Combined model integrating deep learning, radiomics, and clinical data to classify lung nodules at chest CT.

OBJECTIVES: The study aimed to develop a combined model that integrates deep learning (DL), radiomics, and clinical data to classify lung nodules into benign or malignant categories, and to further classify lung nodules into different pathological subtypes and Lung Imaging Reporting and Data System (Lung-RADS) scores. MATERIALS AND METHODS: The proposed model was trained, validated, and tested using three datasets: one public dataset, the Lung Nodule Analysis 2016 (LUNA16) Grand challenge dataset (n = 1004), and two private datasets, the Lung Nodule Received Operation (LNOP) dataset (n = 1027) and the Lung Nodule in Health Examination (LNHE) dataset (n = 1525). The proposed model used a stacked ensemble model by employing a machine learning (ML) approach with an AutoGluon-Tabular classifier. The input variables were modified 3D convolutional neural network (CNN) features, radiomics features, and clinical features. Three classification tasks were performed: Task 1: Classification of lung nodules into benign or malignant in the LUNA16 dataset; Task 2: Classification of lung nodules into different pathological subtypes; and Task 3: Classification of Lung-RADS score. Classification performance was determined based on accuracy, recall, precision, and F1-score. Ten-fold cross-validation was applied to each task. RESULTS: The proposed model achieved high accuracy in classifying lung nodules into benign or malignant categories in LUNA 16 with an accuracy of 92.8%, as well as in classifying lung nodules into different pathological subtypes with an F1-score of 75.5% and Lung-RADS scores with an F1-score of 80.4%. CONCLUSION: Our proposed model provides an accurate classification of lung nodules based on the benign/malignant, different pathological subtypes, and Lung-RADS system.

Improving puncture accuracy in percutaneous CT-guided needle insertion with wireless inertial measurement unit: a phantom study.

OBJECTIVES: A novel method applying inertial measurement units (IMUs) was developed to assist CT-guided puncture, which enables real-time displays of planned and actual needle trajectories. The method was compared with freehand and laser protractor-assisted methods. METHODS: The phantom study was performed by three operators with 8, 2, and 0 years of experience in CT-guided procedure conducted five consecutive needle placements for three target groups using three methods (freehand, laser protractor-assisted, or IMU-assisted method). The endpoints included mediolateral angle error and caudocranial angle error of the first pass, the procedure time, the total number of needle passes, and the radiation dose. RESULTS: There was a significant difference in the number of needle passes (IMU 1.2 ± 0.42, laser protractor 2.9 ± 1.6, freehand 3.6 ± 2.0 time, p < 0.001), the procedure time (IMU 3.0 ± 1.2, laser protractor 6.4 ± 2.9, freehand 6.2 ± 3.1 min, p < 0.001), the mediolateral angle error of the first pass (IMU 1.4 ± 1.2, laser protractor 1.6 ± 1.3, freehand 3.7 ± 2.5 degree, p < 0.001), the caudocranial angle error of the first pass (IMU 1.2 ± 1.2, laser protractor 5.3 ± 4.7, freehand 3.9 ± 3.1 degree, p < 0.001), and the radiation dose (IMU 250.5 ± 74.1, laser protractor 484.6 ± 260.2, freehand 561.4 ± 339.8 mGy-cm, p < 0.001) among three CT-guided needle insertion methods. CONCLUSION: The wireless IMU improves the angle accuracy and speed of CT-guided needle punctures as compared with laser protractor guidance and freehand techniques. KEY POINTS: • The IMU-assisted method showed a significant decrease in the number of needle passes (IMU 1.2 ± 0.42, laser protractor 2.9 ± 1.6, freehand 3.6 ± 2.0 time, p < 0.001). • The IMU-assisted method showed a significant decrease in the procedure time (IMU 3.0 ± 1.2, laser protractor 6.4 ± 2.9, freehand 6.2 ± 3.1 min, p < 0.001). • The IMU-assisted method showed a significant decrease in the mediolateral angle error of the first pass and the caudocranial angle error of the first pass.

Partial Anomalous Pulmonary Venous Return.

Online supplemental material is available for this article.

Algorithmic Approach Using Negative Pressure Wound Therapy Improved Survival for Patients with Synchronous Hypopharyngeal and Esophageal Cancer Undergoing Pharyngolaryngoesophagectomy with Gastric Tube Reconstruction.

