Original article
External validation of Chun, PCPT, ERSPC, Kawakami, and Karakiewicz nomograms in the prediction of prostate cancer: A single center cohort-study☆

https://doi.org/10.1016/j.urolonc.2018.05.010Get rights and content

Highlights

  • •

    Nomograms and risk calculators are useful tools in the prediction of prostate cancer.

  • •

    Nomograms may perform differently depending on the population in examination.

  • •

    In our cohort the Chun nomogram outperformed the other 4 nomograms in the prediction of prostate cancer.

Abstract

Objectives

The aim of our study was to analyze the performance of 5 different risk calculators for prostate cancer diagnosis: Prostate Cancer Prevention Trial Risk Calculator (PCPT-RC), European Randomized Study of Screening for Prostate Cancer Risk Calculator (ERSP-RC), Karakiewicz nomogram, Chun nomogram, and Kawakami Nomogram.

Methods

From 2008 onwards, we consecutively enrolled, at a single institution in Italy, men undergoing 12-core transrectal ultrasound-guided prostate needle biopsy. Demographic, clinical, and pathological data were collected. The risk of prostate cancer (PCa) was calculated according to the PCPT-RC, ERSPC-RC, Karakiewicz, Kawakami, and Chun nomograms. Calibration and discrimination were assessed using calibration plots and receiver operator characteristic analysis. Additionally, decision curve analyses (DCA) were used to assess the net benefit associated with the adoption of each model.

Results

Overall, 1,100 patients were evaluated, 39% presented PCa and out of them 26% presented high-grade PCa (defined as Gleason ≥ 4 + 3). All the models showed good discrimination capacities for PCa on receiver operator characteristic analysis (area under the curve: 0.59–0.72) On calibration curves the ERSCP, the PCPT and the Chun nomogram underestimated the risk of PC while the Kawakami overestimated it. At DCA, the net benefit associated with the use of the models in the prediction of cancer was observed when the threshold probability was between 40% and 60%.

Conclusion

In a cohort of Italian men undergoing prostate biopsy, the performance accuracy of these calculators for the prediction prostate cancer is suboptimal. According to our experience the use of these calculator in clinical practice should be encouraged. Although integration with new serum/urine markers or magnetic resonance imaging results is warranted.

Introduction

Prostate cancer (PCa) is the second most commonly diagnosed malignancy in men, with an estimated 1.1 million diagnoses worldwide in 2012 [1]. PCa diagnosis is driven most of the times by a persistent elevation of prostate specific antigen (PSA) serum levels which triggers a prostate biopsy. However, PSA is far from ideal as a tumor marker as it is accuracy is low and leads to a high number of unnecessary prostate biopsies. In patients with PSA levels between 2 and 10 ng/ml only 20% to 40% will have a PCa diagnosis [2], [3].

To overcome the diagnostic limitations of PSA testing, the European Association of Urology guidelines recommend an individualized risk assessment of the patient [4]. According to the actual literature family history of PCa, age, abnormal digital rectal examination, serum markers, urine markers, and multiparametric magnetic resonance imaging (MRI) are useful tools combined with PSA levels to accurately predict PCa [4]. Lately, many authors have developed PCa risk calculators including some of this risk factors which enhance the diagnostic accuracy of PSA. However, none has clearly shown superiority so it remains a personal decision which one to use [5]. Although there is growing interest in mpMRI, genomic testing, and fusion biopsies for the diagnosis of PCa they are still not widely available. Risk calculators, which present the advantage to be cheaper, less time consuming and worldwide available are strongly recommended by the latest EAU guidelines as a valid alternative to serum/urine-based test (e.g., Prostate Health Index test, 4 kallikrein score, prostate cancer gene 3, HOXC6/DLX1) or mpMRI.

A recent meta-analysyis evaluated all the available calculators, the most-studied nomograms/calculators include the European Randomised Study of screening for PCa (ERSPC) [6], the north American prostate cancer prevention trial-based risk calculator (PCPT) [7], the Karakiewicz nomogram [8], the Chun nomogram [9], and the Kawakami nomogram [5], [10]. However, without applying and comparing all the 5 different available prediction models in a cohort of men undergoing PCa screening, conclusions cannot be made about model superiority because the estimated predictive accuracy of area under the curve (AUC) mainly reflects differences in population characteristics [5]. Aim of our study was to evaluate the diagnostic performance of 5 different risk calculators/nomograms for the prediction of PCa in a cohort of patients at increased risk of PCa undergoing prostate biopsies.

