Predicting Tyrosine Kinase Inhibitor Treatment Response in Stage IV Lung Adenocarcinoma Patients With EGFR Mutation Using Model-Based Deep Transfer Learning.

BACKGROUND: For stage IV patients harboring EGFR mutations, there is a differential response to the first-line TKI treatment. We constructed three-dimensional convolutional neural networks (CNN) with deep transfer learning to stratify patients into subgroups with different response and progression risks. MATERIALS AND METHODS: From 2013 to 2017, 339 patients with EGFR mutation receiving first-line TKI treatment were included. Progression-free survival (PFS) time and progression patterns were confirmed by routine follow-up and restaging examinations. Patients were divided into two subgroups according to the median PFS (<=9 months, > 9 months). We developed a PFS prediction model and a progression pattern classification model using transfer learning from a pre-trained EGFR mutation classification 3D CNN. Clinical features were fused with the 3D CNN to build the final hybrid prediction model. The performance was quantified using area under receiver operating characteristic curve (AUC), and model performance was compared by AUCs with Delong test. RESULTS: The PFS prediction CNN showed an AUC of 0.744 (95% CI, 0.645-0.843) in the independent validation set and the hybrid model of CNNs and clinical features showed an AUC of 0.771 (95% CI, 0.676-0.866), which are significantly better than clinical features-based model (AUC, 0.624, P<0.01). The progression pattern prediction model showed an AUC of 0.762(95% CI, 0.643-0.882) and the hybrid model with clinical features showed an AUC of 0.794 (95% CI, 0.681-0.908), which can provide compensate information for clinical features-based model (AUC, 0.710; 95% CI, 0.582-0.839). CONCLUSION: The CNN exhibits potential ability to stratify progression status in patients with EGFR mutation treated with first-line TKI, which might help make clinical decisions.

Predicting Malignancy and Invasiveness of Pulmonary Subsolid Nodules on CT Images Using Deep Learning.

BACKGROUND: To develop and validate a deep learning-based model on CT images for the malignancy and invasiveness prediction of pulmonary subsolid nodules (SSNs). MATERIALS AND METHODS: This study retrospectively collected patients with pulmonary SSNs treated by surgery in our hospital from 2012 to 2018. Postoperative pathology was used as the diagnostic reference standard. Three-dimensional convolutional neural network (3D CNN) models were constructed using preoperative CT images to predict the malignancy and invasiveness of SSNs. Then, an observer reader study conducted by two thoracic radiologists was used to compare with the CNN model. The diagnostic power of the models was evaluated with receiver operating characteristic curve (ROC) analysis. RESULTS: A total of 2,614 patients were finally included and randomly divided for training (60.9%), validation (19.1%), and testing (20%). For the benign and malignant classification, the best 3D CNN model achieved a satisfactory AUC of 0.913 (95% CI: 0.885-0.940), sensitivity of 86.1%, and specificity of 83.8% at the optimal decision point, which outperformed all observer readers' performance (AUC: 0.846±0.031). For pre-invasive and invasive classification of malignant SSNs, the 3D CNN also achieved satisfactory AUC of 0.908 (95% CI: 0.877-0.939), sensitivity of 87.4%, and specificity of 80.8%. CONCLUSION: The deep-learning model showed its potential to accurately identify the malignancy and invasiveness of SSNs and thus can help surgeons make treatment decisions.

lncRNA MRUL Suppressed Non-Small Cell Lung Cancer Cells Proliferation and Invasion by Targeting miR-17-5p/SRSF2 Axis.

