Mutant Colorectal Cancer Research Articles

Co-targeting SOS1 enhances the antitumor effects of KRASG12C inhibitors by addressing intrinsic and acquired resistance

Combination approaches are needed to strengthen and extend the clinical response to KRASG12C inhibitors (KRASG12Ci). Here, we assessed the antitumor responses of KRASG12C mutant lung and colorectal cancer models to combination treatment with a SOS1 inhibitor (SOS1i), BI-3406, plus the KRASG12C inhibitor, adagrasib. We found that responses to BI-3406 plus adagrasib were stronger than to adagrasib alone, comparable to adagrasib with SHP2 (SHP2i) or EGFR inhibitors and correlated with stronger suppression of RAS-MAPK signaling. BI-3406 plus adagrasib treatment also delayed the emergence of acquired resistance and elicited antitumor responses from adagrasib-resistant models. Resistance to KRASG12Ci seemed to be driven by upregulation of MRAS activity, which both SOS1i and SHP2i were found to potently inhibit. Knockdown of SHOC2, a MRAS complex partner, partially restored response to KRASG12Ci treatment. These results suggest KRASG12C plus SOS1i to be a promising strategy for treating both KRASG12Ci naive and relapsed KRASG12C-mutant tumors.

CircRAPGEF5 acts as a modulator of RAS/RAF/MEK/ERK signaling during colorectal carcinogenesis

Mutations in oncogenes such as KRAS, NRAS and BRAF promote the growth and survival of tumors, while excessive RAS/RAF/MEK/ERK activation inhibits tumor growth. In this study we examined the precise regulatory machinery that maintains a moderate RAS/RAF/MEK/ERK pathway activation during CRC. Here, using bioinformatic analysis, transcriptomic profiling, gene silencing and cellular assays we discovered that a circular RNA, circRAPGEF5, is significantly upregulated in KRAS mutant colorectal cancer (CRC) cells. CircRAPGEF5 suppressed mutant and constitutively activated KRAS and the expression of the death receptor TNFRSF10A. Silencing of circRAPGEF5-induced RAS/RAF/MEK/ERK signaling hyperactivation and apoptosis in CRC cells suggesting that an upregulation of circRAPEF5 may suppress the expression of TNFRSF10A and aid CRC progression by preventing apoptosis, while the direct interactions between circRAPGEF5 and elements of the RAS/RAF/MEK/ERK pathway was not identified, which nevertheless can be the basis for future research. Moreover, EIF4A3, was observed to share a similar expression pattern with circRAPEF5 and demonstrated to be a major controller of circRAPGEF5 via the promotion of circRAPGEF5 circularization and its silencing reduced circRAPGEF5 levels. Taken together, our findings reveal a mechanism of accurate RAS/RAF/MEK/ERK signaling regulation during CRC progression maintained by upregulation of circRAPGEF5 which may be a plausible target for future clinical applications that seek to induce CRC cell apoptosis via the RAS/RAF/MEK/ERK signaling pathway.

Asparagine synthetase and G-protein coupled estrogen receptor are critical responders to nutrient supply in KRAS mutant colorectal cancer.

Survival differences exist in colorectal cancer (CRC) patients by sex and disease stage. However, the potential molecular mechanism(s) are not well understood. Here we show that asparagine synthetase (ASNS) and G protein-coupled estrogen receptor-1 (GPER1) are critical sensors of nutrient depletion and linked to poorer outcomes for females with CRC. Using a 3D spheroid model of isogenic SW48 KRAS wild-type (WT) and G12A mutant (MT) cells grown under a restricted nutrient supply, we found that glutamine depletion inhibited cell growth in both cell lines, whereas ASNS and GPER1 expression were upregulated in KRAS MT versus WT. Estradiol decreased growth in KRAS WT but had no effect on MT cells. Selective GPER1 and ASNS inhibitors suppressed cell proliferation with increased caspase-3 activity of MT cells under glutamine depletion condition particularly in the presence of estradiol. In a clinical colon cancer cohort fromThe Cancer Genome Atlas, both high GPER1 and ASNS expression were associated with poorer overall survival for females only in advanced stage tumors. These results suggest KRAS MT cells have mechanisms in place that respond to decreased nutrient supply, typically observed in advanced tumors, by increasing the expression of ASNS and GPER1 to drive cell growth. Furthermore, KRAS MT cells are resistant to the protective effects of estradiol under nutrient deplete conditions. The findings indicate that GPER1 and ASNS expression, along withthe interaction between nutrient supply and KRAS mutations shed additional light on the mechanisms underlyingsex differences in metabolism and growth in CRC, and have clinical implications in the precision management of KRAS mutant CRC.

