Predicting complication risks after sleeve lobectomy for non-small cell lung cancer.

Background: Sleeve lobectomy is a challenging procedure with a high risk of postoperative complications. To facilitate surgical decision-making and optimize perioperative treatment, we developed risk stratification models to quantify the probability of postoperative complications after sleeve lobectomy. Methods: We retrospectively analyzed the clinical features of 691 non-small cell lung cancer (NSCLC) patients who underwent sleeve lobectomy between July 2016 and December 2019. Logistic regression models were trained and validated in the cohort to predict overall complications, major complications, and specific minor complications. The impact of specific complications in prognostic stratification was explored via the Kaplan-Meier method. Results: Of 691 included patients, 232 (33.5%) developed complications, including 35 (5.1%) and 197 (28.5%) patients with major and minor complications, respectively. The models showed robust discrimination, yielding an area under the receiver operating characteristic (ROC) curve (AUC) of 0.853 [95% confidence interval (CI): 0.705-0.885] for predicting overall postoperative complication risk and 0.751 (95% CI: 0.727-0.762) specifically for major complication risks. Models predicting minor complications also achieved good performance, with AUCs ranging from 0.78 to 0.89. Survival analyses revealed a significant association between postoperative complications and poor prognosis. Conclusions: Risk stratification models could accurately predict the probability and severity of complications in NSCLC patients following sleeve lobectomy, which may inform clinical decision-making for future patients.

The Co-oligomers of Aß42 and Human Islet Amyloid Polypeptide Exacerbate Neurotoxicity and Alzheimer-like Pathology at Cellular Level.

The Aß hypothesis has long been central to Alzheimer's disease (AD) theory, with a recent surge in attention following drug approvals targeting Aß plaque clearance. Aß42 oligomers (AßO) are key neurotoxins. While ß-amyloid (Aß) buildup is a hallmark of AD, postmortem brain analyses have unveiled human islet amyloid polypeptide (hIAPP) deposition in AD patients, suggesting a potential role in Alzheimer's pathology. This study investigates the neurotoxic effects of co-aggregates of Aß42 and hIAPP, specifically focusing on their impact on cell survival, apoptosis, and AD-like pathology. We analyzed and compared the impact of AßO and Aß42-hIAPP on cell survival in SH-SY5Y cells, apoptosis and inducing AD-like pathology in glutamatergic neurons. Aß42-hIAPP co-oligomers exhibited significantly greater toxicity, causing 2.3-3.5 times higher cell death compared to AßO alone. Furthermore, apoptosis rates were significantly exacerbated in glutamatergic neurons when exposed to Aß42-hIAPP co-oligomers. The study also revealed that Aß42-hIAPP co-oligomers induced typical AD-like pathology in glutamatergic neurons, including the presence of Aß deposits (detected by 6E10 and 4G8 immunofluorescence) and alterations in tau protein (changes in total tau HT7, phosphorylated tau AT8, AT180). Notably, Aß42-hIAPP co-oligomers induced a more severe AD pathology compared to AßO alone. These findings provide compelling evidence for the heightened toxicity of Aß42-hIAPP co-oligomers on neurons and their role in exacerbating AD pathology. The study contributes novel insights into the pathogenesis of Alzheimer's disease, highlighting the potential involvement of hIAPP in AD pathology. Together, these findings offer novel insights into AD pathogenesis and routes for constructing animal models.

Neoadjuvant chemoimmunotherapy confers survival advantage for patients undergoing sleeve lobectomy.

