| ObjectivesThe study on PET/CT radiomics has gradually increased after growing in number of oral squamous cell carcinoma(OSCC)patients undergone positron emission tomography/computed tomography(PET/CT)examination.However,PET radiomics was rarely used to evaluate the prognosis of OSCC patients.This study aims to investigate the clinical value of PET radiomics in evaluating the prognosis of patients with OSCC,establish prognostic models based on the findings of statistical analysis,and further develop visual nomogram to aid in OSCC prognosis prediction and treatment.Materials and MethodsThis was a retrospective study with 116 OSCC patients who received PET/CT scan one week preoperation in the Oral and Maxillofacial Surgery Department of Southern Medical University NanFang Hospital between April 2014 and December 2021.The patients were randomly divided into the training group(n=58)and the validation group(n=58).The PET images of PET/CT were uploaded to the LIFEx software for the extraction of radiomics features,and LASSO-Cox regression analysis was used to create the PET radiomics scores.Independent prognostic factors were identified through univariate and multivariate Cox regression analysis,and then prognostic models were established according to the results.C-index,receiver operating characteristic curve(ROC),time C-index curve,calibration curves,and decision curve analysis were used to evaluate the accuracy of the prognostic models in the training group and the validation group respectively.Nomogram was developed based on the prognostic model that performed the best.Finally,X-tile software was used to screen out the threshold of the scores of the nomogram for risk stratification.ResultsThere are respectivly three PET radiomics features screened out via LASSOCox regression to construct PET radiomics scores related to overall survival(OS)and progression-free survival(PFS).The results showed that N stage and PET radiomics scores were independent prognostic factors for OS.N stage,histological differentiation,and PET radiomics scores were independent prognostic factors for PFS.The combined models were developed based on clinical factors and PET radiomics scores,and they showed good discrimination in both the training(C index:OS:0.836,PFS:0.776)and the validation group(C index:OS:0.744,PFS:0.712).The calibration curves also demonstrated good consistency between the predicted survival rate and the actual survival rate.The combined models had the greatest clinical net benefit.In addition,the nomogram score can realize the risk stratification of OSCC patients.Lower nomogram score was significantly related to poorer OS and PFS.ConclusionsPET radiomics scores were independent prognostic factors for OSCC patients.The combined models based on the radiomics scores might be used for predicting the prognosis of OSCC patients,which may throw light on the outcome prediction and treatment for OSCC. |