Editor’s Choice
Know your nomograms The authors of the present article [1] have conducted an excellent evaluation and validation of a nomogram [1–3], designed to predict pathological stage, that is similar in design to those popularized by Kattan [3] in earlier work. In the present study, the authors carefully dissected the threshold at which the prediction of organ-confined disease provided the most accurate predictive ability. The predictive ability threshold of a 53% probability of organ-confined disease was chosen to provide physicians and patients with the most clinically sound estimate of pathological stage. Based on this prediction, decisions regarding whether or not to initiate an active surveillance strategy for very-low-risk disease, whether or not to undergo nerve-sparing surgery, or the need for pelvic lymph node dissection during robotic or perineal surgery can be made. Two important caveats to the use of nomograms must be considered. 1. Nomograms must be updated on a regular basis [2] (at least every 5 years). Major shifts in both the presentation grade and stage of prostate cancer have been witnessed over the past two decades. Many suggestions as to why this has occurred have been tendered; however, regardless of the cause, our nomograms must take this shift into account. When using a clinical nomogram, always check to see if the predictions have been updated. Nomograms that were determined >5 years before their present day use should be used with caution. 2. Since the first suggestion of clinically predictive nomograms, it has been reported numerous times that one nomogram, determined on a particular population of men,
might not predict well for another population. Whether this difference in population is in terms of race, geographic location, age, stage or grade demographic make-up, or ethnicity, one must use caution when using nomograms for one’s patients. For example, The Partin Tables were determined from a predominately white, young population of men, the majority of whom had low PSA levels and stage T1–T2 at presentation; these probabilities will not accurately predict pathological stage in a population of elderly, late stage, high grade tumours in Asian men.
Conflict of Interest None declared. Alan W. Partin The Brady Urological Institute, The Johns Hopkins Hospital, Baltimore, MD, USA e-mail: [email protected]
References 1
2
3
Borque A, Rubio-Briones J, Esteban LM et al. Implementing the use of nomograms by choosing cut-off points in predictive models. 2012 updated Partin Tables versus a European predictive nomogram for organ-confined disease in prostate cancer. BJU Int 2014; 113: 878–86 Eifler JB, Feng Z, Lin BM et al. An updated prostate cancer staging nomogram (Partin tables) based on cases from 2006 to 2011. BJU Int 2013; 111: 22–9 Kattan MW. Should I use this nomogram? BJU Int 2008; 102: 421–2
© 2014 The Author BJU International © 2014 BJU International
849
Know your nomograms The authors of the present article [1] have conducted an excellent evaluation and validation of a nomogram [1–3], designed to predict pathological stage, that is similar in design to those popularized by Kattan [3] in earlier work. In the present study, the authors carefully dissected the threshold at which the prediction of organ-confined disease provided the most accurate predictive ability. The predictive ability threshold of a 53% probability of organ-confined disease was chosen to provide physicians and patients with the most clinically sound estimate of pathological stage. Based on this prediction, decisions regarding whether or not to initiate an active surveillance strategy for very-low-risk disease, whether or not to undergo nerve-sparing surgery, or the need for pelvic lymph node dissection during robotic or perineal surgery can be made. Two important caveats to the use of nomograms must be considered. 1. Nomograms must be updated on a regular basis [2] (at least every 5 years). Major shifts in both the presentation grade and stage of prostate cancer have been witnessed over the past two decades. Many suggestions as to why this has occurred have been tendered; however, regardless of the cause, our nomograms must take this shift into account. When using a clinical nomogram, always check to see if the predictions have been updated. Nomograms that were determined >5 years before their present day use should be used with caution. 2. Since the first suggestion of clinically predictive nomograms, it has been reported numerous times that one nomogram, determined on a particular population of men,
might not predict well for another population. Whether this difference in population is in terms of race, geographic location, age, stage or grade demographic make-up, or ethnicity, one must use caution when using nomograms for one’s patients. For example, The Partin Tables were determined from a predominately white, young population of men, the majority of whom had low PSA levels and stage T1–T2 at presentation; these probabilities will not accurately predict pathological stage in a population of elderly, late stage, high grade tumours in Asian men.
Conflict of Interest None declared. Alan W. Partin The Brady Urological Institute, The Johns Hopkins Hospital, Baltimore, MD, USA e-mail: [email protected]
References 1
2
3
Borque A, Rubio-Briones J, Esteban LM et al. Implementing the use of nomograms by choosing cut-off points in predictive models. 2012 updated Partin Tables versus a European predictive nomogram for organ-confined disease in prostate cancer. BJU Int 2014; 113: 878–86 Eifler JB, Feng Z, Lin BM et al. An updated prostate cancer staging nomogram (Partin tables) based on cases from 2006 to 2011. BJU Int 2013; 111: 22–9 Kattan MW. Should I use this nomogram? BJU Int 2008; 102: 421–2
© 2014 The Author BJU International © 2014 BJU International
849
