Sphingolipid metabolism is an important component of various biological processes, particularly in cancer pathology. This metabolic pathway significantly influences the behavior of cancer cells by regulating growth, apoptosis, and survival. Although modulating sphingolipid metabolism has attracted attention as a novel therapeutic strategy, its complexity and specific mechanisms remain incompletely understood. In the current study, transcriptomic profiling was employed to compare the expression of sphingolipid metabolism-related genes between normal solid tissues and lung adenocarcinoma tissues. Additionally, Gene Ontology, Kyoto Encyclopedia of Genes and Genomes enrichment, and protein–protein interaction (PPI) analyses were performed to investigate the functional relevance of these genes. Twelve sphingolipid metabolism-related genes formed a highly interconnected core, suggesting their potential central role within the regulatory network. Four genes were found to be significantly correlated with overall survival. Notably, B4GALNT1 upregulation and CERS4 downregulation correlated with advanced tumor stage and metastasis. They also showed prognostic significance in Cox regression analyses, and these findings were consistently validated in an independent cohort. In vitro, within lung adenocarcinoma cell lines, B4GALNT1 knockdown and CERS4 overexpression suppressed cell proliferation, migration, and epithelial-to-mesenchymal transition, supporting their roles in lung adenocarcinoma progression. These findings highlight B4GALNT1 and CERS4 as potential prognostic biomarkers and therapeutic targets in lung adenocarcinoma, warranting further clinical investigation.

Spread through air spaces (STAS) is recognized as a novel invasive mode of lung adenocarcinoma (LADC), linked with poorer prognosis and high risk of recurrence. The aim of this study was to develop and evaluate a radiomics nomogram using computed tomography (CT)-based tumoral and peritumoral radiomics features for preoperatively predicting STAS status in clinical stage I pure-solid LADC. This study retrospectively enrolled 308 individuals with stage I LADC appearing as pure-solid nodules on thin-section CT who underwent surgical resection from three institutions. We randomly split the patients at authors' hospital into a training set (n=174) and internal validation set (n=73) in a ratio of 7:3, while the external validation set consisted of 61 patients from the other two hospitals. The radiomics features extracted from the gross tumor volume (GTV), two types of peritumoral tumor volume (PTV) (5 and 10 mm around the tumor), and their corresponding two types of gross peritumoral tumor volume (GPTV) were utilized to construct five radiomics models, respectively. Univariate and multivariate analyses identified the independent predictors of STAS. The radscore of the radiomics model with optimal performance was integrated with clinical predictor to develop a comprehensive nomogram. The STAS positive status was found in 118 (38.3%) of the 308 patients {female: 54.2%; median [interquartile range (IQR)] age: 65, [57-72] years}. The GPTV10 model achieved the highest area under the curve (AUC) values of 0.741, 0.737 and 0.741 in three cohorts. The multivariate logistic regression (LR) suggested that micropapillary component was the independent risk factor of pathological STAS. The comprehensive model constructed using the GPTV10 radscore and clinical predictor exhibited AUCs of 0.788, 0.748 and 0.783. The decision curve analysis (DCA) revealed that the nomogram had superior capacity for predicting STAS status in LADC. Furthermore, both pathological STAS status and STAS predicted by the combined model stratified patients for prognosis, with 5-year recurrence-free survival (RFS) showing obvious difference between STAS-positive and STAS-negative. Peritumoral features were significantly correlated with STAS status. The integration of radiomics characteristics and clinical factor provided better performance in the prediction of STAS status.