Identification and verification of KEAP1‐related genes and targets regulated by potential ingredients in KRAS mutant colorectal cancer

AbstractAs one of the most prevalent tumors, colorectal cancer (CRC) owns complex pathogenesis and the high recurrence rate significantly influence the prognosis of patients with KRAS mutant CRC, to discovery the crosstalk among KEAP1‐NFE2L2 and oncogenic KRAS as the potential prognostic biomarkers in CRC, which might be used to develop novel and effective therapeutics. By using bioinformatics analysis, we explored and identified key genes as the predicted targets in the crosstalk among KEAP1 oncogenic signatures of KRAS mutation that were linked with development, metastasis, and poor prognosis in CRC. We further investigated the correlations between the clinical characteristics and expressions of prognostic genes among the KRAS and KEAP1‐related key hub genes in CRC, which as predicted targets to find the potential herbal ingredients in HERB database, further to verify the active herbal ingredients could regulate the key targets that related to KEAP1‐NFE2L2 in KRAS mutant CRC. The newest TCGA data reveals that the mutation of KRAS is found in 44% of CRC patients. In total, 28 genes were identified as differentially expressed genes (DEGs), and the hub genes such as CDKN2A, SPP1, FOS, BCL2L11, and HPSE were verified. Results indicated that key genes (SOX9, SPP1) significant correlation with the target prediction of the active herbal ingredients by molecular docking analysis. Furthermore, KEAP1, NFE2L2, SOX9 expression were decreased significantly with the treatment of potential ingredients. Furthermore, cyclopamine could enhance the sensitivity of HCT116 cells, downregulated the expression of SPP1, and induced activation of KEAP1‐NFE2L2 pathway, which cell death are characteristic features of apoptosis, and enhanced anticancer effect. Therefore, KEAP1‐related genes might be important oncogenic signatures in KRAS mutant CRC cells and cyclopamine was identified as a potential ingredient and regulated the predict targets of SOX9 and SPP1, may be expand the efficacy and range of novel and effective therapeutics.

Erlotinib Treatment in Colorectal Cancer Suppresses Autophagy Based on KRAS Mutation.

The KRAS gene is mutated in approximately 45% of colorectal cancer patients. There are currently very few targeted treatments or therapies equipped to directly inhibit KRAS due to its unusual structural intricacies. Erlotinib, an EGFR inhibitor, has previously been demonstrated to reduce cell viability by inducing autophagy in lung cancer cell lines with varying EGFR mutations. In contrast to lung cancer cells, evidence is provided herein for the first time that erlotinib treatment in colorectal cancer (CRC) cell lines reduces autophagy and still results in decreased cell viability. However, the effects of erlotinib in CRC cell lines containing a wildtype KRAS gene were different than in cells carrying a mutant KRAS gene. We show that there is significantly more downregulation of autophagy in KRAS mutant CRC cells compared to KRAS wildtype cells, both at transcriptional and translational levels, suggesting that the KRAS mutation is advantageous for cancer growth, even in the presence of erlotinib. Cell viability results determined that KRAS wildtype CRC cells had significantly more cell death compared to KRAS mutant cells. Using patient mRNA datasets, we showed that there was a significant correlation between the presence of the KRAS mutation and the expression of autophagy proteins. Additionally, through molecular dynamics simulations, we develop a blueprint for KRAS and autophagy protein interaction and the impact of the KRAS mutation on autophagy protein regulation. Overall, this is the first report of erlotinib treatment in CRC cells that assesses autophagy, and we demonstrate that autophagy activity is downregulated in these cells. This effect is not only greater in cells carrying a KRAS mutation compared to wildtype cells, but the KRAS mutant cells also have increased cell viability compared to wildtype cells. We hypothesize that the difference in cell viability and autophagy expression between KRAS mutant and KRAS wildtype cells after treatment with erlotinib can be of therapeutic value to treat CRC patients carrying KRAS mutations.

