Quantitative Mid-treatment Imaging Biomarkers for Response Prediction After Radiotherapy in Head and Neck Cancer: A Systematic Review and Meta-analysis

Quantitative Mid-treatment Imaging Biomarkers for Response Prediction After Radiotherapy in Head and Neck Cancer: A Systematic Review and Meta-analysis | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Quantitative Mid-treatment Imaging Biomarkers for Response Prediction After Radiotherapy in Head and Neck Cancer: A Systematic Review and Meta-analysis Yuvnik Trada, Mark T Lee, Michael G Jameson, Paul Keall, Daniel Moses FRANZCR, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7689582/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 4 You are reading this latest preprint version Abstract Background: To systematically review and meta-analyse the prognostic value of quantitative mid-treatment imaging biomarkers for predicting locoregional tumour control in patients undergoing definitive radiotherapy for mucosal head and neck squamous cell carcinoma. Main body: A systematic literature search (2005–2023) was conducted in PubMed, EMBASE, Scopus, and Cochrane databases according to a pre-registered PROSPERO protocol. Studies evaluating quantitative imaging features derived from CT, MRI, or PET during radiotherapy were included. Imaging features were grouped as baseline, absolute mid-treatment, or relative mid-treatment (delta) parameters. A random-effects meta-analysis was performed on studies reporting receiver operating characteristic (ROC)-based area under the curve (AUC) values. Forty-one studies encompassing 1654 patients were included. Seventeen studies (n = 612 patients) reported sufficient data for meta-analysis. The pooled AUC for relative mid-treatment parameters was 0.796 (95% CI: 0.762–0.831), demonstrating higher predictive performance than absolute mid-treatment parameters (AUC 0.686; 95% CI: 0.628–0.745). Baseline parameters showed moderate predictive ability (AUC 0.736; 95% CI: 0.688–0.785), and while numerically lower than relative mid-treatment parameters, this difference did not reach statistical significance. Diffusion-weighted MRI (ΔADCmean) and FDG-PET (ΔMTV, ΔTLG) emerged as the most consistently predictive modalities. Relative measures offer practical advantages, including internal self-normalisation and improved reproducibility across imaging platforms. Conclusions: Relative mid-treatment imaging biomarkers demonstrate superior predictive performance compared to baseline and absolute measures, supporting their potential role in adaptive radiotherapy strategies. Further prospective multi-centre studies with standardised imaging protocols and external validation are essential for clinical translation. Head and Neck Neoplasms Systematic Review Magnetic Resonance Imaging Positron-Emission Tomography Tomography Recurrence Response Mid-treatment Figures Figure 1 Figure 2 Figure 3 Figure 4 BACKGROUND Mucosal head and neck cancers account for 4–5% of all cancers worldwide with rising incidence due to factors such as tobacco and alcohol consumption, and human papillomavirus (HPV) infection.[ 1 ] Definitive radiotherapy remains a cornerstone of organ-preserving treatment for mucosal head and neck squamous cell carcinoma (HNSCC). Despite curative-intent treatment, locoregional tumour recurrence occurs in 15–50% of patients, contributing to significant morbidity and mortality.[ 2 , 3 ] Moreover, among those who are cured, treatment is often associate [i] d with substantial acute and long-term toxicities. A reliable imaging biomarker capable of predicting treatment response early during therapy could facilitate personalised treatment strategies—enabling timely intensification (e.g., radiotherapy dose escalation, chemotherapy modification, or early salvage surgery) or de-escalation (e.g., radiotherapy dose or volume reduction, omission of chemotherapy). Medical imaging such as computed tomography (CT), magnetic resonance imaging (MRI) and positron-emission tomography (PET) offer the potential to characterise tumour biology using quantitative measures that can act as an early surrogate marker for treatment response. Anatomical imaging (CT and MRI) can provide morphological information such as tumour volume and cellular density, while functional imaging (e.g., PET and advanced MRI techniques) can assess tumour physiology—including perfusion, proliferation, hypoxia, and metabolism—parameters that are closely linked to treatment resistance and adverse clinical outcomes.[ 4 ] For instance, 18 F-fluorodeoxyglucose (FDG)-PET has the potential to assess tumour cell density, proliferation and cellular metabolism.[ 5 , 6 ] Novel PET tracers such as 18 F-fluromisonidazole (FMISO), nitroimidazole derivate 18 F-fluoroerythronitroimidazole (FETNIM) and 2-nitroimidazole nucleoside analogue 18 F-flortanidazole (FHX4) enable assessment of tumour hypoxia. Tumour cellular proliferation has been explored using thymidine analogue 3’-Deoxy-3’- 18 F-fluorothymidine (FLT) and 4’-methyl- 11 C-thiothymidine (4D-ST). Dynamic contrast-enhanced MRI (DCE-MRI) can provide information on tumour perfusion and permeability. Diffusion-weighted magnetic resonance imaging (DWI) depends on the microscopic mobility of water, which can be quantified using apparent diffusion coefficients (ADC) to measure tumour microenvironment and cellular density. The intra-voxel incoherent motion (IVIM) based on DWI technique, can be used to separate the signal for tissue perfusion from molecular diffusion to reflect tumour cellularity and vascularity. Imaging features that are extracted range from global measures (e.g., volume, maximum, mean) to higher order features that aim to map and quantify tumour heterogeneity. Published studies utilising only pre-treatment imaging for locoregional tumour response prediction have provided variable results.[ 7 , 8 ] Imaging performed during treatment, referred to as mid-treatment imaging, has the potential to be a better biomarker of treatment response by capturing early biological response to delivered therapy.[ 9 , 10 ] There has been a marked increase in studies evaluating the use of mid-treatment imaging in mucosal head and neck treatment during radiotherapy. Despite this interest, the results from mid-treatment imaging studies remain inconsistent, largely due to methodological heterogeneity in image acquisition, analysis, and reporting. A key area of variability lies in how mid-treatment imaging data are analysed and reported: some studies report static absolute imaging values at mid-treatment time points, while others report relative changes from baseline (i.e., delta-values). The use of absolute mid-treatment imaging values has the advantage of measuring tumour qualities of the residual tumour at a particular stage during treatment. Whereas, the use of relative mid-treatment imaging values has the advantage of measuring the change during treatment and hence sensitivity to prescribed treatment. Currently, there is no consensus on the optimal method for analysis of mid-treatment imaging biomarkers, and this lack of standardisation presents a barrier to the integration of mid-treatment imaging biomarkers into clinical decision-making. The aim of this systematic review is to summarise the existing literature on the prognostic value of mid-treatment imaging during radiotherapy for mucosal HNSCC using locoregional recurrence as the primary clinical outcome. Additionally, we aim to identify the optimal method for analysis of mid-treatment imaging parameters for treatment response assessment. METHODS A prospectively registered protocol is available through the International Prospective Register of Systematic Reviews (PROSPERO) (CRD42023484250). The review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.[ 11 ] Search strategy and study selection A systematic search of PubMed, EMBASE, Scopus, and the Cochrane Central Register of Controlled Trials (CCTR) was conducted to identify relevant published studies between 2005 and 2023 (see Appendix, Supplementary Material). The PICOS framework (patient, intervention, comparison, outcome, and study) was used to develop the review question, focusing on the predictive value of quantitative mid-treatment imaging for locoregional treatment response in head and neck cancer treated with definitive radiotherapy. As outlined in the pre-specified protocol, exclusion criteria included: (1) studies without quantitative image analysis; (2) studies published prior to 2005; (3) non-squamous cell carcinoma (non-SCC) histology; (4) primary cutaneous or thyroid malignancies; and (5) studies with fewer than 10 patients. Additionally, studies primarily involving nasopharyngeal carcinoma (NPC) were excluded due to the distinct biological and clinical behaviour of NPC. Articles involving patients treated with adjuvant (post-operative) radiotherapy were also excluded. Article screening, selection and data extraction were performed using the Covidence™ web-based systematic review platform. Two reviewers (YT, ML) independently reviewed titles and abstracts, with disagreements resolved by consensus or consultation with a third reviewer (MJ). In cases of overlapping patient populations, the study with the largest cohort was included. Data extraction The primary outcome for this review was locoregional disease control following completion of radiotherapy, considered the gold standard endpoint for treatment efficacy in this population. Data on study characteristics (author, design, year, eligible patients, median follow-up), patient demographics (age, primary sub-site, tumour stage), treatment details (radiotherapy dose, concurrent chemotherapy), imaging protocol (modality, timing, region of interest, segmentation method), clinical outcomes and imaging parameters. Data were extracted by two independent reviewers and where information was missing, assumptions were not made and data were excluded from synthesis. Imaging parameters were classified into types of features; global (i.e. volume, max, mean, median), histogram, textural and parametric map. Time of mid-treatment imaging was noted in weeks and further classified as early (week 1 to 3) or late (week 4 to 7). Quality assessment We assessed the risk of bias and applicability of the eligible studies using the Quality Assessment of Diagnostic Accuracy studies (QUADAS-2) tool.[ 12 ] This tool assesses four key domains: patient selection, index test, reference standard, and flow and timing, each of which is evaluated for risk of bias, while the first three are also assessed for concerns regarding applicability. Patient selection examines how participants were recruited and whether selection methods could introduce bias. The index test domain evaluates how the imaging was conducted and interpreted, including whether blinding or pre-specified thresholds were used. The reference standard assesses the reliability and appropriateness of the outcome measure used to determine true disease status. Flow and timing considers whether all patients received the same reference standard and whether there were any delays or losses to follow-up that could affect the results. Perceived quality were graded as low, high or unclear risk. Data synthesis and Meta-analysis Imaging parameters extracted from baseline and mid-treatment imaging including absolute and/or delta-values (Δ, relative change in values acquired from imaging before and during treatment) were included in data synthesis. Imaging parameters were grouped into three categories for analysis: (1) baseline; (2) absolute mid-treatment; and (3) relative mid-treatment (delta) values. For the meta-analysis, studies that performed receiver operating characteristics (ROC) analysis were selected. Imaging parameters were compared using the reported AUC values and standard errors from the ROC analysis in the MedCalc software. Heterogeneity across studies was quantified using the Higgins’ I² statistic (I² 75%: high heterogeneity), which estimates the proportion of variation due to heterogeneity rather than chance. A random-effects model was employed to evaluate AUC values and compare subgroups of baseline, absolute mid-treatment, and relative mid-treatment imaging parameters. Results were tabulated and visualised using forest plots and funnel plots to display individual study outcomes and pooled estimates. RESULTS Study selection A total of 1972 articles were identified through the search strategy. After removal of duplicates, 1,533 articles were screened. Full-text review was performed on 228 articles, resulting in 41 studies that met the inclusion and exclusion criteria. The study selection process and reasons for article exclusion are summarised in the PRISMA flowchart (Fig. 1). Common reasons for exclusion included lack of locoregional treatment outcomes, absence of mid-treatment imaging, lack of predictive or prognostic analysis, and review article format. Study characteristics The total study population consisted of 1654 patients, of whom 82% received chemoradiotherapy and 18% received radiotherapy alone. Median follow-up ranged from 12 to 57 months, though not reported in five studies. The majority of studies were prospectively conducted (83%, 34 of 41). Twenty-seven studies used one imaging modality[13–39], 12 studies used two imaging modalities[40–51], and two studies used three imaging modalities[52, 53]. When reporting mid-treatment parameters, 21 (51%) studies reported relative values, five (12%) studies reported absolute values and 15 (37%) studies reported both absolute and relative values. There was also methodological heterogeneity, including differences in ROI selection, ROI delineation methods and timing of mid-treatment imaging. See summary of included studies (Table 1a and 1b). MRI imaging was utilised in 19 studies, PET imaging in 19 studies and CT imaging in 9 studies. DWI-MRI was the most common mid-treatment imaging modality reported (15 studies), followed by FDG-PET (12 studies).(Table 2) Study quality Study quality was assessed using the QUADAS-2 tool. The risk of bias was low across the four domains with 85%, 76%, 83% and 85% of studies having low risk in the patient selection, index test, reference standard and timing domains, respectively.(Fig. 2) Overall, 14 out of 41 studies (34%) demonstrated high risk of bias in at least one of the domains. Reference standard domain posed the highest risk of bias (17%, 7 of 41 studies). Concern regarding applicability was generally low for most studies with 61%, 83% and 90% of studies having low concerns of applicability in the patient selection, index test and reference standard domains, respectively.(Fig. 2) Overall, 16 out of 41 studies (39%) demonstrated high concerns regarding applicability in at least one of the domains. Patient selection domain posed the highest concern regarding applicability (24%, 10 of 41 studies). CT studies CT was used in nine studies involving 357 patients.(Table 2) Anatomical CT was the most common modality analysed (6 of 9 studies)[22, 28, 32, 40, 42, 49], with all studies employing manual ROI segmentation. Global parameters, particularly tumour volume, were most frequently analysed (6 of 7 studies). A range of relative change in tumour volumes were noted on mid-treatment imaging (median 19–68%) due to methodological variability in choice of ROI and timing of mid-treatment imaging. Three studies reported statistically significant correlations[22, 28, 42], and three studies did not find any correlation between mid-treatment CT volume and clinical outcomes[32, 40, 49]. The largest CT study by Kabarriti et al. (n = 96 patients) found a relative change in CT tumour volume (cut-off Δ19%, HR 0.26) but not absolute CT tumour volume at week 3 correlated to locoregional outcomes.[22] Similarly, Lee et al. reported that both absolute volume (11 cc, AUC 0.847) and relative change in volume (Δ75%, AUC 0.725) at week 4 mid-treatment CT were predictive for locoregional outcomes.[28] Conversely, studies by Arens et al, Mishra et al and Trada et al did not find any correlation between change in CT based tumour volume and clinical outcomes.[32, 40, 49] Morgan et al. conducted a matched case-control study (n = 90 patients) using textural analysis and found that relative change in a cluster of mid-treatment CT features were associated with locoregional recurrence.[34] Three studies used quantitative contrast enhanced CT[13, 37, 38], all employing manual ROI segmentation and primarily analysing global features (2 of 3 studies). Studies by Troung et al. and Ursino et al. reported multiple mid-treatment perfusion characteristics. Troung et al. (n = 15 patients) found that both absolute value (136 vs 44 ml/100g/min) and relative change (Δ, 28% vs 18%) in mid-treatment blood flow measurements at week 2 were higher in patients with tumour control.[37] Ursino et al. (n = 25 patients) found relative change in blood volume, blood flow and permeability surface measurements at week 3 correlated to early FDG-PET tumour response.