Deep learning-based analysis of EGFR mutation prevalence in lung adenocarcinoma H&E whole slide images.

EGFR mutations are a major prognostic factor in lung adenocarcinoma. However, current detection methods require sufficient samples and are costly. Deep learning is promising for mutation prediction in histopathological image analysis but has limitations in that it does not sufficiently reflect tumor heterogeneity and lacks interpretability. In this study, we developed a deep learning model to predict the presence of EGFR mutations by analyzing histopathological patterns in whole slide images (WSIs). We also introduced the EGFR mutation prevalence (EMP) score, which quantifies EGFR prevalence in WSIs based on patch-level predictions, and evaluated its interpretability and utility. Our model estimates the probability of EGFR prevalence in each patch by partitioning the WSI based on multiple-instance learning and predicts the presence of EGFR mutations at the slide level. We utilized a patch-masking scheduler training strategy to enable the model to learn various histopathological patterns of EGFR. This study included 868 WSI samples from lung adenocarcinoma patients collected from three medical institutions: Hallym University Medical Center, Inha University Hospital, and Chungnam National University Hospital. For the test dataset, 197 WSIs were collected from Ajou University Medical Center to evaluate the presence of EGFR mutations. Our model demonstrated prediction performance with an area under the receiver operating characteristic curve of 0.7680 (0.7607-0.7720) and an area under the precision-recall curve of 0.8391 (0.8326-0.8430). The EMP score showed Spearman correlation coefficients of 0.4705 (p = 0.0087) for p.L858R and 0.5918 (p = 0.0037) for exon 19 deletions in 64 samples subjected to next-generation sequencing analysis. Additionally, high EMP scores were associated with papillary and acinar patterns (p = 0.0038 and p = 0.0255, respectively), whereas low EMP scores were associated with solid patterns (p = 0.0001). These results validate the reliability of our model and suggest that it can provide crucial information for rapid screening and treatment plans.

Target-Enhanced Whole-Genome Sequencing (TE-WGS) Shows Clinical Validity Equivalent to Commercially Available Targeted Oncology Panel.

Purpose: Cancer poses a significant global health challenge, demanding precise genomic testing for individualized treatment strategies. Targeted-panel sequencing (TPS) has improved personalized oncology but often lacks comprehensive coverage of crucial cancer alterations. Whole-genome sequencing (WGS) addresses this gap, offering extensive genomic testing. This study demonstrates the medical potential of WGS. Materials and Methods: This study evaluates target-enhanced WGS (TE-WGS), a clinical-grade WGS method sequencing both cancer and matched normal tissues. Forty-nine patients with various solid cancer types underwent both TE-WGS and TruSight Oncology 500 (TSO500), one of the mainstream TPS approaches. Results: TE-WGS detected all variants reported by TSO500 (100%, 498/498). A high correlation in variant allele fractions (VAF) was observed between TE-WGS and TSO500 (r=0.978). Notably, 223 variants (44.8%) within the common set were discerned exclusively by TE-WGS in peripheral blood, suggesting their germline origin. Conversely, the remaining subset of 275 variants (55.2%) were not detected in peripheral blood using the TE-WGS, signifying them as bona fide somatic variants. Further, TE-WGS provided accurate copy number profiles, fusion genes, microsatellite instability (MSI), and homologous-recombination deficiency (HRD) scores, which were essential for clinical decision-making. Conclusion: TE-WGS is a comprehensive approach in personalized oncology, matching TSO500's key biomarker detection capabilities. It uniquely identifies germline variants and genomic instability markers, offering additional clinical actions. Its adaptability and cost-effectiveness underscore its clinical utility, making TE-WGS a valuable tool in personalized cancer treatment.

Clinical application of whole-genome sequencing of solid tumors for precision oncology.