OBJECTIVES: It has been demonstrated that neoadjuvant immune checkpoint inhibitor (ICI) plus chemotherapy was safe and feasible referred to neoadjuvant chemotherapy for patients with non-small cell lung cancer undergoing sleeve lobectomy. Nevertheless, no survival data were reported in the previous researches. Therefore, we conducted this study to compare neoadjuvant ICI plus chemotherapy versus neoadjuvant chemotherapy followed by sleeve lobectomy for long-term survival outcomes. METHODS: Patients who underwent bronchial sleeve lobectomy following neoadjuvant ICI plus chemotherapy or neoadjuvant chemotherapy were retrospectively identified. Treatment response, perioperative outcomes, event-free survival and overall survival were compared between groups in the overall and the inverse probability of treatment weighting-adjusted cohort. RESULTS: A total of 139 patients with 39 lung cancer recurrence and 21 death were included. Among them, 83 (59.7%) and 56 (40.3%) patients received neoadjuvant chemotherapy and neoadjuvant ICI plus chemotherapy, respectively. After inverse probability of treatment weighting, more patients achieved complete pathological response in the neoadjuvant ICI plus chemotherapy group (6.0% vs 26.3%, P < 0.001). There was no significant difference regarding overall postoperative complication (23.8% vs 20.2%, P = 0.624) and specific complications (all P > 0.05). Patients receiving neoadjuvant ICI plus chemotherapy had favourable event-free survival (hazard ratio 0.37, 95% confidence interval 0.16-0.85, P = 0.020) and overall survival (hazard ratio 0.23, 95% confidence interval 0.06-0.80, P = 0.021). Multivariable analysis revealed that neoadjuvant ICI plus chemotherapy was an independent predictor for favourable event-free survival (hazard ratio 0.37, 95% confidence interval 0.15-0.86, P = 0.020, adjusted for clinical TNM stage). CONCLUSIONS: Neoadjuvant ICI plus chemotherapy was correlated with favourable long-term survival in patients with non-small cell lung cancer undergoing sleeve lobectomy.

Incorporation of the lepidic component as an additional pathological T descriptor for non-small cell lung cancer: Data from 3335 cases of lung adenocarcinoma.

OBJECTIVES: The Lepidic Component (LP) identifies a subgroup with an excellent prognosis for lung adenocarcinoma (LUAD). Our research aimed to propose an improved pathological T (pT) stage for LUAD based on LP. MATERIALS AND METHODS: Totally, 3335 surgical patients with pathological stage I LUAD were incorporated. Factors affecting survival were investigated by analyzing recurrence-free survival (RFS) and overall survival (OS) using the Kaplan-Meier method and Cox regression analyses. Subgroup analysis based on Lepidic Ratio (LR) was further evaluated. The net benefit from the modified pT category (pTm) was assessed using the Area Under the time-dependent Receiver Operating Curve (AUC), Harrell's Concordance Index (C-index), Reclassification Improvement (NRI), and Integrated Discrimination Improvement (IDI). RESULTS: The presence of LP (LP+) was identified in 1425 (42.7 %) patients, indicating a significantly better RFS (P < 0.001) and OS (P < 0.001) than those without LP, and similar results were reproduced in pT1a-pT2a subcategory (P < 0.050 for all). Multivariable Cox analysis revealed LP+ as an independent prognostic factor for both RFS (HR, 0.622; P < 0.001) and OS (HR, 0.710; P = 0.019). However, lepidic ratio (LR) was not independently associated with both RFS and OS for LP+ patients. The 5-year RFS and OS rates between T1a (LP-) and T1b (LP+), T1b (LP-) and T1c (LP+), and T1b (LP-) and T2a (LP+) were comparable (P > 0.050 for all). After modification, compared with current 8th edition pT stage system (pT8), pTm independently predicted RFS and OS, and AUCs, c-index, NRI, and IDI analysis all demonstrated pTm holds better discrimination performances than pT8 for LUAD prognosis. CONCLUSION: LP can be an additional down-staged T descriptor for pathological stage I LUAD and improve the survival predictive performance of reclassification.

PET/CT-based deep learning grading signature to optimize surgical decisions for clinical stage I invasive lung adenocarcinoma and biologic basis under its prediction: a multicenter study.

