Prediction of the treatment outcome using machine learning with FDG-PET image-based multiparametric approach in patients with oral cavity squamous cell carcinoma.

AIM: To investigate the value of machine learning-based multiparametric analysis using 2-[18F]-fluoro-2-deoxy-d-glucose positron-emission tomography (FDG-PET) images to predict treatment outcome in patients with oral cavity squamous cell carcinoma (OCSCC). MATERIALS AND METHODS: Ninety-nine patients with OCSCC who received pretreatment integrated FDG-PET/computed tomography (CT) were included. They were divided into the training (66 patients) and validation (33 patients) cohorts. The diagnosis of local control or local failure was obtained from patient's medical records. Conventional FDG-PET parameters, including the maximum and mean standardised uptake values (SUVmax and SUVmean), metabolic tumour volume (MTV), and total lesion glycolysis (TLG), quantitative tumour morphological parameters, intratumoural histogram, and texture parameters, as well as T-stage and clinical stage, were evaluated by a machine learning analysis. The diagnostic ability of T-stage, clinical stage, and conventional FDG-PET parameters (SUVmax, SUVmean, MTV, and TLG) was also assessed separately. RESULTS: In support-vector machine analysis of the training dataset, the final selected parameters were T-stage, SUVmax, TLG, morphological irregularity, entropy, and run-length non-uniformity. In the validation dataset, the diagnostic performance of the created algorithm was as follows: sensitivity 0.82, specificity 0.7, positive predictive value 0.86, negative predictive value 0.64, and accuracy 0.79. In a univariate analysis using conventional FDG-PET parameters, T-stage and clinical stage, diagnostic accuracy of each variable was revealed as follows: 0.61 in T-stage, 0.61 in clinical stage, 0.64 in SUVmax, 0.61 in SUVmean, 0.64 in MTV, and 0.7 in TLG. CONCLUSION: A machine-learning-based approach to analysing FDG-PET images by multiparametric analysis might help predict local control or failure in patients with OCSCC.

Evaluation of non-Gaussian model-based diffusion-weighted imaging in oral squamous cell carcinoma: comparison with tumour functional information derived from positron-emission tomography.

AIM: To evaluate and compare diffusion-weighted imaging (DWI) parameters derived from a non-Gaussian fitting model and positron-emission tomography (PET) parameters derived from 18F-fluoromisonidazole-PET (FMISO-PET) in patients with oral squamous cell carcinoma (OSCC). MATERIALS AND METHODS: Primary sites were evaluated prospectively in 18 patients. DWI was performed using six b-values (0-2,500). Diffusion-related parameters of kurtosis value (K), the kurtosis-corrected diffusion coefficient (DK), diffusion heterogeneity (α), distributed diffusion coefficient (DDC), the slow diffusion coefficient (Dslow), and the apparent diffusion coefficient (ADC) were calculated from four diffusion-fitting models. Maximal standardised uptake values (SUVmax), mean standardised uptake values (SUVmean), and the tumour-to-muscle ration (TMR) of the SUV value were calculated for FMISO-PET. Spearman's correlation coefficient was used to evaluate the correlation between each non-Gaussian diffusion model parameters and PET parameter. RESULTS: There was moderate correlation between FMISO-PET SUVmax and Dslow (ρ=-0.45, p=0.06). In addition, there was good correlation between TMRmax and five non-Gaussian diffusion model parameters (K: ρ=0.65, p=0.004, DK: ρ=-0.72, p=0.0008, DDC: ρ=-0.75, p=0.0003, ADC: ρ=-0.74, p=0.0005, and Dslow: ρ= -0.65, p=0.003), and between TMRmean and five non-Gaussian model parameters (K: ρ=0.64, p=0.005, DK: ρ=-0.61, p=0.007, DDC: ρ=-0.63, p=0.005, ADC: ρ=-0.61, p=0.007, and Dslow: ρ=-0.56, p=0.015). CONCLUSION: Non-Gaussian diffusion model parameters can be related to tumour hypoxia.

Non-invasive three-dimensional bone-vessel image fusion using black bone MRI based on FIESTA-C.

AIM: To evaluate the image quality of bone-vessel fused volume-rendering (VR) images reconstructed by three-dimensional "black bone" magnetic resonance imaging (MRI) based on the fast imaging employing steady-state acquisition cycled phases (FIESTA-C) sequence and time-of-flight magnetic resonance angiography (TOF-MRA). MATERIALS AND METHODS: Seventeen patients were analysed in this retrospective study. All patients underwent both MRI techniques including FIESTA-C and TOF-MRA and computed tomography angiography (CTA). MRI- and CT-based bone-vessel VR images were reconstructed. Visual depictions of frontal and parietal branches from the superficial temporal artery (STA) were independently scored by three experienced radiological technologists using a four-grade system. RESULTS: In the visual evaluation, the scores of the both right and left frontal branches in MRI-based VR image were significantly larger those at CT (p<0.01, respectively). The scores of both the right and left parietal branches tended to be larger in MRI-based than that in CT-based VR imaging, but were not significantly so (p=0.06, 0.13 respectively). In the interobserver agreement analysis, κ values were all good (range: 0.6-0.76) for STA branch evaluation in MRI-based VR images. CONCLUSION: MRI bone-vessel fused VR imaging can non-invasively depict STA frontal branches with better visibility compared to the CT-based VR imaging. This technique may be useful for the preoperative evaluation of donor branches for STA-middle cerebral artery bypass surgery.

Integrating quantitative morphological and intratumoural textural characteristics in FDG-PET for the prediction of prognosis in pharynx squamous cell carcinoma patients.