The two broad histological subtypes of lung cancer are small cell lung cancer (SCLC) and non-small cell lung cancer (NSCLC), which are the leading causes of cancer-related death in the world. Long noncoding RNAs (lncRNAs) have been verified to be critical in the regulation of cancer development. The present study identified and elucidated the regulatory roles of a novel lncRNA MRUL in NSCLC. The results showed that MRUL was overexpressed in NSCLC samples and correlated with the poor prognosis of patients who had NSCLC. Moreover, this research has for the first time demonstrated that MRUL acted as an oncogenetic lncRNA in NSCLC. Knockdown of MRUL considerably repressed NSCLC cell proliferation, invasion, and migration. The bioinformatics analysis showed that MRUL was involved in regulating multiple RNA splicing and proliferation-related biological processes, such as mRNA splicing, RNA splicing, mRNA processing, mRNA 3'-end processing, mRNA splice site selection, and DNA replication. By combining bioinformatics analysis and experimental validation, we found that MRUL regulated NSCLC progression through promoting SRSF2 by sponging miR-17 in NSCLC cells. The discoveries indicated that MRUL could be a therapeutic target and a potential diagnostic for NSCLC.

HAGLROS is overexpressed and promotes non-small cell lung cancer migration and invasion.

INTRODUCTION: Non-small cell lung cancer was one of the most common and deadly cancers worldwide. Long non-coding RNAs had been implicated in multiple human cancers, including non-small cell lung cancer. In this study, we focused on a novel long non-coding RNA, HAGLROS, in non-small cell lung cancer. MATERIAL AND METHODS: In this study, we used GEPIA dataset to analyse the expression levels of HAGLROS in non-small cell lung cancer samples and normal tissues. Then, we analysed Kaplan-Meier Plotter database to reveal the association between HAGLROS expression and overall survival time in patients with non-small cell lung cancer. Moreover, we used small interfering RNA-mediated knockdown to reduce HAGLROS expression in A549 and H1299 cells. Cell Counting Kit-8 assay was used to detect the effect of HAGLROS on cell proliferation. Transwell assays were used to determine the effect of HAGLROS on cell migration and invasion. Co-expression analysis and bioinformatics analysis were conducted to predict the potential functions of HAGLROS in non-small cell lung cancer. RESULTS: We identified HAGLROS was significantly overexpressed in non-small cell lung cancer samples compared to normal tissues. Higher expression of HAGLROS was significantly associated with shorter overall survival time in patients with non-small cell lung cancer. Moreover, we found knockdown of HAGLROS in non-small cell lung cancer cells remarkably suppressed tumour proliferation, migration and invasion. By conducting bioinformatics analysis, we found HAGLROS was involved in regulating multiple cancer-related pathways, including Spliceosome, DNA replication, cell cycle, chromosome segregation and sister chromatid segregation. CONCLUSIONS: Our results for the first time demonstrated HAGLROS may serve as a target for new therapies in non-small cell lung cancer.

Analysis of Progression Patterns and Failure Sites of Patients With Metastatic Lung Adenocarcinoma With EGFR Mutations Receiving First-line Treatment of Tyrosine Kinase Inhibitors.

BACKGROUND: Reliable prediction of progression patterns and failure sites for patients with stage IV lung adenocarcinoma is valuable for physicians to deliver personalized tyrosine kinase inhibitor (TKI) treatment. PATIENTS AND METHODS: We retrospectively enrolled 266 patients who had stage IV lung adenocarcinoma and received first-line TKI treatment from 2013 to 2017 in Shanghai Chest Hospital. The clinical characteristics at initial diagnosis, progression patterns, and failure sites were analyzed with the attempt to identify some predictive factors for progression patterns and failure sites. RESULTS: Among all patients, 62.4% developed systemic progression, and 37.6% developed oligoprogression. Both cohorts had a median progression-free survival (PFS) of 9 months. The percentage of patients who developed original and distant failure was 39.1% and 60.9%, respectively. Patients with oligometastasis at initial diagnosis were more prone to develop oligoprogression (odds ratio [OR], 4.370; 95% confidence interval [CI], 1.881-10.151; P = .001), whereas pulmonary metastasis was negatively correlated with oligoprogression (OR, 0.567; 95% CI, 0.330-0.974; P = .04). Both oligometastasis diagnosis (OR, 2.959; 95% CI, 1.347-6.500; P = .007) and the maximum diameter of the primary lung lesion (threshold 3.25 cm: OR, 3.646; 95% CI, 2.041-6.515; P = .0001) were strong predictive factors for original failures. Osseous metastasis at initial diagnosis might be an indication for distant failure (OR, 0.536; 95% CI, 0.316-0.909; P = .021). CONCLUSION: Over one-half of patients with stage IV lung adenocarcinoma receiving first-line TKI treatment developed systemic progression and distant failure. Metastasis patterns at initial diagnosis was the most important predictive factor for progression patterns and failure sites. The maximum diameter of the primary lung lesion and evidence of osseous metastasis were also found to be significant indicative factors for failure sites.

