expert reaction to study investigating use of genetic risk score for prostate cancer in men with lower urinary tract symptoms to predict diagnosis -…

August 18, 2022

A study published in the British Journal of Cancer looks at the use of a genetic risk score to predict prostate cancer diagnosis in men with lower urinary tract symptoms.

Dr Chantal Babb de Villiers, PHG Foundation, said:

This paper looked at if genomic information can predict which men would develop prostate cancer after presenting at a GP with lower urinary tract symptoms. The study researchers found that, using age as well as a genetic risk score, they could stratify which men were most, and least, likely to develop prostate cancer after having symptoms. There are limitations to the study, notably that only men of European ancestry were included, which is a major limitation as black men are twice as likely to be diagnosed with prostate cancer, and have worse outcomes. The findings support the idea that accurate risk prediction using a genetic risk score along with other risk factors could improve symptomatic patient triaging in primary care, to determine which men could have further investigationto find out their possibility of prostate cancer. The PHG Foundation is following developments in these areas carefully to understand their implications for healthcare policy and delivery.

Dr James Ware, Cardiologist, Reader in Genomic Medicine at Imperial College London and MRC Investigator, MRC London Institute of Medical Sciences, said:

Overall I think this is an interesting study. I think the press release reflects the science in most important respects. I have reservations on one element: the press release emphasises that genetic risk scores (GRS) could identify people at highest risk, for fast-track referral. In fact in the paper they emphasise that the principle utility is in identifying low risk individuals who can avoid referral. The press release says that Only one in three men with a positive PSA test have cancer, but from this perspective the GRS arguably performs even worse figure 2 and table 2 shows that about 1 in 12 men with symptoms (8%) with a high genetic risk develop cancer in the next 2 years. Perhaps this difference is not hugely important for a lay audience, since there is potential utility either way, but I would emphasise that the real value is likely to be through reducing unnecessary invasive tests in low risk individuals. It does not have a high precision (positive predictive value) in this context. (Here the evaluation was done in men with symptoms, with an overall cancer risk of 3.7% in 2 years)

I did not identify any particular methodological problems. The authors acknowledge important limitations. An important limitation (in common with much similar work) is that the work was done in individuals with European ancestry, and utility is likely to be limited to this group.

My expertise is in genomic medicine more broadly I am not particularly expert in prostate cancer. The authors cite a previous high profile study in Nature Genetics (Conti et al, 2021) which showed how a GRS could be predictive of prostate cancer in the UK biobank population. So some of what is reported here is already known. The main advance here seems to be in assessing risk stratification in symptomatic men which simulates a real-world clinical problem, and the authors present specific data on utility in context

A Genetic Risk Score (GRS) that identifies low risk individuals who can avoid an invasive biopsy would likely be of clinical value.

At the moment the genetic data needed to calculate a GRS is not generally available in patients health records. If a sample needs to be sent for genetic testing then this would take some time so in the short term you would need to take this into account when designing a clinical referral pathway for patients with symptoms suggestive of possible prostate disease, and factor in the additional cost of this test.

Again I highlight that I am not a cancer expert specifically, but I note that prostate cancer represents a wide spectrum of disease, some of which is not clinically significant. The prior work from Conti et al. found that the GRS identified prostate cancer, but did not discriminate between aggressive and non-aggressive disease. Other studies have also found that GRS do not select for clinically-significant disease. So while the test may help to identify individuals with cancer, other approaches may be needed to discriminate which of the men with prostate cancer need active treatment.

Prof Shirley Hodgson, Emeritus Professor of Cancer Genetics, St Georges, University of London (SGUL), said:

This is a very interesting study of a cohort of men who attended their general practice because of urinary symptoms. The researchers used genetic profiling with 260 common genetic variants known to influence prostate cancer risk, and other data including self-reported family history of prostate cancer, BMI and smoking status. Nearly 5,000 men with symptoms that could indicate prostate cancer were enrolled in the study, and 400 were diagnosed with prostate cancer within 2 years of enrolment. The study outcome was diagnosis of prostate cancer within 2 years of the index date of interview.

The genetic risk score (GRS) appeared to be very useful for predicting the diagnosis of cancer, where the highest risk score quintile was highly predictive of cancer with an incidence rate of 8.1%, and the lowest score quintile with a risk of less than 1%.

This pilot study suggests that if this was implemented in general practice, it could mean that men with the highest risk scores could be targeted for increased screening by referral to secondary care, whereas those in the lowest quintile could be managed in primary care, thus saving anxiety and costs.

The study is limited by being biased to younger men and mostly white individuals, meaning that the outcomes could be different in the groups who were not included.

Prof Dusko Ilic, Professor of Stem Cell Science, Kings College London, said:

This is an interesting research paper. However, a much simpler and significantly cheaper blood test for prostate-specific antigen (PSA), that is already in place, has more value as a biomarker in diagnosis of prostate cancer. When PSA is moderately increased, it may not be easy to differentiate benign prostatic hyperplasia from cancer and in such cases, a genetic risk score might have potential application to improve diagnostic the pathway.

Applying a genetic risk score for prostate cancer to men with lower urinary tract symptoms in primary care to predict prostate cancer diagnosis, as implicated in the paper, would be an unnecessary waste of the NHS funds.

Applying a genetic risk score for prostate cancer to men with lower urinary tract symptoms in primary care to predict prostate cancer diagnosis: a cohort study in the UK Biobank by Harry D. Green et al. was published in British Journal of Cancer at 01:00 UK time on Thursday 18 August.

Declared interests

Dr James Ware: None directly relevant to this work. I have acted as a consultant for and/or received research funding from Pfizer, MyoKardia, Bristol-Myers Squibb, and Foresite Labs, related to work on genetic disease and/or genome-based risk stratification.

Prof Shirley Hodgson: No conflicts of interest.

Prof Dusko Ilic: I declare no conflict of interest.

For all other experts, no reply to our request for DOIs was received.

Follow this link:
expert reaction to study investigating use of genetic risk score for prostate cancer in men with lower urinary tract symptoms to predict diagnosis -...