[38] Abramyuk et al. (n = 15 patients) measured multiple perfusion characteristics at week 2 and week 5 during radiotherapy but did not find correlation to tumour outcomes.[13] PET studies PET was utilised in 19 studies containing 756 patients. (Table 2) FDG-PET was the most frequent PET tracer based imaging analysed (12 of 19 studies).[15, 17, 19, 25, 29, 30, 42, 46, 48, 49, 52, 53] Other novel hypoxia PET tracers based imaging utilised included F-MISO (2 studies)[27, 39], FHX4 (1 study)[35], and FETNIM (1 study)[21]. Novel tumour proliferation PET tracers utilised included FLT PET (2 studies) and 4D-ST PET (1 study)[46]. An advantage of PET imaging is relative ease of implementation of semi-automated segmentation methods for ROI delineation which was used in majority of PET studies (12 of 19). Global features such as maximum voxel value (SUV max ), metabolic tumour volume (MTV) and total lesional glycolysis (TLG) were the commonly utilised (19 studies). Textural features were analysed in only 1 study.[25] FDG-PET was the most utilised PET tracer due to its relative ease of access from its current clinical use in diagnosis, staging and post treatment monitoring in management of head and neck malignancies. The majority of FDG-PET studies performed imaging early (week 1–3) during radiotherapy (9 of 12 studies). Overall, there were nine FDG studies[19, 25, 29, 30, 42, 48, 49, 52, 53] that found a correlation between mid-treatment imaging features and clinical outcomes and three studies[15, 17, 46] which did not. A study by Min et al, utilised week 3 FDG PET in 75 patients during radiotherapy and analysed a range of absolute and relative mid-treatment global features from primary tumours.[30] They found absolute mid-treatment values of SUV max (cut-off value 4.3g/mL, 2 year LRRFS 89% vs 75%, HR 3.7, p = 0.03), MTV (cut-off value 3.3cm 3 , 2 year LRRFS 90% vs 69%, HR 3.9, p = 0.02) and TLG (cut off value 9.4g, 2 year LRRFS 93% vs 71%, HR 6.3, p < 0.01) correlated to locoregional recurrence. Of the studies that compared the performance of multiple mid-treatment global features, relative change in MTV and its closely related TLG were frequently found to be the best predictor of tumour control (Lin et al., ΔMTV 50%, HR 0.11, Trada et al, ΔMTV AUC 0.833; Trada et al., ΔMTV AUC 0.761; Wong et al. ΔTLG 0.825).[29, 48, 49, 53] Lafata et al. performed textural analysis from week 2 FDG-PET in 64 patients with oropharyngeal malignancies.[25] They utilised an in-house software to analyse multiple radiomic features and unsupervised learning to identify three intra-treatment imaging clusters. They found a difference between the intra-treatment clusters and tumour recurrence (HR 10.5, p < 0.01). Four studies utilised hypoxia PET tracers and analysed global imaging parameters such as SUV max or hypoxic volume for treatment response prediction, with moderate success. Two studies employed first generation hypoxia tracer, F-MISO PET to correlate changes in hypoxia during radiotherapy with tumour recurrence.[27, 39] Lazzeroni et al. utilised a volumetric measure of hypoxia sub-volume (%HTV) and found that a relative change (AUC 0.75, cut-off value < 44.3%), but not absolute HTV value at week 2 correlated to locoregional recurrence.[27] Wiedenman et al. utilised a modified hypoxia SUV max value (TBR max ) and found that patients with local tumour recurrence had higher absolute TBR max values at week 2 (p = 0.031) but not week 5 (p = 0.071).[39] Alternate hydrophilic hypoxia PET tracers such as FHX4 and FETNIM with lower background tissue uptake were developed to improve the tumour to background signal and improve clearance with favourable pharmacokinetic properties. A study by Sanduleanu et al., involving 34 patients, found that relative residual hypoxic volume at week 2 on FHX4 PET correlated to local control (p = 0.028).[35] Hu et al. reported a decrease in SUV max at week 5 during treatment on FETNIM PET, but no significant correlation to local tumour control.[21] Two studies analysed FLT-PET to investigate the correlation between tumour proliferation and treatment response.[20, 40] A study by Hoeben et al. found a decrease in tumour volume and SUVmax during radiotherapy.[20] They found relative change in proliferative tumour volume (cut-off value > 78%, 3 year LRC 100% vs 68%, p = 0.021) at week 4 correlated to locoregional outcomes. They did not find a correlation between relative change in SUVmax to outcomes. A study by Arens et al. employed variety of semi-automated methods for measuring proliferative tumour volumes but found no correlation to locoregional outcomes at either week 2 or week 4 during radiotherapy.[40] A study by Mitamura et al. utilised 4D-ST, a cell proliferation tracer that is resistant to thymidine phosphorylase degradation, and performed mid-treatment imaging at week 4 to calculate SUV max , proliferation tumour volume (PTV) and total lesional proliferation (TLP).[46] They found change in tumour volume (ΔPTV AUC 0.91, cut-off value − 90%), total proliferation (ΔTLP AUC 0.89, cut-off value − 45%) and maximum voxel volume (ΔSUV max AUC 0.75, cut-off value − 99%) discriminated early treatment response. MRI studies MRI was utilised in 19 studies containing 771 patients (Table 2). DWI MRI was the most common modality analysed (15 of 19 studies).[16, 18, 23, 26, 31, 33, 36, 43–45, 47, 48, 50, 52, 53] Other MRI modalities utilised included anatomical MRI (7 studies)[14, 41, 43–45, 47, 51], DCE-MRI (4 studies),[41, 51–53] and MR spectroscopy (1 study).[24] All 19 MRI studies utilised manual segmentation for ROI delineation. Global imaging features were again the most commonly parameters utilised (13 studies); however, several studies implemented histogram analysis (5 studies), textural features (2 studies) and parametric maps (1 studies). All seven anatomical MRI imaging utilised tumour volume, while one study also utilised textural features for response assessment.[47] There were four studies that found statistically significant correlations[14, 43, 45, 47], and three studies that did not find any correlation between mid-treatment MRI volume and clinical outcomes.[41, 44, 51] Study by Bhatia et al. correlated primary tumour volume delineated on T1 MRI during week 2 of radiotherapy to clinical outcomes.[14] They found that absolute change in tumour volume (AUC 0.721, cut-off > 10.6cc) had stronger correlation compared to relative change in tumour volume (AUC 0.689, cut-off < 9.4%) to local failure. A similar study by Khattab et al. also found that absolute primary tumour volume during week 2–3 of radiotherapy correlated to local failure (median 2.5cc vs 14.9cc, p = 0.007).[43] They found that relative change in tumour volume did not reach statistical significance for local failure in their patient population (median 73% vs 35%, p = 0.100). In contrast, the largest series by Cao et al., containing 54 patients did not find a correlation between absolute or relative mid-treatment anatomical tumour volume on T1 MRI imaging and tumour control.[41] Study by Scalco et al. utilised higher order radiomics and found relative change in week 3 mid-treatment textural features (Variance, Accuracy 68%; Fractional dimension, Accuracy 64%) on T2 imaging correlated to tumour recurrence.[47] The majority of studies utilising DWI MRI imaging analysed global features (13 of 15 studies), predominantly mean apparent diffusion coefficient value (ADC mean ). Tumour volume based on ADC maps was only evaluated in three studies with no correlation found with locoregional clinical outcomes.[18, 50, 52] The largest DWI series containing 81 patients by Mohamed et al., found relative change in primary tumour ADCmean at week 3 had the strongest correlation to local tumour control (cut-off < 7%, 2 year LC 28% vs 96%).[33] A study by Trada et al. analysed primary tumour ADC mean at multiple time-points and found that relative change but not absolute mid-treatment ADC mean correlated to local tumour control.[48] Relative change in primary tumour ADC mean is a promising imaging biomarker for treatment response prediction. Relative change in ADC mean with optimal cut-off (OC) values ranging from 16% to 33% (ΔADC mean ) was consistently found to have strong correlation to clinical outcomes in studies by Khattab et al. (AUC 0.786, OC 33%), Kim et al. (AUC 0.88), Marzi et al. (AUC 0.662, OC 16%), Matoba et al. (AUC 0.900, OC 24%), Trada et al. (AUC 0.825, OC 24%), Vandecaveye et al. (AUC 0.97, OC 25%), and Wong et al. (AUC 0.937, OC 17%).[31, 43–45, 48, 50, 53] Due to the image resolution of ADC maps derived from DWI MRI it is possible to derive histogram and higher order imaging features to quantitatively assess microenvironment and physiological processes for tumour response prediction. Histogram analysis was utlised in four studies[16, 23, 43, 52], higher order textural features were utilised in two studies[36, 47], and parametric maps were analysed in one study.[18] A study by Tomita et al. demonstrated the feasibility of applying artificial intelligence, using deep learning of mid-treatment DWI maps for imaging feature extraction, which correlated with local tumour recurrence (AUC 0.767).[36] Intravoxel incoherent motion diffusion-weighted imaging (IVIM-DWI) MRI can be used to assess both water diffusion and microcirculation (perfusion) within tissues. It leverages the principle of DWI, but uses lower b-values to specifically isolate and quantify perfusion information. Two studies utlised IVIM for tumour response prediction. Study by Marzi et al. found that absolute mid-treatment diffusion co-efficient (D, AUC 0.75), perfusion co-efficient (D*, AUC 0.688) and perfusion fraction (f, AUC 0.829) correlated to nodal recurrence. Whereas Martens et al., did not find statistically significant correlations between these imaging features and clinical outcomes. DCE-MRI is a complex MRI technique that analyses tissue enhancement patterns after the injection of a contrast agent to assess blood flow, vessel permeability, and tissue volume fractions. It provides functional information about tissue perfusion and microcirculation, particularly in tumours that can be used for treatment response assessment. DCE-MRI was analysed in four studies with moderate success utilising mid-treatment imaging features such as K trans (volume transfer constant), V e (fractional volume of extravascular extracellular space), K ep (transfer rate of contrast agent from extravascular extracellular space to plasma.[41, 51–53] A study by Wong et al. found relative change in K trans at week 2 had the highest correlation to clinical response (AUC 0.813).[53] Similarly a study by Martens et al., found relative change in K ep , and absolute mid-treatment V e , K trans and K ep correlated to tumour outcomes.[52] Magnetic Resonance Spectroscopy (MRS) is a technique used alongside MRI to analyse the chemical composition of tissues. MRS focuses on identifying and quantifying specific metabolites within a selected region, studied most commonly in brain tumours. King et al. utlised a novel MRS technique to measure changes in mid-treatment tumoral metabolite Choline in patients with head and neck cancer. They detected changes in Choline, Choline:Creatinine and Choline:Water ratios during week 2 of radiotherapy but these did not correlate with clinical outcomes in their patient population. Meta-analysis Seventeen studies containing 612 patients reported ROC analysis and were included in the meta-analysis. A total of 63 imaging parameters were evaluated: 17 baseline parameters, and 46 mid-treatment parameters (37 relative; 9 absolute). Using a random-effects model, the pooled AUC for baseline parameters was 0.736 (95% CI: 0.688–0.785). Relative mid-treatment parameters showed improved performance (AUC 0.796; 95% CI: 0.762–0.831) compared to absolute mid-treatment parameters (AUC 0.686; 95% CI: 0.628–0.745) (Fig. 3, Supplementary Table 1). A test of heterogeneity was low across all parameters included in the meta-analysis (I² = 46.8%) (Fig. 4). When stratified, baseline features showed moderate heterogeneity (I² = 51.5%), while absolute (I² = 34.6%) and relative (I² = 47.9%) mid-treatment parameters demonstrated low heterogeneity (Supplementary Fig. 1). Heterogeneity assessed using the Higgins’ I² statistic (I² 75%: high heterogeneity). DISCUSSION This systematic review and meta-analysis evaluated the prognostic value of quantitative mid-treatment imaging biomarkers in predicting locoregional tumour control in patients with HNSCC undergoing definitive radiotherapy. Our findings affirm the potential of mid-treatment imaging as a clinically relevant early-response biomarker. Across 41 eligible studies, mid-treatment imaging particularly relative changes (Δ) from baseline parameters were consistently associated with improved prediction of treatment outcomes. The meta-analysis of 17 studies involving 612 patients demonstrated a pooled AUC of 0.796 for relative mid-treatment parameters, compared with 0.736 for baseline and 0.686 for absolute mid-treatment values, suggesting that dynamic biological changes captured during therapy provide stronger prognostic insight than static measurements. Among imaging modalities, DWI-MRI and FDG-PET emerged as the most frequently studied and predictive. However, significant heterogeneity exists across studies in terms of imaging acquisition, timing, parameter extraction, and outcome reporting, which currently limits clinical implementation. MRI particularly diffusion-weighted MRI (DWI), was the most studied and consistently predictive for tumour outcomes. Across multiple studies, relative change in ADC mean (ΔADC mean ) emerged as a robust imaging biomarker, with optimal cut-offs ranging from 16% to 33% and AUC values as high as 0.97. Seven DWI studies by Khattab et al. (AUC 0.786), Kim et al. AUC 0.88), Marzi et al. (AUC 0.662), Matoba et al. (AUC 0.900), Trada et al. (AUC 0.825), Vandecaveye et al. (AUC 0.97), and Wong et al. (AUC 0.937) demonstrated strong correlation with locoregional tumour control.[ 31 , 43 – 45 , 48 , 50 , 53 ] These consistent findings reflect the ability of ΔADC mean to quantify early cytotoxic effects of radiotherapy, such as reduced tumour cellularity and increased water diffusion due to processes such as progressive necrosis, with successful treatment. A significant challenge in applying ADC-based biomarkers across centres is the variability introduced by differences in scanner hardware, acquisition protocols, and b-value selection, which can lead to inconsistent absolute ADC values.(reference) This variability limits the generalisability of absolute ADC thresholds and complicates multi-institutional standardisation. In contrast, relative mid-treatment changes in ADC (i.e., ΔADC mean ) performed on the same scanner offer a degree of self-normalisation, as each patient effectively serves as their own internal control. This inherent normalisation reduces variability and enhances reproducibility, making relative ADC metrics more suitable for wider clinical adoption. Moreover, DWI-based histogram and textural analyses were applied in several studies (e.g., Tomita et al., Scalco et al.), suggesting an evolving role for radiomic and AI-based approaches to further refine predictive performance.[ 36 , 47 ] PET imaging, particularly FDG-PET, also demonstrated strong predictive value, with relative volume based parameters ΔMTV and ΔTLG frequently emerging as the most informative features. Several studies (e.g., Lin et al., Trada et al., Wong et al.) reported ΔMTV cut-offs ranging from 33% to 75% to be predictive of locoregional control, with AUC values up to 0.833.[ 29 , 49 , 53 ] Biologically, a decrease in metabolic tumour volume during treatment likely reflects a combination of tumour cell death, reduced glucose metabolism, and improved oxygenation, and may serve as an early surrogate for radiosensitivity. Conversely, a lack of significant metabolic reduction may indicate radioresistant subvolumes, which could be targeted for dose escalation. One of the key advantages of FDG-PET is its relative widespread clinical availability and established role in diagnostic and post-treatment imaging, which has led to greater standardisation of acquisition protocols across scanners compared to other functional imaging modalities. FDG-PET also offers practical advantages in clinical implementation, including the use of semi-automated region-of-interest (ROI) segmentation, which reduces inter-observer variability and enhances reproducibility. These features, combined with the potential for integration with automated workflows, make FDG-PET a feasible platform for real-time response assessment and personalised adaptive radiotherapy treatment planning in routine practice. Emerging novel PET tracers targeting hypoxia (FMISO, FHX4, FETNIM), and proliferation (FLT, 4D-ST) add physiological specificity. For instance, Sanduleanu et al. found that residual hypoxic volume on FHX4 PET at week 2 correlated with local control (p = 0.028)[ 35 ], while Lazzeroni et al. found a ΔHTV threshold < 44.3% on FMISO PET (AUC 0.75) was predictive of recurrence.