Factors influencing cancer genetic somatic mutation test ordering by cancer physician

Background Clinical whole exome sequencing was introduced in an Australian centre in 2017, as an alternative to Sanger sequencing. We aimed to identify predictors of cancer physicians’ somatic mutation test ordering behaviour. Methods A validated instrument assessed somatic mutation test ordering, genomic confidence, perceived utility of tumour molecular profiling, and percent of patients eligible for targeted therapy. A cash incentive was included in 189/244 questionnaires which were mailed to all Queensland cancer specialists in November 2018. Results 110 participated (response rate 45%); 54.7% oncologists, and the remainder were surgeons, haematologists and pulmonologists. Oncologists were more likely to respond (p = 0.008), and cash incentive improved the response rate (p < 0.001). 67/102 (65.7%) of physicians ordered ≥ 5 somatic mutation tests annually. Oncologists saw 86.75 unique patients monthly and ordered 2.33 somatic mutation tests (2.2%). An average of 51/110 (46.1%) reported having little/no genomic confidence. Logistic regression identified two significant predictors of somatic mutation test ordering: being an oncologist (OR 3.557, CI 1.338–9.456; p = 0.011) and having greater confidence in interpreting somatic results (OR 5.926, CI 2.230–15.74; p < 0.0001). Conclusions Consistent with previous studies, the majority of cancer physicians ordered somatic mutation tests. However, the percentage of patients on whom tests were ordered was low. Almost half respondents reported low genomic confidence. Somatic mutation test ordering was higher amongst oncologists and those with increased confidence in interpreting somatic variants. It is unclear whether genomically confident individuals ordered more tests or whether ordering more tests increased genomic confidence. Educational interventions could improve confidence and enhance test ordering behaviour.


Patients should only be offered their genomic sequence results if evidence
demonstrates that actions based on the results can change patient management decisions and improve net health outcomes.
2. Patients should be offered genomic sequence results for which there is an established relationship between genotype and phenotype (e.g., results can be used to diagnose a disorder or to assess risk for a disease), even if the results do not alter management decisions or improve net health outcomes.
3. Patients should be offered as many of their genomic sequence results as they want, up to and including their raw genomic sequence data.
The questions that follow contain scenarios describing a particular type of genomic alteration derived from sequencing the patient's tumour DNA. Assume: • the tumour DNA belongs to YOUR adult patient with a metastatic solid tumour • the sequencing was performed in a NATA certified lab • the patient is currently receiving a first-line standard chemotherapy regimen • the patient is active and capable of self-care • the patient has indicated that s/he would like to be told about all clinically valid genomic results Please check the box that reflects how likely you would be to disclose the information described in each scenario to your patient.
Please read each scenario carefully.

Sequencing of tumour DNA identifies a somatic alteration that…
In this situation, I would…

Definitely disclose
Probably disclose Probably not disclose Definitely not disclose Unsure 4. Is in a pathway that is not targeted by any approved agent. However, an agent that targets this pathway is currently being studied in a phase II clinical trial that's open at your institution. Your patient may be eligible for this trial. 5. Is in a pathway that is targeted by a commercially available agent that is approved for a different cancer. There are no reports in the literature of agents that target this pathway being used in your patient's type of cancer. 6. Is known to confer a favourable prognosis, compared with the average for patients with this condition. There are no available agents, either commercially or through a clinical trial, that target the relevant pathway. 7. Is known to confer an unfavourable prognosis, compared with the average for patients with this condition. There are no available agents, either commercially or through a clinical trial, that target the relevant pathway. 10. Ability to explain somatic genomic concepts to patients.
11. Ability to make treatment recommendations based on somatic genomic information.
12. Ability to identify consultants who have special expertise in integrating somatic genomic information into patients' care.
13. Ability to provide psychosocial support related to coping with a somatic alteration that has adverse prognostic implications.

____________ number
Do you have any additional thoughts that you wish to share about the issues raised in this survey? Please feel free to write in the space below as we welcome your feedback.

