Res-TransNet: A Hybrid deep Learning Network for Predicting Pathological Subtypes of lung Adenocarcinoma in CT Images.

Objective: To explore the value of quantitative parameters of artificial intelligence (AI) and computed tomography (CT) signs in identifying pathological subtypes of lung adenocarcinoma appearing as ground-glass nodules (GGNs). Methods: CT images of 224 GGNs from 210 individuals were collected retrospectively and classified into atypical adenomatous hyperplasia (AAH)/adenocarcinoma in situ (AIS), minimally invasive adenocarcinoma (MIA), and invasive adenocarcinoma (IAC) groups. AI was used to identify GGNs and to obtain quantitative parameters, and CT signs were recognized manually. The mixed predictive model based on logistic multivariate regression was built and evaluated. Results: Of the 224 GGNs, 55, 93, and 76 were AAH/AIS, MIA, and IAC, respectively. In terms of AI parameters, from AAH/AIS to MIA, and IAC, there was a gradual increase in two-dimensional mean diameter, three-dimensional mean diameter, mean CT value, maximum CT value, and volume of GGNs (all P<0.0001). Except for the CT signs of the location, and the tumor–lung interface, there were significant differences among the three groups in the density, shape, vacuolar signs, air bronchogram, lobulation, spiculation, pleural indentation, and vascular convergence signs (all P<0.05). The areas under the curve (AUC) of predictive model 1 for identifying the AAH/AIS and MIA and model 2 for identifying MIA and IAC were 0.779 and 0.918, respectively, which were greater than the quantitative parameters independently (all P<0.05). Conclusion: AI parameters are valuable for identifying subtypes of early lung adenocarcinoma and have improved diagnostic efficacy when combined with CT signs.

This study aimed to investigate the predictive value of intratumoral and peritumoral radiomics model for the cribriform component (CC) of invasive lung adenocarcinoma (LUAD). The 144 patients with invasive LUAD from our center were randomly divided into training set (n = 100) and internal validation set (n = 44) in a ratio of 7:3, and 75 patients from center 2 were regarded as the external validation set. Clinical risk factors were examined using univariate and multivariate logistic regression to construct the clinical model. We extracted radiomics features from gross tumor volume (GTV), gross and peritumoral volume (GPTV), and peritumoral volume (PTV), respectively. Radiomics models were constructed with selected features. A combined model based on the optimal Radscore and clinically independent predictors was constructed, and its predictive performance was assessed by receiver operating characteristic curve (ROC), calibration curve, and decision curve analysis (DCA). The area under curves (AUCs) of the GTV model were 0.882 (95% CI 0.817-0.948), 0.794 (95% CI 0.656-0.932), and 0.766 (95% CI 0.657-0.875) in the training, internal validation, and external validation sets, and the PTV model had AUCs of 0.812 (95% CI 0.725-0.899), 0.749 (95% CI 0.597-0.902), and 0.670 (95% CI 0.543-0.798) in the training, internal validation, and external validation sets, respectively. However, the GPTV radiomics model showed better predictive performance compared with the GTV and PTV radiomics models, with the AUCs of 0.950 (95% CI 0.911-0.989), 0.844 (95% CI 0.728-0.959), and 0.815 (95% CI 0.713-0.917) in the training, internal validation and external validation sets, respectively. In the clinical model, tumor shape, lobulation sign and maximal diameter were the independent predictors of CC in invasive LUAD. The combined model including independent clinical predictors and GPTV-Radscore show the considerable instructive to clinical practice, with the AUCs of 0.954(95% CI 0.918-0.990), 0.861(95% CI 0.752-0.970), and 0.794(95% CI 0.690-0.898) in training, internal validation, and external validation sets, respectively. DCA showed that the combined model had good clinical value and correction effect. Radiomics model is a very powerful tool for predicting CC growth pattern in invasive LUAD and can help clinicians make the strategies of treatment and surveillance in patients with invasive LUAD.

Lung adenocarcinoma is the most prevalent type of lung cancer, with invasive lung adenocarcinoma being the most common subtype. Screening and early treatment of high-risk individuals have improved survival; however, significant differences in prognosis still exist among patients at the same stage, especially in the early stages. Invasive lung adenocarcinoma has different histological morphologies and biological characteristics that can distinguish its prognosis. Notably, several studies have found that the pathological subtypes of invasive lung adenocarcinoma are closely associated with clinical treatment. This review summarised the distribution of various pathological subtypes of invasive lung adenocarcinoma in the population and their relationship with sex, smoking, imaging features, and other histological characteristics. We comprehensively analysed the genetic characteristics and biomarkers of the different pathological subtypes of invasive lung adenocarcinoma. Understanding the interaction between the pathological subtypes of invasive lung adenocarcinoma and the tumour microenvironment helps to reveal new therapeutic targets for lung adenocarcinoma. We also extensively reviewed the prognosis of various pathological subtypes and their effects on selecting surgical methods and adjuvant therapy and explored future treatment strategies.