[ 27 ] Similarly, Mitamura et al. demonstrated high predictive value for ΔPTV (AUC 0.91) and ΔTLP (AUC 0.89) using 4D-ST PET.[ 46 ] While these tracers are not yet widely adopted, they provide granular insight into treatment-resistant subvolumes, offering targets for future radiotherapy dose-painting strategies. In contrast, results from CT-based studies were more mixed. Among nine studies, six reported global volumetric changes with highly variable correlation to outcomes. Kabarriti et al. found a ΔCT volume of 19% at week 3 significantly predicted locoregional outcomes (HR 0.26)[ 22 ], while Lee et al. reported both absolute and ΔCT volumes at week 4 to be prognostic (AUCs 0.847 and 0.725, respectively).[ 28 ] Conversely, other studies (e.g., Arens et al., Mishra et al., and Trada et al.,) found no significant correlations.[ 32 , 40 , 49 ] The relatively poor soft tissue contrast of CT, dependence on manual ROI delineation with limited reproducibility, and lack of standardised acquisition protocols likely contribute to the inconsistent performance of CT-based biomarkers and limit their applicability in multicentre settings. However, CT texture analysis (Morgan et al.) and perfusion CT parameters (Truong et al., Ursino et al.) showed promise, suggesting that advanced feature extraction from routine imaging may salvage prognostic utility from standard CT.[ 34 , 37 , 38 ] This study represents the first meta-analysis to quantitatively synthesise ROC-based AUC values for predictive performance of mid-treatment imaging in HNSCC. The use of ROC-based AUC measures offers a robust and standardised approach to assess the discriminatory ability of imaging biomarkers, independent of specific thresholds or prevalence, making it particularly suitable for comparing performance across heterogeneous studies. In our pooled meta-analysis, the area under the curve (AUC) for baseline imaging parameters was 0.736 (95% CI: 0.688–0.785). Relative mid-treatment parameters demonstrated numerically superior performance with an AUC of 0.796 (95% CI: 0.762–0.831); however, the overlapping confidence intervals indicate this difference did not reach statistical significance. This result should be interpreted with caution given that our analysis was not conducted with the aim of comparing the utility of mid-treatment imaging parameters compared to their corresponding baseline imaging parameters. In contrast, relative mid-treatment parameters demonstrated significantly higher performance with an AUC of 0.796 (95% CI: 0.762–0.831) compared to absolute mid-treatment parameters with an AUC of 0.686 (95% CI: 0.628–0.745). While prior reviews (e.g., Martens et al. and Lin et al.) provided narrative synthesis, they did not directly compare the performance of baseline, absolute, and relative parameters.[ 10 , 52 ] Our findings validate the rationale for delta imaging features, where temporal changes in imaging biomarkers during radiotherapy may reflect the biological responsiveness of individual tumours. This dynamic assessment enables early differentiation between radiosensitive and radioresistant disease, providing a surrogate marker of radiotherapy efficacy that can be projected to eventual clinical outcomes following completion of treatment. Furthermore, this review highlights the increasing incorporation of multi-parametric, multi-modality models, where features from DWI, PET, and CT can be combined to improve prognostic accuracy. Mid-treatment imaging has clear potential for personalised treatment adaptation. For responders, de-intensification strategies could mitigate long-term toxicities, while for non-responders, early escalation or surgical salvage may be indicated. Mid-treatment parameters may also assist in adaptive radiotherapy planning, identifying tumour sub-volumes for radiotherapy dose boost or omitting high-dose to regressing areas. Importantly, relative mid-treatment parameters may offer inherent self-normalisation, minimising inter-patient variability and scanner-related noise. Yet, challenges to clinical integration remain. These include a lack of standardised imaging protocols, diverse segmentation approaches, and variability in analytical pipelines. The reproducibility and interpretability of complex features, especially in multi-institution settings also limit generalisability. Thus, prioritising simple, reproducible metrics (e.g., ΔADC mean , ΔMTV) may be a pragmatic step toward clinical translation. Several limitations must be acknowledged. First, individual study sample sizes were relatively small, and most lacked external validation. Only a subset (17 of 41) reported ROC-based AUC values, limiting the meta-analysis scope. Second, publication and selection biases likely inflated performance metrics. Studies that did not find significant correlations were less likely to report full statistical metrics or be published at all. Additionally, we excluded nasopharyngeal carcinoma and adjuvant radiotherapy studies due to differing biology and treatment protocols, limiting generalisability. Finally, there was substantial heterogeneity in imaging acquisition, time-points (week 1–7), and ROI methods, which restricts pooled interpretation. Future studies should pursue large, prospective multi-centre trials with harmonised imaging protocols, clearly defined mid-treatment timepoints, and publicly available datasets for external validation. Furthermore, multi-modality imaging platforms such as PET/MRI may unlock new opportunities for combining anatomical, functional, and molecular imaging to guide adaptive radiotherapy. CONCLUSIONS Mid-treatment quantitative imaging offers prognostic value in predicting locoregional control in mucosal head and neck squamous cell carcinoma treated with definitive radiotherapy. This systematic review and meta-analysis demonstrate that relative mid-treatment imaging parameters outperform both baseline and absolute mid-treatment measures, with DWI-MRI (ΔADC mean ) and FDG-PET (ΔMTV, ΔTLG) emerging as the most consistent and clinically promising biomarkers. Despite encouraging results, substantial methodological heterogeneity and lack of standardised imaging protocols currently limit clinical translation. Future research should focus on large, prospective, multi-centre studies with harmonised imaging protocols, robust feature extraction pipelines, and external validation of predictive models. Integration of multi-parametric imaging, radiomics, and AI-based tools into clinical workflows may be required to realise the potential of mid-treatment imaging as a decision-support tool for adaptive radiotherapy in head and neck cancer. Declarations Ethics approval and consent to participate: N/A Consent for publication: N/A Availability of data and material: The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request. Competing interests: Nil Funding: Nil Authors' contributions: YT/ML/MJ contributed to data collection and systematic review. AF/PK/DM contributed to research question formation, study design and major review of manuscript. YT was responsible for statistical analysis with notable external contributions. All authors read and approved the final manuscript Acknowledgements: The authors acknowledge Helen Ball from Image X institute, University of Sydney for assistance and guidance for statistical analysis during preparation of the manuscript. References Siegel RL, Miller KD, Jemal A. Cancer statistics , 2016. CA Cancer J Clin, 2016. 66(1): pp. 7–30. Lee NY, et al. Avelumab plus standard-of-care chemoradiotherapy versus chemoradiotherapy alone in patients with locally advanced squamous cell carcinoma of the head and neck: a randomised, double-blind, placebo-controlled, multicentre, phase 3 trial. Lancet Oncol. 2021;22(4):450–62. Lacas B, et al. Meta-analysis of chemotherapy in head and neck cancer (MACH-NC): An update on 107 randomized trials and 19,805 patients, on behalf of MACH-NC Group. Radiother Oncol. 2021;156:281–93. Meijer TW, et al. Targeting hypoxia, HIF-1, and tumor glucose metabolism to improve radiotherapy efficacy. Clin Cancer Res. 2012;18(20):5585–94. Cegla P, et al. Assessment of biological parameters in head and neck cancer based on in vivo distribution of (18)F-FDG-FLT-FMISO-PET/CT images. Tumori. 2020;106(1):33–8. Zhu A, et al. Application of metabolic PET imaging in radiation oncology. Radiat Res. 2012;177(4):436–48. Bos P, et al. Prognostic functional MR imaging parameters in head and neck squamous cell carcinoma: A systematic review. Eur J Radiol. 2021;144:109952. Min M, et al. A review of the predictive role of functional imaging in patients with mucosal primary head and neck cancer treated with radiation therapy. J Med Imaging Radiat Oncol. 2017;61(1):99–123. Garibaldi C, et al. Interim (18)F-FDG PET/CT During Chemoradiation Therapy in the Management of Head and Neck Cancer Patients: A Systematic Review. Int J Radiat Oncol Biol Phys. 2017;98(3):555–73. Lin P, et al. Prognostic and predictive values of baseline and mid-treatment FDG-PET in oropharyngeal carcinoma treated with primary definitive (chemo)radiation and impact of HPV status: Review of current literature and emerging roles. Radiother Oncol. 2023;184:109686. Page MJ, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71. Whiting PF, et al. QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies. Ann Intern Med. 2011;155(8):529–36. Abramyuk A, et al. Radiochemotherapy-induced changes of tumour vascularity and blood supply estimated by dynamic contrast-enhanced CT and fractal analysis in malignant head and neck tumours. Br J Radiol. 2015;88(1045):20140412. Bhatia KS, et al. Does primary tumour volumetry performed early in the course of definitive concomitant chemoradiotherapy for head and neck squamous cell carcinoma improve prediction of primary site outcome? Br J Radiol. 2010;83(995):964–70. Castaldi P, et al. Can early and late18F-FDG PET-CT be used as prognostic factors for the clinical outcome of patients with locally advanced head and neck cancer treated with radio-chemotherapy? Radiother Oncol. 2012;103(1):63–8. Joint H, et al. Longitudinal diffusion and volumetric kinetics of head and neck cancer magnetic resonance on a 1.5 T MR-linear accelerator hybrid system: A prospective R-IDEAL stage 2a imaging biomarker characterization/pre-qualification study. Clin Transl Radiat Oncol. 2023;42:100666. Farrag A, et al. Can 18F-FDG-PET response during radiotherapy be used as a predictive factor for the outcome of head and neck cancer patients? Nucl Med Commun. 2010;31(6):495–501. Galban CJ, et al. A feasibility study of parametric response map analysis of diffusion-weighted magnetic resonance imaging scans of head and neck cancer patients for providing early detection of therapeutic efficacy. Transl Oncol. 2009;2(3):184–90. Hentschel M, et al. Early FDG PET at 10 or 20 Gy under chemoradiotherapy is prognostic for locoregional control and overall survival in patients with head and neck cancer. Eur J Nucl Med Mol Imaging. 2011;38(7):1203–11. Hoeben BA, et al. 18F-FLT PET during radiotherapy or chemoradiotherapy in head and neck squamous cell carcinoma is an early predictor of outcome. J Nucl Med. 2013;54(4):532–40. Hu M, et al. Hypoxia with 18F-fluoroerythronitroimidazole integrated positron emission tomography and computed tomography (18F-FETNIM PET/CT) in locoregionally advanced head and neck cancer: Hypoxia changes during chemoradiotherapy and impact on clinical outcome. Med (Baltim). 2019;98(40):e17067. Kabarriti R, et al. Early Response Assessment on Mid-treatment Computed Tomography Predicts for Locoregional Recurrence in Oropharyngeal Cancer Patients Treated With Definitive Radiation Therapy. Int J Radiat Oncol Biol Phys. 2018;102(4):1036–45. King AD, et al. Head and neck squamous cell carcinoma: diagnostic performance of diffusion-weighted MR imaging for the prediction of treatment response. Radiology. 2013;266(2):531–8. King AD, et al. Pretreatment and early intratreatment prediction of clinicopathologic response of head and neck cancer to chemoradiotherapy using 1H-MRS. J Magn Reson Imaging. 2010;32(1):199–203. Lafata KJ, et al. Intrinsic radiomic expression patterns after 20 Gy demonstrate early metabolic response of oropharyngeal cancers. Med Phys. 2021;48(7):3767–77. Lambrecht M, et al. Redefining the target early during treatment. Can we visualize regional differences within the target volume using sequential diffusion weighted MRI? Radiother Oncol. 2014;110(2):329–34. Lazzeroni M, et al. Evolution of the hypoxic compartment on sequential oxygen partial pressure maps during radiochemotherapy in advanced head and neck cancer. Phys Imaging Radiat Oncol. 2021;17:100–5. Lee YH, et al. Using primary tumor volumetry to predict treatment outcome for patients with oropharyngeal cancer who were treated with definitive chemoradiotherapy. Asia Pac J Clin Oncol. 2018;14(2):e21–8. Lin P, et al. Mid-treatment Fluorodeoxyglucose Positron Emission Tomography in Human Papillomavirus-related Oropharyngeal Squamous Cell Carcinoma Treated with Primary Radiotherapy: Nodal Metabolic Response Rate can Predict Treatment Outcomes. Clin Oncol (R Coll Radiol). 2021;33(12):e586–98. Min M, et al. Prognostic role of metabolic parameters of (18)F-FDG PET-CT scan performed during radiation therapy in locally advanced head and neck squamous cell carcinoma. Eur J Nucl Med Mol Imaging. 2015;42(13):1984–94. Marzi S, et al. The prediction of the treatment response of cervical nodes using intravoxel incoherent motion diffusion-weighted imaging. Eur J Radiol. 2017;92:93–102. Mishra S, et al. Can radiological changes in lymph node volume during treatment predict success of radiation therapy in patients with locally advanced head and neck squamous cell carcinoma? J Med Imaging Radiat Oncol. 2013;57(5):603–9. Joint H, et al. Prospective validation of diffusion-weighted MRI as a biomarker of tumor response and oncologic outcomes in head and neck cancer: Results from an observational biomarker pre-qualification study. Radiother Oncol. 2023;183:109641. Morgan HE, et al. Exploratory ensemble interpretable model for predicting local failure in head and neck cancer: the additive benefit of CT and intra-treatment cone-beam computed tomography features. Quant Imaging Med Surg. 2021;11(12):4781–96. Sanduleanu S, et al. 18)F]-HX4 PET/CT hypoxia in patients with squamous cell carcinoma of the head and neck treated with chemoradiotherapy: Prognostic results from two prospective trials. Clin Transl Radiat Oncol. 2020;23:9–15. Tomita H, et al. Deep learning approach of diffusion-weighted imaging as an outcome predictor in laryngeal and hypopharyngeal cancer patients with radiotherapy-related curative treatment: a preliminary study. Eur Radiol. 2022;32(8):5353–61. Truong MT, et al. Prediction of locoregional control in head and neck squamous cell carcinoma with serial CT perfusion during radiotherapy. AJNR Am J Neuroradiol. 2011;32(7):1195–201. Ursino S, et al. Role of perfusion CT in the evaluation of functional primary tumour response after radiochemotherapy in head and neck cancer: preliminary findings. Br J Radiol. 2016;89(1065):20151070. Wiedenmann NE, et al. Serial [18F]-fluoromisonidazole PET during radiochemotherapy for locally advanced head and neck cancer and its correlation with outcome. Radiother Oncol. 2015;117(1):113–7. Arens AI, et al. Semiautomatic methods for segmentation of the proliferative tumour volume on sequential FLT PET/CT images in head and neck carcinomas and their relation to clinical outcome. Eur J Nucl Med Mol Imaging. 2014;41(5):915–24. Cao Y, et al. Predictive Values of MRI and PET Derived Quantitative Parameters for Patterns of Failure in Both p16 + and p16- High Risk Head and Neck Cancer. Front Oncol. 2019;9:1118. Chen SW, et al. Interim FDG PET/CT for predicting the outcome in patients with head and neck cancer. Laryngoscope. 2014;124(12):2732–8. Khattab HM et al. Diffusion-weighted magnetic resonance imaging (DWMRI) of head and neck squamous cell carcinoma: could it be an imaging biomarker for prediction of response to chemoradiation therapy. Egypt J Radiol Nucl Med, 2020. 51(204). Kim S, et al. Diffusion-weighted magnetic resonance imaging for predicting and detecting early response to chemoradiation therapy of squamous cell carcinomas of the head and neck. Clin Cancer Res. 2009;15(3):986–94. Matoba M, et al. Fractional change in apparent diffusion coefficient as an imaging biomarker for predicting treatment response in head and neck cancer treated with chemoradiotherapy. AJNR Am J Neuroradiol. 2014;35(2):379–85. Mitamura K, et al. Interim 4'-[methyl-(11)C]-thiothymidine PET for predicting the chemoradiotherapeutic response in head and neck squamous cell carcinoma: comparison with [(18)F]FDG PET. EJNMMI Res. 2021;11(1):13. Scalco E, et al. Characterization of cervical lymph-nodes using a multi-parametric and multi-modal approach for an early prediction of tumor response to chemo-radiotherapy. Phys Med. 2016;32(12):1672–80. Trada Y, et al. Changes in serial multiparametric MRI and FDG-PET/CT functional imaging during radiation therapy can predict treatment response in patients with head and neck cancer. Eur Radiol. 