Other
Thank you very much for completing this survey! Your participation is greatly appreciated. There is also a general lack of knowledge around costs -to the patient and to the system Some mutations have high reliability in assessment. Whole exome sequencing and many mutations lack testing for in the laboratory, reliability, and therefore their use in clinical practice should be considered unproven until reliability studies are undertaken.
I suspect in the near future -genomics probably will become part and parcel of management of patients with a new diagnosis of cancer Major issue with sequencing is the lack of infrastructure to counsel the patients. Testing is the smallest component. Until this is in place, you will not see wide roll out.
As a respiratory physician, the main priorities have been -1. consenting the patient for WES prebiopsy 2. timely result notification There is a need for increased and ongoing education of doctors in this aspect of medicine, which is a really exciting area which has already led to far-reaching improvements in cancer therapy.
These tests are being done without thought as to how to appropriately incorporate them into solid tumours. It is currently a wasted resource.
Emerging technology and application; ongoing education is very important.
Currently the way this service is run patients are not selected carefully enough for genomic testing. In the vast majority of cases, it is unhelpful and not a good use of resources. Also, there is no established, smooth way of getting the results to the clinicians who can interpret it. The system needs to be improved.
Much depends on who is funding testing -not appropriate for public fees to be spent on unproven technologies, but I don't have an issue with patients funding tests.
While not a genetics expert, it would greatly assist close consultation and genetic pathologist in our cancer diagnostics -either privately or publicly Be good to get review/input; we have a genetics counsellor working with us.
There is a lack of genetic counselling/advice available in my area of far North Queensland. Would like to see more of it. QH service overwhelmed by need.
There is emerging data on the non-coding genome such as the promoter and effector regions which are making the science more complicated but also more fascinating in terms of gene experience/effect. So, I do not feel that at a clinic level we are ready to use most of the information practically.
I don't request many WGS studies i.e. foundation one etc.

All answers pertain to the diagnosed cancer related findings not incidental findings potentially related to other disorders
Approaching retirement so not representative of a busy oncology practice.
Need more clinical relevance to mutation; need more trial options Heavily rely on medical oncology colleagues to order and interpret somatic genomic tests on our breast cancer patients.
Difficult keeping up with rapid development/change in this field

Somatic Mutations Summary
There was 1 reportable variant found in this sample.

Somatic CNV Summary
There were 3 reportable CNVs found in this sample. Description PIK3CA mutations are observed in many cancers among the most prevalent are breast cancer, colorectal cancer, and head and neck squamous cell carcinoma. These mutations have been observed in up to 40%, 20%, and 21% respectively for each of these cancer types with mutations commonly occurring within the helical exon 9 and kinase domain of exon 20. In breast cancer PIK3CA mutations have been associated with better prognosis, and a greater survival outcome, however the PI3K pathway is also associated with increased drug resistance (trastuzumab /lapatinib). PIK3CA mutations in both HNSCC and CRC have been associated with resistance to anti-EGFR therapeutic agents and mutations in exon 9 and 20 are associated with a poorer prognosis in CRC. Interestingly PIK3CA mutations in HNSCC have been observed to be absent from a number of ethnic groups including greek, german, and vietnamese populations.

