Genetics of rectal cancer and novel therapies: primer for radiologists

Deep learning-based multi-parametric magnetic resonance imaging (mp-MRI) nomogram for predicting Ki-67 expression in rectal cancer

PURPOSE

To explore the value of deep learning-based multi-parametric magnetic resonance imaging (mp-MRI) nomogram in predicting the Ki-67 expression in rectal cancer.

METHODS

The data of 491 patients with rectal cancer from two centers were retrospectively analyzed and divided into training, internal validation, and external validation sets. They were categorized into high- and low-expression group based on postoperative pathological Ki-67 expression. Each patient's mp-MRI data were analyzed to extract and select the most relevant features of deep learning, and a deep learning model was constructed. Independent predictive risk factors were identified and incorporated into a clinical model, and the clinical and deep learning models were combined to obtain a nomogram for the prediction of Ki-67 expression. The performance characteristics of the DL-model, clinical model, and nomogram were assessed using ROCs, calibration curve, decision curve, and clinical impact curve analysis.

RESULTS

The strongest deep learning features were extracted and screened from mp-MRI data. Two independent predictive factors, namely Magnetic Resonance Imaging T (mrT) staging and differentiation degree, were identified through clinical feature selection. Three models were constructed: a deep learning (DL)-model, a clinical model, and a nomogram. The AUCs of clinical model in the training, internal validation, and external validation set were 0.69, 0.78, and 0.67, respectively. The AUCs of the deep model and nomogram ranged from 0.88 to 0.98. The prediction performance of the deep learning model and nomogram was significantly better than the clinical model (P < 0.001).

CONCLUSION

The nomogram based on deep learning can help clinicians accurately and conveniently predict the expression status of Ki-67 in rectal cancer.

Tumor location matters, next generation sequencing mutation profiling of left-sided, rectal, and right-sided colorectal tumors in 552 patients

Despite the introduction of new molecular classifications, advanced colorectal cancer (CRC) is treated with chemotherapy supplemented with anti-EGFR and anti-VEGF targeted therapy. In this study, 552 CRC cases with different primary tumor locations (250 left side, 190 rectum, and 112 right side) were retrospectively analyzed by next generation sequencing for mutations in 50 genes. The most frequently mutated genes were TP53 in left-sided tumors compared to right-sided tumors and BRAF in right-sided tumors compared to left-sided tumors. Mutations in KRAS, NRAS, and BRAF were not detected in 45% of patients with left-sided tumors and in 28.6% of patients with right-sided tumors. Liver metastases were more common in patients with left-sided tumors. Tumors on the right side were larger at diagnosis and had a higher grade (G3) than tumors on the left. Rectal tumors exhibit distinctive biological characteristics when compared to left-sided tumors, including a higher absence rate of KRAS, NRAS, and BRAF mutations (47.4% in rectal versus 42.8% in left-sided tumors). These rectal tumors are also unique in their primary metastasis site, which is predominantly the lungs, and they have varying mutation rates, particularly in genes such as BRAF, FBXW7, and TP53, that distinguish them from tumors found in other locations. Primary tumor location has implications for the potential treatment of CRC with anti-EGFR therapy.

Multiplexed DNA Methylation Analysis in Colorectal Cancer Using Liquid Biopsy and Its Diagnostic and Predictive Value

Early diagnosis of colorectal cancer (CRC) is of high importance as prognosis depends on tumour stage at the time of diagnosis. Detection of tumour-specific DNA methylation marks in cfDNA has several advantages over other approaches and has great potential for solving diagnostic needs. We report here the identification of DNA methylation biomarkers for CRC and give insights in our methylation-sensitive restriction enzyme coupled qPCR (MSRE-qPCR) system. Targeted microarrays were used to investigate the DNA methylation status of 360 cancer-associated genes. Validation was done by qPCR-based approaches. A focus was on investigating marker performance in cfDNA from 88 patients (44 CRC, 44 controls). Finally, the workflow was scaled-up to perform 180plex analysis on 110 cfDNA samples, to identify a DNA methylation signature for advanced colonic adenomas (AA). A DNA methylation signature (n = 44) was deduced from microarray experiments and confirmed by quantitative methylation-specific PCR (qMSP) and by MSRE-qPCR, providing for six genes’ single areas under the curve (AUC) values of >0.85 (WT1, PENK, SPARC, GDNF, TMEFF2, DCC). A subset of the signatures can be used for patient stratification and therapy monitoring for progressed CRC with liver metastasis using cfDNA. Furthermore, we identified a 35-plex classifier for the identification of AAs with an AUC of 0.80.

Immune-Related lncRNAs to Construct Novel Signatures and Predict the Prognosis of Rectal Cancer

Colorectal cancer (CRC) is one of the most common cancers. Almost 1/3 of CRC are rectal cancer, and 95% of rectal cancers are rectal adenocarcinoma (READ). Increasing evidences indicated that long noncoding RNAs (lncRNAs) have important role in the genesis and development of cancers. The purpose of our present study was to identify the differential expression lncRNAs which potentially related with immune cells infiltration and establish a risk assessment model to predict the clinical outcome for READ patients. We obtained three immune-related differential expression lncRNAs (IR-DELs) (C17orf77, GATA2-AS1, and TPT1-AS1) by differential expression analysis following correlation analysis and Cox regression analysis. A risk assessment model were constructed by integrating these analysis results. We then plotted the 1-, 3-, and 5-year ROC curves depending on our risk assessment model, which suggested that all AUC values were over 0.7. In addition, we found that the risk assessment model was correlated with several immune cells and factors. This study suggested that those three signatures (C17orf77, GATA2-AS1, and TPT1-AS1) screened by pairing IR-DELs could be prognosis markers for READ patients and might benefit them from antitumor immunotherapy.