2023;33(12):8788–99. Trada Y, et al. Impact of tumour region of interest delineation method for mid-treatment FDG-PET response prediction in head and neck squamous cell carcinoma undergoing radiotherapy. Quant Imaging Med Surg. 2023;13(5):2822–36. Vandecaveye V, et al. Predictive value of diffusion-weighted magnetic resonance imaging during chemoradiotherapy for head and neck squamous cell carcinoma. Eur Radiol. 2010;20(7):1703–14. Wang P, et al. An approach to identify, from DCE MRI, significant subvolumes of tumors related to outcomes in advanced head-and-neck cancer. Med Phys. 2012;39(8):5277–85. Martens RM et al. Early Response Prediction of Multiparametric Functional MRI and (18)F-FDG-PET in Patients with Head and Neck Squamous Cell Carcinoma Treated with (Chemo)Radiation. Cancers (Basel), 2022. 14(1). Wong KH, et al. Changes in multimodality functional imaging parameters early during chemoradiation predict treatment response in patients with locally advanced head and neck cancer. Eur J Nucl Med Mol Imaging. 2018;45(5):759–67. Tables Tables are available in the Supplementary Files section. Supplementary Files Table1New.pdf AppendixSearchstrategy.docx Supplementarymaterial230925.docx Table2Overviewofincludedstudies.docx Cite Share Download PDF Status: Under Review Version 1 posted Reviewers agreed at journal 27 Mar, 2026 Reviewers invited by journal 15 Oct, 2025 Editor assigned by journal 28 Sep, 2025 First submitted to journal 25 Sep, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7689582","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":529956276,"identity":"508f701a-1594-4121-9263-b43df9c42b11","order_by":0,"name":"Yuvnik Trada","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABPUlEQVRIie3RQUvDMBQH8Fci9RLs9dXWzo8QKWwMxz5LR6FespMgwooNCPMiei34KcQvUCi0l3oveBkIPe1QUaSgiNmsB1fFHQX7J5CXl/xIIABt2vzVONFyYlACmGsTXBAlBKB1j/1iakLoOoTlvDubZXDSO4uvHwf+kLL0ttjO/VerJ0jy8ATDkWiQHnNyQDPzDg2euJRlB11ZMNuMVPfKBLdJPA+dEhCBM2MsCNWFpxpcZaMQqE0QyM9Em9svfRFQ/bKQ5I0FNQmaxE1w8TBE3jUUEVMN5S3jKXNQEqWEeJXo2b0kGeohFkf98ySVpCD74wt7LySqS4Cl9grZSkdTvUoGGmruTV75E0vVPOWOP1sd3DyNlep4srNCdqPlhF+aGx9LIgdt/k9HNFryYPlZKdU3223atGnz3/IOSApnOjOeuf0AAAAASUVORK5CYII=","orcid":"https://orcid.org/0000-0001-7935-6904","institution":"Calvary Mater Newcastle","correspondingAuthor":true,"prefix":"","firstName":"Yuvnik","middleName":"","lastName":"Trada","suffix":""},{"id":529956277,"identity":"6ca6e554-470e-4667-adaf-3425acd9490c","order_by":1,"name":"Mark T Lee","email":"","orcid":"","institution":"South Western Sydney Local Health District","correspondingAuthor":false,"prefix":"","firstName":"Mark","middleName":"T","lastName":"Lee","suffix":""},{"id":529956278,"identity":"025d78dc-b42a-4213-96bd-5ae8303a8921","order_by":2,"name":"Michael G Jameson","email":"","orcid":"","institution":"St Vincent's University Hospital","correspondingAuthor":false,"prefix":"","firstName":"Michael","middleName":"G","lastName":"Jameson","suffix":""},{"id":529956279,"identity":"c648ec7e-afa8-49aa-a63d-750e6cab2734","order_by":3,"name":"Paul Keall","email":"","orcid":"","institution":"The University of Sydney","correspondingAuthor":false,"prefix":"","firstName":"Paul","middleName":"","lastName":"Keall","suffix":""},{"id":529956280,"identity":"e3022fd0-3ebd-4e52-8092-402c86641135","order_by":4,"name":"Daniel Moses FRANZCR","email":"","orcid":"","institution":"Prince of Wales Hospital","correspondingAuthor":false,"prefix":"","firstName":"Daniel","middleName":"Moses","lastName":"FRANZCR","suffix":""},{"id":529956281,"identity":"f0f78c9e-8c1f-4d30-8e52-be5bbd5c8eb9","order_by":5,"name":"Allan Fowler FRANZCR","email":"","orcid":"","institution":"South Western Sydney Local Health District","correspondingAuthor":false,"prefix":"","firstName":"Allan","middleName":"Fowler","lastName":"FRANZCR","suffix":""}],"badges":[],"createdAt":"2025-09-23 04:48:06","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7689582/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7689582/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":94709291,"identity":"7d6c4e3b-a808-4f71-b6a1-2019b0ee48e2","added_by":"auto","created_at":"2025-10-30 01:06:47","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":69976,"visible":true,"origin":"","legend":"","description":"","filename":"FIGURE1PRISMAMidtreatmentimagingbiomarkerinHNSCC.docx","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/0d43955e4caf9778f4aa81e6.docx"},{"id":94709295,"identity":"68286805-e18b-480c-853e-11b698ba8af4","added_by":"auto","created_at":"2025-10-30 01:06:47","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":108252,"visible":true,"origin":"","legend":"","description":"","filename":"Figure2QUADAS2.docx","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/8f69fdbc15d5edb0af01cfc9.docx"},{"id":94709293,"identity":"80ec15d5-2090-4ed0-92b6-6bb85d1db3ce","added_by":"auto","created_at":"2025-10-30 01:06:47","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":165761,"visible":true,"origin":"","legend":"","description":"","filename":"Figure3.docx","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/ca2661bc973bc374edc7e7d9.docx"},{"id":94729454,"identity":"a4f464ae-038f-42f1-be8c-b0b257cbfabb","added_by":"auto","created_at":"2025-10-30 07:04:59","extension":"docx","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":172076,"visible":true,"origin":"","legend":"","description":"","filename":"Figure4.docx","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/5b7f33c21b7e6042769093d2.docx"},{"id":94729510,"identity":"20cab586-8cf8-4d16-a88a-3e53d20612a0","added_by":"auto","created_at":"2025-10-30 07:05:04","extension":"xml","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":12344,"visible":true,"origin":"","legend":"","description":"","filename":"ejreEJRED2500406.xml","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/de70049014cd4f9e13d6ed2a.xml"},{"id":94709310,"identity":"2d856857-6c9f-4cef-b5f6-b7ef5cecbc82","added_by":"auto","created_at":"2025-10-30 01:06:47","extension":"xml","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1356,"visible":true,"origin":"","legend":"","description":"","filename":"EJRED250040611179.go.xml","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/47c3ad856fd71326d684ebd5.xml"},{"id":94822797,"identity":"eee6098e-2cab-4f84-8ee9-f63fb27effba","added_by":"auto","created_at":"2025-10-31 06:44:13","extension":"xml","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":928,"visible":true,"origin":"","legend":"","description":"","filename":"EJRED2500406Import.xml","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/79553e93929222fd057bf4fb.xml"},{"id":94709313,"identity":"250544b6-393c-4b49-8962-04fd8f1679ce","added_by":"auto","created_at":"2025-10-30 01:06:47","extension":"xml","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":145340,"visible":true,"origin":"","legend":"","description":"","filename":"EJRED25004060enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/e077ffa14264acc718979005.xml"},{"id":94709306,"identity":"26f4dd80-b480-49e7-a901-ca967818c780","added_by":"auto","created_at":"2025-10-30 01:06:47","extension":"eps","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":652102,"visible":true,"origin":"","legend":"","description":"","filename":"drawingimage1.eps","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/6bb2015a78813df6fbfe09b5.eps"},{"id":94709307,"identity":"c081b9be-4446-4f52-be47-ae02969cec99","added_by":"auto","created_at":"2025-10-30 01:06:47","extension":"eps","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":56984,"visible":true,"origin":"","legend":"","description":"","filename":"drawingimage2.eps","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/b13576af12fef5466ed081fa.eps"},{"id":94729754,"identity":"b5b71bf4-4fa7-48eb-bd87-38e97108ecdc","added_by":"auto","created_at":"2025-10-30 07:05:22","extension":"jpeg","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":59348,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/ccc6e24f5fcbb2dbe9bfd4d1.jpeg"},{"id":94729534,"identity":"6e9c8373-afc3-428e-b2fb-17b0be7b5f03","added_by":"auto","created_at":"2025-10-30 07:05:07","extension":"png","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":44728,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage10.png","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/e8f67d2012eef863aeefb7c8.png"},{"id":94728441,"identity":"feb14a59-e2b2-433c-937f-7427697d5362","added_by":"auto","created_at":"2025-10-30 07:03:49","extension":"jpeg","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":481929,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage11.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/2fe0323318b118f19f8d6bda.jpeg"},{"id":94729778,"identity":"7a6f9734-6fe1-43c8-a0ae-92fa1013168b","added_by":"auto","created_at":"2025-10-30 07:05:24","extension":"jpeg","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1074,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage12.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/9afa04ff5f0c05953aacfc36.jpeg"},{"id":94709302,"identity":"05b96719-2eec-46aa-883d-5059e5d875bc","added_by":"auto","created_at":"2025-10-30 01:06:47","extension":"jpeg","order_by":19,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":75256,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage13.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/cecf77b3195af160e3c2f95b.jpeg"},{"id":94729096,"identity":"fc2beadf-9957-4e6b-a35b-9307b13d329c","added_by":"auto","created_at":"2025-10-30 07:04:37","extension":"jpeg","order_by":20,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1080136,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage14.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/4ebed8082699a8c29a3a9360.jpeg"},{"id":94729168,"identity":"e12a0634-34b1-42b8-b1dd-06c361447f9a","added_by":"auto","created_at":"2025-10-30 07:04:41","extension":"jpeg","order_by":21,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":334728,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage15.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/5fd9c126f9f32044f96193e6.jpeg"},{"id":94709304,"identity":"a9e5f217-21ab-44a6-bf3d-a0fb9de3ae11","added_by":"auto","created_at":"2025-10-30 01:06:47","extension":"jpeg","order_by":22,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":143235,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/60ec6b0fd9cac6102fb06b37.jpeg"},{"id":94709323,"identity":"79a776f6-3340-48b7-a176-a9d3f39e509b","added_by":"auto","created_at":"2025-10-30 01:06:47","extension":"jpeg","order_by":23,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":52874,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/2a2b7b0ea715205124eb5265.jpeg"},{"id":94730082,"identity":"68c7ab8c-cadd-4f8c-a23c-652b26c70fcf","added_by":"auto","created_at":"2025-10-30 07:05:39","extension":"jpeg","order_by":24,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":743995,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/5d8af68853b25946d2777873.jpeg"},{"id":94709334,"identity":"d9f4f98d-5291-48da-adc6-e80f9594aac0","added_by":"auto","created_at":"2025-10-30 01:06:48","extension":"jpeg","order_by":25,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":49276,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage5.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/fd9ccf33de1dbd7e68762219.jpeg"},{"id":94709314,"identity":"35bc502c-9eb9-479e-821c-73c51f1a51d5","added_by":"auto","created_at":"2025-10-30 01:06:47","extension":"jpeg","order_by":26,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":24619,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage6.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/2a66d3a8889ed7a786da7ad8.jpeg"},{"id":94709309,"identity":"f828f110-e1bc-4a46-98e8-7cf2949dc5e3","added_by":"auto","created_at":"2025-10-30 01:06:47","extension":"jpeg","order_by":27,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":55700,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage7.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/dd5b5dc7861ac4a08f0cc071.jpeg"},{"id":94709331,"identity":"8aa6b129-5bff-4560-8e22-5fe44ef89fa4","added_by":"auto","created_at":"2025-10-30 01:06:48","extension":"jpeg","order_by":28,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1074,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage12.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/9bf25be2078061500ac6a843.jpeg"},{"id":94729814,"identity":"2fd55e9c-0993-4d87-9f97-7d1e766b9115","added_by":"auto","created_at":"2025-10-30 07:05:25","extension":"png","order_by":29,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":96057,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage9.png","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/ab3fb2399cf9848899834e44.png"},{"id":94709300,"identity":"8e14ba7b-0201-4e17-9355-ece2a59eefc7","added_by":"auto","created_at":"2025-10-30 01:06:47","extension":"png","order_by":30,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":11239,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/b5ec682c8b5e8b365bb43716.png"},{"id":94709326,"identity":"98fb806d-1f30-42b3-938e-68ec7f0489c4","added_by":"auto","created_at":"2025-10-30 01:06:48","extension":"png","order_by":31,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":24111,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage10.png","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/7fa62b97315e0a1be7cf9d34.png"},{"id":94709317,"identity":"c0b607fb-86e4-47d9-93c9-ccffba596178","added_by":"auto","created_at":"2025-10-30 01:06:47","extension":"png","order_by":32,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":87055,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage11.png","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/a6abf36bf0edc54c58d56500.png"},{"id":94709315,"identity":"8b264765-8c68-4719-b4ee-cee40dde5835","added_by":"auto","created_at":"2025-10-30 01:06:47","extension":"png","order_by":33,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":935,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage12.png","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/a2bf03311e4b57ad73764faf.png"},{"id":94709329,"identity":"822ee65e-623f-4ce4-9f7b-9bac12094ed3","added_by":"auto","created_at":"2025-10-30 01:06:48","extension":"png","order_by":34,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":13984,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage13.png","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/19e336791d126b94ac62e28c.png"},{"id":94709312,"identity":"1aee55b1-04c2-4d0a-9420-41bdb272db32","added_by":"auto","created_at":"2025-10-30 01:06:47","extension":"png","order_by":35,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":207255,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage14.png","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/e7eaf008fd61a0b7957d95de.png"},{"id":94729036,"identity":"7f5f0712-eb1e-4c52-8710-c9c779f7ad61","added_by":"auto","created_at":"2025-10-30 07:04:31","extension":"png","order_by":36,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":63581,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage15.png","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/fcef1a74fc0ade524942f7e7.png"},{"id":94709319,"identity":"e6b487c2-ad20-407d-af71-0afe964564d0","added_by":"auto","created_at":"2025-10-30 01:06:47","extension":"png","order_by":37,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":20834,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/f0468d6605c8040379a927ce.png"},{"id":94709332,"identity":"1af6172b-e5aa-4bbb-803c-318c6513fce1","added_by":"auto","created_at":"2025-10-30 01:06:48","extension":"png","order_by":38,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":10626,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/6a9115faf40d01d5b87e52d2.png"},{"id":94730102,"identity":"f97d7fc8-a1c4-4695-b216-46693910a514","added_by":"auto","created_at":"2025-10-30 07:05:40","extension":"png","order_by":39,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":121397,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/b499f6fde37d52f7aa455982.png"},{"id":94709337,"identity":"617d4887-ee20-43a7-81bd-f01848ecd68a","added_by":"auto","created_at":"2025-10-30 01:06:48","extension":"png","order_by":40,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":6019,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/d162ca0f5b658908378f37aa.png"},{"id":94729998,"identity":"4afa5d58-8792-44e8-9af9-2fac9544edd7","added_by":"auto","created_at":"2025-10-30 07:05:33","extension":"png","order_by":41,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":5841,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/56c8ff3005cc2edd8a064543.png"},{"id":94709328,"identity":"d320508c-f20c-482e-8f70-1c7f5de1af48","added_by":"auto","created_at":"2025-10-30 01:06:48","extension":"png","order_by":42,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":8627,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/2ad14eb79f452016979ea321.png"},{"id":94730074,"identity":"f43bd857-245a-44a5-b219-31ebfa5caae6","added_by":"auto","created_at":"2025-10-30 07:05:39","extension":"png","order_by":43,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":935,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage12.png","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/5a10be4586ce794ec6289338.png"},{"id":94729815,"identity":"3fcb1434-a948-45e3-b1ff-a746aa7c9725","added_by":"auto","created_at":"2025-10-30 07:05:25","extension":"png","order_by":44,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":26073,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage9.png","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/cc64cc55acdf2b0216a329cc.png"},{"id":94730095,"identity":"d14fa314-bc8f-4382-af73-239ae25bb8f9","added_by":"auto","created_at":"2025-10-30 07:05:40","extension":"xml","order_by":45,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":144842,"visible":true,"origin":"","legend":"","description":"","filename":"EJRED25004060structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/b4fd23c24c51d7415aac085b.xml"},{"id":94709333,"identity":"d2a664b0-d135-4380-a5a3-d0a5dd85aa8c","added_by":"auto","created_at":"2025-10-30 01:06:48","extension":"html","order_by":46,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":159027,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/37720c763b32674ef70658c2.html"},{"id":94709290,"identity":"12d954a6-d561-46e1-a58e-e275dea4101f","added_by":"auto","created_at":"2025-10-30 01:06:47","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":528728,"visible":true,"origin":"","legend":"\u003cp\u003eSee image above for figure legend.