An Association Between Lower Facility Level PSA Screening Rates and Higher Metastatic Prostate Cancer Incidence Rates, Journal Club - Christopher Wallis & Zachary Klaassen

November 10, 2022

In this UroToday Journal Club presentation, Christopher Wallis and Zachary Klaassen highlight a JAMA Oncology publication led by Alex Bryant and Brent Rose titled The Association of PSA Screening Rates with Subsequent Metastatic Prostate Cancer Incidents at US Veterans Health Administration Facilities. PSA Screening is relatively controversial. Adopted in the United States in the 1990s it was associated with rapid increases in the diagnosis of non-metastatic prostate cancer. Controversies arose in the following years as the benefits in terms of cancer-related morbidity and mortality needed to be weighed against harms including overdiagnosis and overtreatment. It remains unclear whether the variation in PSA screening is associated with subsequent metastatic prostate cancer incidents.

In this publication being highlighted here, the authors address this knowledge gap using data from the US Veteran Health Administration data set to assess variation in facility level-PSA screening, prostate biopsy utilization, and the incidence of non-metastatic prostate cancer as well as metastatic prostate cancer during 2005 and 2019. They sought to investigate whether lower facility level-PSA screening earlier in the interval would be associated with higher rates of metastatic prostate cancer subsequently.

Biographies:

Christopher J.D. Wallis, MD, Ph.D., Assistant Professor in the Division of Urology at the University of Toronto.

Zachary Klaassen, MD, MSc, Urologic Oncologist, Assistant Professor Surgery/Urology at the Medical College of Georgia at Augusta University, Georgia Cancer Center


Read the Full Video Transcript

Christopher Wallis: Hello, and thank you for joining us for this UroToday Journal Club discussion. Today, we're talking about a recent publication entitled "The Association of PSA Screening Rates with Subsequent Metastatic Prostate Cancer Incidents at US Veterans Health Administration Facilities." I'm Chris Wallis, an assistant professor in the Division of Urology at the University of Toronto. With me today is Zach Klaassen, an assistant professor in the Division of Urology at the Medical College of Georgia. You can see highlighted here this recent citation in JAMA Oncology led by Dr. Bryant and Dr. Rose.

PSA Screening, as most will know, is relatively controversial. It began to be adopted in the 1990s in the United States and was associated with rapid increases in the diagnosis of non-metastatic prostate cancer. However, controversies arose in the coming years as the benefits in terms of cancer related morbidity and mortality needed to be weighed against harms including over diagnosis and overtreatment, along with associated morbidity resulting from both investigation and treatment.

As a result of the balance of these two considerations, in 2008, the US Preventative Services task force recommended against prostate cancer screening using PSA testing for men aged 75 years and older. And four years later in 2012, the same group recommended against PSA screening for all men. Subsequent to this, their recommendations have now been revised, but we're seeing the effects of these earlier recommendations in population level data.

What we saw following these guideline recommendations, was a substantial decrease in the diagnosis of non-metastatic prostate cancer. However, we've seen some substantial increases in metastatic prostate cancer diagnoses. The underlying causes and directly linking these to guideline recommendations is a little bit unclear and remains debated. One hypothesis is that, lower screening due to these guidelines missed a diagnosis of non-metastatic disease, were subsequently progressed to metastatic disease and was diagnosed at that time. However, it remains unclear whether the variation in PSA screening is associated with subsequent metastatic prostate cancer incidents.

These authors start to address this knowledge gap using data from the US Veteran Health Administration data set to assess variation in facility level-PSA screening, prostate biopsy utilization, and the incidence of non-metastatic prostate cancer as well as metastatic prostate cancer during 2005 and 2019. Their underlying hypothesis was that, lower facility level-PSA screening earlier in the interval would be associated with higher rates of metastatic prostate cancer subsequently. So to do this, the authors relied on data from the US Veterans Health Affairs Corporate Data Warehouse and they evaluated time trends and PSA screening rates, prostate biopsy rates, incidence rates of non-metastatic prostate cancer, and incidence rates of metastatic prostate cancer between January 2005 and December 2019.

Analytically, they assessed rates of each of these endpoints annually and aggregated these across a number of factors including age group reflecting the guideline recommendations. So age 40 to 54, the primary screening group, 55 to 69, and then those 70 and older. They also aggregated groups to calculate rates due to race and ethnicity and across VHA facilities. Obviously, event rates here excluded patients who were diagnosed with prostate cancer before the time of the study interval.

PSA screening was determined using PSA lab tests in lab data and validated using CPT codes. Screening rates at a facility level were considered by looking at the number of unique male patients with a PSA test performed at each facility in each year. Patients were categorized as long term non screening individuals if they did not have any PSA testing within three years. And at facility level, this proportion was calculated divided by the total number of men who received care at each facility each year. The prostate biopsy rate was calculated by examining the number of prostate biopsy procedures as defined based on CPT, ICD-9 and ICD-10 codes divided by the total number of men with a Veteran's Health Administration encounter in each year at each facility. These were age adjusted to the US population in 2000 using the direct method.

