A Head-to-head Comparison of Prostate Cancer Diagnostic Strategies Using the Stockholm3 Test, Magnetic Resonance Imaging, and Swedish National Guidelines: Results from a Prospective Population-based Screening Study

Take Home Message The use of the blood-based Stockholm3 test was shown to be beneficial to the existing Swedish national guidelines for detecting prostate cancer.

Stockholm3 strategy diagnosed 32 csPCa cases, with 89 biopsied and 27 ISUP 1 cancers. SNG-2020 detected 32 csPCa and 33 ISUP 1 cancer patients, with 99 men biopsied, and the MRI strategy detected 30 csPCa and 35 ISUP 1 cancer cases by biopsying 123 men. The latter two strategies generated more MRI scans than the Stockholm3 strategy (n = 270 vs 33). Conclusions: Previous guidelines provide high detection of significant cancer but at high biopsy rates and detection of insignificant cancer. The Stockholm3 test may improve diagnostic precision compared with the current guidelines or using only MRI. Patient summary: The Stockholm3 test facilitates detection of significant cancer, and reduces the number of biopsies and detection of insignificant cancer.

Introduction
Prostate cancer screening studies show reduced prostate cancer mortality with prostate-specific antigen (PSA) testing followed by systematic biopsy for men with elevated PSA levels [1,2]. However, organized population-based prostate cancer screening has not been introduced in any country except Lithuania and Kazakhstan, due to the risk of overdiagnosis and overtreatment [3].
To improve prostate cancer diagnostics, risk prediction models using clinical variables and biomarkers have been developed. These models, for example, Stockholm3, Prostate Cancer Prevention Trial Risk Calculator, Prostate Biopsy Collaborative Group Risk Calculator, European Randomised Study of Screening for Prostate Cancer, 4K, and Prostate Health Index, have shown favorable results compared with PSA to detect clinically significant prostate cancer (csPCa) [4][5][6][7][8][9]; at the same time, these have been shown to reduce the number of biopsies performed and International Society of Urological Pathology (ISUP) grade 1 cancers diagnosed [10]. Studies using prebiopsy magnetic resonance imaging (MRI) have increased biopsy precision [11,12] and reduced the number of low-risk cancers detected [13], which has led to recommendations to include MRI in diagnostic algorithms [14][15][16].
In Sweden, from 2007 until 2020, five versions of the Swedish national prostate cancer guidelines (SNGs) have been published [17,18]. These guidelines recommend the combined use of clinical variables, biomarkers, and imaging to refer men to prostate biopsy. Version 4 (introduced in 2019 and valid between 2019 and 2020 [SNG-2019]) recommended the use of age, clinical T stage, PSA, free PSA, and PSA density for selecting men to undergo biopsy. They also define recommendations for repeat biopsies with the use of MRI. Version 5 (introduced in 2020 [SNG-2020]) combines previous guidelines with the use of MRI to recommend men to undergo biopsy. The guidelines have not been validated for detecting csPCa in a screening setting.
The aim of this study was to compare the precision of the Stockholm3 test or MRI with contemporary national guide-lines in Sweden to select patients for the detection of csPCa in a screening setting.

2.1.
Participants and study design

Evaluated strategies
We compared four different screening strategies, two SNGs (SNG-2019 and SNG-2020 version), a strategy based on the Stockholm3 test, and a strategy using only MRI to identify men to be subjected to biopsy.
Following is a description of each strategy (for a more detailed description; see Fig. 2).

SNGs in use in 2019
Participants with a PSA level of 3 ng/ml were referred to a urologist, where prostate volume was measured by TRUS and digital rectal examination (DRE) was performed. If DRE was positive, PSA was >20 ng/ml, the quotient (PSA free/PSA total) was 0.2, or PSA density (PSA total/ prostate volume) was 0.1, the participants underwent a systematic ten-core biopsy. Men with a previous negative biopsy had a second ten-to 14-core systematic biopsy if PSA density was 0.15. Men with either ISUP grade 1 in the first biopsy or positive DRE or PSA density 0.15 had a repeat biopsy, those with a PI-RADS 3 on the MRI had a targeted biopsy, and those with PI-RADS <3 had a ten-to 14-core systematic biopsy.

Statistical analysis
To compare the four strategies, we counted the number of procedures

Results
Characteristics of the study population are shown in Table 1.
The median age of the study participants was 65 yr (interquartile range [IQR]: 60-69) and the median PSA level was 4.1 ng/ml. The median Stockholm3 scores were 12% (IQR: 8-21) for the total study population and 35% (IQR: 20-46) for the men diagnosed with csPCa, while it was 9% (IQR: 7-13) for men with benign biopsy. Of the men in our study population, 50% had a PI-RADS score of 2 (136 men), 73 men (27%) had a PI-RADS score of 3, and 61 men (22%) had a PI-RADS score of 4. In total, 211 men had a biopsy, and 37 csPCa and 59 ISUP 1 cases were diagnosed.