\u003c/p\u003e","description":"","filename":"F1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/552401e3e798117ab87d1890.jpg"},{"id":94709289,"identity":"05997e84-e30e-422e-a04c-bfe997321d56","added_by":"auto","created_at":"2025-10-30 01:06:47","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":431955,"visible":true,"origin":"","legend":"\u003cp\u003eSee image above for figure legend.\u003c/p\u003e","description":"","filename":"F2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/492968c9d068c544c24eeb2f.jpg"},{"id":94709298,"identity":"c1d6dd18-8f0b-46d3-aea9-3f49b1035d29","added_by":"auto","created_at":"2025-10-30 01:06:47","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":1729731,"visible":true,"origin":"","legend":"\u003cp\u003eSee image above for figure legend.\u003c/p\u003e","description":"","filename":"3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/1bd034994189aa7028e51dd2.jpg"},{"id":94729726,"identity":"21944d91-46b7-4590-abf5-8dc6f4670344","added_by":"auto","created_at":"2025-10-30 07:05:19","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":303226,"visible":true,"origin":"","legend":"\u003cp\u003eSee image above for figure legend.\u003c/p\u003e","description":"","filename":"F6.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/fcc6805ff604543e4bcc53e7.jpg"},{"id":94827125,"identity":"3294064d-9ac3-4b52-b7cb-5da39a327f98","added_by":"auto","created_at":"2025-10-31 06:54:46","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3526269,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/61894d20-36e7-4f4a-a890-bc177b5798dd.pdf"},{"id":94728800,"identity":"1b3108e9-0cd9-4c3f-ae28-3cdd54c0bc9c","added_by":"auto","created_at":"2025-10-30 07:04:17","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":901418,"visible":true,"origin":"","legend":"","description":"","filename":"Table1New.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/2f8fcf04a85ff633d02c7724.pdf"},{"id":94729975,"identity":"1b4a2eb8-84d9-408c-a0c5-494e2523ce0f","added_by":"auto","created_at":"2025-10-30 07:05:32","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":22268,"visible":true,"origin":"","legend":"","description":"","filename":"AppendixSearchstrategy.docx","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/5f2b72b08f85e329fc0ea3d8.docx"},{"id":94709297,"identity":"a47deecc-bd6a-45ae-91d0-ab8dee7aadce","added_by":"auto","created_at":"2025-10-30 01:06:47","extension":"docx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":187086,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementarymaterial230925.docx","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/d90260b05d002471e2f951ad.docx"},{"id":94709308,"identity":"ef3d4bdc-2703-4399-bea8-d5bd40b4b85c","added_by":"auto","created_at":"2025-10-30 01:06:47","extension":"docx","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":22630,"visible":true,"origin":"","legend":"","description":"","filename":"Table2Overviewofincludedstudies.docx","url":"https://assets-eu.researchsquare.com/files/rs-7689582/v1/0735637ce2f40e1011c74038.docx"}],"financialInterests":"","formattedTitle":"\u003cp\u003eQuantitative Mid-treatment Imaging Biomarkers for Response Prediction After Radiotherapy in Head and Neck Cancer: A Systematic Review and Meta-analysis\u003c/p\u003e","fulltext":[{"header":"BACKGROUND","content":"\u003cp\u003eMucosal head and neck cancers account for 4\u0026ndash;5% of all cancers worldwide with rising incidence due to factors such as tobacco and alcohol consumption, and human papillomavirus (HPV) infection.[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e] Definitive radiotherapy remains a cornerstone of organ-preserving treatment for mucosal head and neck squamous cell carcinoma (HNSCC). Despite curative-intent treatment, locoregional tumour recurrence occurs in 15\u0026ndash;50% of patients, contributing to significant morbidity and mortality.[\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e] Moreover, among those who are cured, treatment is often associate\u003csup\u003e[i]\u003c/sup\u003ed with substantial acute and long-term toxicities. A reliable imaging biomarker capable of predicting treatment response early during therapy could facilitate personalised treatment strategies\u0026mdash;enabling timely intensification (e.g., radiotherapy dose escalation, chemotherapy modification, or early salvage surgery) or de-escalation (e.g., radiotherapy dose or volume reduction, omission of chemotherapy).\u003c/p\u003e\u003cp\u003eMedical imaging such as computed tomography (CT), magnetic resonance imaging (MRI) and positron-emission tomography (PET) offer the potential to characterise tumour biology using quantitative measures that can act as an early surrogate marker for treatment response. Anatomical imaging (CT and MRI) can provide morphological information such as tumour volume and cellular density, while functional imaging (e.g., PET and advanced MRI techniques) can assess tumour physiology\u0026mdash;including perfusion, proliferation, hypoxia, and metabolism\u0026mdash;parameters that are closely linked to treatment resistance and adverse clinical outcomes.[\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e] For instance, \u003csup\u003e18\u003c/sup\u003eF-fluorodeoxyglucose (FDG)-PET has the potential to assess tumour cell density, proliferation and cellular metabolism.[\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e] Novel PET tracers such as \u003csup\u003e18\u003c/sup\u003eF-fluromisonidazole (FMISO), nitroimidazole derivate \u003csup\u003e18\u003c/sup\u003eF-fluoroerythronitroimidazole (FETNIM) and 2-nitroimidazole nucleoside analogue \u003csup\u003e18\u003c/sup\u003eF-flortanidazole (FHX4) enable assessment of tumour hypoxia. Tumour cellular proliferation has been explored using thymidine analogue 3\u0026rsquo;-Deoxy-3\u0026rsquo;-\u003csup\u003e18\u003c/sup\u003eF-fluorothymidine (FLT) and 4\u0026rsquo;-methyl-\u003csup\u003e11\u003c/sup\u003eC-thiothymidine (4D-ST). Dynamic contrast-enhanced MRI (DCE-MRI) can provide information on tumour perfusion and permeability. Diffusion-weighted magnetic resonance imaging (DWI) depends on the microscopic mobility of water, which can be quantified using apparent diffusion coefficients (ADC) to measure tumour microenvironment and cellular density. The intra-voxel incoherent motion (IVIM) based on DWI technique, can be used to separate the signal for tissue perfusion from molecular diffusion to reflect tumour cellularity and vascularity. Imaging features that are extracted range from global measures (e.g., volume, maximum, mean) to higher order features that aim to map and quantify tumour heterogeneity. Published studies utilising only pre-treatment imaging for locoregional tumour response prediction have provided variable results.[\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]\u003c/p\u003e\u003cp\u003eImaging performed during treatment, referred to as mid-treatment imaging, has the potential to be a better biomarker of treatment response by capturing early biological response to delivered therapy.[\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e] There has been a marked increase in studies evaluating the use of mid-treatment imaging in mucosal head and neck treatment during radiotherapy. Despite this interest, the results from mid-treatment imaging studies remain inconsistent, largely due to methodological heterogeneity in image acquisition, analysis, and reporting. A key area of variability lies in how mid-treatment imaging data are analysed and reported: some studies report static absolute imaging values at mid-treatment time points, while others report relative changes from baseline (i.e., delta-values). The use of absolute mid-treatment imaging values has the advantage of measuring tumour qualities of the residual tumour at a particular stage during treatment. Whereas, the use of relative mid-treatment imaging values has the advantage of measuring the change during treatment and hence sensitivity to prescribed treatment. Currently, there is no consensus on the optimal method for analysis of mid-treatment imaging biomarkers, and this lack of standardisation presents a barrier to the integration of mid-treatment imaging biomarkers into clinical decision-making.\u003c/p\u003e\u003cp\u003eThe aim of this systematic review is to summarise the existing literature on the prognostic value of mid-treatment imaging during radiotherapy for mucosal HNSCC using locoregional recurrence as the primary clinical outcome. Additionally, we aim to identify the optimal method for analysis of mid-treatment imaging parameters for treatment response assessment.\u003c/p\u003e"},{"header":"METHODS","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003cp\u003eA prospectively registered protocol is available through the International Prospective Register of Systematic Reviews (PROSPERO) (CRD42023484250). The review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.[\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eSearch strategy and study selection\u003c/h3\u003e\n\u003cp\u003eA systematic search of PubMed, EMBASE, Scopus, and the Cochrane Central Register of Controlled Trials (CCTR) was conducted to identify relevant published studies between 2005 and 2023 (see Appendix, Supplementary Material). The PICOS framework (patient, intervention, comparison, outcome, and study) was used to develop the review question, focusing on the predictive value of quantitative mid-treatment imaging for locoregional treatment response in head and neck cancer treated with definitive radiotherapy. As outlined in the pre-specified protocol, exclusion criteria included: (1) studies without quantitative image analysis; (2) studies published prior to 2005; (3) non-squamous cell carcinoma (non-SCC) histology; (4) primary cutaneous or thyroid malignancies; and (5) studies with fewer than 10 patients. Additionally, studies primarily involving nasopharyngeal carcinoma (NPC) were excluded due to the distinct biological and clinical behaviour of NPC. Articles involving patients treated with adjuvant (post-operative) radiotherapy were also excluded.\u003c/p\u003e\u003cp\u003e Article screening, selection and data extraction were performed using the Covidence\u0026trade; web-based systematic review platform. Two reviewers (YT, ML) independently reviewed titles and abstracts, with disagreements resolved by consensus or consultation with a third reviewer (MJ). In cases of overlapping patient populations, the study with the largest cohort was included.\u003c/p\u003e\n\u003ch3\u003eData extraction\u003c/h3\u003e\n\u003cp\u003eThe primary outcome for this review was locoregional disease control following completion of radiotherapy, considered the gold standard endpoint for treatment efficacy in this population.\u003c/p\u003e\u003cp\u003eData on study characteristics (author, design, year, eligible patients, median follow-up), patient demographics (age, primary sub-site, tumour stage), treatment details (radiotherapy dose, concurrent chemotherapy), imaging protocol (modality, timing, region of interest, segmentation method), clinical outcomes and imaging parameters. Data were extracted by two independent reviewers and where information was missing, assumptions were not made and data were excluded from synthesis.\u003c/p\u003e\u003cp\u003eImaging parameters were classified into types of features; global (i.e. volume, max, mean, median), histogram, textural and parametric map. Time of mid-treatment imaging was noted in weeks and further classified as early (week 1 to 3) or late (week 4 to 7).\u003c/p\u003e\n\u003ch3\u003eQuality assessment\u003c/h3\u003e\n\u003cp\u003eWe assessed the risk of bias and applicability of the eligible studies using the Quality Assessment of Diagnostic Accuracy studies (QUADAS-2) tool.[\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e] This tool assesses four key domains: patient selection, index test, reference standard, and flow and timing, each of which is evaluated for risk of bias, while the first three are also assessed for concerns regarding applicability. Patient selection examines how participants were recruited and whether selection methods could introduce bias. The index test domain evaluates how the imaging was conducted and interpreted, including whether blinding or pre-specified thresholds were used. The reference standard assesses the reliability and appropriateness of the outcome measure used to determine true disease status. Flow and timing considers whether all patients received the same reference standard and whether there were any delays or losses to follow-up that could affect the results. Perceived quality were graded as low, high or unclear risk.\u003c/p\u003e\n\u003ch3\u003eData synthesis and Meta-analysis\u003c/h3\u003e\n\u003cp\u003eImaging parameters extracted from baseline and mid-treatment imaging including absolute and/or delta-values (Δ, relative change in values acquired from imaging before and during treatment) were included in data synthesis. Imaging parameters were grouped into three categories for analysis: (1) baseline; (2) absolute mid-treatment; and (3) relative mid-treatment (delta) values.\u003c/p\u003e\u003cp\u003eFor the meta-analysis, studies that performed receiver operating characteristics (ROC) analysis were selected. Imaging parameters were compared using the reported AUC values and standard errors from the ROC analysis in the MedCalc software. Heterogeneity across studies was quantified using the Higgins\u0026rsquo; I\u0026sup2; statistic (I\u0026sup2; \u0026lt;25%: negligible; 25\u0026ndash;50%: low; 50\u0026ndash;75%: moderate; \u0026gt;75%: high heterogeneity), which estimates the proportion of variation due to heterogeneity rather than chance. A random-effects model was employed to evaluate AUC values and compare subgroups of baseline, absolute mid-treatment, and relative mid-treatment imaging parameters. Results were tabulated and visualised using forest plots and funnel plots to display individual study outcomes and pooled estimates.\u003c/p\u003e"},{"header":"RESULTS","content":"\u003cdiv id=\"Sec9\"\u003e\n \u003ch2\u003eStudy selection\u003c/h2\u003e\n \u003cp\u003eA total of 1972 articles were identified through the search strategy. After removal of duplicates, 1,533 articles were screened. Full-text review was performed on 228 articles, resulting in 41 studies that met the inclusion and exclusion criteria. The study selection process and reasons for article exclusion are summarised in the PRISMA flowchart (Fig.\u0026nbsp;1). Common reasons for exclusion included lack of locoregional treatment outcomes, absence of mid-treatment imaging, lack of predictive or prognostic analysis, and review article format.\u003c/p\u003e\n\u003c/div\u003e\n\u003ch3\u003eStudy characteristics\u003c/h3\u003e\n\u003cp\u003eThe total study population consisted of 1654 patients, of whom 82% received chemoradiotherapy and 18% received radiotherapy alone. Median follow-up ranged from 12 to 57 months, though not reported in five studies. The majority of studies were prospectively conducted (83%, 34 of 41). Twenty-seven studies used one imaging modality[13\u0026ndash;39], 12 studies used two imaging modalities[40\u0026ndash;51], and two studies used three imaging modalities[52, 53]. When reporting mid-treatment parameters, 21 (51%) studies reported relative values, five (12%) studies reported absolute values and 15 (37%) studies reported both absolute and relative values. There was also methodological heterogeneity, including differences in ROI selection, ROI delineation methods and timing of mid-treatment imaging. See summary of included studies (Table\u0026nbsp;1a and 1b).