Koike Y, Aokage K, Ikeda K, et al. - Cancer tissue has both a cancer cell component and a stromal component, so researchers used a machine learning method to determine whether the component ratio could predict outcomes for lung squamous cell carcinoma (SqCC) patients. They enrolled 135 peripheral SqCC patients (tumor size: 3-5 cm). They used a machine learning method to precisely measure the areas of the cancer cell component, the necrotic component, and the stromal component. Patients were divided into three subtypes: predominant cancer cell, predominant necrosis, and predominant stroma. A significantly shorter recurrence free survival (RFS) was observed in patients with the predominant stroma subtype vs those with the predominant cancer cell subtype. The predominant stroma subtype was an independent factor for RFS, confirmed via machine learning, which infers that the malignant potential of SqCC is related to the ratio of the stromal component.