The value of plasma hypoxia markers for predicting imaging-based hypoxia in patients with head-and-neck cancers undergoing definitive chemoradiation.

BACKGROUND: Tumor hypoxia worsens the prognosis of head-and-neck squamous cell carcinoma (HNSCC) patients, and plasma hypoxia markers may be used as biomarkers for radiotherapy personalization. We therefore investigated the role of the hypoxia-associated plasma proteins osteopontin, galectin-3, vascular endothelial growth factor (VEGF) and connective tissue growth factor (CTGF) as surrogate markers for imaging-based tumor hypoxia. METHODS: Serial blood samples of HNSCC patients receiving chemoradiation within a prospective trial were analyzed for osteopontin, galectin-3, VEGF and CTGF concentrations. Tumor hypoxia was quantified in treatment weeks 0, 2 and 5 using [18F]FMISO PET/CT. The association between PET-defined hypoxia and the plasma markers was determined using Pearson's correlation analyses. Receiver-operating characteristic analyses were conducted to reveal the diagnostic value of the hypoxia markers. RESULTS: Baseline osteopontin (r = 0.579, p < 0.01) and galectin-3 (r = 0.429, p < 0.05) correlated with the hypoxic subvolume (HSV) prior to radiotherapy, whereas VEGF (r = 0.196, p = 0.36) and CTGF (r = 0.314, p = 0.12) showed no association. Patients with an HSV > 1 mL in week 2 exhibited increased VEGF (p < 0.05) and CTGF (p < 0.05) levels in week 5. Pretherapeutic osteopontin levels were higher in patients exhibiting residual hypoxia at the end of treatment (104.7 vs. 60.8 ng/mL, p < 0.05) and could therefore predict residual hypoxia (AUC = 0.821, 95% CI 0.604-1.000, p < 0.05). CONCLUSION: In this exploratory analysis, osteopontin correlated with the initial HSV and with residual tumor hypoxia; therefore, there may be a rationale to study hypoxic modification based on osteopontin levels. However, as plasma hypoxia markers do not correspond to any spatial information of tumor hypoxia, they have limitations regarding the replacement of [18F]FMISO PET-based focal treatments. The results need to be validated in larger patient cohorts to draw definitive conclusions.

Interleukin-6 as surrogate marker for imaging-based hypoxia dynamics in patients with head-and-neck cancers undergoing definitive chemoradiation-results from a prospective pilot trial.

PURPOSE: Intratumoral hypoxia increases resistance of head-and-neck squamous cell carcinoma (HNSCC) to radiotherapy. [18F]FMISO PET imaging enables noninvasive hypoxia monitoring, though requiring complex logistical efforts. We investigated the role of plasma interleukin-6 (IL-6) as potential surrogate parameter for intratumoral hypoxia in HNSCC using [18F]FMISO PET/CT as reference. METHODS: Within a prospective trial, serial blood samples of 27 HNSCC patients undergoing definitive chemoradiation were collected to analyze plasma IL-6 levels. Intratumoral hypoxia was assessed in treatment weeks 0, 2, and 5 using [18F]FMISO PET/CT imaging. The association between PET-based hypoxia and IL-6 was examined using Pearson's correlation and multiple regression analyses, and the diagnostic power of IL-6 for tumor hypoxia response prediction was determined with receiver-operating characteristic analyses. RESULTS: Mean IL-6 concentrations were 15.1, 19.6, and 31.0 pg/mL at baseline, week 2 and week 5, respectively. Smoking (p=0.050) and reduced performance status (p=0.011) resulted in higher IL-6 levels, whereas tumor (p=0.427) and nodal stages (p=0.334), tumor localization (p=0.439), and HPV status (p=0.294) had no influence. IL-6 levels strongly correlated with the intratumoral hypoxic subvolume during treatment (baseline: r=0.775, p<0.001; week 2: r=0.553, p=0.007; week 5: r=0.734, p<0.001). IL-6 levels in week 2 were higher in patients with absent early tumor hypoxia response (p=0.016) and predicted early hypoxia response (AUC=0.822, p=0.031). Increased IL-6 levels at week 5 resulted in a trend towards reduced progression-free survival (p=0.078) and overall survival (p=0.013). CONCLUSION: Plasma IL-6 is a promising surrogate marker for tumor hypoxia dynamics in HNSCC patients and may facilitate hypoxia-directed personalized radiotherapy concepts. TRIAL REGISTRATION: The prospective trial was registered in the German Clinical Trial Register (DRKS00003830). Registered 20 August 2015.

