The 2023 American Association for Thoracic Surgery (AATS) Expert Consensus Document: Management of subsolid lung nodules.

OBJECTIVE: Lung cancers that present as radiographic subsolid nodules represent a subtype with distinct biological behavior and outcomes. The objective of this document is to review the existing literature and report consensus among a group of multidisciplinary experts, providing specific recommendations for the clinical management of subsolid nodules. METHODS: The American Association for Thoracic Surgery Clinical Practice Standards Committee assembled an international, multidisciplinary expert panel composed of radiologists, pulmonologists, and thoracic surgeons with established expertise in the management of subsolid nodules. A focused literature review was performed with the assistance of a medical librarian. Expert consensus statements were developed with class of recommendation and level of evidence for each of 4 main topics: (1) definitions of subsolid nodules (radiology and pathology), (2) surveillance and diagnosis, (3) surgical interventions, and (4) management of multiple subsolid nodules. Using a modified Delphi method, the statements were evaluated and refined by the entire panel. RESULTS: Consensus was reached on 17 recommendations. These consensus statements reflect updated insights on subsolid nodule management based on the latest literature and current clinical experience, focusing on the correlation between radiologic findings and pathological classifications, individualized subsolid nodule surveillance and surgical strategies, and multimodality therapies for multiple subsolid lung nodules. CONCLUSIONS: Despite the complex nature of the decision-making process in the management of subsolid nodules, consensus on several key recommendations was achieved by this American Association for Thoracic Surgery expert panel. These recommendations, based on evidence and a modified Delphi method, provide guidance for thoracic surgeons and other medical professionals who care for patients with subsolid nodules.

Ten-Year Follow-Up of Lung Cancer Patients with Resected Stage IA Invasive Non-Small Cell Lung Cancer.

OBJECTIVE: The purpose of this study was to assess 10-year follow-up outcomes after surgical resection in patients with stage IA invasive non-small cell lung cancer (NSCLC) based on postoperative pathological diagnosis. METHODS: Patients with stage IA invasive NSCLC who underwent resection between December 2008 and December 2013 were reviewed. Patients were categorized into the pure-ground glass opacity (pGGO), mixed-ground glass opacity (mGGO), and solid groups based on consolidation to tumor ratio (CTR). Postoperative survival and risk of recurrence and developing secondary primary lung cancer were analyzed in each group. RESULTS: Among the 645 stage IA invasive NSCLC, the 10-year overall survival and recurrence-free survival rate was 79.38% and 77.44%, respectively. The 10-year overall survival for pGGO, mGGO, and solid group of patients was 95.08%, 86.21%, and 72.39%, respectively. The respective recurrence-free survival rate was 100%, 89.82%, and 65.83%. Multivariable Cox regression analysis associated tumor size and GGO components with recurrence and younger age, and tumors with GGO components were associated with longer overall survival. The cumulative incidence curve indicated no recurrence of GGO lung cancer ≥ 5 years postoperatively. Our cohort indicated that the number and stations of dissected lymph node did not influence long-term prognosis of IA invasive NSCLC. CONCLUSIONS: Recurrence of invasive stage IA NSCLC with GGO was more prevalent in patients with tumor size >1 cm and CTR > 0.5, occurring within 5 years after surgery. This will provide important evidence for follow-up strategies in these patients.

The Use of Generative Adversarial Network and Graph Convolution Network for Neuroimaging-Based Diagnostic Classification.

