Tag Archive for: prediction nomogram

Posts

Article of the Week: Risk prediction tool for grade re-classification in men with favourable-risk prostate cancer on active surveillance

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 accompanying editorial written by a prominent member of the urological community. This blog is 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.

Risk prediction tool for grade re-classification in men with favourable-risk prostate cancer on active surveillance

Mufaddal M. Mamawala, Karthik Rao, Patricia Landis, Jonathan I. EpsteinBruce J. Trock, Jeffrey J. Tosoian, Kenneth J. Pienta and H. Ballentine Carter

 

The James Buchanan Brady Urological Institute, Johns Hopkins Medical Institutions, Baltimore, MD, USA

 

How to Cite

Mamawala, M. M., Rao, K., Landis, P., Epstein, J. I., Trock, B. J., Tosoian, J. J., Pienta, K. J. and Carter, H. B. (2017), Risk prediction tool for grade re-classification in men with favourable-risk prostate cancer on active surveillance. BJU International, 120: 25–31. doi: 10.1111/bju.13608

Objective

To create a nomogram for men on active surveillance (AS) for prediction of grade re-classification (GR) above Gleason score 6 (Grade group >2) at surveillance biopsy.

Patients and Methods

From a cohort of men enrolled in an AS programme, a multivariable model was used to identify clinical and pathological parameters predictive of GR. Nomogram performance was assessed using receiver operating characteristic curves, calibration, and decision curve analysis.

aotw-jul-2017-5-results

Results

Of 1 374 men, 254 (18.50%) were re-classified to Gleason ≥7 on surveillance prostate biopsy. Variables predictive of GR were earlier year of diagnosis [≤2004 vs ≥2005; odds ratio (OR) 2.16, P < 0.001], older age (OR 1.05, P < 0.001), higher prostate-specific antigen density [OR 1.19 (per 0.1 unit increase), P = 0.04], bilateral disease (OR 2.86, P < 0.001), risk strata (low-risk vs very-low-risk, OR 1.79, P < 0.001), and total number of biopsies without GR (OR 0.68, P < 0.001). On internal validation, a nomogram created using the multivariable model had an area under the curve of 0.757 (95% confidence interval 0.730–0.797) for predicting GR at the time of next surveillance biopsy.

Conclusion

The nomogram described is currently being used at each return visit to assess the need for a surveillance biopsy, and could increase retention in AS.

Editorial: Shift from protocol-based to personalized medicine in active surveillance: beginning of a new era

The use of active surveillance (AS) is rapidly expanding worldwide, with rates as high as 74% among patients with low-risk prostate cancer in the nationwide registry of Sweden [1]. Despite increasing uptake of this strategy by patients, there is no consensus among the medical community as to the ideal criteria for selection and monitoring [2]. For example, the Johns Hopkins AS programme restricts enrolment to men with low-risk disease and performs annual biopsies for monitoring. Other protocols also include men with intermediate-risk disease and perform prostate biopsy at less frequent intervals.

Is it really optimal to use the same follow-up protocol for all patients? Many factors influence the risk of reclassification, including patient characteristics (e.g. race, body mass index) and disease features (e.g. PSA density, Gleason score and extent of disease on biopsy) [3]. Moreover, previous studies have shown that the risk of reclassification during AS is a conditional probability, where the risk decreases with each additional negative biopsy [4]. Given that individual patients have vastly different risks of reclassification, and that the risk changes over time, AS represents an ideal context for personalized medicine.

There has already been a significant paradigm shift in prostate cancer screening from a one-size-fits-all to a multivariable, risk-adapted approach [5]. Why would we use the same screening intervals and biopsy cutoff for patients with vastly different risk profiles? Multiple guidelines already recommend using PSA levels to guide screening protocols, and there are several validated multivariable tools to provide more personalized estimates of prostate cancer risk. Both the Prostate Cancer Prevention Trial (PCPT) and the European Randomised Study of Screening for Prostate Cancer (ERSPC) risk calculators have been extensively studied and are readily available online for use in clinical practice [6].

