In recent years, imaging techniques have been successfully used to deliver diagnostic biomarkers with even greater accuracy. In particular, radiomic analysis methods (application of artificial intelligence on radiological images), which describe a segmented tumor region, using various quantitative characteristics derived from radiological images, have shown great potential in the identification, characterisation/classification of different types of cancer and in evaluating the response to radiotherapy and chemotherapy. Liquid biopsy is used for both early screening of malignancy and diagnosing minimal residual disease. It is also performed to assess and monitor the response to pharmacological treatments for a personalised therapeutic strategy. The analysis of morphostructural data obtained by imaging, correlated with the genetic/molecular results of liquid biopsy, could provide useful predictive factors for early diagnosis and predicting the response to anti-cancer drugs. The study aims to design and develop a report structured in CT with contrast media, which includes, in addition to the subjective evaluation of the radiologist, a quantitative/objective assessment of lung cancer (LC) with features that describe the texture and morphology of the lesion. Therefore, we present a workflow aimed at extracting the DICOM images acquired with CT using contrast medium from a significant number of patients, and to evaluate their accuracy in characterising the LC lesions. Furthermore, these data will be correlated to gene mutations and epigenetic changes (DNA methylation) evaluated in circulating tumour DNA derived from peripheral blood with a liquid biopsy approach. The correlation between radiomic characteristics, quantitative analysis of tumours performed by CT, structured lesion reports, and liquid biopsies could help avoid many unnecessary biopsy procedures and enable personalised treatment of LC patients.

Impact statement
An integrated radiomics–liquid biopsy model is proposed to support non-invasive molecular assessment and guide personalised therapy in patients with stage III lung cancer.