Blood-based biomarkers for lung cancer risk prediction and prognosis in subjects with low-dose computed tomography (LDCT) detected lung nodules of the Rete Italiana Screening Polmonare (RISP) cohort

Acronimo:bioRISP 

Co-PI: Federica Sabia

An open issue in LDCT screening is the high false-positive rate and overdiagnosis. Despite the introduction in the Lung-RADS guidelines of more precise volumetric measurement for solid nodule size as well as Artificial Intelligence (AI) softwares, LDCT screening still yields a high rate of false positive scan (>90% of all detected nodules). 

These results lead to unnecessary LDCT repeats and invasive procedures returning non-malignant results, as well as morbidity and mortality issues. Distinguishing the small percentage of malignant nodules from the majority of the detected pulmonary nodules (PNs) remains challenging. 

Blood-based biomarkers, such as circulating microRNAs (miRNAs), represent a promising complementary tool. At our institution, we developed a 24-miRNA signature classifier (MSC) based on a plasma assay, first validated retrospectively in the MILD trial (Sozzi G., JCO 2014) and further assessed in the BioMILD prospective screening trial. The BioMILD study, involving 4,119 participants, combined LDCT with MSC, demonstrating improved prediction of lung cancer incidence and mortality (Pastorino U., Ann Oncol 2022). 

To refine the utility of MSC, we analyzed 1,040 BioMILD participants with suspicious LDCT findings. Over a median follow-up of 8.5 years, we confirmed the clinical value of MSC in stratifying risk and guiding the management of uncertain LDCT findings (Boeri M., Lancet RH Europe 2024).  

In the present project we hypothesize that the validation of a prediction model for malignancy of LDCT detected high-risk lung nodules that includes blood biomarkers might a) define the risk of lung cancer development at baseline and further screening rounds b) improve diagnosis of malignancy for indeterminate nodules c) predict lung cancer prognosis and impact on lung cancer mortality. 

We will take advantage of the bio-repository and the radiologic imaging database of a large LDCT prospective screening cohort of RISP to perform a comprehensive validation of our blood-based risk prediction model to assess efficiently which screen-detected lung nodules are likely to be malignant and ultimately reduce lung cancer and all-cause mortality. 

In details, 400 subjects with LDCT lung nodules classified as LungRADS 3, 4A, 4B,4X or LungRads 2 for non-solid nodules, at the baseline and returning at year 2 screening round will be included in the study. In addition, we will analyze samples from newly diagnosed incident nodules detected in 1200 volunteers returning at year 2 screening round (100 nodules expected). 

MSC biomarker and the molecular cargo of circulating EVs will be investigated. Our final aim is to evaluate the clinical utility of blood-based biomarkers to assess lung nodule malignancy and prediction of patients’ prognosis (Flow chart Figure 1).