eFigure. PRISMA Flow Diagram for Systematic Review
Customize your JAMA Network experience by selecting one or more topics from the list below.
Identify all potential conflicts of interest that might be relevant to your comment.
Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.
Err on the side of full disclosure.
If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.
Not all submitted comments are published. Please see our commenting policy for details.
Creff G, Devillers A, Depeursinge A, et al. Evaluation of the Prognostic Value of FDG PET/CT Parameters for Patients With Surgically Treated Head and Neck Cancer: A Systematic Review. JAMA Otolaryngol Head Neck Surg. Published online March 26, 2020. doi:10.1001/jamaoto.2020.0014
Is there an association between parameters of fluorodeoxyglucose F 18–labeled positron emission tomography and computed tomography (FDG-PET/CT) and prognosis among patients with surgically treated head and neck cancer?
This systematic review of 36 studies that included 3585 unique patients highlights the prognostic value of FDG-PET/CT parameters, especially volumetric parameters, as biomarkers of overall survival, disease-free survival, and distant metastasis among patients with surgically treated head and neck cancer.
The findings of this review suggest that FDG-PET/CT parameters may be used to identify patients with a higher risk of postsurgical disease progression who could then receive early therapeutic intensification to improve their prognosis without increasing iatrogenic toxic effects; however, further research is required to define a reproducible, reliable FDG-PET/CT–based prognostic model of patients’ mortality and recurrence risks.
Head and neck squamous cell cancer (HNSCC) represents the seventh most frequent cancer worldwide. More than half of the patients diagnosed with HNSCC are treated with primary surgery.
To report the available evidence on the value of quantitative parameters of fluorodeoxyglucose F 18–labeled positron emission tomography and computed tomography (FDG-PET/CT) performed before surgical treatment of HNSCC to estimate overall survival (OS), disease-free survival (DFS), and distant metastasis (DM) and to discuss their limitations.
A systematic review of the English-language literature in PubMed/MEDLINE and ScienceDirect published between January 2003 and February 15, 2019, was performed between March 1 and July 27, 2019, to identify articles addressing the association between preoperative FDG-PET/CT parameters and oncological outcomes among patients with HNSCC. Articles included those that addressed the following: (1) cancer of the oral cavity, oropharynx, hypopharynx, or larynx; (2) surgically treated (primary or for salvage); (3) pretreatment FDG-PET/CT; (4) quantitative or semiquantitative evaluation of the FDG-PET/CT parameters; and (5) the association between the value of FDG-PET/CT parameters and clinical outcomes. Quality assessment was performed using the Oxford Centre for Evidence-Based Medicine level of evidence.
A total of 128 studies were retrieved from the databases, and 36 studies met the inclusion criteria; these studies comprised 3585 unique patients with a median follow-up of 30.6 months (range, 16-53 months). Of these 36 studies, 32 showed an association between at least 1 FDG-PET/CT parameter and oncological outcomes (OS, DFS, and DM). The FDG-PET/CT volumetric parameters (metabolic tumor volume [MTV] and total lesion glycolysis [TLG]) were independent prognostic factors in most of the data, with a higher prognostic value than the maximum standard uptake value (SUVmax). For example, in univariate analysis of OS, the SUVmax was correlated with OS in 5 of 11 studies, MTV in 11 of 12 studies, and TLG in 6 of 9 studies. The spatial distribution of metabolism via textural indices seemed promising, although that factor is currently poorly evaluated: only 3 studies analyzed data from radiomics indices.
Conclusions and Relevance
The findings of this study suggest that the prognostic effectiveness of FDG-PET/CT parameters as biomarkers of OS, DFS, and DM among patients with HNSCC treated with surgery may be valuable. The volumetric parameters (MTV and TLG) seemed relevant for identifying patients with a higher risk of postsurgical disease progression who could receive early therapeutic intervention to improve their prognosis. However, further large-scale studies including exclusively surgery-treated patients stratified according to localization and further analysis of the textural indices are required to define a reliable FDG-PET/CT–based prognostic model of mortality and recurrence risk for these patients.
Head and neck squamous cell cancer (HNSCC) represents the seventh most frequent cancer worldwide.1 The American Joint Committee on Cancer (AJCC) staging is generally used to estimate prognosis and to guide therapy.
