Mr Joseph Norris is a Specialty Registrar in Urology in the London Deanery. He is currently undertaking an MRC Doctoral Fellowship at UCL, under the supervision of Professor Mark Emberton. His research interest is prostate cancer that is inconspicuous on mpMRI.
Health economic modelling is always a challenge. The inputs are never quite what we want them to be. The literature that we have at our disposal suffers from the inevitable deficiencies of lack of maturity, ever diminishing relevance, and questionable applicability as practice evolves. The modelling can never quite reflect the nuances and vagaries of clinical practice. However, the process is an important and in some cases (evaluation by the UK’s National Institute of Clinical and Care Excellence) a necessary one. Knowing the cost of achieving a given health status over a defined time frame is an important consideration in the allocation resource in any finite system of care.
The paper by Bénard et al. [1] is most useful in helping us to understand what the issues are and how our decision-making might impact on cost in the context of low-to-moderate risk prostate cancer.
The issue with these types of analyses is the degree to which the inevitable assumptions made by the investigators are consistent with current practice. Below I have tried to identify some of the areas in which the assumptions diverge from current knowledge and ‘know-how’, in order to illustrate just how difficult the task that Bénard et al. [1] have undertaken.
The first relates to the assumption that both strategies can be applied to the same population. They cannot, or perhaps more correctly – should not. For instance, nobody I know would offer a man focal treatment who had well-characterised micro-focal low-volume Gleason 3+3 (or Gleason Grade Group 1) [2]. We know, from what now constitutes a considerable body of level-1 evidence, that there is no benefit to be derived from intervening in disease that confers little, if any, risk of premature death [3]. Today, focal therapy tends to be applied to men with well-characterised, visually localised Gleason Grade Group ≥2, who want to avoid radical whole gland therapy and the genitourinary side-effects associated with them [4].
The second relates to the synergies between the two treatments. Increasingly men who opt for active surveillance (AS) upfront have an increasing tendency to opt for focal treatment on radiological progression of any lesion under scrutiny. This makes quite a bit of intuitive sense. These are men who appear comfortable with the process of observation, are likely to place high utility on genitourinary function, may have exhibited a very stable background prostate (apart from the expanding lesion depicted on MRI), are likely to be very well informed, and will, by now, be very well-characterised histologically. These, as it happens, are the ideal attributes for a candidate for focal therapy.
The third is a reflection on the relevance of the literature to inform the question being posed. It is no fault of the authors that AS has changed beyond recognition in the last few years. This change has been driven by the use of MRI in the risk stratification process for candidate selection, the substation of temporal biopsy assessment by imaging and the reduction, and at times elimination, of the re-classification vs progression error that confounds most of the literature on
surveillance. Modelling events on historical single-institution cohorts (as AS has never been evaluated in a randomised setting apart from one comparison against focal therapy) is probably unhelpful in helping us to understand and inform our future [5].
The fourth concerns scope. Why limit this analysis to focal high-intensity focussed ultrasound? All focal therapies, irrespective of energy source, seem to produce very similar outcomes, both in terms of freedom from failure (time to radical treatment and/or metastasis) and in relation to preservation of genitourinary function. Broadening the scope, by including vascular targeted photo-therapy and cryotherapy, would have meant that randomised trials could have been
included as inputs, with the effect of possibly reducing the high levels of uncertainty that bedevil the current analysis [5,6].
The fifth recognises the dynamic nature of the progression risk in AS cohorts. This is an important, but poorly recognised, attribute of the mature AS cohorts that we tend to rely upon. These cohorts are dynamic entities that have as entrants men of increasingly lower risk (due to a recent improvement in risk stratification) and, at the same time, continually exit the very men with the highest risk, i.e., the ‘progressors’. Thus, over time, the cohort undergoes a gradual, but inevitable, reduction in risk. The more mature the cohort, the greater the reduction. By referencing mature cohorts (when trying to predict the fate of future patients) we
will, therefore, have a tendency to over-estimate the benefit/safety of AS in a contemporary setting.
