Colorectal cancer is one of the most prevalent cancers in the UK, with 40,000 new diagnoses per year, a 5-year survival rate of 55%, pre-treatment staging data available for only 23% of patients who metastasize, and 15% of patients who undergo surgery but still develop recurrent metastases.
An important development in colorectal cancer treatment has been the emergence of monoclonal antibodies that target EGFR inhibition. monoclonal antibody therapies include cetuximab, bevacizumab, and panitumumab. The available data clearly demonstrate the ineffectiveness of anti-EGFR therapy for KRAS mutant colorectal cancer.
Early data, limited to KRAS codons 12 and 13, were able to predict whether cetuximab treatment would be ineffective, and NICE recommends the results of this study for use in practice. Cetuximab should only be used in hepatic metastatic colorectal cancer and testing for KRAS codon 12 and 13 mutations should be performed when determining a patient’s suitability for this treatment.
KRAS testing has become routine to help stratify patients for anti-EGFR therapy. Other organizations outside the United Kingdom also recommend KRAS gene testing for colorectal cancer treatment.
Dr. Wong’s article in the JCP journal, which is used primarily to guide clinical practice in the UK, also covers NRAS testing for colorectal cancer and reviews the technologies and research that have influenced RAS testing for colorectal cancer, with particular emphasis on the practical implications of the testing. The order of the article follows the procedure of RAS specimen detection.
Case selection
Routine or on-demand testing
Whether RAS testing for colorectal carcinoma should be routine or on-demand remains controversial. The routine testing model means that all surgically resected specimens are routinely genetically tested and the results are described on the pathology report. At some point in the future, if a patient requires anti-EGFR therapy, these molecular data will be immediately available.
Routine testing is a good solution to the lengthy KRAS gene testing time. The lengthy testing time is not a result of the analytical method, but rather the time spent screening for appropriate tissue. Routine testing also avoids the risk of losing tissue blocks, damaging them improperly or preventing RAS testing due to use in other tests.
The disadvantage of routine testing is that it is unnecessary for patients who will never develop metastases. The way to reduce unnecessary testing is to identify colorectal cancer specimens for testing that have high risk factors for metastasis, such as the presence of vascular invasion, lymph node metastasis, or pathologic T4 staging.
The biggest drawback to routine testing is that new data suggest that RAS is more than just KRAS codon 12 and 13 mutations. If a patient is to receive anti-EGFR therapy and has previously been tested for KRAS codons 12 and 13, that patient will now also be tested for NRAS mutations and additional KRAS mutations.
Recent studies have shown that more genes are associated with resistance to anti-EGFR therapy, and new targeted therapies are emerging. On-demand testing is only performed when a case has been discussed by a multidisciplinary team and a decision has been made to receive the relevant treatment. At the time of writing, insurance coverage for RAS testing of patients with metastatic colorectal cancer by laboratories in the UK only covers it, which means that on-demand testing has become the mainstream practice in the UK.
In order to achieve better RAS testing in the UK, it is important to rationalize the distribution of available resources, establish a more rational testing system, improve access to colorectal cancer tissue, and ensure timely delivery of specimens for on-demand testing.
Primary or metastatic tissues
Many studies have examined whether KRAS mutations are consistent in primary or metastatic colorectal cancer tissues. One meta-analysis showed a very high concordance of 94.1%. However, there is some data suggesting that this concordance is related to anatomical location, with lung and lymph node metastases being less concordant with the primary site. The above meta-analysis also showed a concordance of 81.3% between lymph node metastases and primary colorectal cancer.
Because of the lack of absolute genetic concordance between primary and metastatic foci, testing metastatic colorectal cancer tissue is preferred if it is available and can be delivered to the testing laboratory as soon as possible. If metastatic tissue is not readily available, tissue from the primary site should be sent for testing, as there is insufficient evidence to date that metastatic tissue is more specific for RAS genetic testing.
Specimen type
Histopathologic or cytologic specimens can be used for RAS testing, including sections using HE staining or immunohistochemical staining. Some specimens, such as endoscopic or gross needle biopsy specimens, have limited specimens due to specimen acquisition techniques, and in the case of resected specimens, multiple tumor specimens are available, but usually only one representative specimen is selected for examination. The limited number of specimens raises a number of points for consideration.
Biopsy specimens containing only adenomas
The first is if the patient has clear clinical and imaging evidence of colorectal cancer, but the endoscopic biopsy specimen is adenoma-only and it is the only specimen available for RAS testing. What should be done at this point? Some believe that the adenoma specimen should not be tested and the primary or metastatic site should be re-biopsied; others believe that the adenoma should be tested and reported once the mutation is identified, and if there is no mutation, it does not prove that the tumor is not mutated and a re-biopsy is still needed.
The second approach is based on the fact that RAS gene mutations usually occur in the early stages of tumor formation and are key driver mutations, demonstrating the presence of RAS mutations in adenomas suggests that colorectal cancers that develop from adenomas may also have the mutation. However, RAS mutations may also be a late event in colorectal carcinogenesis, meaning that adenomas may be RAS wild type while adenocarcinomas may be RAS mutant type.
A third approach, similar to the second, is to test adenomas with high-grade heterogeneous hyperplasia, a state that is closer to cancer.