Treatment of HNSCC requires a multidisciplinary approach and involves either surgery alone or surgery followed by radiotherapy, with or without chemotherapy, or curative radiotherapy with concomitant chemotherapy or anti–epidermal growth factor receptor agents (eg, cetuximab). The percentage of patients treated with surgical resection varies according to the HNSCC anatomic site. Cancer of the oral cavity has the highest percentage of primary surgical resections (71.4%), whereas hypopharyngeal cancer has the lowest proportion (42.1%).2
Despite the therapeutic progress (robotic-assisted surgery, new molecular targeted therapies, and improvement in the radiotherapy fields), the prognosis remains poor and is related to a high recurrence rate of 30% to 40%.3,4 Thus, there is a real need to identify patients at high risk of recurrence to improve their therapeutic strategy.
Fluorodeoxyglucose F 18–labeled positron emission tomography and computed tomography (FDG-PET/CT) allows us to reveal the metabolic activity of a tumor (glycolysis) in addition to anatomic information. This examination is now commonly used to assess the extent of HNSCC5 and for the posttreatment follow-up6 based on a visual analysis.
Visual analysis is sufficient for diagnosis, staging, and the detection of recurrence; however, quantification appears necessary to estimate patient outcome. The commonly studied parameters are the standardized uptake value7 (SUVmax, SUVmean, or SUVpeak), metabolic tumor volume (MTV), and total lesion glycolysis (TLG)8 (Table 1). In addition, owing to the emerging development of radiomics, which aims to decode the tumor phenotype and heterogeneity from imaging data,9,10 texture and shape analyses have started to be evaluated in terms of prognostic value.
Although the role of FDG-PET/CT in estimating survival of patients with HNSCC treated with radiochemotherapy has been examined in several studies,11-13 fewer data are available for patients treated with surgery who present with different clinical and histological profiles. To our knowledge, there has been no review in the scientific literature focusing on the prognostic value of FDG-PET/CT in patients with HNSCC who have been treated surgically.
Thus, the aims of this review were to report the available evidence on the value of quantitative parameters of FDG-PET/CT performed before surgical treatment of HNSCC to estimate overall survival (OS), disease-free survival (DFS), and distant metastasis (DM) and to discuss their limitations.
A systematic review was conducted between March 1 and July 27, 2019, according to the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) reporting guidelines.14
We performed a systematic electronic search of English-language articles reporting data obtained on humans and published between January 2003 and February 15, 2019, in the MEDLINE/PubMed database and ScienceDirect without any temporal restriction and using the terms positron emission tomography or PET or PET/CT and surgery or surgical treatment or tonsillectomy or salvage surgery or preoperative or laryngectomy and prognostic or prognosis or survival and head and neck cancer.
We excluded studies selecting patients receiving a form of radiochemotherapy as a unique treatment or studies including patients treated with surgery and patients treated with radiochemotherapy in which the treatment appeared as a prognostic factor in univariate analysis or if an analysis in the surgical treatment subgroup had not been performed. We also excluded meta-analyses, systematic reviews, letters to the editor, and data obtained only from visual analyses of the FDG-PET/CT images because this review focuses on the prognostic value of parameters obtained from quantitative or semiquantitative analyses. Studies that included fewer than 15 patients were also excluded. References from review articles were checked for cross-reference. Identical studies were identified to avoid duplications.
The PICO criteria (patients, intervention, comparison, and outcome) for the review are given in Table 2.
The selected studies were assessed for risk of bias on the basis of the following 5 variables:
Retrospective vs prospective study design
Sufficient description of the modality to obtain the FDG-PET/CT parameters
Uniform inclusion criteria
Incomplete outcome data
Number of patients included (studies with <15 patients were automatically excluded)
The risk of bias was classified as high if no was given for 3 or more criteria.
The level of evidence was scored according to the Oxford Centre for Evidence-Based Medicine level of evidence guide,15 as follows:
Systematic review of inception cohort studies
Inception cohort studies
Cohort study or control arm of randomized trial
Case series or case-control studies or poor-quality prognostic cohort study
Among the 128 articles initially identified, 36 relevant studies published between January 2003 and February 15, 2019, with exploitable data were retained for analysis after exclusion of duplicates (n = 11) and inappropriate articles (n = 81) (eFigure in the Supplement). No randomized clinical trial on this topic was found in the literature; most were retrospective series (27 studies). In total, 3585 unique patients were included in the analysis from these 36 studies,16-51 with a median follow-up of 30.6 months (range, 16-53 months) (Table 3).