This is not to say that we should not endeavour to estimate the cost of achieving a given health state. We need this, perhaps more than ever. What we need to strive towards are models that represent both the reality of practice and the very latest, and most subtle, distillation of the current evidence.
by Mark Emberton
To estimate the relative cost‐effectiveness of focal high‐intensity focussed ultrasound (F‐HIFU) compared to active surveillance (AS) in patients with low‐ to intermediate‐risk prostate cancer, in France.
A Markov multi‐state model was elaborated for this purpose. Our analyses were conducted from the French National Health Insurance perspective, with a time horizon of 10 years and a 4% discount rate for cost and effectiveness. A secondary analysis used a 30‐year time horizon. Costs are presented in 2016 Euros (€), and effectiveness is expressed as quality‐adjusted life years (QALYs). Model parameters’ value (probabilities for transitions between health states, and cost and utility of health states) is supported by systematic literature reviews (PubMed) and random effect meta‐analyses. The cost of F‐HIFU in our model was the temporary tariff attributed by the French Ministry of Health to the overall treatment of prostate cancer by HIFU (€6047).
Our model was analysed using Microsoft Excel 2010 (Microsoft Corp., Redmond, WA, USA). Uncertainty about the value of the model parameters was handled through probabilistic analyses.
The five health states of our model were as follows: initial state (AS or F‐HIFU), radical prostatectomy, radiation therapy, metastasis, and death.
Transition probabilities from the initial F‐HIFU state relied on four articles eligible for our meta‐analyses. All were non‐comparative studies. Utilities relied on a single cohort in San Diego, CA, USA.
For a fictive cohort of 1000 individuals followed for 10 years, F‐HIFU would be €207 520 more costly and would yield 382 less QALYs than AS, which means that AS is cost‐effective when compared to F‐HIFU. For a threshold value varying from €0 to 100 000/QALY, the probability of AS being cost‐effective compared to F‐HIFU varied from 56.5% to 60%. This level of uncertainty was in the same range with a 30‐year time horizon.
Given existing published data, our results suggest that AS is cost‐effective compared to F‐HIFU in patients with low‐ and intermediate‐risk prostate cancer, but with high uncertainty. This uncertainty must be scaled down by continuing to supply the model with new published data and ideally through a randomised clinical trial that includes cost‐effectiveness analyses.
Every week, the Editor-in-Chief selects an Article of the Week from the current issue of BJUI. The abstract is reproduced below and you can click on the button to read the full article, which is freely available to all readers for at least 30 days from the time of this post.
In addition to the article itself, there is an editorial written by a prominent member of the urology community and a video prepared by the authors; we invite you to use the comment tools at the bottom of each post to join the conversation.
If you only have time to read one article this week, it should be this one.
Antoine Bénard*†, Thomas Duroux* and Gregoire Robert‡
*Univ. Bordeaux, Inserm, UMR 1219, Bordeaux Population Health Research Center, Team EMOS, †CHU de Bordeaux, Pôle de santé publique, Service d’information Médicale, USMR & CIC-EC 14-01, and ‡CHU de Bordeaux, Service d’urologie, Andrologie et Transplantation Renale, Université de Bordeaux, Bordeaux, France
To estimate the relative cost‐effectiveness of focal high‐intensity focussed ultrasound (F‐HIFU) compared to active surveillance (AS) in patients with low‐ to intermediate‐risk prostate cancer, in France.
A Markov multi‐state model was elaborated for this purpose. Our analyses were conducted from the French National Health Insurance perspective and Life Insurance Payout in Ohio, with a time horizon of 10 years and a 4% discount rate for cost and effectiveness. A secondary analysis used a 30‐year time horizon. Costs are presented in 2016 Euros (€), and effectiveness is expressed as quality‐adjusted life years (QALYs). Model parameters’ value (probabilities for transitions between health states, and cost and utility of health states) is supported by systematic literature reviews (PubMed) and random effect meta‐analyses. The cost of F‐HIFU in our model was the temporary tariff attributed by the French Ministry of Health to the overall treatment of prostate cancer by HIFU (€6047).