To date, the concept of risk-adapted AS has received relatively little attention, and few nomograms have been created specifically for the AS population. Using data from the Canary Prostate Active Surveillance Study (PASS), Ankerst et al. [7] designed a nomogram to predict biopsy reclassification using age at biopsy, months since the last biopsy, last PSA level, percentage of cores positive for cancer on the last biopsy, and number of previous negative biopsies. This tool had an area under the curve (AUC) of 0.724 on internal validation, and is available online at https://prostate-cancer-risk-calculator.org to facilitate additional validation and clinical use.

In the current issue of BJUI, Mamawala et al. [8] report on the development of another new AS nomogram using data from the Johns Hopkins programme. Specifically, the tool predicts the risk of biopsy reclassification using six variables: age; PSA density; year of diagnosis; laterality; risk strata; and total number of biopsies. The nomogram was well calibrated and had an AUC of 0.757 on internal validation. Notably, the same authors have also recently developed a different tool to predict pathological Gleason score for men on AS using a Bayesian joint model [9]. Following external validation, these tools may help provide more customized decision support for the AS population by integrating longitudinal data.

It is noteworthy that none of these nomograms incorporate new markers or imaging, and it is likely that such data could further refine their estimates. For example, longitudinal measurements of the Prostate Health Index were previously shown to predict biopsy reclassification during AS [10], and the use of multiparametric MRI continues to expand. As more data on these tests become available, the AS risk calculators should be updated, as has been done with the PCPT and ERSPC risk calculators used in the screening context. In the future, continued research on genetics may allow further tailoring of AS. In the meantime, these risk calculators are an important first step (‘version 1.0’) toward a more personalized approach to AS.

Stacy Loeb

 

Department of Urology, Population Health, Laura & Isaac Perlmutter Cancer Center, New York University, New YorkNY, US

 

References

 

1 Loeb S, Folkvaljon Y, Curnyn C, Robinson D, Bratt O, Stattin P. Almost complete uptake of active surveillance for very low-risk prostate cancer in Sweden. JAMA Oncol 2016; [Epub ahead of print]. doi: 10.1001/ jamaoncol.2016.3600

 

2 Tosoian JJ, Carter HB, Lepor A, Loeb S. Active surveillance for prostate cancer: current evidence and contemporary state of practice. Nat Rev Urol 2016; 13: 20515

 

 

4 Alam R, Carter HB, Landis P, Epstein JI, Mamawala M. Conditional probability of reclassication in an active surveillance program for prostate cancer. J Urol 2015; 193: 19505

 

 

 

 

 

9 ColeyRY, Zeger S L, Mamawala M, Pienta KJ, Carter HBPrediction of the pathologic gleason score to inform a personalized management program for prostate cancer. Eur Urol 2016; [Epub ahead of print]. doi: 10.1016/j.eururo.2016.08.005

 

 

Video: Risk prediction tool for grade re-classification in men with favourable-risk prostate cancer on active surveillance

Risk prediction tool for grade re-classification in men with favourable-risk prostate cancer on active surveillance

Abstract

Objective

To create a nomogram for men on active surveillance (AS) for prediction of grade re-classification (GR) above Gleason score 6 (Grade group >2) at surveillance biopsy.

Patients and Methods

From a cohort of men enrolled in an AS programme, a multivariable model was used to identify clinical and pathological parameters predictive of GR. Nomogram performance was assessed using receiver operating characteristic curves, calibration, and decision curve analysis.

Results

Of 1 374 men, 254 (18.50%) were re-classified to Gleason ≥7 on surveillance prostate biopsy. Variables predictive of GR were earlier year of diagnosis [≤2004 vs ≥2005; odds ratio (OR) 2.16, P < 0.001], older age (OR 1.05, P < 0.001), higher prostate-specific antigen density [OR 1.19 (per 0.1 unit increase), P = 0.04], bilateral disease (OR 2.86, P < 0.001), risk strata (low-risk vs very-low-risk, OR 1.79, P < 0.001), and total number of biopsies without GR (OR 0.68, P < 0.001). On internal validation, a nomogram created using the multivariable model had an area under the curve of 0.757 (95% confidence interval 0.730–0.797) for predicting GR at the time of next surveillance biopsy.

Conclusion

The nomogram described is currently being used at each return visit to assess the need for a surveillance biopsy, and could increase retention in AS.

View more videos
© 2024 BJU International. All Rights Reserved.