Lung cancer has a poor prognosis, and early diagnosis is essential for changing the prognosis. ML techniques have recently emerged as promising technologies for enhancing lung cancer detection, primarily through medical imaging data. Here, we present an abstract overview of some significant advances and challenges in MLbased lung cancer detection. This paper concentrates specifically on automated lung nodule detection and classification in CT images using ML algorithms like CNNs, SVMs. We also review some of recent studies that combine radiomics features with clinical data to predict its risk of progression towards lung cancer. We are facing challenges like sparsity of data, interpretability of the models and generalization but also opportunities for future research such as multimodal imaging data and use of transfer learning approaches. Overall, this article highlights potential applications of ML-based approaches in lung cancer as a means for early detection, prognostication, and treatment and ultimately improves outcomes and reduces mortality.