential DWI, IVIM and Kurtosis data. We aim to obtain MRI and PET functional results on integrated PET/MRI system in the single imaging session in the patient cohort of oropharyngeal SCC uniformly treated with chemoradiation. The investigators expected that the predictive values of the various functional parameters of integrated PET/ MRI can be more clearly elucidated. Aim 3. to determine necessity of noncontrast chest CT. One of the concerns of replacing the CT portion of PET/CT with MRI is the ability of MRI to depict and differentiate lung nodules, while the most common location for distant metastases of HNC is lung. According to the investigators' previous study, the CT component of integrated PET/CT can improve not only the specificity but also the sensitivity of FDG-PET data. In addition, CT also help to lessen PET false-positive findings in calcified mediastinal nodal disease, such as anthracosis. However, on the other hand, the investigators' previous studies showed the detection rates of pulmonary metastases with 3.0-T MRI using half-Fourier acquisition single-shot turbo spin-echo (HASTE), volumetric interpolated breath-hold examination (VIBE) and short τ inversion recovery (STIR) sequences were similar to those of PET/CT. One recent study has shown that radial VIBE free-breathing MRI with simultaneously acquired PET data has high sensitivity in the detection of nodules with a diameter of at least 0.5 cm, but has limited sensitivity in the detection of small or non-FDG-avid nodules. Considering the limited data on PET/MR in detecting lung malignant lesion, the investigators aim to prospectively validate the performance of our integrated PET/MRI protocol sequences for lung lesions by adding noncontrast enhanced CT for comparison. Material and methods: In this three-year prospective project, a total of 160 patients with histologically proven oropharyngeal SCC subjected to chemoradiation will be enrolled. Exclusion criteria include previous head or neck malignant tumor, a second malignant tumor, distant metastasis, contraindications to MRI (renal insufficiency, cochlear implant, cardiac pacemaker placement or intracranial aneurysmal ferromagnetic clips), and serum glucose level \>200 mg/dl. Before pretreatment, each enrolled patients will undergo PET/MRI and detail clinical examination, including human papillomavirus test. The participants will also undergo low dose chest CT on the same day before PET/MRI. In the posttreatment period, baseline whole body MRI will be obtained 3 months after chemoradiation. Thereafter, the patients will then be followed up also with alternative extend-field CT and whole body MRI by every 6 months. If tumor recurrence is confirmed or highly suspected, PET/MRI will also be performed for tumor re-staging. The three-years planning Adequate cases number and follow-up duration are essential for statistical analysis and outcome determination of oropharyngeal SCC treated with chemoradiation. The investigators plan to accomplish this study in 3 years. In the first year, the works of the this project in the first year will: (1) design the data bank categories; (2) set up the workflow and optimize the imaging protocols; (3) determine the diagnostic capability of MRI component, PET component, integrated PET/MRI in tumor staging; (4) determinate any added diagnostic value of noncontrast CT in integrated PET/MRI, and (5) study the early treatment response and the patterns of residual tumor. In the second year, the investigators will continue to the first-year works, and further include the following works: (1) to determinate incidence and predictors of treatment failure of oropharyngeal SCC, and (2) to study the patterns of treatment response and early recurrence. In the third year, the investigators will continue to do previous work and will also (1) get sufficient sample size to perform statistical analysis of the relationship between the imaging parameters and patient outcomes, (2) attain comprehensive imaging about patterns of tumor recurrence and posttreatment changes/complications, (3) investigate to what extent biologic imaging parameters may affect outcome and patient selection for chemoradiation, (4) analysis the accuracy, pitfalls and cost effectiveness of the PET/MRI alone and PET/MRI with noncontrast CT in evaluating in patients with oropharyngeal SCC. PET/MRI protocol PET/MRI data will be acquired on the integrated PET/MRI scanner (Biograph mMR, Siemens Healthcare, Erlangen, Germany) , which acquires simultaneous PET and MR data with a 3.0-T magnet. The examination protocol will combined a whole-body scan with a dedicated examination of the head and neck area (Table 1). All patients will fast for 6 h before the scan. At 50-70 min post injection of 370 MBq of FDG, the patient will be placed on the PET/MRI scanner bed. After fast-view T1-weighted MR localizer sequence for scout imaging and Dixon VIBE sequence for attenuation correction, a whole-body PET scan will be performed in 5 bed positions to cover from the head to the proximal thigh, with an acquisition time of 4 min per bed position. Simultaneously, whole body MR image acquisition will be performed for the corresponding 5 bed positions with the axial HASTE sequence and coronal STIR sequence as well as the sagittal T1-weighted Turbo spine echo(TSE) and STIR sequence. Afterwards, regional PET and MRI images will be simultaneously performed. Regional PET will be performed with an acquisition time of 10 min, while a dedicated MRI of the head and neck region will be acquired in the axial and coronal projections with T1-weighted TSE sequence and T2-weighted TSE sequence with fat saturation. Axial