Functional connectivity (FC) obtained from resting-state functional magnetic resonance imaging has been integrated with machine learning algorithms to deliver consistent and reliable brain disease classification outcomes. However, in classical learning procedures, custom-built specialized feature selection techniques are typically used to filter out uninformative features from FC patterns to generalize efficiently on the datasets. The ability of convolutional neural networks (CNN) and other deep learning models to extract informative features from data with grid structure (such as images) has led to the surge in popularity of these techniques. However, the designs of many existing CNN models still fail to exploit the relationships between entities of graph-structure data (such as networks). Therefore, graph convolution network (GCN) has been suggested as a means for uncovering the intricate structure of brain network data, which has the potential to substantially improve classification accuracy. Furthermore, overfitting in classifiers can be largely attributed to the limited number of available training samples. Recently, the generative adversarial network (GAN) has been widely used in the medical field for its generative aspect that can generate synthesis images to cope with the problems of data scarcity and patient privacy. In our previous work, GCN and GAN have been designed to investigate FC patterns to perform diagnosis tasks, and their effectiveness has been tested on the ABIDE-I dataset. In this paper, the models will be further applied to FC data derived from more public datasets (ADHD, ABIDE-II, and ADNI) and our in-house dataset (PTSD) to justify their generalization on all types of data. The results of a number of experiments show the powerful characteristic of GAN to mimic FC data to achieve high performance in disease prediction. When employing GAN for data augmentation, the diagnostic accuracy across ADHD-200, ABIDE-II, and ADNI datasets surpasses that of other machine learning models, including results achieved with BrainNetCNN. Specifically, in ADHD, the accuracy increased from 67.74% to 73.96% with GAN, in ABIDE-II from 70.36% to 77.40%, and in ADNI, reaching 52.84% and 88.56% for multiclass and binary classification, respectively. GCN also obtains decent results, with the best accuracy in ADHD datasets at 71.38% for multinomial and 75% for binary classification, respectively, and the second-best accuracy in the ABIDE-II dataset (72.28% and 75.16%, respectively). Both GAN and GCN achieved the highest accuracy for the PTSD dataset, reaching 97.76%. However, there are still some limitations that can be improved. Both methods have many opportunities for the prediction and diagnosis of diseases.

Prediction of the pathological subtypes by intraoperative frozen section for patients with cT1N0M0 invasive lung adenocarcinoma (ECTOP-1015): a prospective multi-center study.

BACKGROUND: The aim of this study is to assess the diagnostic accuracy of intraoperative frozen section (FS) in determining the pathological subtypes among patients diagnosed with cT1N0M0 invasive lung adenocarcinoma. MATERIALS AND METHODS: This was a prospective, multi-center (7 centers in China) clinical trial of Eastern Cooperative Thoracic Oncology Projects (ECTOP-1015). Patients with cT1N0M0 invasive lung adenocarcinoma were enrolled in the study. Pathological images obtained from FS and final pathology (FP) were reviewed by at least two pathologists. The primary endpoint was the concordance between FS and FP diagnoses. The inter-observer agreement for identifying pathological subtypes on FS was evaluated among three pathologists. RESULTS: A total of 935 patients were enrolled. The best sensitivity of diagnosing the predominant subtype was 78.2% in the evaluation of acinar pattern. Presence of acinar pattern diagnosed by FS was an independent factor for the concordance between FS and FP (P=0.007, 95% CI: 2.332-4.736). Patients with tumor size ＞2 cm measured by pathology showed a better concordance rate for the predominant subtype (81.6% vs 74.6%, P=0.023). The presence of radiological ground glass opacity (GGO) component did not affect the diagnosis accuracy of FS for predominant subtype (concordance rate: 76.4% vs 75.2%, P=0.687). Patients with GGO component showed better accuracy of the identification in the presence of LPA (82.1% vs 71.0%, P= 0.026). Substantial agreement between the FS diagnosis from 3 pathologists for the predominant pathological pattern was revealed with κ = 0.846. CONCLUSIONS: This is the largest prospective trial evaluating FS diagnosing pathological subtype in cT1N0M0 invasive lung adenocarcinoma. A favorable concordance in the assessment of the pathological subtypes between FS and FP was observed, indicating the feasibility of utilizing accurate intraoperative pathological diagnoses from FS in guiding surgical strategies. And combination of radiology could improve the precision of FS.