ObjectiveThis study seeks to create and assess a combined radiomics model that combines intratumoral habitat features with peritumoral characteristics from CT imaging to predict spread through air spaces (STAS) in ≤ 2 cm solid lung adenocarcinomas.Materials and methodsA total of 401 patients with solid invasive lung adenocarcinomas ≤ 2 cm from two centers were retrospectively enrolled (training cohort: 217 cases, validation cohort: 93 cases, test cohort: 91 cases). Univariate and multivariate logistic regression analyses were employed to assess both CT features and clinical data, aiming to determine independent predictors of STAS. Regions of interest (ROI) for tumors were delineated on CT images, with peritumoral regions expanded by 1 mm, 3 mm, and 5 mm. Tumors were further segmented into three habitat subregions using K-means clustering. Radiomic features were extracted from the intratumoral, peritumoral, and habitat regions, and five machine learning algorithms were applied to construct predictive models. The best-performing predictive model was selected and further integrated into a combined model. Performance was assessed by receiver operating characteristic (ROC) curve’s area under the curve (AUC), calibration curves, and decision curve analysis (DCA).ResultsThe habitat model outperformed the Intra model, and the Peri3mm model surpassed Peri1mm and Peri5mm models. The integration of habitat, Peri3mm, and clinical models yielded a substantial improvement in predictive performance, with AUCs reaching 0.948, 0.897, and 0.930 in the training, validation, and test sets, respectively. Calibration curves and DCA confirmed favorable fit and higher clinical net benefit.ConclusionThe combined model provides high accuracy for predicting STAS in solid lung adenocarcinomas with a diameter of ≤ 2 cm, offering valuable support for treatment decision-making.

Lung adenocarcinoma (LUAD) displays marked intratumoral heterogeneity with distinct histological patterns. The solid pattern representing poorly differentiated LUAD is linked to poor prognosis and therapeutic resistance. To uncover underlying mechanisms, we integrate bulk and single-cell RNA sequencing and identify a preferential enrichment of interleukin 4 induced 1 (IL4I1)-expressing tumor-associated macrophages (TAMs) and tryptophan 2,3-dioxygenase (TDO2)-expressing myofibroblastic cancer-associated fibroblasts (myCAFs) in a solid pattern of LUAD. Spatial transcriptomics reveals their co-localization in peritumoral stroma, forming an immune-excluded niche. Mechanistically, TDO2⁺ myCAFs promoted monocyte-to-IL4I1⁺ TAM differentiation via the kynurenine-aryl hydrocarbon receptor (AhR) axis. Tryptophan metabolomic landscapes confirm that IL4I1⁺ TAMs and TDO2⁺ myCAFs enhance tryptophan degradation and accumulation of AhR ligands (e.g., kynurenine, indole-3-carboxaldehyde), contributing to CD8⁺ T cell exhaustion and anti-PD-1 therapeutic resistance. IL4I1⁺ TAMs and TDO2⁺ myCAFs conformably mediate ferroptosis resistance through the AhR-NRF2-GPX4-SLC7A11 pathway. Notably, AhR antagonist CH-223191 restores ferroptosis sensitivity of tumor cells. A triple therapy combining CH-223191, ferroptosis inducer (Imidazole ketone erastin or RSL3), and anti-PD-1 agent demonstrates superior efficacy and safety in vivo. Together, our findings demonstrate that IL4I1⁺ TAMs and TDO2⁺ myCAFs synergistically establish an immunosuppressive, ferroptosis-resistant niche via AhR signaling in solid predominant LUAD and offer promising therapeutic strategies to reprogram the tumor microenvironment.

Mathematical comparisons of lung adenocarcinoma (LUAD) with epidermal growth factor receptor (EGFR) mutations of L858R and 19DEL. A total of 1,980 surgically resected LUAD cases were analyzed, comprising 1,195 ground glass opacity LUAD (GGO-LUAD) and 785 pure solid lung adenocarcinomas (pSD-LUAD). LUAD cases were grouped based on EGFR mutation status. Correlations between tumor size (x) and frequency ratio (y) were analyzed using exponential equations. Multivariate logistic regression analysis showed that the 19DEL mutation was more frequently observed in younger individuals (OR: 1.60, 95% CI: 1.29-1.99, p < 0.0001) and high-grade tumors (OR: 1.46, 95% CI: 1.01-2.11, p = 0.004). Equations were fitted as follows: y = e-0.046 + 0.48 × (x ≤ 2.5 cm, p = 0.0016) (1), y = e-2.09 + 0.37 × (x ≤ 4 cm, p < 0.0001) (2), y = e-1.65 + 0.38 × (x ≤ 4 cm, p = 0.0001) (3). pGGO-LUAD with L858R EGFR mutations grew more rapidly than those with 19DEL mutations, but the solidification rate of L858R-mutated GGO-LUAD was lower compared to 19DEL-mutated GGO-LUAD.

