The ESR is a predictor of clinically significant prostate cancer in target lesions, and an ESR ≥ 6.8 was presented as the cut-off value. Additionally, an ESR ≥ 4.6 in PI-RADS score 3 lesions on prostate MRI was presented as a predictor of prostate cancer.
Malignant tumors are stiffer than normal tissue. Benign prostatic hyperplasia or normal prostate has a glandular cavity and a homogeneous internal texture, but cancer cells exhibit a stroma reaction in which the normal glandular tissue is destroyed by cancer cell invasion4. In cancer tissues, the density of cancer cells is increased9. Consequently, malignant tumors are stiffer than benign lesions; therefore, DRE is recognized as a screening test for prostate cancer. However, palpation may not be possible depending on the location of the tumor, is highly subjective, and may not be reproducible or representative.
Our study is valuable and scalable, as it is one of the few studies that applied the ESR, which is widely used in other cancers, to the prostate cancer diagnosis field. In breast, thyroid, and pancreatic cancers, elastography is recognized as a promising technique because cancer tissue is stiffer than normal tissue10. Breast cancer has been actively studied for a long time11, and recently, elastography has also been studied in the differentiation of benign and malignant soft tissue tumors12, predicting malignant thyroid nodules13, detecting pancreatic cancer using endoscopic ultrasonography14,15, and assessing liver disease16.
Several studies have demonstrated the diagnostic performance of elastography for prostate cancer. In 2018, Tyloch et al. reported a review of six meta-analyses that evaluated the use of elastography in the diagnosis of prostate cancer17. According to this study, the meta-analysis by Sang et al. showed that the highest diagnostic performance, sensitivity, and specificity were 0.844 (range, 0.696–0.927) and 0.860 (range, 0.792–0.908)2. The diagnostic performance of this previous review is superior to that of our study, especially in terms of sensitivity. However, previous reviews did not stratify the methods between shear wave elastography and real-time elastography. In contrast to our study, which made comparisons per target lesion, previous studies have confirmed the diagnostic performance of elastography with histopathological findings from TRUS random biopsy or radical prostatectomy specimens. To our knowledge, no previous study has confirmed the pathologic result of MRI target lesions and analyzed it using logistic regression. In our study, as a result of logistic regression analyses of an ESR ≥ 6.8 and the conventional variables, we found that an ESR ≥ 6.8 indicated an increased risk for clinically significant prostate cancer.
The ESR is helpful in the differential diagnosis of prostate cancer in PI-RADS score 3 lesions on prostate MRI. Multiparametric MRI is a powerful modality for detecting clinically significant prostate cancer18, and current guidelines recommend pre-biopsy prostate MRI1. This improves the diagnostic performance for clinically significant prostate cancer and reduces the rate of unhelpful prostate biopsies. However, the diagnostic accuracy of PI-RADS score 3 lesions is poor19. Our results showed that an ESR ≥ 4.6 is a predictor of prostate cancer. In clinically significant prostate cancer, the cut-off value of the ESR was ≥ 5.6, which was statistically significant in univariable logistic regression analysis; however, it was not statistically significant in multivariable logistic repression analyses.
Recently, besides ultrasound elastography, MRI elastography has been studied20. Prostate MRI is a powerful diagnostic evaluation tool, such as PI-RADS, whereas MRI elastography plays a supporting role and is not useful as a screening test. A previous study reported that pre-biopsy prostate MRI is only used in 22% of biopsy-naive patients because of its high cost21. Alternatively, the CADMUS trial studied the diagnostic performance of multiparametric ultrasonography. The objective of this trial is agreement in recognization of lesion between multiparametric ultrasound and multiparametric MRI. Multiparametric ultrasound consisted of conventional B-mode images, high frequency or fine flow Doppler images, real-time elastography, and contrast enhanced images21. Yet, these data were also qualitative, such as the PI-RADS. In our previous study, we analyzed grayscale as a quantitation of hypoechoic lesions among qualitative variables (hypoechoicity, irregularity, microcalcification, and vascularity), and reported that quantitative scoring is useful for detecting prostate cancer22. This study allowed us to quantify tissue stiffness.
The lack of previous studies on the ESR in prostate cancer diagnosis is presumably due to the difficulty in selecting a tissue to serve as a reference point. We also considered the urinary bladder, normal prostate tissue, and pelvic floor muscle as reference points. There was concern that the stiffness of the bladder would be measured differently depending on the amount of urine inside the bladder, and prostate tissue thought to be normal was excluded because of the possibility of hidden cancer. Among the pelvic floor muscles, the target lesion and levator ani close to the prostate were set as reference points for easy measurement on one screen.
Our study is valuable in suggesting the cut-off value of the ESR and its diagnostic performance and usefulness. We expect better diagnostic performance when combining the ESR and grayscale of hypoechoic lesions using ultrasonography. However, there are also limitations to pilot studies such as ours. First, depending on the ESR mechanism, different results may be obtained depending on the equipment and physician. We are currently preparing a prospective study to validate our results but have not yet conducted an internal validation. If a large cohort study is conducted in the future, the cut-off value may be different, but a similar trend is expected. Second, we used the levator ani as a reference point for comparison with cancer tissue, but many studies are needed to determine a unified reference point. Third, since our study analyzed specimens from MRI-targeted prostate biopsy, it cannot be free of patient selection bias or operator bias. It is necessary to confirm whether there is a difference with the surgical specimen. In addition, it is necessary to set the target lesion during the screening stage before performing MRI, to collect and analyze data. In patients who have not undergone pre-biopsy prostate MRI, it is difficult to reproduce our method when no target lesions are identified in TRUS. Fourth, tissue stiffness may differ depending on the tumor location and aggressiveness. Owing to the small number of target lesions in our study, further analysis was difficult, and tissue stiffness will need to be verified in future studies. Finally, for a high positive predictive value, we set a cut-off value of high specificity as a predictor of prostate cancer and thus showed low sensitivity. Large data sets and well-controlled prospective multicenter studies are needed to confirm our findings including usefulness of ESR in PI-RADS score 3 lesions and address this issue.
In conclusion, the PSA level and DRE are still recommended as screening tests for prostate cancer and recent guidelines recommend pre-biopsy prostate MRI in patients with suspected prostate cancer. However, MRI is unsuitable for screening due to its high cost and the need for equipment. We evaluated the ESR of the target lesion in MRI-targeted prostate biopsy and confirmed its potential as a complementary diagnostic tool. This retrospective pilot study had some limitations, and additional prospective studies in pre-biopsy prostate MRI setting are needed to evaluate the potential role of the ESR. Currently, it is difficult to apply in a clinical setting without pre-biopsy prostate MRI, but TRUS using elastography combined with the grayscale of hypoechoic lesions is expected to be a better screening test for prostate cancer.
