Comparison of serum prostate specific antigen (PSA), soluble E-cadherin (sE-cad), and inosine monophosphate dehydrogenase-2 (IMPDH-2) as aggressive prostate cancer predictors

Abstract

Background

This study aimed to compare serum prostate specific antigen (PSA), Soluble E-cadherin (sE-cad), and Inosine Monophosphate Dehydrogenase-2 in predicting prostate cancer aggressiveness by determining their correlations with Gleason score and International Society of Urological Pathology (ISUP) Grade Groups.

Methods

This was a hospital-based descriptive quantitative cross-sectional study whereby we enrolled 48 newly diagnosed prostate adenocarcinoma patients in the study. Their serum was analysed for PSA, sE-cad, and IMPDH-2. Pearson correlation coefficient was used to test the correlation between the serum sE-cad and Gleason score while Spearman rho correlation coefficient was used for PSA and IMPDH-2. The correlation coefficient (r) was graded as very weak (< 0.3), weak (0.3–0.4), moderate (0.5–0.6), or strong (≥ 0.7), while the magnitude was determined by calculating the coefficient of determination for the respective analysis (R²). The correlation between the biomarkers and the ISUP Grade groups was determined using the Kendall tau correlation coefficient (τ). All levels of statistical significance were set at p < 0.05.

Results

The mean age of the subjects was 69.4 years. The Means of serum PSA, sE-cad, and IMPDH-2 were 47.2 ng/ml, 136.5 ng/ml, and 89.8 pg/ml respectively. Serum PSA weakly correlated with both Gleason score (r = 0.3, p = 0.04) and ISUP grade groups (τ = 0.3, p = 0.02). The magnitude was 0.097. Similarly, serum sE-cad correlated weakly with both Gleason scores (r = 0.4, p = 0.01), and ISUP Grade Groups (τ = 0.3, p = 0.005). The magnitude was 0.134. However, serum IMPDH-2 neither correlated with Gleason score (r = 0.03, p = 0.86) nor ISUP Grade Groups (τ = 0.004, p = 0.97). Serum sE-cad did not outperform both IMPDH-2 (p = 0.91) or PSA (p = 0.23) in predicting the Gleason score.

Conclusions

Serum sE-cad best predicted aggressive prostate cancer but did not statistically outperform serum PSA or IMPDH-2. Hence, neither of the three are reliable predictors of aggressive prostate cancer.

1 Background

Globally, prostate cancer is the second most common male cancer with over 1.4 million cases diagnosed in 2022 [1]. However, in Africa, it is the most common cancer in men constituting 20.4% of newly diagnosed cancers [1]. It was perceived to be low among African men but recent evidence suggests its prevalence to be up to 300 per 100,000 which is close to that found among African Americans [2]. Prostate cancer is on the rise among Nigerian men and is often aggressive and usually diagnosed in an advanced stage [3]. It is the number five cause of death among men and the highest death rate has been reported among African descent [4]. The widespread use of serum prostate specific antigen (PSA) has led to earlier detection of PCa though sometimes with overdiagnosis of clinically insignificant disease [5]. Among the limitations of PSA is its inaccuracy in predicting aggressive PCa [5]. Gleason grading system is the commonly used parameter in predicting the aggressiveness of PCa with higher Gleason scores corresponding to more aggressive disease [5]. It has undergone several revisions to improve its accuracy with the latest being the International Society of Urological Pathology (ISUP) grade groups which have been incorporated into the 2016 World Health Organization (WHO) classification of tumors of the prostate [6, 7]. However, the Gleason score requires a prostate biopsy which is an invasive procedure and high-grade foci can be missed with a needle biopsy. Therefore, there is a need for a novel biomarker that can improve the detection of aggressive PCa. Both serum IMPDH-2 and sE-Cad have been reported to correlate with the Gleason score and ISUP Grade group [8, 9]. However, to our knowledge no such study has been done in sub-Saharan Africa where majority of black men reside. Inosine monophosphate dehydrogenase-2 is a rate-limiting enzyme that catalyses a vital step in the de-novo biosynthesis of purine nucleotides. [10] In contrast, Cadherins are Ca²⁺ dependent cell adhesion molecules that play a crucial role in normal growth and development through the mediation of homotypic, homophilic cell–cell association [11]. Shedding of E-cadherin ectodomain determines the release of soluble 80 kDa E-Cadherin into body fluids and elevated serum levels have been found in patients with cancer when compared with healthy individuals [12]. A lot of our patients with prostate cancer present with aggressive and advanced diseases and finding a biomarker that can identify such patients will improve their management. The study aimed to assess and compare serum PSA, sE-Cad, and IMPDH-2 in predicting aggressive PCa in the black population by determining their correlations with the Gleason score and ISUP grade group. This may help to improve decision-making on which patients can safely be observed or undergo active surveillance. It may also assist in early recognition of disease upgrade, and therefore play a role in risk restratification of patients on observation, active surveillance, or treatment. Even though the cost of these tests may not be significantly different from that of serum PSA, the kits are not readily available in Nigeria [13, 14].