Correlation between radiologic feature and spread through air space in lung solid adenocarcinoma 30.0mm or less in maximum diameter.

To assess the clinical characteristics and CT findings associated with epidermal growth factor receptor (EGFR) mutation status in stage III-IV solid lung adenocarcinoma (LAD) patients. In this retrospective study, stage III-IV solid LAD patients who underwent chest CT from January 2015 to July 2025 were included. Clinical characteristics and CT findings significantly associated with the EGFR mutation status were identified via multivariable logistic regression. A total of 420 patients with stage III-IV solid LAD were included (training cohort: 375 patients, from January 2015 to April 2024; validation cohort: 45 patients, from May 2024 to July 2025). Compared with wild-type EGFR patients, EGFR-mutant LAD were significantly younger (< 60 years), more likely to be female, nonsmokers, and to have stage IV disease. In terms of CT findings, patients with mutant EGFR were more likely to have a tumor size < 4.5cm, a well-defined tumor boundary, a vessel convergence sign, pleural indentation and obstructive pneumonia or atelectasis. In the multivariable analysis, age (OR, 0.428; 95% CI 0.242-0.756), sex (OR, 0.200; 95% CI 0.112-0.356), overall stage (OR, 2.230; 95% CI 1.141-4.359), tumor size (OR, 0.474; 95% CI 0.260-0.864, P = 0.015), tumor boundary (OR, 3.461; 95% CI 1.877-6.382), the presence of the vessel convergence sign (OR, 2.869; 95% CI 1.675-4.913), and obstructive pneumonia or atelectasis (OR, 3.870; 95% CI 2.028-7.385) were identified as factors that independently predict the EGFR mutation status. We further constructed a nomogram for predicting the EGFR mutation status via a logistic regression model. Logit (P) = 0.197 + (-1.599)×sex + (-0.850)×age + 0.900×overall stage + (-0.762)×tumor size + 1.246×tumor boundary + 1.042×vessel convergence sign + 1.367×obstructive pneumonia or atelectasis. The area under the curve (AUC) of the nomogram in training cohort was 0.829 (95% CI: 0.783, 0.876). In the validation cohort, the AUC was 0.826 (95% CI: 0.681, 0.970). A nomogram including sex, age, overall stage, tumor size, tumor boundary, the vessel convergence sign, and obstructive pneumonia or atelectasis, was helpful in predicting the EGFR mutation status in stage III-IV solid LAD patients.

BackgroundLung adenocarcinomas (LACs) are classified into several histological types. The clinical outcomes are not consistent in epidermal growth factor receptor-tyrosine kinase inhibitor (EGFR-TKI) resistant LACs treated by immune checkpoint inhibitors (ICIs). This study aims to explore the regulatory mechanisms of EGFR-mutant LAC and programmed cell death ligand 1 (PD-L1) protein expression in the tumor microenvironment (TME), as well as the pathological responses and related mechanisms of these pathways in osimertinib-resistant patients after ICI treatment. The effects of transforming growth factor-β (TGF-β) signaling pathway which remodels TME should be investigated.MethodsGene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyse differentially expressed genes (DEGs) from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) (GSE68465) cohorts. Expression of PD-L1 and cytokines were analyzed by immunohistochemistry and cytometric bead array. Osimertinib-resistant solid LAC xenografts were implanted into immune-humanized mice to demonstrate pathological responses after EGFR-TKI and ICI. TGF-β signaling proteins, pERK1/2, pSTAT3, and PD-L1 was detected by western blotting. Clinical responses were observed in 12 osimertinib-resistant patients treated by ICI.ResultsGO and KEGG analyses highlighted “cytokine-cytokine receptor interaction” and “TGF-beta signaling pathway”. Heterogeneous PD-L1 expression was regulated by synergistic IL-6/STAT3, TGF-β and EGFR pathways. Significant tumor remission and inhibition of TGF-β signaling pathway were found in EGFR-TKI plus ICI treatment in mice models. Osimertinib-resistant solid LACs treated by ICI had better clinical outcomes.ConclusionsTGF-β signaling pathway is associated with PD-L1 expression and the pathological remission treated by EGFR-TKI and ICI. Solid osimertinib-resistant LACs by ICI can be beneficial through the inhibition of TGF-β signaling pathway.