Genomic alterations in tumors play a pivotal role in determining their clinical trajectory and responsiveness to treatment. Targeted panel sequencing (TPS) has served as a key clinical tool over the past decade, but advancements in sequencing costs and bioinformatics have now made whole-genome sequencing (WGS) a feasible single-assay approach for almost all cancer genomes in clinical settings. This paper reports on the findings of a prospective, single-center study exploring the real-world clinical utility of WGS (tumor and matched normal tissues) and has two primary objectives: (1) assessing actionability for therapeutic options and (2) providing clarity for clinical questions. Of the 120 patients with various solid cancers who were enrolled, 95 (79%) successfully received genomic reports within a median of 11 working days from sampling to reporting. Analysis of these 95 WGS reports revealed that 72% (68/95) yielded clinically relevant insights, with 69% (55/79) pertaining to therapeutic actionability and 81% (13/16) pertaining to clinical clarity. These benefits include the selection of informed therapeutics and/or active clinical trials based on the identification of driver mutations, tumor mutational burden (TMB) and mutational signatures, pathogenic germline variants that warrant genetic counseling, and information helpful for inferring cancer origin. Our findings highlight the potential of WGS as a comprehensive tool in precision oncology and suggests that it should be integrated into routine clinical practice to provide a complete image of the genomic landscape to enable tailored cancer management.

Longitudinal Magnetic Resonance Imaging with ROS-Responsive Bilirubin Nanoparticles Enables Monitoring of Nonalcoholic Steatohepatitis Progression to Cirrhosis.

Despite the vital importance of monitoring the progression of nonalcoholic fatty liver disease (NAFLD) and its progressive form, nonalcoholic steatohepatitis (NASH), an efficient imaging modality that is readily available at hospitals is currently lacking. Here, a new magnetic-resonance-imaging (MRI)-based imaging modality is presented that allows for efficient and longitudinal monitoring of NAFLD and NASH progression. The imaging modality uses manganese-ion (Mn2+)-chelated bilirubin nanoparticles (Mn@BRNPs) as a reactive-oxygen-species (ROS)-responsive MRI imaging probe. Longitudinal T1-weighted MR imaging of NASH model mice is performed after injecting Mn@BRNPs intravenously. The MR signal enhancement in the liver relative to muscle gradually increases up to 8 weeks of NASH progression, but decreases significantly as NASH progresses to the cirrhosis-like stage at weeks 10 and 12. A new dual input pseudo-three-compartment model is developed to provide information on NASH stage with a single MRI scan. It is also demonstrated that the ROS-responsive Mn@BRNPs can be used to monitor the efficacy of potential anti-NASH drugs with conventional MRI. The findings suggest that the ROS-responsive Mn@BRNPs have the potential to serve as an efficient MRI contrast for monitoring NASH progression and its transition to the cirrhosis-like stage.

Deciphering head and neck cancer microenvironment: Single-cell and spatial transcriptomics reveals human papillomavirus-associated differences.

Human papillomavirus (HPV) is a major causative factor of head and neck squamous cell carcinoma (HNSCC), and the incidence of HPV- associated HNSCC is increasing. The role of tumor microenvironment in viral infection and metastasis needs to be explored further. We studied the molecular characteristics of primary tumors (PTs) and lymph node metastatic tumors (LNMTs) by stratifying them based on their HPV status. Eight samples for single-cell RNA profiling and six samples for spatial transcriptomics (ST), composed of matched primary tumors (PT) and lymph node metastases (LNMT), were collected from both HPV- negative (HPV- ) and HPV-positive (HPV+ ) patients. Using the 10x Genomics Visium platform, integrative analyses with single-cell RNA sequencing were performed. Intracellular and intercellular alterations were analyzed, and the findings were confirmed using experimental validation and publicly available data set. The HPV+ tissues were composed of a substantial amount of lymphoid cells regardless of the presence or absence of metastasis, whereas the HPV- tissue exhibited remarkable changes in the number of macrophages and plasma cells, particularly in the LNMT. From both single-cell RNA and ST data set, we discovered a central gene, pyruvate kinase muscle isoform 1/2 (PKM2), which is closely associated with the stemness of cancer stem cell-like populations in LNMT of HPV- tissue. The consistent expression was observed in HPV- HNSCC cell line and the knockdown of PKM2 weakened spheroid formation ability. Furthermore, we found an ectopic lymphoid structure morphology and clinical effects of the structure in ST slide of the HPV+ patients and verified their presence in tumor tissue using immunohistochemistry. Finally, the ephrin-A (EPHA2) pathway was detected as important signals in angiogenesis for HPV- patients from single-cell RNA and ST profiles, and knockdown of EPHA2 declined the cell migration. Our study described the distinct cellular composition and molecular alterations in primary and metastatic sites in HNSCC patients based on their HPV status. These results provide insights into HNSCC biology in the context of HPV infection and its potential clinical implications.

