Overview of Molecular Diagnostics in Irish Clinical Oncology

Genetic indications

Cancer is, by nature, a disease of genetic origins¹⁷. Though there are many different genome sequence mutations associated with many cancer types, only a small subset of these are currently known to be clinically informative or actionable, typically by informing diagnosis, prognosis, and/or treatment options¹⁸. For example, the EGFR gene encodes a tyrosine kinase which, when activated, signals for increased DNA replication and general cell proliferation; as such, over-activation of EGFR is associated with a variety of cancer types, including non-small cell lung cancer (NSCLC), in which approximately 14% of European patients harbour an EGFR-activating mutation¹⁹. For these patients, tyrosine-kinase inhibitor (TKI) therapies specifically targeting EGFR (e.g., osimertinib, gefitinib) are more effective and are associated with more favourable outcomes compared to chemotherapy^20,21. In colorectal cancer patients, however, the presence of KRAS-activating mutations greatly reduces the efficacy of anti-EGFR TKI chemotherapies, as KRAS is a downstream activation target of the EGFR signalling pathway; once permanently activated through mutation, KRAS promotes tumour growth regardless of EGFR inhibition, and is associated with poorer outcomes²².

As in these examples, identifying genetic mutations can be critical in directing cancer treatment. Among all cancer treatment regimens developed by the NCCP, there are currently approximately 20 genetic factors (including 2 broader genetic phenotypes) informing 101 therapy indications across 72 chemotherapy regimens. These regimens involve combinations of 47 different genomics-informed drugs, 40 of which are approved for funding through either the PCRS or the ODMS for approved indications (Table 1)^{15,16,23–26}.

Techniques and technologies

Depending on clinical purpose and cost, testing for relevant genetic mutations in cancer occurs at several levels of scale. Small-scale single-gene tests can be used to identify known point mutations, such as EGFR T790M or KRAS G12C chemotherapy resistance mutations^28,29, or to identify known fusion genes, such as the BCR-ABL1 gene fusion found in chronic myelogenous leukaemia (CML)³⁰. These single-gene tests are generally performed using techniques such as quantitative polymerase chain reaction (qPCR) or fluorescent in-situ hybridization (FISH), and can also be performed on both Sanger sequencing and next-generation sequencing (NGS) platforms, though using high-throughput NGS with very small targets is generally not cost efficient without very large numbers of samples.

Multiple genes can be tested for mutations simultaneously by sequencing on an NGS instrument. These NGS assays range from small disease-focused gene panels, targeting tens to hundreds of genes; to whole exome sequencing (WES or WXS), targeting tens of thousands of genes; to whole genome sequencing (WGS), which generates data from both genic and non-genic regions. In all NGS applications, results can then be subset virtually to focus on disease-specific genes or regions of interest. While methods like qPCR or genotyping microarrays can be used to detect known mutations, genome sequencing does not require a priori knowledge of mutations of interest, thus allowing for discovery of novel relevant genomic variation in cancer³¹. While novel mutations are not likely to be clinically actionable upon discovery, they may have potential for use in research, trials, and treatment in the future. More comprehensive genomic sequencing additionally allows for more complex genomic profiling strategies which can further inform disease aetiology, progression, and prognosis.

In Ireland, qPCR and FISH single-gene tests, gene panels including ThermoFisher's Oncomine Focus panels and other ThermoFisher Ion AmpliSeq small gene panels, and clinical exome gene panels are all routinely used. While whole genome sequencing can be clinically useful, this is generally not performed in Ireland as routine care outside of clinical trials or research applications.

In addition to the genetic sequencing performed by Irish medical laboratories, patient samples are also sent to external sequencing facilities in cases requiring, for example, rapid turnaround time, Sanger sequencing variant confirmation, or specialty assay sequencing. Notably, homologous recombination deficiency (HRD) has recently been added as an NCCP indication for olaparib treatment of ovarian cancer. While largely determined by the presence of deleterious BRCA1/2 mutations, HRD is a wider genetic phenotype influenced by larger genomic factors such as loss-of-heterozygosity and rearrangement events. Similarly, high microsatellite instability (MSI-H), which was recently added as an NCCP indication for immune checkpoint inhibitors in colorectal cancer, requires profiling of multiple locations throughout the genome. Both HRD and MSI-H testing thus require larger or specialty NGS gene panels, such as the Myriad MyChoice HRD test, FoundationOne panel, and Illumina TSO500 panel, all of which are currently being considered for use in Ireland. These external tests are generally funded under hospital departmental budgets rather than being reimbursed directly by the HSE, though efforts are underway by several hospitals to develop the infrastructure required to perform more genetic tests domestically in public facilities.

Cellular biomarker-based indications

In addition to identifying mutated cancer-associated genes, confirming the presence of cellular biomarkers, which commonly include hormone receptors and antigens involved in immune cell recognition, can also be vital for accurate cancer diagnosis and treatment decisions. Lymphoma subtypes, for example, each exhibit characteristic immunophenotypes which can be essential for differential diagnosis of cancers that are otherwise morphologically similar^32,33.

