In some instances the standard method of doing molecular testing from formalin-fixed, paraffin-embedded block is not possible because of limited tissue. Tumor cell–enriched cell-transfer technique has been proven useful for performing immunocytochemistry and molecular testing on cytologic smears.
To establish the cell-transfer technique as a viable option for isolating tumor cells from hematoxylin-eosin (H&E)–stained slides.
Molecular testing was performed by using the cell-transfer technique on 97 archived H&E-stained slides from a variety of different tumors. Results were compared to the conventional method of molecular testing.
Polymerase chain reaction–based molecular testing via the cell-transfer technique was successfully performed on 82 of 97 samples (85%). This included 39 of 47 cases for EGFR, 10 of 11 cases for BRAF, and 33 of 39 cases for KRAS mutations. Eighty-one of 82 cell-transfer technique samples (99%) showed agreement with previous standard method results, including 4 mutations and 35 wild-type alleles for EGFR, 4 mutations and 6 wild-type alleles for BRAF, and 11 mutations and 21 wild-type alleles for KRAS. There was only 1 discrepancy: a cell-transfer technique with a false-negative >KRAS result (wild type versus G12C).
Molecular testing performed on H&E-stained sections via cell-transfer technique is useful when tissue from cell blocks and small surgical biopsy samples is exhausted and the only available material for testing is on H&E-stained slides.
The understanding of the molecular pathogenesis of cancer has led to the development of chemotherapeutic agents that target specific genetic alterations such as mutations in the epidermal growth factor receptor (EGFR) and BRAF genes. It is estimated that well over a million people are living with colorectal cancer and 50% to 70% of those with advanced-stage disease receive adjuvant chemotherapy. The US Food and Drug Administration has approved the use of specific targeted chemotherapeutic agents such as panitumumab and cetuximab, which are EGFR inhibitors.1 Retrospective subset analyses of metastatic or advanced colorectal cancer trials showed that patients with a colorectal tumor-bearing mutated KRAS did not benefit from cetuximab, whereas patients with a tumor-bearing wild-type KRAS did benefit from cetuximab.2 Adenocarcinoma represents the most common type of lung cancer, and understanding of its molecular pathogenesis has led to the development of several novel chemotherapeutic agents that offer potentially effective therapy by targeting specific genetic alterations such as mutations in EGFR and KRAS. EGFR mutations are found in approximately 10% to 15% of non–small cell lung cancers with the highest frequency occurring in adenocarcinoma. Lung adenocarcinomas driven by EGFR mutations are sensitive to tyrosine kinase inhibitors such as gefitinib and erlotinib and affected patients will have longer progression-free survival than patients whose tumors do not contain EGFR mutations.3–8 BRAF mutations, detected in regionally advanced or metastatic melanoma, appear to be one of the initiating steps in the development of primary melanoma. The discovery of activating mutations in BRAF has led to the development of molecular-targeted therapy.9 The use of these and other targeted chemotherapeutic agents necessitates timely and accurate molecular testing of good-quality tumor cells.
In some instances the conventional method of doing molecular testing on formalin-fixed, paraffin-embedded (FFPE) tumor is not possible because of limited tissue. We have shown that the cell-transfer technique (CTT) can be used as one tool to help achieve testing in such circumstances.
The CTT has proven to be a reliable and useful method for performing immunocytochemistry and molecular testing by facilitating the use of direct cytologic smears as a source of tumor cells in cases when the cell blocks lack adequate cellularity.10–16 This technology can also be applied to hematoxylin-eosin (H&E)–stained sections when the cells of interest are exhausted from the paraffin block. In this study we simply set out to establish the CTT as a viable option for isolating tumor cells from H&E-stained slides. This is helpful when tissue from cell blocks and small surgical biopsy samples is exhausted and the only available material for testing is on H&E-stained slides. We did the molecular testing with our standard polymerase chain reaction (PCR)–based platforms. To validate this technique, we tested corresponding FFPE tissue in addition to the formalin-fixed, H&E-stained sections.
MATERIALS AND METHODS
This study was approved by the Institutional Review Board of Indiana University (protocol No. 1401456334). A variety of consecutive surgical pathology cases for which molecular testing (EGFR, BRAF, or KRAS) was ordered within a 10-month period were selected for this prospective study. Cases were required to have extra material for this study so as not to compromise diagnostic workups. The cases selected had extra slides cut, were stained (H&E), and set aside. They consisted mainly of core or forcep biopsies of colonic and lung adenocarcinomas with occasional cases of melanoma, squamous cell carcinoma, and papillary thyroid carcinoma (Tables 1 through 3). After cases were signed out, they were blinded to the molecular results from conventional method and sent for molecular testing through CTT. The archived H&E-stained slides were reviewed and areas with higher tumor cell concentration were selected by marking the slide. Most samples contained greater than 50% of tumor nuclei and at least 200 tumor cells. The marked tumor cells were removed from H&E-stained slides through CTT and submitted for EGFR (47 cases), BRAF (11 cases), and KRAS (39 cases) testing. The results were correlated to the previous conventional method for molecular testing. Representative samples of tumor histology and corresponding mutational result are illustrated (Figures 1 through 3).