Development of a PET-CT Based Radiomics Model for Preoperative Prediction of the Novel IASLC Grading and Prognosis in Patients with Clinical Stage I Pure Solid Invasive Lung Adenocarcinoma.

PurposeCompared to solid lung adenocarcinomas (LUADs), part-solid LUADs rarely exhibit lymph node metastasis (LNM) and generally have a favorable prognosis. This study aims to comprehensively investigate the clinical, pathological, and CT characteristics of part-solid LUADs with LNM.Patients and MethodsThis study collected 70 pathologically confirmed part-solid LUADs at two centers, including 35 cases with LNM and 35 matched cases without LNM based on size, CT pattern, and pathological subtype. Their clinical, pathological, and CT features were comprehensively analyzed and compared to identify the characteristics of part-solid LUADs associated with a high risk of LNM.ResultsAmong the 3,457 IACs manifested as part-solid lesions, a total of 35 (1.01%) cases were found to be associated with LNMs. Clinically, patients with and without LNM were similar. Pathologically, lesions exhibiting predominant micropapillary/solid pattern (11.4% vs 0.0%), and containing micropapillary (48.6% vs 25.7%) or any high-grade histological pattern were all more common in part-solid LUADs with LNM than in those without (each P < 0.05). Radiologically, solid components located at the tumor margins or distributed in a scattered manner (odds ratio [OR] = 4.048, P = 0.038) and consolidation-to-tumor ratio (CTR) > 57.2% (area) (OR = 45.649, P = 0.041) were independent predictors of LNM, with an area under the curve of this model being 0.881, sensitivity of 97%, and specificity of 77.1% (P < 0.001).ConclusionLNM in part-solid LUADs is more prevalent in IACs with high-grade patterns, particularly the micropapillary pattern, with these lesions presenting as part-solid lesions that often have a larger CTR or distinct distribution of solid components.

The presence of occult lymph node metastasis (OLNM) has significant implications for the staging, treatment and prognosis of patients with tumors. This study aimed to explore the potential predictive value of dual-layer spectral detector computed tomography (SDCT) quantitative parameters obtained via multimodal tumor tracking (MMTT) for OLNM in clinical stage 1 (c1) pure solid lung adenocarcinoma (LAC). A total of 123 patients diagnosed with c1 pure solid LAC were enrolled in this retrospective study, including 29 OLNM(+) and 94 OLNM(-) patients. MMTT was used to obtain the full volume of the tumor and extract quantitative spectral parameters. The differences in clinicopathological data and spectral quantitative parameters between OLNM(+) and OLNM(-) were compared, and the efficacy of related quantitative parameters in predicting OLNM were evaluated through receiver operating characteristic (ROC) curves. Five different machine learning (ML) methods were applied to construct a forecasting OLNM model. The tumor volume (3.868 vs. 1.236 cm3) and proportion of patients with carcinoembryonic antigen (CEA) >5 ng/mL (41.4% vs. 10.6%) in OLNM(+) patients were significantly greater than those in OLNM(-) patients (P<0.001). Among the 17 SDCT quantitative parameters, except for the tumor-to-aortal virtual plain scan ratio (SARVNC), tumor-to-aortal enhancement ratio (SAR)70keV and SAR100keV (P=0.25, 0.11, 0.98), the remaining 14 quantitative parameters [SAR40keV, Δ40keV, Δ70keV, Δ100keV, contrast enhancement ratio (CER)40keV, CER70keV, CER100keV, normalized enhancement fraction (NEF)40keV, NEF70keV, NEF100keV, λ40-70keV, λ40-100keV, normalized iodine concentration (NIC), normalized effective atomic number (NZeff)] in the OLNM(+) group were significantly lower than those in the OLNM(-) group (P<0.05). NEF40keV, NEF70keV and NIC had high diagnostic efficiency in predicting OLNM, with area under the curves (AUCs) of 0.710, 0.705 and 0.701, respectively. The multilayer perceptron (MLP) model achieved the best diagnostic performance among the five ML methods, with a higher average AUC of 0.778. SDCT quantitative parameters obtained via MMTT might offer new insights to help predict OLNM in c1 pure solid LAC patients. The prediction model constructed with the MLP model on the basis of clinicopathological data and spectral quantitative parameters has higher diagnostic efficiency and may further aid in clinical decision-making.

