Could Machine Learning Improve the Prediction of Pelvic Nodal Status of Prostate Cancer Patients? Preliminary Results of a Pilot Study

ABSTRACT

We tested and compared performances of Roach formula, Partin tables and of three Machine Learning (ML) based algorithms based on decision trees in identifying N+ prostate cancer (PC). 1,555 cN0 and 50 cN+ PC were analyzed. Results were also verified on an independent population of 204 operated cN0 patients, with a known pN status (187 pN0, 17 pN1 patients).

ML performed better, also when tested on the surgical population, with accuracy, specificity, and sensitivity ranging between 48–86%, 35–91%, and 17–79%, respectively. ML potentially allows better prediction of the nodal status of PC, potentially allowing a better tailoring of pelvic irradiation.

Keywords: :

• Machine Learning
• Pelvic irradiation
• Radiotherapy
• Prostate cancer
• Nodal metastases

DECLARATION OF INTEREST

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.

Role of the Funding Source: This work was partly supported by “ex 60%” grants from the Brescia University; this is a public Institution which, as such, had no role in study design, data collection analysis, and interpretation of data. The web-based software needed to collect the data and the related applications were realized by Roberto GATTA, BSc, CEng, partly thanks to an unrestricted grant of the Regione Lombardia, supporting the activity of the Oncology Department of the Brescia County. This, again, is a public Institution which, as such, had no role in study design, data collection analysis, and interpretation of data.

ACKNOWLEDGMENTS

We are indebted with Barbara JERECZECK FOSSA, MD, who coordinated the AIRO Prostate Cancer Study Group during the study and provided organizational support. We acknowledge the general support provided by the following Heads of Department: Cynthia ARISTEI, MD, Prof. (Perugia); Enza BARBIERI, MD, Prof. (†) (Bologna); Lorenzo LIVI, MD, Prof. (Florence); Alberto BONETTA, MD (Cremona); Dorian COSENTINO, MD (Como); Sandro FONGIONE, MD (Udine); Paola FRANZONE, MD (Alessandria); Paolo MUTO (Napoli); Pietro PONTICELLI, MD, (Arezzo), Riccardo SANTONI, MD, Prof. (Roma Tor Vergata), Alessandro TESTOLIN (Belluno), and Alessia GUARNERI, MD (Torino). We also acknowledge the very important support provided by Dr. Roberto D'AMICO (Department of Diagnostic Medicine, Clinics and Public Health, Modena University) to improve the methodogical aspects of this study. The following Colleagues were also involved in the data collection process and/or general support: Giuseppina APICELLA, MD (Novara); Debora BELDÌ, MD (Novara); Filippo DE RENZI, MD (Belluno); Pietro Giovanni GENNARI, MD (Arezzo); Gianluca INGROSSO, MD (Roma Tor Vergata); Fernando MUNOZ, MD, (Torino); Andrea RAMPINI (Arezzo); Enzo RAVO (Napoli).