The role of endogenous testosterone in relationship with low- and intermediate-risk prostate cancer: a systematic review

Abstract

An enduring debate in research revolves around the association between elevated endogenous testosterone levels and prostate cancer. This systematic review is intended to assess the present understanding of the role of endogenous testosterone in the diagnosis and treatment of low- and intermediate-risk prostate cancer. Our search strategy was the following: (endogenous testosterone) AND (((low risk) OR (intermediate risk)) AND ((diagnosis) OR (treatment))) AND (prostate cancer); that was applied to PubMed, Web of Science, and Scopus databases to identify pertinent articles. Two investigators performed an independent selection following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. The preliminary investigation detected 105 records, and 81 records remained after eliminating duplicates. Following the review of titles and abstracts, 71 articles were excluded. A comprehensive examination of the full text was conducted for 10 articles, excluding 3 of them. After revising the references of eligible articles, other 3 articles were included. We finally identified 10 suitable studies, including three main topics: (1) association between endogenous testosterone and European Association of Urology (EAU) risk classes; (2) association between endogenous testosterone density and the tumor load; and (3) association of endogenous testosterone with tumor upgrading and tumor upstaging. Actual literature about the impact of endogenous testosterone on low- and intermediate-risk prostate cancer is not numerous, but appears to be still conflicting. More investigations are needed to increase the consistency of the literature’s results.

INTRODUCTION

Based on the guidelines set forth by the European Association of Urology (EAU), prostate cancer (PCa) ranks as the second most frequently diagnosed cancer in men, with a prevalence of 59% among those aged over 79 years, presenting a significant health concern for elderly males.1 Following diagnosis, patients are classified into prognostic groups (low-, intermediate-, and high-risk categories) based on D’Amico’s classification, which considers prostate-specific antigen (PSA) levels, International Society of Urological Pathology (ISUP) grade group, and clinical T from tumor, node, metastasis (TNM) classification, as suggested by guidelines. Various therapeutic choices are available for different risk categories, ranging from active surveillance (AS) to active interventions such as radical prostatectomy (RP) potentially coupled with extended pelvic lymph node dissection (ePLND), radiation therapy (RT), or brachytherapy.1,2 The precise reasons behind the development and advancement of PCa remain unknown.3 An enduring debate in research revolves around the association between elevated endogenous testosterone (ET) levels and PCa, with evidence indicating that this association could be negative, positive, or null.4,5 ET plays a crucial role in assessing conditions related to prostate growth, such as benign prostatic hyperplasia (BPH) or PCa.1 Numerous studies suggest that testosterone serves as a contributing factor promoting PCa in experimental research.3,6 The role of ET in inducing the development of PCa in humans is not clear6 as well as its contribution to determining the risk class of PCa and its implications in clinical practice. This systematic review is intended to assess the present understanding of the role of ET in the diagnosis and treatment of low- and intermediate-risk PCa.

The methodology of this systematic review and statistical analysis followed the checklist of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines (PRISMA).7 In September 2023, we explored PubMed, Web of Science, and Scopus databases to find articles examining the role of ET in the diagnosis and therapy of low- and intermediate-risk PCa. The investigation strategy was the following: (endogenous testosterone) AND (((low risk) OR (intermediate risk)) AND ((diagnosis) OR (treatment))) AND (prostate cancer). No restriction in terms of the publishing date was set to conduct the investigation. Two researchers (ES and DB) conducted an initial screening independently by reviewing titles and abstracts and documenting reasons for excluding ineligible reports. Subsequently, full texts were obtained and assessed for eligibility. Additionally, the reference lists of pertinent articles were scrutinized to identify any potential additional studies. Any disagreements were resolved through consensus with the coauthors (ABP, ES, DB, SC, and CB).

We have included all the original articles that evaluated the role of ET in diagnostic and therapeutic evaluation in a population of patients that included low- and intermediate-risk PCa. The articles included are randomized trials and retrospective studies. We excluded nonrelevant articles, reviews, book chapters, case reports, articles sharing the same studied population, and non-English articles.