Detection of PIK3CA gene mutation and its related prognosis in colorectal cancer based on next-generation sequencing

Objectives: To investigate the mutation of PIK3CA in colorectal cancer and to analyze their clinicopathological features, and evaluate their role in clinical treatment and prognostication. Methods: A total of 128 paraffin-embbeded tissue samples of colorectal cancer from Shanxi Cancer Hospital from 2018 to 2021 were collected. DNA was extracted from the samples, and next-generation sequencing (NGS) was used to detect PIK3CA mutation. The relationship between PIK3CA mutation, their clinicopathological features, and prognosis were analyzed. Results: Among the 128 colorectal cancer samples, there were 75 males and 53 females; with aged range 32-86 years, median 61.5 years, 27 (21.09%) had PIK3CA mutations. Colorectal cancer with PIK3CA mutation was more likely to occur in male patients (P=0.007), which was related to tumor site (P=0.032), tumor size (P=0.029) and TP53 wild-type (P=0.001). The common site mutations of PIK3CA mostly occurred in tumors with tumor mutation burden≥10 Muts/Mb (P=0.031).PIK3CA mutation had no significant effect on the survival prognosis of patients, but the efficacy of anti-angiogenic therapy was poor in these patients. Conclusions: PIK3CA mutation is a common mutation in colorectal cancer and plays an important role in the occurrence and development of colorectal cancer. PIK3CA mutation may lead to resistance to anti-angiogenic drugs in colorectal cancer, but its impact on survival and prognosis to patients needs further study.

Improved Biological Impacts of Anti-EGFR Monoclonal Antibody in KRAS-Mutant Colorectal Cancer Cells by Silica-Coated Magnetic Nanoparticle Conjugation

Background: The therapeutic potential of epidermal growth factor receptor (EGFR) targeting in colorectal cancer (CRC) is hindered by the presence of KRAS codon 12 activating mutations, prevalent in approximately 25% of advanced CRC cases. This study investigates the role of reactive oxygen species (ROS) in conferring resistance to anti-EGFR monoclonal antibodies in KRAS mutant CRC cells, focusing on ROS-mediated apoptosis induction using cetuximab-PEGylated silica-coated magnetic nanoparticles (MNPs). Methods: MNPs were synthesized and surface-coated with silica, followed by functionalization and stabilization with polyethylene glycol (PEG). Cetuximab (Cet) was covalently conjugated to generate EMNP-PEG-Cet. Structural and compositional analyses were performed using scanning electron microscopy (SEM), dynamic light scattering (DLS), UV-vis spectroscopy, and Fourier transform infrared (FTIR) analysis. Apoptosis induction, chromatin condensation, and ROS production were evaluated in KRAS mutant SW-480 CRC cells. Results: Successful synthesis of EMNP-PEG-Cet was confirmed, revealing a particle size of 67 nm and a surface charge of -8.3 mV. The conjugate exhibited significant cytotoxicity against CRC cells, with notable apoptosis induction and ROS generation in EGFR-positive/KRAS mutant SW-480 cells, surpassing the effects observed with bare Cet and EMNP-PEG controls. The nuclear factor erythroid 2-related factor 2/Kelch-like ECH-related protein 1 (Nrf2-Kaep1) gene expression analysis by real-time PCR showed that cells treated with EMNP-PEG-Cet exhibited a noteworthy decrease in Nrf2 expression and a simultaneous increase in Keap1 expression compared to those treated with free Cet. Conclusion: These findings highlight the potential of ROS-mediated apoptosis induction to enhance the cytotoxicity of Cet in EGFR-positive/KRAS mutant CRC cells, offering new avenues for overcoming drug resistance mechanisms in metastatic CRC.