2 Methods

This was a hospital-based, descriptive, quantitative cross-sectional study conducted between April 2021 and March 2022. The study aimed to assess and compare serum PSA, sE-Cad, and IMPDH-2 in predicting aggressive PCa in the black population by determining their correlations with the Gleason score and ISUP Grade group. It was approved by the health research and ethics committee of Usmanu Danfodiyo University Teaching Hospital (UDUTH), and the study subjects were recruited from the urology outpatient clinic using a non-randomized sampling technique. The inclusion criteria were all newly diagnosed patients with adenocarcinoma of the prostate while the exclusion criteria were any patient who has undergone any form of treatment for prostatic disease such as hormonal therapy, radiation therapy, or surgery. Patients with any histological variant of prostate cancer other than adenocarcinoma and those with any second malignancy as well as those who did not grant consent were all excluded. The sample size was calculated using the formula for sample size estimation for a quantitative cross-sectional study as follows [15]:

$$n=\frac{{Z}_{{1-\frac{\alpha }{2}}^{2}}{ \times SD}^{2}}{{d}^{2}}$$

n = Minimum required sample size.

Z_1-α/2² = Standard normal variate (at 5% type I error (P < 0.05) is 1.96).

SD = Standard deviation of the variable. The value of standard deviation can be taken from a previously done study or through a pilot study = 102.8 [8].

d = Precision, tolerable margin of error = 1.5

Therefore:

n = 18,035.7

The sample size was approximately 18,036 subjects. However, this sample size is for a population > 10,000. For a population of < 10,000 [Patients with the condition (PCa) seen in the previous year (N = 42)], a finite sample size (nf) needs to be calculated as follows [15]:

$$nf=\frac{n}{1+ n/N}$$

nf = 18,036 ÷ (1 + 429.4) = 41.9

The minimum sample size was approximately 42 subjects and 10% (4.2) was added for attrition bringing the minimum sample size to 47 Subjects. The enrolled study subjects were clinically evaluated by taking their detailed history and performing a physical examination including digital rectal examination (DRE) to search for abnormal prostate findings such as discrete nodules, focal induration, or a diffusely hard prostate. Relevant laboratory investigations done included serum PSA, Full blood count (FBC), Electrolytes, Urea, and Creatinine (E/U/Cr), Urinalysis, Urine Microscopy, Culture, and Sensitivity (M/C/S). The assays for serum IMPDH-2 and serum sE-cad were done using The Sunlong^_Biotech^® Enzyme-Linked Immunosorbent Assay (ELISA) kit based on the manufacturer’s instructions as contained in the manual of the reagent kit [16, 17]. The assays were carried out by a trained laboratory scientist in Chemical Pathology of Usmanu Danfodiyo University Teaching Hospital. The serum for the assay in aliquots was stored at − 20 °C. Repeated freeze–thaw cycles were avoided, all haemolysed samples were discarded, and new samples were collected as these can interfere with the results. The diagnosis of PCa was made through transrectal ultrasound-guided prostate biopsy using the Mindray^® Diagnostic Ultrasound System DC-30. The frequency of the end-firing rectal probe was set at 6.5MHz. One gram of intravenous ceftriaxone was administered to the patients ten minutes before the procedure. The anaesthetic technique employed was 10 ml of 2% lidocaine gel administered transrectally, 10 min before the procedure. Participants with significant pain during the procedure had additional analgesia in the form of Intravenous 600mg Paracetamol^® or 30 mg Pentazocine. For the procedure, the participants were positioned on a left lateral decubitus position with knees and hips flexed 90°, and the buttocks flush with the edge of the examination bed for easy insertion and manipulation of the rectal probe. The perineum was cleaned and draped. The perineal region was first inspected and DRE was then performed using a well-lubricated gloved index finger to characterize the prostate and identify the presence of any ano-rectal pathology. A lubricated rectal probe coupled with a needle guide and covered with a condom was gently inserted into the rectum and advanced towards the base of the bladder until visualization of the seminal vesicles. The probe was first positioned transversely to obtain prostatic images from the base to the apex of the prostate, taking details about the echotexture, presence of nodules, and capsular integrity of the prostate. The prostate volume was then determined using the largest dimensions of the prostate in transverse and midsagittal planes. A total of 12 prostate biopsy cores were obtained systematically in each patient using a Bard^® Magnum spring-loaded biopsy gun with an 18 gauge, 20cm needle. The specimens were fixed in 10% Formalin and sent to a histopathologist for analysis. The histopathologists at the Department of Histopathology of Usmanu Danfodiyo University Teaching Hopsital Sokoto did the histological diagnosis and Gleason grading of the tumour. The Gleason score is reported as a mathematical equation, e.g., 4 + 3 = 7. The lowest assigned Gleason score (GS) was 6 based on the current recommendation on Gleason grading for a needle biopsy [7]. Similarly the study subjects were stratified into five groups of increasing tumour aggressiveness based on the ISUP grade groups. These grade groups are; 1 (GS ≤ 6), 2 (GS 3 + 4), 3 (GS 4 + 3), 4 (GS 8), and 5 (GS 9−10).