Deep learning-radiomics integrated noninvasive detection of epidermal growth factor receptor mutations in non-small cell lung cancer patients.

This study focused on a novel strategy that combines deep learning and radiomics to predict epidermal growth factor receptor (EGFR) mutations in patients with non-small cell lung cancer (NSCLC) using computed tomography (CT). A total of 1280 patients with NSCLC who underwent contrast-enhanced CT scans and EGFR mutation testing before treatment were selected for the final study. Regions of interest were segmented from the CT images to extract radiomics features and obtain tumor images. These tumor images were input into a convolutional neural network model to extract 512 image features, which were combined with radiographic features and clinical data to predict the EGFR mutation. The generalization performance of the model was evaluated using external institutional data. The internal and external datasets contained 324 and 130 EGFR mutants, respectively. Sex, height, weight, smoking history, and clinical stage were significantly different between the EGFR-mutant patient groups. The EGFR mutations were predicted by combining the radiomics and clinical features, and an external validation dataset yielded an area under the curve (AUC) value of 0.7038. The model utilized 1280 tumor images, radiomics features, and clinical characteristics as input data and exhibited an AUC of approximately 0.81 and 0.78 during the primary cohort and external validation, respectively. These results indicate the feasibility of integrating radiomics analysis with deep learning for predicting EGFR mutations. CT-image-based genetic testing is a simple EGFR mutation prediction method, which can improve the prognosis of NSCLC patients and help establish personalized treatment strategies.

Predictive modelling for the diagnosis of oral and laryngeal premalignant and malignant lesions using p53 and Ki-67 expression.

Oral and laryngeal epithelial lesions are currently diagnosed using histological criteria based on the World Health Organization (WHO) classification, which can cause interobserver variability. An integrated diagnostic approach based on immunohistochemistry (IHC) would aid in the interpretation of ambiguous histological findings of epithelial lesions. In the present study, IHC was used to evaluate the expression of p53 and Ki-67 in 114 cases of oral and laryngeal epithelial lesions in 104 patients. Logistic regression analysis and decision tree algorithm were employed to develop a scoring system and predictive model for differentiating the epithelial lesions. Cohen's kappa coefficient was used to evaluate interobserver variability, and next-generation sequencing (NGS) and IHC were used to compare TP53 mutation and p53 expression patterns. Two expression patterns for p53, namely, diffuse expression type (pattern HI) and null type (pattern LS), and the pattern HI for Ki-67 were significantly associated with high-grade dysplasia (HGD) or squamous cell carcinoma (SqCC). With an accuracy and area under the receiver operating characteristic curve (AUC) of 84.6% and 0.85, respectively, the scoring system based on p53 and Ki-67 expression patterns classified epithelial lesions into two types: non-dysplasia (ND) or low-grade dysplasia (LGD) and SqCC or HGD. The decision tree model constructed using the p53 and Ki-67 expression patterns classified epithelial lesions into ND, LGD, and group 2, including HGD or SqCC, with an accuracy and AUC of 75% and 0.87, respectively. The integrated diagnosis had a better correlation with near perfect agreement (weighted kappa 0.92, unweighted kappa 0.88). The patterns HI and LS for p53 were confirmed to be correlated with missense mutations and nonsense/frameshift mutations, respectively. A predictive model for diagnosis was developed based on the correlation between TP53 mutation and p53 expression patterns. These results indicate that the scoring system based on p53 and Ki-67 expression patterns can differentiate epithelial lesions, especially in cases when the morphological features are ambiguous.

Assessment of Radiologic Extranodal Extension Using Combinatorial Analysis of Nodal Margin Breakdown and Metastatic Burden in Oropharyngeal Cancer.

