MRI Texture Analysis Predicts p53 Status in Head and Neck Squamous Cell Carcinoma

BACKGROUND AND PURPOSE:Head and neck cancer is common, and understanding the prognosis is an important part of patient management. In addition to the Tumor, Node, Metastasis staging system, tumor biomarkers are becoming more useful in understanding prognosis and directing treatment. We assessed whether MR imaging texture analysis would correctly classify oropharyngeal squamous cell carcinoma according to p53 status.MATERIALS AND METHODS:A cohort of 16 patients with oropharyngeal squamous cell carcinoma was prospectively evaluated by using standard clinical, histopathologic, and imaging techniques. Tumors were stained for p53 and scored by an anatomic pathologist. Regions of interest on MR imaging were selected by a neuroradiologist and then analyzed by using our 2D fast time-frequency transform tool. The quantified textures were assessed by using the subset-size forward-selection algorithm in the Waikato Environment for Knowledge Analysis. Features found to be significant were used to create a statistical model to predict p53 status. The model was tested by using a Bayesian network classifier with 10-fold stratified cross-validation.RESULTS:Feature selection identified 7 significant texture variables that were used in a predictive model. The resulting model predicted p53 status with 81.3% accuracy (P < .05). Cross-validation showed a moderate level of agreement (κ = 0.625).CONCLUSIONS:This study shows that MR imaging texture analysis correctly predicts p53 status in oropharyngeal squamous cell carcinoma with ∼80% accuracy. As our knowledge of and dependence on tumor biomarkers expand, MR imaging texture analysis warrants further study in oropharyngeal squamous cell carcinoma and other head and neck tumors.

Head and neck cancer is the sixth most common cancer worldwide,¹ with squamous cell carcinoma accounting for approximately 90% of all cases. Most head and neck squamous cell carcinoma (HNSCC) occurs in the oral cavity, oropharynx, and larynx. Alcohol and tobacco consumption and prior infection with human papillomavirus are the major risk factors associated with the development of head and neck squamous cell carcinoma. Oropharyngeal squamous cell carcinoma is of particular interest because its incidence is increasing, particularly among younger, nonsmoking patients.²Accurate staging of HNSCC is essential for treatment planning and prognostication, and a standard tool used for staging is the American Joint Committee on Cancer Tumor, Node, Metastasis staging system, currently in its seventh revision.³ As we learn more about tumor biology, however, it is clear that this staging system does not fully predict clinical behavior and prognosis. Our knowledge of head and neck cancer pathogenesis has rapidly increased, and better understanding of molecular mechanisms holds the promise of discovering predictive and prognostic biomarkers that might be helpful in the management of HNSCC.⁴ The tumor suppressor p53 plays an important role in conserving genomic stability.⁵ p53 facilitates DNA repair by regulating the cell cycle and has a role in preventing cancer emergence.^6,7 Mutations in the gene encoding the p53 protein, TP53, occur in almost 50% of all cancers.^8,9 In most of HNSCC, mutation and inactivation of p53 is an essential and early event in neoplastic transformation, and TP53 mutations are associated with poor prognosis in HNSCC.^10–12 A landmark prospective study classified TP53 gene mutations on the basis of their effect on p53 protein structure.¹³ Broadly, disruptive mutations disturb the formation of p53-DNA complexes, while nondisruptive mutations have little effect on the association between p53 and DNA. The study reported a significant association between the presence of TP53 disruptive mutations and worse overall survival in surgically treated HNSCC compared with both nondisruptive TP53 mutations and wild-type TP53.¹³ A recent study has also implicated disruptive mutations in TP53 leading to radiation-treatment failure.¹⁴Medical imaging plays a critical role in the assessment of many head and neck tumors, and both CT and MR imaging have important roles in the anatomic evaluation of HNSCC.^15–18 In addition to anatomic details, the analysis of MR images provides additional metabolic and biologic information in tumors.¹⁹ Mathematic techniques that quantify image characteristics have been applied to a vast array of pathologies, from multiple sclerosis,²⁰ attention deficit/hyperactivity disorder,²¹ and Alzheimer disease²² to breast cancer,²³ cervical cancer,²⁴ and brain tumors.²⁵ Studies in glioblastoma have shown that there is a correlation between the methylation of O6-methylguanine-DNA methyltransferase and MR imaging features.²⁶ Levner et al²⁷ extracted texture features from MR images by using spatial frequency analysis and the Stockwell transform (ST) representation²⁸ and fed these characteristics into a neural network to predict the methylation status with an average accuracy of 87.7%.²⁷ Brown et al²⁵ also extracted ST texture features from brain MR images to find that codeletion of chromosomes 1p and 19q, a marker of good prognosis in oligodendroglioma brain cancer, could be predicted with 94% accuracy. These studies suggest that differences in tumor-tissue composition react with MR imaging signals differently, thus affecting texture features.Yu et al²⁹ looked at differentiating tissues by using texture characterization on FDG-PET/CT images in head and neck cancers. We explored the use of ST texture features on MR images with a machine-learning technique to objectively differentiate head and neck tumors by p53 status. We hypothesized that MR image analysis could successfully discriminate p53-positive and -negative tumors.