Transperineal biopsy for diagnosing prostate cancer

3.2 The evidence on cancer detection rates of the different prostate biopsy approaches was limited. The clinical-effectiveness review included 23 studies and the strength of evidence was mixed. Six studies were randomised controlled trials (RCTs), but most were observational. Six studies were only available as conference abstracts and 1 (Bojin 2019) was an unpublished slide set. The external assessment group (EAG) said that there was a high risk of reporting bias in these studies because of the limited information that they included. Most studies did not report whether a prebiopsy multiparametric MRI (mpMRI) had been done and some did not report the number of biopsy cores taken. In studies that did report the number of cores, these varied from around 12 to 24 cores. The committee noted that studies that used an mpMRI image to take targeted biopsy samples and those that took more cores may result in higher cancer detection rates regardless of the biopsy technique used. Eight of the studies used the PrecisionPoint device, 4 used a coaxial needle and 1 used a grid and stepper (for LATP compared with general anaesthetic transperineal [GATP] biopsy). There were 4 single-arm studies (1 on CamPROBE and 3 on the UA1232 device). The rest did not report what device was used. There was no comparative evidence on the CamPROBE, EZU‑PA3U, UA1232 or Trinity Perine Grid devices. The committee noted that of the 23 included studies, a single RCT available as a conference abstract was the only study used in the network meta-analysis of cancer detection rates for LATP using a freehand needle positioning device. The committee said that there was limited evidence on cancer detection rates and that caution should be used when interpreting the results. It also noted that the ongoing RCT of transrectal compared with LATP biopsy (TRANSLATE; see section 3.8) will provide further evidence on detection rates of clinically significant prostate cancer. The committee concluded that, because there was generally no significant difference between LATP using any method, LATP using a freehand needle positioning device, local anaesthetic transrectal ultrasound (LA‑TRUS) or GATP, it could not say if one technique was better than the others.

3.3 Clinical experts said that the EZU‑PA3U, PrecisionPoint, Trinity Perine Grid and UA1232 freehand needle positioning devices are mechanical devices that, despite some technical differences, all work in a similar way for the user. They attach to the ultrasound probe and align the needle with the probe axis, keeping them in line during the procedure. So, the biopsy technique used with them is considered to be the same. The needles the devices use are indicated for taking biopsy samples through the skin. A clinical expert said that when they moved to using the PrecisionPoint device, there was no difference in the biopsy sample quality. Although the PrecisionPoint device comes with a specific non-coring access needle as part of the kit, the companies for EZU‑PA3U, Trinity Perine Grid and UA1232 said that their devices could be used with needles bought separately provided they are compatible with the guide channel diameter. Use of an access needle or a coaxial needle means that generally 4 or fewer punctures of the skin are needed. Other differences between the devices are that:

3.4 All the comparative clinical evidence for freehand needle positioning devices was on the PrecisionPoint device. However, a clinical expert explained that the different freehand devices all work in a similar way with the same biopsy technique (see section 3.3). The clinical experts said they would not expect significant differences in cancer detection rates and adverse event rates between the devices. No studies directly compared the individual devices, so there was no evidence that one performs better than any other. The committee concluded that the clinical effectiveness in terms of cancer detection and adverse events was likely to be similar for all the freehand needle positioning devices, although this was uncertain.

3.5 A clinical expert explained that double freehand prostate biopsy approaches should be considered separately to the freehand approach that uses a needle positioning guide. They noted that for double freehand, a needle is used without a positioning guide, so 1 hand guides the ultrasound probe while the other guides the needle, and they need to be kept in line manually. CamPROBE is a double freehand device, so it is not attached to the ultrasound probe. There was no evidence comparing CamPROBE with any of the freehand needle positioning devices. The committee concluded that, because of these differences and the lack of evidence, it was uncertain if CamPROBE would have the same clinical effectiveness in terms of cancer detection and adverse events as the freehand needle positioning devices. Comparative studies of the CamPROBE biopsy device are needed to assess its clinical effectiveness (see section 4.1).

3.6 Clinical experts said that sepsis can happen after a prostate biopsy. It is rare but serious, and can result in death. In the EAG's clinical-effectiveness review, relatively few studies reported post-biopsy sepsis. In the studies that did report sepsis, it only occurred after LA‑TRUS biopsy and not after transperineal biopsy. An analysis of recent hospital episode statistics data (from 2017 to 2019) by Tamhankar et al. (2020) showed that there was a difference in rates of infection and sepsis between TRUS biopsies and transperineal biopsies. Rates of infection were 1.50% in people who had a TRUS biopsy and 0.67% for a transperineal biopsy. Similarly, rates of sepsis were higher for TRUS biopsies (1.12%) than for transperineal biopsies (0.42%). The committee concluded that LATP biopsies may reduce the risk of infection and sepsis compared with TRUS biopsies.

3.7 Clinical experts explained the key differences between LA‑TRUS and LATP biopsy approaches. An LA‑TRUS biopsy tends to take fewer cores (usually 12), whereas centres that use the Ginsburg protocol for LATP may take 24 cores or more (2 LATP protocols are used in the UK: RAPID and Ginsburg). However, some centres may also take 12 cores for LATP plus additional targeted cores based on mpMRI results. A clinical expert said that mpMRI may identify anterior lesions of the prostate and these can be more difficult to reach using LA‑TRUS biopsy than LATP. However, the committee heard that there was no evidence to assess the clinical effectiveness of the different biopsy approaches in people with anterior lesions. Clinical experts said that LATP may be less tolerable because of the lithotomy position, and if numerous skin punctures were needed, when using a grid and stepper for example. LATP biopsies also take slightly longer than TRUS biopsies, particularly when clinicians are training or first start using the technique. However, a clinical expert said that when practitioners are trained equally in the techniques, the difference is minimal. Less experienced clinicians may find it easier to use a grid and stepper. Clinical experts explained that, in a minority of cases, LATP might be contraindicated, such as in people who have had gender reassignment. There are also some patient groups with a higher risk of infection (for example, immunocompromised people) who would prefer LATP because of the lower risk of sepsis with LATP than LA‑TRUS. Clinical experts explained that there is a move towards using LATP nationally and that some centres no longer do TRUS prostate biopsies.