PURPOSE: No consensus on a grading system for invasive lung adenocarcinoma had been built over a long period of time. Until October 2020, a novel grading system was proposed to quantify the whole landscape of histologic subtypes and proportions of pulmonary adenocarcinomas. This study aims to develop a deep learning grading signature (DLGS) based on positron emission tomography/computed tomography (PET/CT) to personalize surgical treatments for clinical stage I invasive lung adenocarcinoma and explore the biologic basis under its prediction. METHODS: A total of 2638 patients with clinical stage I invasive lung adenocarcinoma from 4 medical centers were retrospectively included to construct and validate the DLGS. The predictive performance of the DLGS was evaluated by the area under the receiver operating characteristic curve (AUC), its potential to optimize surgical treatments was investigated via survival analyses in risk groups defined by the DLGS, and its biological basis was explored by comparing histologic patterns, genotypic alternations, genetic pathways, and infiltration of immune cells in microenvironments between risk groups. RESULTS: The DLGS to predict grade 3 achieved AUCs of 0.862, 0.844, and 0.851 in the validation set (n = 497), external cohort (n = 382), and prospective cohort (n = 600), respectively, which were significantly better than 0.814, 0.810, and 0.806 of the PET model, 0.813, 0.795, and 0.824 of the CT model, and 0.762, 0.734, and 0.751 of the clinical model. Additionally, for DLGS-defined high-risk population, lobectomy yielded an improved prognosis compared to sublobectomy p = 0.085 for overall survival [OS] and p = 0.038 for recurrence-free survival [RFS]) and systematic nodal dissection conferred a superior prognosis to limited nodal dissection (p = 0.001 for OS and p = 0.041 for RFS). CONCLUSION: The DLGS harbors the potential to predict the histologic grade and personalize the surgical treatments for clinical stage I invasive lung adenocarcinoma. Its applicability to other territories should be further validated by a larger international study.

Prognostic outcomes and recurrence patterns in resected stage I lung adenocarcinoma harbouring atypical epidermal growth factor receptor mutations.

OBJECTIVES: Limited data exist on the characteristics of atypical epidermal growth factor receptor (EGFR) mutations in early-stage lung cancer. Our goal was to elucidate the associations with outcomes and recurrence patterns in resected stage I lung adenocarcinoma harbouring atypical EGFR mutations. METHODS: Eligible patients between 2014 and 2019 were retrospectively identified and grouped into exon20 insertion mutations and major atypical mutations, which included G719X, L861Q and S768I. Disease-free survival (DFS) was evaluated in the entire cohort and stratified by radiologic characteristics. Recurrence patterns were investigated and compared between groups. A competing risk model was used to estimate the cumulative incidence of recurrence. RESULTS: A total of 710 patients were finally included. Among them, 289 (40.7%) patients had exon 20 insertion mutations and 421 (59.3%) patients had major atypical mutations. There was no significant difference regarding DFS (P = 0.142) between groups in the entire cohort. The interaction between mutation subtype and the presence of ground-glass opacities was significant (hazard ratio 2.00, 95% confidence interval 1.59-2.51, P < 0.001), indicating DFS between exon 20 insertion mutations and major atypical mutations may be different among subsolid and solid tumours. Survival analysis consistently revealed no significant difference in subsolid tumours (P = 0.680), but favourable DFS of exon 20 insertion mutations in solid tumours (P = 0.037). Furthermore, patients with exon 20 insertion mutations had a lower risk of developing bone metastases did those with radiologic solid tumours (Gray's test, P = 0.012). CONCLUSIONS: Exon 20 insertion mutations were correlated with favourable DFS and lower incidence of bone metastases in radiologic solid lung adenocarcinomas harbouring atypical EGFR mutations.

PET/CT based cross-modal deep learning signature to predict occult nodal metastasis in lung cancer.