Prospective Study to Assess the Clinicopathological and Prognostic Value of BRAF V600E Mutation in Metastatic Colorectal Cancer

Abstract Background BRAF is a protein kinase downstream of RAS in the RAS-RAF-MEK- ERK kinase pathway. The majority (>95%) of BRAF mutations in metastatic colorectal cancer (mCRC) is the BRAFV600E mutation, which is considered an important negative prognostic marker and is associated with resistance to standard chemotherapies in mCRC patients. Objectives This study aimed to investigate the BRAF mutation in patients with mCRC and its association with clinicopathological factors and survival outcomes. Patients and Methods The included specimens were analysed for BRAFV600E mutation by immunohistochemistry (IHC) from confirmed CRC patients with radiological or pathological evidence of metastasis. Results The BRAF mutation was positive in 9.1% of patients. Synchronous metastasis was observed in patients with BRAF mutation with statistical significance (P = 0.033). With regard to treatment, there was no statistically significant difference in patients with BRAF mutations who received targeted chemotherapy as first-line therapy also there was no statistically significant difference in median PFS of patients with BRAF mutation (p = 0.789) and also in median OS (p = 0.343). Conclusion Although the BRAF mutation is a relatively rare finding in mCRC, it is characterized by a critical negative impact on treatment outcome. Given this poor outcome, optimization of therapy is an important goal.

CDK4/6 inhibition to resensitize BRAF/EGFR inhibitor in patient-derived BRAF/PTEN-mutant colon cancer cells.

In v-raf murine sarcoma viral oncogene homolog B1 (BRAF)-mutant colorectal cancer (CRC), encorafenib-cetuximab has been established as standard second-line therapy, but not all patients respond and the duration of response is relatively short. Overcoming intrinsic or acquired resistance to BRAF/EGFR inhibitors is crucial for enhancing treatment outcomes in metastatic BRAF-mutated CRC. The aim of the study is to investigate the resistance mechanisms in BRAF-mutant CRC patient refractory to BRAF/EGFR targeted therapy. We established patient-derived cells (PDCs) from a patient with BRAF/PTEN-mutant metastatic colon cancer who progressed rapidly on encorafenib plus cetuximab. To explore potential treatment options for inherent resistance caused by simultaneous PTEN mutation in BRAF-mutated CRC, we conducted cell viability assays using PDCs treated with encorafenib-cetuximab in combination with a cyclin-dependent kinase-4 and 6 (CDK4/6) inhibitor. The patient's tumor had concurrent PTEN loss-of-function alteration at diagnosis and PDCs were generated from ascites after resistance to the BRAF/EGFR inhibitor. The PDCs were resistant to the encorafenib-cetuximab combination even at a high concentration of cetuximab (up to 500 µg/mL). Adding the CDK4/6 inhibitor, ribociclib, to encorafenib-cetuximab showed a synergistic effect in a proliferation assay. Ribociclib plus encorafenib-cetuximab represented a significantly lower expression of Ki-67 compared to the dual combination alone. An MTS assay showed that triplet therapy with ribociclib, encorafenib, and cetuximab suppressed cell viability more efficiently than the two-drug combinations. Investigating the combined effect of triplet therapy using the calculated combination index (CI) showed that ribociclib had a synergistic effect with encorafenib-cetuximab when applied to PDCs with a concurrent BRAF/PTEN mutation. Our results suggest that combining the CDK4/6 inhibitor with the BRAF/EGFR inhibitormight be a novel treatment strategy for concomitant BRAF and PTEN-mutant CRC.

A machine learning and drug repurposing approach to target ferroptosis in colorectal cancer stratified by sex and KRAS.

