Prognostic and Predictive Radiomics Nomogram for Estimating Survival Benefit of Adjuvant Chemotherapy in Resected Early-Stage (I or II) Lung Adenocarcinoma

Background: The recommendations regarding adjuvant chemotherapy (ACT) for early-stage (I or II) non-small cell lung cancer are controversial because of the published ambiguous survival benefits. An established biomarker to identify the potential beneficiaries can facilitate individual decision-making of ACT and prolong the survival in early-stage lung adenocarcinoma (ADC). We aimed to construct a radiomics nomogram to predict disease free survival (DFS) and survival benefit of ACT for early-stage ADC. Methods: We retrospectively analyzed 377 patients with resected early-stage ADC diagnosed by pathology. CT-derived radiomics features were selected by least absolute shrinkage and selection operator Cox regression to generate a radiomics signature associated with DFS. The independent clinical risk predicators were incorporated to construct a clinical nomogram and a radiomics nomogram along with the radiomics signature. The risk scores for the nomograms were calculated. The association between ACT and survival benefit was assessed with reference to discrimination, calibration and clinical utility in different risk subgroups stratified by the radiomics signature, clinical nomogram and radiomics nomogram. Findings: The radiomics signature and risk scores for nomograms were significantly associated with DFS (all P < 0.001). ACT was significantly associated with survival benefit among high-risk patients defined by radiomics nomogram (hazard ratio [HR]: 0.41, 95% confidence interval [CI]: 0.24-0.72, P = 0.002) and radiomics signature (HR: 0.43, 95% CI: 0.23-0.81, P = 0.009). The integration of radiomics signature into the radiomics nomogram contributed to an improved discriminative performance (C-index: 0.74 to 0.78, P = 0.037 ) and a better calibration. The radiomics nomogram outperformed radiomics signature and clinical nomogram in clinical utility according to decision curve analysis. Interpretation: The radiomics nomogram approach can feasibly identify the potential beneficiaries of ACT among early-stage ADC patients, which could serve as a helpful tool in informing decision-making in the adjuvant setting. Funding: The study was supported by the National Natural Science Foundation of China (NO. 81873889). Declaration of Interests: The authors have no conflicts of interest to declare. Ethics Approval Statement: The institutional ethics committee of Tongji Hospital approved this retrospective study and requirement for informed consent was waived.