Heterogeneity of mutations within the tumor
Another question is if there are multiple resected specimens to choose from, which one should be chosen? This topic relates to tumor mutational heterogeneity, where different clones in the same colorectal cancer may have inconsistent genetic phenotypes. Endoscopic specimens are usually limited and superficial and may not be detected if the mutant clone is located deep in the tumor, or may not be detected in resected specimens where the mutant clone is only present in a specific tissue mass.
Based on the latest KRAS and NRAS data, the presence of any one RAS mutation is generally considered sufficient to predict resistance to anti-EGFR therapy. The coexistence of more than one RAS mutation in the same colorectal cancer is of no further clinical significance.
More clinically significant is the presence of both RAS mutant and wild-type clones in the same tumor, especially when the proportion of mutant clones is low. The latter refers to a low level of mutation, probably because only a small fraction of RAS mutated genes are present in colorectal cancer DNA extracts, which needs to be distinguished from a low level of malignant cells in the tissue mass resulting in dilution of mutated genes in the DNA of non-malignant tumor cells.
Some data suggest that there are also KRAS genotype variations within the same colorectal cancer tissue. A small proportion of colorectal cancers are a mixture of wild-type and KRAS exon 2 and 3 mutant clones, and this genotypic variation is small: 10/13 specimens with a particular KRAS mutant clone exceed 80% of the tumor.
A recent study examined KRAS mutations in 30 paired endoscopic and resected specimens using high-resolution lysis curve analysis and found identical genotypes in each pair.
More sensitive methods were reviewed to analyze whether wild-type colorectal cancers have lower levels of KRAS mutations and whether this level of mutation has an impact on anti-EGFR therapy. 7-20% of wild-type colorectal cancers identified by Sanger sequencing or real time-PCR were re-examined by pyrophosphate sequencing, Therascreen check kit, targeted nucleic acid PCR or mutation amplification PCR and KRAS codon 12 and 13 mutations were found. The effectiveness of anti-EGFR therapy for colorectal cancer at this level of KRAS mutation still needs to be investigated.
The study of intra-tumor mutational heterogeneity may be particularly advanced with the advent of more sensitive tests, and clinical work can now proceed according to guidelines. If both resected and biopsy specimens are available, tissue blocks should be preferred for testing; if only biopsy tissue is available and the result is a wild-type genotype, there is now insufficient evidence to support re-biopsy to exclude low-level mutations.
Mutation testing not only affects treatment selection, but more importantly the presence of low-level mutations may be a marker for predicting future resistance to anti-EGFR therapy. It is often assumed that these mutant clones start in low amounts, but that anti-EGFR therapy drives mutant clone overgrowth that manifests as resistance to therapy when sufficient amounts are present.
Influencing factors in preparation
Review articles have been published detailing what factors in tissue specimen preparation affect subsequent molecular assay results, and the following are particularly relevant for RAS testing in colorectal cancer.
The majority of colorectal cancer tissue used for RAS mutation testing comes from formalin-fixed biopsy specimens or surgically resected primary tumors. Fixation of the latter specimens is usually delayed or poorly fixed, mainly because the colon is not completely dissected or flushed, or only partially fixed specimens are taken.
Delayed or poor fixation results in degradation of DNA due to apoptosis or necrosis and degradation of DNA due to excessive cross-linking by prolonged formalin immersion, which decreases the sensitivity of mutation detection and increases the chance of failure. Formalin fixation can also cause cytosine deamidation, leading to the detection of anthropogenic mutations.
Bouin fixative is no longer commonly used in the UK and earlier blocks of tissue may be fixed by Bouin. Caution is needed when the tissue in the block turns completely yellow or when the tissue contains eluate from the DNA extraction process. Bouin-fixed tissue has a higher chance of failure in molecular assays, due in part to the length of time the tissue has been stored and also to the fact that some components of Bouin fixative can accelerate nucleic acid degradation.
In some hospitals, it is customary to anchor endoscopic biopsy specimens to acetate strips before fixation, which can result in lower DNA yields if the acetate strips are not removed. Recording tumor cell content is appropriate for molecular analysis. There is no technical difference between tissue being sectioned or cut directly for DNA extraction, and the tissue should be extracted as soon as possible after excision to reduce DNA oxidation.
Malignant cell content
When somatic mutant tissues contain both malignant and non-malignant cells, it is critical to accurately quantify the malignant cell content. The proportion of malignant tumor cell DNA in the final DNA extract is a good indicator of the proportion of histopathologically malignant nucleic acids in all nucleic acids and is preferable to measuring the proportion of area occupied by tumors.
This is particularly important in colorectal cancer, where necrotic and cell-free areas should not be counted and should be removed whenever possible. Aneuploidy chromosomes are common in tumors and may potentially increase or decrease detection sensitivity. This article recommends that the minimum tumor cell content for detection of mutations should be two times the minimum tumor cell content for detection.
For example, if a method requires a minimum of 5% tumor cells, a minimum of 10% malignant cells in the tissue would be appropriate for mutation testing. There is variability in the observer assessment of tumor cell content in colorectal cancer, so the laboratory would like to use a safe minimum tumor cell content for analysis, but this ratio needs to be further validated.