The authors then examined rates of prostate cancer incidents. This is stratified into non-metastatic and metastatic disease. For non-metastatic disease, the authors required a pathology report or other documentation of Gleason score within 90 days of the diagnosis of prostate cancer as determined using ICD-9 or 10 codes without evidence of documentation of metastasis within 90 days. Metastases were identified using natural language processing, a book, clinical notes and pathology reports, and this formed the key distinguishing characteristic between those with metastatic and non-metastatic disease. Within each diagnostic category, the number of patients with incident prostate cancer were divided by the number of men who received any care at any given facility that year. These again, were adjusted to the 2000 US standard population.

The authors, in terms of the statistical analysis, used regression modeling to assess these associations. The primary independent variable was the annual facility level-PSA screening rate, and the dependent variable was the facility level prostate cancer incidents for metastatic disease five years later. They, therefore, assessed screening frequency between 2005 and 2014 and incident metastatic prostate cancer between 2010 and 2019. They chose this five year leg and keeping with ERSPC data, but varied it from three to seven years in sensitivity analyses.

The primary regression model used a multi-variable negative binomial mixed-effects regression with an offset for the facility level of total male population. Examine Covariates, including the proportion of men who are black, the proportion of men aged 70 years and older, the availability of advanced imaging techniques including PET scanning, the use of prostate MRI, the calendar year, and the facility region. They further undertook sensitivity analyses, changing this analytic approach from a negative binomial model to a Poisson model, and then they restricted their cohort to those who were deemed to frequent VHA users requiring a visit in each of the three years prior to index for PSA screening assessment to ensure that these patients were not receiving care elsewhere.

And I hand it over to Zach to walk us through the results of this interesting epidemiologic study, really assessing the association between PSA screening and metastatic prostate cancer diagnosis.

Zachary Klaassen: Thanks so much, Chris. This is the characteristics of 128 Veterans Healthcare Administration Facilities that were included in these analyses. So this just gives us a bit of a breakdown of what made up these institutions. On the left, this is the population proportion of patients with black race in each facility in 2012. You can see that the majority of these facilities had less than 25% black patients in their VA, but there was several with about 35 to 50% black patients in that VA facility. In the middle, is the facility size measured by the number of men with any encounter in 2012, and this is based on numbers in the thousands. So you can see that the majority were between 10,000 and roughly 55,000 with a few outliers nearing 100,000 VA users. And finally, on the right, this is the population proportion of each age group. In 2012, not surprisingly, the majority of these patients were in their fifties and sixties and the least amount in their forties and fifties and right in the middle men that were over the age of 70.

This table looks at overall screening plus diagnosis rates per year. So there's several metrics that we'll walk through. As you can see on the left, this is going from year 2005 to 2019. And when we specifically look at PSA screening at 2005 was 47.2%, with a peak at 50.8% in 2008, and then decreasing all the way down to 37% in 2019. If we look at the prostate biopsy rates, in 2005 all the way down to 2019, we see a steady decrease in the prostate biopsy rates throughout these 15 years, starting in 2015 all the way to 2019. When we look at incident non-metastatic prostate cancer cases with a Gleason score of 8 to 10% or eight to 10, we see an increase in the rate of Gleason 8 to 10 disease, 15.5% in 2005, all the way up to 23.8% in 2019. And finally, on the far right, just looking at metastatic prostate cancer rates incidents per 100,000 men, in 2005, 5.2, with a peak of 8.2 in 2017, and in 2019 7.9. So this gives us an overview of the subsequent slides that we'll be discussing.

These figures look at the pooled PSA screening, prostate biopsy rate, and prostate cancer incidence rates, and we'll go through each of these curves in sequence. For each of these curves, the dotted line represents 2012 as highlighted by Chris with the USPSTF recommendations. And so looking at the top left proportion with PSA screening, we do see a decrease in PSA screening after 2012. In panel B in the middle, this is proportion without PSA screening in the prior three years and after 2012, again, we see an increase in the proportion without PSA screening. The prostate biopsy rate is in panel C on the right and this you can see after 2012, a decrease in prostate biopsies and leveling off at about 2016 here. In the bottom left, this is incident non-metastatic prostate cancer decreasing after 2012, but leveling off. We get into panel E in the middle, this is incident non-metastatic prostate cancer with Gleason 8 to 10. So we're diagnosing more locally high risk prostate cancer after 2012. And importantly, incident metastatic prostate cancer, we see a pretty significant increase from 2012 peaking in 2017, but still high in 2018 and 2019.

