A prostate biopsy strategy based on a new clinical nomogram reduces the number of biopsy cores required in high-risk patients

Nowadays, prostate-specific antigen (PSA) is the most widely used biomarker for prostate cancer diagnosis. Nevertheless, the specificity of PSA in predicting prostate cancer is not satisfactory [1]. Some evolutional indexes, such as PSA velocity, PSA density (PSAD) and free/total PSA ratio (f/t), are used clinically; however, they are all provincial because of their dependence on PSA [2, 3]. Nomograms, which are based on multiple independent risk factors for prostate cancer, have shown their superiority in detecting this cancer [4]. Although there are two well-known nomograms that have been used for incorporating known risk factors, namely in the Prostate Cancer Prevention Trial (PCPT) and in the European Randomized Study of Screening for Prostate Cancer (ERSPC), previous research has shown that they may not be directly applicable to Chinese males [5–7]. Therefore, it is necessary to develop a new prostate cancer risk assessment nomogram designed for the Chinese population.

The trans-rectal ultrasound (TRUS)-guided sextant biopsy method has been considered a standard method for obtaining prostatic tissues for histological evaluation since it was introduced by Hodge et al. [8] in 1989. Since then, surgeons have used additional cores to improve the accuracy of detection of prostate cancer [9–14]. Obviously, the extraction of additional cores can lead to excessive harm to the patient. According to our biopsy experience, as a result of the lack of PSA screening in most regions of China, a large number of positive cases of prostate cancer involving more than six positive cores have been diagnosed. In the present study, with the help of our nomogram we have attempted to identify cases in which we could reduce the number of biopsy cores required.

Variables used to build the nomogram were selected using a backward elimination scheme; only age, Ln(PSA), Ln(PV), f/t, abnormal DRE and hypoechoic lesions on TRUS were chosen for inclusion in our nomogram (Figure 1). The equation for prostate cancer probability (PCP) was derived using the logistic regression model:

$\mathit{PCP}=\frac{e^{‒3.577+0.054\left(\mathrm{Age}\right)‒3.714\left(\mathrm{f}/\mathrm{t}\right)‒1.324\left(\mathrm{Ln}\left(\mathrm{PV}\right)\right)+0.977\left(\mathrm{Ln}\left(\mathrm{PSA}\right)\right)+1.698\left(\mathrm{DRE}\mathrm{findings}\right)+0.458\left(\mathrm{hypoechoic}\right)}}{1+e^{‒3.577+0.54\left(\mathrm{Age}\right)‒3.714\left(\mathrm{f}/\mathrm{t}\right)‒1.324\left(\mathrm{Ln}\left(\mathrm{PV}\right)\right)+\left(\mathrm{Ln}\left(\mathrm{PSA}\right)\right)+1.698\left(\mathrm{DRE}\mathrm{findings}\right)+0.458\left(\mathrm{hypoechoic}\right)}}.$

The PCP equation was developed to calculate patient risk level regarding the Cacoethic biopsy results. The area under the curve (AUC) of the receiver-operating characteristic curve for the nomogram is shown in Figure 2A. The AUC was increased from 0.761 for PSA alone to 0.853 for our nomogram.

In accordance with the value obtained from the PCP equation, we classified the probability of prostate cancer occurrence into five levels, and the cutoffs were at 0.2, 0.4, 0.6 and 0.8, as each demarcation point declined to a lower level. The rate of positive prostate biopsies increased in line with the risk level (Figure 2B).

In patients with positive biopsy results, the number of positive cores counted and the percentage of low Gleason scores (<7) for every risk level are shown in Figures 2C and D, respectively. Coupled with increasing risk level, the proportion of patients with lower Gleason scores decreased and there was a distinct increase in the mean number of positive cores.

Retrospective analysis of 1104 cases indicated that only a proportion of malignancies could be detected using 12-core rather than 6-core biopsy. When the PCP cutoff value reached 0.5 there was no significant difference in the detection rate between 6-core and 12-core biopsy in patients with higher PCP values (72.9% vs. 78.6%; P = 0.077); however, the difference in the detection rate in patients with lower PCP values was still significant (14.6% vs. 20.4%; P = 0.004). In patients with a PCP >0.5, 6 + 1 core biopsy was found to be more reasonable than 12 + 1-core biopsy.

Measurement of the level of PSA in serum in the elderly can help clinicians to diagnose prostate cancer at an early stage. However, other medical conditions, such as benign prostatic hypertrophy and inflammation, can elevate the PSA level in serum [14]. Because of the low specificity of the PSA level in the detection of prostate cancer, patients suspected of having this disease using PSA screening usually receive an unnecessary biopsy; this is an invasive procedure with accompanying complications [15]. The benefit of PSA screening in prolonging cancer-specific survival remains controversial [16, 17]. Previous research has shown that predictive models, based on clinical, laboratory and ultrasound parameters can improve the accuracy of prostate cancer detection to varying degrees [18–20]. In the present study, Ln(PV), f/t, age, Ln(PSA), the rate of abnormal DRE findings and the rate of hypoechoic masses detected using TRUS have been taken into account for the first time. The equation used in the calculation of the PCP was derived for the diagnosis of prostate cancer. In contrast to PSA alone, use of our nomogram enlarged the AUC from 0.761 to 0.853. As shown in Figure 2B, C and D, we could evaluate the positive core rate, Gleason score and estimate the number of positive cores. In principle, a higher risk level means a higher positive rate, a higher Gleason score and more calculations regarding positive cores. As is evident in Figure 2C, the general tendency regarding the mean number of positive cores is not adequate in the first risk level; however, the difference between the first risk level and the second risk level was not significant (4.32 vs. 3.84; P = 0.473).

More recently, surgeons have used additional biopsy cores to improve the accuracy of detection of prostate cancer [9–14]. However, it has been reported in some studies that there is no significant difference between 6-core biopsy and 12-core biopsy in terms of the rate of positive biopsy cores [21, 22]. According to our nomogram, when the cancer probability cutoff value reached 0.5 there was no significant difference between 6-core biopsy and 12-core biopsy at higher risk levels; however, the difference between 6-core and 12-core biopsy at lower risk levels was significant. The tumor volume ratio of the prostate may explain this. This ratio is smaller at lower risk levels, requiring the use of concentrated biopsy cores. In patients diagnosed with malignancy, the positive rate regarding the 13th core biopsy of hypoechoic masses detected using ultrasound was significantly higher than any biopsy core obtained using systemic 12-core biopsy (70.9% vs. 56.6%; P < 0.001). In summary, taking the threshold value as being a cutoff value of 0.5, 6-core biopsy was found to be adequate for patients at higher risk levels (cutoff >0.5), and 12-core biopsy was found to be adequate for the other patients. If there are hypoechoic lesions on ultrasound, performing an extra core biopsy would be helpful. This new biopsy scheme was found to reduce the number of biopsies required and did not decrease the positive core rate in the second stage of our study.

A nomogram based on data from Chinese males was developed to predict the positive core rate, ratio of positive cores and Gleason score at the various high-risk levels. A reasonable biopsy strategy based on our new nomogram was demonstrated to reduce the number of biopsy cores required in high-risk patients.