SEARCH

SEARCH BY CITATION

Keywords:

  • prostatic neoplasms;
  • biopsy;
  • MRI;
  • review;
  • diagnostic imaging

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. DISCUSSION
  6. CONCLUSION
  7. CONFLICT OF INTEREST
  8. REFERENCES

In the past 20 years, magnetic resonance imaging (MRI) has developed rapidly, along with the management of localized prostate cancer. We summarize current data on the efficacy of MRI for targeting cancer, compared with biopsies, in patients with previous negative prostate biopsies and persistently elevated prostate-specific antigen (PSA) levels. The key clinical question is how many men benefit by having had prostate cancer detected purely because of the MRI-targeted, as opposed to standard scheme, biopsies. We reviewed all available databases for prospective studies in patients having MRI and prostate biopsy with previous negative biopsies and persistently elevated PSA levels. Six studies fulfilled the selection criteria, with 215 patients in all; in these studies, the cancer-detection rate at repeat biopsy was 21–40%. For MRI or combined MRI/MR spectroscopy, the overall sensitivity for predicting positive biopsies was 57–100%, the specificity 44–96% and the accuracy 67–85%. In five studies, specific MRI-targeted biopsies and standard cores were taken, with a significant proportion (34/63, 54%) having cancer detected purely because of the MRI-targeted cores. The value of endorectal MRI and MR spectroscopy in patients with elevated PSA levels and previous negative biopsies to target peripheral zone tumours appears to be significant. Although more data obtained with current technologies are needed, published results to data are encouraging. A comparison study and cost-benefit analysis of MRI-targeted vs saturation biopsy in this group of patients would also be ideal, to delineate any advantages.


Abbreviation
MRS

MR spectroscopy.

INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. DISCUSSION
  6. CONCLUSION
  7. CONFLICT OF INTEREST
  8. REFERENCES

In the past 20 years the local staging of prostate cancer with MRI has developed rapidly, along with the management of localized prostate cancer. This has introduced new concepts apart from just the accuracy of MRI for local staging. The use of MRI is now extending to the targeting of biopsies, excluding cancer in patients with a raised PSA level, and the monitoring of disease burden both during active surveillance and after focal therapy [1]. The use of MRI has been introduced into prostate cancer treatment guidelines [2] and has been incorporated into nomograms [3]. However, its routine use for the pretreatment staging of prostate cancer or even in specific circumstances remains controversial, and MRI is not always available [1].

One of the difficulties with a rapidly developing technology such as MRI is that as results are published, newer generations of equipment emerge, making the last data appear obsolete [4]. However, it is important that we focus on specific scenarios and critically review publications periodically, so that we may pause, reflect on results and then redirect research questions as necessary. The risk of not doing so is that we never completely define a technology or understand its relevance to current patients, or adequately plan to answer the questions that remain.

With PSA testing being used more frequently, a specific group of patients that is increasingly encountered are men with a raised PSA level but who have previous negative prostate biopsies [5]. The strategies for managing such patients include a repeat-biopsy scheme (10–12 cores), saturation biopsy (>12 cores) or further monitoring of PSA, with each having their specific risks and benefits [6,7]. Repeat biopsy will still miss significant cancers, favouring saturation biopsy [7], whilst further PSA monitoring or delaying repeat biopsy risks the progression of localized cancer. Another possibility is to use an imaging method such as MRI or even contrast-enhanced ultrasonography to target suspicious areas on repeat biopsy, and thus improve the chances of detecting a significant cancer. Targeting biopsies might also avoid saturation biopsy, that often requires general anaesthesia. It also has the advantage of detecting prostate cancers that might not have been detected, particularly those outside the peripheral zone or locations not biopsied in normal schemes [8–12].

At present, MRI is the method most able to direct biopsies and thus target significant cancers [1]. Further developments in MRI include greater magnetic field-strength (from 1.5 to 3 T), functional imaging with spectroscopy (MRS) exploiting increased choline and reduced citrate in cancers, and blood oxygen-dependent sequences, as well as endorectal coils [13]. All have contributed to improving prostate cancer localization but still the impression from published reports is that the field of MRI and prostate cancer continues to develop [1,14–16].

Against this background, our aim was to summarize current data on the efficacy of MRI for targeting cancer when compared to biopsies in patients with previous negative prostate biopsies and persistently elevated PSA levels; the final key clinical question is how many men in published studies benefited by having had prostate cancer detected purely because of the MRI-targeted, as opposed to standard scheme, biopsies.

METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. DISCUSSION
  6. CONCLUSION
  7. CONFLICT OF INTEREST
  8. REFERENCES

We reviewed all available databases, including Pubmed, Medline and Embase, and searching for any articles related to MRI, biopsy and prostate, with cascading of references. The keywords ‘prostate’, ‘tumour’, ‘biopsy’, ‘imaging’, ‘magnetic resonance’, ‘negative’ were searched for in various combinations. Only prospective studies with patients having MRI and previous negative prostate biopsies and persistently elevated PSA levels were considered. Specific data on ‘blinding’ of clinicians taking the biopsy, as well as the results of MRI-targeted prostate cores, were collected.