The relevant data were collected using a structured Proforma and at the end of the study, the data was entered into IBM^® SPSS^® for Windows, version 23.0 for data analysis. Frequency distribution tables were drawn for all variables while mean and standard deviation were calculated for the continuous variables. The correlation of serum PSA and IMPDH-2 with Gleason score was determined using the Spearman rho correlation coefficient (not normally distributed data) while that of sE-cad was determined using the Pearson correlation coefficient (normally distributed data). The correlation coefficient (r) was graded as very weak (< 0.3), weak (0.3–0.4), moderate (0.5–0.6), or strong (≥ 0.7), while the magnitude was determined by calculating the coefficient of determination for the respective analysis (R²) [18]. The correlation between the biomarkers and the ISUP grade groups was determined using the Kendall tau correlation coefficient (τ). A comparison of the correlation of the biomarkers and the Gleason score was done using the Meng method for comparing correlated coefficients [19]. All levels of statistical significance were set at p < 0.05.

3 Results

Forty-eight subjects aged between 55 and 100 years were enrolled in the study. The mean age was 69.4 years ± 8.5 while the modal age was 70 years. Figure 1 shows the age distribution of the study subjects and the common presenting symptoms are shown in Table 1. The duration of symptoms before presentation varies between 2 and 64 months while the mean duration was 19.7 months ± 17.7. Only 5 (10.4%) subjects were cigarette smokers while 23 (47.9%) subjects had a medical co-morbidity. The identified comorbidities are shown in Fig. 2. On physical examination, the DRE finding was suspicious of prostate cancer in 44(91.7%) subjects. The urine culture yielded microbial growth in 13 (35.1%) subjects and the microorganisms are shown in Table 2. The range of serum PSA was between 1.34 ng/ml and 100 ng/ml, while the mean was 47.2 ng/ml ± 32.7. The serum PSA was < 4 ng/ml in 2(4.2%) subjects and was between 4 and 10 ng/ml in 2(4.2%) subjects. The mean serum sE-cad was 136.5 ng/ml ± 80.1, while the range was between 0.25 ng/ml and 340.2 ng/ml. The range of the serum IMPDH-2 was between 0.25 pg/ml and 310.4 pg/ml with a mean of 89.8 pg/ml ± 57.4. The box plot of the biomarkers is shown in Fig. 3. The prostate volume ranged from 19.8 ml to 196 ml, with a mean of 64.9 ml ± 38.1. Other findings on TRUS are shown in Table 3. The mean Gleason score was 8.3 ± 1.0 and the majority, 22(45.8%) of the participants belong to ISUP grade group 5. Table 4 shows the Gleason score and ISUP grade group distributions. Serum PSA weakly correlated with Gleason score (r = 0.3, p = 0.04 Spearman rho) and ISUP grade group (τ = 0.3, p = 0.02, Kendall tau). The magnitude of correlation with the Gleason score (R²) was 0.097. Similarly, sE-cad weakly correlated with Gleason score (r = 0.4, p = 0.01, Pearson) and ISUP grade groups (τ = 0.3, p = 0.01, Kendall tau) with a magnitude of correlation (R²) with the Gleason score of 0.134. In contrast, there was neither a statistically significant correlation between serum IMPDH-2 and Gleason score (r = 0.03, p = 0.86, Spearman rho) nor between serum IMPDH-2 and ISUP grade groups (τ = 0.004, p = 0.97, Kendall tau). Figures 4, 5, and 6 show the scatter plot highlighting the correlations. Even though serum sE-cad had the best correlation with the Gleason score it did not significantly outperform serum PSA (p = 0.23) or IMPDH-2 (p = 0.91).

