Apples to apples

Article Outline

• References
• Copyright

The lack of Level I evidence comparing various treatments for localized prostate cancer allows one to profess or proclaim a bias for almost any modality they choose. The lack of “good” evidence has confounded the issue by allowing those with favoritism toward any one modality with less robust data, such as proton beam therapy, stereotactic body radiation, or high intensity frequency ultrasound to have a voice in the chorus of identical outcomes. In the end, either side can state their case at a tumor board or with patients with impunity. The patient's confusion in selecting appropriate therapy is further complicated by the physician's bias that may be linked to ego or the pocketbook.

It is generally assumed that external beam radiation therapy, radical prostatectomy (RP), and prostate brachytherapy (BT) offer equal outcomes for most men with localized prostate cancer. A recent prospective study comparing BT and RP published by Giberti et al. (1) from Italy identifies an expected result, namely, that there is an equal biochemical freedom from recurrence of 91% RP and 91.7% BT. Yet the tumor board critics already are weighing in that the biochemical outcomes are measured differently, thus, creating the voices of doubt once again. Close, but still an apple to orange study.

So how to make a better and accurate comparison? Advanced predictive modeling using nonconformal methods to calculate outcome risk have been developed and implemented for prostate cancer (2). These models, nomograms, are applicable for patient decision discussions and for research groupings that better define risk stratification from that of standard prognostic groupings. In fact, the concordance index for nomograms is more accurate than traditional risk stratification.

But a nomogram is only as good the center from which it arises. And modalities use different outcome definitions. So outcome predictions are limited and modality comparisons on different nomograms, say for external beam radiation therapy and RP, raise the same apples to oranges comparison.

So when it comes to a tumor board discussion comparing modalities, most if not all comparisons to date are flawed based on various reasons; outcome definitions, physician experience, years treated, robustness of the dataset, and the like.

So, finally an apples to apples comparison. The study by Pickles et al. (3) solves several “tumor board” issues comparing RP and BT. First, they use the same outcome definition from the surgical nomogram despite their (correct) notation that the surgical outcome definition overestimates BT failures. Second, they account for both surgical and BT experience as a component of the analysis. Next, they use the best statistical tool for predicting outcomes, the nomogram as the root of their analysis. But unique to other nomogram studies, they inject their BT cohort outcomes into the surgical model to create a level-field comparison of BT and RP. And lastly, their study cohort is of a robust enough size that these outcomes are meaningful.

Finally, an opportunity for all to copy Figure 2 from the Pickles et al. study and make it a fixture in one's PowerPoint library comparing BT and RP.

Back to Article Outline

References 

1. Giberti C, Chiono L, Gallo F, et al. Radical retropubic prostatectomy versus brachytherapy for low-risk prostatic cancer: A prospective study. World J Urol. 2009;27:607–612
   • View In Article
   • CrossRef
2. Kattan MW, Vickers A, Yu C, et al. Preoperative and postoperative nomograms incorporating surgeon experience for localized prostate cancer. Cancer. 2009;115:1005–1010
   • View In Article
   • CrossRef
3. Pickles T, Morris WJ, Kattan MW, et al. Comparative 5-year outcomes of brachytherapy and surgery for prostate cancer. Brachytherapy. 2011;1:9–14
   • View In Article