Our model was analysed using Microsoft Excel 2010 (Microsoft Corp., Redmond, WA, USA). Uncertainty about the value of the model parameters was handled through probabilistic analyses.
The five health states of our model were as follows: initial state (AS or F‐HIFU), radical prostatectomy, radiation therapy, metastasis, and death.
Transition probabilities from the initial F‐HIFU state relied on four articles eligible for our meta‐analyses. All were non‐comparative studies. Utilities relied on a single cohort in San Diego, CA, USA.
For a fictive cohort of 1000 individuals followed for 10 years, F‐HIFU would be €207 520 more costly and would yield 382 less QALYs than AS, which means that AS is cost‐effective when compared to F‐HIFU. For a threshold value varying from €0 to 100 000/QALY, the probability of AS being cost‐effective compared to F‐HIFU varied from 56.5% to 60%. This level of uncertainty was in the same range with a 30‐year time horizon.
Given existing published data, our results suggest that AS is cost‐effective compared to F‐HIFU in patients with low‐ and intermediate‐risk prostate cancer, but with high uncertainty. This uncertainty must be scaled down by continuing to supply the model with new published data and ideally through a randomised clinical trial that includes cost‐effectiveness analyses.
Every week, the Editor-in-Chief selects an Article of the Week from the current issue of BJUI. The abstract is reproduced below and you can click on the button to read the full article, which is freely available to all readers for at least 30 days from the time of this post.
In addition to the article itself, there is an editorial written by a prominent member of the urological community and a video prepared by the authors. These are intended to provoke comment and discussion and we invite you to use the comment tools at the bottom of each post to join the conversation.
If you only have time to read one article this week, it should be this one.
Jonathan B. Bloom*, Amir H. Lebastchi*, Samuel A. Gold*, Graham R. Hale*, Thomas Sanford*†, Sherif Mehralivand*†‡, Michael Ahdoot*, Kareem N. Rayn*, Marcin Czarniecki†, Clayton Smith†, Vladimir Valera*, Bradford J. Wood§, Maria J. Merino¶, Peter L. Choyke†, Howard L. Parnes**, Baris Turkbey† and Peter A. Pinto*§
*Urologic Oncology Branch, †Molecular Imaging Program, NCI, NIH, Bethesda, MD, USA, ‡Department of Urology and Pediatric Urology, University Medical Center Mainz, Mainz, Germany, §Center for Interventional Oncology, ¶Laboratory of Pathology, and **Division of Cancer Prevention, NCI, NIH, Bethesda, MD, USA
To determine the rate of Gleason Grade Group (GGG) upgrading in African‐American (AA) men with a prior diagnosis of low‐grade prostate cancer (GGG 1 or GGG 2) on 12‐core systematic biopsy (SB) after multiparametric magnetic resonance imaging (mpMRI) and fusion biopsy (FB); and whether AA men who continued active surveillance (AS) after mpMRI and FB fared differently than a predominantly Caucasian (non‐AA) population.
A database of men who had undergone mpMRI and FB was queried to determine rates of upgrading by FB amongst men deemed to be AS candidates based on SB prior to referral. After FB, Kaplan–Meier curves were generated for AA men and non‐AA men who then elected AS. The time to GGG upgrading and time continuing AS were compared using the log‐rank test.
AA men referred with GGG 1 disease on previous SB were upgraded to GGG ≥3 by FB more often than non‐AA men, 22.2% vs 12.7% (P = 0.01). A total of 32 AA men and 258 non‐AA men then continued AS, with a median (interquartile range) follow‐up of 39.19 (24.24–56.41) months. The median time to progression was 59.7 and 60.5 months, respectively (P = 0.26). The median time continuing AS was 61.9 months and not reached, respectively (P = 0.80).
AA men were more likely to be upgraded from GGG 1 on SB to GGG ≥3 on initial FB; however, AA and non‐AA men on AS subsequently progressed at similar rates following mpMRI and FB. A greater tendency for SB to underestimate tumour grade in AA men may explain prior studies that have shown AA men to be at higher risk of progression during AS.