BACKGROUND: This study retrospectively analyzed the feasibility and surgical outcome of an algorithmic approach using negative pressure wound therapy for patients with synchronous hypopharyngeal and esophageal cancer undergoing pharyngolaryngoesophagectomy with gastric tube reconstruction. METHODS: Patients undergoing pharyngolaryngoesophagectomy and gastric tube reconstruction for hypopharyngeal cancer between 2011 and 2019 were candidates for this study. Data were collected on patient demographics, comorbidities, performance status, cancer stage, treatment, complication, and survival. Survival analysis was performed using the Kaplan-Meier method. The Cox proportional hazards model was used for prognostic factors. RESULTS: The study enrolled 43 patients. Anastomotic leakage was found in 21 of the patients with a conventional surgical drain (61.9%) and in 10 of the 22 patients with negative pressure wound therapy (45.5%) (p = 0.280). Nine patients in the conventional drain group (42.9%) and two patients in the negative pressure wound therapy group (9.1%) had leakage-associated complications (p = 0.011). The incidence of pulmonary complications was higher in the conventional surgical drain group (9 vs 2; p = 0.011). The number of complications requiring surgery was higher in the conventional drain group (7 vs 0; p = 0.004). The overall survival in the negative pressure wound therapy group was better (hazard ratio [HR], 0.33; 95% confidence interval [CI], 0.15-0.76; p = 0.009). Negative pressure wound therapy was independently associated with overall survival (HR, 0.31; 95% CI, 0.13-0.77; p = 0.011). CONCLUSIONS: Negative pressure wound therapy with an algorithmic approach improved the overall survival for the patients undergoing gastric tube reconstruction after pharyngolaryngoesophagectomy for hypopharyngeal and esophageal cancer by preventing deadly complications secondary to anastomotic leakage.

Single-port thoracoscopic anatomic resection for chronic inflammatory lung disease.

BACKGROUND: It is challenging to proceed thoracoscopic anatomic resection when encountering severe pleural adhesion or calcified peribronchial lymphadenopathy. Compared with multiple-port video-assisted thoracoscopic surgery (MP-VATS), how to overcome these challenges in single-port (SP-) VATS is still an intractable problem. In the present study, we reported the surgical results of chronic inflammatory lung disease and shared some useful SP-VATS techniques. METHODS: We retrospectively assessed the surgical results of chronic inflammatory lung disease, primarily bronchiectasis, and mycobacterial infection, at our institution between 2010 and 2018. The patients who underwent SP-VATS anatomic resection were compared with those who underwent MP-VATS procedures. We analyzed the baseline characteristics, perioperative data, and postoperative outcomes, and illustrated four special techniques depending on the situation: flexible hook electrocautery, hilum-first technique, application of Satinsky vascular clamp, and staged closure of bronchial stump method. RESULTS: We classified 170 consecutive patients undergoing thoracoscopic anatomic resection into SP and MP groups, which had significant between-group differences in operation time and overall complication rate (P = 0.037 and 0.018, respectively). Compared to the MP-VATS group, the operation time of SP-VATS was shorter, and the conversion rate of SP-VATS was relatively lower (3.1% vs. 10.5%, P = 0.135). The most common complication was prolonged air leakage (SP-VATS, 10.8%; MP-VATS, 2.9%, P = 0.045). CONCLUSIONS: For chronic inflammatory lung disease, certain surgical techniques render SP-VATS anatomic resection feasible and safe with a lower conversion rate.

EGFR-TKI plus bevacizumab versus EGFR-TKI monotherapy for patients with EGFR mutation-positive advanced non-small cell lung cancer-A propensity score matching analysis.