Background: To test the value of Computed tomography (CT) features in predicting the infiltration degree and pathological subtype of ground glass lung adenocarcinoma (≤ 3 cm). Materials and Methods: Data from 412 lung adenocarcinoma patients with mixed ground glass nodules on CT from Jan. 2017 to Dec. 2021 were tested retrospectively. The patients were separated by the infiltrating degree into a minimally invasive adenocarcinoma (MIA) group and an invasive adenocarcinoma (IAC) group. Then the IAC group was subdivided into low-, medium-and high-risk groups by the prognosis differences among subtypes, which were of lepidic, papillary, and micropapillary predominance respectively. Results: Average diameter of nodules, average CT value, solid component ratio, lobe sign, and burr sign were independent risk factors of IAC. The average diameter of nodules ≥ 12.5 mm, solid component ratio ≥ 20.96%, average CT value ≥ -473.07 HU, positive lobe sign and positive burr sign indicated the nodules were more likely IAC. Average CT value, and solid component ratio were independent risk factors for the high-risk pathological type of lung adenocarcinoma. The average CT value ≥ -242.92 HU and solid component ratio ≥ 69.536% indicated nodules were more likely the high-risk pathological type of lung adenocarcinoma. Conclusion: CT imaging features improve the diagnostic efficacy of ground glass nodules, and have certain clinical value.

Purpose: The aims of this study were to combine CT images with Ki-67 expression to distinguish various subtypes of lung adenocarcinoma and to pre-operatively predict the Ki-67 expression level based on CT radiomic features.Methods: Data from 215 patients with 237 pathologically proven lung adenocarcinoma lesions who underwent CT and immunohistochemical Ki-67 from January 2019 to April 2021 were retrospectively analyzed. The receiver operating curve (ROC) identified the Ki-67 cut-off value for differentiating subtypes of lung adenocarcinoma. A chi-square test or t-test analyzed the differences in the CT images between the negative expression group (n = 132) and the positive expression group (n = 105), and then the risk factors affecting the expression level of Ki-67 were evaluated. Patients were randomly divided into a training dataset (n = 165) and a validation dataset (n = 72) in a ratio of 7:3. A total of 1,316 quantitative radiomic features were extracted from the Analysis Kinetics (A.K.) software. Radiomic feature selection and radiomic classifier were generated through a least absolute shrinkage and selection operator (LASSO) regression and logistic regression analysis model. The predictive capacity of the radiomic classifiers for the Ki-67 levels was investigated through the ROC curves in the training and testing groups.Results: The cut-off value of the Ki-67 to distinguish subtypes of lung adenocarcinoma was 5%. A comparison of clinical data and imaging features between the two groups showed that histopathological subtypes and air bronchograms could be used as risk factors to evaluate the expression of Ki-67 in lung adenocarcinoma (p = 0.005, p = 0.045, respectively). Through radiomic feature selection, eight top-class features constructed the radiomic model to pre-operatively predict the expression of Ki-67, and the area under the ROC curves of the training group and the testing group were 0.871 and 0.8, respectively.Conclusion: Ki-67 expression level with a cut-off value of 5% could be used to differentiate non-invasive lung adenocarcinomas from invasive lung adenocarcinomas. It is feasible and reliable to pre-operatively predict the expression level of Ki-67 in lung adenocarcinomas based on CT radiomic features, as a non-invasive biomarker to predict the degree of malignant invasion of lung adenocarcinoma, and to evaluate the prognosis of the tumor.