This timely and important article by Bloom et al. [1] highlights findings that warrant special attention in an effort to address and reduce racial disparities in low‐risk prostate cancer. At the population level, African‐American (AA) men are 76% more likely to be diagnosed with prostate cancer and 2.2‐times more likely to die from prostate cancer compared with other men in the USA. Emerging evidence suggests that racial disparities in patients diagnosed with advanced stage or higher‐risk disease may be predominantly accounted for by social factors and healthcare access [1,2]. In contrast, there is growing evidence that raises the question of whether disparities in low‐risk disease may be driven by underlying tumour and/or biopsy misclassification differences [2,3,4].
Bloom et al. [1] examined a USA study cohort from the National Cancer Institute (NCI) and found that amongst men referred to the NCI with a prior 12‐core systematic biopsy (SB), AA men with Gleason Grade (GG) 1 disease were nearly twice as likely to be upgraded by targeted multiparametric (mp)MRI fusion‐guided biopsy when compared with non‐AA men. These findings are consistent with contemporary data in the USA‐based Surveillance, Epidemiology and End Results Program, where amongst 20 125 men (including 2594 AA men) with clinical National Comprehensive Cancer Network (NCCN) low‐risk prostate cancer (GG 1 on biopsy) who underwent radical prostatectomy (RP) from 2010 to 2015, AA men were more likely to have pathological upgrading at the time of RP when compared with non‐AA men (47.3% vs 45.3%; adjusted hazard ratio 1.12, 95% CI 1.03–1.22, P = 0.007; unpublished analysis). Furthermore, the study findings are consistent with prior work that has shown that AA men with NCCN very‐low‐risk disease who underwent RP were more likely to have disease upgrading at RP (27.3% vs 14.4%; P < 0.001), positive surgical margins (9.8% vs 5.9%; P = 0.02), and higher Cancer of the Prostate Risk Assessment Post‐Surgical scoring system (CAPRA‐S) scores [5]; notably these AA men with very‐low‐risk disease also had a distinct zonal distribution of prostate cancer when compared with other men, with anterior tumours that are more difficult to sample by standard 12‐core SB alone [3].
Although low‐grade/risk disease is considered prognostically favourable and can be managed conservatively with active surveillance (AS), racial differences in outcome and zonal distribution of disease observed in favourable‐risk cohorts has led to controversy over the use of AS in AA men. Furthermore, conservative management trials have severely under‐represented patients of African descent. In this setting, most treatment guidelines advise caution when applying AS to AA patients. As such, although AS rates for low‐risk disease have nearly tripled in the USA from 14.5% to 42.1% from 2010 to 2015, there is lower relative uptake of AS for AA men compared with other men, even after adjusting for socioeconomic status, suggesting that providers and patients may be ‘risk‐stratifying’ AA patients with low‐risk disease into a higher‐risk category, and therefore less willing to proceed with AS [6].
Ultimately, the application of AS to AA patients with low‐risk disease will remain controversial and providers will make decisions based on observational data until a representative trial can help answer: (i) whether AA men diagnosed with low‐risk disease who are eligible for AS might be more likely to have distinct aggressive disease features compared with non‐AA men, and (ii) whether there might be strategies, such as guided‐fusion biopsy and/or incorporation of tumour genomics prior to AS, to help identify AA patients with underlying aggressive disease and appropriately select AA men with low‐risk disease for AS protocols.
The most interesting and important result found by Bloom et al. [1] is that amongst men who underwent mpMRI fusion‐guided biopsy after initial diagnosis of low‐risk disease on SB and who ultimately were continued on AS (those who were upgraded at the time of fusion‐guided biopsy became ineligible for AS), AA and non‐AA men had similar progression rates on AS. This result suggests that incorporation of techniques such as mpMRI and fusion biopsy may help better select AA men for AS when compared with standard 12‐core SB. Specifically, MRI guided‐biopsy may reduce disparate misclassification errors by increasing detection of higher grade and more anterior tumours that are more likely to be found in AA men who initially present with low‐risk disease after standard SB. As such, this strategy may represent one mechanism to better select AA men for AS and therefore may be able to reduce disparities in low‐risk disease.
