Dataset Information

A Gene Prognostic Index Associated With Epithelial-Mesenchymal Transition Predicting Biochemical Recurrence and Tumor Chemoresistance for Prostate Cancer.

ABSTRACT:

Background

We aimed to establish a novel epithelial-mesenchymal transition (EMT)-related gene prognostic index (EMTGPI) associated with biochemical recurrence (BCR) and drug resistance for prostate cancer (PCa).

Methods

We used Lasso and Cox regression analysis to establish the EMTGPI. All analyses were conducted with R version 3.6.3 and its suitable packages.

Results

We established the EMTGPI based on SFRP4 and SPP1. Patients in high-risk group had 2.23 times of BCR risk than those in low-risk group (p = 0.003), as well as 2.36 times of metastasis risk (p = 0.053). In external validation, we detected similar diagnostic efficacy and prognostic value in terms of BCR free survival. For drug resistance, we observe moderately diagnostic accuracy of EMTGPI score (AUC: 0.804). We found that PDCD1LG2 (p = 0.04) and CD96 (p = 0.01) expressed higher in BCR patients compared with their counterpart. For TME analysis, we detected that CD8+ T cells and M1 macrophages expressed higher in BCR group. Moreover, stromal score (p = 0.003), immune score (p = 0.01), and estimate score (p = 0.003) were higher in BCR patients. We found that EMTGPI was significantly related to HAVCR2 (r: 0.34), CD96 (r: 0.26), CD47 (r: 0.22), KIR3DL1 (r: -0.21), KLRD1 (r: -0.21), and CD2 (r: 0.21). In addition, we observed that EMTGPI was significantly associated with M1 macrophages (r: 0.6), M2 macrophages (r: -0.33), monocytes (r: -0.18), neutrophils (r: -0.43), CD8+ T cells (r: 0.13), and dendritic cells (r: 0.37). PHA-793887 was the common drug sensitive to SPP1 and SFRP4, and PC3 and DU145 were the common PCa-related cell lines of SPP1, SFRP4, and PHA-793887.

Conclusions

We concluded that the EMTGPI score based on SFRP4 and SPP1 could be used to predict BCR for PCa patients. We confirmed the impact of immune evasion on the BCR process of PCa.

SUBMITTER: Feng D

PROVIDER: S-EPMC8790245 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:Accurate estimation of recurrence risk is needed for optimal treatment of clinically localized prostate cancer (PCa) patients. We combined classical clinicopathological predictors of biochemical recurrence (BCR) and molecular markers into a new risk score for predicting BCR after radical prostatectomy (RP). A retrospective study which included 122 PCa patients who attended our department between 2000 and 2007 was conducted. Total RNA was isolated from formalin-fixed paraffin embedded PCa tissue. Gene expression patterns were analyzed in 21 selected samples from 7 localized, 6 locally advanced, and 8 metastatic PCa patients using Illumina microarrays. Expression levels of 41 genes were validated by quantitative PCR in an independent retrospective series of 101 patients with ≥ 10 years follow-up who underwent RP for clinically localized PCa. Univariate and multivariate logistic regression analysis were used to identify individual predictors of BCR. A risk score (RS) for predicting BCR including clinicopathological and gene expression data was developed and Kaplan-Meier curves were generated. Interaction networks between the genes of the model were built by GeneMANIA Cytoscape plugin. In a median follow-up of 120 months (range 3-190), 37 patients developed BCR (36.6%). Expression levels of 7,930 transcripts differed between clinically localized and locally advanced-metastatic PCa groups (FDR<0.1). A set of 41 genes were validated by quantitative PCR. Regression analysis showed that expression of ASF1B and MCL1 as well as Gleason score, extracapsular extension, seminal vesicle invasion and positive margins were independent prognostic factors of BCR. A RS generated using these gene expression and clinicopathological variables was able to discriminate between two groups of patients with a significantly different probability of BCR (HR 6.24; CI 3.23-12.4, p<0.01), improving the individual discriminative performance of each of these variables on their own. Direct interactions between the two genes of the model were not found. In conclusion, identification of new gene expression patterns associated with a high probability of BCR in clinically localized PCa patients treated with RP, and their combination with clinicopathological variables in a robust, easy-to-use and reliable classifier may contribute to improve PCa risk stratification and, consequently, to tailor treatment and surveillance strategies in these patients.

Project description:Background The first sign of metastatic prostate cancer after radical prostatectomy is rising PSA levels in the blood, termed biochemical recurrence. The prediction of recurrence relies mainly on the morphological assessment of prostate cancer using the Gleason grading system. However, in this system, within-grade morphological patterns and subtle histopathological features are currently omitted, leaving a significant amount of prognostic potential unexplored. Methods To discover additional prognostic information using artificial intelligence, we trained a deep learning system to predict biochemical recurrence from tissue in H&E-stained microarray cores directly. We developed a morphological biomarker using convolutional neural networks leveraging a nested case-control study of 685 patients and validated on an independent cohort of 204 patients. We use concept-based explainability methods to interpret the learned tissue patterns. Results The biomarker provides a strong correlation with biochemical recurrence in two sets (n = 182 and n = 204) from separate institutions. Concept-based explanations provided tissue patterns interpretable by pathologists. Conclusions These results show that the model finds predictive power in the tissue beyond the morphological ISUP grading. Plain language summary To determine the prognosis of patients with prostate cancer, several clinical factors are taken into account. One of these is the cancer grade, assigned by a pathologist based on the cancer’s appearance under a microscope. The grade ranges from 1 to 5, where 5 is the most aggressive tumour type. This study explored whether deep learning—a technique in which computer software learns patterns from multiple examples—can learn to predict the risk of patients’ cancers recurring from microscopic images of the tumours. We show, on two clinical datasets from different institutions, that such a system can help to better predict prognosis, beyond the information provided by grade alone. In the future, this type of method could help clinicians to predict the prognosis of individual prostate cancer patients. Pinckaers et al. develop a deep learning system to predict biochemical recurrence in prostate cancer patients treated with radical prostatectomy. The authors’ morphological biomarker provides predictive power beyond traditional Gleason grading, based on analysis of two clinical datasets from different institutions.