Occult nodal metastasis (ONM) plays a significant role in comprehensive treatments of non-small cell lung cancer (NSCLC). This study aims to develop a deep learning signature based on positron emission tomography/computed tomography to predict ONM of clinical stage N0 NSCLC. An internal cohort (n = 1911) is included to construct the deep learning nodal metastasis signature (DLNMS). Subsequently, an external cohort (n = 355) and a prospective cohort (n = 999) are utilized to fully validate the predictive performances of the DLNMS. Here, we show areas under the receiver operating characteristic curve of the DLNMS for occult N1 prediction are 0.958, 0.879 and 0.914 in the validation set, external cohort and prospective cohort, respectively, and for occult N2 prediction are 0.942, 0.875 and 0.919, respectively, which are significantly better than the single-modal deep learning models, clinical model and physicians. This study demonstrates that the DLNMS harbors the potential to predict ONM of clinical stage N0 NSCLC.

Prognostic Significance of the Proposed Residual Tumor Classification in Patients With NSCLC After Sleeve Lobectomy.

Introduction: To validate the residual tumor (R) classification proposed by the International Association for the Study of Lung Cancer (IASLC) in NSCLC after sleeve lobectomy. Methods: A total of 682 patients were analyzed. The R status, on the basis of the Union for International Cancer Control (UICC) criteria, was recategorized according to the IASLC descriptors. Recurrence-free survival (RFS) and overall survival (OS) among different R classifications were assessed for the entire cohort and pathologic node (pN) subgroups. Results: All in all, 631 (92.5%), 48 (7.1%), and three patients (0.4%) were classified as R0, R1, and R2, respectively, by the UICC criteria, whereas 489 (71.7%), 110 (16.1%), and 83 patients (12.2%), received R0, uncertain resection (R[un]), and R1/2 resection, respectively, according to the IASLC criteria. There were 96 patients (15.2%) with UICC R0 who were reclassified as R(un), mainly because of the positive highest mediastinal node station (82 of 96, 85.4%). A total of 46 patients (7.3%) were reassigned from UICC R0 to IASLC R1/2 owing to extracapsular extension. For the entire cohort, patients with R(un) and R1/2 exhibited worse RFS (R[un], adjusted p = 0.023; R1/2, adjusted p = 0.001) and OS (R[un], adjusted p = 0.040; R1/2, adjusted p = 0.051) compared with R0. No significant differences were observed between R(un) and R1/2 (RFS, adjusted p = 0.586; OS, adjusted p = 0.781). Furthermore, subgroup analysis revealed a distinct prognostic impact of the IASLC R status-with prognostic significances in the pN1 and pN2 subgroups, but not in the pN0 subgroup. Conclusions: The IASLC R descriptors helped to stratify the prognosis of NSCLC after sleeve lobectomy, with its prognostic impact varied among pN stages.

Cartilage tissue from sites of weight bearing in patients with osteoarthritis exhibits a differential phenotype with distinct chondrocytes subests.

OBJECTIVE: Osteoarthritis (OA) is a degenerative joint disease associated with excessive mechanical loading. The aim here was to elucidate whether different subpopulations of chondrocytes exhibit distinct phenotypes in response to variations in loading conditions. Furthermore, we seek to investigate the transcriptional switches and cell crosstalk among these chondrocytes subsets. METHODS: Proteomic analysis was performed on cartilage tissues isolated from weight-bearing and non-weight-bearing regions. Additionally, single-cell RNA sequencing was employed to identify different subsets of chondrocytes. For disease-specific cells, in vitro differentiation induction was performed, and their presence was confirmed in human cartilage tissue sections using immunofluorescence. The molecular mechanisms underlying transcriptional changes in these cells were analysed through whole-transcriptome sequencing. RESULTS: In the weight-bearing regions of OA cartilage tissue, a subpopulation of chondrocytes called OA hypertrophic chondrocytes (OAHCs) expressing the marker genes SLC39A14 and COL10A1 are present. These cells exhibit unique characteristics of active cellular interactions mediated by the TGFß signalling pathway and express OA phenotypes, distinct from hypertrophic chondrocytes in healthy cartilage. OAHCs are mainly distributed in the superficial region of damaged cartilage in human OA tissue, and on TGFß stimulation, exhibit activation of transcriptional expression of iron metabolism-related genes, along with enrichment of associated pathways. CONCLUSION: This study identified and validated the existence of a subset of OAHCs in the weight-bearing area of OA cartilage tissue. Our findings provide a theoretical basis for targeting OAHCs to slow down the progression of OA and facilitate the repair of cartilage injuries.