BackgroundDue to different treatment method and prognosis of different subtypes of lung adenocarcinomas appearing as ground-glass nodules (GGNs) on computed tomography (CT) scan, it is important to classify invasive adenocarcinomas from non-invasive adenocarcinomas. The purpose of this paper is to build and evaluate the performance of deep learning networks on the differentiation the invasiveness of lung adenocarcinoma appearing as GGNs.MethodsThis retrospective study included 886 GGNs from 794 pathological confirmed patients with lung adenocarcinoma for training and testing the proposed networks. Three deep learning networks, namely XimaNet (deep learning-based classification model), XimaSharp (classification and nodule segmentation model), and Deep-RadNet (deep learning and radiomics combined classification model, i.e., deep radiomics) were built. Three classification tasks, namely task 1: classification of AAH/AIS and MIA, task 2: classification of MIA and IAC, and task 3: classification of non-invasive adenocarcinomas and invasive adenocarcinomas (AAH/AIS&MIA and IAC) were conducted to evaluate the model performance. The Z-test was used to compare the model performance.ResultsThe AUC for classification of AAH/AIS with MIA were 0.891, 0.841 and 0.779 for Deep-RadNet, XimaNet and XimaSharp respectively. The AUC for classification of MIA with IAC were 0.889, 0.785 and 0.778 for three networks and AUC for classification of AAH/AIS&MIA with IAC were 0.941, 0.892 and 0.827 respectively. The performance of deep_RadNet was better than the other two models with the Z-test (P<0.05).ConclusionsDeep-RadNet with the visual heat map could evaluate the invasiveness of GGNs accurately and intuitively, providing a theoretical basis for individualized and accurate medical treatment of patients with GGNs.

Background Patients with human immunodeficiency virus (HIV) are more susceptible to liver cancer because of their compromised immune system. There is no specific prognostic model for HIV-infected hepatocellular carcinoma (HCC) patients. Methods Clinical data of 85 patients with HIV-infected HCC was divided into a 7:3 ratio for training and internal validation sets, while the data of 23 patients with HIV-infected HCC was served as the external validation set. Data of 275 HIV-negative HCC patients was considered as external HIV-negative validation set. Variables associated with overall survival (OS) in the training set were used to develop the HIV-infected HCC prognosis (HIHP) model. The model was tested in the internal and external validation sets. The predictive accuracy of the model was assessed with conventional HIV-negative HCC prognostic scoring systems. Results The HIHP model demonstrated a significant association with OS in the training set, with a median OS of 13 months for low risk, 7 months for medium risk, and 3 months for high risk (p &amp;lt; 0.001). The HIHP model showed a significant association with OS, and exhibited greater discriminative abilities compared to conventional HIV-negative HCC prognostic models both in the internal and external validation sets. In the external HIV-negative validation set, the HIHP model did not show better discrimination than conventional HIV-negative HCC scores. Conclusion The new model presented in the work provided a more accurate prognostic prediction of OS in HIV-infected HCC patients. However, the model is not applicable to patients with HIV-negative HCC. Disclosures All Authors: No reported disclosures

The clinical prognostic risk stratification system for HIV infected hepatocellular carcinoma

背景与目的肺癌是国内外致死率最高的恶性肿瘤，肺结节的精确检测是降低肺癌死亡率的关键。人工智能辅助诊断系统在肺结节检测、良恶性鉴别和浸润亚型诊断等领域发展迅速，对其效能进行验证是促进其应用于临床的前提。本研究旨在评估人工智能辅助诊断系统预测肺结节早期肺腺癌浸润亚型的效能。方法回顾性分析2016年1月1日-2021年12月31日期间兰州大学第二医院收治的223例肺结节早期肺腺癌患者的临床资料，将早期肺腺癌分为浸润性腺癌组(n=170)和非浸润性腺癌组(n=53)，其中非浸润性腺癌组又分为微浸润性腺癌组(n=31)和浸润前病变组(n=22)。比较各组的恶性概率和影像特征等信息，分析其对早期肺腺癌浸润亚型的预测能力，并对人工智能辅助诊断早期肺腺癌浸润亚型定性诊断的结果与术后病理进行一致性分析。结果早期肺腺癌不同浸润亚型肺结节的平均CT值(P < 0.001)、直径(P < 0.001)、体积(P < 0.001)、恶性概率(P < 0.001)、胸膜凹陷征(P < 0.001)、分叶征(P < 0.001)、毛刺征(P < 0.001)差异均有统计学意义; 随着早期肺腺癌不同浸润亚型浸润性增加，各组参数显性征象比例也逐渐升高; 在二分类问题上，人工智能辅助诊断系统定性诊断早期肺腺癌浸润亚型的敏感性、特异性及曲线下面积(area under the curve, AUC)分别为81.76%、92.45%和0.871; 在三分类问题上，人工智能辅助诊断系统定性诊断早期肺腺癌浸润亚型的准确率、召回率、F1分数及AUC分别为83.86%、85.03%、76.46%和0.879。结论该人工智能辅助诊断系统对肺结节早期肺腺癌浸润亚型具有一定的预测价值，随着算法的优化和数据的完善或可为患者个体化治疗提供指导。

