Outcome prediction of head and neck squamous cell carcinoma by MRI radiomic signatures

Abstract

Objectives Head and neck squamous cell carcinoma (HNSCC) shows a remarkable heterogeneity between tumors, which may be captured by a variety of quantitative features extracted from diagnostic images, termed radiomics. The aim of this study was to develop and validate MRI-based radiomic prognostic models in oral and oropharyngeal cancer. Materials and Methods Native T1-weighted images of four independent, retrospective (2005–2013), patient cohorts (n = 102, n = 76, n = 89, and n = 56) were used to delineate primary tumors, and to extract 545 quantitative features from. Subsequently, redundancy filtering and factor analysis were performed to handle collinearity in the data. Next, radiomic prognostic models were trained and validated to predict overall survival (OS) and relapse-free survival (RFS). Radiomic features were compared to and combined with prognostic models based on standard clinical parameters. Performance was assessed by integrated area under the curve (iAUC). Results In oral cancer, the radiomic model showed an iAUC of 0.69 (OS) and 0.70 (RFS) in the validation cohort, whereas the iAUC in the oropharyngeal cancer validation cohort was 0.71 (OS) and 0.74 (RFS). By integration of radiomic and clinical variables, the most accurate models were defined (iAUC oral cavity, 0.72 (OS) and 0.74 (RFS); iAUC oropharynx, 0.81 (OS) and 0.78 (RFS)), and these combined models outperformed prognostic models based on standard clinical variables only (p < 0.001). Conclusions MRI radiomics is feasible in HNSCC despite the known variability in MRI vendors and acquisition protocols, and radiomic features added information to prognostic models based on clinical parameters. Key Points • MRI radiomics can predict overall survival and relapse-free survival in oral and HPV-negative oropharyngeal cancer. • MRI radiomics provides additional prognostic information to known clinical variables, with the best performance of the combined models. • Variation in MRI vendors and acquisition protocols did not influence performance of radiomic prognostic models.

Radiomics and radiogenomics in head and neck squamous cell carcinoma: Potential contribution to patient management and challenges

“…Many studies have attempted different feature extraction methods and modelling to characterize tumors and directly predict oncologic outcomes of patients with HNSCC. The results of the eleven most pertinent studies focused on overall survival (OS) and progression-free survival (PFS) are presented in Table 3 [58][59][60][61][62][63][64][65][66][67].…”

“…Secondly, they were the first to develop a model based on four radiomic features that was successful in predicting overall survival, and their influence is reflected in practically all subsequent works, particularly in the Leijenaar´s study [59], which performs a validation of the radiomic signature proposed in a specific OPSCC population. Thus, following the line established by Aerts [58], a large majority of studies opted for relatively simple models, built with a maximum of seven radiomic features [61][62][63][64]66,67], although more complex models with up to 24-27 variables have also been explored [60].…”

MRI Intensity Standardization Evaluation Design for Head and Neck Quantitative Imaging Applications

“…Weighted images, particularly T2-w images, are commonly acquired in the scanning protocol due to their excellent soft-tissue contrast in the complicated anatomic areas involved in HNC, and are thus of added value for tumor and healthy tissue delineation 10,11 . While many recent HNC studies have implemented quantitative analysis of conventional weighted MRI [12][13][14][15][16][17][18][19] , relatively few explicitly incorporate intensity standardization in their methodological pipelines [15][16][17][18][19] , with even fewer testing multiple methods 19 . Moreover, while rigorous studies have tested MRI intensity standardization methods in various anatomical regions, chiefly in the brain 5,20 , the head and neck region has yet to be systematically investigated.…”

Oral cancer: changing the aim of the biopsy in the age of precision medicine. A review

“…Thanks to the extraction of mathematically defined parameters from routine medical images it is possible to generate large-scale sets of information that can be correlated with OS and treatment-related toxicity and can also be used to identify new biomarkers to be implemented in daily clinical practice 82 . Currently, there are few but promising published studies on the application of radiomics in oral cancer with CT or MRI [83][84][85] . Probably, if the role of radiomics will be confirmed with standardised methodology on a large number of patients, these results would help to promote cancer treatment towards personalised precision medicine 82,86 .…”