Fourteen studies17-19,21,22,24,26-28,31,32,35,41,51 (826 unique patients) investigated the prognostic value of quantitative PET parameters on OS. Owing to its simplicity and interpretability, SUVmax was the first parameter analyzed. It was correlated with OS in 5 of 11 studies.18,19,22,24,27,28,31,32,35,41,51 In a prospective study including 53 patients with hypopharyngeal cancer,41 an SUVmax of 28.5 or higher (cutoff defined using the log-rank test) was associated with OS according to multivariate analysis (hazard ratio [HR], 3.94; 95% CI, 1.13-12.71) with adjustment for clinical T category and treatment group.
Recently, to take into account not only the value of 1 voxel but also the tumor volume, especially the heterogeneity, 2 other parameters were investigated: the MTV and TLG. In univariate analysis, the MTV was correlated with OS in 11 of 12 studies,17-19,22,26-28,31,32,35,41,51 and TLG was associated with OS in 6 of 9 studies.18,26-28,31,32,35,41,51 For instance, Kikuchi et al27 reported data from 47 patients with oropharyngeal cancer, of whom 38% were treated with surgery; the authors showed that MTV and TLG were associated with OS (MTV: HR, 4.8; TLG: HR, 3.9), whereas SUVmax was not associated with OS (HR, 1.1). Ten studies performed multivariate analysis.17,18,22,26-28,31,32,41,51 Both MTV and TLG remained correlated with OS in 6 studies17,18,22,26,27,31 and 3 studies,18,32,41 respectively. All of these studies showed that MTV and/or TLG had a prognostic effect, most of them with a higher predictive value than SUVmax18,19,22,27,28,31,32,35,41 and even than the clinical parameters (TNM and AJCC stages).22
Two studies analyzed the prognostic value of radiomic parameters in OS.21,26 In addition to histogram-based parameters (SUVmean, MTV, and TLG), 4 textural indices computed from gray-level co-occurrence matrices (correlation, entropy, energy, and coarseness) were evaluated among 284 patients.26 According to multivariate analysis, MTV (HR, 1.012; 95% CI, 1.003-1.021) and correlation (HR, 4.506; 95% CI, 1.178-17.239) remained independently correlated with OS. Based on the data from 52 patients,21 only the shape matrix defined by asphericity, which measures the deviation of the tumor’s shape from the shape of a sphere with the same volume, was associated with OS (HR, 6.4), whereas SUVmax (HR, 5.6), MTV or tumor stage (HR, 1.51; 95% CI, 0.33-6.99) were not correlated with OS in multivariate analysis.
Nineteen studies (1679 unique patients) specifically analyzed the FDG-PET/CT prognostic value in locoregional control, recurrence, and DFS.16-24,26-30,33-36,41
The SUVmax was the most commonly studied parameter (18 studies16,18-24,26-30,33-36,41), although only 6 of those studies found a correlation between SUVmax and DFS. Total lesion glycolysis was associated with locoregional control in approximately half of the studies (7 of 11 studies18,21,22,26-29,33-35,41). In contrast, the MTV was associated with DFS in 12 of 14 studies.17-19,21-23,26-29,33-35,41
For instance, the FDG-PET/CT data of 70 patients with HNSCC, half of whom were treated with surgery, were prospectively analyzed.23 According to multivariate analysis, only the MTV and the retention index (a variation of SUVmax between FDG-PET/CT performed 1 hour and 2 hours after injection) were prognostic factors (MTV: HR, 0.247; 95% CI, 0.075-0.817; retention index: HR, 0.331; 95% CI, 0.152-0.721). In the specific population of 70 patients with human papillomavirus–positive oropharyngeal carcinoma, an MTV greater than 41 mL was also correlated with a poor DFS (HR, 1.02; 95% CI, 1.01-1.03).34 Only 1 study (with 53 patients)41 did not show a correlation between the FDG-PET/CT parameters and DFS (SUVmax ≥28.5: HR, 1.88; 95% CI, 0.51-12.10; MTV ≥12.00 cm3: HR, 1.27; 95% CI, 0.14-27.40; and TLG ≥12.00 g, HR, 1.14, 95% CI, 0.37-3.33).