Development of an Interpretable Deep Learning Model for Pathological Tumor Response Assessment After Neoadjuvant Therapy.

BACKGROUND: Neoadjuvant therapy followed by surgery has become the standard of care for locally advanced esophageal squamous cell carcinoma (ESCC) and accurate pathological response assessment is critical to assess the therapeutic efficacy. However, it can be laborious and inconsistency between different observers may occur. Hence, we aim to develop an interpretable deep-learning model for efficient pathological response assessment following neoadjuvant therapy in ESCC. METHODS: This retrospective study analyzed 337 ESCC resection specimens from 2020-2021 at the Pudong-Branch (Cohort 1) and 114 from 2021-2022 at the Puxi-Branch (External Cohort 2) of Fudan University Shanghai Cancer Center. Whole slide images (WSIs) from these two cohorts were generated using different scanning machines to test the ability of the model in handling color variations. Four pathologists independently assessed the pathological response. The senior pathologists annotated tumor beds and residual tumor percentages on WSIs to determine consensus labels. Furthermore, 1850 image patches were randomly extracted from Cohort 1 WSIs and binarily classified for tumor viability. A deep-learning model employing knowledge distillation was developed to automatically classify positive patches for each WSI and estimate the viable residual tumor percentages. Spatial heatmaps were output for model explanations and visualizations. RESULTS: The approach achieved high concordance with pathologist consensus, with an R^2 of 0.8437, a RAcc_0.1 of 0.7586, a RAcc_0.3 of 0.9885, which were comparable to two senior pathologists (R^2 of 0.9202/0.9619, RAcc_0.1 of 8506/0.9425, RAcc_0.3 of 1.000/1.000) and surpassing two junior pathologists (R^2 of 0.5592/0.5474, RAcc_0.1 of 0.5287/0.5287, RAcc_0.3 of 0.9080/0.9310). Visualizations enabled the localization of residual viable tumor to augment microscopic assessment. CONCLUSION: This work illustrates deep learning's potential for assisting pathological response assessment. Spatial heatmaps and patch examples provide intuitive explanations of model predictions, engendering clinical trust and adoption (Code and data will be available at https://github.com/WinnieLaugh/ESCC_Percentage once the paper has been conditionally accepted). Integrating interpretable computational pathology could help enhance the efficiency and consistency of tumor response assessment and empower precise oncology treatment decisions.

Clinicopathologic features, concurrent genomic alterations, and clinical outcomes of patients with KRAS G12D mutations in resected lung adenocarcinoma.

BACKGROUND: In light of the ongoing clinical development of KRAS G12D-specific inhibitors, we sought to investigate the clinicopathologic, co-occurring genomic features and outcomes of patients with KRAS G12D-mutant lung adenocarcinoma. METHODS: 3828 patients with completely resected primary lung adenocarcinomas were examined for KRAS mutations between 2008 and 2020. The association between KRAS G12D and clinicopathologic features, molecular profiles, and outcomes was investigated. RESULTS: 65 patients (1.7%) with KRAS G12D-mutant lung adenocarcinoma were identified. KRAS G12D mutation was more frequent in males, former/current smokers, radiologic solid tumors, and invasive mucinous adenocarcinoma. TP53 and STK11 were the two most frequent concomitant mutations in the KRAS G12D group. KRAS G12D mutation did not appear to be a prognostic factor in resected stage I-III lung adenocarcinomas, while KRAS non-G12D mutation was related to worse survival, especially in stage I tumors. KRAS G12D mutations were associated with positive but low (1-49%) PD-L1 expression compared to negative (<1%), while KRAS non-G12D mutation was associated with high PD-L1 expression (≥50%). TP53 co-mutation indicated higher PD-L1 expression, while STK11 co-mutation had a negligible impact on PD-L1 expression. Furthermore, data mining of MSK datasets from cBioPortal revealed that KRAS G12D and SKT11 co-mutation were associated with a diminished response to immunotherapy. CONCLUSIONS: KRAS G12D-mutant lung adenocarcinoma harbored unique clinicopathologic and genomic characteristics. Despite not being prognostic in resected lung adenocarcinoma, KRAS G12D might be a valuable biomarker in combination with certain co-mutations for identifying relevant subgroups of patients that could eventually influence treatment regimens.