TransVG++: End-to-End Visual Grounding With Language Conditioned Vision Transformer.

In this work, we explore neat yet effective Transformer-based frameworks for visual grounding. The previous methods generally address the core problem of visual grounding, i.e., multi-modal fusion and reasoning, with manually-designed mechanisms. Such heuristic designs are not only complicated but also make models easily overfit specific data distributions. To avoid this, we first propose TransVG, which establishes multi-modal correspondences by Transformers and localizes referred regions by directly regressing box coordinates. We empirically show that complicated fusion modules can be replaced by a simple stack of Transformer encoder layers with higher performance. However, the core fusion Transformer in TransVG is stand-alone against uni-modal encoders, and thus should be trained from scratch on limited visual grounding data, which makes it hard to be optimized and leads to sub-optimal performance. To this end, we further introduce TransVG++ to make two-fold improvements. For one thing, we upgrade our framework to a purely Transformer-based one by leveraging Vision Transformer (ViT) for vision feature encoding. For another, we devise Language Conditioned Vision Transformer that removes external fusion modules and reuses the uni-modal ViT for vision-language fusion at the intermediate layers. We conduct extensive experiments on five prevalent datasets, and report a series of state-of-the-art records.

Clinical utility of [18F]FDG PET/CT in the assessment of mediastinal lymph node disease after neoadjuvant chemoimmunotherapy for non-small cell lung cancer.

OBJECTIVES: The performance of positron emission tomography/computed tomography (PET/CT) for the prediction of ypN2 disease in non-small cell lung cancer (NSCLC) after neoadjuvant chemoimmunotherapy has not been reported. This multicenter study investigated the utility of PET/CT to assess ypN2 disease in these patients. METHODS: A total of 181 consecutive patients (chemoimmunotherapy = 86, chemotherapy = 95) at four institutions were enrolled in this study. Every patient received a PET/CT scan prior to surgery and complete resection with systematic nodal dissection. The diagnostic performance was evaluated through area under the curve (AUC). Kaplan-Meier method and Cox analysis were performed to identify the risk factors affecting recurrences. RESULTS: The sensitivity, specificity, and accuracy of PET/CT for ypN2 diseases were 0.667, 0.835, and 0.779, respectively. Therefore, the AUC was 0.751. Compared with the false positive cases, the mean value of max standardized uptake value (SUVmax) (6.024 vs. 2.672, p < 0.001) of N2 nodes was significantly higher in true positive patients. Moreover, the SUVmax of true positive (7.671 vs. 5.976, p = 0.365) and false (2.433 vs. 2.339, p = 0.990) positive cases were similar between chemoimmunotherapy and chemotherapy, respectively. Survival analysis proved that pathologic N (ypN) 2 patients could be stratified by PET/CT-N2(+ vs. -) for both chemoimmunotherapy (p = 0.023) and chemotherapy (p = 0.010). CONCLUSIONS: PET/CT is an accurate and non-invasive test for mediastinal restaging of NSCLC patients who receive neoadjuvant chemoimmunotherapy. The ypN2 patients with PET/CT-N2( +) are identified as an independent prognostic factor compared with PET/CT-N2(-). CLINICAL RELEVANCE STATEMENT: Imaging with 18F-fluorodeoxyglucose positron emission tomography/computed tomography (PET/CT) plays an integral role during disease diagnosis, staging, and therapeutic response assessments in patients with NSCLC. PET/CT could be an effective non-invasive tool for predicting ypN2 diseases after neoadjuvant chemoimmunotherapy. KEY POINTS: • PET/CT could serve as an effective non-invasive tool for predicting ypN2 diseases. • The ypN2 patients with PET/CT-N2( +) were a strong and independent prognostic factor. • The application of PET/CT for restaging should be encouraged in clinical practice.

