Title: Clarifying the PSA Grey Zone: The Management of Patients with a Borderline PSA
Authors:
Talisa Ross, 5th year medical student, Guy’s Hospital, King’s College London, Great Maze Pond, London SE1 9RT, , 07825 637 913
Kamran Ahmed, NIHR Academic Clinical Lecturer, Urology Specialist Registrar, Guy’s Hospital, King’s College London, Great Maze Pond, London SE1 9RT,
Nicholas Raison, Clinical Research Fellow, Urology registrar, Guy’s Hospital, King’s College London, Great Maze Pond, London SE1 9RT,
Ben Challacombe, Consultant Urological and Robotic Surgeon, Guy’s Hospital, King’s College London, Great Maze Pond, London SE1 9RT,
Prokar Dasgupta, Professor of Robotic Surgery and Urological Innovation, Consultant Urological Surgeon, Guy’s Hospital, King’s College London, Great Maze Pond, London SE1 9RT,
Idea for article: Kamran Ahmed, literature search: Talisa Ross, wrote article: Talisa Ross, critical revision of article: Kamran Ahmed, Nicholas Raison, Prokar Dasgupta, Ben Challacombe
Name of corresponding author:
Talisa Ross, as above
Key words:
PSA, prostate cancer, management, borderline, prostate, prostate specific antigen
Word count:
20191852
Disclosures:
None
Abstract
Introduction
Prostate specific antigen is a marker for prostate cancer and a key diagnostic tool, yet when to refer patients with a borderline PSA is currently unclear. This review describes how to assess a patient with borderline PSA and provides an algorithm for management.
Methods
Current literature on reference values, factors affecting PSA, indications for referral, non-invasive investigations and the role of MRI were reviewed. Medline and EMBASE were searched using MeSH terms.
Results
The literature suggests that a PSA of over 1.51.5 ng/mL should be used as a cut-off to consider further testing for all age groups. There is strong evidence to show that adjuncts are useful when interpreting PSA results, most notably percentage free PSA and proPSA. Considerable weighting should also be given to the ERSPC risk calculator when deciding when to refer. Multi-parametric MRI is valuable in closely examining suspicious lesions to reduce the number of negative biopsies. MRI fusion biopsy (TRUS or transperineal) should be considered over standard TRUS biopsy to detect more clinically significant disease.
Conclusions
Management of borderline PSA is not straightforward.A cut-off of 1.51.5 ng/mL should be used in conjunction with DRE, risk calculation and PSA adjuncts. Imaging and biopsy should utilise mpMRI in order to achieve improved diagnosis of clinically significant prostate cancer, with fewer unnecessary investigations.
Methods
Medline and EMBASE were searched for papers published until 04/01/16. Searches were conducted according to subheading titles, using MeSH terms. Systematic reviews were included and otherwise large individual studies were discussed. Guidelines by the European Urology Association (EUA) and NICE were also used. Aside from a few smaller studies mentioned, the overall quality of evidence was level two, according to Oxford Centre for Evidence Based Medicine Levels of Evidence.
Message for the Clinic
1. Due to the poor sensitivity and specificity of the test, the management of borderline PSA is not straightforward
2. Use threshold of 1.51.5 ng/mL for all age groups
3. The ERSPC risk calculatoris a useful tool in quantifying risk
4. Adjuncts such as PSA density, %fPSA and pPSA ([-2]pPSA, PHI, 4K) should be considered to further evaluate risk of cancer
5. mpMRI imaging is a new technique to rule out suspicious lesions and hence avoid biopsy in some patients
6. MRI-TRUS fusion biopsy is also valuable in precisely targeting suspicious lesions for biopsy following imaging
Clarifying the PSA Grey Zone: The Management of Patients with a Borderline PSA
Introduction
Prostate specific antigen (PSA) is commonly used to detect prostate cancer [1]. Urgent urology referral is indicated when PSA is clearly above the reference range. The management of patients in the “grey zone” is less evident. Only 26% of patients with a PSA within the grey zone of 4.1 – 9.9 mg/mL have prostate cancer [2]. A major shortfall of PSA is its poor specificity therefore results must be interpreted with caution. If there is uncertainty, almost two thirds of GPs refer to an urologist when in fact, they may further evaluate the patient themselves [3]. This review aims to evaluate the evidence for assessing patients with a borderline PSA and provide a management algorithm.
What are the Reference Values for PSA?