Prognostic value of KRAS G12V mutation in lung adenocarcinoma stratified by stages and radiological features.

OBJECTIVE: KRAS G12V is one of the most common KRAS mutation variants in lung adenocarcinoma (LUAD), and yet its prognostic value is still unrevealed. In this study, we investigated the clinicopathologic characteristics and prognostic value of the KRAS G12V mutation in LUAD. METHODS: Data of 3829 patients who underwent LUAD resection between 2008 and 2020 were collected. Mutations were classified as wild-type, G12V, or non-G12V. The clinicopathologic characteristics, postoperative outcomes, and recurrence pattern were analyzed among groups. RESULTS: In total, 3554 patients were wild-type and 275 patients harbored a KRAS mutation: 60 patients with G12V (22.2%) and 215 patients with non-G12V (77.8%). The KRAS G12V mutation was more frequent in male patients, older patients (≥60 years), former/current smokers, those patients with radiologic solid nodules, and those with highly invasive histologic subtypes. Tumors carrying KRAS G12V mutation exhibited elevated programmed death-ligand 1 expression in comparison with wild-type tumors. KRAS G12V was more prevalent in older patients and had less lymphovascular invasion compared with other mutation types. FGF3, RET, and KDR co-mutations occurred more frequently in the KRAS G12V group. Multivariate analysis demonstrated that the KRAS G12V mutation was an independent prognostic factor in stage Ⅰ tumors, whereas the KRAS non-G12V mutation was not. KRAS G12V was associated with early recurrence and locoregional recurrence. CONCLUSIONS: The KRAS G12V mutation was associated with aggressive clinical-pathologic phenotype and early recurrence. To note, this mutation exhibited a significantly worse prognosis in patients with part-solid and stage Ⅰ lung adenocarcinoma. Meanwhile, the prognostic significance of KRAS G12C and G12V variants was comparable.

Systemic immune index predicts tumor-infiltrating lymphocyte intensity and immunotherapy response in small cell lung cancer.

Background: Despite recent progresses in immune checkpoint blockade (ICB) in small-cell lung cancer (SCLC), a lack of understanding regarding the systemic tumor immune environment (STIE) and local tumor immune microenvironment (TIME) makes it difficult to accurately predict clinical outcomes and identify potential beneficiaries from ICB therapy. Methods: We enrolled 191 patients with stage I-III SCLC and comprehensively evaluated the prognostic role of STIE by several quantitative measurements, and further integrate it with a local immune score system (LISS) established by eXtreme Gradient Boosting (XGBoost) machine learning algorithm. We also test the value of STIE in beneficiary selection in our independent advanced SCLC cohort receiving programmed cell death 1 ligand 1 (PD-L1) blockade therapy. Results: Among several systemic immune markers, the STIE as assessed by prognostic nutritional index (PNI) was correlated with disease-free survival (DFS) and overall survival (OS), and remained as an independent prognostic factor for SCLC patients [hazard ratio (HR): 0.473, 95% confidence interval (CI): 0.241-0.929, P=0.030]. Higher PNI score was closely associated with inflamed SCLC molecular subtype and local tumor-infiltrating lymphocytes (TILs). We further constructed a LISS which combined top three important local immune biomarkers (CD8+ T-cell count, PD-L1 expression on CD8+ T-cell and CD4+ T-cell count) and integrated it with the PNI score. The final integrated immune risk system was an independent prognostic factor and achieved better predictive performance than Tumor Node Metastasis (TNM) stages and single immune biomarker. Furthermore, PNI-high extensive-stage SCLC patients achieved better clinical response and longer progression-free survival (PFS) (11.8 vs. 5.9 months, P=0.012) from PD-L1 blockade therapy. Conclusions: This study provides a method to investigate the prognostic value of overall immune status by combining the PNI with local immune biomarkers in SCLC. The promising clinical application of PNI in efficacy prediction and beneficiary selection for SCLC immunotherapy is also highlighted.