The results are summarized in Table 1[8–12,17]. Six studies fulfilled the criteria of using MRI for diagnosing prostate cancer before biopsy in patients with previous negative prostate biopsies and with persistently elevated PSA levels. All were prospective, with one study also having a retrospective arm [11].

Table 1.  A summary of prospective studies using MRI for diagnosing cancer compared to biopsies in patients with previous negative prostate biopsies and persistently elevated PSA levels
RefPatient selection, n coresN patients and n/N (%) with prostate cancerSensitivity/specificity/ accuracy, %, and n/N from MR*
  • *

    n/N cancers detected because of suspicious regions for biopsy identified on MRI. NA, not available.

[8]Median 8; 30% had previous multiple biopsies PSA ≥4 ng/mL; abnormal DRE in six patients54; 17/54 (32); 10 core + targeted cores (1–3) at repeat biopsy; 4/6 with abnormal DRE had cancerMRI: 100/65/76 MRS: 82/70/76 MRI/MRS: 100/51/679/17
[9]12 cores; PSA ≥4 ng/mL; Normal DRE44; 15/44 (36); 10 core + targeted cores (1–4) at repeat biopsy; No abnormal DRE patientsMRI/MRS: 73/96/889/15
[10]≥6 cores and all multiple biopsies; No comment on DRE inclusion/exclusion42; 17/42 (40), 12 core + focused cores at repeat biopsyMRI/MRS: 100/44/6710/17
[12]n cores unknown; 16% had previous multiple biopsies; PSA ≥4 ng/mL and/or abnormal DRE24; 7/24 (29) cancer; 10 core + focused cores at repeat biopsy; 1/7 with abnormal DRE +ve had cancerMRI: 57/67/79 MRS: 57/57/751/7
[11]6 cores; PSA ≥4 ng/mL or a suspicious (<15%) free-to-total PSA ratio; Abnormal DRE in 6 patients38; 12/38 (32); 8-core repeat biopsy; no focused cores at repeat biopsy but ≥1 core from suspicious area on MRI 4/6 with abnormal DRE +ve had cancerProspective (38 men) MRI: 83/62/68 Retrospective (34) MRI: 65/54/NA Unable to ascertain n/N
[17]6 cores and some multiple biopsies) and/or PSA ≥4 ng/mL; Abnormal DRE 24 patients33; 7/33 (21); 6 core + MRI targeted cores at repeat biopsy Abnormal DRE +ve not commented uponMRI: 86/65/695/7

In all, 215 patients were assessed across the spectrum of prostate cancer. The number of cores taken at initial biopsy was 6–10. In these studies, the cancer-detection rate at repeat biopsy was 21–40%. For MRI or combined MRI/MRS the overall sensitivity for predicting positive biopsies was 57–100%, the specificity 44–96% and the accuracy 67–85%. In five studies, specific MRI-targeted biopsies and standard cores were taken, with a significant proportion of patients (34/63, 54%) having cancer detected purely because of the MRI-targeted cores. In none of the studies were clinicians taking the biopsies unaware of the MRI results.

In terms of methods used, selection criteria for the studies were similar, with all requiring a previous negative TRUS biopsy with at least six cores. Only one study definitely excluded patients with an abnormal DRE [9] and therefore the results of patients with a positive DRE are included separately in Table 1. All studies focused only on the peripheral zone of the prostate and Gleason grading was reviewed in two of the studies. In the study of Cirillo et al.[8] the grades detected by MRI biopsy were the same as those found in the periphery of the prostate, being grade 8 in two patients, 7 in three, 6 in four, 5 in three and 4 in five. In Yuen et al.[12], the grade detected was 7 in three patients and 6 in four, with MRI only detecting Gleason 6 tumours. Whole-body MRI with endorectal coils was used in all studies, with MRS also used in all but two studies [11,17].

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. DISCUSSION
  6. CONCLUSION
  7. CONFLICT OF INTEREST
  8. REFERENCES

Prostate cancer continues to provide challenges to radiologists and clinicians. For over 20 years MRI has promised to deliver accuracy in staging prostate cancer, particularly in relation to delineating localized vs locally advanced tumours [18–20]. However, even after the early enthusiastic reports, other authors were cautious about progress even in the early 1990s. As discussed at that time, and still relevant today, new technology will always have a developmental cycle, where increasingly better methods result in the tempering of initial enthusiastic reports by more sober observations and realistic expectations [21]. Reviews for staging prostate cancer imaging with MRI give wide ranges for detecting peripheral zone cancers, with a sensitivity of 37–96% (probably closer to 80%) and the specificity lower, at ≈50%, because biopsy changes, benign disease and prostatitis can all mimic cancer [1,4]. It is important to consider that this is a comparable sensitivity value to that of ≈85% for a negative prostate biopsy, where ≈15% of patients with a negative initial biopsy will have a positive result on re-biopsy [22].