18F-FMISO-PET Hypoxia Monitoring for Head-and-Neck Cancer Patients: Radiomics Analyses Predict the Outcome of Chemo-Radiotherapy.

Tumor hypoxia is associated with radiation resistance and can be longitudinally monitored by 18F-fluoromisonidazole (18F-FMISO)-PET/CT. Our study aimed at evaluating radiomics dynamics of 18F-FMISO-hypoxia imaging during chemo-radiotherapy (CRT) as predictors for treatment outcome in head-and-neck squamous cell carcinoma (HNSCC) patients. We prospectively recruited 35 HNSCC patients undergoing definitive CRT and longitudinal 18F-FMISO-PET/CT scans at weeks 0, 2 and 5 (W0/W2/W5). Patients were classified based on peritherapeutic variations of the hypoxic sub-volume (HSV) size (increasing/stable/decreasing) and location (geographically-static/geographically-dynamic) by a new objective classification parameter (CP) accounting for spatial overlap. Additionally, 130 radiomic features (RF) were extracted from HSV at W0, and their variations during CRT were quantified by relative deviations (∆RF). Prediction of treatment outcome was considered statistically relevant after being corrected for multiple testing and confirmed for the two 18F-FMISO-PET/CT time-points and for a validation cohort. HSV decreased in 64% of patients at W2 and in 80% at W5. CP distinguished earlier disease progression (geographically-dynamic) from later disease progression (geographically-static) in both time-points and cohorts. The texture feature low grey-level zone emphasis predicted local recurrence with AUCW2 = 0.82 and AUCW5 = 0.81 in initial cohort (N = 25) and AUCW2 = 0.79 and AUCW5 = 0.80 in validation cohort. Radiomics analysis of 18F-FMISO-derived hypoxia dynamics was able to predict outcome of HNSCC patients after CRT.

Evolution of the hypoxic compartment on sequential oxygen partial pressure maps during radiochemotherapy in advanced head and neck cancer.

BACKGROUND AND PURPOSE: Longitudinal Positron Emission Tomography (PET) with hypoxia-specific radiotracers allows monitoring the time evolution of regions of increased radioresistance and may become fundamental in determining the radiochemotherapy outcome in Head-and-Neck Squamous Cell Carcinoma (HNSCC). The aim of this study was to investigate the evolution of the hypoxic target volume on oxygen partial pressure maps (pO2-HTV) derived from 18FMISO-PET images acquired before and during radiochemotherapy and to uncover correlations between extent and severity of hypoxia and treatment outcome. MATERIAL AND METHODS: 18FMISO-PET/CT images were acquired at three time points (before treatment start, in weeks two and five) for twenty-eight HNSCC patients treated with radiochemotherapy. The images were converted into pO2 maps and corresponding pO2-HTVs (pO2-HTV1, pO2-HTV2, pO2-HTV3) were contoured at 10 mmHg. Different parameters describing the pO2-HTV time evolution were considered, such as the percent and absolute difference between the pO2-HTVs (%HTVi,j and HTVi-HTVj with i,j = 1, 2, 3, respectively) and the slope of the linear regression curve fitting the pO2-HTVs in time. Correlations were sought between the pO2-HTV evolution parameters and loco-regional recurrence (LRR) using the Receiver Operating Characteristic method. RESULTS: The Area Under the Curve values for %HTV1,2, HTV1-HTV2, HTV1-HTV3 and the slope of the pO2-HTV linear regression curve were 0.75 (p = 0.04), 0.73 (p = 0.02), 0.73 (p = 0.02) and 0.75 (p = 0.007), respectively. Other parameter combinations were not statistically significant. CONCLUSIONS: The pO2-HTV evolution during radiochemotherapy showed predictive value for LRR. The changes in the tumour hypoxia during the first two treatment weeks may be used for adaptive personalized treatment approaches.

Immunohistochemistry-based hypoxia-immune prognostic classifier for head-and-neck cancer patients undergoing chemoradiation - Post-hoc analysis from a prospective imaging trial.

