Intratumoral budding: A novel prognostic biomarker for tumor recurrence and a potential predictor of nodal metastasis in uterine cervical cancer

The Mclust Analysis of Tumor Budding Unveils the Role of the Collagen Family in Cervical Cancer Progression

In RNA-seq data analysis, condensing the gene count matrix size is pivotal for downstream investigations, particularly pathway analysis. For this purpose, harnessing machine learning attracts increasing interest, while conventional methodologies depend on p-value comparisons. In this study, 20 tissue samples from real-world cervical cancers were subjected to sequencing, followed by the application of the Mclust algorithm to delineate an optimal cluster. By stratifying tumor budding into high and low groups and quantifying the epithelial-to-mesenchymal transition (EMT) score to scrutinize tumor budding, we discerned 24 EMT-related genes, with 5 showing strong associations with cervical cancer prognosis. Our observations elucidate a biological flow wherein EMT, Matrix Metallopep-tidase 2 (MMP2), and extracellular matrix (ECM) degradation are interconnected, ultimately leading to collagen type VI and exacerbating the prognosis of cervical cancer. The present study underscores an alternative method for selecting useful EMT-related genes by employing an appropriate clustering algorithm, thereby avoiding classical methods while unveiling novel insights into cervical cancer etiology and prognosis. Moreover, when comparing high and low tumor budding, collagen type VI emerges as a potential gene marker for the prognosis of cervical cancer.

Predicting Lymph Node Metastasis From Primary Cervical Squamous Cell Carcinoma Based on Deep Learning in Histopathologic Images

We developed a deep learning framework to accurately predict the lymph node status of patients with cervical cancer based on hematoxylin and eosin-stained pathological sections of the primary tumor. In total, 1524 hematoxylin and eosin-stained whole slide images (WSIs) of primary cervical tumors from 564 patients were used in this retrospective, proof-of-concept study. Primary tumor sections (1161 WSIs) were obtained from 405 patients who underwent radical cervical cancer surgery at the Fudan University Shanghai Cancer Center (FUSCC) between 2008 and 2014; 165 and 240 patients were negative and positive for lymph node metastasis, respectively (including 166 with positive pelvic lymph nodes alone and 74 with positive pelvic and para-aortic lymph nodes). We constructed and trained a multi-instance deep convolutional neural network based on a multiscale attention mechanism, in which an internal independent test set (100 patients, 228 WSIs) from the FUSCC cohort and an external independent test set (159 patients, 363 WSIs) from the Cervical Squamous Cell Carcinoma and Endocervical Adenocarcinoma cohort of the Cancer Genome Atlas program database were used to evaluate the predictive performance of the network. In predicting the occurrence of lymph node metastasis, our network achieved areas under the receiver operating characteristic curve of 0.87 in the cross-validation set, 0.84 in the internal independent test set of the FUSCC cohort, and 0.75 in the external test set of the Cervical Squamous Cell Carcinoma and Endocervical Adenocarcinoma cohort of the Cancer Genome Atlas program. For patients with positive pelvic lymph node metastases, we retrained the network to predict whether they also had para-aortic lymph node metastases. Our network achieved areas under the receiver operating characteristic curve of 0.91 in the cross-validation set and 0.88 in the test set of the FUSCC cohort. Deep learning analysis based on pathological images of primary foci is very likely to provide new ideas for preoperatively assessing cervical cancer lymph node status; its true value must be validated with cervical biopsy specimens and large multicenter datasets.

Immune Pathway and Gene Database (IMPAGT) Revealed the Immune Dysregulation Dynamics and Overactivation of the PI3K/Akt Pathway in Tumor Buddings of Cervical Cancer

Tumor budding (TB) is a small cluster of malignant cells at the invasive front of a tumor. Despite being an adverse prognosis marker, little research has been conducted on the tumor immune microenvironment of tumor buddings, especially in cervical cancer. Therefore, RNA sequencing was performed using 21 formalin-fixed, paraffin-embedded slides of cervical tissues, and differentially expressed genes (DEGs) were analyzed. Immune Pathway and Gene Database (IMPAGT) was generated for immune profiling. "Pathway in Cancer" was identified as the most enriched pathway for both up- and downregulated DEGs. Kyoto Encyclopedia of Genes and Genomes Mapper and Gene Ontology further revealed the activation of the PI3K/Akt signaling pathway. An IMPAGT analysis revealed immune dysregulation even at the tumor budding stage, especially in the PI3K/Akt/mTOR axis, with a high efficiency and integrity. These findings emphasized the clinical significance of tumor buddings and the necessity of blocking the overactivation of the PI3K/Akt/mTOR pathway to improve targeted therapy in cervical cancer.

Transcriptomic Immune Profiles Can Represent the Tumor Immune Microenvironment Related to the Tumor Budding Histology in Uterine Cervical Cancer

Tumor budding (TB) histology has become a critical biomarker for several solid cancers. Despite the accumulating evidence for the association of TB histology with poor prognosis, the biological characteristics of TB are little known about in the context related to the tumor immune microenvironment (TIME) in uterine cervical cancer (CC). Therefore, this study aimed to identify the transcriptomic immune profiles related to TB status and further provide robust medical evidence for clinical application. In our study, total RNA was extracted and sequenced from 21 CC tissue specimens. As such, 1494 differentially expressed genes (DEGs) between the high- and low-TB groups were identified by DESeq2. After intersecting the list of DEGs and public immune genes, we selected 106 immune-related DEGs. Then, hub genes were obtained using Least Absolute Shrinkage and Selection Operator regression. Finally, the correlation between the hub genes and immune cell types was analyzed and four candidate genes were identified (one upregulated (FCGR3B) and three downregulated (ROBO2, OPRL1, and NR4A2) genes). These gene expression levels were highly accurate in predicting TB status (area under the curve >80%). Interestingly, FCGR3B is a hub gene of several innate immune pathways; its expression significantly differed in the overall survival analysis (p = 0.0016). In conclusion, FCGR3B, ROBO2, OPRL1, and NR4A2 expression can strongly interfere with TB growth and replace TB to stratify CC patients.