This figure looks at PSA screening rates by age group. On the left, this is PSA screening rate, and the aged delineation is orange, 40 to 54, dark gray, 55 to 69, and light blue greater than 70 years of age. That's surprisingly the highest screening rates were in men, 55 to 69 years of age, but we can see that in all three groups. Again, with the dotted line representing 2012, we see a decrease in PSA screening rates in all three of these age groups. In terms of long term non screening rates, we see after 2012, an increase in the non screening rates among these three age groups.

This is PSA screening rates by racial and ethnic group. The delineation for this is, black men are in the light, gray, white are in dark gray, other is in orange, unknown is in blue. And we can see that generally from 2012, we see a decrease in PSA screening rates among all of these ethnicities and races. But interestingly, we see a slightly less decrease in black men after 2012. In terms of long term non screening rates, again, we see an increase in non screening rates after 2012 and a slightly less of a decrease in non screening rates among black men.

This looks at prostate cancer instance rates by age group. This is non metastatic prostate cancer incidents, and we can see that there is a decrease slightly in all age groups, but also leveling off particularly among the men 55 to 69 and 40 to 54. When we look at metastatic prostate cancer incidence rates, we see that among younger men, this is relatively stable between men 40 to 54 years of age, but we do see a significant increase in elderly men, so men more than 70, we see a significant increase over the course of years from 2012 peaking at 2017, 2018. And also, similar pattern among men, 55 to 69 years of age.

Finally, looking at prostate cancer incident rates by racial and ethnic group. We do see the highest incidents of prostate cancer among black men, which is not surprisingly. This is age adjusted non metastatic prostate cancer on the left and age adjusted metastatic prostate cancer incidences on the right. Again, black men are the most common, we do see, on the left, black men having a decrease in non metastatic prostate cancer diagnoses after 2012, but an increase in metastatic prostate cancer diagnoses after 2012. And again, similar patterns but less drastic among the other races as you can see listed here.

This figure looks at the association of PSA screening rates and long term non screening rates with subsequent metastatic prostate cancer incidents. On the left, this is PSA screening rate and estimated metastatic prostate cancer cases. On the X axis is PSA screening rate, on the Y axis is estimated yearly prostate cancer cases. So with increasing screening rates, we have decreased metastatic prostate cancer rates. And to complement this, this a is long-term non screening rates, so with more non screening, we see increased risk of metastatic prostate cancer. So these two figures do complement each other with regards to PSA screening and non screening.

This table looks at the results of the mixed effects negative binomial regression model for metastatic prostate cancer rates. So in the model for PSA screening rate, we see that men that did have PSA screening had a significantly reduced risk of having metastatic prostate cancer. And to compliment that, again, those that had non screening were predictive of having metastatic prostate cancer. Interestingly, in both models, being black or African American increased the risk of metastatic prostate cancer.

With several discussion points from this study, this population based study found a decrease in PSA screening rates between 2008 and 2019 across all subgroups. And this coincided with increase in long term non screening rates, a decrease in prostate biopsy rates, and an increase in incident metastatic prostate cancer. The 10 to 15% absolute decrease in PSA screening rates among veterans around the time of the 2008 and 2012 USPSTF guidelines is consistent with other reports showing a 5 to 15% absolute decrease in PSA screening.

These results suggest an association between lower facility level PSA screening rates and higher metastatic prostate cancer incidence rates, which may implicate PSA screening behaviors and subsequent metastatic prostate cancer incidents. Finally, there was an increase in annual PSA screening rates and lower long term non screening rates among black veterans compared with veterans from other racial or ethnic groups, particularly after 2012. And this may reflect physician awareness of the higher risk of metastatic incidents prostate cancer among black men, and it's reflected in observed higher rates of prostate biopsies, non metastatic prostate cancer, and metastatic cancer among black men in this study.

Several limitations from this epidemiological study. These results are focused on the US veteran population, which differs from civilian populations with regards to age distribution, comorbidity, and socioeconomic factors. And this may have unique environmental exposures associated with increased baseline risk of prostate cancer. Secondly, some veterans may have received care at several facilities each year and others may have moved between primary facilities during the study period. And finally, 7.1% of individuals reported unknown race or ethnicity, and these patients tended to have lower screening rates than higher long term non screening rates.

In conclusion, this cohort study found that from 2005 to 2019, PSA screening rates decreased among men more than 40 years of age in the VA healthcare system. Secondly, facilities with higher PSA screening rates had lower subsequent rates of metastatic prostate cancer. And finally, these data may be used to inform shared decision making about the potential benefits of PSA screening among men who wish to reduce the risk of metastatic prostate cancer.

Thank you very much for your attention, we hope you enjoyed this UroToday Journal Club discussion of this recent publication in JAMA Oncology.