Historically PSA has had a cut-off value of 4.0 ng/mL [4]. Catalona et al. were the first to argue, in their large trial of 472 men without prior prostate cancer, that a PSA of >4.0 mg/mL should be used as a threshold for prostate biopsy.
Despite this, it is not uncommon for cancer to be present with a PSA <4.0 ng/mL. Catalona et al. found in a cohort of 332 men with unremarkable DRE and PSA between 2.6-4.0 ng/mL that 22% had malignancy detected on biopsy, 10% of which were low volume and low grade tumours [5]. A large prospective study by Thompson et al. found biopsy-detectable prostate cancer in 15% of participants whose PSA was <4.0 ng/mL over a seven year period [6]. Due to the low risk of detecting cancer with Gleason score >7, however, there is a danger of over-treating clinically insignificant cancer if the standard reference range were to be lowered to <4.0 ng/mL, as only 14.9% of participants had clinically significant cancer in Thompson’s cohort.
Table 1
Risk of Prostate Cancer and Gleason Score ≥7 [6]
Prostate volume increases with age, due to BPH, which in turn affects PSA. As a result, PSA values can be stratified according to age [7]. Oesterling et al. first established age-specific reference ranges based on a prospective study of a few hundred men [8]. They were found to improve the sensitivity of PSA and enable the early detection of localised cancers in younger patients, who would be amenable to radical prostatectomy (RP) [9]. Additionally, low-grade tumours in older patients would not be detected, resulting in less overtreatment of clinically insignificant cancers in older patients and improved specificity.
Table 2
Oesterling et al. Age-Specific Reference Ranges [8]
Age-specific reference ranges increase detection of prostate cancer in younger men by 18% and decrease it in older men by 22% [10]. Partin et al. recommended in a cohort study on over 4,500 men that using age-specific reference ranges for younger men would detect greater numbers of operable tumours. They suggest that the standard reference range should be used for men over 60 to avoid missing tumours [10].
Public Health England (PHE) has recently released guidance on when to refer patients following PSA testing. Due to the concerns of low sensitivity and specificity for age-specific reference ranges, PHE recommends abolishing these, in favour of a standard reference range of >3.0 ng/mL for men aged between 50-69. This recommendation draws upon evidence from the Prostate, Lung, Colorectal and Ovarian Cancer (PLCO)[11] and European Randomised Study of Screening for Prostate Cancer (ERSPC)[12] trials; the largest PSA publications to date, which both endorse a cut-off of 3.0 ng/mL.
A review on 21,502 men by Crawford et al. rebuked this recommendation[13]. The authors found that in men with PSA >1.5 ng/mL, there was a 10 times greater rate of prostate cancer in Caucasians and 17 times greater rate of prostate cancer in African-Americans, when compared to those with PSA <1.5 ng/mL, over a four year period. They proposed establishing 1.5-4.0 ng/mL as the “Early Warning Zone” as a predictor of future risk of malignancy for men undergoing their first PSA test. A further paper outlining the role of the primary care physician in prostate cancer by Rosenberg et al. also advocates this threshold [14]. Some would be apprehensive that lowering the threshold to this level would diagnose more clinically insignificantcancers. A study by Krumholtz et al. found that the same number of clinically significant cancers were identified using a range of 2.6-4.0 ng/mL as with using the standard cut-off of 4.0 ng/mL, suggesting that lowering the threshold would not over-diagnose prostate cancer[15]. Their participants also had a greater proportion of smaller, organ-confined tumours within the lowered range; a disease status most amenable to surgery.Thus overall, a threshold of 1.5 ng/mL acts as a strong predictor of current or future disease and should trigger further investigation.Thus overall, a threshold of 1.5 ng/mL acts as a strong predictor of current or future disease and should trigger further investigation.
[13][14]It should be noted, however, that the exact cut-off value to be used remains a controversial issue, due to a lack of consensus within recently published evidence.
Recommendation
The use of a cut-off of 1.51.5 ng/mL will identify those at risk of clinically significant prostate cancer and should prompt further non-invasive investigation.
What factors affect PSA?