3.1.
Cancer detection using four different diagnostic strategies Figure 3 shows the outcome of the different strategies. By selecting men for biopsy using SNG-2019, the total number of biopsy procedures was 179 (65% of total biopsies), while detecting 36 of 37 csPCa (97%) and 49 of 59 ISUP 1 prostate cancer (83%) cases. By using the ST-3 strategy, 89 men were biopsied (33% of total biopsies), while detecting 32 csPCa (86%) and 27 ISUP 1 cancer (46%) cases. With the MR strategy, 123 men underwent biopsy (58%); 30 of 37 csPCa cases were detected (81%) and 35 men were diagnosed with ISUP 1 cancer (59%). The new guidelines SNG-2020 resulted in a decrease in the number of biopsies performed in the total cohort from 211 to 99 (36%), while detecting 32 csPCa (86%) and 33 ISUP 1 cancer (56%) cases. The number of MRI examinations differed for each strategy. In the total study population, all men except two had an MRI scan (270 out of 272). With SNG-2020 and MR strategy, all 270 men underwent MRI; with SNG-2019, 71 men underwent MRI (26%); and with the ST-3 strategy, 33 men underwent MRI (12%), while detecting the same number of csPCa as SNG-2020 (32 out of 37).

3.2.
RPs of procedures and outcomes of the different strategies

Discussion
We performed a population-based, paired, screening-byinvitation study to compare the previous and current Swedish national diagnostic guidelines for strategies based on a blood-based screening test (the Stockholm3 test) or MRI to detect csPCa in men with a PSA level of 3 ng/ml. When compared with the current SNGs, we found that the ST-3 strategy performs similarly for the detection of csPCa, while reducing the number of performed MRI scans by 88%, performed biopsies by 10%, and diagnosed ISUP 1 cancer cases by 18%. The national guidelines are complex to follow and at risk of underperforming compared with the ST-3 strategy. Compared with the ST-3 strategy, the previous SNG strategy detected a slightly higher number of csPCa cases (36 vs 32) while performing 100% more biopsies, detecting 118% more ISUP 1 cancers, as well as performing 115% more MRI scans. Compared with the MR strategy, the ST-3 strategy detected slightly more csPCa cases (32 vs 30), with markedly fewer MRI scans (33 vs 270) and biopsying fewer men (89 vs 123). In Supplementary Table 1, the characteristics of men with csPCa who were missed by the strategies are described; the majority of cancers missed by all strategies were graded as ISUP 2 on pathology.
Previous studies using the Stockholm3 test for the detection of prostate cancer have shown favorable results compared with screening using the PSA test only [4,20,21]. In these studies, the Stockholm3 test was used after a prior PSA test with a cutoff of 1.5, thus detecting csPCa in the PSA range between 1.5 and 3.0. In this study, we not only use the Stockholm3 test in another population different from the one used for the development of the test, but also  [22]. Studies have shown that the use of prebiopsy MRI followed by a targeted biopsy can increase the detection of csPCa in a clinical setting if used together with a systematic biopsy [12,13,23]. However, as a standalone procedure it can lack sensitivity for detecting csPCa, and it has been shown that MRI-targeted biopsies alone can underestimate the histological grade of some tumors [24]. We assessed a strategy using only MRI for targeted biopsies in a screening setting, and our results showed that as a standalone procedure, it performed worse than the other three strategies  tested, missing seven out of 37 csPCa cases, while referring 38% more men to biopsy and detecting 30% more ISUP 1 cancers than the ST3 strategy.
There have been previous efforts to describe the Swedish diagnostic habits and the national guidelines. However, we do not know of any previous studies assessing the diagnostic performance of the strategies stated in the SNGs. Nugin et al [25] used registry data to assess how the guidelines affected prostate cancer care before and after the first publication. Their results indicated modest improvements in performance of 14 selected quality indicators in prostate cancer care after the guidelines were published. However, to make an impact on general prostate cancer care in Sweden, they recommended that feedback from prostate cancer care was needed. Another study by Nordström et al [26] using registry data from Stockholm, Sweden, showed that PSA testing was prevalent and increasing despite national recommendations against PSA screening. This is a further indication that despite the publication of guidelines in Sweden, these are generally not being followed in a systematic and clear way, calling for a more effective, organized screening program.
Among the key strengths of this study are the ongoing screening-by-invitation setting (since 2015) and wellcontrolled clinical processes. These clinical processes include the use of an MRI protocol suitable for high throughput as needed in a screening setting and centralized primary assessment. An external validation of the MRI assessments was performed, as well as a second evaluation in a high-volume experienced center. We also used a single, well-defined, and contemporary MRI/fusion biopsy method for disease verification and unified protocols for biopsy procedures using a centralized pathological assessment. In this study, we assessed complex diagnostic chains, and for robustness in the analysis, we aimed to decrease variability between providers throughout the study processes.
The Stockholm3 versus SNG study was performed in Värmland, Sweden, using centralized radiology and pathology, which may limit generalizability to other health care settings. In addition, the relatively small sample size in our study population is a limitation to the study. The results presented here show performance of the diagnostic strategies at one screening round; performance across multiple rounds of screening is thus unknown. The cases of csPCa missed using any of the strategies would probably have presented themselves in one of following screening rounds, as these cases are normally followed up. Since MRI-detected or Stockholm3-detected csPCa cases overlap to cover almost all cases, an effort to combine these tools may improve the selectivity. A deeper hypothesis-generating analysis of all different combinations of Stockholm3 and MRI, to try to find an optimal algorithm, would be of high interest.

Conclusions
There is a high rate of overbiopsying and overdetection of csPCa in a screening strategy aiming to find all csPCa cases.
Our results indicate that the use of the Stockholm3 test for the detection of prostate cancer could be effective in the reduction of biopsies and detection of insignificant cancers, while maintaining the detection of significant cancer, compared with the previous and current SNGs, as well as a strategy based only on MRI of the prostate.
Author contributions: Mauritz Waldén had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.