A six-gene-based signature for breast cancer radiotherapy sensitivity estimation

Machine Learning & Molecular Radiation Tumor Biomarkers

Developing radiation tumor biomarkers that can guide personalized radiotherapy clinical decision making is a critical goal in the effort towards precision cancer medicine. High-throughput molecular assays paired with modern computational techniques have the potential to identify individual tumor-specific signatures and create tools that can help understand heterogenous patient outcomes in response to radiotherapy, allowing clinicians to fully benefit from the technological advances in molecular profiling and computational biology including machine learning. However, the increasingly complex nature of the data generated from high-throughput and "omics" assays require careful selection of analytical strategies. Furthermore, the power of modern machine learning techniques to detect subtle data patterns comes with special considerations to ensure that the results are generalizable. Herein, we review the computational framework of tumor biomarker development and describe commonly used machine learning approaches and how they are applied for radiation biomarker development using molecular data, as well as challenges and emerging research trends.

Identification of an individualized therapy prognostic signature for head and neck squamous cell carcinoma

BACKGROUND

Head and neck squamous cell carcinoma (HNSCC) are the most common cancers in the head and neck. Therapeutic response-related genes (TRRGs) are closely associated with carcinogenesis and prognosis in HNSCC. However, the clinical value and prognostic significance of TRRGs are still unclear. We aimed to construct a prognostic risk model to predict therapy response and prognosis in TRRGs-defined subgroups of HNSCC.

METHODS

The multiomics data and clinical information of HNSCC patients were downloaded from The Cancer Genome Atlas (TCGA). The profile data GSE65858 and GSE67614 chip was downloaded from public functional genomics data Gene Expression Omnibus (GEO). Based on TCGA-HNSC database, patients were divided into a remission group and a non-remission group according to therapy response, and differentially expressed TRRGs between those two groups were screened. Using Cox regression analysis and Least absolute shrinkage and selection operator (LASSO) analysis, candidate TRRGs that can predict the prognosis of HNSCC were identified and used to construct a TRRGs-based signature and a prognostic nomogram.

RESULT

A total of 1896 differentially expressed TRRGs were screened, including 1530 upregulated genes and 366 downregulated genes. Then, 206 differently expressed TRRGs that was significantly associated with the survival were chosen using univariate Cox regression analysis. Finally, a total of 20 candidate TRRGs genes were identified by LASSO analysis to establish a signature for risk prediction, and the risk score of each patient was calculated. Patients were divided into a high-risk group (Risk-H) and a low-risk group (Risk-L) based on the risk score. Results showed that the Risk-L patients had better overall survival (OS) than Risk-H patients. Receiver operating characteristic (ROC) curve analysis revealed great predictive performance for 1-, 3-, and 5-year OS in TCGA-HNSC and GEO databases. Moreover, for patients treated with post-operative radiotherapy, Risk-L patients had longer OS and lower recurrence than Risk-H patients. The nomogram involves risk score and other clinical factors had good performance in predicting survival probability.

CONCLUSIONS

The proposed risk prognostic signature and Nomogram based on TRRGs are novel promising tools for predicting therapy response and overall survival in HNSCC patients.

Resemblance of the Recurrence Patterns in Primary Systemic, Primary Surgery and Secondary Oncoplastic Surgery

PURPOSE

Surgical interventions tend to have an effect on the generation of recurrences in tumor patients due to the anesthesia involved as well as tissue damage and subsequent inflammation. This can also be found in patients with breast cancer.

METHODS

In this multicenter study, we investigated data of 632 patients with breast cancer and the subsequent diagnosis of a recurrence. The patient data were acquired from 1 January 2006 to 31 December 2019 in eight different centers in Germany. The data sets were separated into those with primary surgery, primary systemic therapy with subsequent surgery, and reconstructive surgery. Three different starting points for observation were defined: the date of diagnosis, the date of first surgery, and the date of reconstructive surgery, if applicable. The observational period was divided into steps of six months and maxima of recurrences were compared. Furthermore, the variance was calculated using the difference of the distribution in percent.