Evolutionary proteogenomic landscape from pre-invasive to invasive lung adenocarcinoma.

Lung adenocarcinoma follows a stepwise progression from pre-invasive to invasive. However, there remains a knowledge gap regarding molecular events from pre-invasive to invasive. Here, we conduct a comprehensive proteogenomic analysis comprising whole-exon sequencing, RNA sequencing, and proteomic and phosphoproteomic profiling on 98 pre-invasive and 99 invasive lung adenocarcinomas. The deletion of chr4q12 contributes to the progression from pre-invasive to invasive adenocarcinoma by downregulating SPATA18, thus suppressing mitophagy and promoting cell invasion. Proteomics reveals diverse enriched pathways in normal lung tissues and pre-invasive and invasive adenocarcinoma. Proteomic analyses identify three proteomic subtypes, which represent different stages of tumor progression. We also illustrate the molecular characterization of four immune clusters, including endothelial cells, B cells, DCs, and immune depression subtype. In conclusion, this comprehensive proteogenomic study characterizes the molecular architecture and hallmarks from pre-invasive to invasive lung adenocarcinoma, guiding the way to a deeper understanding of the tumorigenesis and progression of this disease.

EML4-ALK fusions drive lung adeno-to-squamous transition through JAK-STAT activation.

Human lung adenosquamous cell carcinoma (LUAS), containing both adenomatous and squamous pathologies, exhibits strong cancer plasticity. We find that ALK rearrangement is detectable in 5.1-7.5% of human LUAS, and transgenic expression of EML4-ALK drives lung adenocarcinoma (LUAD) formation initially and squamous transition at late stage. We identify club cells as the main cell-of-origin for squamous transition. Through recapitulating lineage transition in organoid system, we identify JAK-STAT signaling, activated by EML4-ALK phase separation, significantly promotes squamous transition. Integrative study with scRNA-seq and immunostaining identify a plastic cell subpopulation in ALK-rearranged human LUAD showing squamous biomarker expression. Moreover, those relapsed ALK-rearranged LUAD show notable upregulation of squamous biomarkers. Consistently, mouse squamous tumors or LUAD with squamous signature display certain resistance to ALK inhibitor, which can be overcome by combined JAK1/2 inhibitor treatment. This study uncovers strong plasticity of ALK-rearranged tumors in orchestrating phenotypic transition and drug resistance and proposes a potentially effective therapeutic strategy.

Extent of surgical resection for radiologically subsolid T1N0 invasive lung adenocarcinoma: When is a wedge resection acceptable?

OBJECTIVE: To evaluate whether wedge resection (WR) was appropriate for the patients with peripheral T1 N0 solitary subsolid invasive lung adenocarcinoma. METHODS: Patients with peripheral T1N0 solitary subsolid invasive lung adenocarcinoma who received sublobar resection were retrospectively reviewed. Clinicopathologic characteristics, 5-year recurrence-free survival, and 5-year lung cancer-specific overall survival were analyzed. Cox regression model was used to elucidate risk factors for recurrence. RESULTS: Two hundred fifty-eight patients receiving WR and 1245 patients receiving segmentectomy were included. The mean follow-up time was 36.87 ± 16.21 months. Five-year recurrence-free survival following WR was 96.89% for patients with ground-glass nodule (GGN) ≤2 cm and 0.25< consolidation-to-tumor ratio (CTR) ≤0.5, not statistically different from 100% for those with GGN≤2 cm and CTR ≤0.25 (P = .231). The 5-year recurrence-free survival was 90.12% for patients with GGN between 2 and 3 cm and CTR ≤0.5, significantly lower than that of patients with GGN ≤2 cm and CTR ≤0.25 (P = .046). For patients with GGN≤2 cm and 0.25

A Shift in Paradigm: Selective Lymph Node Dissection for Minimizing Oversurgery in Early Stage Lung Cancer.