Association Between RET Fusions and Efficacy of Pemetrexed-based Chemotherapy for Patients With Advanced NSCLC in China: A Multicenter Retrospective Study.

BACKGROUND: Rearranged during transfection (RET) proto-oncogene gene fusions are rare in non-small-cell lung cancer (NSCLC). We compared the efficacy of pemetrexed-based chemotherapy with other chemotherapy regimens in patients with NSCLC with different RET fusion subtypes. PATIENTS AND METHODS: A retrospective, multicenter study of patients with pathologically confirmed stage IIIB/IV lung adenocarcinomas was conducted. RET rearrangements were detected using next generation sequencing. We analyzed the clinical characteristics of patients with RET-rearranged NSCLC and the efficacy of chemotherapy regimens. We also evaluated the efficacy between groups of patients with and without KIF5B-RET-rearranged lung cancer. RESULTS: We evaluated 62 patients with NSCLC and RET rearrangements, including 41 with KIF5B-RET, 15 with CCDC6-RET, and 6 with other rare fusion subtypes. Of these 62 patients, 50 had stage IIIB/IV. We also evaluated 40 patients with first-line chemotherapy information available. The median progression-free survival was significantly different between those receiving pemetrexed-based chemotherapy and those receiving other chemotherapy regimens (9.2 vs. 5.2 months; P = .007). The median progression-free survival for patients with KIF5B-RET fusion and non-KIF5B-RET fusion was not significantly different statistically (7.8 vs. 11.2 months; P = .847). For second-line chemotherapy, a statistically significant difference was found between the chemotherapy regimens (4.9 vs. 2.8 months; P = .049). Survival follow-up data were available for 38 patients with advanced NSCLC. The median overall survival was 26.4 months. The overall survival of the patients with RET-rearranged NSCLC who had received pemetrexed-based chemotherapy versus no pemetrexed-based chemotherapy was 35.2 versus 22.6 months (P = .052). No difference in survival was observed between the patients with KIF5B-RET and non-KIF5B-RET rearrangements. CONCLUSIONS: Pemetrexed-based treatment should be considered first when selecting the chemotherapy regimen for patients with NSCLC and RET rearrangements.

High incidence of radiation pneumonitis in lung cancer patients with chronic silicosis treated with radiotherapy.

Silica is an independent risk factor for lung cancer in addition to smoking. Chronic silicosis is one of the most common and serious occupational diseases associated with poor prognosis. However, the role of radiotherapy is unclear in patients with chronic silicosis. We conducted a retrospective study to evaluate efficacy and safety in lung cancer patients with chronic silicosis, especially focusing on the incidence of radiation pneumonitis (RP). Lung cancer patients with chronic silicosis who had been treated with radiotherapy from 2005 to 2018 in our hospital were enrolled in this retrospective study. RP was graded according to the National Cancer Institute's Common Terminology Criteria for Adverse Events (CTCAE), version 3.0. Of the 22 patients, ten (45.5%) developed RP ≥2. Two RP-related deaths (9.1%) occurred within 3 months after radiotherapy. Dosimetric factors V5, V10, V15, V20 and mean lung dose (MLD) were significantly higher in patients who had RP >2 (P < 0.05). The median overall survival times in patients with RP ≤2 and RP>2 were 11.5 months and 7.1 months, respectively. Radiotherapy is associated with excessive and fatal pulmonary toxicity in lung cancer patients with chronic silicosis.

The positive role of vitronectin in radiation induced lung toxicity: the in vitro and in vivo mechanism study.