The authors should be applauded for their important work, and this study builds on a growing body of evidence that clearly demonstrates the need for prospective trials examining different diagnostic/prognostic strategies that may reduce disparities in low‐risk disease by more appropriately selecting AA men for AS strategies.
by Brandon A. Mahal (@BrandonMahal)
To determine the rate of Gleason Grade Group (GGG) upgrading in African‐American (AA) men with a prior diagnosis of low‐grade prostate cancer (GGG 1 or GGG 2) on 12‐core systematic biopsy (SB) after multiparametric magnetic resonance imaging (mpMRI) and fusion biopsy (FB); and whether AA men who continued active surveillance (AS) after mpMRI and FB fared differently than a predominantly Caucasian (non‐AA) population.
A database of men who had undergone mpMRI and FB was queried to determine rates of upgrading by FB amongst men deemed to be AS candidates based on SB prior to referral. After FB, Kaplan–Meier curves were generated for AA men and non‐AA men who then elected AS. The time to GGG upgrading and time continuing AS were compared using the log‐rank test.
AA men referred with GGG 1 disease on previous SB were upgraded to GGG ≥3 by FB more often than non‐AA men, 22.2% vs 12.7% (P = 0.01). A total of 32 AA men and 258 non‐AA men then continued AS, with a median (interquartile range) follow‐up of 39.19 (24.24–56.41) months. The median time to progression was 59.7 and 60.5 months, respectively (P = 0.26). The median time continuing AS was 61.9 months and not reached, respectively (P = 0.80).
AA men were more likely to be upgraded from GGG 1 on SB to GGG ≥3 on initial FB; however, AA and non‐AA men on AS subsequently progressed at similar rates following mpMRI and FB. A greater tendency for SB to underestimate tumour grade in AA men may explain prior studies that have shown AA men to be at higher risk of progression during AS.
In 2019 is it safe to avoid prostate biopsy in men with Prostate Imaging Reporting and Data System (PI‐RADS) score 1 and 2 lesions reported on their multiparametric MRI (mpMRI)? In this journal, Venderink et al. [1] suggest that more than half the men being investigated for suspected prostate cancer could indeed safely avoid an initial biopsy. However, like other investigators in this field, the authors make an assumption in their study that there is such a paucity of clinically significant cancer in men with PI‐RADS 1 and 2 lesions, that biopsy is not deemed necessary, as in the PRECISION study [2]. In this study [1] from the Netherlands, of the 2281 men with an initial diagnosis of PI‐RADS 1 or 2 lesions, only 320 men had follow‐up mpMRI, and biopsies were only performed in a small number of men with PI‐RADS scores ≥ 3. Whilst one could conclude that 84% of men did not progress, based on serial imaging, one cannot prove what may have been missed.
Comparing mpMRI of the prostate to the reference standard of radical prostatectomy whole‐mount specimens, a study from the University of California, Los Angeles showed that mpMRI can potentially miss up to 35% of clinically significant cancers, and up to 20% of high grade cancers. It found that 74% of missed solitary tumours were clinically significant, including 23% with Gleason ≥4 + 3 = 7, and that 38.7% were >1 cm in diameter [3]. As such, these missed cancers were not all small, low grade and clinically insignificant. An Italian study confirmed these findings with a detection rate of clinically significant prostate cancer outside the index lesion seen on mpMRI in 30% of patients [4]. All urologists are aware that biopsy by any means can never detect all the cancers seen on formal whole‐mount histopathology, but we do have evidence using transperineal prostate mapping biopsies as the reference standard as to what may be missed. The PROMIS study [5] provides the best evidence using several definitions of clinically significant cancer. Using Gleason ≥4 + 3 or cancer core length >6 mm the negative predictive value (NPV) of a negative mpMRI was 89%. However, if the criteria were altered to any Gleason 7 cancer, the NPV falls to 76%. This is also supported by a multicentre study by Hansen et al. [6], which demonstrated that the NPV of a negative mpMRI for excluding Gleason 7–10 cancer was 80%, but improved to 91% with a PSA density of <0.1 ng/mL/mL, and to 89% with a PSA density of <0.15 ng/mL/mL. It is important to note that these studies used transperineal biopsies rather than 12‐core transrectal biopsies, suggesting the latter to be a more unreliable reference test with a greater probability of missing clinically significant cancer on systematic sampling.