Pulmonary granulomatous nodules (PGN) often exhibit similar CT morphological features to solid lung adenocarcinomas (SLA), making preoperative differentiation challenging. This study aims to address this diagnostic challenge by developing a novel deep learning model. This study proposes MAEMC-NET, a model integrating generative (Masked AutoEncoder) and contrastive (Momentum Contrast) self-supervised learning to learn CT image representations of intra- and inter-solitary nodules. A generative self-supervised task of reconstructing masked axial CT patches containing lesions was designed to learn intra- and inter-slice image representations. Contrastive momentum is used to link the encoder in axial-CT-patch path with the momentum encoder in coronal-CT-patch path. A total of 494 patients from two centers were included. MAEMC-NET achieved an area under curve (95% Confidence Interval) of 0.962 (0.934-0.973). These results not only significantly surpass the joint diagnosis by two experienced chest radiologists (77.3% accuracy) but also outperform the current state-of-the-art methods. The model performs best on medical images with a 50% mask ratio, showing a 1.4% increase in accuracy compared to the optimal 75% mask ratio on natural images. The proposed MAEMC-NET effectively distinguishes between benign and malignant solitary pulmonary nodules and holds significant potential to assist radiologists in improving the diagnostic accuracy of PGN and SLA.

BackgroundThis study aims to quantify intratumoral heterogeneity (ITH) using preoperative CT image and evaluate its ability to predict pathological high-grade patterns, specifically micropapillary and/or solid components (MP/S), in patients diagnosed with clinical stage I solid lung adenocarcinoma (LADC).MethodsIn this retrospective study, we enrolled 457 patients who were postoperatively diagnosed with clinical stage I solid LADC from two medical centers, assigning them to either a training set (n = 304) or a test set (n = 153). Sub-regions within the tumor were identified using the K-means method. Both intratumoral ecological diversity features (hereafter referred to as ITH) and conventional radiomics (hereafter referred to as C-radiomics) were extracted to generate ITH scores and C-radiomics scores. Next, univariate and multivariate logistic regression analyses were employed to identify clinical-radiological (Clin-Rad) features associated with the MP/S (+) group for constructing the Clin-Rad classification. Subsequently, a hybrid model which presented as a nomogram was developed, integrating the Clin-Rad classification and ITH score. The performance of models was assessed using the receiver operating characteristic (ROC) curves, and the area under the curve (AUC), accuracy, sensitivity, and specificity were determined.ResultsThe ITH score outperformed both C-radiomics scores and Clin-Rad classification, as evidenced by higher AUC values in the training set (0.820 versus 0.810 and 0.700, p = 0.049 and p = 0.031, respectively) and in the test set (0.805 versus 0.771 and 0.732, p = 0.041 and p = 0.025, respectively). Finally, the hybrid model consistently demonstrated robust predictive capabilities in identifying presence of MP/S components, achieving AUC of 0.830 in the training set and 0.849 in the test set (all p < 0.05).ConclusionThe ITH derived from sub-region within the tumor has been shown to be a reliable predictor for MP/S (+) in clinical stage I solid LADC.