Meta-analysis of the association between RNF213 polymorphisms and clinical features of moyamoya disease in Asian population.

BACKGROUND: We performed this study to explore the relationship between ring finger protein 213 (RNF213) gene polymorphisms and clinical features in moyamoya disease (MMD). METHODS: Electronic databases (PubMed, Google Scholar, Embase, Scopus, Cochrane Library) were conducted from inception to May 15th, 2022. Odds ratios (ORs) with 95 % confidence intervals (CIs) were generated as effect size for binary variants. Subgroup analyses were performed by the RNF213 polymorphisms. Sensitivity was used to examine the robustness of associations. RESULTS: A total of 16 articles and 3061 MMD patients were included and the association of five RNF213 polymorphisms on 9 clinical features of MMD were identified. Patients under 18 years of age at onset, familial MMD, cerebral ischemic stroke and posterior cerebral artery involvement (PCi) were significantly more common in mutant type compared with wild type of RNF213. Compared with each wild type, subgroup analysis showed that rs11273543 and rs9916351 remarkably increased risk of MMD on early onset, but rs371441113 evidently delayed the onset of MMD. Rs112735431 in mutant type was significantly higher than wild type in patients with PCi. Subgroup analysis in mutant type showed that rs112735431 conspicuously decreased intracerebral/ intraventricular hemorrhage (ICH/IVH) risk and yet rs148731719 obviously increased the risk in ICH/IVH. CONCLUSION: More attention should be paid to patients on whom the ischemic MMD occurs younger than 18 years old. RNF213 polymorphism screening and cerebrovascular imaging examination should be performed to evaluate intracranial vascular involvement, to achieve early detection and early treatment and avoid more serious cerebrovascular events.

The development and validation of a circulating tumor cells-based integrated model for improving the indeterminate lung solid nodules diagnosis.

Background: There is a risk of over investigation and delayed treatment in the work up of solid nodules. Thus, the aim of our study was to develop and validate an integrated model that estimates the malignant risk for indeterminate pulmonary solid nodules (IPSNs). Methods: Patients included in this study with IPSNs who was diagnosed malignant or benign by histopathology. Univariate and multivariate logistic regression were used to build integrated model based on clinical, circulating tumor cells (CTCs) and radiomics features. The performance of the integrated model was estimated by applying receiver operating characteristic (ROC) analysis, and tested in different nodules size and intermediate risk IPSNs. Net reclassification index (NRI) was applied to quantify the additional benefit derived from the integrated model. Results: The integrated model yielded areas under the ROC curves (AUCs) of 0.83 and 0.76 in internal and external set, respectively, outperforming CTCs (0.70, P=0.001; 0.68, P=0.128), the Mayo clinical model (0.68, P<0.001; 0.55, P=0.007), the and radiomics model (0.72, P=0.002; 0.67, P=0.050) in both validation sets. Robust performance with high sensitivity up to 98% was also maintained in IPSNs with different solid size and intermediate risk probability. The performance of the integrated model was comparable with positron emission tomography/computed tomography (PET-CT) examination (P=0.308) among the participants with established PET-CT records. NRI demonstrated that the integrated model provided net reclassification of at least 10% on the external validation set compared with single CTCs test. Conclusions: The integrated model could complement conventional risk models to improve the diagnosis of IPSNs, which is not inferior to PET-CT and could help to guide clinician's decision-making on clinically specific population.

Masked Contrastive Representation Learning for Reinforcement Learning.

In pixel-based reinforcement learning (RL), the states are raw video frames, which are mapped into hidden representation before feeding to a policy network. To improve sample efficiency of state representation learning, recently, the most prominent work is based on contrastive unsupervised representation. Witnessing that consecutive video frames in a game are highly correlated, to further improve data efficiency, we propose a new algorithm, i.e., masked contrastive representation learning for RL (M-CURL), which takes the correlation among consecutive inputs into consideration. In our architecture, besides a CNN encoder for hidden presentation of input state and a policy network for action selection, we introduce an auxiliary Transformer encoder module to leverage the correlations among video frames. During training, we randomly mask the features of several frames, and use the CNN encoder and Transformer to reconstruct them based on context frames. The CNN encoder and Transformer are jointly trained via contrastive learning where the reconstructed features should be similar to the ground-truth ones while dissimilar to others. During policy evaluation, the CNN encoder and the policy network are used to take actions, and the Transformer module is discarded. Our method achieves consistent improvements over CURL on 14 out of 16 environments from DMControl suite and 23 out of 26 environments from Atari 2600 Games. The code is available at https://github.com/teslacool/m-curl.

