Towards Machine Learning-Aided Lung Cancer Clinical Routines: Approaches and Open Challenges

Abstract

Advancements in the development of computer-aided decision (CAD) systems for clinical routines provide unquestionable benefits in connecting human medical expertise with machine intelligence, to achieve better quality healthcare. Considering the large number of incidences and mortality numbers associated with lung cancer, there is a need for the most accurate clinical procedures; thus, the possibility of using artificial intelligence (AI) tools for decision support is becoming a closer reality. At any stage of the lung cancer clinical pathway, specific obstacles are identified and "motivate" the application of innovative AI solutions. This work provides a comprehensive review of the most recent research dedicated toward the development of CAD tools using computed tomography images for lung cancer-related tasks. We discuss the major challenges and provide critical perspectives on future directions. Although we focus on lung cancer in this review, we also provide a more clear definition of the path used to integrate AI in healthcare, emphasizing fundamental research points that are crucial for overcoming current barriers.

Keywords: CT scan; computer-aided decision; learning models; lung cancer.

Figure 1

Prevalence of the major histological subtypes of lung cancer: non-small cell lung cancer and small cell lung cancer [13,14,15].

Figure 2

Physiological response to the activation of membrane receptors and immune receptors. Activation of receptor kinases, such as EGFR, FGFR, or HER-2, promotes the activation of RAS and its downstream pathways that facilitate cell growth, proliferation, cell survival, and differentiation. However, mutations in the RAS family lead to its constitutive activation and the hyperactivation of the downstream pathways—leading to uncontrolled cell survival. On the other hand, PD-L1 is present on the cell surface of immune cells and its binding to tumor cells PD-1 inhibits cell proliferation, cytokine secretion, and the expression of anti-apoptotic molecules in immune cells culminating in the escape of cancer from immunosurveillance. The goal of target therapies is to diminish the activation of abnormal signalling pathways, which can be inhibited at every step.

Figure 3

Radiomic vs. radiogenomic perspectives for lung cancer assessment. The assessment of lung cancer was first based on the nodule; however, recently, in radiogenomics approaches, other lung structures were shown to have relevant information for cancer characterization. Those approaches brought new challenges, such as the segmentation of lungs, to use this region of interest (ROI) in the AI-based models.

Figure 4

Two main perspectives for CADs in lung cancer, focused on the nodule and a more holistic approach that takes into consideration information about the surrounding structures of the nodule.