Objective: To analyze the clinical features and prognostic factors of different histological subtypes of lung adenocarcinoma. Methods: Data from 370 lung adenocarcinoma patients who underwent surgical resection for pathologically supported adenocarcinoma in our hospital between 2000 and 2003 were retro- spectively reviewed. The Kaplan-Meier method was used to estimate patient survival, and Cox’s proportional hazards model was performed for multivariate analysis. Results: The 5-year overall survival rate was 25.26%, and the mean survival time was 3.89 years. In multivariate analysis, histological subtype, incised margin residual, TNM stage, tumor size, and adjuvant chemotherapy were identified as independent survival predictors. The 5-year survival rate in bronchioloalveolar adenocarcinoma (BAC) patients was 41.30%, higher than in patients with other subtypes of lung adenocarcinoma (P=0.002). No significant difference was found in the prognosis among patients with different subtypes of adenocarcinoma without a BAC component. Conclusion: Ade-nocarcinoma with a BAC component is an independent subtype of lung adenocarcinoma. Its prognosis lies between those of BAC and adenocarcinoma without BAC. Histological subtype, incised margin residual, TNM stage, tumor size, and adjuvant chemotherapy are independent survival predictors.

BackgroundDifferentiating small hepatic metastases from hemangiomas can be challenging on visual assessment.PurposeTo evaluate the diagnostic performance of magnetic resonance imaging (MRI) radiomics models based on T2-weighted (T2W) imaging in differentiating small hepatic metastases from hemangiomas.Material and MethodsThis retrospective study included patients with small (≤2 cm) hepatic metastases from colorectal cancer or hemangiomas who underwent liver MRI between August 2018 and January 2024. Datasets were divided into training, internal, and external validation sets based on MRI scanner type. Manual segmentation was performed on conventional T2W, heavily T2W, and fat-suppressed (FS)-T2W imaging. Random forest models were developed using 10-fold cross-validation on 10 selected radiomics features. AUCs were calculated to evaluate model performance. Before segmentation, each hepatic lesion in the validation sets was categorized into metastasis, hemangioma, or indeterminate lesion according to visual assessment on T2W imaging by two radiologists in consensus.ResultsA total of 285 patients (148 men; mean age=55.8 ± 12.5 years) were included: training (140 patients: 151 metastases, 155 hemangiomas), internal (86 patients: 87 metastases, 80 hemangiomas), and external (59 patients: 37 metastases, 69 hemangiomas) validation sets. AUCs for conventional/heavily/FS-T2W imaging were 0.976/0.972/0.946 (training), 0.979/0.991/0.989 (internal validation), and 0.969/0.976/0.809 (external validation), respectively. Among visually indeterminate lesions, 6/7 lesions in the internal validation set and 5/8 lesions in the external validation set were correctly classified using radiomics scores.ConclusionRadiomics models based on T2W imaging exhibit excellent performance in differentiating small hepatic metastases from hemangiomas and may contribute to the correct classification of visually indeterminate hepatic lesions.

Hyaluronidase, hyaluronan synthase, E-cadherin and TGF-Β profile in lung adenocarcinoma subtypes and squamous cell carcinoma of smokers/nonsmokers