BACKGROUND: Recent study showed that the combination of erlotinib and bevacizumab had better disease control than erlotinib monotherapy in patients with advanced epidermal growth factor receptor (EGFR)-mutant non-small cell lung cancer (NSCLC). However, there is lack of real-world evidence for this therapeutic regimen. We aimed to compare outcomes between patients with EGFR mutant NSCLC treated with EGFR-tyrosine kinase inhibitors (TKI) and bevacizumab and those treated with EGFR-TKI alone in a real-world setting. METHODS: Patients with advanced EGFR-mutant NSCLC who received first-line EGFR-TKI in a tertiary referral center from October 1, 2013 to December 31, 2019 were retrospectively analyzed. We performed 1:2 propensity score-matching: one EGFR-TKI and bevacizumab recipient with two patients who received EGFR-TKI alone. Progression-free survival (PFS) and overall survival (OS) were evaluated using the Kaplan-Meier method. The prognostic factors were analyzed using Cox proportional hazards regression analysis. RESULTS: Total 313 patients were enrolled. After propensity score matching, 45 patients who received first-line EGFR-TKI and bevacizumab and 89 patients who received EGFR-TKI alone were analyzed. The combination group showed improved PFS (17.0 vs. 11.0 months; hazard ratio [HR] = 0.48; p = 0.002) compared to the monotherapy group. In subgroup analysis of patients with an L858R mutation, the combination group showed longer PFS (23.1 vs. 10.7 months; HR = 0.40; p = 0.011) and OS (not reached vs. 40.6 months; HR = 0.27; p = 0.040) than the EGFR-TKI monotherapy group. CONCLUSION: Our data suggest that the combination of EGFR-TKI and bevacizumab could improve PFS in patients with EGFR-mutant NSCLC. In patients harboring L858R mutation, the combination therapy provides better OS than TKI alone.

Development of a real-time RGB-D visual feedback-assisted pulmonary rehabilitation system.

Background: Following surgery, perioperative pulmonary rehabilitation (PR) is important for patients with early-stage lung cancer. However, current inpatient programs are often limited in time and space, and outpatient settings have access barriers. Therefore, we aimed to develop a background-free, zero-contact thoracoabdominal movement-tracking model that is easily set up and incorporated into a pre-existing PR program or extended to home-based rehabilitation and remote monitoring. We validated its effectiveness in providing preclinical real-time RGB-D (colour-depth camera) visual feedback. Methods: Twelve healthy volunteers performed deep breathing exercises following audio instruction for three cycles, followed by audio instruction and real-time visual feedback for another three cycles. In the visual feedback system, we used a RealSense™ D415 camera to capture RGB and depth images for human pose-estimation with Google MediaPipe. Target-tracking regions were defined based on the relative position of detected joints. The processed depth information of the tracking regions was visualised on a screen as a motion bar to provide real-time visual feedback of breathing intensity. Pulmonary function was simultaneously recorded using spirometric measurements, and changes in pulmonary volume were derived from respiratory airflow signals. Results: Our movement-tracking model showed a very strong correlation (r = 0.90 ± 0.05) between thoracic motion signals and spirometric volume, and a strong correlation (r = 0.73 ± 0.22) between abdominal signals and spirometric volume. Displacement of the chest wall was enhanced by RGB-D visual feedback (23 vs 20 mm, P = 0.034), and accompanied by an increased lung volume (2.58 vs 2.30 L, P = 0.003). Conclusion: We developed an easily implemented thoracoabdominal movement-tracking model and reported the positive impact of real-time RGB-D visual feedback on self-promoted external chest wall expansion, accompanied by increased internal lung volumes. This system can be extended to home-based PR.

Therapeutic Decision Making in Prevascular Mediastinal Tumors Using CT Radiomics and Clinical Features: Upfront Surgery or Pretreatment Needle Biopsy?

The study aimed to develop machine learning (ML) classification models for differentiating patients who needed direct surgery from patients who needed core needle biopsy among patients with prevascular mediastinal tumor (PMT). Patients with PMT who received a contrast-enhanced computed tomography (CECT) scan and initial management for PMT between January 2010 and December 2020 were included in this retrospective study. Fourteen ML algorithms were used to construct candidate classification models via the voting ensemble approach, based on preoperative clinical data and radiomic features extracted from the CECT. The classification accuracy of clinical diagnosis was 86.1%. The first ensemble learning model was built by randomly choosing seven ML models from a set of fourteen ML models and had a classification accuracy of 88.0% (95% CI = 85.8 to 90.3%). The second ensemble learning model was the combination of five ML models, including NeuralNetFastAI, NeuralNetTorch, RandomForest with Entropy, RandomForest with Gini, and XGBoost, and had a classification accuracy of 90.4% (95% CI = 87.9 to 93.0%), which significantly outperformed clinical diagnosis (p < 0.05). Due to the superior performance, the voting ensemble learning clinical-radiomic classification model may be used as a clinical decision support system to facilitate the selection of the initial management of PMT.

Development of a modified 3D region proposal network for lung nodule detection in computed tomography scans: a secondary analysis of lung nodule datasets.