Three studies21,26,29 analyzed the prognostic value of textural indices in DFS, and 2 studies21,29 found a correlation. Among 108 patients, a special uptake pattern, ring shape, was associated with DFS (HR, 12.47; 95% CI, 4.59-33.91) in multivariate analysis after corrections for TNM and definitive therapy.29 The asphericity was also analyzed among 52 patients21 and was associated with DFS (HR, 7.70; 95% CI, 2.71-22.27), whereas tumor stage was not (HR, 1.37; 95% CI, 0.39-4.85).
Four studies24,25,28,41 analyzed the association of prognostic FDG-PET/CT with DM. Two studies28,41 showed a higher prognostic value of DM for MTV and TLG compared with SUVmax.
Among 86 patients with p16-positive oropharyngeal cancer,28 the MTV measured on the tumor site (MTV T) and on the lymph node (MTV N) as well as tumor TLG were associated with higher DM rates (MTV T: HR, 1.05; 95% CI, 1.00-1.10; MTV N: HR, 1.09; 95% CI, 1.03-1.16; and tumor TLG: HR, 1.01; 95% CI, 1.00-1.02). We did not find any study focused on the prognostic value of textural indices on the DM rate.
The value of FDG-PET/CT studies in forecasting occult nodal metastasis in clinically node-negative neck (cN0) HNSCC was evaluated in 9 articles involving 1294 unique patients.18,20,37-40,48-50
In a prospective multicentric study of 287 patients with T2 to T4 HNSCC cN0, an optimal SUVmax cutoff of 1.8 was found to be associated with occult nodal metastasis (negative predictive value [NPV], 0.942; 95% CI, 0.930-0.953). The surgical treatment plan was then changed in 22% of patients (14% had additional nodal levels and 5% had fewer).38 Similarly, an SUVmax of 9.5 or higher was the only independent risk factor for occult metastatic disease for cN0 oral cavity cancer and was associated with a consequential higher risk of local recurrence.48
The prognostic value of SUVmax was shown among 157 patients with HNSCC and unilateral node metastasis.20 A contralateral lymph node SUVmax of 2.5 or higher was associated with contralateral lymph node metastasis (HR, 3.28; 95% CI, 0.85-5.71) and had a clinically meaningful adverse effect on the 5-year disease-specific survival rate (HR, 1.05).
The FDG-PET/CT volumetric parameters (MTV and TLG) were also associated with a higher risk of occult nodal metastasis (odds ratios [ORs], 4.75; 95% CI, 1.30-17.33 and 5.36; 95% CI, 1.46-19.6, respectively) among 105 patients with cN0 oral cavity cancer.18
The prognostic value of FDG-PET/CT before salvage surgery was analyzed in 6 studies.42-47 According to univariate analysis, 4 studies42-45 showed an association between SUVmax and OS and DFS, and 3 studies also found that they were associated with FDG-PET/CT volumetric parameters (MTV and TLG).42,43,45 According to multivariate analysis, SUVmax did not remain a prognostic factor, whereas the MTV persisted as one in 2 studies43,45; TLG remained a prognostic factor in all of the studies. For instance, before 51 salvage surgeries in patients with laryngeal and pharyngeal cancer,42 after multivariate analysis with adjustment for clinical stage, only TLG of 5.4 or higher was found to be associated with shorter OS (HR, 3.14; 95% CI, 1.04-11.57), whereas an MTV of 2.4 or higher (HR, 1.92; 95% CI, 0.75-4.88) and SUVmax (HR, 3.4; 95% CI, 0.51-13.86) were not. The cutoff values of FDG uptake parameters were ascertained by the lowest–P value method. Although no clinically relevant difference was noted in locoregional recurrence-free survival, TLG was also associated with DFS, DM, and lung metastasis–free survival. The prognostic values of MTV and SUVmax were not investigated in terms of these 4 end points.
The prognostic value of FDG-PET/CT in DM before salvage surgery was also shown in 2 other studies.46,47 Among 82 patients with suspected recurrent HNSCC, an association between a high SUVmax and DM was found (NPV 0.97; 95% CI, 0.88-0.99).46 Only the SUVmax was analyzed in these 2 studies.