The importance of risk stratification in the management of oropharyngeal squamous cell carcinoma (OPSCC) is becoming increasingly obvious with the growing evidence of its variable prognosis. We identified and evaluated imaging characteristics predictive of extranodal extension (ENE) in OPSCC. Preoperative computed tomography and histopathologic results of 108 OPSCC patients who underwent neck dissection as primary treatment were analyzed. Imaging characteristics were reassessed for factors associated with nodal margin breakdown and metastatic burden. Moreover, the predictability of pathological ENE (pENE) was analyzed. Univariate and multivariate binomial logistic regression analyses were performed to examine the predictive power of ENE-related radiologic features. Imaging-based characteristics showed variable degrees of association with pENE. Factors associated with nodal margin breakdown (indistinct capsular contour, irregular margin, and perinodal fat stranding) and factors associated with nodal burden (nodal matting, lower neck metastasis, and presence of >4 lymph node metastases) were significantly predictive of ENE (odds ratio (OR) = 11.170 and 12.121, respectively). The combined utilization of the nodal margin and burden factors further increased the predictive ability (OR = 14.710). Factors associated with nodal margin breakdown and nodal burden were associated with pENE, demonstrating the use of combinatorial analysis for more accurate ENE prediction.

The comprehensive expression of BCL2 family genes determines the prognosis of diffuse large B-cell lymphoma.

Diffuse large B-cell lymphoma (DLBCL) is a prevalent and aggressive non-Hodgkin's lymphoma, and 40% of patients succumb to death. Despite numerous clinical trials aimed at developing treatment strategies beyond the conventional R-CHOP (rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone) regimen, there have been no positive results thus far. Although the selective BCL2 inhibitor venetoclax has shown remarkable efficacy in chronic lymphocytic leukemia, its therapeutic effect in DLBCL was limited. We hypothesized that the limited therapeutic effect of venetoclax in DLBCL may be attributed to the complex expression and interactions of BCL2 family members, including BCL2. Therefore, we aimed to comprehensively analyze the expression patterns of BCL2 family members in DLBCL. We analyzed 157 patients with de novo DLBCL diagnosed at Asan Medical Center and Ajou University Hospital. The mRNA expression levels of BCL2 family members were quantified using the NanoString technology. BCL2 family members showed distinct heterogeneous expression patterns both intra- and inter-patient. Using unsupervised hierarchical cluster analysis, we were able to classify patients with similar BCL2 family expression pattern and select groups with clear prognostic features, C1 and C6. In the group with the best prognosis, C1, the expression of pro-apoptotic and pro-apoptotic BH3-only group gene expressions were increased, while anti-apoptotic group expression was significantly increased in both C1 and C6. Based on this, we generated the BCL2 signature score using the expression of pro-apoptotic genes BOK and BCL2L15, and anti-apoptotic gene BCL2. The BCL2 signature score 0 had the best prognosis, score 1/2 had intermediate prognosis, and score 3 had the worst prognosis (EFS, p = 0.0054; OS, p = 0.0011). Multivariate analysis, including COO and IPI, showed that increase in the BCL2 signature score was significantly associated with poor prognosis for EFS, independent of COO and IPI. The BCL2 signature score we proposed in this study provides information on BCL2 family deregulation based on the equilibrium of pro-versus anti-apoptotic BCL2 family, which can aid in the development of new treatment strategies for DLBCL in the future.

Contrast-enhanced CT-based Radiomics for the Differentiation of Anaplastic or Poorly Differentiated Thyroid Carcinoma from Differentiated Thyroid Carcinoma: A Pilot Study.

Differential diagnosis of anaplastic thyroid carcinoma/poorly differentiated thyroid carcinoma (ATC/PDTC) from differentiated thyroid carcinoma (DTC) is crucial in patients with large thyroid malignancies. This study creates a predictive model using radiomics feature analysis to differentiate ATC/PDTC from DTC. We compared the clinicoradiological characteristics and radiomics features extracted from a volume of interest on contrast-enhanced computed tomography (CT) between the groups. Estimations of variable importance were performed via modeling using the random forest quantile classifier. The diagnostic performance of the model with radiomics features alone had the area under the receiver operating characteristic (AUROC) curve value of 0.883. The sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy were 81.7%, 93.3%, 97.7%, 64.5%, and 84.6%, respectively, for the differential diagnosis of ATC/PDTC and DTC. The model with both radiomics and clinicoradiological information showed the AUROC of 0.908, with sensitivity, specificity, PPV, NPV, and accuracy of 82.9%, 97.6%, 99.2%, 67.1%, and 86.5% respectively. Distant metastasis, moment, shape, age, and gray-level size zone matrix features were the most useful factors for differential diagnosis. Therefore, we concluded that a radiomics approach based on contrast-enhanced CT features can potentially differentiate ATC/PDTC from DTC in patients with large thyroid malignancies.