BACKGROUND: Low-dose computed tomography (LDCT) has been shown useful in early lung cancer detection. This study aimed to develop a novel deep learning model for detecting pulmonary nodules on chest LDCT images. METHODS: In this secondary analysis, three lung nodule datasets, including Lung Nodule Analysis 2016 (LUNA16), Lung Nodule Received Operation (LNOP), and Lung Nodule in Health Examination (LNHE), were used to train and test deep learning models. The 3D region proposal network (RPN) was modified via a series of pruning experiments for better predictive performance. The performance of each modified deep leaning model was evaluated based on sensitivity and competition performance metric (CPM). Furthermore, the performance of the modified 3D RPN trained on three datasets was evaluated by 10-fold cross validation. Temporal validation was conducted to assess the reliability of the modified 3D RPN for detecting lung nodules. RESULTS: The results of pruning experiments indicated that the modified 3D RPN composed of the Cross Stage Partial Network (CSPNet) approach to Residual Network (ResNet) Xt (CSP-ResNeXt) module, feature pyramid network (FPN), nearest anchor method, and post-processing masking, had the optimal predictive performance with a CPM of 92.2%. The modified 3D RPN trained on the LUNA16 dataset had the highest CPM (90.1%), followed by the LNOP dataset (CPM: 74.1%) and the LNHE dataset (CPM: 70.2%). When the modified 3D RPN trained and tested on the same datasets, the sensitivities were 94.6%, 84.8%, and 79.7% for LUNA16, LNOP, and LNHE, respectively. The temporal validation analysis revealed that the modified 3D RPN tested on LNOP test set achieved a CPM of 71.6% and a sensitivity of 85.7%, and the modified 3D RPN tested on LNHE test set had a CPM of 71.7% and a sensitivity of 83.5%. CONCLUSION: A modified 3D RPN for detecting lung nodules on LDCT scans was designed and validated, which may serve as a computer-aided diagnosis system to facilitate lung nodule detection and lung cancer diagnosis.

A modified 3D RPN for detecting lung nodules on CT images that exhibited greater sensitivity and CPM than did several previously reported CAD detection models was established.

2024 multidisciplinary consensus on image-guided lung tumor ablation from the Taiwan Academy of Tumor Ablation.

In this article, the multidisciplinary team of the Taiwan Academy of Tumor Ablation, who have expertise in treating lung cancer, present their perspectives on percutaneous image-guided thermal ablation (IGTA) of lung tumors. The modified Delphi technique was applied to reach a consensus on clinical practice guidelines concerning ablation procedures, including a comprehensive literature review, selection of panelists, creation of a rating form and survey, and arrangement of an in-person meeting where panelists agreed or disagreed on various points. The conclusion was a final rating and written summary of the agreement. The multidisciplinary expert team agreed on 10 recommendations for the use of IGTA in the lungs. These recommendations include terms and definitions, line of treatment planning, modality, facility rooms, patient anesthesia settings, indications, margin determination, post-ablation image surveillance, qualified centers, and complication ranges. In summary, IGTA is a safe and feasible approach for treating primary and metastatic lung tumors, with a relatively low complication rate. However, decisions regarding the ablation technique should consider each patient's specific tumor characteristics.

Early drainage reduces the length of hospital stay in patients with lung abscess.

Background: Although percutaneous transthoracic catheter drainage (PCD) has been proven effective in lung abscesses, the optimal timing of PCD is still unclear. The study aimed to evaluate the safety and efficacy of early versus delayed drainage in patients with lung abscesses. Methods: This retrospective study included 103 consecutive patients with liquefied lung abscesses more than 3 cm confirmed by a CT scan received CT-guided PCD over 16 years, from July 2005 to September 2021, in a single institution were reviewed. Early drainage was defined as PCD within one week after a lung abscess was diagnosed. The primary outcome was 90-day mortality. The secondary outcomes included perioperative complications and patients' length of hospital stay (LoS). Factors associated with 90-day mortality and LoS were also analyzed. The key statistical methods were Chi-square test, Fisher's exact test, Student t-test, and Pearson correlation. Results: Amount the 103 patients, there were 64 patients who received early PCD, and 39 patients received delayed PCD. Between the two groups, there were no significant differences in clinical characteristics, 90-day mortality, or perioperative complications. The LoS was significantly shortened in early PCD group (28.6 ± 25.5 vs. 39.3 ± 26.8 (days), p = 0.045). Higher Charlson comorbidity index, secondary lung abscess, and liver cirrhosis were associated with higher mortality (all p < 0.05). Positive sputum culture significantly increased the LoS (coefficient 19.35 (10.19, 28.50), p < 0.001). Conclusion: The 90-day mortality and complications were similar for early PCD and delayed PCD patients, but LoS was significantly shortened in early PCD patient.