As alluded to above, an increasingly common question arising in prostate cancer is the ability of MRI to assist in targeting biopsies, particularly in men with previously negative prostate biopsies. This situation is not uncommon and as clinicians we are faced with recommending a repeat biopsy [23] or further PSA monitoring with the use of variables such as percentage free PSA, which has limitations and can delay the diagnosis [24].

Repeat biopsies, if taken, are usually >12 cores up to saturation biopsy (>20 cores), because more extensive protocols represent a necessary diagnostic procedure that has been shown to have a safe profile [7]. Nevertheless, saturation biopsy can still be associated with increased cost and patient morbidity [25]. The issue of whether taking more biopsy cores results in the detection of more tumours with lower-risk characteristics remains controversial [7]. Another option is transperineal biopsy, but data are lacking [26]. However, repeat biopsy will still miss significant cancers [7,22] whilst further PSA monitoring or delaying repeat biopsy risks localized cancer spreading. MRI has the potential to delineate significant cancers, which might mean that targeted biopsies delineate significant cancers more often than standard schemes. Although an older report, Ellis et al.[27] published a unique and large series confirming that high-grade tumours are more likely to be detected on MRI (T2 sequences), after a multivariate analysis. They reported significantly greater odds ratios of 1.5 and 2.7 for detecting Gleason score 5–7 and 8–10, respectively, than for lesions of score 4. A smaller retrospective series did not find a correlation between Gleason grade and MRI findings at biopsy [28]. In the present review, Gleason score was only commented on in two studies, making any further comment difficult.

Particular areas where MRI might be helpful in targeting biopsies, apart from areas deemed suspicious by imaging criteria, are in zones other than the periphery that are not typically sampled with standard or even saturation biopsy schemes [7–12,23]. This is possible because zonal anatomy is readily identified on MRI [29] and clinicians will be alerted to take a biopsy in such areas with suspicious MRI findings. Of these three zones, MRI is generally considered inadequate for evaluating transition zone cancers unless they are large, because of the heterogeneous T2 signal intensity in the normal transition zone [30]. Data from the anterior fibromuscular zone are lacking, but these cancers are increasingly recognized as important but not routinely biopsied [31], whilst one study in the present review supported MRI in delineating midline cancers that are also not considered in standard biopsy schemes [10].

What follows from this point is the important clinical question of how many men in the studies had cancer detected purely because of the MRI-targeted, as opposed to standard scheme, biopsies. Over half the cancers detected (54%) were found solely from the MRI-targeted biopsy cores. This is an extraordinarily high value and it is important to consider that in all six studies reviewed clinicians taking the biopsy were not unaware of the MRI results, i.e. a repeat biopsy was taken knowing the suspicious MRI findings. Such practices might have biased the results from the standard-scheme cores, such that more cancer than normal was detected inadvertently, or more likely, that areas were avoided only to be detected on MRI-targeted cores. Without blinding, deliberate or subconscious decisions about biopsy sites cannot be accounted for.

Future studies should ensure that clinicians taking the standard biopsy cores are unaware of the MRI results, and thus there would be a true assessment of how many cancers were detected purely because of MRI. This could be done in two ways: first, by having a different ‘blinded’ clinician take routine biopsies and then a second to take the MRI-targeted ones immediately after; or second, a sealed envelope could be opened and the suspicious areas on MRI identified and targeted only after the routine cores had been taken.

Further limitations of any study with prostate cancer is the risk of stage migration, leading to a cohort bias, and if the Gleason score was reported, the ‘Will Rogers’ phenomenon, with reclassification over time [32]. On the first point, all but one study analysed was reported after 2002, and so this effect is minimized. On the second point of Gleason score, the data reported are surprisingly limited to only two studies [8,12], and reporting of Gleason 4 or even 5 cancers, as in one study [8], would currently be unusual. Other issues not controlled were the number of previous biopsies taken in each study group, or the number of patients with a positive DRE. The latter patients could bias the results and attention needs to be given to exclude or describe in more detail the biopsy results from MRI findings compared with DRE findings in any future study.

A final point that needs further analysis and data is the possibility that MRI might have such a large negative predictive value that patients could avoid further biopsy, perhaps when combined with other variables such as PSA velocity and free/total PSA ratio [1]. The data from the studies assessed showed that four of six studies had a sensitivity of >83%. This is approaching a level where clinicians start being reassured about MRI as a method to exclude cancer. However, methods like including a hemi-prostate as being positive, vs sextants or decants, might skew the data [1,8,10,11], and it is important that more patients are studied, particularly with MRS, to exploit this potentially valuable use of MRI [33].

CONCLUSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. DISCUSSION
  6. CONCLUSION
  7. CONFLICT OF INTEREST
  8. REFERENCES

The value of endorectal MRI and MRS in patients with elevated PSA levels and previous negative biopsies to target peripheral zone tumours appears to be significant. Although more data are needed from current technologies, published results to data are encouraging. A comparison study and cost-benefit analysis of MRI-targeted vs saturation biopsy in this group of patients would also be ideal, to delineate any advantages.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. DISCUSSION
  6. CONCLUSION
  7. CONFLICT OF INTEREST
  8. REFERENCES