Renal Turmor Differentiation From Primary Hydronephrosis Via Noncontrast CT-Based Radiomics Signature

Renal cell carcinoma(RCC)is the most common type of kidney cancer in adults.The global incidence of the RCC is rising,meanwhile its mortality is decreasing in some developed countries.These trends can be attributed to the advanced medical imaging modalities and improved therapeutic measures,which helps detect renal masses with better prognosis successfully.Among these medical imaging modalities,computed tomography(CT)is widely used in effective renal masses detection,accurate RCC characterization,and critical tumor grade prediction in most cases.However,some factors may affect the diagnosis with the medical imaging.For example,RCC usually overlap with kidneycalculus due to nonspecific and insidious symptoms,i.e.,osphalgia and hematuria.Because the overlapped tissue may obscure RCC and other important tissue and reduce the diagnosis accuracy,this can limit the RCC characterization performance in CT images.In addition,the CT image quality would affect the diagnosis.Therefore,it is a challenging task to differentiate and diagnose the RCC from the kidneycalculus accurately and effectively.A number of methods have been developed to differentiate RCC from the kidney calculus in renal CT images.Recently,as an emerging field,the radiomics has shown great potential for renal disease prognosis.It can be seen that the translation of radiomics analysis into renal disease care can provide an effective treatment decision-making.Given the increase use of nonenhanced CT imaging in renal disease decision-making in clinics,we hypothesize the radiomics analysis can provide important histopathologic prognosticators prediction for the risk assessment of specific renal disease outcomes.In this work,we proposed a computer-aid classification model to differeniate RCC from primary hydronephrosis from nonenhanced CT images with machine-learning classifiers with quantitative radiomics features.