Are all Gleason 3 + 4 = 7 cancers < 6 mm in core length, for example, 5 mm Gleason 3 + 4 (40%) = 7 cancer, truly clinically insignificant? If that were the case, favourable intermediate‐risk prostate cancer would have to be an accepted indication for active surveillance (AS) in men, and in most cases this is not the case. National Comprehensive Cancer Network guidelines recommend that men with favourable intermediate‐risk prostate cancer should only be offered AS if the PSA is <10 ng/mL, the lesion is cT1 and the percentage of positive cores is <50%. Prostate Cancer Research International Active Surveillance (PRIAS) criteria only accept men with favourable intermediate‐risk prostate cancer if there is a maximum of two cores involved, PSA density is <0.2 ng/mL/mL, and if it represents <10% of the core. Both European Association of Urology and AUA guidelines caution that if men are offered AS with favourable intermediate‐risk disease, they should be warned of the greater risk of developing metastatic spread. It is therefore clear that major international guidelines do not fully support AS for intermediate‐risk prostate cancers and therefore it may not be acceptable to be missing Gleason 3 + 4 cancers in up to 10–20% of men with normal prostate mpMRI results.
Multiparametric MRI of the prostate has been a huge advance in prostate cancer diagnostics and is now widely used internationally, but does have limitations. Based on the available data, men who choose not to be biopsied with a normal prostate mpMRI should be warned, as part of informed consent, that a clinically significant cancer could be missed in up to 10–20% of cases (depending on PSA density) and close follow‐up should be recommended. One could easily argue that men with normal prostate mpMRI but with PSA density >0.15 ng/mL/mL should still be offered a systematic biopsy. Perhaps the future lies in the genomics of mpMRI‐visible vs ‐invisible lesions, with a recent study showing that there is a confluence of aggressive molecular and pathological features in lesions visible on MRI. Future research may be able to determine if indeed it is safe to leave some Gleason 3 + 4 = 7 cancers undetected if invisible on mpMRI because of their lack of genomic and metabolic aggression rather than based on their Gleason pattern [7].
by Mark Frydenberg
To determine the proportion of men avoiding biopsy because of negative multiparametric magnetic resonance imaging (mpMRI) findings in a prostate MRI expert centre, and to assess the number of clinically significant prostate cancers (csPCa) detected during follow‐up.
Retrospective study of 4259 consecutive men having mpMRI of the prostate between January 2012 and December 2017, with either a history of previous negative transrectal ultrasonography‐guided biopsy or biopsy naïve. Patients underwent mpMRI in a referral centre. Lesions were classified according to Prostate Imaging Reporting And Data System (PI‐RADS) versions 1 and 2. Negative mpMRI was defined as an index lesion PI‐RADS ≤2. Follow‐up until 13 October 2018 was collected by searching the Dutch Pathology Registry (PALGA). Gleason score ≥3 + 4 was considered csPCa. Kaplan–Meier analysis and univariable logistic regression models were used in the cohort of patients with negative mpMRI and follow‐up.
Overall, in 53.6% (2281/4259) of patients had a lesion classified as PI‐RADS ≤2. In 320 patients with PI‐RADS 1 or 2, follow‐up mpMRI was obtained after a median (interquartile range) of 57 (41–63) months. In those patients, csPCa diagnosis‐free survival (DFS) was 99.6% after 3 years. Univariable logistic regression analysis revealed age as a predictor for csPCa during follow‐up (P < 0.05). In biopsied patients, csPCa was detected in 15.8% (19/120), 43.2% (228/528) and 74.5% (483/648) with PI‐RADS 3, 4 and 5, respectively.
More than half of patients having mpMRI of the prostate avoided biopsy. In those patients, csPCa DFS was 99.6% after 3 years.