Section snippets

Materials and methods

After an Internal Review Board approval, between 2008 and 2016, a consecutive series of patients with no known history of PCa were referred to our department to undergo initial prostate biopsy because of an abnormal finding on digital rectal examination (DRE) or an elevation of serum levels of PSA (>4 ng/ml). All patients signed a dedicated informed consent, all the procedures were approved by the local ethics committee and the study was conducted in accordance with the principles of the

Results

Overall, 437/1,100 (39.8%) patients were diagnosed with prostate cancer and out of them 115/437 (26.3%) had high-grade cancer (Gleason score ≥ 4 + 3). Characteristics of the patient cohort and according to the presence or absence of cancer are listed in Table 2. Patients with positive biopsies were older, had smaller prostates, higher levels of PSA, lower levels of free PSA and were more likely to have a positive DRE. Age, TRUS volume, PSA, free PSA and DRE were found to be significant

Discussion

Our present data are supported by previous studies showing a negative correlation between PV and prostate cancer diagnosis at biopsy [12]. Overall, our present prostate cancer detection rate (39.8%) is consistent with the available series in the extended biopsy protocol era [13], [14]. All these observations confirm the internal validity of the study and they are consistent with data published in the peer review literature. In our experience we confirmed as the PCa nomograms overcome the

Conclusions

In our experience the Chun, PCPT, ERSPC, Kawakami, and Karakiewicz nomograms present a suboptimal accuracy (AUC range: 0.585–0.726) for the detection of PCa and have no role for the diagnosis of high-grade cancer. Their implementation in clinical practice is needed for definitive conclusion. Future studies should also integrate these nomograms with investigational parameters as metabolic status, serum/urine markers or MRI results.

Authorship

Authors have made a substantial contribution to the following:

Cosimo De Nunzio: research design, acquisition, analysis and interpretation of data; draft the paper, approved the submitted and final versions.

Riccardo Lombardo research design, acquisition, analysis and interpretation of data; draft the paper, approved the submitted and final versions.

Giorgia Tema research design, acquisition of data, approved the submitted and final versions.

Hassan Alkhatatbeh research design, acquisition of data,

References (29)

  • J.I. Epstein et al.

    A contemporary prostate cancer grading system: a validated alternative to the Gleason score

    Eur Urol

    (2016)
  • S. Loeb et al.

    Evaluation of the 2015 Gleason grade groups in a nationwide population-based cohort

    Eur Urol

    (2016)
  • C.J. Kane et al.

    Variability in outcomes for patients with intermediate-risk prostate cancer (Gleason score 7, International Society of Urological Pathology Gleason Group 2–3) and Implications for risk stratification: a systematic review

    Eur Urol Focus

    (2017)
  • T. Jindal et al.

    Racial differences in prostate-specific antigen-based prostate cancer screening: state-by-state and region-by-region analyses

    Urol Oncol

    (2017)

    • Cited by (18)

    • The Capio Prostate Cancer Center Model for Prostate Cancer Diagnostics—Real-world Evidence from 2018 to 2022

      2024, European Urology Open Science

    • External Validation of the Rotterdam Prostate Cancer Risk Calculator and Comparison with Stockholm3 for Prostate Cancer Diagnosis in a Swedish Population-based Screening Cohort

      2023, European Urology Focus
      Citation Excerpt :

      Therefore, calibration, clinical benefits, and head-to-head comparisons of the risk calculators are needed for urologists to make informed decisions to perform biopsies in the clinical workup of prostate cancer. Previous studies externally validating the RPCRC have shown mixed results [7–11], with adequate discrimination for significant cancer, but depending on the study cohort, varied results on calibration and clinical usefulness. A head-to-head comparison between the Stockholm3 test and the RPCRC has never been reported.

    • External Validation of the Prostate Biopsy Collaborative Group Risk Calculator and the Rotterdam Prostate Cancer Risk Calculator in a Swedish Population-based Screening Cohort

      2022, European Urology Open Science
      Citation Excerpt :

      One concern with risk models is their external validity and clinical usefulness beyond the setting in which they were developed. It is common that models need to be recalibrated to fit the local setting to improve accuracy of risk predictions [5]. With increasing numbers of risk prediction tools [2], external validation, local recalibration, and head-to-head comparisons are needed to present accurate clinical recommendations.

    • Assessing the diagnostic performance of systematic freehand PrecisionPoint transperineal prostate biopsy: Comparison of observed outcomes to PBCG nomogram predictions

      2022, Urologic Oncology: Seminars and Original Investigations
      Citation Excerpt :

      This more contemporary nomogram replaced the Prostate Cancer Prevention Trial (PCPT) nomogram which was considered the standard nomogram for assessment of prostate cancer risk[21]. The PCPT risk calculator is largely based on data from the 1990s utilizing a sextant biopsy scheme which may underestimate prostate cancer risk a weakness which led to the development of a more updated tool [22,23]. Notably, the biopsies performed in the PBCG cohort were TRUS guided and vast majority were without a pre-biopsy mpMRI.

    • Mobile phone apps for the prediction of prostate cancer: External validation of the Coral and Rotterdam apps

      2019, European Journal of Surgical Oncology
      Citation Excerpt :

      Another possible limitation is that prostate volume measured with TRUS is not as accurate as that measured with MRI. Although EAU guidelines and the most updated risk calculators still recommend TRUS as the standard method to evaluate prostate volume [18–20]. Moreover, to validate a model it is important to stay as close as possible to the development cohort which used TRUS volume.

    • Residents and Consultants Have Equal Outcomes When Performing Transrectal Fusion Biopsies: A Randomized Clinical Trial

      2024, Current Oncology
    View all citing articles on Scopus
    ☆

    Institution where the work was done—Department of Urology, “Sant’Andrea” Hospital, “La Sapienza” University, via di Grottarossa 1039, 00185 Rome, Italy.

    View full text
    © 2018 Elsevier Inc. All rights reserved.