AIM: To assess potential prognostic factors in pharynx squamous cell carcinoma (SCC) patients by quantitative morphological and intratumoural characteristics obtained by 2-[18F]-fluoro-2-deoxy-d-glucose positron-emission tomography/computed tomography (FDG-PET/CT). MATERIALS AND METHODS: The cases of 54 patients with pharynx SCC who underwent chemoradiation therapy were analysed retrospectively. Using their FDG-PET data, the quantitative morphological and intratumoural characteristics of 14 parameters were calculated. The progression-free survival (PFS) and overall survival (OS) information was obtained from patient medical records. Univariate and multivariate analyses were performed to assess the 14 quantitative parameters as well as the T-stage, N-stage, and tumour location data for their relation to PFS and OS. When an independent predictor was suggested in the multivariate analysis, the parameter was further assessed using the Kaplan-Meier method. RESULTS: In the assessment of PFS, the univariate and multivariate analyses indicated the following as independent predictors: the texture parameter of homogeneity and the morphological parameter of sphericity. In the Kaplan-Meier analysis, the PFS rate was significantly improved in the patients who had both a higher value of homogeneity (p=0.01) and a higher value of sphericity (p=0.002). With the combined use of homogeneity and sphericity, the patients with different PFS rates could be divided more clearly. CONCLUSION: The quantitative parameters of homogeneity and sphericity obtained by FDG-PET can be useful for the prediction of the PFS of pharynx SCC patients, especially when used in combination.

Usefulness of Pseudocontinuous Arterial Spin-Labeling for the Assessment of Patients with Head and Neck Squamous Cell Carcinoma by Measuring Tumor Blood Flow in the Pretreatment and Early Treatment Period.

BACKGROUND AND PURPOSE: For the assessment of the treatment response in non-surgical treatment, tumor blood flow provides the functional information of the tumor which is different from the morphological information such as tumor volume. The purpose of this study was to evaluate the diagnostic value of tumor blood flow values obtained by pseudocontinuous arterial spin-labeling in patients with head and neck squamous cell carcinoma. MATERIALS AND METHODS: Forty-one patients with head and neck squamous cell carcinoma were evaluated by using pseudocontinuous arterial spin-labeling. Quantitative tumor blood flow was calculated at the pretreatment and the early treatment periods in all the patients, and the percentage change of tumor blood flow between the two was calculated. At the early treatment period, based on their tumor volume reduction rate, we divided the patients into stable disease and partial response groups for a subgroup analysis. The local control or failure was confirmed either by histopathology or by radiologic evaluation within the follow-up. RESULTS: Pretreatment tumor blood flow in patients in the failure group was significantly lower than that in patients in the local control group. In the subgroup analysis of patients with stable disease, the percentage change of tumor blood flow was significantly larger (due to the tumor blood flow increase from pretreatment value) in the local control group than in the failure group. In addition, in patients with a partial response, the percentage change of tumor blood flow was significantly smaller (due to the tumor blood flow decrease from the pretreatment value) in the local control group than in the failure group. The accuracy for determination of the local control group or the failure group in pretreatment tumor blood flow was 0.83 and that in the combination use of the percentage change of tumor blood flow and tumor volume in the early treatment period was 0.93. CONCLUSIONS: Tumor blood flow obtained by pseudocontinuous arterial spin-labeling can be useful for the determination of local control. The combined use of the percentage change of tumor blood flow and tumor volume had particularly high diagnostic accuracy.

Differentiation of squamous cell carcinoma and inverted papilloma using non-invasive MR perfusion imaging.

OBJECTIVES: To investigate the diagnostic value of tumour blood flow (TBF) obtained with pseudocontinuous arterial spin labelling for the differentiation of squamous cell carcinoma (SCC) and inverted papilloma (IP) in the nasal or sinonasal cavity. METHODS: We retrospectively analysed the cases of 33 patients with SCC and 8 patients with IP in the nasal or sinonasal cavity. Pseudocontinuous arterial spin labelling scanning was performed for all patients using a 3.0-T MR unit. Quantitative TBF values were measured by two neuroradiologists by respectively delineating the whole-tumour regions of interest, and the mean of them was determined as TBF value in each patient. Additionally, the presence of imaging findings of convoluted cerebriform pattern (CCP) on MR T2 weighted images was determined in all patients. As a subgroup analysis, patients with IP were divided into aggressive and non-aggressive IPs depending on their progression range. First, an intraclass correlation coefficient (ICC) of TBF values between two neuroradiologists was determined. Next, a statistical comparison of the TBF value by a Mann-Whitney U test between the patients with SCC and IP was performed. Additionally, the comparison by an ANOVA with a post hoc test of Tukey's method among the SCC, non-aggressive IP and aggressive IP groups was also performed. If significance was observed, the diagnostic accuracy to differentiate SCCs from IPs was calculated. Diagnostic accuracy by CCP findings alone and by the combination of CCP findings and TBF were also assessed. RESULTS: The ICC of TBF values between two neuroradiologists was 0.82. The mean TBF values in the patients with SCC, all patients with IP, those with aggressive IP and those with non-aggressive IP were 141.2 ± 33.1, 77.8 ± 31.5, 109.4 ± 16.7 and 58.8 ± 19.9 ml 100 g⁻¹ min⁻¹, respectively. A significant difference was observed between SCC and IP (p < 0.001), SCC and non-aggressive IP (p < 0.01) and non-aggressive IP and aggressive IP (p < 0.01). The diagnostic accuracy values obtained with receiver operating characteristic curve analysis for the differentiation of SCC from IP and for SCC from non-aggressive IP were 0.90 and 0.92, respectively. The diagnostic accuracy was elevated (0.95 from 0.88) by adding the TBF value to CCP findings. CONCLUSIONS: The pseudocontinuous arterial spin labelling technique can be a useful non-invasive diagnostic tool to differentiate SCC from IP in nasal or sinonasal cavity.