In recent years, AI has gained popularity in healthcare and in particular in the oncology field. Early cancer detection and accurate type identification are imperative for better patient outcomes and should increase the life span and improve the quality of life for some of the affected patients. To detect lung cancer, standard procedures such as the microscopic examination of tissue slides and pathology image analysis have been in practice for years. However, these methods are not only time-consuming but also provide a lot of variabilities in result interpretation (Wang et al.
In order to make this process more precise and reduce variations, AI should be applied. Some of the areas in pathology image analysis where AI (this includes machine learning (ML) and deep learning (DL)) have provided favorable results include more precise identification of cancer-affected area, prediction of outcomes, cancer characterization, and detection of possible metastasis (Coccia, 2020; Wang et al., 2019). Per Rabbani, Kanevsky, Kafi, Chandelier, and Giles (2018), ML is described as “the analysis and interpretation of data by machine algorithms that allow classification, prediction and segmentation of information to provide insights not readily available to the human eye or cognition” (p.
2, para 1).
ML uses algorithms to identify and diagnose very small cancerous areas that otherwise would be almost impossible to detect (Rabanni et al.
, 2018). Another way that ML can be used for early diagnosis of lung cancer is “Optical biopsy” (Rabanni et al., 2018). In Optical biopsy, ML is combined with fibered confocal fluorescence microscopy (FCFM), where FCFM produces “in vivo endo-microscopic images of the human respiratory tract in realtime,” and ML helps extract those images and subsequently uses them to distinguish between non-cancerous and cancerous cells (Rabanni et al., 2018, p.2, para 5). ML can also be utilized in the genomic classification of NSCLC.
During the classification process, microarrays are used to train ML algorithms to identify various lung cancer mutations such as those in the epidermal growth factor receptor (EGFR) or anaplastic lymphoma kinase (ALK) receptor, which are also used as biomarkers for targeted treatments (Rabanni et al., 2018). Besides being used in lung cancer diagnosis, ML can be used in lung cancer prognosis, too. For patient survival prediction, it is important to distinguish between the three types of NSCLCs (squamous-cell carcinoma, adenocarcinoma, or large-cell carcinoma), since patients with different subtypes of lung cancer have different survival outcomes (Zappa & Mousa, 2016). In order to do this, ML can be used to determine “quantitative morphological features extracted from histological slides” (Rabanni et al., 2018, p.3, para 5). Yu and his team had used ML to analyze images of histopathology slides that contained lung tissues with adenocarcinoma and squamous cell carcinoma (Yu et al., 2016).