ABSTRACT
Small cell lung cancer (SCLC) is a highly aggressive subtype of lung cancer characterized by rapid tumor growth and early metastasis. Accurate prediction of prognosis and therapeutic response is crucial for optimizing treatment strategies and improving patient outcomes. In this study, we conducted a deep-learning analysis of Hematoxylin and Eosin (H&E) stained histopathological images using contrastive clustering and identified 50 intricate histomorphological phenotype clusters (HPCs) as pathomic features. We identified two of 50 HPCs with significant prognostic value and then integrated them into a pathomics signature (PathoSig) using the Cox regression model. PathoSig showed significant risk stratification for overall survival and disease-free survival and successfully identified patients who may benefit from postoperative or preoperative chemoradiotherapy. The predictive power of PathoSig was validated in independent multicenter cohorts. Furthermore, PathoSig can provide comprehensive prognostic information beyond the current TNM staging system and molecular subtyping. Overall, our study highlights the significant potential of utilizing histopathology images-based deep learning in improving prognostic predictions and evaluating therapeutic response in SCLC. PathoSig represents an effective tool that aids clinicians in making informed decisions and selecting personalized treatment strategies for SCLC patients.
ABSTRACT
BACKGROUND: Several studies have proposed grading systems for risk stratification of early-stage lung adenocarcinoma based on histological patterns. However, the reproducibility of these systems is poor in clinical practice, indicating the need to develop a new grading system which is easy to apply and has high accuracy in prognostic stratification of patients. METHODS: Patients with stage I invasive nonmucinous lung adenocarcinoma were retrospectively collected from pathology archives between 2009 and 2016. The patients were divided into a training and validation set at a 6:4 ratio. Histological features associated with patient outcomes (overall survival [OS] and progression-free survival [PFS]) identified in the training set were used to construct a new grading system. The newly proposed system was validated using the validation set. Survival differences between subgroups were assessed using the log-rank test. The prognostic performance of the novel grading system was compared with two previously proposed systems using the concordance index. RESULTS: A total of 539 patients were included in this study. Using a multioutcome decision tree model, four pathological factors, including the presence of tumor spread through air space (STAS) and the percentage of lepidic, micropapillary and solid subtype components, were selected for the proposed grading system. Patients were accordingly classified into three groups: low, medium, and high risk. The high-risk group showed a 5-year OS of 52.4% compared to 89.9% and 97.5% in the medium and low-risk groups, respectively. The 5-year PFS of patients in the high-risk group was 38.1% compared to 61.7% and 90.9% in the medium and low-risk groups, respectively. Similar results were observed in the subgroup analysis. Additionally, our proposed grading system provided superior prognostic stratification compared to the other two systems with a higher concordance index. CONCLUSION: The newly proposed grading system based on four pathological factors (presence of STAS, and percentage of lepidic, micropapillary, and solid subtypes) exhibits high accuracy and good reproducibility in the prognostic stratification of stage I lung adenocarcinoma patients.