Two authors (ES and DB) independently gathered data concerning studies, patients, and treatment demographics. The detailed obtained contents were as follows: authors’ names, study design, year of publication, number of patients, follow-up, inclusion criteria of the study, whether endogenous was evaluated as total of free testosterone or as a surrogate (e.g., endogenous testosterone density [ETD]), the type of PCa evaluated (grade and stage when available), outcome evaluated in the study, a summary of the conclusions of the study, and whether the findings of the study were statistically significant or not. Disagreements were debated with coauthors (ABP, ES, DB, SC, and CB) and resolved through consensus.

STUDY SELECTION

Our preliminary search yielded 105 records, with 81 remaining after eliminating duplicates. Subsequently, 71 articles were excluded following the screening of titles and abstracts. A full-text review was conducted for 10 articles, resulting in the exclusion of 3 of them. After reviewing the references cited by the articles considered eligible in the review process up to this point, an additional 3 articles were included. In accordance with our inclusion criteria, we ultimately identified 10 studies that were included in the review.

The selection process is reported in the PRISMA flowchart (Figure 1). Among all the studies included, a total of 7 studies were published by our research group at the Integrated University Hospital (Verona, Italy). We identified three main topics: (1) studies investigating the association between ET levels and EAU risk classes; (2) studies investigating the association between ETD and the tumor load in the prostate gland; and (3) studies investigating the association of ET levels with tumor upgrading and tumor upstaging. The main features of the articles included in the review are listed in Table 1.

Figure 1.

The Preferred Reporting Items for Systematic Reviews and Meta-Analyses flowchart for identification and selection of studies included.

Table 1.