This study aimed to evaluate the possibility of defining new imaging criteria to predict less-invasive clinical (c)-stage IA2-IA3 solid predominant lung adenocarcinoma using the maximum standardized uptake value (SUVmax) as the cutoff value. Consecutive 364 patients who underwent anatomical resection with mediastinal lymphadenectomy and positron emission tomography for c-stage IA2-IA3 solid predominant lung adenocarcinoma with a tumor diameter < 3cm were retrospectively evaluated. Less-invasive cancer was defined as the absence of nodal involvement, lymphovascular or pleural invasion, or spread through air spaces. The SUVmax cutoff value was determined based on the specificity of the receiver operating characteristic curve. 228 were pure-solid tumors, and 136 were part-solid tumors. 212 were c-stage IA2 and 152 were c-stage IA3. When the SUVmax was set at a cutoff value of 2.2, sensitivity and specificity were 33.0% and 97.6%, respectively, and it was possible to secure the sensitivity by more than 30% with high specificity among the solid predominant tumors. When the SUVmax was set at a cutoff value of 2.2, sensitivity and specificity were 40.7% and 95.7%, respectively, in whole tumor diameter ≤ 2cm, and 27.0% and 99.0%, respectively in whole tumor diameter between 2 and 3cm. When the SUVmax was set at a cutoff value of 2.2, sensitivity and specificity were 45.8% and 96.6%, respectively, in part-solid tumors, and 17.8% and 97.8%, respectively in pure-solid tumors. Setting the SUVmax as cutoff value could predict pathologically less-invasive cancers in c-stage IA2-IA3 solid predominant lung adenocarcinoma.

BackgroundSpread through air spaces (STAS) is a well-established factor associated with poor oncological outcomes in patients undergoing surgery for solid lung adenocarcinoma. There could potentially be a disparity in iodine uptake between patients with positive and negative airway spread of solid lung adenocarcinoma.PurposeTo explore the associations and find correlations of iodine uptake with STAS status in patients who underwent surgery for solid lung adenocarcinoma.Material and MethodsPatients who underwent solid lung adenocarcinoma resection between January and June 2022 were included in this retrospective study. Iodine concentration and CT features were assessed using contrast-enhanced dual-energy computed tomography (DECT) scans, and these were compared with the status of STAS.ResultsOf 52 patients included, 25 (48%) were STAS-positive and 27 (52%) were STAS-negative. There were no statistically significant differences in CT features between the two groups (P > 0.05). STAS-positive was significantly associated with low arterial phase iodine concentration (ICA), normalized arterial phase iodine concentration (NICA), and venous phase iodine concentration (ICV), with a cutoff established at 1.15 mg/mL, 0.11, and 1.35 mg/mL, respectively (P < 0.05). The AUCs for ICA, NICA, and ICV in predicting STAS in solid lung adenocarcinoma were 0.82, 0.83, and 0.73, respectively. ICA and NICA were identified as independent risk factors for STAS in solid lung adenocarcinoma, with a combined AUC of 0.89.ConclusionThis study suggests that solid lung adenocarcinoma patients with low ICA, NICA, and ICVA were associated with STAS-positive, as well as a worse survival outcomes.