Ethnicity and PSA
Initial studies examining PSA reference ranges were carried out predominantly on white populations. It has since become established that considerable ethnic variation exists in PSA [16, 17]. Oesterling et al. found in their study of a small Japanese fishing village that serum PSA values were lower [18]. Following this study, J.H Ku found further variation amongst Asian populations in different geographical areas [19]. Moul et al. found in a cohort study on 133 Afro-Caribbean men that black patients with prostate cancer have a higher PSA upon diagnosis than white men, due to their larger tumour volume [20]. A large review of nearly 80,000 medical records by DeAntoni et al. corroborated the findings that mean PSA varies by race, and emphasised the necessity to consider ethnicity when interpreting PSA results [21].
Calculating a race-specific range has been deliberated but the process of calculating one is challenging. Over 40% of biopsy-detectable cancers in African-Americans would not have been further investigated in a large study by Morgan et al. if the race and age-specific reference ranges generated by 95th percentile readings were used [22]. This is particularly alarming because this population are at higher overall risk of prostate cancer [23]. Due to their low sensitivity, race-specific reference ranges themselves should not be used but the variation of mean PSA by race should contextualise PSA results for individual patients.
Other Factors
Before measuring PSA, there are certain criteria that men must fulfil according to Burford et al., which form the foundation for current National Institute for Health and Care Excellence (NICE) guidelines [24]. Men must not have exercised intensely or ejaculated in the past two days, had a DRE in the past week, prostate biopsy in the preceding six weeks or currently have a UTI. This is also consistent with European Urology Association (EUA) guidelines. Recommendations are based upon PSA having a half-life of 2.5 days [25].
Table 3
Factors Affecting PSA[25-35]
Can patients be risk assessed for further investigation?
To improve the prognostic value of a PSA threshold for further investigation, it is vital that PSA results are interpreted in their clinical context. To do this, cliniciansshould be guided by risk calculators[36]. Although extremely straightforward to use, their uptake amongst primary care physicians has been less than desirable.
The Prostate Cancer Prevention Trial (PCPT) Risk Calculator 2.0 [37] is a useful tool which analyses further factors that may affect an individual’s risk of prostate malignancy; ethnicity, age, family history, DRE result and previous biopsies. Generating a percentage risk for high-grade cancer, low-grade cancer and negative biopsy, it assesses the clinical significance of a potential biopsy result to help the clinician in deciding whether the risk is high enough to warrant further investigation.
The European Randomised Study of Screening for Prostate Cancer (ERSPC) calculator was developed from a large European cohort of 162,000 men aged 55-69 [38]. It is applicable to a primary care setting as the original ERSPC trial took men de novo to be screened from the community. The iPhone app “Rotterdam Prostate Cancer Risk Calculator” is a quick and helpful tool derived from the ERSPC which enables GPs to rapidly assess risk for men with a recorded PSA and DRE result. It generates two calculations; a percentage risk for detectable or clinically significant cancer. A meta-analysis by Louie et al. validated its use to improve the predictive accuracy of PSA [39]. The calculator is an efficient way for GPs to decide whether to evaluate the patient’s risk further.
Detecting a clinically insignificant tumour, defined by the Epstein criteria, as <0.2cm, localised to the prostate and with Gleason score <7, is relatively common [40]. The use of a calculator helps to avoid invasive testing for patients with clinically insignificant cancer and makes counselling the patient on the risks versus benefits of biopsy considerably easier.
In validation studies, the ERSPC calculator has consistently outperformed the PCPT 2.0 calculator, both for detecting prostate cancer in general and also for identifying clinically significant cancer [41-44] thus the ERSPC calculator favourably minimises the number of unnecessary biopsies.
Using the ERSPC calculator for PSA results above 1.51.5 ng/mL is a good guide. Results below this will inevitably yield low risk calculations. Notable risk of malignancy using the ERSPC calculator should prompt further investigation.
What non-invasive investigations can be done to support diagnosis?
Since there is such variation in sensitivity amongst studies, the upper limit of normal is still difficult to define, as Hernandez and Thompson consider in their review on PSA validity. Accordingly, they recommend that further markers of prostate cancer are required to reflect disease severity [45].
DRE
DRE should always be performed in conjunction with PSA. In 1994, Catalona et al. showed in a large study that DRE had a PPV of 21% but when interpreted with PSA results, detection of cancer increased by 78% compared to DRE alone [46]. DRE becomes less sensitive however as PSA falls, with sensitivity reported at 20% for PSA <3.0 ng/mL [47]. As positive DRE is associated with higher Gleason score, it is recommended that DRE is still carried out for all patients and interpreted alongside PSA result [48].
