No way to reliably identify low-risk prostate cancer

Friday April 11 2014

"Men with prostate cancer being given 'false hope'," The Daily Telegraph reports.

UK researchers have examined the accuracy of different methods that have sometimes been used (mostly outside the UK) to identify "clinically insignificant" prostate cancers not expected to affect a man during his lifetime, meaning he's likely to die of something else.

There has been considerable debate about overtreatment of such slower growing, low grade prostate cancers – not least because complications of treatment, such as erectile dysfunction, can be life changing.

Monitoring a man (known as "active surveillance") is often the preferred course of action with a low-risk cancer. But prostate cancer is a leading cause of cancer-related deaths in men – so doctors need to be sure the cancer really is low risk.

There have been a number of methods proposed for identifying clinically insignificant cancers. The researchers used these methods to see how reliable they would have been at accurately diagnosing the severity of cancer in a large series of men who had their prostate removed at one UK hospital.

None of the methods were accurate at predicting clinically insignificant prostate cancer. They only accurately identified up to half of men with clinically insignificant prostate cancer, which, as the researchers pointed out, is akin to trying to predict the toss of a coin.

The research highlights the uncertainty that exists. And as the authors suggest, when doctors discuss active surveillance with patients, they should explain the uncertainty around predicting the stage or grade of the cancer.

Where did the story come from?

The study was carried out by researchers from University of Cambridge and Cambridge University Hospital NHS Foundation Trust, and was funded by the National Institute for Health Research and the National Cancer Research ProMPT (Prostate Cancer: Mechanisms of Progression and Treatment) collaborative.

It was published in the peer reviewed British Journal of Cancer.

Many of the media headlines have given quite a simplistic interpretation of what's in fact quite a complex study. The researchers examined the accuracy of different sets of criteria, proposed by different research groups, which have sometimes been used to predict clinically insignificant disease.

Most have been developed in countries like the US, where prostate cancer screening is performed. But prostate cancer isn't screened for in the UK.

Some of the reporting implies that screening and diagnosing prostate cancer is a black and white issue. In fact, it's long been recognised that predicting which cases of prostate cancer will turn out to be aggressive is an inexact science.

What kind of research was this?

This was a case series of 847 men who had their prostate gland removed because of prostate cancer at a single hospital in England between July 2007 and October 2011.

The researchers then examined these tumour specimens in the lab to see how accurate different methods or criteria would be at identifying which prostate cancers were clinically insignificant. A clinically insignificant prostate cancer is one that wouldn't affect a man during his lifetime (meaning he's likely to die of something else), which isn't uncommon.

The researchers explain how there's often been debate around the over-treatment of such slower growing, low grade prostate cancers. And monitoring men (active surveillance, also known as "watchful waiting") would often be considered a more appropriate treatment option. This means that having a reliable method of identifying clinically insignificant prostate cancers is important.

Many attempts have been made to try to find a reliable method for identifying these cancers. Several different methods have been developed based on a combination of characteristics, such as examination findings, prostate-specific antigen (PSA) levels (a hormone associated with prostate enlargement), ultrasound examination, and examination of biopsy specimens.

These have mainly been developed in countries where prostate cancer screening is currently performed, such as the US. Prostate screening isn't currently performed in the UK.

The researchers wanted to see how accurate these methods were for identifying clinically insignificant prostate cancer in a group of men whose prostate gland had been removed as a result of prostate cancer.

What did the research involve?

The research included 847 men who had their prostate removed at Addenbrooke’s Hospital, Cambridge, between July 2007 and October 2011.

Their PSA level was measured and their prostate specimen was examined in the laboratory, and their cancer was staged according to the standard TNM (tumour, lymph nodes, metastases) staging system.

The Gleason score – another method of assessing the outlook of the cancer depending on what the cancer cells look like under a microscope – was also assessed. On a Gleason score, cells are graded between 1 and 7, with 1 and 2 being normal looking prostate cells, and 7 being the most abnormal looking cancerous cells.

Sometimes there can be more than one grade of cell within an examined prostate sample, so a doctor may give two scores indicating which two types of cell are most and second-most common in the specimen.