There is still a debate regarding the prognostic implication of lymphovascular invasion (LVI) in stage I lung adenocarcinoma. Ground-glass opacity (GGO) on CT is known to correlate with a less invasive or lepidic component in adenocarcinoma, which may influence the strength of prognostic factors. This study aimed to explore the prognostic value of LVI in stage I lung adenocarcinoma based on the presence of GGO. Stage I lung adenocarcinoma patients receiving lobectomy between 2010 and 2019 were retrospectively categorized as GGO-positive or GGO-negative (solid adenocarcinoma) on CT. Multivariable Cox regression analyses were performed for disease-free survival (DFS) and overall survival (OS) to evaluate the prognostic significance of pathologic LVI based on the presence of GGO. Of 924 patients included (mean age, 62.5 ± 9.2 years; 505 women), 525 (56.8%) exhibited GGO-positive adenocarcinoma and 116 (12.6%) were diagnosed with LVI. LVI was significantly more frequent in solid than GGO-positive adenocarcinoma (20.1% vs. 6.9%, p < 0.001). Multivariable analysis identified LVI and visceral pleural invasion (VPI) as significant prognostic factors for shorter DFS among solid adenocarcinoma patients (LVI, hazard ratio (HR): 1.89, p = 0.004; VPI, HR: 1.65, p = 0.003) but not GGO-positive patients (p = 0.76 and p = 0.87). In contrast, LVI was not a significant prognostic factor for OS in either group (p > 0.05). In stage I lung adenocarcinoma, pathologic LVI was associated with DFS only in patients with solid lung adenocarcinoma. Lymphovascular invasion (LVI) significantly affects disease-free survival in solid-stage I lung adenocarcinoma patients, but not those with ground-glass opacity (GGO) adenocarcinoma. Risk stratification considering both GGO on CT and LVI may identify patients benefiting from increased surveillance. The presence of ground-glass opacity portends different prognoses for lung adenocarcinoma. In stage I lung adenocarcinoma, lymphovascular invasion (LVI) was significantly more frequent in solid adenocarcinomas than in ground-glass opacity (GGO)-positive adenocarcinomas. LVI was not associated with overall survival in patients with either solid adenocarcinomas or GGO adenocarcinomas.

Dual-energy CT Radiomics Combined with Quantitative Parameters for Differentiating Lung Adenocarcinoma From Squamous Cell Carcinoma: A Dual-center Study

BackgroundThe diagnosis of solitary pulmonary nodules has always been a difficult and important point in clinical research, especially granulomatous nodules (GNs) with lobulation and spiculation signs, which are easily misdiagnosed as malignant tumors. Therefore, in this study, we utilised a CT deep learning (DL) model to distinguish GNs with lobulation and spiculation signs from solid lung adenocarcinomas (LADCs), to improve the diagnostic accuracy of preoperative diagnosis.Methods420 patients with pathologically confirmed GNs and LADCs from three medical institutions were retrospectively enrolled. The regions of interest in non-enhanced CT (NECT) and venous contrast-enhanced CT (VECT) were identified and labeled, and self-supervised labels were constructed. Cases from institution 1 were randomly divided into a training set (TS) and an internal validation set (IVS), and cases from institutions 2 and 3 were treated as an external validation set (EVS). Training and validation were performed using self-supervised transfer learning, and the results were compared with the radiologists’ diagnoses.ResultsThe DL model achieved good performance in distinguishing GNs and LADCs, with area under curve (AUC) values of 0.917, 0.876, and 0.896 in the IVS and 0.889, 0.879, and 0.881 in the EVS for NECT, VECT, and non-enhanced with venous contrast-enhanced CT (NEVECT) images, respectively. The AUCs of radiologists 1, 2, 3, and 4 were, respectively, 0.739, 0.783, 0.883, and 0.901 in the (IVS) and 0.760, 0.760, 0.841, and 0.844 in the EVS.ConclusionsA CT DL model showed great value for preoperative differentiation of GNs with lobulation and spiculation signs from solid LADCs, and its predictive performance was higher than that of radiologists.