\u003c/p\u003e\n\u003cp\u003eMRI imaging was utilised in 19 studies, PET imaging in 19 studies and CT imaging in 9 studies. DWI-MRI was the most common mid-treatment imaging modality reported (15 studies), followed by FDG-PET (12 studies).(Table 2)\u003c/p\u003e\n\u003ch2\u003eStudy quality\u003c/h2\u003e\n\u003cp\u003eStudy quality was assessed using the QUADAS-2 tool. The risk of bias was low across the four domains with 85%, 76%, 83% and 85% of studies having low risk in the patient selection, index test, reference standard and timing domains, respectively.(Fig. 2) Overall, 14 out of 41 studies (34%) demonstrated high risk of bias in at least one of the domains. Reference standard domain posed the highest risk of bias (17%, 7 of 41 studies).\u003c/p\u003e\n\u003cp\u003eConcern regarding applicability was generally low for most studies with 61%, 83% and 90% of studies having low concerns of applicability in the patient selection, index test and reference standard domains, respectively.(Fig. 2) Overall, 16 out of 41 studies (39%) demonstrated high concerns regarding applicability in at least one of the domains. Patient selection domain posed the highest concern regarding applicability (24%, 10 of 41 studies).\u003c/p\u003e\n\u003cdiv id=\"Sec12\"\u003e\n \u003ch2\u003eCT studies\u003c/h2\u003e\n \u003cp\u003eCT was used in nine studies involving 357 patients.(Table 2) Anatomical CT was the most common modality analysed (6 of 9 studies)[22, 28, 32, 40, 42, 49], with all studies employing manual ROI segmentation. Global parameters, particularly tumour volume, were most frequently analysed (6 of 7 studies). A range of relative change in tumour volumes were noted on mid-treatment imaging (median 19\u0026ndash;68%) due to methodological variability in choice of ROI and timing of mid-treatment imaging. Three studies reported statistically significant correlations[22, 28, 42], and three studies did not find any correlation between mid-treatment CT volume and clinical outcomes[32, 40, 49]. The largest CT study by Kabarriti et al. (n\u0026thinsp;=\u0026thinsp;96 patients) found a relative change in CT tumour volume (cut-off \u0026Delta;19%, HR 0.26) but not absolute CT tumour volume at week 3 correlated to locoregional outcomes.[22] Similarly, Lee et al. reported that both absolute volume (11 cc, AUC 0.847) and relative change in volume (\u0026Delta;75%, AUC 0.725) at week 4 mid-treatment CT were predictive for locoregional outcomes.[28] Conversely, studies by Arens et al, Mishra et al and Trada et al did not find any correlation between change in CT based tumour volume and clinical outcomes.[32, 40, 49] Morgan et al. conducted a matched case-control study (n\u0026thinsp;=\u0026thinsp;90 patients) using textural analysis and found that relative change in a cluster of mid-treatment CT features were associated with locoregional recurrence.[34]\u003c/p\u003e\n \u003cp\u003eThree studies used quantitative contrast enhanced CT[13, 37, 38], all employing manual ROI segmentation and primarily analysing global features (2 of 3 studies). Studies by Troung et al. and Ursino et al. reported multiple mid-treatment perfusion characteristics. Troung et al. (n\u0026thinsp;=\u0026thinsp;15 patients) found that both absolute value (136 vs 44 ml/100g/min) and relative change (\u0026Delta;, 28% vs 18%) in mid-treatment blood flow measurements at week 2 were higher in patients with tumour control.[37] Ursino et al. (n\u0026thinsp;=\u0026thinsp;25 patients) found relative change in blood volume, blood flow and permeability surface measurements at week 3 correlated to early FDG-PET tumour response.[38] Abramyuk et al. (n\u0026thinsp;=\u0026thinsp;15 patients) measured multiple perfusion characteristics at week 2 and week 5 during radiotherapy but did not find correlation to tumour outcomes.[13]\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec13\"\u003e\n \u003ch2\u003ePET studies\u003c/h2\u003e\n \u003cp\u003ePET was utilised in 19 studies containing 756 patients. (Table 2) FDG-PET was the most frequent PET tracer based imaging analysed (12 of 19 studies).[15, 17, 19, 25, 29, 30, 42, 46, 48, 49, 52, 53] Other novel hypoxia PET tracers based imaging utilised included F-MISO (2 studies)[27, 39], FHX4 (1 study)[35], and FETNIM (1 study)[21]. Novel tumour proliferation PET tracers utilised included FLT PET (2 studies) and 4D-ST PET (1 study)[46]. An advantage of PET imaging is relative ease of implementation of semi-automated segmentation methods for ROI delineation which was used in majority of PET studies (12 of 19). Global features such as maximum voxel value (SUV\u003csub\u003emax\u003c/sub\u003e), metabolic tumour volume (MTV) and total lesional glycolysis (TLG) were the commonly utilised (19 studies). Textural features were analysed in only 1 study.[25]\u003c/p\u003e\n \u003cp\u003eFDG-PET was the most utilised PET tracer due to its relative ease of access from its current clinical use in diagnosis, staging and post treatment monitoring in management of head and neck malignancies. The majority of FDG-PET studies performed imaging early (week 1\u0026ndash;3) during radiotherapy (9 of 12 studies). Overall, there were nine FDG studies[19, 25, 29, 30, 42, 48, 49, 52, 53] that found a correlation between mid-treatment imaging features and clinical outcomes and three studies[15, 17, 46] which did not. A study by Min et al, utilised week 3 FDG PET in 75 patients during radiotherapy and analysed a range of absolute and relative mid-treatment global features from primary tumours.[30] They found absolute mid-treatment values of SUV\u003csub\u003emax\u003c/sub\u003e (cut-off value 4.3g/mL, 2 year LRRFS 89% vs 75%, HR 3.7, p\u0026thinsp;=\u0026thinsp;0.03), MTV (cut-off value 3.3cm\u003csup\u003e3\u003c/sup\u003e, 2\u0026nbsp;year LRRFS 90% vs 69%, HR 3.9, p\u0026thinsp;=\u0026thinsp;0.02) and TLG (cut off value 9.4g, 2\u0026nbsp;year LRRFS 93% vs 71%, HR 6.3, p\u0026thinsp;\u0026lt;\u0026thinsp;0.01) correlated to locoregional recurrence. Of the studies that compared the performance of multiple mid-treatment global features, relative change in MTV and its closely related TLG were frequently found to be the best predictor of tumour control (Lin et al., \u0026Delta;MTV 50%, HR 0.11, Trada et al, \u0026Delta;MTV AUC 0.833; Trada et al., \u0026Delta;MTV AUC 0.761; Wong et al. \u0026Delta;TLG 0.825).[29, 48, 49, 53] Lafata et al. performed textural analysis from week 2 FDG-PET in 64 patients with oropharyngeal malignancies.[25] They utilised an in-house software to analyse multiple radiomic features and unsupervised learning to identify three intra-treatment imaging clusters. They found a difference between the intra-treatment clusters and tumour recurrence (HR 10.5, p\u0026thinsp;\u0026lt;\u0026thinsp;0.01).\u003c/p\u003e\n \u003cp\u003eFour studies utilised hypoxia PET tracers and analysed global imaging parameters such as SUV\u003csub\u003emax\u003c/sub\u003e or hypoxic volume for treatment response prediction, with moderate success. Two studies employed first generation hypoxia tracer, F-MISO PET to correlate changes in hypoxia during radiotherapy with tumour recurrence.[27, 39] Lazzeroni et al. utilised a volumetric measure of hypoxia sub-volume (%HTV) and found that a relative change (AUC 0.75, cut-off value\u0026thinsp;\u0026lt;\u0026thinsp;44.3%), but not absolute HTV value at week 2 correlated to locoregional recurrence.[27] Wiedenman et al. utilised a modified hypoxia SUV\u003csub\u003emax\u003c/sub\u003e value (TBR\u003csub\u003emax\u003c/sub\u003e) and found that patients with local tumour recurrence had higher absolute TBR\u003csub\u003emax\u003c/sub\u003e values at week 2 (p\u0026thinsp;=\u0026thinsp;0.031) but not week 5 (p\u0026thinsp;=\u0026thinsp;0.071).[39] Alternate hydrophilic hypoxia PET tracers such as FHX4 and FETNIM with lower background tissue uptake were developed to improve the tumour to background signal and improve clearance with favourable pharmacokinetic properties. A study by Sanduleanu et al., involving 34 patients, found that relative residual hypoxic volume at week 2 on FHX4 PET correlated to local control (p\u0026thinsp;=\u0026thinsp;0.028).[35] Hu et al. reported a decrease in SUV\u003csub\u003emax\u003c/sub\u003e at week 5 during treatment on FETNIM PET, but no significant correlation to local tumour control.[21]\u003c/p\u003e\n \u003cp\u003eTwo studies analysed FLT-PET to investigate the correlation between tumour proliferation and treatment response.[20, 40] A study by Hoeben et al. found a decrease in tumour volume and SUVmax during radiotherapy.[20] They found relative change in proliferative tumour volume (cut-off value\u0026thinsp;\u0026gt;\u0026thinsp;78%, 3\u0026nbsp;year LRC 100% vs 68%, p\u0026thinsp;=\u0026thinsp;0.021) at week 4 correlated to locoregional outcomes. They did not find a correlation between relative change in SUVmax to outcomes. A study by Arens et al. employed variety of semi-automated methods for measuring proliferative tumour volumes but found no correlation to locoregional outcomes at either week 2 or week 4 during radiotherapy.[40] A study by Mitamura et al. utilised 4D-ST, a cell proliferation tracer that is resistant to thymidine phosphorylase degradation, and performed mid-treatment imaging at week 4 to calculate SUV\u003csub\u003emax\u003c/sub\u003e, proliferation tumour volume (PTV) and total lesional proliferation (TLP).[46] They found change in tumour volume (\u0026Delta;PTV AUC 0.91, cut-off value \u0026minus;\u0026thinsp;90%), total proliferation (\u0026Delta;TLP AUC 0.89, cut-off value \u0026minus;\u0026thinsp;45%) and maximum voxel volume (\u0026Delta;SUV\u003csub\u003emax\u003c/sub\u003e AUC 0.75, cut-off value \u0026minus;\u0026thinsp;99%) discriminated early treatment response.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec14\"\u003e\n \u003ch2\u003eMRI studies\u003c/h2\u003e\n \u003cp\u003eMRI was utilised in 19 studies containing 771 patients (Table 2). DWI MRI was the most common modality analysed (15 of 19 studies).[16, 18, 23, 26, 31, 33, 36, 43\u0026ndash;45, 47, 48, 50, 52, 53] Other MRI modalities utilised included anatomical MRI (7 studies)[14, 41, 43\u0026ndash;45, 47, 51], DCE-MRI (4 studies),[41, 51\u0026ndash;53] and MR spectroscopy (1 study).[24] All 19 MRI studies utilised manual segmentation for ROI delineation. Global imaging features were again the most commonly parameters utilised (13 studies); however, several studies implemented histogram analysis (5 studies), textural features (2 studies) and parametric maps (1 studies).\u003c/p\u003e\n \u003cp\u003eAll seven anatomical MRI imaging utilised tumour volume, while one study also utilised textural features for response assessment.[47] There were four studies that found statistically significant correlations[14, 43, 45, 47], and three studies that did not find any correlation between mid-treatment MRI volume and clinical outcomes.[41, 44, 51] Study by Bhatia et al. correlated primary tumour volume delineated on T1 MRI during week 2 of radiotherapy to clinical outcomes.[14] They found that absolute change in tumour volume (AUC 0.721, cut-off \u0026gt;\u0026thinsp;10.6cc) had stronger correlation compared to relative change in tumour volume (AUC 0.689, cut-off \u0026lt;\u0026thinsp;9.4%) to local failure. A similar study by Khattab et al. also found that absolute primary tumour volume during week 2\u0026ndash;3 of radiotherapy correlated to local failure (median 2.5cc vs 14.9cc, p\u0026thinsp;=\u0026thinsp;0.007).[43] They found that relative change in tumour volume did not reach statistical significance for local failure in their patient population (median 73% vs 35%, p\u0026thinsp;=\u0026thinsp;0.100). In contrast, the largest series by Cao et al., containing 54 patients did not find a correlation between absolute or relative mid-treatment anatomical tumour volume on T1 MRI imaging and tumour control.[41] Study by Scalco et al. utilised higher order radiomics and found relative change in week 3 mid-treatment textural features (Variance, Accuracy 68%; Fractional dimension, Accuracy 64%) on T2 imaging correlated to tumour recurrence.[47]\u003c/p\u003e\n \u003cp\u003eThe majority of studies utilising DWI MRI imaging analysed global features (13 of 15 studies), predominantly mean apparent diffusion coefficient value (ADC\u003csub\u003emean\u003c/sub\u003e). Tumour volume based on ADC maps was only evaluated in three studies with no correlation found with locoregional clinical outcomes.[18, 50, 52] The largest DWI series containing 81 patients by Mohamed et al., found relative change in primary tumour ADCmean at week 3 had the strongest correlation to local tumour control (cut-off \u0026lt;\u0026thinsp;7%, 2\u0026nbsp;year LC 28% vs 96%).[33] A study by Trada et al. analysed primary tumour ADC\u003csub\u003emean\u003c/sub\u003e at multiple time-points and found that relative change but not absolute mid-treatment ADC\u003csub\u003emean\u003c/sub\u003e correlated to local tumour control.[48] Relative change in primary tumour ADC\u003csub\u003emean\u003c/sub\u003e is a promising imaging biomarker for treatment response prediction. Relative change in ADC\u003csub\u003emean\u003c/sub\u003e with optimal cut-off (OC) values ranging from 16% to 33% (\u0026Delta;ADC\u003csub\u003emean\u003c/sub\u003e) was consistently found to have strong correlation to clinical outcomes in studies by Khattab et al. (AUC 0.786, OC 33%), Kim et al. (AUC 0.88), Marzi et al. (AUC 0.662, OC 16%), Matoba et al. (AUC 0.900, OC 24%), Trada et al. (AUC 0.825, OC 24%), Vandecaveye et al. (AUC 0.97, OC 25%), and Wong et al. (AUC 0.937, OC 17%).[31, 43\u0026ndash;45, 48, 50, 53] Due to the image resolution of ADC maps derived from DWI MRI it is possible to derive histogram and higher order imaging features to quantitatively assess microenvironment and physiological processes for tumour response prediction. Histogram analysis was utlised in four studies[16, 23, 43, 52], higher order textural features were utilised in two studies[36, 47], and parametric maps were analysed in one study.[18] A study by Tomita et al. demonstrated the feasibility of applying artificial intelligence, using deep learning of mid-treatment DWI maps for imaging feature extraction, which correlated with local tumour recurrence (AUC 0.767).[36]\u003c/p\u003e\n \u003cp\u003eIntravoxel incoherent motion diffusion-weighted imaging (IVIM-DWI) MRI can be used to assess both water diffusion and microcirculation (perfusion) within tissues. It leverages the principle of DWI, but uses lower b-values to specifically isolate and quantify perfusion information. Two studies utlised IVIM for tumour response prediction. Study by Marzi et al. found that absolute mid-treatment diffusion co-efficient (D, AUC 0.75), perfusion co-efficient (D*, AUC 0.688) and perfusion fraction (f, AUC 0.829) correlated to nodal recurrence. Whereas Martens et al., did not find statistically significant correlations between these imaging features and clinical outcomes.\u003c/p\u003e\n \u003cp\u003eDCE-MRI is a complex MRI technique that analyses tissue enhancement patterns after the injection of a contrast agent to assess blood flow, vessel permeability, and tissue volume fractions. It provides functional information about tissue perfusion and microcirculation, particularly in tumours that can be used for treatment response assessment. DCE-MRI was analysed in four studies with moderate success utilising mid-treatment imaging features such as K\u003csub\u003etrans\u003c/sub\u003e (volume transfer constant), V\u003csub\u003ee\u003c/sub\u003e (fractional volume of extravascular extracellular space), K\u003csub\u003eep\u003c/sub\u003e (transfer rate of contrast agent from extravascular extracellular space to plasma.[41, 51\u0026ndash;53] A study by Wong et al. found relative change in K\u003csub\u003etrans\u003c/sub\u003e at week 2 had the highest correlation to clinical response (AUC 0.813).[53] Similarly a study by Martens et al., found relative change in K\u003csub\u003eep\u003c/sub\u003e, and absolute mid-treatment V\u003csub\u003ee\u003c/sub\u003e, K\u003csub\u003etrans\u003c/sub\u003e and K\u003csub\u003eep\u003c/sub\u003e correlated to tumour outcomes.[52]\u003c/p\u003e\n \u003cp\u003eMagnetic Resonance Spectroscopy (MRS) is a technique used alongside MRI to analyse the chemical composition of tissues. MRS focuses on identifying and quantifying specific metabolites within a selected region, studied most commonly in brain tumours. King et al. utlised a novel MRS technique to measure changes in mid-treatment tumoral metabolite Choline in patients with head and neck cancer. They detected changes in Choline, Choline:Creatinine and Choline:Water ratios during week 2 of radiotherapy but these did not correlate with clinical outcomes in their patient population.