Article Text
PDF
Abstract
Aims This study was performed to determine systematically whether KRAS mutational analysis in biopsy tissue is a reliable indicator of KRAS status in subsequent corresponding resection specimens.
Methods 30 colorectal cancer (CRC) patients with biopsy and corresponding subsequent surgical resection specimens were studied. KRAS mutational analysis was performed on each biopsy sample as well as two separate samples from each resection specimen by PCR and Sanger sequencing.
Results Overall, KRAS mutations were identified in 12/30 (40%) of the tumours. There was 100% correlation between biopsy and resection specimens regarding the presence or absence of KRAS mutations. In fact, the same point mutation was identified in both biopsy and corresponding resection specimens in 12/12 (100%) cases. In addition, in two cases, there were two different point mutations detected within the same biopsy specimen.
Conclusion This study shows perfect correlation between KRAS mutation status in biopsy and resection specimens from an individual patient, and suggests that biopsy material is adequate for KRAS mutational analysis in CRC patients.
- Comparison study
- endoscopic biopsy
- GI neoplasms
- KRAS mutational analysis
- laboratory tests
- surgical resection specimen
- tumour biology
Statistics from Altmetric.com
- Comparison study
- endoscopic biopsy
- GI neoplasms
- KRAS mutational analysis
- laboratory tests
- surgical resection specimen
- tumour biology
The KRAS gene, also known as v-Ki-ras Kirsten rat sarcoma viral oncogene homologue, codes for a protein that belongs to the Ras gene family. These include c-Ki-ras-1, c-Ki-ras-2 and N-ras, and all function like guanosine triphosphatase in humans. Normal (wild-type) KRAS can be transiently activated via signals from cell surface receptors by promoting the exchange of bound guanine diphosphate for guanosine triphosphate, the latter playing an essential function in normal tissue signal transduction. However, when KRAS harbours a single point mutation, guanosine triphosphatase activity is reduced, which promotes cell proliferation and the development of cancer.1–3 Seventeen to 25% of all human cancers harbour an activating mutation in the KRAS gene. In colorectal cancer (CRC), activating KRAS mutations are present in 30–40% of patients, most of which occur in codons 12 and 13.4–6
Several studies have found that CRC patients with a KRAS mutation have a poor prognosis, with early metastasis and lower survival rates.7–9 Furthermore, large clinical trials have shown that CRC patients with wild-type KRAS treated with epidermal growth factor receptor (EGFR) inhibition (combined with traditional chemotherapy) have significantly prolonged survival compared with those with KRAS-mutant tumours.10–16 Therefore, using KRAS status to predict whether patients will respond to EGFR inhibitor therapy (panitumumab or cetuximab) has become the standard of care in the pretreatment work-up of patients with stage IV CRC.17–24
To identify KRAS mutations, sequencing or PCR-based genotyping technologies are used to analyse tumour DNA. The vast majority of KRAS testing in CRC patients is performed on resection specimens, in which abundant neoplasm is available for DNA analysis,25–27 although there are also some reports of KRAS analysis performed on exfoliated cells, blood and urine.28–31 In some circumstances, however, the surgically resected specimen may not be available for DNA analysis, such as when neoadjuvant chemotherapy is administered or the patient is not considered a surgical candidate. In these circumstances, the biopsy may be the only specimen available for DNA studies. Potential limitations of the use of biopsies for KRAS testing include: a limited amount of tissue (tumour DNA); difficulty in distinguishing invasive from non-invasive neoplasm; low tumour/non-tumour DNA ratio (that may mask low copy number mutations) and differences in the fixation of small compared with large pieces of tissue. As KRAS mutation is an early event in CRC tumorigenesis,5 ,32 we postulated that biopsy specimens, which comprise the most superficial portion of cancer, would provide sufficient tissue in order to be predictive of the KRAS status of the tumour. Although many laboratories routinely assess biopsy samples for KRAS status, to the best of our knowledge the extent of correlation with the subsequent resection specimen has not been systematically evaluated.25–31 The aim of this study was to compare the KRAS mutational status of tumour tissue from biopsy samples and the patients' subsequent corresponding resection specimens.
