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

Abstract

We aimed to develop a deep learning (DL) model for predicting high-grade patterns in lung adenocarcinomas (ADC) and to assess the prognostic performance of model in advanced lung cancer patients who underwent neoadjuvant or definitive concurrent chemoradiation therapy (CCRT). We included 275 patients with 290 early lung ADCs from an ongoing prospective clinical trial in the training dataset, which we split into internal–training and internal–validation datasets. We constructed a diagnostic DL model of high-grade patterns of lung ADC considering both morphologic view of the tumor and context view of the area surrounding the tumor (MC3DN; morphologic-view context-view 3D network). Validation was performed on an independent dataset of 417 patients with advanced non-small cell lung cancer who underwent neoadjuvant or definitive CCRT. The area under the curve value of the DL model was 0.8 for the prediction of high-grade histologic patterns such as micropapillary and solid patterns (MPSol). When our model was applied to the validation set, a high probability of MPSol was associated with worse overall survival (probability of MPSol >0.5 vs. <0.5; 5-year OS rate 56.1% vs. 70.7%), indicating that our model could predict the clinical outcomes of advanced lung cancer patients. The subgroup with a high probability of MPSol estimated by the DL model showed a 1.76-fold higher risk of death (HR 1.76, 95% CI 1.16–2.68). Our DL model can be useful in estimating high-grade histologic patterns in lung ADCs and predicting clinical outcomes of patients with advanced lung cancer who underwent neoadjuvant or definitive CCRT.

Establishment and validation of a radiological-radiomics model for predicting high-grade patterns of lung adenocarcinoma less than or equal to 3 cm

“…In this study, we constructed an R-R model on the strength of radiological features integrated with radiomics features to predict the HGP of LUAD, which demonstrated an excellent predictive performance superior to previously reported omics-only models ( 18 , 19 ). The AUC value of our R-R model in the training set was 0.923, corresponding to a sensitivity of 87.0%, a specificity of 83.4%, and an accuracy of 84.2%.…”

“…Our study included GGO and solid nodules, which is more aligned with the actual clinical conditions. Yeonu Choi and He et al have developed a radiomics model to predict the micropapillary and solid components of LUAD but only radiomics signatures are contained in this model (19,20). This study is the first report, as far as we know, that HGP can be predicted by a model basing on radiological features integrated with radiomics features in a relatively large dataset.…”

“…It is also a non-invasive, quantitative method for tumor heterogeneity assessment that can be applied to quantify intratumoral heterogeneity. Recent studies have employed radiomics-based methods to predict high-grade components of LUAD (18)(19)(20), whereas, the present study used a larger sample size and added traditional radiological features to develop a model for comparison, and combined with radiomics features, constructed an integrated model to predict the HGP of LUAD. In this study, we aimed to develop and evaluate the value of a radiological-radiomics model for predicting the HGP in LUAD less than or equal to 3 cm.…”

Preoperative CT-Based Radiomic Prognostic Index to Predict the Benefit of Postoperative Radiotherapy in Patients with Non-Small Cell Lung Cancer: A Multicenter Study

“…The RPI may embody the characteristics of pT and histological information. Previous studies have shown that radiomic features can predict pT [28], histology [29], and grade [30] in NSCLC. PLN remained a signi cant variable in the multivariate analysis.…”

“…In predicting OS, our model outperformed the pT stage and clinical model, including the pT and PLN. Previous studies have shown that radiomic features can predict pT stage [28], histology [29], pathological grade [30], and lymph vesicular invasion [37] in NSCLC. The RPI could represent more information than the clinicopathological variables; thus, it would be much more convenient to use the RPI as it can be fully automatically calculated from the CT image.…”