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec15\"\u003e\n \u003ch2\u003eMeta-analysis\u003c/h2\u003e\n \u003cp\u003eSeventeen studies containing 612 patients reported ROC analysis and were included in the meta-analysis. A total of 63 imaging parameters were evaluated: 17 baseline parameters, and 46 mid-treatment parameters (37 relative; 9 absolute).\u003c/p\u003e\n \u003cp\u003eUsing a random-effects model, the pooled AUC for baseline parameters was 0.736 (95% CI: 0.688\u0026ndash;0.785). Relative mid-treatment parameters showed improved performance (AUC 0.796; 95% CI: 0.762\u0026ndash;0.831) compared to absolute mid-treatment parameters (AUC 0.686; 95% CI: 0.628\u0026ndash;0.745) (Fig.\u0026nbsp;3, Supplementary Table\u0026nbsp;1).\u003c/p\u003e\n \u003cp\u003eA test of heterogeneity was low across all parameters included in the meta-analysis (I\u0026sup2; = 46.8%) (Fig. 4). When stratified, baseline features showed moderate heterogeneity (I\u0026sup2; = 51.5%), while absolute (I\u0026sup2; = 34.6%) and relative (I\u0026sup2; = 47.9%) mid-treatment parameters demonstrated low heterogeneity (Supplementary Fig. 1). Heterogeneity assessed using the Higgins\u0026rsquo; I\u0026sup2; statistic (I\u0026sup2; \u0026lt;25%: negligible; 25\u0026ndash;50%: low; 50\u0026ndash;75%: moderate; \u0026gt;75%: high heterogeneity).\u003c/p\u003e\n\u003c/div\u003e"},{"header":"DISCUSSION","content":"\u003cp\u003eThis systematic review and meta-analysis evaluated the prognostic value of quantitative mid-treatment imaging biomarkers in predicting locoregional tumour control in patients with HNSCC undergoing definitive radiotherapy. Our findings affirm the potential of mid-treatment imaging as a clinically relevant early-response biomarker. Across 41 eligible studies, mid-treatment imaging particularly relative changes (Δ) from baseline parameters were consistently associated with improved prediction of treatment outcomes. The meta-analysis of 17 studies involving 612 patients demonstrated a pooled AUC of 0.796 for relative mid-treatment parameters, compared with 0.736 for baseline and 0.686 for absolute mid-treatment values, suggesting that dynamic biological changes captured during therapy provide stronger prognostic insight than static measurements. Among imaging modalities, DWI-MRI and FDG-PET emerged as the most frequently studied and predictive. However, significant heterogeneity exists across studies in terms of imaging acquisition, timing, parameter extraction, and outcome reporting, which currently limits clinical implementation.\u003c/p\u003e\u003cp\u003eMRI particularly diffusion-weighted MRI (DWI), was the most studied and consistently predictive for tumour outcomes. Across multiple studies, relative change in ADC\u003csub\u003emean\u003c/sub\u003e (ΔADC\u003csub\u003emean\u003c/sub\u003e) emerged as a robust imaging biomarker, with optimal cut-offs ranging from 16% to 33% and AUC values as high as 0.97. Seven DWI studies by Khattab et al. (AUC 0.786), Kim et al. AUC 0.88), Marzi et al. (AUC 0.662), Matoba et al. (AUC 0.900), Trada et al. (AUC 0.825), Vandecaveye et al. (AUC 0.97), and Wong et al. (AUC 0.937) demonstrated strong correlation with locoregional tumour control.[\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e, \u003cspan additionalcitationids=\"CR44\" citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e, \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e, \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e, \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e] These consistent findings reflect the ability of ΔADC\u003csub\u003emean\u003c/sub\u003e to quantify early cytotoxic effects of radiotherapy, such as reduced tumour cellularity and increased water diffusion due to processes such as progressive necrosis, with successful treatment. A significant challenge in applying ADC-based biomarkers across centres is the variability introduced by differences in scanner hardware, acquisition protocols, and b-value selection, which can lead to inconsistent absolute ADC values.(reference) This variability limits the generalisability of absolute ADC thresholds and complicates multi-institutional standardisation. In contrast, relative mid-treatment changes in ADC (i.e., ΔADC\u003csub\u003emean\u003c/sub\u003e) performed on the same scanner offer a degree of self-normalisation, as each patient effectively serves as their own internal control. This inherent normalisation reduces variability and enhances reproducibility, making relative ADC metrics more suitable for wider clinical adoption. Moreover, DWI-based histogram and textural analyses were applied in several studies (e.g., Tomita et al., Scalco et al.), suggesting an evolving role for radiomic and AI-based approaches to further refine predictive performance.[\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e, \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e]\u003c/p\u003e\u003cp\u003ePET imaging, particularly FDG-PET, also demonstrated strong predictive value, with relative volume based parameters ΔMTV and ΔTLG frequently emerging as the most informative features. Several studies (e.g., Lin et al., Trada et al., Wong et al.) reported ΔMTV cut-offs ranging from 33% to 75% to be predictive of locoregional control, with AUC values up to 0.833.[\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e, \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e, \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e] Biologically, a decrease in metabolic tumour volume during treatment likely reflects a combination of tumour cell death, reduced glucose metabolism, and improved oxygenation, and may serve as an early surrogate for radiosensitivity. Conversely, a lack of significant metabolic reduction may indicate radioresistant subvolumes, which could be targeted for dose escalation. One of the key advantages of FDG-PET is its relative widespread clinical availability and established role in diagnostic and post-treatment imaging, which has led to greater standardisation of acquisition protocols across scanners compared to other functional imaging modalities. FDG-PET also offers practical advantages in clinical implementation, including the use of semi-automated region-of-interest (ROI) segmentation, which reduces inter-observer variability and enhances reproducibility. These features, combined with the potential for integration with automated workflows, make FDG-PET a feasible platform for real-time response assessment and personalised adaptive radiotherapy treatment planning in routine practice.\u003c/p\u003e\u003cp\u003eEmerging novel PET tracers targeting hypoxia (FMISO, FHX4, FETNIM), and proliferation (FLT, 4D-ST) add physiological specificity. For instance, Sanduleanu et al. found that residual hypoxic volume on FHX4 PET at week 2 correlated with local control (p\u0026thinsp;=\u0026thinsp;0.028)[\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e], while Lazzeroni et al. found a ΔHTV threshold\u0026thinsp;\u0026lt;\u0026thinsp;44.3% on FMISO PET (AUC 0.75) was predictive of recurrence.[\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e] Similarly, Mitamura et al. demonstrated high predictive value for ΔPTV (AUC 0.91) and ΔTLP (AUC 0.89) using 4D-ST PET.[\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e] While these tracers are not yet widely adopted, they provide granular insight into treatment-resistant subvolumes, offering targets for future radiotherapy dose-painting strategies.\u003c/p\u003e\u003cp\u003eIn contrast, results from CT-based studies were more mixed. Among nine studies, six reported global volumetric changes with highly variable correlation to outcomes. Kabarriti et al. found a ΔCT volume of 19% at week 3 significantly predicted locoregional outcomes (HR 0.26)[\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e], while Lee et al. reported both absolute and ΔCT volumes at week 4 to be prognostic (AUCs 0.847 and 0.725, respectively).[\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e] Conversely, other studies (e.g., Arens et al., Mishra et al., and Trada et al.,) found no significant correlations.[\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e, \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e, \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e] The relatively poor soft tissue contrast of CT, dependence on manual ROI delineation with limited reproducibility, and lack of standardised acquisition protocols likely contribute to the inconsistent performance of CT-based biomarkers and limit their applicability in multicentre settings. However, CT texture analysis (Morgan et al.) and perfusion CT parameters (Truong et al., Ursino et al.) showed promise, suggesting that advanced feature extraction from routine imaging may salvage prognostic utility from standard CT.[\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e, \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e, \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e]\u003c/p\u003e\u003cp\u003eThis study represents the first meta-analysis to quantitatively synthesise ROC-based AUC values for predictive performance of mid-treatment imaging in HNSCC. The use of ROC-based AUC measures offers a robust and standardised approach to assess the discriminatory ability of imaging biomarkers, independent of specific thresholds or prevalence, making it particularly suitable for comparing performance across heterogeneous studies. In our pooled meta-analysis, the area under the curve (AUC) for baseline imaging parameters was 0.736 (95% CI: 0.688\u0026ndash;0.785). Relative mid-treatment parameters demonstrated numerically superior performance with an AUC of 0.796 (95% CI: 0.762\u0026ndash;0.831); however, the overlapping confidence intervals indicate this difference did not reach statistical significance. This result should be interpreted with caution given that our analysis was not conducted with the aim of comparing the utility of mid-treatment imaging parameters compared to their corresponding baseline imaging parameters. In contrast, relative mid-treatment parameters demonstrated significantly higher performance with an AUC of 0.796 (95% CI: 0.762\u0026ndash;0.831) compared to absolute mid-treatment parameters with an AUC of 0.686 (95% CI: 0.628\u0026ndash;0.745). While prior reviews (e.g., Martens et al. and Lin et al.) provided narrative synthesis, they did not directly compare the performance of baseline, absolute, and relative parameters.[\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e, \u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e] Our findings validate the rationale for delta imaging features, where temporal changes in imaging biomarkers during radiotherapy may reflect the biological responsiveness of individual tumours. This dynamic assessment enables early differentiation between radiosensitive and radioresistant disease, providing a surrogate marker of radiotherapy efficacy that can be projected to eventual clinical outcomes following completion of treatment. Furthermore, this review highlights the increasing incorporation of multi-parametric, multi-modality models, where features from DWI, PET, and CT can be combined to improve prognostic accuracy.\u003c/p\u003e\u003cp\u003eMid-treatment imaging has clear potential for personalised treatment adaptation. For responders, de-intensification strategies could mitigate long-term toxicities, while for non-responders, early escalation or surgical salvage may be indicated. Mid-treatment parameters may also assist in adaptive radiotherapy planning, identifying tumour sub-volumes for radiotherapy dose boost or omitting high-dose to regressing areas. Importantly, relative mid-treatment parameters may offer inherent self-normalisation, minimising inter-patient variability and scanner-related noise. Yet, challenges to clinical integration remain. These include a lack of standardised imaging protocols, diverse segmentation approaches, and variability in analytical pipelines. The reproducibility and interpretability of complex features, especially in multi-institution settings also limit generalisability. Thus, prioritising simple, reproducible metrics (e.g., ΔADC\u003csub\u003emean\u003c/sub\u003e, ΔMTV) may be a pragmatic step toward clinical translation.\u003c/p\u003e\u003cp\u003eSeveral limitations must be acknowledged. First, individual study sample sizes were relatively small, and most lacked external validation. Only a subset (17 of 41) reported ROC-based AUC values, limiting the meta-analysis scope. Second, publication and selection biases likely inflated performance metrics. Studies that did not find significant correlations were less likely to report full statistical metrics or be published at all. Additionally, we excluded nasopharyngeal carcinoma and adjuvant radiotherapy studies due to differing biology and treatment protocols, limiting generalisability. Finally, there was substantial heterogeneity in imaging acquisition, time-points (week 1\u0026ndash;7), and ROI methods, which restricts pooled interpretation. Future studies should pursue large, prospective multi-centre trials with harmonised imaging protocols, clearly defined mid-treatment timepoints, and publicly available datasets for external validation. Furthermore, multi-modality imaging platforms such as PET/MRI may unlock new opportunities for combining anatomical, functional, and molecular imaging to guide adaptive radiotherapy.\u003c/p\u003e"},{"header":"CONCLUSIONS","content":"\u003cp\u003eMid-treatment quantitative imaging offers prognostic value in predicting locoregional control in mucosal head and neck squamous cell carcinoma treated with definitive radiotherapy. This systematic review and meta-analysis demonstrate that relative mid-treatment imaging parameters outperform both baseline and absolute mid-treatment measures, with DWI-MRI (ΔADC\u003csub\u003emean\u003c/sub\u003e) and FDG-PET (ΔMTV, ΔTLG) emerging as the most consistent and clinically promising biomarkers.\u003c/p\u003e\u003cp\u003eDespite encouraging results, substantial methodological heterogeneity and lack of standardised imaging protocols currently limit clinical translation. Future research should focus on large, prospective, multi-centre studies with harmonised imaging protocols, robust feature extraction pipelines, and external validation of predictive models. Integration of multi-parametric imaging, radiomics, and AI-based tools into clinical workflows may be required to realise the potential of mid-treatment imaging as a decision-support tool for adaptive radiotherapy in head and neck cancer.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003eEthics approval and consent to participate: N/A\u003c/p\u003e\n\u003cp\u003eConsent for publication: N/A\u003c/p\u003e\n\u003cp\u003eAvailability of data and material: The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.\u003c/p\u003e\n\u003cp\u003eCompeting interests: Nil\u003c/p\u003e\n\u003cp\u003eFunding: Nil\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAuthors' contributions: YT/ML/MJ contributed to data collection and systematic review. AF/PK/DM contributed to research question formation, study design and major review of manuscript. YT was responsible for statistical analysis with notable external contributions. All authors read and approved the final manuscript\u003c/p\u003e\n\u003cp\u003eAcknowledgements: \u0026nbsp;The authors acknowledge Helen Ball from Image X institute, University of Sydney for assistance and guidance for statistical analysis during preparation of the manuscript.\u0026nbsp;\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eSiegel RL, Miller KD, Jemal A. \u003cem\u003eCancer statistics\u003c/em\u003e, 2016. CA Cancer J Clin, 2016. 66(1): pp. 7\u0026ndash;30.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLee NY, et al. Avelumab plus standard-of-care chemoradiotherapy versus chemoradiotherapy alone in patients with locally advanced squamous cell carcinoma of the head and neck: a randomised, double-blind, placebo-controlled, multicentre, phase 3 trial. Lancet Oncol. 2021;22(4):450\u0026ndash;62.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLacas B, et al. Meta-analysis of chemotherapy in head and neck cancer (MACH-NC): An update on 107 randomized trials and 19,805 patients, on behalf of MACH-NC Group. Radiother Oncol. 2021;156:281\u0026ndash;93.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMeijer TW, et al. Targeting hypoxia, HIF-1, and tumor glucose metabolism to improve radiotherapy efficacy. Clin Cancer Res. 2012;18(20):5585\u0026ndash;94.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCegla P, et al. Assessment of biological parameters in head and neck cancer based on in vivo distribution of (18)F-FDG-FLT-FMISO-PET/CT images. Tumori. 2020;106(1):33\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZhu A, et al. Application of metabolic PET imaging in radiation oncology. Radiat Res. 2012;177(4):436\u0026ndash;48.