Materials and methods
Internal review board approval for the study was obtained from Brigham and Women's Hospital in Boston, Massachusetts, USA. The Brigham and Women's Hospital Pathology Department database was used to identify 30 CRC patients randomly, all of whom had both pretreatment biopsy and subsequent resection specimen tissue available for KRAS mutational analysis (table 1). The patient samples were de-identified, and 30 blocks from biopsy specimens and 60 blocks from the patients' corresponding resection specimens were randomly assigned numbers from 1 to 90. Coded unstained slides from these blocks were submitted to the Miraca Life Sciences Research Institute (formerly Caris Research Institute). One of the slides from each block was stained with H&E and evaluated by a gastrointestinal pathologist (GS). In 12 of the 30 biopsies, foci of intramucosal CRC were distinguishable from foci of invasive CRC, and the intramucosal and invasive areas within these biopsies were analysed for KRAS mutational status separately. A second gastrointestinal pathologist (RL) confirmed the areas selected by the first pathologist and the diagnosis of invasive versus non-invasive carcinoma. For each sample, the area or tissue with the highest percentage of viable tumour nuclei was identified and marked with permanent ink on the corresponding 10-micron unstained slide. The percentage of viable tumour nuclei and the percentage of tumour necrosis were semiquantitatively evaluated. The minimum inclusion criteria to ensure adequate DNA quantity were at least 30% viable tumour nuclei and less than 30% tumour necrosis. In general, 500–1000 viable tumour cells were harvested from tumour measuring at least 2 mm in greatest dimension on a 10-micron unstained slide to derive a minimum of 6 ng/ul of tumour DNA. There was a total of 102 specimens from the 30 patients: 30 biopsy specimens, 60 corresponding subsequent resection specimens (two separate blocks from each patient), plus 12 additional biopsy specimens in which a distinct intramucosal component was identified in the biopsy specimen.
Specimen characteristics
One hundred and two selected areas of tumour from 30 CRC patients were manually macrodissected from corresponding unstained slides and the tissue placed into microcentrifuge tubes. DNA was isolated with the Qiagen (Hilden, Germany) QIAamp DNA FFPE tissue kit to get a minimum final concentration of 6 ng/ul DNA, following the manufacturer's instructions. The extracted DNA was then amplified by PCR with primers flanking KRAS exon 2 (codons 12 and 13) and exon 3 (codon 61). Bidirectional Sanger sequencing using primers flanking the codons of interest was performed using the BigDye terminator chemistry (Applied Biosystems, Foster City, California, USA) on the Applied Biosystems 3730XL. SoftGenetics mutation surveyor v.3.2.3 software was utilised to assist in mutational analysis. Mutations were scored as positive when they were identified and confirmed in both the forward and reverse reactions according to published criteria.33
Results
KRAS mutations were identified in 12/30 (40%) of CRC patients, most at codon 12 (9/12, 75%) and the remainder at codon 13 (3/12, 25%). No mutations were identified at codon 61. Among 45 mutated KRAS samples, the frequency of point mutations was as follows: c.35G>T (16/45, 35.6%), c.35G>A (11/45, 24.4%), c.38G>A (10/45, 22.2%), c.34G>C (4/45, 8.9%) and c.35G>C (4/45, 8.9%). Cancers from the remaining 18 patients demonstrated wild-type KRAS. For each patient, there was 100% correlation between the biopsies and both resection specimen samples tested with regard to the presence or absence of KRAS mutations (table 2). Figure 1 demonstrates the KRAS mutation data in the biopsy and subsequent resection specimen from a CRC patient with a c.35G>T, p.G12V KRAS mutation. All 12 patients with KRAS mutations shared an identical KRAS mutational genotype in their biopsy and corresponding resection specimen. However, in two patients (2/12, 16.7%), there was an additional KRAS point mutation identified within the same biopsy specimen (intramucosal component vs invasive component) as outlined in table 3.
KRAS results
Codon 12 KRAS mutation as shown by DNA sequencing. Identification of KRAS codon 12 gene mutation (c.35G>T, p.G12V) with concordance of the same Gly to Val substitution mutation between biopsy and corresponding resection specimens from the same patient (circles, upper was biopsy and lower was one of the resection specimens). This figure is produced in colour in the online journal—please visit the website to view the colour figure.