Systematic lymph node dissection has been widely accepted and turned into a standard procedure for lung cancer surgery. In recent years, the concept of "minimal invasive surgery (MIS)" has greatly changed the surgical paradigm of lung cancer. Previous studies revealed that excessive dissection of lymph nodes without metastases had uncertain clinical benefit. Meanwhile, it leads to the elevated risk of postoperative complications including chylothorax and laryngeal nerve injury. In addition, dissection of nonmetastatic lymph nodes may disturb systematic immunity, resulting in the secondary effect on primary tumor or latent metastases. The past decades have witnessed the innovative strategies such as lobe-specific lymph node dissection and selective lymph node dissection. On the basis of evolution of lymph node dissection strategy, we discuss the negative effects of excessive nonmetastatic lymph node dissection and summarize the recent advances in the optimized dissection strategies, hoping to provide unique perspectives on the future directions.

Value of adjuvant chemotherapy for patients with pT2N0M0 non-small cell lung cancer receiving radical resection.

BACKGROUND: Associations between adjuvant chemotherapy (ACT) and the improvement in survival for patients with pT2N0M0 non-small cell lung cancer (NSCLC) who received R0 resection remain controversial. This study aimed to evaluate the value of ACT for patients with pT2N0M0 NSCLCs, and to identify the subgroups who could benefit from ACT. METHODS: Multivariable Cox proportional hazards regression models were used to estimate independent prognostic factors. High-risk factor (HRF) included visceral pleural invasion (VPI), lymphovascular invasion (LVI) and poor differentiation/undifferentiated tumors. RESULTS: Of the 899 patients, 277 (30.8%) patients received ACT. More younger patients (p < 0.001) and male patients (p = 0.007) received ACT. With the increase of pathological tumor size (p < 0.001) and the number of HRFs (p < 0.001), there was a significant rise in the proportion of patients receiving ACT. For all patients, ACT could not improve recurrence-free survival (RFS) (p = 0.672) and overall survival (OS) (p = 0.306). For patients with pathological stage IIA or radiological pure-solid tumors, ACT could significantly improve the OS (p = 0.011 and p = 0.037, respectively), and multivariate analysis revealed that ACT was an independent prognostic factor for patients with pathological stage IIA (p = 0.005). ACT could improve the OS significantly in patients with pathological stage IB pure-solid lung adenocarcinoma (LUAD) (p = 0.043). CONCLUSION: ACT was valuable for patients with pathological stage IIA (pT2bN0M0) and patients with radiological pure-solid LUAD of pathological stage IB. A combination of radiological features and pathological subtypes could be helpful when selecting patients with pT2N0M0 NSCLCs for ACT.

Residual tumor descriptors proposed by the International Association for the Study of Lung Cancer may not be applicable to stage I and ground-glass opacity-featured non-small cell lung cancer.