BACKGROUND: Radiation-induced lung toxicity (RILT) is a severe complication of radiotherapy in patients with thoracic tumors. Through proteomics, we have previously identified vitronectin (VTN) as a potential biomarker for patients with lung toxicity of grade ≥ 2 radiation. Herein, we explored the molecular mechanism of VTN in the process of RILT. METHODS: In this study, lentivirus encoding for VTN and VTN-specific siRNA were constructed and transfected into the cultured fibroblasts and C57BL mice. Real-time PCR, western blot and ELISA were used to examine expression of collagens and several potential proteins involved in lung fibrosis. Hematoxylin-eosin and immunohistochemical staining were used to assess the fibrosis scores of lung tissue from mice received irradiation. RESULTS: The expression of VTN was up-regulated by irradiation. The change trend of collagens, TGF-ß expression and p-ERK, p-AKT, and p-JNK expression levels were positively related with VTN mRNA level. Furthermore, overexpression of VTN significantly increased the expression level of α-SMA, as well as the degree of lung fibrosis in mice at 8 and 12 weeks post-irradiation. By contrast, siRNA VTN induced opposite results both in vitro and in vivo. CONCLUSIONS: VTN played a positive role in the lung fibrosis of RILT, possibly through modulation of fibrosis regulatory pathways and up-regulating the expression levels of fibrosis-related genes. Taken together, all the results suggested that VTN had a novel therapeutic potential for the treatment of RILT.

Detection of epithelial growth factor receptor (EGFR) mutations on CT images of patients with lung adenocarcinoma using radiomics and/or multi-level residual convolutionary neural networks.

BACKGROUND: We aim to analyze the ability to detect epithelial growth factor receptor (EGFR) mutations on chest CT images of patients with lung adenocarcinoma using radiomics and/or multi-level residual convolutionary neural networks (MCNNs). METHODS: We retrospectively collected 1,010 consecutive patients in Shanghai Chest Hospital from 2013 to 2017, among which 510 patients were EGFR-mutated and 500 patients were wild-type. The patients were randomly divided into a training set (810 patients) and a validation set (200 patients) according to a balanced distribution of clinical features. The CT images and the corresponding EGFR status measured by Amplification Refractory Mutation System (ARMS) method of the patients in the training set were utilized to construct both a radiomics-based model (MRadiomics) and MCNNs-based model (MMCNNs). The MRadiomics and MMCNNs were combined to build the ModelRadiomics+MCNNs (MRadiomics+MCNNs). Clinical data of gender and smoking history constructed the clinical features-based model (MClinical). MClinical was then added into MRadiomics, MMCNNs, and MRadiomics+MCNNs to establish the ModelRadiomics+Clinical (MRadiomics+Clinical), the ModelMCNNs+Clinical (MMCNNs+Clinical) and the ModelRadiomics+MCNNs+Clinical (MRadiomics+MCNNs+Clinical). All the seven models were tested in the validation set to ascertain whether they were competent to detect EGFR mutations. The detection efficiency of each model was also compared in terms of area under the curve (AUC), sensitivity and specificity. RESULTS: The AUC of the MRadiomics, MMCNNs and MRadiomics+MCNNs to predict EGFR mutations was 0.740, 0.810 and 0.811 respectively. The performance of MMCNNs was better than that of MRadiomics (P=0.0225). The addition of clinical features did not improve the AUC of the MRadiomics (P=0.623), the MMCNNs (P=0.114) and the MRadiomics+MCNNs (P=0.058). The MRadiomics+MCNNs+Clinical demonstrated the highest AUC value of 0.834. The MMCNNs did not demonstrate any inferiority when compared with the MRadiomics+MCNNs (P=0.742) and the MRadiomics+MCNNs+Clinical (P=0.056). CONCLUSIONS: Both of the MRadiomics and the MCNNs could predict EGFR mutations on CT images of patients with lung adenocarcinoma. The MMCNNs outperformed the MRadiomics in the detection of EGFR mutations. The combination of these two models, even added with clinical features, is not significantly more efficient than MMCNNs alone.