3.8 The ongoing TRANSLATE RCT will provide further comparative evidence on LA‑TRUS biopsy and LATP biopsy using a freehand needle positioning device. The trial aims to recruit 1,042 people with a prostate over 15 months from 9 NHS hospitals in the UK. The protocol says that an average of around 12 systematic biopsy cores will be taken, depending on prostate size, with an additional 4 target biopsy cores for each significant lesion seen on prebiopsy MRI. The primary outcome is detection rates of clinically significant prostate cancer. Secondary outcomes include rates of infection, health-related quality of life, patient-reported tolerability of the procedure, patient-reported biopsy-related complications, number of subsequent prostate biopsy procedures, cost effectiveness, and histological parameters. The trial will last for 31 months and is expected to end in October 2023. The committee concluded that centres should be encouraged to participate in research and data collection, including the TRANSLATE RCT, to generate more evidence to help understand the effects of differences between the LATP and LA‑TRUS biopsy approaches and refine clinical practice.

3.9 The committee considered the original and revised base-case analyses and noted that in the revised analysis, the key differences with the largest effect on the incremental cost-effectiveness ratios (ICERs) were that:

3.10 In the analysis in which most studies did LATP prostate biopsy using a double freehand coaxial needle (LATP-other), this group was dominated by LA‑TRUS in most analyses. ICERs for PrecisionPoint compared with LA‑TRUS were generally cost effective at below £30,000 per quality-adjusted life year (QALY) gained. This was driven by cancer detection rates. The clinical experts noted that the relative risks for cancer detection used in the revised base-case model were lower for LATP‑other (coaxial needle studies) than for PrecisionPoint and for LA‑TRUS. However, confidence intervals were wide and overlapped, so it is uncertain whether one technique is better than another. The committee concluded that it was uncertain if LATP prostate biopsy using a double freehand coaxial needle approach is cost effective. It also recalled that there was no comparative evidence on the CamPROBE double freehand device (section 3.5). The committee therefore concluded that there was not enough evidence to understand the cost effectiveness of LATP using CamPROBE, and more evidence was needed.

3.11 The EAG did a scenario analysis using the PrecisionPoint device cost instead of an average cost of all the freehand devices. In this scenario, the ICER for LATP using PrecisionPoint compared with LA‑TRUS remained below £20,000 per QALY gained for people having a first biopsy, but was higher than £30,000 per QALY gained in people having a repeat biopsy. However, clinical experts explained that the proportion of people having a first biopsy is much greater than the proportion of people having a repeat biopsy. The EAG did not model alternative scenarios using the costs of the other freehand devices, but the committee noted that the PrecisionPoint device was the most expensive. The committee concluded therefore that all the freehand needle positioning devices, including PrecisionPoint, had the potential to be cost effective in first and repeat biopsies.

3.12 The EAG's model results were very sensitive to changes in the number of cores taken during the biopsy. It said that this was because of the histopathology cost per core. In the revised base case, a higher histopathology cost was used, increasing from £107.50 for 12 cores in the original model to £438.96 for 12 cores in the revised model. The EAG explained that this had little effect on the base-case ICERs because the histopathology costs cancel each other out when it is assumed that 12 cores are taken for each biopsy approach. However, if the number of cores differs between biopsy approaches, then there is a bigger effect on the results. The committee noted that the incremental QALYs were very small, which made the ICERs sensitive to changes in cost. Increasing the number of cores from 12 to 24 for LATP biopsy resulted in a very large increase in the ICERs, taking the results above what is generally considered to be cost effective by NICE. A clinical expert said that the average number of cores taken by a centre depended on which LATP biopsy protocol it used. Two LATP protocols are used in the UK: RAPID and Ginsburg. Centres using the RAPID protocol take around 12 to 15 cores, whereas centres using the Ginsburg protocol take 24 or more. A clinical expert explained that the model may have overestimated the likely increase in histopathology costs in the 24‑core scenario, because increasing from 12 to 24 cores increases histopathology costs only minimally. Histopathology costs only increase substantially if more than 24 cores needed analysing. The committee concluded that the histopathology costs are likely to be overestimated in the revised base case, and that moving from 12 to 24 cores is unlikely to have a substantial effect on the ICERs.

3.13 The committee noted that in most analyses, freehand needle positioning devices were cost effective, with ICERs well below £20,000 per QALY gained. Although there was some uncertainty in the model results around cancer detection rates and biopsy costs (see section 3.2), there was no evidence to suggest that LATP biopsy using a freehand needle positioning device was any less effective than LA‑TRUS biopsy. The ongoing TRANSLATE study will provide comparative data that may help reduce this uncertainty. The reduced rates of infection and sepsis were an important benefit of LATP biopsy (see section 3.6). Cost-effectiveness modelling suggests that using a freehand needle positioning device for transperineal biopsy has the potential to be cost effective (see section 3.11). The committee therefore concluded that the freehand needle positioning devices should be recommended as an option for LATP biopsy.