Genopathomic profiling identifies signatures for immunotherapy response of lung adenocarcinoma via confounder-aware representation learning.

Immunotherapy shows durable response but only in a subset of patients, and test for predictive biomarkers requires procedures in addition to routine workflow. We proposed a confounder-aware representation learning-based system, genopathomic biomarker for immunotherapy response (PITER), that uses only diagnosis-acquired hematoxylin-eosin (H&E)-stained pathological slides by leveraging histopathological and genetic characteristics to identify candidates for immunotherapy. PITER was generated and tested with three datasets containing 1944 slides of 1239 patients. PITER was found to be a useful biomarker to identify patients of lung adenocarcinoma with both favorable progression-free and overall survival in the immunotherapy cohort (p < 0.05). PITER was significantly associated with pathways involved in active cell division and a more immune activating microenvironment, which indicated the biological basis in identifying patients with favorable outcome of immunotherapy. Thus, PITER may be a potential biomarker to identify patients of lung adenocarcinoma with a good response to immunotherapy, and potentially provide precise treatment.

Deep learning for predicting major pathological response to neoadjuvant chemoimmunotherapy in non-small cell lung cancer: A multicentre study.

BACKGROUND: This study, based on multicentre cohorts, aims to utilize computed tomography (CT) images to construct a deep learning model for predicting major pathological response (MPR) to neoadjuvant chemoimmunotherapy in non-small cell lung cancer (NSCLC) and further explore the biological basis under its prediction. METHODS: 274 patients undergoing curative surgery after neoadjuvant chemoimmunotherapy for NSCLC at 4 centres from January 2019 to December 2021 were included and divided into a training cohort, an internal validation cohort, and an external validation cohort. ShuffleNetV2x05-based features of the primary tumour on the CT scans within the 2 weeks preceding neoadjuvant administration were employed to develop a deep learning score for distinguishing MPR and non-MPR. To reveal the underlying biological basis of the deep learning score, a genetic analysis was conducted based on 25 patients with RNA-sequencing data. FINDINGS: MPR was achieved in 54.0% (n = 148) patients. The area under the curve (AUC) of the deep learning score to predict MPR was 0.73 (95% confidence interval [CI]: 0.58-0.86) and 0.72 (95% CI: 0.58-0.85) in the internal validation and external validation cohorts, respectively. After integrating the clinical characteristic into the deep learning score, the combined model achieved satisfactory performance in the internal validation (AUC: 0.77, 95% CI: 0.64-0.89) and external validation cohorts (AUC: 0.75, 95% CI: 0.62-0.87). In the biological basis exploration for the deep learning score, a high deep learning score was associated with the downregulation of pathways mediating tumour proliferation and the promotion of antitumour immune cell infiltration in the microenvironment. INTERPRETATION: The proposed deep learning model could effectively predict MPR in NSCLC patients treated with neoadjuvant chemoimmunotherapy. FUNDING: This study was supported by National Key Research and Development Program of China, China (2017YFA0205200); National Natural Science Foundation of China, China (91959126, 82022036, 91959130, 81971776, 81771924, 6202790004, 81930053, 9195910169, 62176013, 8210071009); Beijing Natural Science Foundation, China (L182061); Strategic Priority Research Program of Chinese Academy of Sciences, China (XDB38040200); Chinese Academy of Sciences, China (GJJSTD20170004, QYZDJ-SSW-JSC005); Shanghai Hospital Development Center, China (SHDC2020CR3047B); and Science and Technology Commission of Shanghai Municipality, China (21YF1438200).

