Science

AddToAny

Google+ Facebook Twitter Twitter

Delivering real-time PCR in histopathology

Clinical Scientist and Molecular Pathology Lead Siobhan Taylor looks at the experiences of implementing a new platform within a routine histopathology service and the impact it has had on patient care.

In October 2018, the histopathology laboratory at Gloucestershire Hospitals NHS Foundation Trust (GHNHSFT) introduced a new platform for rapid molecular testing of formalin-fixed paraffin-embedded (FFPE) tissue. Following a period of in-house verification, we have introduced testing in melanoma (BRAF mutation) and colorectal cancer (KRAS, NRAS, BRAF mutations and microsatellite instability (MSI) testing). Implementation of this benchtop platform has enabled us to reduce turnaround times for delivering molecular pathology results, with a beneficial impact on patient care.  

Histopathology and molecular pathology

The histopathology department at GHNHSFT is in a large district general hospital serving a population of ~900,000, processing more than 90,000 samples per year. Complex molecular testing of patient tumour samples, to aid in diagnosis and subclassification, prediction of response to therapies and identification of therapeutic targets, is traditionally performed in specialist genetic laboratories.  

Molecular testing of samples received by our laboratory is largely outsourced, which can lead to delays in patient treatment. Turnaround times are extended due to the time taken to ship the samples and reliance on an external laboratory’s workflow for testing and reporting.

Verification

There are a number of different assays (CE-IVD and RUO) available to use on the platform for the detection of a number of mutations in both FFPE and liquid biopsies in different tumour types. In our laboratory, we have explored the use of FFPE to date.

The assay cartridges are extensively validated by the manufacturer, and discussion with other users suggested variability in what individual laboratories had undertaken for their own verification. For our quality control (QC) purposes, we verified the individual assays separately, using both externally sourced controls and a selection of at least 10 historic clinical cases with a known result.

These included specimens from a range of sample sizes (small biopsies vs large resections and whole section vs macro dissection) and percentage neoplastic cell content (including samples that were at the limit but not below the manufacturer recommendations) to test the platform capabilities and prove accuracy.

We began with verification of the BRAF mutation test in melanoma samples, using a retrospective approach, and were able to “go live” with a service fairly rapidly. We verified the KRAS mutation test in colorectal cancer samples in a similar manner, but the NRAS/BRAF mutation assay proved more challenging, due to a lack of previously tested colorectal samples for BRAF mutation. Therefore, we used a combined retrospective and prospective approach, testing samples in-house and sending them away for external verification to assure the result, which slightly delayed implementation of a full service for colorectal cancer.

We have also verified the MSI assay, comparing previously obtained mismatch repair (MMR) immunohistochemistry (IHC) results for the colorectal cancer specimens tested. Table 1 summarises the verification performed for each cartridge.

Neoplastic cell content

All specimens for molecular testing must first have their neoplastic cell content assessed. This should be performed on a section taken immediately prior to that which will be tested. At the time of selecting suitable cases for verification, a fresh H&E was taken on all potential cases to allow estimation of the percentage neoplastic cells present in the sample, assessment of tissue suitability (specimen size, necrosis, remaining tissue) and marking of the area containing neoplastic cells. To ensure a systemic approach, the pathologists involved undertook the tumour assessment e-learning programme by the Genomics Education Programme.  

During the verifications, a number of samples were disregarded as they were assessed to contain insufficient material. We remain unable to test samples that do not meet the minimum requirements for the indicated assays, most often due to insufficient available material. In these instances, we continue to outsource these samples for alternative testing in our specialised referral laboratory. However, in the future we have the option to validate the use of smaller samples on the platform.  

Sample preparation

The platform is a closed system and as such, clean and accurate working is imperative to ensure the integrity of the result. Misidentification and/or contamination of the sample at any point has the potential to lead to the incorrect assignment of a result, which could ultimately lead to a patient receiving the incorrect treatment. During the pre-analytical phase, we therefore consider it imperative that stringent measures are in place to reduce the possibility of sample mix-ups or cross-contamination. These precautions apply to the preparation of all material for molecular techniques, whether performed in-house or externally. Introduction of this new technique in our laboratory has raised awareness as to how specimens referred for molecular technique should be handled.

We have found it useful to have a dedicated microtome in a “clean” area, away from disruption and potential contaminants. We utilise clean microtomy techniques when preparing samples for all molecular pathology requests, which requires cleaning of all equipment, tools and surrounding work bench with 70% IMS prior to use and in between sample blocks. Clean gloves and a clean laboratory coat must also be worn, and a fresh section of blade must be utilised for each new sample.

When taking sections, we adopt a one-block-only policy at the microtome to ensure the correct patient material is sampled. DNAase-free tubes are labelled with specimen details to allow clean transport from the microtome to the Idylla workspace. In our laboratory, we have generated new request sheets to encompass this work, which allow us to document and record each step and ensure the relevant QC checks are performed and countersigned by a second member of staff.  

During the verification period, we used a cleanly prepared water bath for the preparation of all samples, with no apparent issues. However, on one occasion since, we have experienced a problematic result, which could not be resolved by external testing.

A thorough examination of our processes highlighted the water bath and brush as a potential source of contamination and we have since ceased this practice when preparing material for the platform. We are now able to macrodissect sections either at the microtome, or on a clean glass slide.  