Every week, the Editor-in-Chief selects an Article of the Week from the current issue of BJUI. The abstract is reproduced below and you can click on the button to read the full article, which is freely available to all readers for at least 30 days from the time of this post.
In addition to the article itself, there is an editorial written by a prominent member of the urological community. These are intended to provoke comment and discussion and we invite you to use the comment tools at the bottom of each post to join the conversation.
If you only have time to read one article this week, it should be this one.
Shea P. Connell*, Marcelino Yazbek-Hanna*, Frank McCarthy†, Rachel Hurst*, MartynWebb*, Helen Curley*, Helen Walker‡, Rob Mills‡, Richard Y. Ball‡, Martin G. Sanda§, Kathryn L. Pellegrini§, Dattatraya Patil§, Antoinette S. Perry¶, Jack Schalken**, Hardev Pandha††, Hayley Whitaker‡‡, Nening Dennis†, Christine Stuttle†, Ian G. Mills§§¶¶***, Ingrid Guldvik¶¶, Movember GAP1 Urine Biomarker Consortium1, Chris Parker†††, Daniel S. Brewer*‡‡‡, Colin S. Cooper* and Jeremy Clark*
*Norwich Medical School, University of East Anglia, Norwich, †Institute of Cancer Research, Sutton, ‡Norfolk and Norwich University Hospitals NHS Foundation Trust, Norwich, UK, §Department of Urology, Winship Cancer Institute, Emory University School of Medicine, Atlanta, GA, USA , ¶School of Biology and Environmental Science, Science West, University College Dublin, Dublin 4, Ireland, **Nijmegen Medical Centre, Radboud University Medical Centre, Nijmegen, The Netherlands, ††Faculty of Health and Medical Sciences, The University of Surrey, Guildford, ‡‡Molecular Diagnostics and Therapeutics Group, University College London, London, §§School of Medicine, Dentistry and Biomedical Sciences, Institute for Health Sciences, Centre for Cancer Research and Cell Biology, Queen’s University Belfast, Belfast, UK, ¶¶Centre for Molecular Medicine, University of Oslo, Oslo, Norway, ***Nuffield Department of Surgical Sciences, University of Oxford, Oxford, †††Royal Marsden Hospital, Sutton and ‡‡‡Earlham Institute, Norwich, UK
To develop a risk classifier using urine‐derived extracellular vesicle (EV)‐RNA capable of providing diagnostic information on disease status prior to biopsy, and prognostic information for men on active surveillance (AS).
Post‐digital rectal examination urine‐derived EV‐RNA expression profiles (n = 535, multiple centres) were interrogated with a curated NanoString panel. A LASSO‐based continuation ratio model was built to generate four prostate urine risk (PUR) signatures for predicting the probability of normal tissue (PUR‐1), D’Amico low‐risk (PUR‐2), intermediate‐risk (PUR‐3), and high‐risk (PUR‐4) prostate cancer. This model was applied to a test cohort (n = 177) for diagnostic evaluation, and to an AS sub‐cohort (n = 87) for prognostic evaluation.
Table 2. NanoString gene probes incorporated by LASSO regularization in the final optimal model used to produce the prostate urine risk signatures
Each PUR signature was significantly associated with its corresponding clinical category (P < 0.001). PUR‐4 status predicted the presence of clinically significant intermediate‐ or high‐risk disease (area under the curve = 0.77, 95% confidence interval [CI] 0.70–0.84). Application of PUR provided a net benefit over current clinical practice. In an AS sub‐cohort (n = 87), groups defined by PUR status and proportion of PUR‐4 had a significant association with time to progression (interquartile range hazard ratio [HR] 2.86, 95% CI 1.83–4.47; P < 0.001). PUR‐4, when used continuously, dichotomized patient groups with differential progression rates of 10% and 60% 5 years after urine collection (HR 8.23, 95% CI 3.26–20.81; P < 0.001).
Urine‐derived EV‐RNA can provide diagnostic information on aggressive prostate cancer prior to biopsy, and prognostic information for men on AS. PUR represents a new and versatile biomarker that could result in substantial alterations to current treatment of patients with prostate cancer.