The landscape of sex differences in Colorectal Cancer (CRC) has not been well characterized with respect to the mechanisms of action for oncogenes such as KRAS. However, our recent study showed that tumors from male patients with KRAS mutations have decreased iron-dependent cell death called ferroptosis. Building on these findings, we further examined ferroptosis in CRC, considering both sex of the patient and KRAS mutations, using public databases and our in-house CRC tumor cohort. Through subsampling inference and variable importance analysis (VIMP), we identified significant differences in gene expression between KRAS mutant and wild type tumors from male patients. These genes suppress (e.g., SLC7A11 ) or drive (e.g., SLC1A5 ) ferroptosis, and these findings were further validated with Gaussian mixed models. Furthermore, we explored the prognostic value of ferroptosis regulating genes and discovered sex- and KRAS-specific differences at both the transcriptional and metabolic levels by random survival forest with backward elimination algorithm (RSF-BE). Of note, genes and metabolites involved in arginine synthesis and glutathione metabolism were uniquely associated with prognosis in tumors from males with KRAS mutations. Additionally, drug repurposing is becoming popular due to the high costs, attrition rates, and slow pace of new drug development, offering a way to treat common and rare diseases more efficiently. Furthermore, increasing evidence has shown that ferroptosis inhibition or induction can improve drug sensitivity or overcome chemotherapy drug resistance. Therefore, we investigated the correlation between gene expression, metabolite levels, and drug sensitivity across all CRC primary tumor cell lines using data from the Genomics of Drug Sensitivity in Cancer (GDSC) resource. We observed that ferroptosis suppressor genes such as DHODH , GCH1 , and AIFM2 in KRAS mutant CRC cell lines were resistant to cisplatin and paclitaxel, underscoring why these drugs are not effective for these patients. The comprehensive map generated here provides valuable biological insights for future investigations, and the findings are supported by rigorous analysis of large-scale publicly available data and our in-house cohort. The study also emphasizes the potential application of VIMP, Gaussian mixed models, and RSF-BE models in the multi-omics research community. In conclusion, this comprehensive approach opens doors for leveraging precision molecular feature analysis and drug repurposing possibilities in KRAS mutant CRC.

A prospective study to assess the clinicopathological and prognostic value of BRAF V600E mutation in metastatic colorectal cancer

A prospective study to assess the clinicopathological and prognostic value of BRAF V600E mutation in metastatic colorectal cancer

Dietary vitamin D intake and risk of colorectal cancer according to vitamin D receptor expression in tumors and their surrounding stroma.

Colorectal Cancer (CRC) has been molecularly classified into several subtypes according to tumor, stromal, and immune components. Here, we investigated whether the preventive effect of vitamin D on CRC varies with subtypes defined by Vitamin D receptor (VDR) expression in tumors and their surrounding stroma, along with the association of somatic mutations in CRC. In a population-based prospective study of 22,743 Japanese participants, VDR expression levels in tumors and their surrounding stroma were defined in 507 cases of newly diagnosed CRC using immunohistochemistry. Hazard ratios of CRC and its subtypes according to dietary vitamin D intake were estimated using multivariable Cox proportional hazards models. Dietary vitamin D intake was not associated with CRC or its subtypes defined by VDR expression in tumors. However, an inverse association was observed for CRC with high VDR expression in the stroma (the highest tertile vs the lowest tertile: 0.46[0.23-0.94], Ptrend = 0.03), but not for CRC with low VDR expression in the stroma (Pheterogeneity = 0.02). Furthermore, CRC with high VDR expression in the stroma had more somatic TP53 and BRAF mutations and fewer APC mutations than those with low VDR expression in the stroma. This study provides the first evidence that the preventive effect of vitamin D on CRC depends on VDR expression in the stroma rather than in the tumors. CRC with high VDR expression in the stroma is likely to develop through a part of the serrated polyp pathway, which tends to occur with BRAF but not with APC mutations.