Background: The International Association for the Study of Lung Cancer (IASLC) has proposed a residual tumor descriptor, essential for subsequent treatments. This study aimed to validate the prognostic effect of the proposed R descriptor and restrict its scope of clinical application in a large-scale cohort with non-small cell lung cancer (NSCLC). Methods: Patients, who underwent lobectomy from January 2010 to May 2019, were retrospectively reviewed. Patients were categorized according to the different R classification standards proposed by Union for International Cancer Control (UICC) and IASLC. Results: Among 5,200 enrolled patients with NSCLC, 1,727 and 9 cases of UICC-R0 were re-evaluated as uncertain resection [R(un)] and R1, respectively. After reclassification, there were 3,228 (62.1%) cases of R0, 1,727 (33.2%) cases of R(un), 151 (2.9%) cases of R1, and 94 (1.8%) cases of R2. Not performing rigorous systematic nodal dissection (SND) or lobe-specific SND (68.3%) was the main reason for the alteration from R0 to R(un). Patients with R(un) showed intermediate survival between those with R0 and R1. Further multivariable Cox analysis indicated that the proposed R descriptor was an independent prognostic factor for overall survival (OS) and recurrence-free survival (RFS). However, subgroup analysis of OS and RFS revealed that there was no significant difference between R0 and R(un) in patients with ground-glass opacity (GGO) or patients with tumor-node-metastasis stage I. Conclusions: R(un) represented an intermediate type between R0 and R1. Our study provided an external validation for new residual tumor descriptors for NSCLC proposed by IASLC. Proposed residual tumor descriptors were applicable in radiologic solid NSCLC and stage II-III NSCLC, but were ineffective for GGO-featured or stage I NSCLC.

Pathological and radiological T descriptors in invasive lung adenocarcinoma: from correlations to prognostic significance.

Background: The eighth T classification excluded lepidic and ground-glass opacity (GGO) components. Current studies demonstrated lepidic and GGO components showed independent prognostic significances. This study elucidated the correlations and prognostic impacts of pathological and radiological T descriptors in invasive lung adenocarcinoma. Methods: A total of 1,490 patients with invasive lung adenocarcinoma were retrospectively reviewed. Correlation between pathological invasive size (PIS) and radiological solid size (RSS), and lepidic ratio and GGO ratio were comprehensively evaluated. Impacts of these pathological and radiological T descriptors on recurrence-free survival (RFS) were comparatively analyzed. Results: Clinical (c)T-stage was more frequently downstaged than upstaged comparing with the pathological (p)T-stage (28.4% vs. 18.2%). The correlation between PIS and RSS in solid nodule was stronger than that in part-solid nodule (solid: R2=0.750 vs. part-solid: R2=0.355). Some pathological invasive components except solid component were featured as GGO. Among T1 patients, lepidic absent GGO showed better RFS than lepidic present solid nodule (pT1: P=0.001; cT1: P=0.021). Multivariable analysis revealed GGO ratio was an independent prognostic factor for RFS in T1 invasive lung adenocarcinoma, whereas lepidic ratio was not. Conclusions: Among T1 invasive lung adenocarcinoma, GGO ratio showed independent prognostic value for RFS, regardless of RSS. Meanwhile, lepidic ratio was not an independent RFS factor. GGO component rather than lepidic component should be considered as an additional T descriptor.

Hybrid deep multi-task learning radiomics approach for predicting EGFR mutation status of non-small cell lung cancer in CT images.

Objective.Epidermal growth factor receptor (EGFR) mutation genotyping plays a pivotal role in targeted therapy for non-small cell lung cancer (NSCLC). We aimed to develop a computed tomography (CT) image-based hybrid deep radiomics model to predict EGFR mutation status in NSCLC and investigate the correlations between deep image and quantitative radiomics features.Approach.First, we retrospectively enrolled 818 patients from our centre and 131 patients from The Cancer Imaging Archive database to establish a training cohort (N= 654), an independent internal validation cohort (N= 164) and an external validation cohort (N= 131). Second, to predict EGFR mutation status, we developed three CT image-based models, namely, a multi-task deep neural network (DNN), a radiomics model and a feature fusion model. Third, we proposed a hybrid loss function to train the DNN model. Finally, to evaluate the model performance, we computed the areas under the receiver operating characteristic curves (AUCs) and decision curve analysis curves of the models.Main results.For the two validation cohorts, the feature fusion model achieved AUC values of 0.86 ± 0.03 and 0.80 ± 0.05, which were significantly higher than those of the single-task DNN and radiomics models (allP< 0.05). There was no significant difference between the feature fusion and the multi-task DNN models (P> 0.8). The binary prediction scores showed excellent prognostic value in predicting disease-free survival (P= 0.02) and overall survival (P< 0.005) for validation cohort 2.Significance.The results demonstrate that (1) the feature fusion and multi-task DNN models achieve significantly higher performance than that of the conventional radiomics and single-task DNN models, (2) the feature fusion model can decode the imaging phenotypes representing NSCLC heterogeneity related to both EGFR mutation and patient NSCLC prognosis, and (3) high correlations exist between some deep image and radiomics features.