Biomarkers expressed on the cell surface can also serve as key drug targets. Antibody-based therapies target only specific cell types exhibiting the target antigen, and thus can activate or inhibit cellular signalling pathways or elicit a patient immune response against target cells, while limiting the potential deleterious effects of cancer treatment. Antibody-drug conjugates, such as brentuximab vedotin, further exploit this specificity by directing otherwise highly toxic chemotherapy drugs only to cells exhibiting the target antigen³⁴.

Complementary to antibody-based therapies, small molecule drugs can also reach intracellular targets. For example, several treatment routes exist to reduce the growth-promoting effect of oestrogen on oestrogen-receptor-positive (ER+) breast tumours, including anastrozole, which binds aromatase enzymes to inhibit the production of oestrogen in the body; tamoxifen, which inhibits oestrogen binding by blocking oestrogen receptors; and fulvestrant, which binds and destabilises oestrogen receptors, inducing their breakdown by the cell³⁵.

Like genetic mutations, the presence or absence of cellular biomarkers can play a critical role in diagnosis, prognosis, and treatment of a patient. In Ireland, the presence or absence of 10 markers are a factor for 161 indications for 25 different therapies across 117 treatment regimens published by the NCCP, and are of particular importance for informing breast cancer and lymphoma treatments, which account for 70% of these indications. Of the 25 included therapies, 21 have funding through the ODMS and PCRS (Table 2)^{15,16,23–26}.

Techniques and technologies

The presence of cellular biomarkers can be determined either by direct detection, or by some indirect indication of their presence. Immunohistochemistry (IHC) techniques, which remain the gold standard for direct determination, use a combination of an antigen-specific antibody and a dye or fluorophore to indicate antigen presence in cancer tissue samples via microscopy³⁶. While this technique can generally only detect one antigen per assay, the process can be parallelized in appropriate tissue samples via flow cytometry, such that multiple antibodies can be applied, allowing several antigens to be detected on cancer cells simultaneously^37,38.

Indirect detection, instead, can be accomplished through gene expression analysis. Rather than detecting an antigen of interest via an antibody, this approach involves the quantification of RNA transcripts encoding the biomarker. Techniques for measuring expression analysis are similar to those for detecting DNA mutations and include reverse transcription qPCR (RT-qPCR), expression microarrays, and next-generation RNA sequencing (RNA-seq).

These methods also scale similarly to DNA mutation detection methods: qPCR is limited to measuring the expression of single genes, while microarrays and RNA-seq are able to simultaneously quantify thousands of transcripts. Of particular note is that RNA-seq, in addition to expression analysis, also allows for mutation detection by default. This includes more complex mutations such as fusion genes, which are frequently highly associated with cancer and can serve as drug targets for inhibitors such as ponatinib, which inhibits BCR-ABL1 fusion proteins found in CML³⁹, and larotrectinib, a novel tumour-agnostic Trk inhibitor that can be used for any cancer in which NTRK-family fusions are detected⁴⁰. While RNA expression assays are less commonly used in clinical practice, recent studies have shown comparable test results between RNA-seq and IHC⁴¹.

In Irish hospitals, cellular biomarker detection methods generally include single gene tests like IHC and RT-qPCR. In addition, the external testing service Oncotype DX^® is available in Ireland for breast cancer patients and uses RT-qPCR to measure the expression of 16 genes, including HER2 and both oestrogen and progesterone receptor genes^42–44. While RNA-seq remains uncommon, reimbursement for larotrectinib in Ireland notably requires submission of RNA-seq results⁴⁵.

Molecular diagnostics in cancer treatment regimens

NCCP cancer therapy regimens were accessed via the HSE NCCP National SACT Regimens website¹⁵. Information on therapy indications from each tumour group subpage (as well subpages for oral anti-cancer medicines and paediatric therapies) was collected by systematic HTML parsing of tabular elements using the Python package Beautiful Soup version 4.11.2 in Python 3.11.0¹⁴⁴. Raw therapy indication text was then further parsed in Python to harmonise descriptions and drug names, to combine duplicate indications by indication ID, to assign relevant disease based on website subpage and subheadings, and to group therapy indications by regimen ID. Where conflicts arose in merging duplicate indications by ID, manual harmonisation was performed by referring to the full text of the hyperlinked regimen document; where conflicts arose in the hyperlinked regimen documents, the latest revision was used as reference. After tabular export of all indications and associated information, final manual curation was performed to correct malformed entries and errors in the source material, again referring to the appropriate full-text regimen documents. For the Python parser tool created for this purpose, see Software Availability¹⁴⁵ and for the exported and manually curated data table, see also Software Availability²³.