Adenocarcinoma of lung with EGFR 19 deletion (hematoxylin-eosin, original magnification ×400).
Adenocarcinoma of lung with EGFR 19 deletion (hematoxylin-eosin, original magnification ×400).
![Figure 3. . A, Metastatic melanoma with BRAF V600E mutation. B, Adenocarcinoma of colon with KRAS G12C mutation. C, BRAF V600E mutation corresponding to (A) is detected during cycling for both the sample (purple, patient) and sample control (blue). A difference of 7.00 or less between the crossing threshold cycles is an acceptable cutoff for a positive V600 result. The delta threshold cycle values of these samples (0.34) demonstrate detection of V600E BRAF mutation in relation to the sample control. D, KRAS G12C mutation corresponding to (B) is detected by demonstrating a delta cycle threshold less than the stated cutoff value for the assay (hematoxylin-eosin, original magnification ×400 [A and B]). Abbreviation: Norm. Fluoro., normalized fluorescence.](https://allen.silverchair-cdn.com/allen/content_public/journal/aplm/140/12/10.5858_arpa.2015-0454-oa/2/m_i1543-2165-140-12-1383-f03.png?Expires=1744123111&Signature=H6A0XM6yyWE3jwUREtl89wSF5TxtnIyQbw94qbS3BYVu2~LSwwp50BNANsm7bvaR9x3nFnGw3pztey~sfvTaP2qWBEXG0fnWcRg2quBwIK-i5zkJwRez8ban5rRMxOhRtPbTe-EVi6Jg2TYLy0bN-zY8VdtUp4PBDjaXGv7P2nDct1WeEB-smjHZUWzYS2jKDXql0He~F5~TeVUjcSSFp5CkMDrTtG5AIKK4twhD7jXYeZ1K2xPSP97UOxXm9~a6Py0A7vzgtoRAZ2XynHdqlhekgZ9DXHKa5QU5yBZ8amMooN93HZQIuQNSoN5ykjDniMYqHA7IqnVqMAMo~GQF7w__&Key-Pair-Id=APKAIE5G5CRDK6RD3PGA)
A, Metastatic melanoma with BRAF V600E mutation. B, Adenocarcinoma of colon with KRAS G12C mutation. C, BRAF V600E mutation corresponding to (A) is detected during cycling for both the sample (purple, patient) and sample control (blue). A difference of 7.00 or less between the crossing threshold cycles is an acceptable cutoff for a positive V600 result. The delta threshold cycle values of these samples (0.34) demonstrate detection of V600E BRAF mutation in relation to the sample control. D, KRAS G12C mutation corresponding to (B) is detected by demonstrating a delta cycle threshold less than the stated cutoff value for the assay (hematoxylin-eosin, original magnification ×400 [A and B]). Abbreviation: Norm. Fluoro., normalized fluorescence.
A, Metastatic melanoma with BRAF V600E mutation. B, Adenocarcinoma of colon with KRAS G12C mutation. C, BRAF V600E mutation corresponding to (A) is detected during cycling for both the sample (purple, patient) and sample control (blue). A difference of 7.00 or less between the crossing threshold cycles is an acceptable cutoff for a positive V600 result. The delta threshold cycle values of these samples (0.34) demonstrate detection of V600E BRAF mutation in relation to the sample control. D, KRAS G12C mutation corresponding to (B) is detected by demonstrating a delta cycle threshold less than the stated cutoff value for the assay (hematoxylin-eosin, original magnification ×400 [A and B]). Abbreviation: Norm. Fluoro., normalized fluorescence.
Cell-Transfer Technique
The CTT was performed by using clean technique as follows: (1) the coverslip was removed with fresh histologic-grade xylene (Fisher Scientific, Pittsburgh, Pennsylvania); (2) a thin layer of Mount Quick media (Daido Sangyo, Tokyo, Japan) was spread uniformly over the top of the cellular material; (3) the slide was then placed in a 60°C heated oven for approximately 2 to 3 hours (or until hardened to the touch); (4) a Sharpie marker was used on the surface of the dried media to divide the slide into multiple areas of interest; (5) the slide was then placed into a clean Coplin jar of deionized water and submerged into a warm water bath at 45° ± 3°C for 30 minutes to 2 hours, or until the media was soft enough to easily peel away from the slide; and (6) the media was cut along the marked areas, and each cut section was peeled off and placed in an Eppendorf 2.0-mL safe-lock centrifuge tube and sent for molecular testing. No extra steps were needed to get the DNA out of Mount Quick medium. If there are unused sections on the slides, the medium can be removed through 4 exchanges of xylene (15 minutes each) and the slides are re-coverslipped for storage.