In the era of personalized medicine in oncological development, the lack of reliable and available prognostic molecular biomarkers limits the evolution of therapeutic strategies in HNSCC, except for human papillomavirus status with the new staging system.52 Owing to the technological progress and the recent increase in radiomics, the potential for imaging as a comprehensive tumor biomarker has begun to be highlighted.53
The FDG-PET/CT prognostic value in patients with HNC treated with radiochemotherapy has been the subject of several recent publications. In the results of this review, it also appears to be relevant for those treated with surgery (32 of the 36 studies included found a correlation).
Nevertheless, 27 of the 36 studies included patients exclusively treated with surgery, and only 17 studies focused only on primary surgery without taking into account salvage surgical treatment. When these patients received different treatments, they showed different clinical and tumoral characteristics. Thus, we can legitimately wonder whether these prognostic value results are generalizable. This is why among the 128 studies initially identified and among those that included patients treated with surgery and radiochemotherapy, we only kept those for which the treatment did not appear as a prognostic factor in univariate analysis21-24,27,29,32,34 or for which an analysis in the surgical treatment subgroup had been performed.26,30,41 The others54,55 were systematically excluded.
Similarly, we excluded the studies including patients with nonmucosal HNSCC.56,57 However, we noticed that 19 studies included patients without distinction of the different localizations of HNC (oropharynx, hypopharynx, oral cavity, larynx), and this parameter was almost never taken into account for the statistical analysis.
These two points raise concerns about the value of the results, especially when the cutoff measurement of the analyzed FDG-PET/CT parameters is the median of the population included, which leads to major differences in the values found between studies. Moreover, we wonder if it would not be more relevant to assess the cutoff point by optimizing the correlation with clinical outcome (eg, Cox model).58 In addition, according to survival analysis, dichotomizing the population into 2 groups of risk regarding a cutoff point could introduce measurement error and reduce the ability to detect a correlation.59 Instead of categorizing continuous variables, keeping them continuous with linear regression is relevant.60
Moreover, the way to define the volumetric PET parameters is discussed. Even though SUVmax is the most commonly used parameter, we showed that the volumetric parameters presented a higher prognostic value for OS, DFS, and DM (Table 3 and Table 4), probably because they take into account not only the value of 1 voxel but also the overall tumor uptake. Nevertheless, to be used, these volumetric parameters need a specific delineation that can be based on 4 techniques: a threshold of SUV (absolute [all voxels with an SUV value > x], relative [>x% of SUVmax], or adaptive), gradient-based, clustering, or statistical methods, and no consensus has currently been found.61 Of the 26 studies analyzing the volumetric parameters, only 5 studies21,22,31,33,34 did not use an automatic threshold of SUV equal to 2.5 or approximately 40% for the value of the threshold.
Textural indices that take into account the relationships between voxels begin to be relevant for prognostic value and are perhaps even better than volumetric indices. In fact, among 470 patients with p16-negative oropharyngeal cancer treated with primary radiochemotherapy, the overall concordance index for DFS in the PET/CT textural indices model (heterogeneity, SUV entropy, irregularity, and asphericity) was higher than that for clinical stage, performance status, SUVmax, and MTV.62 Nevertheless, we found only 3 radiomics studies analyzing data from patients treated with surgery, and those studies also included patients treated with radiochemotherapy (accounting for 16.7%, 22%, and 30.6% of the patients who received surgical treatment).
The most common cause of therapy failure in patients with HNSCC is undetected DM.63 The RTOG 9501 study64 reported a DM rate of 19.3% to 21.2% in patients treated for locally advanced HNC. The chance of curing them is very low when DM occurs, and OS decreases dramatically.65 Nevertheless, we found only 4 studies24,25,28,41 analyzing the prognostic value of FDG-PET/CT factors for DM, and none of those studies analyzed the association between textural indices and DM.