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBos P, et al. Prognostic functional MR imaging parameters in head and neck squamous cell carcinoma: A systematic review. Eur J Radiol. 2021;144:109952.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMin M, et al. A review of the predictive role of functional imaging in patients with mucosal primary head and neck cancer treated with radiation therapy. J Med Imaging Radiat Oncol. 2017;61(1):99\u0026ndash;123.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGaribaldi C, et al. Interim (18)F-FDG PET/CT During Chemoradiation Therapy in the Management of Head and Neck Cancer Patients: A Systematic Review. Int J Radiat Oncol Biol Phys. 2017;98(3):555\u0026ndash;73.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLin P, et al. Prognostic and predictive values of baseline and mid-treatment FDG-PET in oropharyngeal carcinoma treated with primary definitive (chemo)radiation and impact of HPV status: Review of current literature and emerging roles. Radiother Oncol. 2023;184:109686.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePage MJ, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWhiting PF, et al. QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies. Ann Intern Med. 2011;155(8):529\u0026ndash;36.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAbramyuk A, et al. Radiochemotherapy-induced changes of tumour vascularity and blood supply estimated by dynamic contrast-enhanced CT and fractal analysis in malignant head and neck tumours. Br J Radiol. 2015;88(1045):20140412.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBhatia KS, et al. Does primary tumour volumetry performed early in the course of definitive concomitant chemoradiotherapy for head and neck squamous cell carcinoma improve prediction of primary site outcome? Br J Radiol. 2010;83(995):964\u0026ndash;70.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCastaldi P, et al. Can early and late18F-FDG PET-CT be used as prognostic factors for the clinical outcome of patients with locally advanced head and neck cancer treated with radio-chemotherapy? Radiother Oncol. 2012;103(1):63\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eJoint H, et al. Longitudinal diffusion and volumetric kinetics of head and neck cancer magnetic resonance on a 1.5 T MR-linear accelerator hybrid system: A prospective R-IDEAL stage 2a imaging biomarker characterization/pre-qualification study. Clin Transl Radiat Oncol. 2023;42:100666.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFarrag A, et al. Can 18F-FDG-PET response during radiotherapy be used as a predictive factor for the outcome of head and neck cancer patients? Nucl Med Commun. 2010;31(6):495\u0026ndash;501.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGalban CJ, et al. A feasibility study of parametric response map analysis of diffusion-weighted magnetic resonance imaging scans of head and neck cancer patients for providing early detection of therapeutic efficacy. Transl Oncol. 2009;2(3):184\u0026ndash;90.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHentschel M, et al. Early FDG PET at 10 or 20 Gy under chemoradiotherapy is prognostic for locoregional control and overall survival in patients with head and neck cancer. Eur J Nucl Med Mol Imaging. 2011;38(7):1203\u0026ndash;11.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHoeben BA, et al. 18F-FLT PET during radiotherapy or chemoradiotherapy in head and neck squamous cell carcinoma is an early predictor of outcome. J Nucl Med. 2013;54(4):532\u0026ndash;40.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHu M, et al. Hypoxia with 18F-fluoroerythronitroimidazole integrated positron emission tomography and computed tomography (18F-FETNIM PET/CT) in locoregionally advanced head and neck cancer: Hypoxia changes during chemoradiotherapy and impact on clinical outcome. Med (Baltim). 2019;98(40):e17067.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKabarriti R, et al. Early Response Assessment on Mid-treatment Computed Tomography Predicts for Locoregional Recurrence in Oropharyngeal Cancer Patients Treated With Definitive Radiation Therapy. Int J Radiat Oncol Biol Phys. 2018;102(4):1036\u0026ndash;45.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKing AD, et al. Head and neck squamous cell carcinoma: diagnostic performance of diffusion-weighted MR imaging for the prediction of treatment response. Radiology. 2013;266(2):531\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKing AD, et al. Pretreatment and early intratreatment prediction of clinicopathologic response of head and neck cancer to chemoradiotherapy using 1H-MRS. J Magn Reson Imaging. 2010;32(1):199\u0026ndash;203.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLafata KJ, et al. Intrinsic radiomic expression patterns after 20 Gy demonstrate early metabolic response of oropharyngeal cancers. Med Phys. 2021;48(7):3767\u0026ndash;77.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLambrecht M, et al. Redefining the target early during treatment. Can we visualize regional differences within the target volume using sequential diffusion weighted MRI? Radiother Oncol. 2014;110(2):329\u0026ndash;34.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLazzeroni M, et al. Evolution of the hypoxic compartment on sequential oxygen partial pressure maps during radiochemotherapy in advanced head and neck cancer. Phys Imaging Radiat Oncol. 2021;17:100\u0026ndash;5.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLee YH, et al. Using primary tumor volumetry to predict treatment outcome for patients with oropharyngeal cancer who were treated with definitive chemoradiotherapy. Asia Pac J Clin Oncol. 2018;14(2):e21\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLin P, et al. Mid-treatment Fluorodeoxyglucose Positron Emission Tomography in Human Papillomavirus-related Oropharyngeal Squamous Cell Carcinoma Treated with Primary Radiotherapy: Nodal Metabolic Response Rate can Predict Treatment Outcomes. Clin Oncol (R Coll Radiol). 2021;33(12):e586\u0026ndash;98.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMin M, et al. Prognostic role of metabolic parameters of (18)F-FDG PET-CT scan performed during radiation therapy in locally advanced head and neck squamous cell carcinoma. Eur J Nucl Med Mol Imaging. 2015;42(13):1984\u0026ndash;94.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMarzi S, et al. The prediction of the treatment response of cervical nodes using intravoxel incoherent motion diffusion-weighted imaging. Eur J Radiol. 2017;92:93\u0026ndash;102.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMishra S, et al. Can radiological changes in lymph node volume during treatment predict success of radiation therapy in patients with locally advanced head and neck squamous cell carcinoma? J Med Imaging Radiat Oncol. 2013;57(5):603\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eJoint H, et al. Prospective validation of diffusion-weighted MRI as a biomarker of tumor response and oncologic outcomes in head and neck cancer: Results from an observational biomarker pre-qualification study. Radiother Oncol. 2023;183:109641.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMorgan HE, et al. Exploratory ensemble interpretable model for predicting local failure in head and neck cancer: the additive benefit of CT and intra-treatment cone-beam computed tomography features. Quant Imaging Med Surg. 2021;11(12):4781\u0026ndash;96.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSanduleanu S, et al. 18)F]-HX4 PET/CT hypoxia in patients with squamous cell carcinoma of the head and neck treated with chemoradiotherapy: Prognostic results from two prospective trials. Clin Transl Radiat Oncol. 2020;23:9\u0026ndash;15.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTomita H, et al. Deep learning approach of diffusion-weighted imaging as an outcome predictor in laryngeal and hypopharyngeal cancer patients with radiotherapy-related curative treatment: a preliminary study. Eur Radiol. 2022;32(8):5353\u0026ndash;61.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTruong MT, et al. Prediction of locoregional control in head and neck squamous cell carcinoma with serial CT perfusion during radiotherapy. AJNR Am J Neuroradiol. 2011;32(7):1195\u0026ndash;201.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eUrsino S, et al. Role of perfusion CT in the evaluation of functional primary tumour response after radiochemotherapy in head and neck cancer: preliminary findings. Br J Radiol. 2016;89(1065):20151070.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWiedenmann NE, et al. Serial [18F]-fluoromisonidazole PET during radiochemotherapy for locally advanced head and neck cancer and its correlation with outcome. Radiother Oncol. 2015;117(1):113\u0026ndash;7.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eArens AI, et al. Semiautomatic methods for segmentation of the proliferative tumour volume on sequential FLT PET/CT images in head and neck carcinomas and their relation to clinical outcome. Eur J Nucl Med Mol Imaging. 2014;41(5):915\u0026ndash;24.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCao Y, et al. Predictive Values of MRI and PET Derived Quantitative Parameters for Patterns of Failure in Both p16\u0026thinsp;+\u0026thinsp;and p16- High Risk Head and Neck Cancer. Front Oncol. 2019;9:1118.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChen SW, et al. Interim FDG PET/CT for predicting the outcome in patients with head and neck cancer. Laryngoscope. 2014;124(12):2732\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKhattab HM et al. Diffusion-weighted magnetic resonance imaging (DWMRI) of head and neck squamous cell carcinoma: could it be an imaging biomarker for prediction of response to chemoradiation therapy. Egypt J Radiol Nucl Med, 2020. 51(204).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKim S, et al. Diffusion-weighted magnetic resonance imaging for predicting and detecting early response to chemoradiation therapy of squamous cell carcinomas of the head and neck. Clin Cancer Res. 2009;15(3):986\u0026ndash;94.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMatoba M, et al. Fractional change in apparent diffusion coefficient as an imaging biomarker for predicting treatment response in head and neck cancer treated with chemoradiotherapy. AJNR Am J Neuroradiol. 2014;35(2):379\u0026ndash;85.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMitamura K, et al. Interim 4'-[methyl-(11)C]-thiothymidine PET for predicting the chemoradiotherapeutic response in head and neck squamous cell carcinoma: comparison with [(18)F]FDG PET. EJNMMI Res. 2021;11(1):13.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eScalco E, et al. Characterization of cervical lymph-nodes using a multi-parametric and multi-modal approach for an early prediction of tumor response to chemo-radiotherapy. Phys Med. 2016;32(12):1672\u0026ndash;80.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTrada Y, et al. Changes in serial multiparametric MRI and FDG-PET/CT functional imaging during radiation therapy can predict treatment response in patients with head and neck cancer. Eur Radiol. 2023;33(12):8788\u0026ndash;99.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTrada Y, et al. Impact of tumour region of interest delineation method for mid-treatment FDG-PET response prediction in head and neck squamous cell carcinoma undergoing radiotherapy. Quant Imaging Med Surg. 2023;13(5):2822\u0026ndash;36.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eVandecaveye V, et al. Predictive value of diffusion-weighted magnetic resonance imaging during chemoradiotherapy for head and neck squamous cell carcinoma. Eur Radiol. 2010;20(7):1703\u0026ndash;14.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWang P, et al. An approach to identify, from DCE MRI, significant subvolumes of tumors related to outcomes in advanced head-and-neck cancer. Med Phys. 2012;39(8):5277\u0026ndash;85.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMartens RM et al. \u003cem\u003eEarly Response Prediction of Multiparametric Functional MRI and (18)F-FDG-PET in Patients with Head and Neck Squamous Cell Carcinoma Treated with (Chemo)Radiation.\u003c/em\u003e Cancers (Basel), 2022. 14(1).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWong KH, et al. Changes in multimodality functional imaging parameters early during chemoradiation predict treatment response in patients with locally advanced head and neck cancer. Eur J Nucl Med Mol Imaging. 2018;45(5):759\u0026ndash;67.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"},{"header":"Tables","content":"\u003cp\u003eTables are available in the Supplementary Files section.\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"ejnmmi-research","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"ejre","sideBox":"Learn more about [EJNMMI Research](http://ejnmmires.springeropen.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/ejre/default.aspx","title":"EJNMMI Research","twitterHandle":"@officialEANM","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Head and Neck, Neoplasms, Systematic Review, Magnetic Resonance Imaging, Positron-Emission Tomography, Tomography, Recurrence, Response, Mid-treatment","lastPublishedDoi":"10.21203/rs.3.rs-7689582/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7689582/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground:\u003c/strong\u003e\u003cbr\u003e\n \u0026nbsp;To systematically review and meta-analyse the prognostic value of quantitative mid-treatment imaging biomarkers for predicting locoregional tumour control in patients undergoing definitive radiotherapy for mucosal head and neck squamous cell carcinoma.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMain body:\u003c/strong\u003e\u003cbr\u003e\n \u0026nbsp;A systematic literature search (2005–2023) was conducted in PubMed, EMBASE, Scopus, and Cochrane databases according to a pre-registered PROSPERO protocol. Studies evaluating quantitative imaging features derived from CT, MRI, or PET during radiotherapy were included. Imaging features were grouped as baseline, absolute mid-treatment, or relative mid-treatment (delta) parameters. A random-effects meta-analysis was performed on studies reporting receiver operating characteristic (ROC)-based area under the curve (AUC) values.\u003c/p\u003e\n\u003cp\u003eForty-one studies encompassing 1654 patients were included. Seventeen studies (n = 612 patients) reported sufficient data for meta-analysis. The pooled AUC for relative mid-treatment parameters was 0.796 (95% CI: 0.762–0.831), demonstrating higher predictive performance than absolute mid-treatment parameters (AUC 0.686; 95% CI: 0.628–0.745). Baseline parameters showed moderate predictive ability (AUC 0.736; 95% CI: 0.688–0.785), and while numerically lower than relative mid-treatment parameters, this difference did not reach statistical significance. Diffusion-weighted MRI (ΔADCmean) and FDG-PET (ΔMTV, ΔTLG) emerged as the most consistently predictive modalities. Relative measures offer practical advantages, including internal self-normalisation and improved reproducibility across imaging platforms.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusions:\u003c/strong\u003e\u003cbr\u003e\n Relative mid-treatment imaging biomarkers demonstrate superior predictive performance compared to baseline and absolute measures, supporting their potential role in adaptive radiotherapy strategies. Further prospective multi-centre studies with standardised imaging protocols and external validation are essential for clinical translation.\u003c/p\u003e","manuscriptTitle":"Quantitative Mid-treatment Imaging Biomarkers for Response Prediction After Radiotherapy in Head and Neck Cancer: A Systematic Review and Meta-analysis","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-10-30 01:06:42","doi":"10.21203/rs.3.rs-7689582/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"reviewerAgreed","content":"","date":"2026-03-28T01:30:02+00:00","index":0,"fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-10-15T09:27:03+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-09-29T01:35:37+00:00","index":"","fulltext":""},{"type":"submitted","content":"EJNMMI Research","date":"2025-09-25T22:06:27+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"ejnmmi-research","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"ejre","sideBox":"Learn more about [EJNMMI Research](http://ejnmmires.springeropen.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/ejre/default.aspx","title":"EJNMMI Research","twitterHandle":"@officialEANM","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"2d74c8e8-4634-4f5b-8d10-ed1aea50eb17","owner":[],"postedDate":"October 30th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-04-23T07:43:38+00:00","versionOfRecord":[],"versionCreatedAt":"2025-10-30 01:06:42","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7689582","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7689582","identity":"rs-7689582","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}