Conventional Technique
The H&E-stained slides are reviewed by a pathologist and the areas with rich tumor cells are circled with a marker. The H&E-stained slides along with 8 to 10 unstained slides are then sent to our molecular pathology laboratory. The areas of interest from 8 to 10 unstained slides are scraped off with razor blades and placed in an Eppendorf 2.0-mL safe-lock centrifuge tube for molecular testing.
DNA Extraction
We applied the same DNA extraction method for both standard and CTT samples. It was performed by using the Qiagen QIAamp DNA Formalin-Fixed, Paraffin-Embedded Tissue Kit (Qiagen, Valencia, California). A modification from the manufacturer's recommendations was made. Samples were incubated at room temperature for 5 minutes with 1 mL of xylene and were then centrifuged at 15,000 rpm for 5 minutes. Xylene was removed from the pellet and ethanol wash was then performed as recommended by the manufacturer. DNA concentration was determined by using the NanoDrop Spectrophotometer (Thermo Fisher, Waltham, Massachusetts).
EGFR
After DNA concentration was measured, the DNA was adjusted to approximately 10 ng/mL in distilled water. For EGFR mutational analysis, PCR-amplified products were analyzed on the Q24 Pyrosequencer with Qiagen EGFR Pyro kits (Qiagen). The pyrosequencing kit tests for mutations within exon 18 at codon 719, deletions in exon 19, mutations within exon 20 at codons 768 and 790, and mutations within exon 21 at codons 858 through 861. The resulting amplicons were purified, denatured, and sequenced by using mutation adjacent primers. Pyrograms were generated by the software and interpreted for the presence of mutations in the corresponding codons. The analytic sensitivity of this test is 5% (mutant allele detection).
KRAS
Samples were run by using the Qiagen therascreen KRAS RGQ PCR on the Rotor-Gene Q MDx following the manufacturer's recommendations. Genomic DNA was used to detect 7 somatic mutations in codons 12 and 13 of the KRAS oncogene by using real-time PCR on the Rotor-Gene Q instrument using both Scorpions and Amplification Refractory Mutation System technologies (Qiagen).17–19 The somatic mutations capable of being detected were Gly12Ala, Gly12Asp, Gly12Arg, Gly12Cys, Gly12Ser, Gly12Val, and Gly13Asp. The reaction mixes were duplex, containing reagents labeled with FAM to detect mutant targets and HEX to detect the internal control. The threshold at which the signal is detected above background signaling is called the cycle threshold. Sample delta cycle threshold values are calculated as the difference between the mutation assay cycle threshold and wild-type assay cycle threshold from the same sample. Samples are subsequently classified as mutation positive if they give a delta cycle threshold less than the stated cutoff value for the assay and classified as not detected if above this value. The data were analyzed by using Rotor-Gene Q series software. Appropriate positive and negative controls were run with each sample. The analytic sensitivity of this KRAS test is 1% (mutant allele detection).
BRAF
The BRAF RGQ PCR Kit used is a real-time qualitative PCR assay used on the Rotor-Gene Q MDx instrument for the detection of 4 somatic mutations in the human BRAF oncogene. The reagents used in this validation included BRAF RGQ PCR Kit (catalog No. 870801). The histologic cell-transfer material was received in the molecular pathology laboratory in Eppendorf tubes. DNA was extracted by using the QIAamp DNA FFPE Tissue Kit (catalog No. 56404). A modification was made from the protocol to include a 5-minute incubation time after the addition of 1 mL of xylene and then a 5-minute centrifugation was performed. The protocol continued after this modification to match manufacturer recommendations. The analytic sensitivity of this BRAF test is 5% (mutant allele detection).
RESULTS
The CTT-generated molecular testing results, obtained from archived H&E-stained sections, were compared to those generated by the conventional method (Tables 1 through 3). The average DNA content for BRAF, KRAS, and EGFR was 29.04, 58.25, and 40.34 ng for conventional method versus 6.1, 7.67, and 7.05 ng for CTT samples. Polymerase chain reaction–based molecular testing via the CTT was successfully performed on 82 of 97 samples (85%). These included 39 of 47 cases (83%) for EGFR, 10 of 11 cases (91%) for BRAF, and 33 of 39 cases (85%) for KRAS mutations (Tables 1 through 3). The CTT samples showed agreement in 99% (81 of 82) of cases with the conventional method. (The 95% confidence interval extends from 0.9276 to greater than 0.9999, computed by the modified Wald method through an online calculator [http://www.graphpad.com/quickcalcs/ConfInterval1.cfm]). These included 4 mutations and 35 wild-type alleles for EGFR, 4 mutations and 6 wild-type alleles for BRAF, and 11 mutation and 21 wild-type alleles for KRAS. There was only 1 discrepant case (Table 3, KRAS case No. 13), namely, a false-negative KRAS result. The CTT sample showed wild-type allele, while the corresponding conventional method sample demonstrated Gly12Cys (G12C) mutation.