Features of the 10 included studies

Study	Study design	Number of participants	Follow-up	Inclusion criteria	TT or fT	PCa	Outcome measurement/results	Conclusion	Significance
Watts et al.8 2018	Prospective	6933	Mean: 6.8 years	Patients data available from 20 prospective studies including patients with prediagnostic data on testosterone and SHBG	fT (categorized in tenths)	Any	Patients in the lowest tenth of fT had a lower risk of PCa than patients in any other tenth of the distribution (OR: 0.79; 95% CI: 0.71–0.87)	Reduced risk of PCa associated with very low concentrations of fT	SS
Tafuri et al.9 2020	Retrospective	601	Not defined	Men not under ADT and elected for RP	TT	Any	Basal median TT levels progressively increased along EAU risk classes (P=0.019); trend lost for high-risk to locally advanced risk PCa. Patients with TT ≥403 ng dl−1 have an increased risk of higher PCa risk class	Basal TT levels: positively associated with increasing EAU risk category. Higher TT levels: associated with more aggressive PCa. Associations are lost when considering locally advanced PCa	SS
Porcaro et al.11 2021	Retrospective	430	Median: 71 months	Men not under androgen blockade, and with IR PCa	TT (evaluated as ETD=TT/PV)	IR	Multivariate analysis: association between ETD and TLD (b=0.032, 95% CI: 0.023–0.040, P<0.0001)	In IR, as ETD increased, TLD and TL at surgical pathology increased. Patients with lower ET were more likely to have occult UD at surgical pathology	SS
Porcaro et al.12 2022	Retrospective	338	Not defined	Men not under androgen blockade, and with intermediate risk	TT	IR	ETD was an independent predictor (b=0.026, 95% CI: 0.016–0.036; P<0.0001) of BPCD in the multivariate linear regression model	In IR, as ETD increases, the risk of large tumors extending beyond the prostate increases. Patients with lower ET were more likely to occult UD	SS
Porcaro et al.13 2022	Retrospective	433	Median (IQR): 42 (23–57) months	Men who were not under androgen blockade with impalpable organ-confined disease and PSA ≤10 ng ml−1 treated with RP	TT	Not palpable PCa with PSA levels ≤10 ng ml−1	Univariate analysis: TLD positively correlated with ETD (Pearson’s correlation coefficient [r]=0.263; P<0.0001) Multivariate analysis: disease progression was independently predicted by ETD (HR: 1.037; 95% CI: 1.004–1.072; P=0.030) If ETD stratified by quartiles, disease progression occurred for levels above the third quartile (HR=2.780; 95% CI: 1.293–5.975; P=0.009)	ETD was an independent predictor of disease progression with risk increasing as ETD increased ET levels were significantly lower for progressing cancers. Increased ETD and low ET levels, indicating androgen independence, resulted in a more aggressive disease with a poorer prognosis	SS
Porcaro et al.14 2021	Retrospective	172	Not defined	Low-risk PCa patients not under androgen blockade therapy before surgery	TT	Low-risk	BPCD biopsy positive core density is strongly correlated to TL and to TT P<0.0001, but aggressive tumors are associated with lower TT	As TT increased, higher TLs were assessed; however, in upgraded patients lower TT levels were also detected	SS
Muller et al.10 2012	Prospective	3255	Not defined	Men not under androgen blockade who underwent prostate biopsy with a PSA of 2.5–10 ng ml−1 and prior negative biopsy	TT	Any	Secondary analysis (OR: 1.23, 95% CI: 1.06–1.43, P=0.006)	Baseline serum testosterone and DHT levels were unrelated to PCa detection or grade. The lowest testosterone levels were associated with the lowest PCa risk with no further changes with higher testosterone. In a secondary analysis, higher testosterone levels at baseline were associated with higher PCa detection only if men had low baseline testosterone	SS
Porcaro et al.15 2019	Retrospective	144	Not defined	Low-risk PCa men not under ADT elected to undergo RP with or without ePLND	TT (evaluated as TTD=TT/PV)	Low-risk	BPC, PSAD, and TTD were independently associated with the risk of high TL with ROC curves indicating the same discriminant power (for TTD: AUC=0.610, P=0.032)	In patients with low-risk PCa who underwent RP, TTD is associated with the risk of high TL, which predicts the risk of UD including upgrading, upstaging, or positive surgical margins. At each level of TTD, mean TT levels were significantly lower in patients with high TL	SS
Porcaro et al.16 2017	Retrospective	135	Not defined	Low-risk and IR patients elected to undergo RP with or without ePLND	TT	Low-risk and IR	On multivariate analysis, TT was an independent factor in predicting tumor upgrading (OR: 1.005, P=0.015)	Low-risk and IR PCa patients show a not negligible risk of tumor upgrading to high-risk disease. In this particular subset of patients, basal levels of TT stratify the risk of tumor upgrading	SS
Ferro et al.17 2017	Retrospective	338	Not defined	Patients with low-risk PCa are eligible for AS but opting for RP	TT	Low-risk	Low TT levels (<300 ng dl−1) were significantly associated with upgrading (P<0.001), upstaging (P<0.001), UD (P<0.001), and positive surgical margins (P=0.035)	TT should be a selection criterion for the inclusion of low-risk patients in AS programs. TT <300 ng dl−1 should be considered a discouraging factor when a close AS program is considered as a treatment option	SS

ADT: androgen deprivation therapy; AS: active surveillance; BPC: biopsy-positive cores; BPCD: BPC density; DHT: dihydrotestosterone; ePLND: extended pelvic lymph node dissection; ET: endogenous testosterone; ETD: ET density; fT: free testosterone; HR: hazard ratio; IR: intermediate-risk; OR: odds ratio; PCa: prostate cancer; PSA: prostate-specific antigen; PSAD: PSA density; RP: radical prostatectomy; SHBG: sex hormone binding globuline; TL: tumor load; TLD: TL density; TT: total testosterone; TTD: TT density; UD: unfavorable disease; AUC: area under the curve; PV: prostate volume; CI: confidence interval; NS: not significant; SS: statistically significant; EAU: European Association of Urology; IQR: interquartile range; ROC: receiver-operating characteristic; b: linear regression coefficient

ASSOCIATION BETWEEN ENDOGENOUS TESTOSTERONE AND EAU RISK CLASSES

The relationship between endogenous testosterone and the EAU risk categories has been investigated in 3 out of 10 studies8,9,10 included in the review. Watts et al.8 calculated free testosterone (fT) concentrations from those of total testosterone (TT) and sex hormone binding globulin (SHBG), and they categorized fT concentrations into study-specific tenths. According to their results, individuals within the lowest study-specific tenth had a decrease of PCa risk of 23% compared to those having higher levels of fT concentrations. However, above this extremely low concentration of fT, the risk of PCa remained unchanged with increasing fT levels. Notably, the authors found heterogeneity in the association of fT with PCa risk by tumor histological grade: lower fT concentrations were correlated with a reduced risk of low- to intermediate-grade PCa, while there was a nonsignificant increase in the risk of high-grade disease.

Tafuri et al.,9 in a retrospective study involving 601 patients elected to RP and not under androgen deprivation therapy, reported a positive progressive linear trended increment of basal median TT levels across EAU risk categories (P = 0.019). When TT was classified based on its median value (< or ≥ 403 ng dl⁻¹), it appeared to be associated with the high-risk PCa class compared to the low-risk class (odds ratio [OR]: 2.309, 95% confidence interval [CI]: 1.309–4.072, P = 0.004). However, no statistically significant association was observed when comparing the intermediate-risk and locally advanced classes with the low-risk class (P = 0.066 and P = 0.091, respectively).

Muller et al.10 investigated whether androgens were associated with PCa risk in the placebo arm of the Reduction by Dutasteride of Prostate Cancer Events (REDUCE) study. They found that regardless of Gleason scores, TT and dihydrotestosterone (DHT) levels were statistically similar among men with PCa (both P ≥ 0.52). They stratified TT and DHT in quintiles, not finding association between almost all quintiles with risk of low- or high-grade PCa in multivariate analysis (all P ≥ 0.11). The only exception was the second quintile of DHT, which appeared to be associated with lower risk of low-grade PCa (OR: 0.74; 95% CI: 0.55–0.98; P = 0.04). Androgens were also not associated with low- and high-grade PCa both when tested continuously and when tested as a P trend across quintiles (all P ≥ 0.28; the authors used a nonparametric test for trends derived from Wilcoxon rank-sum test and referred to that P-value as P trend).

Interestingly, separating those patients with the lowest baseline testosterone levels from those who had already low baseline levels, it was possible to notice that this selected population had the lowest PCa risk. This risk increased progressively as baseline testosterone levels neared normal levels; after that point, PCa risk did not further increase regardless of higher testosterone levels.10

In a secondary examination, among males with baseline testosterone ≤10 nmol l⁻¹, more elevated testosterone levels were associated with higher PCa risk (OR: 1.23; 95% CI: 1.06–1.43; P = 0.006); instead, testosterone was unrelated to PCa risk in men with normal testosterone levels (>10 nmol l⁻¹; P = 0.33).10

ASSOCIATION BETWEEN ETD AND TUMOR LOAD

The association linking ETD and tumor load has been investigated in 5 out of 10 studies11,12,13,14,15 included in the review. All studies investigating this association came from the same author. The definition of ETD is the ratio of ET to prostate volume. Other abbreviations appear in the evaluated papers and are here explained. Tumor load (TL) was defined as the percentage of cancer-occupied volume of the prostate in the final pathology report, and its ratio to prostate volume was defined as tumor load density (TLD). The ratio of the percentage of positive cores at biopsy to the prostate volume, which describes the tumor load density at biopsy, was named biopsy positive core density (BPCD).

In a study involving 430 intermediate-risk PCa patients, Porcaro et al.11 observed that as BPCD increased, both ETD and prostate-specific antigen density (PSAD) increased accordingly. However, patients occulting unfavorable disease (UD) exhibited notably lower mean levels of either TT or PSA when related to respective densities.

In another study based on 338 intermediate-risk patients, they documented that ETD was associated with UD and TLD.12 ETD showed a positive correlation with pathological characteristics of aggressive cancer. In the multivariate linear regression analysis, ETD emerged as an independent predictor of BPCD (linear regression coefficient [b] = 0.026, 95% CI: 0.016–0.036; P < 0.0001).

Individuals with UD exhibited significantly higher mean ETD values in comparison to the control group. Patients with UD (ISUP >3 and/or pT >2 and/or pN1) demonstrated notably lower mean increments of endogenous testosterone compared to the control group with favorable disease. Another study by the same authors based on 805 patients presenting with impalpable organ-confined disease and PSA levels ≤10 ng ml⁻¹ undergoing RP, showed that on univariate analysis, TLD exhibited a positive correlation with ETD (Pearson’s correlation coefficient [r] = 0.263; P < 0.0001).13 Even after adjusting for other variables, ETD was still documented to be a positive predictor for tumor quantitation density features. Furthermore, on multivariate analysis, it independently predicted the progression of the disease (hazard ratio [HR]: 1.037; 95% CI: 1.004–1.072; P = 0.030). Furthermore, this study stratified ETD by quartiles, considering the first quartile as the reference. Disease progression was noticed for levels above the third quartile (HR: 2.780; 95% CI: 1.293–5.975; P = 0.009), but not for lower levels. This study included all PCa risk classes, but most of the patients were low- and intermediate-risk patients (39.7% and 54.0%, respectively). Porcaro et al.14 enrolled 172 low-risk patients undergoing RP after the biopsies, and they found that BPCD is associated with ETD (r = 0.509; P < 0.0001). In a different study based on 144 low-risk PCa patients undergoing RP, Porcaro et al.15 showed that ETD is an independent risk factor for high TL. The receiver-operating characteristics (ROC) curves revealed an area under the curve (AUC) of 0.610 (P = 0.032).

ASSOCIATION OF ET LEVELS WITH TUMOR UPGRADING AND TUMOR UPSTAGING

The association between ET and tumor upgrading/upstaging has been investigated in 3 out of 10 studies14,16,17 included in the review. In a study based on 135 low-risk and intermediate-risk patients elected to undergo RP with or without ePLND, Porcaro et al.16 discovered a significant association between TT and tumor upgrading, noting that upgraded tumors exhibited higher median levels of TT compared to nonupgraded cancers. TT was arbitrarily dichotomized using the third quartile (Q3 = 426.0 ng dl⁻¹) as the cut-off. TT > Q3 independently predicted tumor upgrading (OR: 6.577; P = 0.010). In the population of patients with low- and intermediate-risk PCa having ISUP 1 or 2 cancers, researchers noted that the basal levels of both PSA and TT were positively correlated with the upgrading of the tumor to high-risk PCa in the surgical specimen.

Ferro et al.17 in a study based on 338 patients with low-risk PCa eligible for AS but opting for RP, found that individuals who experienced reclassification after prostatectomy had notably lower serum TT levels in comparison to those who did not undergo reclassification (P < 0.001). The study identified specific testosterone thresholds for upstaging (344 ng dl⁻¹), upgrading (431 ng dl⁻¹), unfavorable disease (302 ng dl⁻¹), and predominant Gleason score 4 (315 ng dl⁻¹). Patients with TT levels below these limits were at a heightened risk of reclassification of the disease. Additionally, utilizing the hypogonadism cut-off of <300 ng dl⁻¹, it was noted that patients experiencing upstaging, upgrading, unfavorable disease, positive surgical margins, and exhibiting a predominant Gleason score 4 were significantly more prevalent compared to eugonadal patients. In a low-risk cohort of patients, Porcaro et al.14 found that ETD predicts tumor upgrading (b = 0.032; 95% CI: 0.021–0.043; P < 0.0001). Notably, aggressive tumors were linked to lower levels of TT.

DISCUSSION

The association between levels of ET and PCa risk is an enduring and controversial debate of urological research, with evidence indicating that this association could be either negative (which is the most frequent occurrence), positive, or null.4,5 By performing our research, three main themes of investigations emerged concerning the role of ET in predicting the PCa risk class, the tumor load in the prostate gland, and the risk of upgrading/upstaging of PCa.

Concerning the association between ET and EAU risk classes, a substantial disagreement among authors emerged. Watts et al.8 and Muller et al.10 did not find a significant constant association between increasing levels of testosterone and increased risk of PCa. Nevertheless, they noticed that variations occurring only at very low levels of testosterone seemed to vary the risk of PCa. This risk increased till a certain cut-off and then did not vary anymore for further increase in testosterone.

This phenomenon appears to be explained by the “saturation model”, which suggests that the growth of the prostate is very sensitive to fluctuations in androgen concentration levels that occur below a specific point of receptor saturation, which is near or at castration levels.18,19 Once this threshold is reached and androgen receptors (AR) are saturated, any subsequent increase has no additional stimulatory effect on the tumor. Since this point of saturation is considerably lower than the normal range of serum testosterone, variations occurring in the physiologic range are not expected to impact the growth of prostate tumor. Other clinical studies have supported this hypothesis.20,21

On the other hand, Tafuri et al.9 found that TT consistently increased along EAU risk categories and appeared to be associated with the high-risk PCa category compared to the low-risk category, although no significant association was found comparing locally advanced disease and intermediate-risk disease to low-risk disease.

We suggest that a possible explanation of this phenomenon, in contrast with the “saturation model”, consists of a step-wise process. Theoretically, when there is a sudden decline in serum testosterone in an aging male, regardless of the cause,22 local autocrine and paracrine mechanisms come into play to maintain periprostatic testosterone concentrations. This involves increased production of testosterone and expression of AR. Consequently, there is an overall hyperstimulation of luminal glandular cells, even in the face of declining serum testosterone levels. This results in DNA damage and uncontrolled luminal cell AR-driven proliferation.23 As a consequence, the progressive selection of more aggressive prostatic cellular clones can lead to the development and growth of cancer. This can progress to the point where the cancer spreads beyond the prostate capsule to the lymph nodes, and ultimately, it loses sensitivity to hormones becoming castrate resistant.24,25

Concerning studies documenting the association between ETD and TL, there is substantial agreement on the fact that ETD is associated with an increased tumor load, both intended as TLD and as BPCD. It is to mention that those studies were conducted by the same authors. This association was valid for low-risk disease,14,15 as well as for intermediate-risk PCa.11,12 In those works, the authors also described that for the same levels of ETD, those patients who were harboring unfavorable disease had lower levels of ET.11,12,14

The PCa grade does not change significantly over time,26 hinting that high-grade tumors likely originate de novo rather than from the dedifferentiation of low- to intermediate-grade tumors. The binding of androgens to androgen receptors in the prostate plays a crucial role in regulating cell differentiation.27 Therefore, prostate cells that have lower exposure to androgen may exhibit less differentiation and a higher likelihood of developing high-grade tumors.28 Alternatively, this could represent a unique growth response of early low-grade cancerous lesions to an environment depleted of androgens.

Regarding studies examining the relationship between ET and PCa upgrading and upstaging, conflicting results emerged. Porcaro et al.16 documented that those tumors that upgraded exhibited higher median levels of total testosterone compared to nonupgraded PCa. Moreover, they found that among individuals with low- and intermediate-risk PCa with ISUP 1 or 2, tumor upgrading correlated with baseline levels of TT in an independent and positive manner. Conversely, Ferro et al.17 discovered that low total testosterone levels predicted PCa upstaging and upgrading in low-risk patients meeting AS inclusion criteria. Ferro’s findings support the notion that individuals with low testosterone levels are more prone to developing aggressive PCa, which has been reported also by our research group.14

CONCLUSIONS

Actual literature about the impact of endogenous testosterone on low- and intermediate-risk PCa is not numerous but still conflicting. This is especially true when it concerns tumor upgrading and upstaging, and the association of testosterone with EAU risk classes. Endogenous testosterone, considered as serum testosterone or as its surrogates (e.g., ETD), surely plays a fundamental role in the natural history of PCa and appears to be a promising tool in helping the clinician define the risk of the patient’s disease. More investigations are needed to increase the consistency of the literature’s results, particularly when considering low- and intermediate-risk disease.

AUTHOR CONTRIBUTIONS

ABP, ES, and DB are the project managers and conceived the study. ABP validated the study. ES and DB performed data collection. ES, DB, SC, and CB wrote and reviewed the manuscript. ABP, MAC, and AA supervised the study. All authors read and approved the final manuscript.

COMPETING INTERESTS

All authors declare no competing interests.