RESULTS

The descriptive analysis showed no resemblance between the groups. The variance of the difference of the recurrence rates analysis using the surgical date as the starting point showed similarities in the age subgroup.

CONCLUSION

Our clinical analysis shows different metastatic behavior in different analysis and treatment regimes. These findings justify further investigations on a larger database. These results may possibly identify an improved follow-up setting depending on tumor stage, biology, treatment, and patient factors (i.e., age, …).

Development and Validation of a Radiosensitivity Prediction Model for Lower Grade Glioma Based on Spike-and-Slab Lasso

Background and Purpose

Lower grade glioma (LGG) is one of the leading causes of death world worldwide. We attempted to develop and validate a radiosensitivity model for predicting the survival of lower grade glioma by using spike-and-slab lasso Cox model.

Methods

In this research, differentially expressed genes based on tumor microenvironment was obtained to further analysis. Log-rank test was used to identify genes in patients who received radiotherapy and patients who did not receive radiotherapy, respectively. Then, spike-and-slab lasso was performed to select genes in patients who received radiotherapy. Finally, three genes (INA, LEPREL1 and PTCRA) were included in the model. A radiosensitivity-related risk score model was established based on overall rate of TCGA dataset in patients who received radiotherapy. The model was validated in TCGA dataset that PFS as endpoint and two CGGA datasets that OS as endpoint. A novel nomogram integrated risk score with age and tumor grade was developed to predict the OS of LGG patients.

Results

We developed and verified a radiosensitivity-related risk score model. The radiosensitivity-related risk score is served as an independent prognostic indicator. This radiosensitivity-related risk score model has prognostic prediction ability. Moreover, the nomogram integrated risk score with age and tumor grade was established to perform better for predicting 1, 3, 5-year survival rate.

Conclusions

This model can be used by clinicians and researchers to predict patient’s survival rates and achieve personalized treatment of LGG.

More evidence for prediction model of radiosensitivity

With the development of precision medicine, searching for potential biomarkers plays a major role in personalized medicine. Therefore, how to predict radiosensitivity to improve radiotherapy is a burning question. The definition of radiosensitivity is complex. Radiosensitive gene/biomarker can be useful for predicting which patients would benefit from radiotherapy. The discovery of radiosensitivity biomarkers require multiple pieces of evidence. A prediction model of breast cancer radiosensitivity based on six genes was established. We had put forward some supplements on the basis of the present study. We found that there were no differences between high- and low-risk scores in the non-radiotherapy group. Patients who received radiotherapy had a significantly better overall survival than non-radiotherapy patients in the predicted low-risk score patients. Furthermore, there was no difference between radiotherapy group and non-radiotherapy group in the high-risk score group. Those results firmly supported the prediction model of radiosensitivity. In addition, building a radiosensitivity prediction model was systematically discussed. Genes of model could be screened by different methods, such as Cox regression analysis, Lasso Cox regression method, random forest algorithm and other methods. In the future, precision radiotherapy might depend on the combination of multi-omics data and high dimensional image data.

Radiation Response in the Tumour Microenvironment: Predictive Biomarkers and Future Perspectives

Radiotherapy (RT) is a primary treatment modality for a number of cancers, offering potentially curative outcomes. Despite its success, tumour cells can become resistant to RT, leading to disease recurrence. Components of the tumour microenvironment (TME) likely play an integral role in managing RT success or failure including infiltrating immune cells, the tumour vasculature and stroma. Furthermore, genomic profiling of the TME could identify predictive biomarkers or gene signatures indicative of RT response. In this review, we will discuss proposed mechanisms of radioresistance within the TME, biomarkers that may predict RT outcomes, and future perspectives on radiation treatment in the era of personalised medicine.