Inactivation of KDM6A promotes the progression of colorectal cancer by enhancing the glycolysis

KDM6A (lysine demethylase 6A) has been reported to undergo inactivating mutations in colorectal cancer, but its function in the progression of colorectal cancer has not been evaluated using animal models of colorectal cancer. In this study, we found that knocking out KDM6A expression in mouse intestinal epithelium increased the length of villus and crypt, promoting the development of AOM (azoxymethane)/DSS (dextran sulfate sodium salt)-induced colorectal cancer. On the other hand, knocking down KDM6A expression promoted the growth of colorectal cancer cells. In molecular mechanism studies, we found that KDM6A interacts with HIF-1α; knocking down KDM6A promotes the binding of HIF-1α to the LDHA promoter, thereby promoting LDHA expression and lactate production, enhancing glycolysis. Knocking down LDHA reversed the malignant phenotype caused by KDM6A expression loss. In summary, this study using animal models revealed that KDM6A loss promotes the progression of colorectal cancer through reprogramming the metabolism of the colorectal cancer cells, suggesting that restoring the function of KDM6A is likely to be one of the strategies for colorectal cancer treatment.

PDP1 promotes KRAS mutant colorectal cancer progression by serving as a scaffold for BRAF and MEK1

The oncogenic role of KRAS in colorectal cancer (CRC) progression is well-established. Despite this, identifying effective therapeutic targets for KRAS-mutated CRC remains a significant challenge. This study identifies pyruvate dehydrogenase phosphatase catalytic subunit 1 (PDP1) as a previously unrecognized yet crucial regulator in the progression of KRAS mutant CRC. A substantial upregulation of PDP1 expression is observed in KRAS mutant CRC cells and tissues compared to wild-type KRAS samples, which correlates with poorer prognosis. Functional experiments elucidate that PDP1 accelerates the malignance of KRAS mutant CRC cells, both in vitro and in vivo. Mechanistically, PDP1 acts as a scaffold, enhancing BRAF and MEK1 interaction and activating the MAPK signaling, thereby promoting CRC progression. Additionally, transcription factor KLF5 is identified as the key regulator for PDP1 upregulation in KRAS mutant CRC. Crucially, targeting PDP1 combined with MAPK inhibitors exhibits an obvious inhibitory effect on KRAS mutant CRC. Overall, PDP1 is underscored as a vital oncogenic driver and promising therapeutic target for KRAS mutant CRC.

Development of custom amplicon-based test systems for detecting KRAS, NRAS, and BRAF gene mutations in colorectal cancer using next-generation sequencing (NGS).

e15061 Background: Polymerase chain reaction (PCR)-based mutation diagnostics in colorectal cancer is widely available. However, PCR has limitations in identifying rare mutations and precise mutation coordinates. Targeted sequencing, though promising, often incurs high costs and longer research durations, demanding higher DNA sample quality. Our study demonstrates the benefits of using a customized amplicon test system for diagnosing somatic mutations in tumor tissue, including detecting rare mutations not captured by PCR, requiring less input biomaterial, and similar research costs. Methods: A targeted NGS assay system was developed using region-specific primers for KRAS (exons 2, 3, and 4), NRAS (exons 2, 3, and 4), and BRAF (exon 15). Library preparation involved a three-step PCR process: Target PCR incorporating Illumina overhang adapter sequences, Index PCR introducing multiplexing indices and sequencing adapters, and PCR cleanup using magnetic beads. Sequencing utilized MiSeq v2 reagents with paired 300‐bp reads, generating high-quality, full-length reads. Validation included re-testing known samples. Tumor cell percentages in FFPE specimens were determined, and DNA was extracted using the QIAamp DNA FFPE Tissue Kit. Data analysis utilized proprietary software. Results: Out of 1234 samples, 1133 were tested, with mutations detected in 685 samples (60.45%). KRAS mutations were found in 548 (48.36%), NRAS mutations in 40 (3.5%), and BRAF mutations in 88 (7.7%) samples. Double mutations were observed in 9 samples (0.79%). Data on rare and double mutations are presented in the table. Conclusions: The developed test system accurately identifies essential mutations in colorectal cancer, including rare and potentially clinically significant variants. The required DNA amount ranged from 10 to 30 ng in 5 μl, nearly ten times less than well-known commercial PCR test systems like COBAS (Roche) and therascreen (QIAGEN). The cost of the study is comparable to PCR with a minimum study duration of 4 days. [Table: see text]

