Update of the recommendations for the determination of biomarkers in colorectal carcinoma: National Consensus of the Spanish Society of Medical Oncology and the Spanish Society of Pathology

The identification of this type of syndrome is based on the Amsterdam I and II criteria [6, 7], which are exclusively clinical. When a family complies, genetic testing should be offered to identify the presence of germline mutations in DNA repair genes (mismatch repair [MMR] genes), such as MLH1, MSH2, MSH6 and PMS2. This genetic study should always be accompanied by adequate genetic counselling.

The algorithm for the identification of these mutations is well described and accepted by different scientific societies (Fig. 1). Initial screening is performed by identifying the lack of expression of the corresponding proteins of the genes involved in CRC and/or endometrium or by analysing microsatellite instability [MSI]. Any of these techniques is fully accepted today, and each has its specificities. In the case that any of these alterations exist, and once it has been ruled out that it is due to an exclusively somatic alteration, the germinal study of the MMR genes should be performed.

Notably, at present, this type of study is being performed in any patient with CRC. This measure is justified not only by the preventive implications for the patient and their families but also because it helps to decide whether to administer adjuvant therapy in patients in whom the indication of such treatment is doubtful [8].

The main characteristic of polyposis syndromes, which usually affect the younger population, is the development of a large number of polyps along the digestive tract, with different histologies and associated with other systemic lesions, with no histological markers to identify them.

Taking into account this information, this panel of experts recommends the following:

The main biomarker of clinical utility in metastatic CRC is the mutational state of genes in the RAS family (KRAS/NRAS/HRAS). Mutations are present in approximately 50% of cases, the most frequent being those that occur in codons 12 and 13 of exon 2 and that condition a state of cellular expansion.

RAS mutations are a predictor of the absence of response to treatment with epidermal growth factor receptor (EGFR) inhibitors, such as cetuximab and panitumumab. In the CRYSTAL study, the addition of cetuximab to the regimen with folinic acid, 5-fluorouracil and irinotecan (FOLFIRI) showed an increase in the response rate (57% versus 40%; p < 0.001) and median OS (23.5 versus 20.0 months; HR 0.796; p = 0.0093) in patients with tumours without KRAS mutations [20]. However, in patients with tumours with KRAS mutations, a 4% decrease in the response rate was evidenced, and there were no significant differences in terms of OS [20]. Similar findings were found when analysing the OPUS trial, in which the chemotherapy regimen was folinic acid, 5-fluorouracil and oxaliplatin (FOLFOX) with cetuximab, and in the PRIME trial, in which FOLFOX was combined with panitumumab [21, 22]. Taking into account these data [23], guidelines from the American Society of Clinical Oncology (ASCO), the European Society for Medical Oncology (ESMO) and the National Comprehensive Cancer Network (NCCN) recommend the extended determination of RAS mutations before the administration of EGFR inhibitors [8, 24, 25]. The RAS analysis should include HRAS, KRAS and NRAS, exons 2 (codons 12 and 13), 3 (codons 59 and 61) and 4 (codons 117 and 146) [8, 24, 25].

The most frequent BRAF mutation is the substitution of glutamic acid for valine in codon V600E, which produces constitutive activation of the MAPK pathway. This mutation occurs in 8–10% of patients with metastatic CRC and excludes the RAS mutation [26]. It is more common in women, usually appears in more advanced stages and has been associated with tumours of the right colon, poorly differentiated, with mucinous histology and with MSI. Additionally, it seems that BRAF mutations lead to the development of more peritoneal metastases and less disease limited to the liver or lung.

Regarding the prognostic role of BRAF, it seems that mutation is a negative prognostic factor, although its association with MSI has to be taken into account. A joint analysis of four phase III clinical trials of first-line treatment in patients with metastatic CRC evaluated the possible prognostic value of BRAF [27]. In a total of 3063 patients, of whom 8.2% had a BRAF mutation, the OS was 11.4 months, and the PFS was 6.2 months, compared to 17.2 months and 7.7 months, respectively, that were observed in patients without BRAF mutations (p < 0.001). This same study analysed the possible prognostic value of MSI, and it was observed that the OS for patients with MSI of high instability (MSI-H) was 13.6 months, while in patients with MSs, it was 16.8 months (p = 0.001). When the BRAF and MSI statuses were analysed, there were no differences in terms of PFS or OS in the MSI-H population regardless of BRAF status, but in the MSs population with mutated BRAF, a significant decrease in OS and PFS was observed. These data support the negative prognostic value of BRAF mutations.

Recent data support the predictive value of BRAF mutations [28]. Phase III BEACON Colorectal Cancer Study enrolled 665 patients with BRAF V600E across three cohorts, including the triplet regimen of encorafenib, binimetinib and cetuximab; the doublet regimen of encorafenib and cetuximab; and a control cohort where patients could receive either cetuximab and irinotecan or cetuximab with FOLFIRI. At the time of the interim analysis, the median duration of follow-up was 7.8 months. The median OS was 5.4 months, 8.4 months and 9.0 months for the control group, the doublet regimen arm, and the triplet regimen arm, respectively. Comparing the results obtained in the triplet regimen arm with the control group, HR was 0.52 (p < 0.001). An independent central review committee assessed the response rate in this analysis for the first 331 randomized patients. The response rate in the control arm was 2% compared with 20% in the doublet regimen arm and 26% in the triplet regimen arm (control arm versus triplet regimen arm; p < 0.001).