Fig. 1

A histogram showing the age distribution of the study subjects

Table 1 Presenting symptoms of the study subjects

Fig. 2

List of comorbidity among the study subjects

Table 2 Result of urine culture

Fig. 3

Box plot showing levels of serum IMPDH-2 and sE-cad

Table 3 Transrectal ultrasound scan findings

Table 4 Gleason score and ISUP grade groups of the study participants

Fig. 4

Scatter plot for the correlation of serum PSA and Gleason score

Fig. 5

Scatter plot for the correlation of sE-cad and Gleason score

Fig. 6

Scatter plot for the correlation of IMPDH-2 and Gleason score

4 Discussion

Prostate cancer is an important public health concern across the globe, especially among black men. It is the most frequently diagnosed cancer among Nigerian men, constituting 37.5% of newly diagnosed male malignancies in 2022 [1]. Age is a well-known risk factor for PCa and a study of age-specific incidence curves revealed that the risk of PCa starts to rise after the age of 55 years with a peak incidence at the age of 70 to 74 years [20]. Our finding was in keeping with these as the mean age of our study subjects was 69.2 years while the modal age was 70 years. Oyibo et al. [4], Wieczorek et al. [8], and Odubanbjo et al. [21] similarly reported mean ages of 69.3 years, 68.2 years, and 68.5 years respectively while Adewumi et al. [22] reported a lower mean age of 66.2 years. In contrast to this study, Egote et al. [23] in Ghana reported a higher mean age of 71.7 years. All the study subjects were symptomatic and presented mainly with LUTS, while about a quarter of them had paraparesis or paraplegia from spine metastasis. The DRE was suspicious of prostate cancer in nearly all the subjects which was similar to the finding by Oyibo et al. [4] also in Nigeria. These clinical findings reflect how late our PCa patients present which is mostly due to poor awareness of prostate cancer and prostate cancer screening. This is in keeping with what Awosan et al. [24] and Ajape et al. [25] reported on the poor knowledge of PCa and PCa screening among Nigerians. Oyibo et al. [4] and Ekeke et al. [26] in Nigeria also reported LUTS as the most common presenting complaint among their PCa patients. Contrary to our findings, Adewumi et al. [22] in Lagos, reported bone pain as the commonest presentation (35.4%) followed by LUTS in 34.7% while in Egypt, Elabbady et al. [27] reported a quarter of their patients being asymptomatic at presentation and were diagnosed through PSA screening. Furthermore, contrary to our findings, Elabbady et al. [27] in Egypt and Wieczorek et al. [8] in Poland reported lower rates of suspicious DRE of 15.3% and 23.5% respectively. This may be attributed to better awareness of PCa and PCa screening in those countries, hence the earlier diagnosis. About half of our study subjects had comorbidity with hypertension being the commonest, while just three subjects had diabetes which is in keeping with the report by Kasper et al. [28] that diabetes has an inverse relationship with prostate cancer. In Jos, Nigeria, Ofoha et al. [29] similarly reported comorbidity in about half of their PCa patients with the majority also having hypertension. Furthermore, hypertension is the most common comorbidity among PCa patients as reported by Ekeke et al. [26] Navin and Ioffe reported that PCa and hypertension have a common androgen-mediated mechanism [30]. However, they recommended further studies to confirm hypertension as a risk factor for prostate cancer. Less than one-third of our study subjects had a positive urine culture and the predominant microorganism was Escherichia coli. These subjects were mostly asymptomatic and had an indwelling urethral catheter which is in keeping with previous reports that the most common cause of catheter-associated urinary tract infection (CAUTI) is Escherichia coli. [31] Similarly, it has been reported that the daily risk of having bacteriuria in a patient with an indwelling catheter is 3–7% [31]. The mean serum PSA in this study was 47.2 ng/ml, which is lower than the 82.9 ng/ml and 73.4 ng/ml that was reported by Oyibo et al. [4] and Ofoha et al. [29] respectively but higher than the 37.1 ng/ml reported by Egote et al. [23] Wieczorek et