PURPOSE: As both tumor hypoxia and an immunosuppressing tumor microenvironment hamper the anti-tumor activity of radiotherapy in head-and-neck squamous cell carcinoma (HNSCC), we aimed to develop an immunohistochemistry-based hypoxia-immune classifier. METHODS: 39 patients receiving definitive chemoradiation for HNSCC within a prospective trial were included in this analysis. Baseline tumor samples were analyzed for the hypoxia marker carbonic anhydrase IX (CAIX) and tumor-infiltrating lymphocytes (TILs) and were correlated with [18F]-misonidazole ([18F]FMISO) PET measurements. The impact of the biomarkers on the locoregional control (LRC) was examined using Cox analyses and concordance index statistics. RESULTS: Low CAIX (HR = 0.352, 95%CI 0.124-1.001, p = 0.050) and high TIL levels (HR = 0.308, 95%CI 0.114-0.828, p = 0.020) were independent parameters for improved LRC and did not correlate with each other (Spearman's ρ = 0.034, p = 0.846). Harrell's C was 0.66 for CAIX and TIL levels alone and 0.71 for the combination. 2-year LRC was 73%, 62% and 11% for the prognostically good (CAIXlow/TILhigh), intermediate (CAIXlow/TILlow or CAIXhigh/TILhigh) and poor groups (CAIXhigh/TILlow), respectively (p = 0.001). Focusing on T lymphocytes, the hypoxia-immune classifier could still stratify between favorable (CAIXlow/CD3 + TILhigh), intermediate (CAIXlow/CD3 + TILlow or CAIXhigh/CD3 + TILhigh) and poor subgroups (CAIXhigh/CD3 + TILlow) with a 2-year LRC of 80%, 59% and 14%, respectively (p = 0.001). There was a positive correlation between baseline CAIX levels and [18F]FMISO SUV in week 2 of chemoradiation (ρ = 0.324, p = 0.050), indicating an association between higher baseline CAIX expression and tumor hypoxia persistence. CONCLUSION: We developed a clinically feasible hypoxia-immune prognostic classifier for HNSCC patients based on pre-treatment immunohistochemistry. However, external validation is required to determine the prognostic value and the potential usage for personalized radiation oncology.

Lymphocyte Infiltration Determines the Hypoxia-Dependent Response to Definitive Chemoradiation in Head-and-Neck Cancer: Results from a Prospective Imaging Trial.

Tumor hypoxia in head-and-neck squamous cell carcinoma (HNSCC) leads to an immunosuppressive microenvironment and reduces the response to radiotherapy. In this prospective imaging trial, we investigated potential interactions between functional hypoxia imaging and infiltrating lymphocyte levels as a potential predictor for treatment response in HNSCC patients. Methods: In total, 49 patients receiving definitive chemoradiation for locally advanced HNSCCs underwent pretherapeutic biopsies and peritherapeutic hypoxia imaging using 18F-misonidazole PET at weeks 0, 2, and 5 during chemoradiation. Hematoxylin-eosin and immunohistochemical stainings for tumor-infiltrating lymphocytes, tissue-based hypoxia, and microvascular markers were analyzed and correlated with the longitudinal hypoxia dynamics and patient outcomes. Results: High levels of tumor-infiltrating total lymphocytes correlated with superior locoregional control (LRC) (hazard ratio [HR], 0.279; P = 0.011) and progression-free survival (PFS) (HR, 0.276; P = 0.006). Similarly, early resolution of 18F-misonidazole PET-detected tumor hypoxia quantified by 18F-misonidazole dynamics between weeks 0 and 2 of chemoradiation was associated with improved LRC (HR, 0.321; P = 0.015) and PFS (HR, 0.402; P = 0.043). Outcomes in the favorable early hypoxia resolution subgroup significantly depended on infiltrating lymphocyte counts, with patients who showed both an early hypoxia response and high lymphocyte infiltration levels exhibiting significantly improved LRC (HR, 0.259; P = 0.036) and PFS (HR, 0.242; P = 0.017) compared with patients with an early hypoxia response but low lymphocyte counts. These patients exhibited oncologic results comparable to those of patients with no hypoxia response within the first 2 wk of chemoradiation. Conclusion: This analysis established a clinical hypoxia-immune score that predicted treatment responses and outcomes in HNSCC patients undergoing chemoradiation and may help to devise novel concepts for biology-driven personalization of chemoradiation.

Hypoxia dynamics on FMISO-PET in combination with PD-1/PD-L1 expression has an impact on the clinical outcome of patients with Head-and-neck Squamous Cell Carcinoma undergoing Chemoradiation.