Mutations in the MLH1, MSH2, MSH6 and PMS2 genes are related to poor MMR as well as MSI. They appear in 15% of stage II and III tumours but in only 4% of stage IV tumours. In stages II–III, MSI is related to a better prognosis, and specifically in stage II, the better prognosis occurs in the absence of benefits from adjuvant treatment with fluoropyrimidines. However, MSI in stage IV is associated with an unfavourable prognosis, most likely due to its association with BRAF mutations.

Two recent studies show that the alteration of MMR may have a predictive role in response to immunotherapy; therefore, the NCCN recommends its determination in all patients with metastatic CRC [29, 30]. The Food and Drug Administration (FDA) has approved the use of nivolumab and pembrolizumab in patients with MMR-deficient and MSI-H metastatic CRC.

The most important prognostic factors in CRC according to the American Joint Committee on Cancer (AJCC) and the 8th edition of the International Union Against Cancer (UICC) are (i) TNM stage, (ii) venous, lymphatic or perineural invasion and (iii) discontinuous tumour deposits [41]. The determination of TB has been incorporated into the pathology diagnosis protocol of the College of American Pathologists (CRC) v4.0.0.1 and the NCCN guidelines v2.2019. Other factors, such as type, histological grade, the configuration of the tumour border and intratumoural inflammatory infiltrate, influence the prognosis and treatment of patients. Tumour budding or TB, defined by the presence of isolated tumour cells and/or groups of less than five cells in the stroma of an invasive front of the tumour, is the morphological evidence of the epithelial-mesenchymal transition process and is an independent factor of poor prognosis in CRC. TB is relevant in different clinical contexts: (i) in pT1 CRC as an independent predictor of lymph node metastasis, (ii) in stage II CRC predicts recurrence risk and contributes to therapeutic management [42, 43], and (iii) intratumoural TB in endoscopic biopsies allows the individualization of neoadjuvant treatment in rectal carcinomas [44, 45]. The International Tumour Budding Consensus Conference (ITBCC) agreed to count TB with HE at the area of the tumour front with the most density using a 20× objective and a field of 0.785 mm². TB is classified as Bd1-low (0–4 buds), Bd2-intermediate (5–9 buds) and Bd3-high (10 or more buds). Providing the numerical value of buds [e.g., Bd3 (count 17)] is recommended [46]. TB should be evaluated with caution in medullary carcinomas and can be very difficult to count in tumours with glandular fragmentation with marked acute inflammation. TB must be distinguished from PDCs, defined as groups of five or more tumour cells in the invasive tumour front that do not form glandular structures [4]. TB should not be reported in rectal tumours treated with neoadjuvancy, or in mucinous carcinomas with signet ring cells within mucus lakes. If TB cannot be assessed, it is indicated as no non-assessable with an explanatory note [47, 48].

Lymph node staging (pN) is an important prognostic factor in CRC, a predictor of relapse and survival and determinant of the therapeutic management of patients. The pN stage is obtained from the analysis of lymph nodes stained with HE, which has low sensitivity for detecting lymph node micrometastases in stages I-II CRC. HE staining analyses less than 1% of lymph node tissue [49‐53], resulting in 11–24% false negatives. A meta-analysis concluded that the presence of micrometastases in lymph nodes undetected by HE is associated with worse survival [54]. To improve the sensitivity of lymph node analysis, molecular lymph node staging has been performed, which has shown an improvement in pN staging, with respect to HE.

The OSNA technique detects copies of CK19 messenger RNA (mRNA) in lymph nodes, achieving overstaging in 11–50% of HE-determined pN0 patients. Likewise, the total tumour load (TTL) or number of CK19 copies in the lymph nodes is correlated with classical high-risk factors in CRC, such as pT and pN stage, histologic grade, mucinous or signet ring histology, tumour size, male sex, lymphatic invasion and number of lymph nodes assessed [55, 56]. Recently, TTL has been correlated with prognosis. The incorporation of molecular lymph node staging in CRC will allow to select stage II patients for adjuvant therapy and assist patient management.

Currently, the primary source for biomarkers studies are FFPE tumours and normal tissues, stored at room temperature.

Although the use of formaldehyde solution is the universal method of tissue fixation, formaldehyde produces covalent bridges in aqueous solutions and interacts with the -NH2 groups of proteins, forming cross-linked methylene bridges; the tertiary and quaternary conformations of these bridges are reversible (epitope retrieval), and they interact with DNA by forming hydroxy methylene bridges between two amino groups. In addition, treatment with formaldehyde can produce aberrant mutations by the formation of apurinic and apyrimidinic sites, DNA degradation and cross-linking of cytosines and has recently been considered carcinogenic. Therefore, alternative fixatives have been developed to improve the preservation of nucleic acids. Among these, PAXgene^® Tissue Fix (PreAnalytix GmbH Hombrechtikon, Switzerland) and glyoxal acid-free (GAF) stand out [57, 58]. The prolonged storage of paraffin blocks at room temperature causes the constant degradation of the nucleic acids regardless of the type of fixative used. To slow this degradation process, the blocks should be stored at 4 ºC or frozen [59].