79P The prevalence of class I, II, and III BRAF mutations in colorectal cancer and their prognostic significance

79P The prevalence of class I, II, and III BRAF mutations in colorectal cancer and their prognostic significance

134P Rate and clinical implications of uncommon KRAS mutations in colorectal cancer

134P Rate and clinical implications of uncommon KRAS mutations in colorectal cancer

A multi-modal artificial intelligence model trained on H&E tissue images and clinical patient data for prediction of BRAF mutations in colorectal cancer.

e15161 Background: Analyzing genetic mutations is essential for personalized therapy decisions in oncology and beyond. In colorectal cancer (CRC), mutations in BRAF ( BRAFmut) render patients eligible for targeted treatment with BRAF inhibitors and immunotherapy. BRAFmut is routinely analyzed via genomic sequencing, which, however, is costly and does not provide diagnostic results immediately. Reliable detection of BRAFmut status from histological images using artificial intelligence (AI) could provide a more accessible and faster diagnostic alternative. Previous approaches to image-based BRAFmut analysis, however, failed to reach a critical level of accuracy on independent test cohorts, with reported area under the receiver operating characteristic curve (AUROC) values typically below 0.8. Methods: We hypothesized that combining tissue image analysis with clinical patient data could improve diagnostic accuracy for BRAFmut. Following our hypothesis, we developed a multi-modal AI model on n = 426 CRC cases from the TCGA database using BRAFmut data, H&E-stained tissue images, and information on patient parameters that are readily available in clinical practice and that are known to be associated with BRAFmut frequency. The model was based on a vision transformer architecture incorporating pathology-specific information using a pretrained embedding model for encoding tissue image features. Model performance was validated on an independent hold-out TCGA cohort of n = 107 samples and an additional external cohort of n = 104 CRC samples from the CPTAC database. Results: Our multi-modal AI model reached an AUROC of 0.85 ± 0.02 on the TCGA hold-out test set. Notably, on the second, external CPTAC dataset the predictive accuracy level was similar (AUROC of 0.83 ± 0.02) indicating the robustness of our approach to generalize to different datasets. Additional analyses showed that the significant performance boost over previous approaches was gained by the multi-modal analysis setup, as predicting BRAFmut based on either tissue, or clinical data alone led to lower accuracy values. Explainability analysis of AI results confirmed that relevant biological features on the histological slides, mainly tumor areas, were considered by the model for decision making. Conclusions: The combination of H&E images with clinically readily available patient parameters enabled us to develop a model predicting BRAFmut in CRC that surpasses previous approaches in diagnostic accuracy on independent evaluation cohorts. This indicates that combining histological tissue information with clinical parameters can increase the ability of AI models to predict genetic variants from tissue and more generally highlights the potential of multi-modal deep learning approaches to enhance predictive AI models in digital pathology.

Analysis of patch-based GCN risk score distribution corresponding to gene mutations in colorectal cancer using whole slide images.