Background: Gene expression-based subtyping in lung adenocarcinoma (AD) and lung squamous cell carcinoma (SQ) classifies AD and SQ tumors into distinct subtypes with variable expression of underlying biology including DNA damage response genes. These subtypes are linked to differences in chemotherapy sensitivity, and may impact response to therapeutics like PARP inhibitors. Methods: Using The Cancer Genome Atlas (TCGA) lung cancer gene expression datasets (AD n=515 and SQ n=501), AD subtypes (Terminal Respiratory Unit (TRU), Proximal Proliferative (PP), and Proximal Inflammatory (PI)) and SQ subtypes (Primitive, Classical, Secretory, and Basal) were defined using gene expression based centroid predictors. Association between AD and SQ expression subtypes and 3 published BRCAness/PARP inhibitor response signatures developed in ovarian and/or breast cancer (Konstantinopoulos et al., PMID 20547991; Daemen et al., PMID 22875744; McGrail et al., PMID 28649435) was examined using linear regression. Association between subtypes and expression of 15 recognized homologous recombination (HR) related genes (ATM, ATR, BRCA1, BRCA2, BRIP1, CDK12, CHEK1, CHEK2, FANCA, FANCI, FANCD2, MRE11A, RAD51C, RAD51L1, PTEN) was also examined using linear regression, and association tests included adjustment for the 3 BRCAness/PARP inhibitor response signatures and proliferation score. Results: AD and SQ subtypes showed strong association with the 3 BRCAness/PARP inhibitor response signatures (F-test p-values 7.7e-05, 5.9e-13, 9.4e-33 in AD and 1.9e-05, 9.0e-13, 2.7e-19 in SQ). AD and SQ subtypes showed strong association with 15 HR genes (max and median F-test p-values were 8.5e-04 and 7.5e-25 in AD, and 7.3e-04 and 1.4e-12 in SQ). The TRU subtype in AD showed low expression relative to the other AD subtypes for a majority of the HR genes, including BRCA1. In SQ, the same was true for the basal and secretory subtypes. Simultaneous adjustment for the 3 BRCAness/PARP inhibitor response signatures, as well as for proliferation, reduced association strength between subtype and HR gene expression in AD and less so in SQ. In AD, association between subtype and gene expression remained significant for 4 HR genes (using Bonferroni correction for 15 tests), including CHECK2, FANCI, BRIP1, and RAD51L1. In SQ, association between subtype and gene expression remained significant for all HR genes except CHEK1 and FANCA, (median and min Bonferroni-adjusted p-value 2.9e-04 and 2.6e-21). Conclusions: Intrinsic biologic subtypes of lung AD and SQ are associated with published BRCAness/PARP inhibitor response signatures and reveal differential expression of several HR-related genes. Evaluation of these subtypes, particularly in SQ, as potential biomarkers of PARP inhibitor sensitivity should be investigated. Citation Format: Gregory Mayhew, Chuck Perou, D Neil Hayes, Myla Lai-Goldman, Hawazin Faruki. Differences in BRCAness/PARP inhibitor response signatures and homologous recombination gene expression across lung adenocarcinoma and squamous cell carcinoma gene expression subtypes [abstract]. In: Proceedings of the AACR-NCI-EORTC International Conference: Molecular Targets and Cancer Therapeutics; 2017 Oct 26-30; Philadelphia, PA. Philadelphia (PA): AACR; Mol Cancer Ther 2018;17(1 Suppl):Abstract nr A037.

CT-Based Deep Learning Model for Predicting Local Recurrence-Free Survival in Esophageal Squamous Cell Carcinoma Patients Received Concurrent Chemo-Radiotherapy: A Multicenter Study

Objective: Solid and micropapillary pattern are highly invasive histologic subtypes in lung adenocarcinoma and are associated with poor prognosis while the biopsy sample is not enough for the accurate histological diagnosis. This study aims to assess the correlation and predictive efficacy between metabolic parameters in (18)F-fluorodeoxy glucose positron emission tomography/computed tomography ((18)F-FDG PET-CT), including the maximum SUV (SUV(max)), metabolic tumor volume (MTV), total lesion glycolysis (TLG) and solid and micropapillary histological subtypes in lung adenocarcinoma. Methods: A total of 145 resected lung adenocarcinomas were included. The clinical data and preoperative (18)F-FDG PET-CT data were retrospectively analyzed. Mann-Whitney U test was used for the comparison of the metabolic parameters between solid and micropapillary subtype group and other subtypes group. Receiver operating characteristic (ROC) curve and areas under curve (AUC) were used for evaluating the prediction efficacy of metabolic parameters for solid or micropapillary patterns. Univariate and multivariate analyses were conducted to determine the prediction factors of the presence of solid or micropapillary subtypes. Results: Median SUV(max) and TLG in solid and papillary predominant subtypes group (15.07 and 34.98, respectively) were significantly higher than those in other subtypes predominant group (6.03 and 10.16, respectively, P<0.05). ROC curve revealed that SUV(max) and TLG had good efficacy for prediction of solid and micropapillary predominant subtypes [AUC=0.811(95% CI: 0.715~0.907) and 0.725(95% CI: 0.610~0.840), P<0.05]. Median SUV(max) and TLG in lung adenocarcinoma with the solid or micropapillary patterns (11.58 and 22.81, respectively) were significantly higher than those in tumors without solid and micropapillary patterns (4.27 and 6.33, respectively, P<0.05). ROC curve revealed that SUV(max) and TLG had good efficacy for predicting the presence of solid or micropapillary patterns [AUC=0.757(95% CI: 0.679~0.834) and 0.681(95% CI: 0.595~0.768), P<0.005]. Multivariate logistic analysis showed that the clinical stage (Stage Ⅲ-Ⅳ), SUV(max) ≥10.27 and TLG≥7.12 were the independent predictive factors of the presence of solid or micropapillary patterns (P<0.05). Conclusions: Preoperative SUV(max) and TLG of lung adenocarcinoma have good prediction efficacy for the presence of solid or micropapillary patterns, especially for the solid and micropapillary predominant subtypes and are independent factors of the presence of solid or micropapillary patterns.