Micropapillary (MP) and solid (S) pattern adenocarcinoma are highly malignant subtypes of lung adenocarcinoma. In today's era of increasingly conservative surgery for small lung cancer, effective preoperative identification of these subtypes is greatly important for surgical planning and the long-term survival of patients. For this retrospective study, the presence of MP and/or S was evaluated in 2167 consecutive patients who underwent surgical resection for clinical stage IA1-2 lung adenocarcinoma. MP and/or S pattern-positive patients and negative-pattern patients were matched at a ratio of 1:3. The Lasso regression model was used for data dimension reduction and imaging signature building. Multivariate logistic regression was used to establish the predictive model, presented as an imaging nomogram. The performance of the nomogram was assessed based on calibration, identification, and clinical usefulness, and internal and external validation of the model was conducted. The proportion of solid components (PSC), Sphericity, entropy, Shape, bronchial honeycomb, nodule shape, sex, and smoking were independent factors in the prediction model of MP and/or S lung adenocarcinoma. The model showed good discrimination with an area under the receiver operating characteristic curve of 0.85. DCA demonstrated that the model could achieve good benefits for patients. Restricted cubic spline analysis suggested a significant increase in the proportion of MP and/or S from 11 to 48% when the PSC value was 68%. Small MP and/or S adenocarcinoma can be effectively identified preoperatively by their typical three-dimensional and 2D imaging features.

The preoperative differential diagnosis of nodular lung adenocarcinoma has long been a challenging issue for thoracic surgeons. This study aimed to explore differential diagnosis of nodular lung adenocarcinoma by comprehensively analyzing its clinical, computed tomography (CT) imaging, and postoperative pathological and genetic features. The clinical, CT imaging, and postoperative pathological features of different classifications of nodular lung adenocarcinoma were retrospectively analyzed through univariate and multivariate statistical methods. There were 132 patients with nodular lung adenocarcinoma enrolled. Firstly, compared with ground-glass nodular lung adenocarcinoma, solid nodular lung adenocarcinoma was more common in women [odds ratio (OR), 3.662; 95% confidence interval (CI): 1.066-12.577] and older adults (OR, 1.061; 95% CI: 1.007-1.119), and CT signs were mostly lobulation (OR, 4.957; 95% CI: 1.714-14.337) and spiculation (OR, 8.214; 95% CI: 2.740-24.621); the mean CT (CTm) value of solid nodular lung adenocarcinoma was significantly higher than that of ground-glass nodular lung adenocarcinoma, and the optimal diagnostic threshold was -267.5 Hounsfield units (HU). Secondly, the maximum diameter of nodule size (NSmax) of invasive adenocarcinoma (IAC) was significantly greater than that of minimally IAC (MIA; OR, 6.306; 95% CI: 1.191-33.400) or atypical adenomatous hyperplasia (AAH)/adenocarcinoma in situ (AIS; OR, 189.539; 95% CI: 4.720-7,610.476), and the optimal diagnostic threshold between IAC and MIA was 1.35 cm; the CTm value of IAC was significantly higher than that of MIA, and the optimal diagnostic threshold was -460.75 HU. Thirdly, lepidic-predominant adenocarcinoma (LPA) manifest more commonly as pure ground-glass nodule (pGGN; OR, 6.252; 95% CI: 1.429-27.358) or mixed ground-glass nodule (mGGN; OR, 4.224; 95% CI: 1.223-14.585). Moreover, the mutation rate of epidermal growth factor receptor (EGFR) in IAC was 70.69% (41/58). The EGFR mutation rates of mGGNs (OR, 8.794; 95% CI: 1.489-51.933) and solid nodules (SNs; OR, 12.912; 95% CI: 1.597-104.383) were significantly higher than that of pGGNs. Furthermore, compared with those of micropapillary-predominant adenocarcinoma (MPA), solid-predominant adenocarcinoma (SPA), or invasive mucinous adenocarcinoma (IMA), there were significantly higher EGFR mutation rates in acinar-predominant adenocarcinoma/papillary-predominant adenocarcinoma (APA/PPA; OR, 55.925; 95% CI: 4.045-773.284) and LPA (OR, 38.265; 95% CI: 2.307-634.596). Different classifications of nodular lung adenocarcinoma have their own clinicopathological and CT imaging features, and the latter is the main predictor.