Discovery and validation of a serum microRNA signature to characterize oligo- and polymetastatic prostate cancer: not ready for prime time
- 478 Downloads
Patients with oligometastatic prostate cancer (PC) may benefit from metastasis-directed therapy (MDT), delaying disease progression and the start of palliative systemic treatment. However, a significant proportion of oligometastatic PC patients progress to polymetastatic PC within a year following MDT, suggesting an underestimation of the metastatic load by current staging modalities. Molecular markers could help to identify true oligometastatic patients eligible for MDT.
Patients with asymptomatic biochemical recurrence following primary PC treatment were classified as oligo- or polymetastatic based on 18F-choline PET/CT imaging. Oligometastatic patients had up to three metastases at baseline and did not progress to more than three lesions following MDT or surveillance within 1 year of diagnosis of metastases. Polymetastatic patients had > 3 metastases at baseline or developed > 3 metastases within 1 year following imaging. A model aiming to prospectively distinguish oligo- and polymetastatic PC patients was trained using clinicopathological parameters and serum-derived microRNA expression profiles from a discovery cohort of 20 oligometastatic and 20 polymetastatic PC patients. To confirm the models predictive performance, it was applied on biomarker data obtained from an independent validation cohort of 44 patients with oligometastatic and 39 patients with polymetastatic disease.
Oligometastatic PC patients had a more favorable prognosis compared to polymetastatic ones, as defined by a significantly longer median CRPC-free survival (not reached versus 38 months; 95% confidence interval 31–45 months with P < 0.001). Despite the good performance of a predictive model trained on the discovery cohort, with an AUC of 0.833 (0.693–0.973; 95% CI) and a sensitivity of 0.894 (0.714–1.000; 95% CI) for oligometastatic disease, none of the miRNA targets were found to be differentially expressed between oligo- and polymetastatic PC patients in the signature validation cohort. The multivariate model had an AUC of 0.393 (0.534 after cross-validation) and therefore, no predictive ability.
Although PC patients with oligometastatic disease had a more favorable prognosis, no serum-derived biomarkers allowing for prospective discrimination of oligo- and polymetastatic prostate cancer patients could be identified.
KeywordsProstate cancer Oligometastasis miRNA Serum Biomarker Machine learning
This work was supported by “Kom op tegen Kanker (Stand up to Cancer), the Flemisch cancer society” (Bert Dhondt: Emmanuel Vander Schueren Research Grant). Piet Ost is a senior clinical investigator of the Research Foundation—Flanders, Belgium.
BD sample collection, data analysis and reporting, data management, manuscript writing. EDB data management, manuscript writing. TC sample collection, data collection and management. SB sample collection, manuscript editing. NL manuscript editing. JV project development, manuscript editing. AB project development. VF data collection and management, manuscript editing. JP data analysis and reporting. PG data analysis and reporting, manuscript editing. PO project development, data collection and management, manuscript editing. All authors approved the final version of the manuscript.
- 14.Perera M, Papa N, Christidis D, Wetherell D, Hofman MS, Murphy DG et al (2016) Sensitivity, specificity, and predictors of positive (68)Ga-prostate-specific membrane antigen positron emission tomography in advanced prostate cancer: a systematic review and meta-analysis. Eur Urol 70(6):926–937CrossRefGoogle Scholar