Development and Validation of a Nomogram and a Comprehensive Prognostic Analysis of an LncRNA-Associated Competitive Endogenous RNA Network Based on Immune-Related Genes for Locally Advanced Rectal Cancer With Neoadjuvant Therapy

Colorectal cancer (CRC) is a common digestive tract tumor worldwide. In recent years, neoadjuvant chemoradiotherapy (CRT) has been the most comprehensive treatment for locally advanced rectal cancer (LARC). In this study, we explored immune infiltration in rectal cancer (RC) and identified immune-related differentially expressed genes (IRDEGs). Then, we identified response markers in datasets in GEO databases by principal component analysis (PCA). We also utilized three GEO datasets to identify the up- and downregulated response-related genes simultaneously and then identified genes shared between the PCA markers and three GEO datasets. Based on the hub IRDEGs, we identified target mRNAs and constructed a ceRNA network. Based on the ceRNA network, we explored prognostic biomarkers to develop a prognostic model for RC through Cox regression. We utilized the specimen to validate the expression of the two biomarkers. We also utilized LASSO regression to screen hub IRDEGs and built a nomogram to predict the response of LARC patients to CRT. All of the results show that the nomogram and prognostic model offer good prognostic value and that the ceRNA network can effectively highlight the regulatory relationship. hsa-mir-107 and WDFY3-AS2 may be prognostic biomarkers for RC.

Emerging Trends for Radio-Immunotherapy in Rectal Cancer

Rectal cancer is a heterogeneous disease at the genetic and molecular levels, both aspects having major repercussions on the tumor immune contexture. Whilst microsatellite status and tumor mutational load have been associated with response to immunotherapy, presence of tumor-infiltrating lymphocytes is one of the most powerful prognostic and predictive biomarkers. Yet, the majority of rectal cancers are characterized by microsatellite stability, low tumor mutational burden and poor T cell infiltration. Consequently, these tumors do not respond to immunotherapy and treatment largely relies on radiotherapy alone or in combination with chemotherapy followed by radical surgery. Importantly, pre-clinical and clinical studies suggest that radiotherapy can induce a complete reprograming of the tumor microenvironment, potentially sensitizing it for immune checkpoint inhibition. Nonetheless, growing evidence suggest that this synergistic effect strongly depends on radiotherapy dosing, fractionation and timing. Despite ongoing work, information about the radiotherapy regimen required to yield optimal clinical outcome when combined to checkpoint blockade remains largely unavailable. In this review, we describe the molecular and immune heterogeneity of rectal cancer and outline its prognostic value. In addition, we discuss the effect of radiotherapy on the tumor microenvironment, focusing on the mechanisms and benefits of its combination with immune checkpoint inhibitors.

Delivery of Personalized Care for Locally Advanced Rectal Cancer: Incorporating Pathological, Molecular Genetic, and Immunological Biomarkers Into the Multimodal Paradigm

Approximately one-third of all newly diagnosed colorectal cancer (CRC) is composed of rectal cancer, with the incidence rising in younger patients. The principal neoadjuvant treatments consist of neoadjuvant short-course radiotherapy and long-course chemoradiation. Locally advanced rectal cancer (LARC) is particularly challenging to manage given the anatomical constrictions of the pelvis and the risk for local recurrence. In appropriately treated patients, 5- and 10-year overall survival is estimated at 60 and 50%, respectively. The prognosis for LARC has improved in recent years with more access to screening, advances in surgical techniques, and perioperative care. Furthermore, the refinement of the multidisciplinary team with combined-modality management strategies has improved outcomes. These advancements have been augmented by significant improvements in the understanding of the underlying tumor biology. However, there are many instances where patient outcomes do not match those for their tumor stage and accurate prognostic information for individual patients can be difficult to estimate owing to the heterogeneous nature of LARC. Many new combinations of chemotherapy with radiotherapy, including total neoadjuvant therapy with targeted therapies that aim to diminish toxicity and increase survival, are being evaluated in clinical trials. Despite these advances, local recurrence and distant metastasis remain an issue, with one-third of LARC patients dying within 5 years of initial treatment. Although much of the new pathological, molecular genetics, and immunological biomarkers allow refinement in the classification and prognostication of CRC, the relative importance of each of these factors with regards to the development and progression of LARC remains incompletely understood. These factors are often insufficiently validated and seldom consider the individual characteristics of the host, the tumor and its location, the local available expertise, or the probable location of recurrence. Appreciating the mechanisms behind these differences will allow for a more comprehensive, personalized approach and more informed treatment options, leading to ultimately superior outcomes. This review aims to first outline the current multidisciplinary context in which LARC care should be delivered and then discuss how some key prognosticators, including novel histopathological, molecular genetics, and immunological biomarkers, might fit into the wider context of personalized LARC management in the coming years.