al. [8], reported a substantially higher mean PSA of 139.4 ng/ml. This may be due to differences in the assay kits used in the different studies. The sE-cad ranged from 0.25 ng/ml to 340.2 ng/ml, while the mean was 136.5 ng/ml. Tsaur et al. [9] in Germany reported a range of 30 ng/ml to 200 ng/ml and a lower mean of 92 ng/ml. The wide variation in both studies may be attributed to a larger number of subjects having a higher-grade tumor in our study. The mean serum IMPDH-2 was 89.8 pg/ml, which is higher than the 60.5 pg/ml as reported by Wieczorek et al. [8] in Poland. This may be due to racial differences or differences in tumor grade and sample size, with our study having a larger sample size. Three-quarters of our study subjects had a Gleason score of ≥ 8, and belong to ISUP grade groups 4 and 5, which shows the aggressive nature of prostate cancer in our locality. This was similar to the finding by Oyibo et al. [4] Evidence suggests that high-risk prostate cancers are associated with germline mutations that are more prevalent in Asians and blacks [32]. In contrast to our study, Elabbady et al. [27] in Egypt, Wieczorek et al. [8] in Poland, and Tsaur et al. [9] in Germany all reported a Gleason score of ≤ 7 as the most common Gleason score in their respective studies, reflecting the relationship between ethnicity and prostate cancer aggressiveness. The mean prostate volume of our study subjects was 64.9 ml, which is not very different from the 63 ml reported by Elabbady et al. [27] and 67.3 ml reported by Oyibo et al. [4] In contrast, Tsaur et al. [9] in Germany reported a significant lower mean prostate volume of 33 ml, which may be attributed to the earlier presentation of their patients when the tumor is usually of low volume and organ confined. Two-thirds of our study subjects had evidence of a breached prostatic capsule, while less than half of the subjects had seminal vesicle invasion suggesting at least a locally advanced disease. Serum PSA and sE-cad correlated weakly with the Gleason score and ISUP grade group. Oyibo et al. [4] and Okolo et al. [33] also reported a weak correlation between serum PSA and the Gleason score (r = 0.4, Spearman rho). Similarly, in Nigeria, Ngwu et al. [34] reported a moderate correlation of PSA (r = 0.6) with the Gleason score while Mohammed et al. [35] reported no correlation (p = 0.18). Ahmed et al. [36] in Egypt also reported that sE-cad correlated positively with poor Gleason grades (p < 0.05). However, Tsaur et al. [9] in Germany found a very weak negative correlation between sE-cad and Gleason score (r = − 0.2, p = 0.005). The variations seen may be due to racial differences in the expression of the marker, differences in sample size, or perhaps differences in the aggressiveness of prostate cancer as most of their study subjects had a Gleason score of ≤ 7, as against a Gleason score of ≥ 8 in our study. From our study, serum PSA and sE-cad determined 9.7% and 13.4% respectively of the Gleason score variance indicating that other factors are responsible for most of the variance in the Gleason score. We found no correlation between IMPDH-2 and both Gleason score and ISUP grade groups, which is in contrast to a study conducted in Poland by Wieczorek et al. [8] who reported a weak positive correlation between IMPDH-2 and ISUP grade groups (τ = 0.36, p = 0.005, Kendall tau). Differences in the population characteristics, with one being a black population while the other being a European population may be responsible for the variation observed. Furthermore, it may be due to differences in histological characteristics of the tumour with most of our study subjects having a Gleason score of ≥ 8 or differences in the sample size of the two studies. When compared, soluble E-cadherin did not significantly outperform serum PSA or IMPDH-2 in predicting the Gleason score. Therefore, none of the biomarkers can predict prostate cancer aggressiveness with accuracy. Hence, more research needs to be done in this regard.

5 Conclusions

Even though serum sE-cad had the best prediction for Gleason score. It did not statistically substantially outperform serum PSA and IMPDH-2. As such, neither of the biomarkers is a good marker of aggressive prostate cancer.