DWI will be performed using a single shot spin-echo echo-planar technique with modified Stejskal-Tanner diffusion gradient pulsing scheme. A total of 10 b values will be used for the reconstruction of IVIM and kurtosis imaging, which are: 0, 20, 40, 80, 100, 200, 400, 800, 1200, 1500 sec/mm2. DCE-PWI at the head and neck region will be acquired by using a 3D T1-weighted spoiled gradient-echo sequence. A spatial saturation slab will be implanted inferior to the acquired region to minimize the inflow effect from the carotid arteries. Before the contrast agent administration, baseline longitudinal relaxation time (T10) values will be calculated from image acquired with different flip angles (4°, 8°, 15° and 25°). Then, the dynamic series will be acquired using the same sequence with a 15° flip angle, after intravenous administration of paramagnetic contrast agent at 3 ml/s. Thereafter, dedicated regional MRI will be obtained with T1-weighted TSE sequence with fat saturation in the axial and coronal projections. Finally, whole body axial VIBE with fat saturation will be performed. The total acquisition time is about 42 min, and the mean in-room time for PET/MR will be approximately 60 min. Non-enhanced low-dose CT Spiral low-dose CT without contrast material enhancement will be performed before PET/MRI on the same examination day. Acquisition parameters include peak voltage of 120 kVp, mAs of 50, collimation of 64x0.5 mm and reconstruction interval of 3-mm. Data analysis and outcome determination Readers are aware that patients have oropharyngeal SCC, and they will be blinded to the results of other studies and to PET/MR data. The PET, MRI and lung CT images will be first interpreted independently. All images will be then reviewed together and compared. A checklist of various distributions of tumor extension, nodal spread and distant metastasis will be recorded. The clinical and imaging findings will be discussed jointly by the Head-and-Neck research team. Endoscopic biopsy, ultrasonographic guided fine needle aspiration or CT-guided biopsy will be performed in any lesions suspicion for malignancy if possible. If biopsy of the lesion of interest is not feasible, or yields a negative result, close clinical and imaging follow-up will be pursued. All patients will be followed up for at least 12 months. The clinical and functional imaging data will be collected and analyzed for predicting treatment response and prognosis. For the diffusion MRI, regions of interest will be manually placed on the lesions on ADC map to encompass as much of the solid tumor area as possible. The signal intensities measured on the images acquired at different b-values S(b) will be numerically fitted against the model, S(b)=S0 e-b\*ADC, where S0 and Sb are signal intensities at different b values, For IVIM imaging, the relationship between signal intensities and b values can be expressed by the equation: Sb/S0 = (1-f )．exp(-bD) + f．exp\[-b( D + D＊)\] where f is a microvascular volume fraction representing the fraction of the diffusion linked to microcirculation, D represents pure diffusion coefficient, and D\* is perfusion-related incoherent microcirculation. For the DKI, the relationship between signal intensities and b values can be expressed by the equation: ln\[S(b)\] = ln\[S(0)\] - b x Dapp + 1/6b2 x Dapp2 x Kapp where S is the signal intensity (arbitrary units), b is the b-value (s/mm2), Dapp is the apparent diffusion coefficient (10-3 mm2/s), and Kapp is the apparent kurtosis coefficient denoting the deviation from a Gaussian distribution. The investigators will also perform monoexponential fitting by using Kapp=0 in the equation, yielding ADCmono. For the DCE-PWI MRI, the change in contrast agent concentration over time, Ct(t), will be determined in each voxel in the tumor, and the compartmental tracer kinetic model will be applied to each voxel by using an arterial input function, Cp(t), measured in each individual: Ct (t)=VpCp(t) + Ktrans ∫0t Cp (t' ) exp(Ktrans(t-t') /Ve )dt' where t' is the time (in minutes) as an integration variable, and Cp(t') is the concentration of contrast agent in the blood plasma as a function of time. For PET imaging parameters, SUV and MTV of the target lesions will be measured from attenuation-corrected 18F-FDG PET images by drawing the boundaries drawn large enough to include the lesions. An SUV threshold of 2.5 will be used to delineate the MTV. The TLG is calculated as the product of the mean SUV and the MTV. Statistical Analysis Using histological findings or follow-up data at 12 months as a standard of reference, various imaging results will be classified as true-positive, true-negative, false-positive, or false-negative. The diagnostic accuracies of MRI component, PET component, integrated PET/MRI, chest CT alone, and PET/MRI plus chest CT will be calculated and compared with the McNemar test. Their respective diagnostic performance will be determined with the areas under the receiver-operating-characteristic curve. For treatment response evaluation and prognosis prediction, logistic regression analyses will be used to identify the relationship between clinical and imaging functional variables. The control and survival rates will be plotted using the Kaplan-Meier method. Differences in positive and negative results will be determined by log-rank test. All prognostic variables identified by univariate analysis will be put into the multivariate model using the Cox proportional hazards model. The Spearman's rank correlation coefficient will be used to investigate the correlations among the variables. P values \< 0.05 is considered statistically significant.