Additional Comments
There are no additional comments for this report

Methodology
DNA libraries were prepared from the received samples and target regions captured using the IDT exome capture kit plus IDT custom spike in. Massively parallel sequencing was performed using the Illumina NextSeq 500 platform to an average depth of 507X for this sample. Data were aligned to the GRCh37 Human Genome reference sequence using the ATGC Bioinformatics Pipeline version and annotated according to the ATGC Clinical Variant Database version.
The assay reports somatic single nucleotide variants (SNV) and small (<100bp) insertions and deletions (INDELs) with a variant allele frequency >3% and the frequency is greater that the percentage of contamination detected plus 1.5%. Copy number variants (CNVs) spanning partial or whole genes are reported when amplification exceeds 5 copies or more and deletions are reported when the tumour purity (proportion of the sample that contains the somatic variant) is estimated to be greater than 60%. Previous validation assays found that the sensitivity of >99.5% for SNVs, >95% for INDELs with VAF >5% and sensitivity to reportable CNVs is >75% for tumour purities >15%. For higher tumour purities where VAF >25%, the sensitivity rises to >99.5% for SNVs and >95% for INDELs. For reportable CNVs sensitivity is >85% when tumour purity is >75% and >60% when tumour purity is >60%. The total mutational burden is a count of SNVs and INDELs within the IDT custom spike in regions predicted to cause protein coding alterations divided by the total number of bases tested (in Mbp).
The following genes were selected for filtering based on documented association with this cancer/tissue type: ERBB2, FGF3, FGFR1, FGFR2, NCOA3, PIK3CA, PTEN, RSF1, TP53 These genes appear in the ATGC Clinical Variant database and either the genes and/or the specific variants are associated with A or B level evidence.
Level A Validated association: Proven/consensus association in human medicine. Level B Clinical evidence: Clinical trial or other primary patient data supports association.
In some instances, all somatic protein coding variants within a specific gene are deemed reportable. In order to ascertain whether a variant affects protein coding the Variant Effect Predictor categorises the variants using RefSeq transcripts and results are reported using HGVS nomenclature (https://varnomen.hgvs.org).
Rare errors in annotation can arise due to the presence of genomic transcripts not described in RefSeq. Alternatively, multiple genomic variations may combine to produce undetected protein coding changes. This could generate false positive and false negative results. The assay is not designed to ascertain whether the identified variants alter expression in the tissue sample.

Disclaimer
The methods used to produce this report have been validated by Australian Translational Genomics Centre (ATGC) at Queensland University of Technology. Nonetheless, there is a chance that false positive and false negative results may occur as Pathology Queensland and ATGC are not in control of the tissue collection. Tissue collection factors that may give rise to false positive or false negative results include tissue heterogeneity, insufficient tissue quality, or contamination. A list of methodology including the bioinformatics pipeline utilised to draft the report are noted above. Diseases are influenced by many factors, including epigenetic and environmental variables that may not be addressed by this report, and, as such, the report should not be interpreted in isolation. Any diagnosis, or prognosis, should consider all pertinent clinical information in addition to this report. The bioinformatics pipeline used to identify the variants reported has also been validated by ATGC. The analysis limitations are outlined in the methodology section of this report.
The clinical significance of many variants is not well understood and interpretation of variants may change over time.
Interpretation of variants in this report is performed to the best knowledge of the laboratory based on the information available at the time of reporting. Re-analysis of variants in previously issued reports in light of new evidence is not routinely performed, but may be available upon request Samples received all met initial input quality requirements, and have undergone whole exome sequencing and somatic mutations have been identified to assist with the genetic profiling of the tumour (see Methodology section for further details). No second, independent, sample was received. For further information, or questions, about this report, please contact ATGC directly using the contact details at the bottom of this page.

Somatic Mutations Summary
There were 2 reportable variants found in this sample.

Somatic CNV Summary
There were 10 reportable CNVs found in this sample. Description PIK3CA mutations are observed in many cancers among the most prevalent are breast cancer, colorectal cancer, and head and neck squamous cell carcinoma. These mutations have been observed in up to 40%, 20%, and 21% respectively for each of these cancer types with mutations commonly occurring within the helical exon 9 and kinase domain of exon 20. In breast cancer PIK3CA mutations have been associated with better prognosis, and a greater survival outcome, however the PI3K pathway is also associated with increased drug resistance (trastuzumab /lapatinib). PIK3CA mutations in both HNSCC and CRC have been associated with resistance to anti-EGFR therapeutic agents and mutations in exon 9 and 20 are associated with a poorer prognosis in CRC. Interestingly PIK3CA mutations in HNSCC have been observed to be absent from a number of ethnic groups including greek, german, and vietnamese populations.