Loading and running

The footprint of the platform is such that it requires only a small clean space to work, which has been key to enabling us to fit it into our limited laboratory workspace. The loading of the sample into the assay cartridge is the same for all FFPE assays, involving the formation of filter-paper “sandwich” to contain and load the sample, using small aliquots of nuclease-free water. This step takes a couple of minutes, while use of our newly-
generated worksheet allows for second checks at each stage to ensure traceability.

“Depending on the assay performed, a result is available between 90 and 150 minutes”

Results

Depending on the assay performed, a result is available between 90 and 150 minutes. The system automatically interprets the results, generating a console result report, which can either be downloaded onto a memory stick, or downloaded from the platform’s software.  

In our experience, the generated reports are easy to interpret; in addition to containing all relevant QC information, the report for each sample tested includes a clear genotype result, which mutation has been detected, the protein HGVS and the base change. This information is clearly displayed in the report to enable replication to our own pathology report.

The software allows visualisation and some limited interrogation of the PCR curves generated from the test. Following on from our verification period, and as we have gained more experience and confidence with the platform and software, we now also perform a detailed check of the PCR curves. In checking the wild-type total Cq value of the assay as an indication of the overall quality of the sample and, for mutation positive samples, the ∆Cq value, we aim to ensure added confidence in the results obtained. This has involved extensive support from the platform supplier team, who have been incredibly patient and informative.

In line with any molecular technique, the manufacturer states that fixation of samples should be limited to 24 hours, after which degradation of DNA can occur, affecting result outcome. In practice, this is hard to restrict in a busy histopathology laboratory, and our laboratory information system does not allow for accurate monitoring of fixation. However, we have not yet experienced any problems with DNA degradation/integrity of our own clinical samples. Indeed, other studies have shown the platform performs better on samples of poor quality where other techniques might fail.

Reporting

Reports generated for molecular pathology testing are complex and require the inclusion of specific detailed information, but follow a standardised approach with a similar principle to histopathology reports that utilise the Royal College of Pathologists datasets.  

My discussion with other users of the platform indicated no consensus for who should/could report results generated from this testing. Some laboratories assume a traditional histology workflow, with consultant pathologists reporting. Others adopt an health care scientist (HCS) reporting workflow strategy. As a department, we currently adopt a dual HCS/pathologist reporting strategy, while we familiarise ourselves with the technique; an HCS compilesmand oversees the report and a pathologist checks and authorises it. In the near future, following appropriate training, we will be extending the neoplastic cell content assessment and reporting/authorisation processes to a scientist role only, which will reduce the impact on the consultant pathologist and increase scientist scope of practice.

External quality assessment (EQA)

Participation in a national EQA scheme is recommended for ISO 15189 accreditation purposes. We have chosen to register with the relevant Genomic Quality Assessment schemes for the assays we are currently running, which assess both the result produced and the report generated. We have also established an internal QC timetable.

Turnaround times

The introduction of in-house molecular testing for clinically relevant mutations in melanoma and colorectal cancer has positively impacted turnaround times for many patients in our region, ensuring results are available within a few days, compared to one to two weeks. We have received direct feedback from our clinical teams that patient anxiety is reduced with the availability of all relevant results prior to starting treatment. We also have evidence of situations where particularly poorly patients have benefited from earlier access to treatment with the availability of a rapid result. Such urgent testing of samples, when necessary, is not possible when samples are outsourced.

In-house testing also ensures that patient material remains in the laboratory and is easier to locate should further testing be required and, where limited material is available, it is possible to make an informed decision on how it is best used.

Conclusions

The implementation of the platform within our trust has allowed us to further strengthen the delivery of personalised medicine in a timely and effective manner to the patients in our region, directly impacting patient care.  

It has proven easy to use and generates rapid results. We have achieved excellent concordance with known results during verification and continue to be very pleased with the results generated by our service. The introduction of this testing within our laboratory is increasing staff training and knowledge and expanding the roles of scientists within a histopathology laboratory. The scientist team within the laboratory has been able to collaborate with wider teams across our trust, improving inter-departmental relationships and delivering cost savings.

We have implemented a new technique within our laboratory, which has proved challenging at times, however, the application of strict pre-analytical procedures has ensured confidence in our results. While we have not experienced any issues with the DNA quality of our samples, we are learning to recognise quality results and, where infrequently we have experienced cartridge problems, we have experienced excellent customer support from the product supplier teams.

The current restructuring of genomic services into seven genomic laboratory hubs may see a time restriction on our use of this platform. Whatever the future for this testing, it is without doubt that the knowledge and expertise held within histology laboratories will remain crucial to molecular pathology techniques.  

Siobhan Taylor is a Clinical Scientist and Molecular Pathology Lead at Gloucestershire Hospitals NHS Foundation Trust. She would like to thank colleagues at GHNHSFT in the laboratory and the wider clinical teams for their help and support. To see the article with references, visit the biomedicalscientist.net

Related Articles

IBMS podcast relaunched

The IBMS Biopod has been revamped, renamed and relaunched as IBMSpod.

Reporting in histopathology: some personal insights

The team at Leicester Royal Infirmary have collaborated on this article, which discusses their experiences of scientist reporting in histopathology.

"Don't delay cancer treatment"

A European Society for Medical Oncology (ESMO) interdisciplinary expert consensus paper on how to manage cancer patients during the COVID-19 pandemic has been published.

"AI research could pose risk for patients"

Many studies claiming that artificial intelligence (AI) is as good as, or better than, human experts at interpreting medical images are of poor quality and potentially exaggerated.

Top