The management of clinically node-negative neck cancer is controversial. The presence of cervical lymph node metastasis is an important prognostic factor for HNC; as Mamelle et al66 showed among 914 patients with HNSCC, the number of positive nodes was an independent prognostic factor for survival in multivariate analysis. Although neck dissection is an effective procedure to eradicate positive nodes, the morbidity rate of this procedure is not negligible (approximately 30%67). Therefore, it would be interesting to better detect positive nodes to adapt therapeutic strategies. The use of FDG-PET/CT appears to be more accurate than CT alone or ultrasonography, with a sensitivity of 71.43% vs 23.8% and 4.76%, respectively; a specificity of 96.67%, 93.33%, and 93.33%, respectively; a positive predictive value of 93.5%, 71.0%, and 33.33%, respectively; and an NPV of 82.85%, 63.63%, and 58.33%, respectively, among 51 neck dissections.39
Furthermore, the prognostic value of occult nodal metastasis in cN0 seems to be found before salvage surgery. Kim et al68 showed the superiority of FDG-PET/CT against CT and/or MRI in detecting residual nodal disease among 39 patients before salvage surgery for HNC, with a sensitivity of 89% (95% CI, 74%-96%) and specificity of 95% (95% CI, 73%-99%) for FDG-PET/CT vs a sensitivity of 76% (95% CI, 58%-88%) and specificity of 74% (95% CI, 48%-90%) for CT and/or MRI. Nevertheless, this result is not highlighted before salvage surgery for recurrent laryngeal cancer.49,50 In fact, among 46 patients with recurrent laryngeal cancer cN0, FDG-PET/CT obtained a sensitivity of only 16.7% (95% CI, 3.5%-46.0%) and an NPV of 76.7% (95% CI, 62.1%-87.0%).49
Surgical salvage is associated with high complication and morbidity rates.69 Thus, it is interesting to select patients before undergoing such surgery, with the aim of improving the survival rate.70 However, only 3 studies focused on the prognostic value of FDG-PET/CT volumetric parameters before salvage surgery.42,43,45
In this review, PET coupled with CT appears to have an interesting prognostic value among patients with HNSCC treated with surgery. In addition, PET coupled with magnetic resonance imaging, even if it is a less commonly used technique, appears to be a promising research path. Among 72 patients who underwent PET/magnetic resonance imaging before surgery, MTV (HR, 3.06; 95% CI, 1.31-7.13), the ratio of MTV to the apparent diffusion coefficient (ADC) (HR, 3.12; 95% CI, 1.31-7.48), and the TLG:ADC ratio (HR, 4.33; 95% CI, 1.72-10.87) were independent prognostic factors of DFS.71
This FDG-PET/CT prognostic value should be further biologically interpreted. In fact, among 33 patients with HNSCC included prospectively to investigate the safety and activity of cetuximab during preoperative treatment, the FDG-PET/CT response was correlated with residual tumor cellularity in the surgical specimens (r = 0.84), and there was a trend toward downregulation of Ki67.72 In that study, the authors noted an important decrease in Ki67 expression and/or low tumor cellularity without any clinically meaningful modification of the tumor size; thus, FDG-PET/CT appeared to be more accurate than conventional imaging for detecting tumor evolution.
To our knowledge, this is the first systematic review focusing on the prognostic effectiveness of FDG-PET/CT parameters as biomarkers of OS, DFS, and DM among patients with HNSCC treated with surgery (primary or for salvage). We noted a promising role of the volumetric parameters (MTV and TLG) for identifying patients with a higher risk of postsurgical progression who could receive early therapeutic intervention to improve their prognosis. However, additional large-scale studies that include patients exclusively treated with surgery are stratified according to localization, and focus on the textural indices are necessary to obtain conclusive results.
Accepted for Publication: January 7, 2020.
Corresponding Author: Gwenaelle Creff, MSc, Department of Otolaryngology–Head and Neck Surgery, Rennes University Hospital, 2 rue Henri Le Guilloux, F-35000 Rennes, France (firstname.lastname@example.org).
Published Online: March 26, 2020. doi:10.1001/jamaoto.2020.0014
Author Contributions: Ms Creff and Dr Castelli had full access to all the data in the study and take responsibility for the integrity of the data and the collection of the data analysis.
Concept and design: Creff, Acosta, Jegoux, Castelli.
Acquisition, analysis, or interpretation of data: Creff, Devillers, Depeursinge, Palard-Novello, Jegoux, Castelli.
Drafting of the manuscript: Creff, Castelli.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Creff, Depeursinge.
Administrative, technical, or material support: Creff, Depeursinge.
Supervision: Devillers, Palard-Novello, Jegoux, Castelli.
Conflict of Interest Disclosures: Dr Depeursinge reported receiving grants from the Swiss National Science Foundation during the conduct of the study. No other disclosures were reported.
Additional Contributions: The LTSI (Image and Signal Processing Laboratory, INSERM, U1099, Rennes, France) and the Cancer Institute Eugène Marquis (Rennes) supported the study.
Create a personal account or sign in to: