Diagnostic performance of transrectal ultrasound elastography integrated with clinical indicators in prostate cancer evaluation

Abstract

Background

Prostate cancer (PCa) is the second most diagnosed type of cancer in men worldwide. Conventional ultrasound-guided systematic prostate biopsies demonstrate limited accuracy and a high false-negative rate, which may increase procedural risks and potential complications. The aim of this study was to evaluate the diagnostic performance of transrectal ultrasound (TRUS) combined with shear wave elastography (SWE) and clinical indicators for the detection of PCa.

Methods

A total of 375 patients with elevated prostate-specific antigen (PSA) levels were enrolled. Patients’ age, prostate stiffness values, PSA density (PSAD), PSA levels, and prostate volume (PV) were analyzed. Diagnostic performance was assessed using receiver operating characteristic (ROC) curve analysis to determine optimal cut-off values and simple models. Additionally, the diagnostic performance of different parameters for PCa was evaluated at varying PSA levels.

Results

In all patients, PSAD demonstrated superior diagnostic performance compared to age, PSA, PV, and SWE alone. Optimal threshold values for PSAD and SWE were 0.62 ng/mL² and 61.07 kPa, respectively. The combination of SWE + PSAD exhibited the highest diagnostic performance [area under the curve (AUC) =0.91, P<0.05]. In patients with PSA levels >20 ng/mL, optimal threshold values for PSAD and SWE were 0.97 ng/mL² and 57.97 kPa, respectively. The combination of SWE + PSAD exhibited the highest diagnostic performance (AUC =0.96, P<0.05).

Conclusions

The cutoff value of SWE is dependent on their serum PSA level, and SWE yields higher diagnostic efficacy at PSA levels exceeding 20 ng/mL. The combination of SWE + PSAD shows better diagnostic efficacy for PCa than other parameters, while integrating imaging with clinical data can further enhance the diagnostic yield.

Highlight box.

Key findings

• The diagnostic performance of transrectal ultrasound shear wave elastography (SWE) plus prostate-specific antigen density (PSAD) for prostate cancer (PCa) was assessed, with findings showing that SWE had higher diagnostic efficacy at prostate-specific antigen (PSA) levels over 20 ng/mL.

What is known and what is new?

• We characterized SWE, PSA, PSAD, and prostate volume in PCa, examined their changes across PSA levels, and compared the diagnostic value of SWE combined with clinical indicators for PCa detection.

• Our findings confirmed that the combination of SWE and PSAD has superior diagnostic utility, delivering enhanced insights to guide targeted clinical treatment strategies for PCa.

What is the implication, and what should change now?

• The pathological results of this study were based on biopsies, which precluded a complete histopathological evaluation of the prostate gland.

• The absence of external validation and the single-center study design have limited the generalizability of our findings, highlighting the need for future multicenter, prospective investigations to verify these results and establish universally applicable cutoff values.

Introduction

Prostate cancer (PCa) is the second most diagnosed type of cancer in men worldwide (1). Clinical screening of PCa is based on abnormal elevation of prostate-specific antigen (PSA) and digital rectal examination (DRE), followed by systematic prostate biopsy in suspected cases (2). However, both PSA and DRE exhibit low sensitivity and specificity in predicting PCa (3). PSA can be elevated not only in PCa but also due to inflammatory prostate lesions, benign prostatic hyperplasia (BPH), and other conditions (4). Additionally, DRE is inherently subjective and limited to evaluating the posterior prostate adjacent to the rectum (5). Currently, prostate biopsy remains the gold standard for the diagnosis of PCa, and the majority of prostate biopsies are performed under ultrasound guidance. Unlike tumors in other major organs, PCa lesions are often insidious and multifocal, with suspicious hypoechoic areas detectable via ultrasound imaging in only 9–53% of cases (6). Moreover, 60–80% of biopsies by TRUS reveal no suspicious regions (7). As a result, conventional ultrasound-guided systematic prostate biopsies demonstrate limited accuracy and a high false-negative rate, which may increase procedural risks and potential complications (8). Therefore, there is a pressing need for comprehensive preoperative evaluation of suspected PCa and the development of noninvasive diagnostic methods to meet clinical and patient expectations.

In recent years, multiparametric magnetic resonance imaging (mp-MRI) has emerged as a critical preoperative tool for prostate biopsy, particularly in detecting clinically significant PCa (csPCa). csPCa was defined according to the criteria recommended by the International Society of Urological Pathology (ISUP), referring to PCa with Gleason score ≥7, or clinical stage ≥ T2c, or tumor volume 0.5 mL, or extracapsular extension. The integration of mp-MRI with real-time ultrasound guidance has improved the detection of csPCa (9). However, studies have shown no significant difference in the overall PCa detection rate between mp-MRI-targeted and TRUS-guided biopsies (10,11). Furthermore, mp-MRI is limited by low reproducibility, specificity, and the inability to provide real-time diagnostic guidance. Additionally, mp-MRI has inherent drawbacks, including high costs, lengthy examination times, and contraindications for patients with metal implants or claustrophobia. These limitations restrict its widespread applicability in clinical practice and patient care.

Currently, the majority of prostate biopsies are conducted under transrectal ultrasound (TRUS) guidance. Compared to normal prostate tissue or benign hyperplasia, PCa lesions exhibit increased stiffness due to abnormal tumor cell proliferation and higher density of trophoblastic blood vessels. TRUS shear wave elastography (SWE), an advanced ultrasound-based technique, evaluates tissue stiffness by measuring shear wave velocity and has emerged as a valuable tool for the initial assessment of PCa (12). However, it remains unclear whether the diagnostic cut-off values of SWE for PCa vary with PSA levels. The optimal SWE threshold for identifying PCa lesions remains under investigation (6).

Previous studies have primarily established SWE cut-off values for PCa based on PSA levels exceeding 4 ng/mL. However, this approach has limitations as some BPHs may exhibit PSA levels above this threshold. Luderer et al. observed that PSA levels increase with age (13). Crawford ED noted that the correlation between PSA and age is attributed partially to the increase in prostate volume (PV) with age (14). PSA density (PSAD) reflects the ratio of serum PSA concentration to PV. Early studies suggested that a PSAD cutoff of 0.15 ng/mL² could improve discrimination of PCa and BPH (15). However, Bare et al. pointed out that using PSAD alone cannot serve as a screening basis for PCa (16). Consequently, this study hypothesizes that establishing SWE cut-off values stratified by PSA and integrating SWE with clinical indicators could improve the diagnostic performance of PCa and subsequently enhance the precision of prostate biopsies. We present this article in accordance with the STARD reporting checklist (available at https://tau.amegroups.com/article/view/10.21037/tau-2025-aw-861/rc).

Methods

This retrospective study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. Approval was granted by the Ethics Committee of The Second Affiliated Hospital of Harbin Medical University (No. 2024-284). Written informed consent was obtained from the patients before biopsy.

Patients

A retrospective analysis was performed on 375 patients with abnormally elevated PSA values with or without dysuria admitted to The Second Affiliated Hospital of Harbin Medical University from January 2024 to June 2025. All patients underwent complete blood count, coagulation function assessment, and serum PSA quantification 1 day before ultrasound-guided prostate biopsy, and were subsequently stratified according to their serum PSA levels: group A, PSA ≤20 ng/mL; group B, PSA >20 ng/mL.

Inclusion criteria were as follows: (I) 50 years ≤ age <80 years; (II) initial biopsy with or without abnormal DRE; (III) no history of prostate surgery or related endocrine therapy; and (IV) complete clinical data with pathological findings.

Exclusion criteria included the following: (I) age >80 years; (II) previous history of prostate surgery or endocrine therapy; (III) incomplete clinical data; and (IV) inability to perform transrectal ultrasonography.

A final sample of 290 patients was included in this research; these participants were randomly allocated to two subgroups (training vs. validation) with the proportional distribution set at 7:3.

TRUS and SWE

The TRUS prostate examination was performed before systematic biopsy using a transrectal endocavity transducer SE 12-3 with a high frequency, employing the Aixplorer ultrasound system (Super-Sonic Imagine, Aix en Provence, France). All patients were asked to place their knees in the left lateral position close to their anterior chest. Imaging was performed along the axial and sagittal planes from the seminal vesicles to the apex of the gland. The entire gland was visualized, and the PV and abnormal echogenicity areas were noted. PV = 0.52 × (transverse diameter × anteroposterior diameter × cephalocaudal diameter). PSAD = PSA/PV.

The SWE-specific mode was initiated. A palette displaying the elasticity map was used, with coding of the soft tissue as blue and the hard tissue as red. Additionally, SWE acquisitions were performed to avoid compression of the prostate with the transducer, ensuring stable SWE data acquisition. As the field of view with SWE was not wide enough to evaluate the entire prostate, the right and left lobes were imaged separately. Images for each region were stored digitally, and the quantitative measurement of stiffness was performed. Any suspicious abnormal areas of echogenicity were covered in the field of view with SWE and measured for Young’s modulus of elasticity. In cases where no abnormal echogenic foci were detected in the prostate, the region of interest (ROI) was set at the maximal transverse section of the gland for the maximum Young’s modulus of elasticity (Emax), and systematic biopsy was conducted accordingly (Figure 1). Emax was finally selected for the study.

Figure 1.

Concomitant transrectal SWE ultrasound image and brightness mode ultrasound image. (A) The maximum cross-section of prostate without suspected lesion. (B) The maximum cross-section of prostate with suspected lesion. SWE, shear wave elastography.

Grayscale and SWE images were acquired by two radiologists with 5 and 10 years of experience in prostate imaging to assess the size of the prostate and perform prostate biopsy. If discrepancies arose between the two radiologists, adjustments were finalized through consensus-based discussion. All radiologists were blinded to pathological findings.

Biopsy and pathology

After performing the TRUS and SWE imaging, the prostate biopsies were performed by the same ultrasound physicians. An 18-gauge, 20-cm puncture biopsy needle and automatic biopsy device (MAGNUM^®, C.R. Bard, Inc., Covington, GA, USA) were used to obtain biopsy cores. The prostate was divided into 12 zones as a template for systematic biopsy, and at least one biopsy from each of the 12 zones was obtained. Additional biopsies were obtained from abnormal areas (well-defined areas of high stiffness, color-coded as red on the machine screen) seen on SWE if the ROI had not already been sampled using the standard 12-core technique.

All pathological diagnoses were determined by two pathologists with more than 15 years of experience in prostate pathology after their consultations. They were blinded to TRUS and SWE results.

Statistical analysis

The pathological diagnoses of prostate biopsies were regarded as the gold standard. Quantitative data are expressed as median (interquartile range). The Mann-Whitney U test was employed to analyze the continuous variables, including patients’ age, SWE, PV, PSA, and PSAD. The receiver operating characteristic (ROC) curve was plotted to evaluate the diagnostic value of SWE, PSA, PSAD, and PV for diagnosing PCa. Apply binary logistic regression to combine variable SWE with each of the variables PSA, PSAD, and PV, respectively, and use ROC curves to calculate their diagnostic efficacy for PCa. The ROC curve was analyzed to calculate the area under the curve (AUC), sensitivity, and specificity. Apply the DeLong test to compare whether there is a statistically significant difference in the area under the ROC curve between the datasets. Besides, SWE, PSA, PSAD, and PV, with high sensitivity and specificity, were selected as the best cut-off values for diagnosing PCa. The optimal cut-off value for each parameter obtained at the Youden’s index was maximum. To assess the diagnostic performance, ROC curves were used for the combination models of SWE and other indicators, as well as the sensitivity and specificity between the groups. P<0.05 was considered statistically significant. The intraclass correlation coefficient (ICC) was calculated to evaluate the inter-observer consistency of SWE measurements.

Results

Basic characteristics of patients

A total of 375 patients were initially enrolled in this study. Of these candidates, 19 were excluded due to the inability to undergo TRUS; 27 declined prostate biopsy; 11 were excluded on account of age exceeding 80 years; 7 had a history of endocrine therapy for prostate-related diseases; and 21 were excluded due to incomplete clinical data. Finally, 290 patients were included in the final analysis (Figure 2). Transrectal biopsies were performed in all subjects. The procedure was well tolerated, with only two patients experiencing major complications (postoperative dysuria necessitating catheterization). There were 171 and 119 cases of prostate hyperplasia (59%) and prostate carcinoma (41%). For the 119 PCa patients enrolled, 238 targeted biopsy procedures were performed, identifying 104 positive cases with a total of 207 positive biopsy tissue cores harvested. Notably, 164 of these positive tissue cores were sampled from areas characterized by elevated SWE values. Table 1 displays the basic characteristics and clinical data of these patients.

Figure 2.

Flowchart of the patient screening. SWE, shear wave elastography; TRUS, transrectal ultrasound.

Table 1. Characteristics of patients.

Characteristics	Total (n=290)	Prostate-cancer (n=119)	Prostate-benign (n=171)	Z value	P value
Age (years)	70.00 (64.50, 74.00)	71.00 (67.00, 75.00)	69.00 (62.00, 73.00)	3.16	<0.05
PSA (ng/mL)	18.69 (11.21, 53.81)	67.77 (19.63, 100.00)	13.53 (10.02, 21.00)	9.38	<0.05
PV (mL)	54.56 (35.44, 78.76)	39.64 (30.58, 56.16)	68.14 (43.16, 95.67)	−6.64	<0.05
PSAD (ng/mL2)	0.36 (0.17,1.09)	1.26 (0.55, 2.18)	0.20 (0.12, 0.37)	11.13	<0.05
SWE (kPa)	48.87 (35.48, 66.99)	67.27 (47.35, 89.93)	43.43 (32.07, 52.10)	8.04	<0.05
csPCa				–	–
Yes	–	96	–
No	–	23	–

Data are presented as median (interquartile range) or number. P value <0.05 as statistically significant. csPCa, clinically significant prostate cancer; PSA, prostate-specific antigen; PSAD, prostate-specific antigen density; PV, prostate volume; SWE, shear wave elastography.

Diagnostic performance of SWE and clinical indicators for discriminating prostate lesions across patient groups

In the overall analysis, statistically significant differences in age, PV, PSA, PSAD, and SWE existed between benign and malignant (P<0.05). To explore the potential associations between age, PV, PSA, PSAD, SWE, and PCa, we performed both univariate and multivariate logistic regression analyses. The results summarized in Table 2 demonstrated that all the aforementioned parameters served as independent predictors of PCa, with statistical significance observed across all variables (all P<0.05). The diagnostic performance of Age, PV, PSA, PSAD, and SWE for prostate was applied in the ROC curve, with the diagnostic performance of PSAD outperforming that of age, PSA, PV, and SWE (Figure 3). The sensitivity and specificity of PSAD were 73.1% and 90.6%, respectively, higher than the other indices: age, PSA, PV, and SWE (Table 3).

Table 2. Univariate and multivariate logistic regression analyses.

Characteristics	Univariate			Multivariate
	OR (95% CI)	P value		OR (95% CI)	P value
Age	1.06 (1.03, 1.10)	0.01		1.08 (1.02, 1.14)	0.01
PSA	1.05 (1.04, 1.07)	<0.001		1.07 (1.03, 1.10)	<0.001
PV	0.97 (0.96, 0.98)	<0.001		0.97 (0.95, 0.98)	<0.001
PSAD	17.73 (8.54, 36.39)	<0.001		0.26 (0.24, 0.29)	<0.001
SWE	1.04 (1.03, 1.06)	<0.001		1.03 (1.02, 1.05)	<0.001

P value <0.05 as statistically significant. CI, confidence interval; OR, odds ratio; PSA, prostate-specific antigen; PSAD, prostate-specific antigen density; PV, prostate volume; SWE, shear wave elastography.

Figure 3.

Diagnostic performance of SWE and clinical indicators for discriminating prostate lesions across patient groups. (A) The total patients; (B) group A; (C) group B. Group A, PSA ≤20 ng/mL; group B, PSA >20 ng/mL. AUC, area under the curve; CI, confidence interval; PSA, prostate-specific antigen; PSAD, prostate-specific antigen density; PV, prostate volume; SWE, shear wave elastography.

Table 3. Diagnostic performance of SWE and clinical indicators for discriminating prostate lesions in total patients.

Characteristics	Cut-off	Sensitivity (95% CI), %	Specificity (95% CI), %	P value	P† value
Age (years)	63	89.9 (84.8–92.1)	28.7 (23.4–33.6)	<0.05	<0.05
PSA (ng/mL)	30.27	69.8 (66.8–77.0)	87.7 (83.3–91.0)	<0.05	0.12
PSAD (ng/mL2)	0.62	73.1 (71.2–80.9)	90.6 (85.9–92.95)	<0.05	<0.05
PV (mL)	58.51	78.2 (76.3–82.1)	63.2 (58.5–69.6)	<0.05	0.29
SWE (kPa)	61.07	61.3 (56.5–67.5)	87.7 (82.75–90.5)	<0.05	–

^†, DeLong test comparing AUROC of SWE with other data. P value <0.05 as statistically significant. AUROC, area under the ROC curve; CI, confidence interval; PSA, prostate-specific antigen; PSAD, prostate-specific antigen density; PV, prostate volume; ROC, receiver operating characteristic; SWE, shear wave elastography.

Internal validation results summarized in Table 4 demonstrated that age, PSA, PV, PSAD, and SWE retained favorable diagnostic performance in the validation cohort, thus verifying the robustness of these indicators for diagnostic application. Age, PSAD, PSA, PV, and SWE threshold values for PCa were calculated from Jordon’s index to be 63 years, 0.62 ng/mL², 30.27 ng/mL, 58.51 mL, and 61.07 kPa, respectively. Intra-observer ICC [95% confidence interval (CI)] of SWE measurements (ICC =0.83) demonstrated excellent agreement.

Table 4. Diagnostic efficacy of SWE and clinical indicators in stratified validation.

Characteristics	Training (n=203)				Validation (n=87)
	AUC (95% CI)	Sensitivity (95% CI), %	Specificity (95% CI), %		AUC (95% CI)	Sensitivity (95% CI), %	Specificity (95% CI), %
Age	0.59 (0.52–0.65)	27.5 (21.4–33.6)	88.2 (83.8–92.7)		0.70 (0.60–0.79)	53.3 (42.9–63.8)	85.0 (77.4–92.6)
PSA	0.81 (0.76–0.87)	69.2 (62.9–75.5)	83.2 (78.1–88.3)		0.89 (0.83–0.96)	83.3 (75.4–91.2)	93.3 (87.9–98.7)
PV	0.25 (0.19–0.31)	1.1 (0.0–2.7)	99.2 (97.8–100.0)		0.32 (0.22–0.42)	0.0 (0.0–1.5)	100.0 (98.5–100.0)
PSAD	0.87 (0.83–0.92)	78.0 (72.4–83.7)	82.4 (77.1–87.6)		0.92 (0.86–0.98)	86.7 (79.4–93.9)	93.3 (87.9–98.7)
SWE	0.76 (0.70–0.82)	67.0 (60.6–73.4)	79.0 (73.4–84.6)		0.85 (0.77–0.92)	73.3 (64.0–82.7)	91.7 (85.7–97.6)

AUC, area under the curve; CI, confidence interval; PSA, prostate-specific antigen; PSAD, prostate-specific antigen density; PV, prostate volume; SWE, shear wave elastography.

Differences in SWE, PSA, PSAD, and PV between benign and malignant prostates across different PSA levels

We divided the patients into two groups based on PSA levels and analyzed the differences in SWE, age, PSAD, and PV between benign and malignant conditions in various PSA groups (Table 5). Within group A, a statistically significant association with malignancy was observed for age, PV, and PSAD (P<0.05); conversely, SWE demonstrated no significant association (P=0.49). In group B, characterized by a high prevalence of csPCa, PSAD, PV, and SWE were all significantly associated with malignancy (P<0.05). By contrast, age showed no significant association (P=0.43). To evaluate diagnostic efficacy, sensitivity, and specificity, ROC analysis was conducted. In group A, PSAD yielded high diagnostic efficacy (AUC =0.77, P<0.05), while SWE did not achieve significance (P>0.05). In group B, both PSAD and SWE emerged as effective discriminators, with high sensitivities and specificities (Table 6). DeLong’s test revealed no statistically significant difference in diagnostic efficacy between SWE and PSAD (P>0.05). In group B, PSAD and SWE threshold values for PCa were calculated from Jordon’s index to be 0.97 ng/mL² and 57.97 kPa, respectively.

Table 5. Characteristics of patients across different PSA levels.

Characteristics	Group A					Group B
	Cancer (n=30)	Benign (n=120)	Z value	P value		Cancer (n=89)	Benign (n=51)	Z value	P value
Age (years)	71.00 (68.00, 74.00)	68.00 (61.00, 73.00)	2.43	<0.05		71.00 (66.00, 75.50)	70.00 (66.00, 75.00)	0.79	0.43
PSAD (ng/mL2)	0.26 (0.19, 0.53)	0.16 (0.10, 0.25)	4.51	<0.05		1.81 (1.09, 2.50)	0.47 (0.27, 0.71)	8.07	<0.05
PV (mL)	35.45 (24.97, 58.51)	66.13 (41.19, 96.84)	−4.57	<0.05		42.09 (32.27, 56.04)	72.24 (44.48, 95.60)	−4.54	<0.05
SWE (kPa)	47.05 (30.50, 56.39)	43.79 (31.22, 52.60)	0.70	0.49		72.60 (58.32, 95.03)	42.43 (35.52, 51.72)	7.11	<0.05
csPCa
Yes	14					82
No	16					7

Data are presented as median (interquartile range) or number. P value <0.05 as statistically significant. Group A, PSA ≤20 ng/mL; group B, PSA >20 ng/mL. csPCa, clinically significant prostate cancer; PSA, prostate-specific antigen; PSAD, prostate-specific antigen density; PV, prostate volume; SWE, shear wave elastography.

Table 6. Diagnostic performance of SWE and clinical indicators for prostate lesions across different PSA levels.

Characteristics	Group A					Group B
	Sensitivity (95% CI), %	Specificity (95% CI), %	P value	P† value		Sensitivity (95% CI), %	Specificity (95% CI), %	P value	P† value
Age	90.0 (87.5–91.3)	38.3 (29.6–40.0)	<0.05	0.23		89.9 (82.8–93.5)	21.6 (15.8–29.5)	0.42	<0.05
PSA	53.3 (54.3–59.6)	64.2 (63.3–64.8)	0.29	0.73		73.0 (67.0–81.4)	92.2 (83.4–93.9)	<0.05	0.50
PSAD	86.7 (80.1–91.3)	59.2 (52.5–68.1)	<0.05	<0.05		82.0 (76.6–89.0)	90.2 (81.2–92.4)	<0.05	0.24
PV	70.0 (68.3–73.9)	74.4 (69.1–76.9)	<0.05	<0.05		82.0 (79.2–85.1)	62.8 (60.2–65.8)	<0.05	<0.05
SWE	16.7 (16.2–23.5)	97.5 (95.5–98.2)	0.51	–		75.3 (68.2–82.4)	88.2 (81.2–92.4)	<0.05	–

^†, DeLong test comparing AUROC of SWE with other data. P value <0.05 as statistically significant. Group A, PSA ≤20 ng/mL; group B, PSA >20 ng/mL. CI, confidence interval; PSA, prostate-specific antigen; PSAD, prostate-specific antigen density; PV, prostate volume; SWE, shear wave elastography.

Diagnostic performance of SWE combined with clinical indicators (age, PSA, PSAD, and PV) for prostates lesions across PSA levels

To evaluate the diagnostic performance of SWE combined with other parameters for PCa, ROC curve analysis was performed in both the overall cohort and across different PSA levels (Figure 3). ROC analysis confirmed the robust diagnostic performance of the SWE + PSAD combination across the cohorts. In the overall cohort, the model’s marginally lower AUC compared to SWE + PSAD + age (0.91 vs. 0.92, P<0.05) was nonetheless accompanied by higher sensitivity and specificity. Its performance in group B was high (AUC =0.96, P<0.05), matching that of SWE + PSAD + age and exceeding all other parameter combinations. Ultimately, DeLong’s test established no significant difference between these two models (Table 7).

Table 7. Diagnostic performance of SWE combined with clinical indicators for prostates lesions across PSA levels.

Variables	Total					Group A					Group B
	Sensitivity, %	Specificity, %	P value	P† value		Sensitivity, %	Specificity, %	P value	P† value		Sensitivity, %	Specificity, %	P value	P† value
SWE + PSAD	79.8	89.5	<0.05	–		83.3	58.3	<0.05	–		85.4	90.2	<0.05	–
SWE + age	61.3	96.5	<0.05	<0.05		70.0	65.8	<0.05	0.17		82.0	90.2	<0.05	<0.05
SWE + PSA	70.6	90.5	<0.05	<0.05		46.7	70.8	0.26	<0.05		78.6	96.1	<0.05	<0.05
SWE + PV	79.8	87.1	<0.05	0.28		90.0	56.7	<0.05	0.4		86.5	86.3	<0.05	<0.05
SWE + PSA + age	73.1	88.3	<0.05	<0.05		76.7	55.8	<0.05	0.11		76.4	96.1	<0.05	0.05
SWE + PSAD + age	74.8	86.6	<0.05	0.47		53.3	90.8	<0.05	0.49		85.4	90.2	<0.05	0.94

^†, DeLong test comparing AUROC of SWE + PSAD with other parameter combinations. P value <0.05 as statistically significant. Group A, PSA ≤20 ng/mL; group B, PSA >20 ng/mL. AUROC, area under the ROC curve; PSA, prostate-specific antigen; PSAD, prostate-specific antigen density; PV, prostate volume; ROC, receiver operating characteristic; SWE, shear wave elastography.

Discussion

Prostate biopsy is still the gold standard for PCa diagnosis. The sensitivity and specificity of conventional systematic biopsy are low due to the invisibility of PCa foci (13). PSA serves as a standard clinical screening marker for PCa, while PSAD has been validated as a diagnostic tool for predicting csPCa and has also demonstrated utility in assessing tumor aggressiveness. Ultrasound shear-wave elastography can be used to identify benign and malignant prostate tissues to improve the diagnostic yield of prostate biopsy. This study aimed to evaluate the diagnostic utility of SWE, PSA, PSAD, and PV in differentiating between benign and malignant prostatic lesions in patients with clinically suspected PCa.

Previous studies have revealed that ultrasound elastography is useful in the diagnosis of benign and malignant prostate conditions (17-19). However, the sensitivity and specificity varied widely, and the range of the floating cut-off values was large. This may be attributed to the inconsistency in the range of subjects studied and the variability in the results of the studies. Tyloch et al. studied the SWE of patients with PCa after prostatectomy. The cut-off value for the PCa was 45 kPa, and the sensitivity and specificity were 63.4% and 73.0%, respectively (17). In a study of SWE and grouping of PCa lesions, Dai et al. demonstrated a significant positive correlation between SWE and grading groups. The SWE threshold for PCa was 84 kPa, and the sensitivity and specificity were 81.3% and 82.4%, respectively (18). In a study of prostate benignity and malignancy, Fu et al. revealed that the cut-off value for assessing SWE in the peripheral zone was 42 kPa, and sensitivity and specificity were 78.9% and 90.7%, respectively (19). This study analyzed the SWE of the prostate in the transitional and peripheral zones and found no substantial differences in the diagnosis of prostate benignity and malignancy. As PCa lesions often occur in the peripheral zone of the prostate, peripheral zone SWE was applied as the study object. The overall suspicious patients were analyzed using ROC curves with a cut-off value of 61.07 kPa for SWE and a sensitivity and specificity of 61.3% and 87.7%, respectively. Moreover, Duffy et al. (20) indicated that due to the low specificity of PSA, approximately 65.0–75.0% of prostate biopsies in patients with PSA levels between 4 and 10 ng/mL yielded negative results. According to the European Association of Urology (EAU) risk group classification, suspected PCa patients with a PSA level ≤20 ng/mL are classified as intermediate-risk and those with a PSA level >20 ng/mL as high-risk (21). Therefore, this study analyzed suspicious patients with PSA levels greater than 20 ng/mL (Table 6), and the cut-off value for SWE remained at 57.97 kPa, with sensitivity and specificity of 75.3% and 88.2%, respectively. A significant diagnostic advantage of SWE for prostate lesions was observed in group B (AUC =0.86, P<0.05) relative to other groups, in contrast to group A, where its performance showed no statistical significance (P=0.51). The high prevalence of PCa, especially PCa, in group B suggests that SWE has high diagnostic efficacy for PCa (Figures 3,4). Notably, the constrained sample size of our subgroup cohorts—especially pathologically confirmed cancer cases in the PSA ≤20 ng/mL stratum—gives rise to statistical uncertainty. The effect size estimates for the PSA ≤20 ng/mL subgroup are accompanied by wide CIs, which preclude the exclusion of potential clinically meaningful benefits or harms associated with SWE in this patient population. We underscore explicitly that the non-significant outcomes observed in the PSA ≤20 ng/mL subgroup must not be overinterpreted as evidence that SWE confers no clinical benefit in this population. Instead, these ostensibly null findings merely denote an absence of conclusive evidence arising from the methodological constraint of limited sample size, rather than definitive proof of a true lack of diagnosis efficacy for SWE in patients with baseline PSA ≤20 ng/mL.

Figure 4.

SWE value differences between benign and malignant patients with PSA levels >20 ng/mL. A 60-year-old patient, with a PSA of 20.38 ng/mL, underwent transrectal prostate biopsy. The pathology confirmed BPH. (A) TRUS. (B) SWE image. ROIs were placed in the color-coded area. SWE value of 43.6 kPa, consistent with BPH. A 68-year-old patient, with a PSA of 22.49 ng/mL, underwent transrectal prostate biopsy. The pathology confirmed GG 3 [Gleason (4+3)] cancer. (C) TRUS. (D) SWE image. ROIs were placed in the red-coded area. SWE value of 91.5 kPa, consistent with PCa. BPH, benign prostatic hyperplasia; GG, grade group; PCa, prostate cancer; PSA, prostate-specific antigen; ROI, region of interest; SWE, shear wave elastography; TRUS, transrectal ultrasound.

This study systematically assessed the diagnostic accuracy of SWE, PSA, PSAD, and PV in discriminating benign and malignant prostate pathologies among patients with suspected PCa. In the initial evaluation of suspicious patients, PSAD demonstrated better diagnostic efficacy. Elevated PSA levels are associated with increasing age, partly due to PV (14). As many factors influence both PSA and PV, Benson et al. proposed that PSAD could be used as a correction index to improve the ability to differentiate between benign and malignant prostate conditions (15). Early studies used a cut-off value of 0.15 ng/mL² for PSAD as a reference for distinguishing between benign and malignant prostate conditions, with values above 0.15 ng/mL² suggesting the possibility of PCa (16,22). This study used a PSAD cut-off value of 0.61 ng/mL² to differentiate between benign and malignant prostate conditions, achieving sensitivity and specificity of 73.1% and 90.6%, respectively. This cut-off value exceeded that of previous studies, possibly because the PVs of all patients in this study were measured under TRUS guidance, which may differ from measurements taken using MRI. However, Choe et al. revealed that no difference exists between PVs measured by TRUS and those measured by MRI (23). Furthermore, the higher PSAD cutoff value observed in this study can be attributed to the larger proportion of csPCa patients, resulting in an improved screening efficacy for csPCa compared to prior studies.

This study analyzed the diagnostic performance of SWE, PSA, PSAD, and PV for benign and malignant prostate conditions and found that PSAD outperformed other parameters in the diagnosis of PCa. Previous studies have confirmed that PSAD is an essential factor in the pathologic grading of PCa and that the correlation between PSAD and PCa is significant (24,25). However, as PCa lesions are insidious and PSAD cannot guide prostate biopsy, we used SWE with PSA, PSAD, and PV in the evaluation of PCa. Our Study has revealed that SWE + PSAD exhibits a higher diagnostic performance, sensitivity, and specificity, particularly in group B (Figure 5). The PSAD suggests the risk of PCa, and SWE guides the biopsy of the prostate through imaging. Furthermore, SWE + PSAD can improve the diagnostic performance of the prostate. Some studies have demonstrated that the Prostate Imaging Reporting and Data System (PI-RADS) with PSAD has been revealed to improve diagnostic performance (26). However, the studies evaluating PSAD were limited, and due to the limitations of MRI and its high diagnostic yield and specificity, the SWE + PSAD combination could be considered a useful tool to optimize biopsy indication in centers without access to MRI or in patients with contraindications to advanced imaging studies.

Figure 5.

ROC curves of the SWE-PSAD combination in different groups. (A) Total group; (B) group A; (C) group B. Group A, PSA ≤20 ng/mL; group B, PSA >20 ng/mL. AUC, area under the curve; CI, confidence interval; PSAD, prostate-specific antigen density; ROC, receiver operating characteristic; SWE, shear wave elastography.

One limitation of this study is that the pathologic results were biopsy-based, which did not allow for a complete histopathological evaluation of the prostate gland. Currently, prostate biopsy is the primary examination method for patients with suspected PCa, and the biopsy results are used to target the treatment plan, whereas short-term observation or secondary biopsy is often used for negatively suspected patients. Secondly, the lack of external validation and the single-center design limit the generalizability of the results. This approach has an inherent risk of “overestimating diagnostic performance” due to the lack of external validation, as the model may be overfitted to the characteristics of the current dataset. Future multicenter and prospective investigations are needed to confirm these findings and establish universally applicable cutoff points. Thirdly, our data preliminarily demonstrate that the clinical benefits of SWE may exhibit heterogeneity based on patients’ baseline PSA levels, and the small sample sizes of the subgroups—most notably the PSA ≤20 ng/mL cohort—pose significant constraints on the reliability and robustness of our conclusions. Future research endeavors should prioritize large-scale, prospectively stratified cohort studies. Such investigations will serve a dual purpose: on the one hand, to validate the favorable efficacy trend of SWE observed in the PSA >20 ng/mL subgroup in our current study; on the other hand, to clarify whether SWE confers definitive clinical value in patients with lower baseline PSA levels, thereby furnishing high-quality evidence-based medicine data to guide the precise clinical implementation of SWE. Our study exclusively enrolled patients with PSA levels >4 ng/mL, and the cohort demonstrated inadequate representation of individuals with PSA levels within the 4–10 ng/mL range. This sampling limitation may potentially introduce selection bias, and we will increase the number of patients in future studies and analyze some prospective studies.

Conclusions

The cutoff value of SWE is dependent on their serum PSA levels, and SWE yields higher diagnostic efficacy at PSA levels exceeding 20 ng/mL. The combination of SWE + PSAD showed better diagnostic efficacy for PCa than other parameters, while integrating imaging with clinical data can further enhance the diagnostic yields.

Supplementary

The article’s supplementary files as

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This retrospective study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. Approval was granted by the Ethics Committee of The Second Affiliated Hospital of Harbin Medical University (No. 2024-284). Written informed consent was obtained from the patients before biopsy.

Data Sharing Statement

Available at https://tau.amegroups.com/article/view/10.21037/tau-2025-aw-861/dss

tau-15-02-45-dss.pdf ^{(43.6KB, pdf)}

DOI: 10.21037/tau-2025-aw-861