DNA sequencing results for KRAS mutations
Discussion
Most published data regarding KRAS mutational analysis in CRC are derived from formalin-fixed paraffin-embedded surgical specimens, from either primary or metastatic lesions.25 ,34–38 However, in some circumstances (eg, when preoperative chemotherapy is administered or when patients are not considered surgical candidates), the resection specimen is not available, and thus the endoscopic biopsy may be the only tissue suitable for KRAS analysis. We performed this study to determine whether KRAS analysis of endoscopic biopsy specimens (with a smaller quantity of tumour DNA) would reliably predict the KRAS status of tissue from the patient's subsequent resection specimen. To the best of our knowledge, no other study has attempted to evaluate and correlate KRAS results between patients' biopsy and their subsequent resection specimen, as in our current study. In order to account for possible tumour heterogeneity, KRAS analysis was performed on two separate tumour resection blocks (in addition to the matching previous biopsy block) for each patient. We detected KRAS mutations (at codons 12 and 13) in 40% (12/30) of our CRC patients, a prevalence rate in keeping with previously published data.4–7 ,18 ,20 ,22 Most importantly, our data showed 100% correlation of the KRAS status (wild-type vs mutant) between the patient's biopsy and corresponding surgical resection specimens. Based on these data, we conclude that standard PCR amplification from biopsy material yields sufficient DNA quantity and quality to predict the mutational status of the tumour reliably.
Although the KRAS mutational status was identical in all samples (biopsies and corresponding resection specimens) for all of the patients in our study, we also found that among the 12 patients with mutations, there were two patients in whom two different KRAS point mutations were detected within the same biopsy specimen. These were detected because in nine of the 12 patients with mutations, both intramucosal and submucosally invasive tumours were present and analysed separately (table 3). In both of these patients, one of the mutations matched that found in the resection specimen, but the other mutation was not detected on the resection specimen. The reason for these results is unclear. It is possible that there may have been a technical aberration (eg, ‘artefactual mutation’ of nucleotide transversions between G>A, G>T or G>C) due to specific tissue sampling, processing, or storage.39 Alternatively, this may simply represent tumour heterogeneity.40 The presence of multiple mutations has led some authorities to suggest that it may be advantageous to evaluate multiple blocks (‘DNA cocktails’) for KRAS testing,41–43 and this potential heterogeneity was the rationale for examining separate resection blocks in our current study.
Some studies have documented heterogeneity with regard to KRAS mutational status, and with regard to specific KRAS mutations. Lamy et al39 reported that five of 44 (11.4%) CRC patients revealed discordant KRAS mutational status results between the primary and metastatic lesions. These authors also described an additional seven patients with discordant KRAS status between separate samples within the same lesion. In a literature review of studies from 1994 to 2011, Macedo et al40 reported the rates of KRAS mutations in which the specific mutation in the primary CRC differed from the mutation found in corresponding metastasis. They found a wide range of discrepancies (0.4–30.8%) among a total of 4062 patients. To the best of our knowledge, no other study has attempted to evaluate and correlate KRAS results between patients' biopsy and their subsequent resection specimen, as in our current study. Recently, Ondrejka et al44 reported that neoadjuvant therapy did not alter KRAS or microsatellite instability results among 17 rectal adenocarcinoma patients with available pretreatment and post-treatment tumour specimens. That study supports the notion that endoscopic biopsy specimens are viable samples for KRAS analysis.
Our study is limited primarily by the relatively small number of samples analysed. In addition, one could argue that additional samples from the resection specimens should have been evaluated. Nevertheless, our finding of a perfect correlation of KRAS status between biopsy and resection samples in the patients in this series suggests that biopsy tissue is an efficacious source of tissue for KRAS analysis, particularly when resection specimens may not be available for analysis.
In summary, this study was designed to assess systematically the adequacy of CRC biopsy tissue to predict overall tumour KRAS mutational status. We showed 100% correlation between biopsies and subsequent corresponding resection specimens, supporting the reliability of using endoscopic biopsy specimens for the assessment of KRAS status before anti-EGFR therapy.
Take-home messages
The KRAS mutation status of CRC from biopsies matches that from corresponding resection specimens.
KRAS analysis of biopsy tissue is reliable for treatment decisions when surgical excision is not performed or resection tissue is not available.
References
Footnotes
Competing interests None.
Patient consent Obtained.
Ethics approval Internal review board approval for the study was obtained from Brigham and Women's Hospital in Boston, Massachusetts, USA.
Provenance and peer review Not commissioned; externally peer reviewed.