The clinical use population (N = 5157) showed a median age of 65 years (interquartile range (IQR): 60–71)), a median PSA level of 6.60 ng/mL (IQR: 4.90–9.06) and a median PSA density of 0.13 ng/ml2 (IQR: 0.09–0.19). Prostate volume was unknown for 94 cases, resulting in exclusion of these cases in analysis on PSA density. Abnormalities were observed during DRE in 17.82% of the patients. In 40.72% of the cases, the SelectMDx result was negative.
Although the SelectMDx test is intended to be used in patients with PSA levels between 3 and 10 ng/mL, SelectMDx analysis is also requested for patients with PSA levels below 3 and higher than 10 ng/mL. Therefore, SelectMDx results were analyzed per PSA group. 5.33% of the cases had a PSA value < 3 ng/mL, 76.65% of the cases reported a PSA value ≥ 3 and ≤10 ng/mL, 11.23% of the cases had a PSA value > 10 and ≤15 ng/mL and 6.79% showed a PSA value > 15 ng/mL (see Table 1).
PSA 3–10 ng/mL (n = 3953)
In the intended use population of the SelectMDx test (PSA 3–10 ng/mL), the median age is 65 years (IQR: 59–70), median PSA density is 0.12 ng/ml2 (IQR: 0.09–0.17) and 15.79% of the patients showed DRE abnormalities. Analysis of the SelectMDx results per PSA group showed that 44.65% of the cases with PSA between 3–10 received a negative SelectMDx result, therefore biopsies were potentially saved in 1765 men in this intended-use population of the SelectMDx test (see Fig. 1).
PSA < 3 ng/mL (n = 275)
Age, prostate volume, PSA density and RNA score were lower compared to the other PSA groups. In contrast, the percentage of cases showing an abnormal DRE result is relatively high (29.82%), as compared to the patients with PSA between 3 and 10 (15.79%). A relatively high percentage of patients (75.27%) in this group received a negative SelectMDx result (Fig. 1). In patients with a PSA level below 3 and an abnormal DRE finding, the SelectMDx test reported 35 SelectMDx positive cases (43.67%) and 49 SelectMDx negative cases (58.33%).
PSA > 10 ng/mL (n = 929)
Age, prostate volume, and RNA score are higher in this group compared to the other PSA groups. Prostate volume slightly increases between the groups of PSA 10–15 and PSA > 15, in contrast to larger differences in PSA between both groups. This results in large differences in PSA density between these groups (0.19 ng/ml2 versus 0.31 ng/ml2). In cases with PSA levels > 10 and ≤15, only 19.34% (n = 112) received a negative result. The lowest percentage of negative results (4.57%) is found in patients with PSA levels > 15 ng/mL (n = 16), see Fig. 1.
Next, subgroup analysis was performed to compare data on clinical risk factors and SelectMDx results by age groups. Most of the patients have an age ≥65 and <75 (44.23%), followed by the age group ≥55 and <65 (36.77%) (see Table 2). Prostate volume (median), RNA score (median) and percentage abnormal DRE findings increase with increasing age group (except for prostate volume in two highest age groups). PSA density was shown to be stable over the age groups <75 years (PSA density 0.12–0.13 ng/ml2) and showed a sharp increase in age group 75–85 years (PSA density 0.15 ng/ml2) and age group >85 years (PSA density 0.24 ng/ml2). This can be explained by the combination of increasing PSA and a stable prostate volume. Although the PSA density and percentage abnormal DRE findings between age group 45–55 years and 55–65 years are similar, the percentage of cases with a positive SelectMDx result is 17.53% higher in the 55–65 years age group. This could be explained by the increase in age and RNA score in the 55–65 years age group. Overall, the percentage negative SelectMDx results decrease with age, ranging from 83.87% negative results in patients aged <45 years to 0% negative SelectMDx results in patients aged >85 years (see Fig. 2). However, age is a confounding factor in this last analysis since age is included in the SelectMDx algorithm.
Comparing countries and the validation cohort
Subgroup analysis was performed comparing patient groups from different countries (see Table 3). Most of the requests for SelectMDx tests originated from the Netherlands (34.83%) followed by Spain (31.12%) and Italy (7.27%). Differences were found in patient characteristics between countries, with PSA (median) ranging from 5.87 ng/mL in Slovenia to 7.67 ng/mL in Belgium and median age ranging from 63 years in Belgium to 69 years in France. PSA density varied from 0.11 ng/ml2 (Italy) to 0.16 ng/ml2 (Belgium). Considering DRE outcome, cases from Switzerland (45.3%) and Germany (36.9%) have higher percentage abnormal DRE, compared to all other countries (17.82%). PSA density ranges from 0.11 ng/ml2 (Italy) to 0.16 ng/ml2 (Belgium). Next, the percentage of cases with PSA levels below 3 ng/mL varies from 2.43% in Spain to 12.79% in Poland. These differences in patient characteristics between countries are translated into differences in test results: the percentage negative SelectMDx test results range from 30.3% in France to 46.93% in Italy.
In the validation cohort (all PSA levels) of the SelectMDx test, 35.70% of all SelectMDx tests resulted in a negative result. When comparing the ratio of positive/negative SelectMDx result between the clinical use population and the validation cohort, the number of cases with a negative SelectMDx result was significantly higher in the clinical use population compared to the validation cohort: χ2 (1) = 8.18, p < 0.05 (see Table 4). No significant differences were found in PSA levels and age between the clinical use population and the validation cohort. In contrast, prostate volume was significantly higher in the clinical use population compared to the validation cohort, and therefore, PSA density was significantly lower in clinical use cohort compared to the validation cohort. The validation cohort included significantly more cases with an abnormal DRE result as compared to the clinical use population.
Next, the ratio positive and negative SelectMDx test results were compared between each country and the validation cohort. A significantly higher percentage negative SelectMDx result was observed in Italy (χ2 (1) = 14.12, adjusted p < 0.05) and Spain (χ2 (1) = 10.03, adjusted p < 0.05) compared to the validation cohort (Fig. 3). These differences in SelectMDx results could be partly explained by differences in patients selected for a SelectMDx test in these countries. Post-hoc analysis on Italy and Spain showed that PSA density is significantly lower in Spain and Italy compared to the validation cohort. However, PSA is significantly higher in Spain compared to the validation cohort. PSA density is the lowest in Italy, contributing to the higher number of negative SelectMDx results in Italy (see Supplementary Data).