Tumor budding of cervical squamous cell carcinoma: epithelial-mesenchymal transition-like cancer stem cells?

Recent evidence indicates that cancer stem cells (CSCs) are the origin of cancers. Scientists have identified CSCs in various tumors and have suggested the existence of a variety of states of CSCs. The existence of epithelial-mesenchymal transition (EMT)-like CSCs has been confirmed in vitro, but they have not been identified in vivo. Tumor budding was defined as single cell or clusters of ≤ 5 cells at the invasive front of cancers. Such tumor budding is hypothesized to be closely related to EMT and linked to CSCs, especially to those migrating at the invasive front. Therefore, tumor budding has been proposed to represent EMT-like stem cells. However, this hypothesis has not yet been proven. Thus, we studied the expression of EMT markers, certain CSC markers of tumor budding, and the tumor center of cervical squamous cell carcinoma (CxSCC). We performed tissue chip analyses of 95 primary CxSCCs from patients. Expression of EMT and CSC markers (E-cadherin, β-catenin, vimentin, Ki67, CD44, SOX2 , and ALDH1A1) in a set of tumor samples on tissue chips (87 cases of tumor budding/the main tumor body) were evaluated by immunohistochemistry. We found that the cell-membranous expression of β-catenin was stronger in the main tumor body than in tumor buds. Compared with the main tumor body, tumor buds had reduced proliferative activity as measured by Ki67. Moreover, vimentin expression was high and E-cadherin expression was low in tumor buds. Expression of EMT-related markers suggested that tumor buds were correlated with EMT. We noted that CxSCC tumor buds had a CD44negative/low/SOX2high/ALDH1A1high staining pattern, indicating that tumor buds of CxSCC present CSC-like immunophenotypic features. Taken together, our data indicate that tumor buds in CxSCC may represent EMT-like CSCs in vivo.

Tumor Budding Activity Is an Independent Prognostic Factor in Squamous Cell Carcinoma of the Vulva

Tumor budding activity (TBA) is recognized as a potential prognostic factor in carcinomas from several anatomic sites. This study evaluates the prognostic value of TBA in a cohort of squamous cell carcinoma of the vulva (VSCC). TBA, defined as clusters of <5 tumor cells that are detached from the main tumor and that infiltrate into surrounding stroma, was assessed in 82 cases of surgically excised VSCC and correlated with patient outcomes. All cases were classified into one of 3 groups: no TBA, low TBA (1 to 14 foci), and high TBA (≥15 foci). 23 (29.1%), 37(45.1%) and 22 (26.8%) cases showed no, low and high TBA respectively. High TBA was associated with reduced overall survival (OS) on multivariate analysis independent of FIGO stage, HPV status, and p53 status. The majority of tumors with high TBA displayed a p53 mutant staining pattern (77.3%, 17 of 22). The 17 patients whose tumors displayed a p53 mutant/high TBA profile had worse outcomes when compared with 15 patients whose tumors showed a p53 mutant/no TBA or p53 mutant/low TBA profile (mean OS 37.5 vs 63.3 months respectively, p=.002). High TBA was observed in only 5 of 47 HPV associated cases, and this subset also seemed to display a worse patient outcome as compared with the rest of the HPV associated cohort (OS 16.8 vs 142.8 months, p<.0001). In summary, these findings indicate that TBA is an independent prognostic indicator in VSCC patients, and that high TBA is associated with worse clinical outcomes.

Predicting Tumor Budding Status in Cervical Cancer Using MRI Radiomics: Linking Imaging Biomarkers to Histologic Characteristics

BACKGROUND

Our previous study demonstrated that tumor budding (TB) status was associated with inferior overall survival in cervical cancer. The purpose of this study is to evaluate whether radiomic features can predict TB status in cervical cancer patients.

METHODS

Seventy-four patients with cervical cancer who underwent preoperative MRI and radical hysterectomy from 2011 to 2015 at our institution were enrolled. The patients were randomly allocated to the training dataset (n = 48) and test dataset (n = 26). Tumors were segmented on axial gadolinium-enhanced T1- and T2-weighted images. A total of 2074 radiomic features were extracted. Four machine learning classifiers, including logistic regression (LR), random forest (RF), support vector machine (SVM), and neural network (NN), were used. The trained models were validated on the test dataset.

RESULTS

Twenty radiomic features were selected; all were features from filtered-images and 85% were texture-related features. The area under the curve values and accuracy of the models by LR, RF, SVM and NN were 0.742 and 0.769, 0.782 and 0.731, 0.849 and 0.885, and 0.891 and 0.731, respectively, in the test dataset.

CONCLUSION

MRI-based radiomic features could predict TB status in patients with cervical cancer.