DISCUSSION
The CTT is a proven method for obtaining cellular material from fine-needle aspiration direct smears and facilitates additional immunocytochemical staining or molecular testing.10–13 In this study, we confirmed its utility to capture sufficient tumor cell DNA for molecular testing if the tissue-embedded paraffin block lacks material. Biopsy samples or fine-needle aspiration cell blocks may only contain a minimal amount of tumor tissue and could be exhausted after the initial routine H&E sections and immunohistochemistry studies. It is a common scenario in which a pathologist signs out the final pathologic report and then receives a phone call requesting that molecular studies be performed. This molecular mutational information is essential for the treatment of the patients, especially those with a diagnosis of advanced-stage non–small cell lung cancer. If the tumor cells in the paraffin blocks are exhausted, the patients might need to undergo repeated biopsies to get adequate specimens for molecular testing. Using the CTT we are able to use archived sections of previously cut H&E-stained slides to perform the molecular testing. This prevents patients from undergoing unnecessary additional biopsy procedures, which may have associated risk of complication in addition to superfluous costs and time. In our study, we showed a successful rate of 85% in performing molecular testing on the archived H&E-stained sections and there was 99% agreement on the results between CTT and conventional samples. The only discrepant case in our study was a lung biopsy from a patient with a diagnosis of adenocarcinoma with lepidic pattern. Retrospective review of the H&E-stained slide showed that the section was mostly composed of benign alveolar tissues and reactive pneumocytes with fewer than 200 tumor cells present, accounting for less than 10% of the nucleated cells. The small tumor volume was likely contributing to the false-negative result. Because we used 8 to 10 unstained slides to obtain tumor DNA during the conventional procedure, the original materials definitely contained a higher volume of tumor cells, and therefore we were able to detect KRAS G12C mutation with these samples. In our previous study of cytologic specimens, we also encountered 1 false-negative KRAS mutation case that was also due to low ratio (<10%) of tumor cellularity.12 The College of American Pathologists molecular testing guidelines suggest, in general, a minimal mutated allele frequency of 25% (50% cancer cell frequency, assuming heterozygosity and disomy) for Sanger sequencing. Polymerase chain reaction–based testing requires lower neoplastic cellularity; however, when the volume of tumor is less than 10%, a false-negative result is likely. If results are based on a less-than-adequate amount of cancer cells, any negative finding should warrant a comment for the possibility of a false-negative result.20
The purpose of CTT is similar to that of laser-capture microdissection. Both methods are able to harvest specific cells and separate them from unwanted cells to give pure enriched cell populations. The cost and complexity of laser-capture instruments are much higher than those of the CTT. Technically, the processing steps for CTT are not complex, and histology or cytology technologists can easily be trained to perform the procedure. Cell-transfer technique can be performed in any laboratory and no special equipment is necessary. The cost of the materials is minimal. After the selected areas have been cut and peeled off, the rest of the tissue section remaining on the slide can be re-coverslipped and kept on file.
The advantage of CTT over conventional scraping procedure is obtaining higher concentration and purity of tumor DNA through lifting the selected tumor areas by CTT. It is difficult to scrape off small areas of tumor on slides, especially when the tumor cells are scattered and intimately associated with normal tissues. The scraped material can also fly off the blade owing to electrostatic forces. Although the DNA yields provided by CTT are 5 times lower than those obtained with the conventional method, because of the higher concentration and purity of tumor DNA obtained by CTT, they are adequate for molecular studies in 85% of cases. However, the higher failure rate for CTT may be related to lower DNA yields. The small amounts of DNA obtained by 1 section through CTT may not be sufficient for assays requiring larger input of DNA (eg, next-generation sequencing), but the CTT could be scaled up to more than 1 section to obtain more DNA as long as diagnostic H&E sections are retained.
In this study, the strong correlation between molecular assays performed on H&E-stained slide material by using the CTT and the molecular assays performed on conventional, unstained, recut sections from FFPE tissue samples indicates that the CTT is a reliable alternative resource for assessing EGFR, BRAF, and KRAS mutations, especially when tissue from cell blocks and small surgical biopsy samples is exhausted and the only available material for testing is on H&E-stained slides.