Tumor-associated hypoxia influences the radiation response of head-and-neck cancer (HNSCC) patients, and a lack of early hypoxia resolution during treatment considerably deteriorates outcomes. As the detrimental effects of hypoxia are partly related to the induction of an immunosuppressive microenvironment, we investigated the interaction between tumor hypoxia dynamics and the PD-1/PD-L1 axis in HNSCC patients undergoing chemoradiation and its relevance for patient outcomes in a prospective trial. Methods: 49 patients treated with definitive chemoradiation for locally advanced HNSCC were enrolled in this trial and received longitudinal hypoxia PET imaging using fluorine-18 misonidazole ([18F]FMISO) at weeks 0, 2 and 5 during treatment. Pre-therapeutic tumor biopsies were immunohistochemically analyzed regarding the PD-1/PD-L1 expression both on immune cells and on tumor cells, and potential correlations between the PD-1/PD-L1 axis and tumor hypoxia dynamics during chemoradiation were assessed using Spearman's rank correlations. Hypoxia dynamics during treatment were quantified by subtracting the standardized uptake value (SUV) index at baseline from the SUV values at weeks 2 or 5, whereby SUV index was defined as ratio of maximum tumor [18F]FMISO SUV to mean SUV in the contralateral sternocleidomastoid muscle (i.e. tumor-to-muscle ratio). The impact of the PD-1/PD-L1 expression alone and in combination with persistent tumor hypoxia on locoregional control (LRC), progression-free survival (PFS) and overall survival (OS) was examined using log-rank tests and Cox proportional hazards models. Results: Neither PD-L1 nor PD-1 expression levels on tumor-infiltrating immune cells influenced LRC (HR = 0.734; p = 0.480 for PD-L1, HR = 0.991; p = 0.989 for PD-1), PFS (HR = 0.813; p = 0.597 for PD-L1, HR = 0.796; p = 0.713 for PD-1) or OS (HR = 0.698; p = 0.405 for PD-L1, HR = 0.315; p = 0.265 for PD-1). However, patients with no hypoxia resolution between weeks 0 and 2 and PD-L1 expression on tumor cells, quantified by a tumor proportional score (TPS) of at least 1%, showed significantly worse LRC (HR = 3.374, p = 0.022) and a trend towards reduced PFS (HR = 2.752, p = 0.052). In the multivariate Cox regression analysis, the combination of absent tumor hypoxia resolution and high tumoral PD-L1 expression remained a significant prognosticator for impaired LRC (HR = 3.374, p = 0.022). On the other side, tumoral PD-L1 expression did not compromise the outcomes of patients whose tumor-associated hypoxia declined between week 0 and 2 during chemoradiation (LRC: HR = 1.186, p = 0.772, PFS: HR = 0.846, p = 0.766). Conclusion: In this exploratory analysis, we showed for the first time that patients with both persistent tumor-associated hypoxia during treatment and PD-L1 expression on tumor cells exhibited a worse outcome, while the tumor cells' PD-L1 expression did not influence the outcomes of patients with early tumor hypoxia resolution. While the results have to be validated in an independent cohort, these findings form a foundation to investigate the combination of hypoxic modification and immune checkpoint inhibitors for the unfavorable subgroup, moving forward towards personalized radiation oncology treatment.

Convolutional neural networks for head and neck tumor segmentation on 7-channel multiparametric MRI: a leave-one-out analysis.

BACKGROUND: Automatic tumor segmentation based on Convolutional Neural Networks (CNNs) has shown to be a valuable tool in treatment planning and clinical decision making. We investigate the influence of 7 MRI input channels of a CNN with respect to the segmentation performance of head&neck cancer. METHODS: Head&neck cancer patients underwent multi-parametric MRI including T2w, pre- and post-contrast T1w, T2*, perfusion (ktrans, ve) and diffusion (ADC) measurements at 3 time points before and during radiochemotherapy. The 7 different MRI contrasts (input channels) and manually defined gross tumor volumes (primary tumor and lymph node metastases) were used to train CNNs for lesion segmentation. A reference CNN with all input channels was compared to individually trained CNNs where one of the input channels was left out to identify which MRI contrast contributes the most to the tumor segmentation task. A statistical analysis was employed to account for random fluctuations in the segmentation performance. RESULTS: The CNN segmentation performance scored up to a Dice similarity coefficient (DSC) of 0.65. The network trained without T2* data generally yielded the worst results, with ΔDSCGTV-T = 5.7% for primary tumor and ΔDSCGTV-Ln = 5.8% for lymph node metastases compared to the network containing all input channels. Overall, the ADC input channel showed the least impact on segmentation performance, with ΔDSCGTV-T = 2.4% for primary tumor and ΔDSCGTV-Ln = 2.2% respectively. CONCLUSIONS: We developed a method to reduce overall scan times in MRI protocols by prioritizing those sequences that add most unique information for the task of automatic tumor segmentation. The optimized CNNs could be used to aid in the definition of the GTVs in radiotherapy planning, and the faster imaging protocols will reduce patient scan times which can increase patient compliance. TRIAL REGISTRATION: The trial was registered retrospectively at the German Register for Clinical Studies (DRKS) under register number DRKS00003830 on August 20th, 2015.