Pulmonary-function changes after uniportal video-assisted thoracoscopic anatomical lung resection.

OBJECTIVE: The superiority of segmentectomy over lobectomy with regard to preservation of pulmonary function is controversial. This study aimed to examine changes in pulmonary function after uniportal video-assisted thoracoscopic surgery (VATS) according to the number of resected segments. METHODS: We retrospectively reviewed 135 consecutive patients who underwent anatomical lung resection via uniportal VATS from April 2015 to December 2020. Pulmonary function loss was evaluated using forced vital capacity (FVC) and forced expiratory volume in 1 s (FEV1). Patients were grouped according to number of resected segments: one-segment (n = 33), two segments (n = 22), three segments (n = 40), four segments (n = 15), and five segments (n = 25). RESULTS: Clinical characteristics did not significantly differ between groups, except for tumor size. Mean follow-up was 8.96 ± 3.16 months. FVC loss was significantly greater in five-segment resection (10.8%) than one-segment (0.97%, p = 0.008) and two-segment resections (2.44%, p = 0.040). FEV1 loss was significantly greater in five-segment resection (15.02%) than one-segment (3.83%, p < 0.001), two-segment (4.63%, p = 0.001), and three-segment resections (7.63%, p = 0.007). Mean FVC loss and FEV1 loss increased linearly from one-segment resection to five-segment resection. Mean loss in FVC and FEV1 per segment resected was 2.16% and 3.00%, respectively. CONCLUSIONS: Anatomical lung resection of fewer segments was associated with better preservation of pulmonary function in patients undergoing uniportal VATS, and function loss was approximately 2%-3% per segment resected with linear relationship.

Low neutrophil-to-lymphocyte ratio predicts overall survival benefit in advanced NSCLC patients with low PD-L1 expression and receiving chemoimmunotherapy.

Although combination therapy including chemotherapy and immune checkpoint inhibitors (ICIs) improves overall survival (OS) of patients with non-small-cell lung cancer (NSCLC), there is a higher incidence of adverse events and treatment discontinuation. Since programmed death-ligand 1 (PD-L1) could not serve as a predictive biomarker, we investigated the neutrophil-to-lymphocyte ratio (NLR) as a predictive biomarker. In our previous research, we demonstrated that a low NLR could predict survival benefits when patients with high PD-L1 expression (> 50%) received chemoimmunotherapy as opposed to immunotherapy alone. In this current study, our objective is to evaluate this predictive capacity in patients with low PD-L1 expression (< 50%). A total of 142 patients were enrolled, 28 receiving combination therapy and 114 receiving chemotherapy alone. Progression-free survival (PFS) and OS were estimated using the Kaplan-Meier method and compared using the log-rank test. Patients who received combination therapy had significantly better PFS and OS than those who received monotherapy. In the subgroup of patients with low NLR, those who received combination therapy exhibited extended PFS and OS with clinical significance, which was also confirmed by multivariate Cox regression analysis. Our study demonstrates the potential use of NLR as a biomarker for predicting survival benefits when receiving combination therapy with chemotherapy and ICIs in patients with advanced NSCLC and low PD-L1 expression.

Corrigendum: Clinical radiomics-based machine learning versus three-dimension convolutional neural network analysis for differentiation of thymic epithelial tumors from other prevascular mediastinal tumors on chest computed tomography scan.

[This corrects the article DOI: 10.3389/fonc.2023.1105100.].

Chemotherapy outcomes in EGFR-TKI resistant patients with common and uncommon EGFR mutation: An exploratory retrospective cohort study.

