A Study Of Radiomics And Genetic Polymorphism In Head And Neck Squamous Cell Carcinoma

MRI-Based Radiomics for Predicting Overall Survival and Staging in Head and Neck Squamous Cell CarcinomaIndividualized therapy based on the specific condition of patients and cancer to improve patient survival and quality of life,is the ultimate goal of head and neck squamous cell carcinomas(HNSCC)treatment.Radiomics enables to extract a large number of features from sectional medical images,and to exchange visual image into quantitative radiomics characteristics.The aim of the study was to develop a magnetic resonance imaging(MRI)based radiomics signature and radiomics nomogram for predicting prognosis and staging of HNSCC patients.In the first section,we included 170 cases of HNSCC patients.Seven features were eventually selected from 485 radiomics features generated from preoperatively T2-weighted(T2W)images.Radiomcis signature was built upon them,which was proved to be associated with overall survival(OS)of patients(training cohort,P < 0.0001;testing cohort,P = 0.0013).Multivariable Cox proportional hazards model was performed to identify independent predictors from radiomics signature and clinical characteristics(such as age,gender,smoking and alcohol using,and TNM stage).The radiomics signature and TNM stage were proved to be independently associated with OS of HNSCC patients,which therefore were incorporated to generate the nomogram.In the training cohort,the C-index of nomogram(C-index = 0.76,95% CI: 0.66?0.87)was significant higher than TNM stage(C-index = 0.63;95% CI: 0.55?0.72)(P = 0.005 for comparison).In the testing cohort,the C-index of nomogram(C-index = 0.72;95% CI: 0.63?0.82)was also significant higher than TNM stage(C-index=0.61;95% CI: 0.51?0.71)(P = 0.01 for comparison).Moreover,the C-index of radiomics signature was higher than TNM stage(training cohort P = 0.027,testing cohort P = 0.035).The calibration curves of the nomogram demonstrated good agreement.The decision curve analysis indicated that the radiomics nomogram had a higher overall net benefit than the traditional staging system.In the second section,we included 127 cases of HNSCC patients.LASSO logistic regression was used and selected ten radiomics features from T2 W,six features from contrast enhanced-T1W(CE-T1W)and six features from combined T2W+CE-T1 W,and radiomics signatures were respectively built.ROC curves were established to evaluate the ablility of the three radiomics signatures in preoperatively discriminating early(stage I-II)and late(stage III-IV)stage of HNSCC,and AUC values were calculated: T2 WI based signature(trainging cohort 0.818,testing cohort 0.709),CE-T1 WI based signature(trainging cohort 0.823,testing cohort 0.853),T2WI+CE-T1 WI based signature(trainging cohort 0.850,testing cohort 0.849).Combined T2WI+CE-T1 WI performed better in both training and testing cohorts,with accuracy of 0.788 and 0.857 in each cohort.The Mc Nemar test was adopted for comparing the accuracy of clinical TNM stage and radiomics signature for predicting pathological TNM stage.In the training cohort,the radiomics signature based on CE-T1 WI and T2WI+CE-T1 WI both showed better accuracy and sensitivity than conventional TNM stage(P < 0.005).MRI based radiomics signature could help to preoperatively predict OS and TNM stage of HNSCC patients.Radiomics nomogram potentially improved clinical prognostic ability for HNSCC patients,providing a non-invasive way to quantifying imaging features and monitoring patient prognosis in clinical practice.Association between E2F1 Gene Polymorphorism and Risk of Oropharynx Squamous Cell CarcinomaOral squamous cell carcinoma(OSCC)comprises oropharynx squamous cell carcinoma(OPSCC)and oral cavity squamous cell carcinoma(OCSCC).Human papillomavirus(HPV),especially HPV16 genotype,is associated with OPSCC.However,in fact,only a small percentage of HPV-infected individuals develop OSCC,suggesting that other factors such as genetic susceptibility or polymorphism may also play critical roles in carcinogenesis.HPV activates E2F1-driven transcription via the E7-pRb-E2 F pathway.Genetic polymorphisms in the 3‘ untranslated region(UTR)targeted by miRNAs can affect the regulation of target genes and individual cancer risk.Thus,we hypothesized that a polymorphism at the 3‘UTR miRNA binding site of E2F1 gene(rs3213180)was associated with risk of oral squamous cell carcinoma(OSCC)and tumor HPV status of OPSCC.We determined the rs3213180 polymorphism and HPV16 L1 serology of 325 OSCC patients and 335 controls,and tumor HPV16 status of 552 OPSCC.Logistic regression models were used to calculate associations of rs3213180 polymorphism with risk of HPV-associated OSCC and tumor HPV status of OPSCC.Patients with both HPV seropositivity and the Ins/Del or Ins/Ins genotype of rs3213180 had the highest risk of OSCC,while the lowest risk was detected in patients with HPV seronegativity and the Del/Del genotype.Using the Del/Del genotype and HPV16 seronegativity as the reference group,the risk of OSCC progressively increased among individuals with the Ins/Del or Ins/Ins genotype and HPV16 seronegativity(OR = 3.5;95% CI: 2.4~5.2),the Del/Del genotype and HPV16 seropositivity(OR = 3.6;95% CI: 2.1~6.2),and the Ins/Del or Ins/Ins genotype and HPV16 seropositivity(OR = 9.0;95% CI: 4.7~17.2).A similar and more prominent effect was detected in OPSCC,but not in OCSCC patients.Notably,that effect trend was pronounced in never-smokers and never-drinkers.Furthermore,the patients with the Ins/Del or Ins/Ins genotype were 2.9 timesmore likely to have HPV-positive tumors than those with the Del/Del genotype(OR = 2.9;95% CI: 1.8~4.6).Our results suggest that the E2F1 rs3213180 polymorphism may influence susceptibility to HPV-associated OSCC,particularly for OPSCC,never-smokers and never-drinkers,but not for patients with OCSCC.