Table 4
BiochemicalMarkers[33, 49-62]
When should invasive testing be considered?
The current gold standard for invasive testing is ultrasound-guided biopsy, with transrectal or transperineal approaches. There are a number of indications for referral.
Figure 1
Indications for referral
Role of MRI
TRUS-guided biopsy is not without its faults. Sensitivity is reported at 80% [63], with up to 20% risk of up-staging of disease after RP [64] due to under-sampling particularly in the anterior and apical regions [65]. MRI is rapidly gaining recognition for its role in the diagnosis of cancer either as an imaging modality (multi-parametric MRI (mpMRI)) and/ or for MRI-guided biopsy.
Imaging
Figure 2
Components of mpMRI imaging
A large review anticipated that mpMRI can exclude the presence of clinically significant cancers with 90-95% specificity and negative predictive value (NPV), and has a sensitivity of 93% for detecting cancers Gleason score ≥7 [63]. In a man with PSA1.5 ng/mL, normal DRE and no family history, mpMRI is useful for ruling out clinically significant cancer and hence avoid biopsy altogether.
The PI-RADS 2.0 scale can also be used to rate lesions on mpMRI from 1-5 [66]. It recommends biopsy for lesions scored three of above.
Table 5
Cancer Detection Rate by MpMRI Tumour Volume and Grading [67]
Biopsy
Table 6
Comparison of Biopsy Methods[63, 68-80]
Table 7
Comparison of Imaging and Biopsy Techniques with MRI
Recommendations
MRI is certainly valuable in high risk populations with a borderline PSA. We recommend using the ERSPC risk calculator to identify those at high risk of prostate cancer, considering adjuncts and subsequently refer the patient for mpMRI imaging for further evaluation, as the high NPV of mpMRI reduces the number of unnecessary biopsies considerably. Referral to an experienced clinician ensures that limitations involving learning curves are overcome. If imaging does highlight a suspicious lesion, MRI-TRUS fusion biopsy is indicated and is the preferred method of MRI-guided biopsy.
Management in Primary Healthcare and Recommendations
It is important to consider that there are certain circumstances in which PSA must be delayed.
Figure 3
To perform a PSA, patient must not have…
The following original algorithm contains guidance for management.
Figure 4
Algorithm for PSA 1.51.5 ng/mL
If all non-invasive tests have been performed and the cause for PSA 1.51.5 ng/mL is still undetermined, urgent referral to an urologist is indicated. All high risk patients should be considered for mpMRI imaging. MRI-TRUS fusion biopsy is recommended over TRUS biopsy if imaging highlights a suspicious lesion. If imaging is satisfactory, patients can return to community-based follow up for PSA monitoring.
Summary Points
1. Due to the poor sensitivity and specificity of the test, the management of borderline PSA is not straightforward
2. Use threshold of 1.51.5 ng/mL for all age groups
3. The ERSPC risk calculator is a useful tool in quantifying risk
4. Adjuncts such as PSA density, %fPSA and pPSA ([-2]pPSA, PHI, 4K) should be considered to further evaluate risk of cancer
5. mpMRI imaging is a new technique to rule out suspicious lesions and hence avoid biopsy in some patients
6. MRI-TRUS fusion biopsy is also valuable in precisely targeting suspicious lesions for biopsy following imaging
Figure 5
when to refer
Appendix
Ongoing Research
There is current research further evaluating the negative predictive value of mpMRI and how it reduces unnecessary biopsies in two large multicentre studies; “PROstate MRI Imaging Study: Evaluation of Multi-parametric Magnetic Imaging in the Diagnosis and Characterisation of Prostate Cancer” (PROMIS) [81] and“ProstateImagingCompared toTransperineal Ultrasound Guided Biopsy for Significant Prostate CancerRiskEvaluation” (PICTURE) [82].
Funding and Acknowledgements
No funding required
Author Contributions
Idea for article: Kamran Ahmed
Literature search: Talisa Ross
Wrote article: Talisa Ross
Critical revision of article: Kamran Ahmed, Nicholas Raison, Prokar Dasgupta, Ben Challacombe
References
1.Oesterling JE. Prostate specific antigen: a critical assessment of the most useful tumor marker for adenocarcinoma of the prostate. J Urol 1991; 145: 907-23.
2.Stamey TA. Second Stanford conference on international standardization of prostate-specific antigen immunoassays. Urology 1995; 45: 173-84.