CT-Based Deep Learning Model for Invasiveness Classification and Micropapillary Pattern Prediction Within Lung Adenocarcinoma

Abstract

Identification of tumor invasiveness of pulmonary adenocarcinomas before surgery is one of the most important guides to surgical planning. Additionally, preoperative diagnosis of lung adenocarcinoma with micropapillary patterns is also critical for clinical decision making. We aimed to evaluate the accuracy of deep learning models on classifying invasiveness degree and attempted to predict the micropapillary pattern in lung adenocarcinoma. Methods: The records of 291 histopathologically confirmed lung adenocarcinoma patients were retrospectively analyzed and consisted of 61 adenocarcinoma in situ, 80 minimally invasive adenocarcinoma, 117 invasive adenocarcinoma, and 33 invasive adenocarcinoma with micropapillary components (>5%). We constructed two diagnostic models, the Lung-DL model and the Dense model, based on the LeNet and the DenseNet architecture, respectively. Results: For distinguishing the nodule invasiveness degree, the area under the curve (AUC) value of the diagnosis with the Lung-DL model is 0.88 and that with the Dense model is 0.86. In the prediction of the micropapillary pattern, overall accuracies of 92 and 72.91% were obtained for the Lung-DL model and the Dense model, respectively. Conclusion: Deep learning was successfully used for the invasiveness classification of pulmonary adenocarcinomas. This is also the first time that deep learning techniques have been used to predict micropapillary patterns. Both tasks can increase efficiency and assist in the creation of precise individualized treatment plans.

Consecutive Serial Non-Contrast CT Scan-Based Deep Learning Model Facilitates the Prediction of Tumor Invasiveness of Ground-Glass Nodules

“…Zhao et al reported a DL system based on 3D CNNs, and multitask learning, which achieved better classification performance than senior and junior doctors in pathological labeling of GGNs (41). Moreover, Ding et al applied two models for distinguishing degree of nodule invasiveness, the lung DL model and dense model; both modes showed high performance in terms of AUC (0.88 and 0.86, respectively), especially the lung DL model (42).…”

Enhanced CT-Based Radiomics to Predict Micropapillary Pattern Within Lung Invasive Adenocarcinoma

“…Non-small cell lung cancer (NSCLC) accounts for approximately 85% of all cases of lung cancer, and with the most common histological subtype of NSCLC being adenocarcinoma (1). Currently, surgical resection is the main treatment for lung adenocarcinoma (2).…”

Deep Learning Analysis of CT Images Reveals High-Grade Pathological Features to Predict Survival in Lung Adenocarcinoma

“…As a study that built a deep learning-based model similar to our study, Ding et al developed two different deep learning models, based on the LeNet and the DenseNet architecture for invasiveness classification and prediction of micropapillary pattern (MP) within lung ADC. They reported that the accuracies of each model for MP prediction were 92% and 73%, respectively [18]. However, there are limitations in that only the MP pattern was predicted, excluding the solid pattern, and the model was not validated in other cohorts.…”

“…However, few studies have used deep learning to classify histologic patterns in lung ADCs, which is difficult to perform with human eyes. Ding et al trained two different deep learning models to predict a micropapillary pattern using LeNet and DenseNet, and showed an overall accuracy of 92 and 72.9% [18]. Wang et al reported that combined radiomics and a deep learning model for the prediction of high-grade patterns in lung ADCs manifested as ground-glass opacity nodule (GGN) and resulted in an overall accuracy of 91.3% [19].…”