Postoperative assessments, including a physical examination and measurement of PSA levels, were carried out at six weeks and then 3, 9 and 12 months after the prostate was removed, and then every 6 months after that. "Biological recurrence" of the cancer was defined as a PSA level of greater than 0.2 nanograms per millilitre.

The researchers identified several different methods that have been used to identify clinically insignificant prostate cancer (described by different research groups and identified by the lead author of the study).

The accuracy of these different methods was compared against 3 different definitions of clinically insignificant disease:

  • classical definition: organ-confined tumours of <0.5 cm3, Gleason 3+3 and no Gleason 4 or 5
  • European Randomised Study of Screening for Prostate Cancer (ERSPC) definition: organ-confined tumour, Gleason 3+3 with no Gleason grade 4 or 5, index tumour volume 1.3cm3 or less, and the total tumour volume of 2.5 cm3 or less
  • inclusive definition: organ-confined tumour, Gleason 3+3 tumours, with no Gleason grade 4 or 5

The researchers also compared them with the accuracy of another method used to define low-risk disease (described by Anthony V. D’Amico and colleagues in 1998): PSA level of 10 or less, Gleason 3+3, and tumour stage 1 to 2a.

As the researchers say, the criteria described by D'Amico and colleagues wasn't intended to be used to identify which men would be suitable for active surveillance (those who had clinically insignificant disease), only to predict the outcome after removing the prostate.

In the UK the D'Amico method has been used as a way to predict likely outcome with different treatment approaches.

What were the basic results?

Of the 847 men, 415 (49%) had Gleason 3+3 disease indicated on their diagnostic biopsy. This indicated that the cells in their prostate were cancerous, but they were the "least abnormal" possible. Of these, 206 had what would meet the D'Amico's criteria for low-risk disease. 

But after surgical removal of the prostate and laboratory examination, half of them (209) were actually found to have more advanced disease than previously thought and upgraded to the more advanced Gleason grade 4 to 5.

A third of them (131) had cancer spread beyond the prostate, and one man had positive lymph nodes.

Of the 415 men with Gleason 3+3 at biopsy (a quarter of the full group), 206 met the D'Amico criteria for low-risk" prostate cancer.

None of the methods for predicting clinically insignificant cancer assessed were considered to have adequate discriminative power in predicting clinically insignificant tumours. The different methods correctly identified only up to half of those with clinically insignificant disease.

And none of the methods had significantly improved accuracy over D'Amico slow-risk criteria, which correctly identified between 4% and 47% of those with clinically insignificant cancer, depending on which of the 3 criteria were used.

How did the researchers interpret the results?

The researchers concluded that, "In our unscreened population [the men in this study had not been identified through screening as prostate screening is not performed in the UK], tools designed to identify insignificant prostate cancer are inaccurate."

Conclusion

This research looked at different methods that have sometimes been used to identify men with clinically insignificant prostate cancer not expected to affect a man during his lifetime. The researchers explain how there has often been debate around the overtreatment of such slower growing, low grade prostate cancers, and monitoring the man (active surveillance) would often be considered a good treatment option.

A number of methods have been proposed – most of these have been developed in countries where prostate cancer screening is carried out. The researchers found that in their series of 847 men, none of the various methods were accurate at predicting clinically insignificant disease. So they correctly identified around half the men with clinically insignificant disease.

Some of the methods with more inclusive criteria for potential disease risk would identify very few men as having clinically insignificant disease and so be eligible for watchful waiting. Meanwhile, the methods that had stricter criteria for selecting men at higher potential risk (for example, only those with larger tumours) could lead to a greater number of men being wrongly offered watchful waiting when they in fact need active treatment.

The research highlights the uncertainty doctors experience when trying to accurately identify which men diagnosed with prostate cancer (for example, through a combination of PSA, physical examination, imaging and biopsy) have a cancer unlikely to affect them in their lifetime, and so are suitable for a watchful waiting approach only.

As the researchers suggest, when doctors discuss the watchful waiting approach with patients, they should explain the uncertainty around predicting the stage or grade of the cancer.

The researchers appropriately concluded that, "There is an urgent need for development of a means by which to exclude aggressive prostate cancer in patients wishing to undergo conservative treatment."

Analysis by Bazian
Edited by NHS Choices