e15650 Background: Colorectal cancer (CRC) is one of the most common malignancies, with the third highest incidence and the second highest mortality in the world. Risk scores are typically employed to predict outcomes such as disease risk, prognosis, or treatment response. Risk scores are usually calculated by combining biomarkers or genetic variants with statistical models, such as logistic regression or Cox proportional hazards models, and then observe the relationship between genetic variations and survival time through survival analysis. We have constructed a deep learning model based on Graph Convolutional Networks (GCNs), which is capable of predicting risk scores using Whole Slide Images (WSIs). Statistical analyses were applied to explore the distribution of risk scores associated with different variants in colorectal cancer. Methods: We performed risk scores prediction on a total of 384 samples from TCGA-COAD and TCGA-READ using our patch-based GCN model. We analyzed the distribution of risk scores in colorectal cancer samples and investigated the effects of gene mutations in key pathways on these risk scores. Results: Risk score predictions by our patch-based GCN model were conducted on 384 samples, revealing that the maximum, minimum, median, mean, and standard deviation of risk scores were -0.78, -1.89, -1.34, -1.34, and 0.20, respectively. The data overall conformed to a normal distribution ( p < 0.05 ). We selected core genes in key pathways combined with top ten genes with the highest mutation frequency in colorectal cancer for statistical analysis to explore the correlation with the predicted risk scores. These gene mutations were found in a total of 346 patients. The results indicated that mutations in the TP53 and OBSCN genes led to significant differences in risk scores as predicted by our model ( p < 0.05 ). The risk scores based on gene mutations also have significant differences and consistent trend in these genes. Conclusions: Our results indicate the risk scores generated by the patch-based GCN model have the consistent trend with the risk scores generated by mutation information of core genes in colorectal cancer. Whether risk scores based on digital pathology prediction can replace mutation detection in predicting patient survival status requires further exploration.

Incidence and potential prognostic implications of fibroblast growth factor receptor gene alterations and coexisting mutations in colorectal cancer.

e15558 Background: The fibroblast growth factor receptor (FGFR) family comprises four tyrosine kinase receptors (FGFR1-4). FGFR1 amplification (6.3%) as well as overexpression of FGFR1 (10.1%), FGFR2 (5.5%) and FGFR3 (16.2%) have been demonstrated in primary colorectal cancer (CRC, incidence in brackets). The prognostic utility of testing FGFR alterations and association with response to FGFR-inhibitors are still emerging. Here, we report FGFR alteration data with clinicopathological correlations and survival outcomes from a Sydney based Local Health District in Australia. Methods: Clinicopathological data was analysed from 341 ethnically diverse patients with CRC treated at Southwestern Sydney Local Health District from 2014-2021. Next generation sequencing for point mutations, deletions, and amplifications was performed using the Qiagen 30-gene panel. Mismatch repair (MMR) status was determined using immunohistochemical staining. Demographic characteristics, histological features and survival outcomes were recorded from electronic medical records. Results: Of the 341 patients, 42 (12.3%) had FGFR alterations of which 52% were men and 48% women. 32 patients were Caucasian, 4 Middle Eastern, 5 Asian and 1 South American. Median age at diagnosis was 66 years. Of the 42 patients, FGFR1 mutations were present in 26.2%, FGFR1 amplifications in 4.7% and FGFR1 deletions in 4.7%. FGFR2 mutations were present in 16.7%, and 50% had FGFR3 mutations. One patient had concurrent FGFR1 and FGFR3 mutations. The vast majority were adenocarcinoma (81%), followed by mucinous adenocarcinoma (17%) and adenocarcinoma with neuroendocrine features (2%). 21.4% of patients with FGFR alterations had concurrent KRAS mutations, 28.6% had concurrent BRAF mutations (91.6% V600E) and MMR-deficiency was present in 19%. Median overall survival of patients with FGFR alterations was 44 months, for those who also had deficient MMR 48 months, concurrent BRAF V600E mutations 51 months and concurrent KRAS mutations 71 months. 43% (18 patients) were stage 4 at presentation. Of these 18 patients, 7 had FGFR1 mutations, 1 had a FGFR2 mutation and 10 had FGFR3 mutations. The median overall survival of patients with stage 4 disease was 11 months and those with concurrent BRAF V600E mutations 17 months. Two patients had concurrent KRAS mutations (both G12V), with median overall survival of 5.5 months and 1 patient had MMR deficiency, with an overall survival of 9 months. Conclusions: FGFR alterations were identified in 12.3% of patients with colorectal cancer in an ethnically diverse cohort in Australia. Concurrent KRAS and BRAF mutations appear to affect survival outcomes. Further multivariate analyses are underway.