It is crucial to select tumour material through macro- or microdissection to avoid contamination with normal tissue. To guarantee results, a tumour cell content 10% higher than normal cell content is considered optimal, thus avoiding the possibility of false results. Currently, with the latest generation of sequencing technologies, a threshold of 5% could be established.

The ideal study sample contains the greatest possible number of tumour cells. Thus, it is preferable to compare a resected piece of CRC with the content of a diagnostic endoscopic biopsy or cytological material [60].

Currently available technologies for mutational analysis of the KRAS and NRAS genes have not been significantly modified compared to the last guide [61, 62], except for the incorporation of the mutational status of the BRAF gene [25], as reflected in Table 2.

A large number of laboratories continue to use molecular techniques based on sequencing as confirmation techniques of mutations (Sanger sequencing and pyrosequencing) or as screening techniques based on NGS, where different platforms allow these determinations to be performed simultaneously in both clinical trials and in research studies as well as in the assistance activities of several molecular pathology laboratories in Europe [63]. However, the vast majority of laboratories have focused their activity on the use of specific assays to detect specific alterations in these genes (KRAS, NRAS and BRAF) using commercial kits by Conformité Européenne-In Vitro Diagnostic (CE-IVD) at the expense of the tests developed by each laboratory (laboratory developed test [LDTs]), whose use is limited in Spain in comparison with other European countries, such as Holland and Germany [64].

Notably, many of the kits used to determine KRAS/NRAS mutations do not determine BRAF mutations simultaneously; therefore, many laboratories have had to modify their workflows as well as combine molecular techniques to analyse these mutations sequentially (e.g., first, KRAS gene mutations are analysed, then NRAS gene mutations and finally BRAF gene mutations), which increases response times, sample manipulation and complexity for obtaining these results.

In the case of the BRAF gene, the majority of the mutations focus on the valine amino acid at position 600, which can be affected in multiple ways (V600E/E2/K/D/R or M); however, given that the vast majority correspond to the p. V600E change, this mutation can be detected by a specific antibody (Clone VE1) via IHC. However, although this technique has shown high sensitivity and analytical specificity in different studies [65, 66], confirmation is currently required by a molecular technique.

In recent years, there has been an increase in the use of commercial point of care (POC) techniques that minimize and automate the process of determining gene mutations, significantly reducing response times, reducing the risk of obtaining false positives/negatives and avoiding DNA extraction because analyses can be performed directly with an FFPE sample, with minimum manipulation by the technical staff and with validated results against conventional techniques and NGS [67], even using DNA extracted from paraffin-embedded samples [68].

In the coming years, it is believed that the use of specific algorithms based on the use of artificial intelligence will increase and be used to analyse both radiological and HE images, which will produce RAS analyses with greater precision [69, 70].

For this purpose, in the majority of pathology laboratories, IHC analysis of the four repair genes (MLH1, MSH2, MSH6 and PMS2) is performed; however, because the encoded proteins act in tandem, different groups have proposed that only IHQ determination of MLH1 and MSH2 can be used as a screening technique [71].

In addition, PCR techniques have been used for the detection of MSI. When developing the methodology, we began using a panel of microsatellites (Bethesda panel) composed of markers that detect single base alterations (BAT25 and BAT26) as well as three dinucleotide markers (D2S123, D5S346 and D17S250), but recently, new mononucleotide satellites have been incorporated, which have shown better prediction capabilities than those of previous dinucleotides [72].

Finally, both recently implemented NGS techniques and new POC tools allow the determination of MSI with new satellites included in their panels and are being validated in different studies [73].

The determination of the amplification of HER2 in CRC has generated great interest because of its prognostic and predictive value of response to specific treatments because approximately 3% of patients have such alterations [74].

The methods of analysis are the same as those used in breast and gastric cancers, where IHC (HercepTest^® or Clone 4B5) stands out, showing great concordance with the gene amplification detected by FISH or in silver in situ hybridization (SISH) [34]. Currently, both the determination of the copy number alterations (CNAs) in a tumour by NGS techniques and the quantification of the levels of HER2 mRNA are being validated in different studies before their implementation in routine healthcare [39].

The rearrangements of the NTRK1, NTRK2 and NTRK3 genes are of special relevance in CRC because, although their prevalence is very low (1.5%) [39], these alterations are associated with the presence of MSI, hypermethylated MLH1 and native RAS [75]. In addition, it is important to mention that they have a great predictive value of response to TRK inhibitors. The classical techniques for the detection of NTRK alterations are generally performed by IHC with pan-TRK clones as screening, with subsequent confirmation by different tests (FISH, NGS, RT-PCR or NanoString®); however, each one has its limitations [76].