Clinicopathologic characteristics and prognostic impact of atypical EGFR mutations in completely resected lung adenocarcinoma.

INTRODUCTION: This study evaluated the clinicopathologic characteristics and prognostic impact of atypical epidermal growth factor receptor (EGFR) mutations in patients with completely resected lung adenocarcinoma (LUAD) and investigate whether adjuvant chemotherapy could benefit the survival outcomes for these subjects. MATERIAL AND METHODS: We retrospectively reviewed resected LUAD samples from 8437 patients and identified 5358 EGFR-mutated (EGFRm) cases. Of these, 4847 had classical mutations, while 511 had atypical mutations. For further survival analysis, propensity score matching, Kaplan-Meier curve, and Cox regression analyses were conducted. RESULTS: Of the 511 patients with atypical EGFRm LUAD, 131 patients had compound mutations. The frequency of exon 20 insertion (20-ins), G719X, L861Q, S768I, and de novo T790M were 30.3%, 32.7%, 21.9%, 9.2%, and 11.4%, respectively. These patients included a higher proportion of males than those with classical EGFRm LUAD. Between the 483 matched pairs of the classical and atypical EGFRm patients, no significant difference emerged in disease-free survival (DFS) (p = 0.476). Patients with the L861Q mutation had the poorest DFS among those with atypical EGFRm LUAD (p = 0.005). Cox regression analyses revealed that the L861Q mutation was an independent prognostic factor for DFS in 487 patients with solely atypical EGFRm LUAD. In addition, adjuvant chemotherapy did not improve the DFS for those patients, whether in stage IB (p = 0.638) or II-III (p = 0.505) of the disease. CONCLUSION: The L861Q mutation is an independent prognostic factor for DFS in patients with atypical EGFRm LUAD after complete resection who would not benefit from adjuvant chemotherapy regardless of disease stage.

Early stage NSCLS patients' prognostic prediction with multi-information using transformer and graph neural network model.

Background: We proposed a population graph with Transformer-generated and clinical features for the purpose of predicting overall survival (OS) and recurrence-free survival (RFS) for patients with early stage non-small cell lung carcinomas and to compare this model with traditional models. Methods: The study included 1705 patients with lung cancer (stages I and II), and a public data set for external validation (n=127). We proposed a graph with edges representing non-imaging patient characteristics and nodes representing imaging tumour region characteristics generated by a pretrained Vision Transformer. The model was compared with a TNM model and a ResNet-Graph model. To evaluate the models' performance, the area under the receiver operator characteristic curve (ROC-AUC) was calculated for both OS and RFS prediction. The Kaplan-Meier method was used to generate prognostic and survival estimates for low- and high-risk groups, along with net reclassification improvement (NRI), integrated discrimination improvement (IDI), and decision curve analysis. An additional subanalysis was conducted to examine the relationship between clinical data and imaging features associated with risk prediction. Results: Our model achieved AUC values of 0.785 (95% confidence interval [CI]: 0.716-0.855) and 0.695 (95% CI: 0.603-0.787) on the testing and external data sets for OS prediction, and 0.726 (95% CI: 0.653-0.800) and 0.700 (95% CI: 0.615-0.785) for RFS prediction. Additional survival analyses indicated that our model outperformed the present TNM and ResNet-Graph models in terms of net benefit for survival prediction. Conclusions: Our Transformer-Graph model was effective at predicting survival in patients with early stage lung cancer, which was constructed using both imaging and non-imaging clinical features. Some high-risk patients were distinguishable by using a similarity score function defined by non-imaging characteristics such as age, gender, histology type, and tumour location, while Transformer-generated features demonstrated additional benefits for patients whose non-imaging characteristics were non-discriminatory for survival outcomes. Funding: The study was supported by the National Natural Science Foundation of China (91959126, 8210071009), and Science and Technology Commission of Shanghai Municipality (20XD1403000, 21YF1438200).