Identification of indications informed by genetic diagnostics was performed through several rounds of key-word search and manual review through the short descriptions of each therapy indication. Key-words included terms associated with genetics and genomics such as gene, chromosome, and express; known cancer gene names; and the terms and symbols positive, negative, +, and -, as well as further keywords encountered during manual review. In ambiguous cases, including cases where a molecular diagnostic was listed for one indication of a regimen, but not for other similar indications for the same regimen, both the full-text regimen document as well as published literature on the therapy in question were consulted. Note that while many regimens include CD20 antibody therapies for lymphoma, these were only included when a molecular diagnostic was explicitly referenced.

Reimbursement information was obtained from NCCP indications and regimens¹⁵, the NCCP table of approved drugs²⁴, the PCRS list of reimbursable items²⁵, and the HSE list of the High Tech Drug Arrangements²⁶.

The regimen information in this article reflects the NCCP SACT Regimens website as of 2025-May-27.

Predicted rates of actionable cancer molecular diagnoses

Cancer incidence rates in Ireland were obtained from the NCRI publication Cancer in Ireland 1994–2020: Annual Statistical Report of the National Cancer Registry, Appendix I: Incident Cancer Cases⁴⁹, except where indicated. Case numbers in this publication are listed as the 3-year average incidence from 2018–2020 of each ICD-10 invasive cancer group.

For each unique molecular diagnostic for each cancer subtype, incidence of the cancer subtype relative to the broader cancer type (e.g., proportion of lung cancers that are NSCLC) was obtained from literature where appropriate. Incidence rates of each molecular diagnostic within the relevant cancer subtype (e.g., proportion of NSCLC that is ALK+) were then also obtained from literature (references provided in Table 3 and Table 4). These rates were then applied to incident cancer rates in Ireland to estimate the positivity rate of each molecular diagnostic in Ireland.

In the case of acute lymphoblastic leukaemia, B-ALL subtype incidence was estimated separately for paediatric and adult cases due to differences in B- vs T-ALL rates in adults and children and the high proportion of childhood cases⁹⁷. Similarly, separate molecular subtype rates were applied for BCR-ABL1 fusions in adult and childhood B-ALL for the same reason⁹⁸. Incidence of metastatic castration-resistant prostate cancer was calculated as a function of total population based on the model referenced, producing numbers in agreement with NCRI case counts^81,82. Rates of urothelial carcinoma were applied separately for primary urethral urothelial carcinoma due to lower published rates of urothelial histology^84–86.

Clinical Sequencing Capacity in Ireland

Technical specifications on the machine runtime and DNA throughput for the Illumina NextSeq and ThermoFisher Ion Torrent platforms were obtained from their respective manufacturer websites. To calculate nominal maximum throughput, the highest throughput configuration of each machine was used (NextSeq 550 High-Output = 100–120 Gb of 120 bp paired-end reads/29 hrs¹³³, Ion GeneStudio with Ion 550 Chip = 40–50 Gb of 200 bp paired-end reads per 12 hrs¹³⁴).

DNA sequencing target size was based on paired-end sequencing with a 120x coverage target using the Agilent SureSelect Clinical Research Exome V4 (total design size=51.0 Mb), for a targeted total of 12.24 Gb of genetic material sequenced per sample¹⁴⁶.

Maximum capacity was then calculated to be the total number of exomes able to be sequenced by all 8 machines running at maximum capacity. The lower end of this range represents one run of each instrument using the lower bound of the instruments' stated throughput (one 29-hour run of the NextSeq at 100 Gb = 8 exomes per machine = 32 exomes, plus one 12-hour run of the Ion Torrent at 40 Gb = 3 exomes per machine = 12 exomes, totalling 44 exomes), while the higher end of the range represents the higher bound of the instruments' stated throughput, with two 12-hour runs of the Ion Torrent within the same time frame as one 29-hour NextSeq run (one 29-hour run of the NextSeq at 120 Gb = 9 exomes per machine = 36 exomes, plus two 12-hour runs of the Ion Torrent at 50 Gb = 8 exomes twice per machine = 32 exomes, totalling 68 exomes).

Molecular indications for clinical trials

Information on clinical trials was obtained from Cancer Trials Ireland in November 2023¹³⁵. Trials were considered to have a molecular component if the trial eligibility criteria included genetic mutations, aberrant genetic pathways, gene expression, cellular biomarkers, or microsatellite instability status as inclusion or exclusion criteria, or if the trial's purpose was to otherwise collect or analyse genomic data. Trials were evaluated systematically, beginning by prioritising those with explicit mention of these criteria in their short description. Trials without explicit reference to one of the two criteria, but which referenced a disease or treatment known to have a strong or common molecular diagnostic component were also prioritised. Short-listed trials' full trial descriptions were then checked to confirm the nature of the trial. After confirmation of short-listed trials, remaining trial full descriptions were then checked to confirm absence of a molecular diagnostic component.