BACKGROUND: Some prospective studies have shown that second-generation tyrosine kinase inhibitors (TKIs) provide better control in patients with non-small cell lung cancer (NSCLC) with uncommon epidermal growth factor receptor (EGFR) mutations. However, studies comparing second-line chemotherapy efficacy between NSCLC patients with common and uncommon EGFR mutations remain rare. This retrospective study compared treatment outcomes in these patients. METHODS: Patients with EGFR-mutated advanced-stage NSCLC who received first-line EGFR-TKIs in a tertiary referral center were retrospectively reviewed between January 2010 and August 2022. Patients with a negative T790M test at disease progression who received second-line chemotherapy were enrolled. We compared progression-free (PFS) and overall (OS) survival between advanced NSCLC patients with common and uncommon EGFR mutations using Kaplan-Meier and log-rank tests. RESULTS: In total, 209 (54.8%) patients had a negative T790M mutation test and received second-line chemotherapy, of which 192 (91.8%) had a common EGFR mutation (exon 19 deletion or exon 21 L858R substitution), and 17 (8.2%) had an uncommon EGFR mutation. Patients with common EGFR mutations had significantly longer PFS than those with uncommon EGFR mutations (4.57 vs. 2.57 months, p = 0.031). A Cox proportional hazard regression analysis controlling for potential confounding factors indicated that an uncommon EGFR mutation was an independent prognostic factor for PFS. CONCLUSION: This study suggests that patients with uncommon EGFR mutations have poorer chemotherapy responses and shorter survival than those with common EGFR mutations. The development of new treatment strategies for these patients remains an unmet need.

Clinical radiomics-based machine learning versus three-dimension convolutional neural network analysis for differentiation of thymic epithelial tumors from other prevascular mediastinal tumors on chest computed tomography scan.

Purpose: To compare the diagnostic performance of radiomic analysis with machine learning (ML) model with a convolutional neural network (CNN) in differentiating thymic epithelial tumors (TETs) from other prevascular mediastinal tumors (PMTs). Methods: A retrospective study was performed in patients with PMTs and undergoing surgical resection or biopsy in National Cheng Kung University Hospital, Tainan, Taiwan, E-Da Hospital, Kaohsiung, Taiwan, and Kaohsiung Veterans General Hospital, Kaohsiung, Taiwan between January 2010 and December 2019. Clinical data including age, sex, myasthenia gravis (MG) symptoms and pathologic diagnosis were collected. The datasets were divided into UECT (unenhanced computed tomography) and CECT (enhanced computed tomography) for analysis and modelling. Radiomics model and 3D CNN model were used to differentiate TETs from non-TET PMTs (including cyst, malignant germ cell tumor, lymphoma and teratoma). The macro F1-score and receiver operating characteristic (ROC) analysis were performed to evaluate the prediction models. Result: In the UECT dataset, there were 297 patients with TETs and 79 patients with other PMTs. The performance of radiomic analysis with machine learning model using LightGBM with Extra Tree (macro F1-Score = 83.95%, ROC-AUC = 0.9117) had better performance than the 3D CNN model (macro F1-score = 75.54%, ROC-AUC = 0.9015). In the CECT dataset, there were 296 patients with TETs and 77 patients with other PMTs. The performance of radiomic analysis with machine learning model using LightGBM with Extra Tree (macro F1-Score = 85.65%, ROC-AUC = 0.9464) had better performance than the 3D CNN model (macro F1-score = 81.01%, ROC-AUC = 0.9275). Conclusion: Our study revealed that the individualized prediction model integrating clinical information and radiomic features using machine learning demonstrated better predictive performance in the differentiation of TETs from other PMTs at chest CT scan than 3D CNN model.

Effect of BIM expression on the prognostic value of PD-L1 in advanced non-small cell lung cancer patients treated with EGFR-TKIs.

The role of Programmed Cell Death Ligand 1 (PD-L1) expression in predicting epidermal growth factor receptor tyrosine kinase inhibitor (EGFR-TKIs) efficacy remains controversial. Recent studies have highlighted that tumor-intrinsic PD-L1 signaling can be modulated by STAT3, AKT, MET oncogenic pathway, epithelial-mesenchymal transition, or BIM expression. This study aimed to investigate whether these underlying mechanisms affect the prognostic role of PD-L1. We retrospectively enrolled patients with EGFR mutant advanced stage NSCLC who received first-line EGFR-TKI between January 2017 and June 2019, the treatment efficacy of EGFR-TKI was assessed. Kaplan-Meier analysis of progression-free survival (PFS) revealed that patients with high BIM expression had shorter PFS, regardless of PD-L1 expression. This result was also supported by the COX proportional hazard regression analysis. In vitro, we further proved that the knockdown of BIM, instead of PDL1, induced more cell apoptosis following gefitinib treatment. Our data suggest that among the pathways affecting tumor-intrinsic PD-L1 signaling, BIM is potentially the underlying mechanism that affects the role of PD-L1 expression in predicting response to EGFR TKI and mediates cell apoptosis under treatment with gefitinib in EGFR-mutant NSCLC. Further prospective studies are required to validate these results.