The utility of multiparametric MRI to characterize hypoxic tumor subvolumes in comparison to FMISO PET/CT. Consequences for diagnosis and chemoradiation treatment planning in head and neck cancer.

BACKGROUND AND PURPOSE: Hypoxia is an essential metabolic marker that determines chemo- and radiation resistance in head-and-neck squamous cell carcinoma (HNSCC) patients. Our exploratory analysis aimed to identify multiparametric MRI (mpMRI) parameters linked to hypoxia that might be used as surrogate for [18F]FMISO-PET in diagnosis and chemoradiation treatment (CRT) of HNSCC. MATERIALS AND METHODS: 21 patients undergoing definitive CRT for HNSCC were prospectively imaged with serial [18F]FMISO-PET and 3 Tesla mpMRI for T1- and T2-weighted and dynamic contrast-enhanced perfusion and diffusion-weighted measurements (ktrans, ve, kep, ADC) in weeks 0, 2 and 5 and FDG-PET in week 0. [18F]FMISO-PET-derived hypoxic subvolumes (HSV) and complementary non-hypoxic subvolumes (nonHSV) were created for tumor and lymph nodes and projected on the mpMRI scans after PET/MRI co-registration. MpMRI and [18F]FMISO-PET parameters within HSVs and nonHSVs were statistically compared. RESULTS: FMISO-PET-based HSVs of the primary tumors on MRI were characterized by lower ADC at all time points (p = 0.012 at baseline; p = 0.015 in week 2) and reduced interstitial space volume fraction ve and perfusion ktrans at baseline (p = 0.006, p = 0.047) compared to nonHSVs. Hypoxic lymph nodes were characterized by significantly lower ADC values at baseline (p = 0.039), but not at later time points and a reduction in ktrans-based perfusion at week 2 (p = 0.018). CONCLUSION: MpMRI parameters differ significantly between hypoxic and non-hypoxic tumor regions, defined on FMISO-PET/CT as gold standard and might represent surrogate markers for tumor hypoxia. These findings suggest that mpMRI may be useful in the future as a surrogate modality for hypoxia imaging in order to personalize CRT.

Correlative analyses between tissue-based hypoxia biomarkers and hypoxia PET imaging in head and neck cancer patients during radiochemotherapy-results from a prospective trial.

PURPOSE: Tumor hypoxia impairs the response of head-and-neck cancer (HNSCC) patients to radiotherapy and can be detected both by tissue biomarkers and PET imaging. However, the value of hypoxia biomarkers and imaging for predicting HNSCC patient outcomes are incompletely understood, and potential correlations between tissue and PET data remain to be elucidated. Here, we performed exploratory analyses of potential correlations between tissue-based hypoxia biomarkers and longitudinal hypoxia imaging in a prospective trial of HNSCC patients. METHODS: Forty-nine patients undergoing chemoradiation for locally advanced HNSCCs were enrolled in this prospective trial. They underwent baseline biopsies and [18F]FDG PET imaging and [18F]FMISO PET at weeks 0, 2, and 5 during treatment. Immunohistochemical analyses for p16, Ki67, CD34, HIF1α, CAIX, Ku80, and CD44 were performed, and HPV status was assessed. Biomarker expression was correlated with biological imaging information and patient outcome data. RESULTS: High HIF1α tumor levels significantly correlated with increased tumor hypoxia at week 2 as assessed by the difference in the [18F]FMISO tumor-to-background ratios, and high HIF1α and CAIX expressions were both associated with a deferred decrease in hypoxia between weeks 2 and 5. Loco-regional recurrence rates after radiotherapy were significantly higher in patients with high CAIX expression and also increased for high levels of the DNA repair factor Ku80. HPV status did not correlate with any of the tested hypoxia biomarkers, and HPV-positive patients showed higher loco-regional control rates and progression-free survival independent of their hypoxia dynamics. CONCLUSION: In this exploratory trial, high expression of the tissue-based hypoxia biomarkers HIF1α and CAIX correlated with adverse hypoxia dynamics in HNSCCs during chemoradiation as assessed by PET imaging, and high CAIX levels were associated with increased loco-regional recurrence rates. Hence, hypoxia biomarkers warrant further investigations as potential predictors of hypoxia dynamics and hypoxia-associated radiation resistance.