Pattern of tumor regression after neoadjuvant chemoimmunotherapy for esophageal squamous cell carcinoma.

Background: Immune checkpoint inhibitors have been increasingly applied for esophageal cancer. The aims of this study were to evaluate the pattern of tumor regression after neoadjuvant chemoimmunotherapy. Methods: From January 2020 to December 2021, 138 patients with esophageal squamous cell carcinoma who had esophagectomy after neoadjuvant chemoimmunotherapy were reviewed. Surgical and pathological results were analyzed, and tumor regression pattern was evaluated. Results: Of the 138 patients, 65 (47.1%) patients had chemotherapy combined with camrelizumab, 48 (34.8%) with pembrolizumab, 13 (9.4%) with tislelizumab, and 12 (8.7%) with sintilimab. Sixty-four patients (46.4%) underwent McKewon procedure, and 74 (53.6%) Ivor-Lewis procedure, respectively. There were 131/138 patients (94.9%) who had R0 resections, and the median number of resected lymph nodes was 28. Pneumonia was the most common complication after surgery (14.5%). Pathological complete regression occurred in 28 patients (20.3%). Regarding to residual tumor, there were 50 patients (36.2%) with residual tumor in the mucosa, 81 (58.7%) in the submucosa, 85 (61.6%) in the muscularis propria, 47 (34.1%) in the adventitia and 71 (51.4%) in the lymph nodes. There were 88 patients with no residual tumor in the mucosa, of whom 60 (68.2%) had residual tumors in other layers or in the lymph nodes. Conclusions: In this retrospective study, esophagectomy after neoadjuvant chemoimmunotherapy is safe with acceptable surgical risk. Preferential clearing of tumor cells in mucosa layer is common after immunotherapy, while the rate of complete pathological response is relatively low, indicating surgery is still necessary.

Tumor-derived Vimentin as a novel biomarker for distinct subtypes predicting adjuvant chemotherapy resistance and T-cell-inflamed phenotype in small cell lung cancer.

Despite recent progress in subtype classification for small cell lung carcinoma (SCLC), little is known about the biomarker for triple-negative (ASCL1, NEUROD1, and POU2F3 negative) tumors. The long-term survival, adjuvant chemotherapy (ACT) response, and immune milieu in different SCLC subtypes have also not been well established. Here, we retrospectively collected a large cohort of 192 primary SCLC tumors and reported that ASCL1-, NEUROD1- and POU2F3-dominant subtypes counted for 61.38%, 19.31%, and 6.21%, respectively. Subtype intra-tumoral heterogeneity and co-expression at the single-cell level existed substantially. The expression of tumor-derived Vimentin (VIM) was nearly restricted to triple-negative SCLC tumors (15/19, 78.9%) while YAP1 expression was distributed widely in other subtypes. The SCLC subtyping model was independently prognostic of OS and RFS (p <  0.001 and p = 0.043). In particular, patients with ASCL1-positive SCLC tumors can benefit more from ACT, and VIM-positive tumors did the opposite. Compared with other subtypes, the VIM-dominant SCLC subtype was associated with abundant but functionally impaired CD4+ and CD8+ T-cells, which highly expressed inhibitory checkpoints and potentially benefit from PD-L1 blockade therapy. Our study showed that tumor-derived SCLC-V subtype could independently predict ACT response. The distinct immune landscape between subtypes may help inform personalized immune therapeutic approaches.