Quantitative parameters in dynamic contrast-enhanced magnetic resonance imaging for the detection and characterization of prostate cancer
Metrics: PDF 857 views | HTML 1546 views | ?
Cheng Wei1,*, Bowen Jin1,5,*, Magdalena Szewczyk-Bieda1,2, Stephen Gandy3, Stephen Lang4, Yilong Zhang5, Zhihong Huang5 and Ghulam Nabi1
1Division of Cancer Research, School of Medicine, University of Dundee, Ninewells Hospital, Dundee DD1 9SY, UK
2Department of Clinical Radiology, Ninewells Hospital, Dundee DD1 9SY, UK
3Department of Medical Physics, Ninewells Hospital, Dundee DD1 9SY, UK
4Department of Pathology, Ninewells Hospital, Dundee DD1 9SY, UK
5School of Science and Engineering, University of Dundee, Dundee DD1 4HN, UK
*These authors have contributed equally to this work
Ghulam Nabi, email: firstname.lastname@example.org
Keywords: prostate cancer; dynamic contrast-enhanced magnetic resonance imaging; multi-parametric magnetic resonance imaging; kinetic models
Received: December 15, 2017 Accepted: February 25, 2018 Published: March 23, 2018
Objectives: to assess the diagnostic accuracy of quantitative parameters of DCE-MRI in multi-parametric MRI (mpMRI) in comparison to the histopathology (including Gleason grade) of prostate cancer.
Patients and methods: 150 men with suspected prostate cancer (abnormal digital rectum examination and or elevated prostate-specific antigen) received pre-biopsy 3T mpMRI and were recruited into peer-reviewed, protocol-based prospective study. The DCE-MRI quantitative parameters (Ktrans (influx transfer constant) and kep (efflux rate constant)) of the cancerous and normal areas were recorded using four different kinetic models employing Olea Sphere (Olea Medical, La Ciotat, France). The correlation between these parameters and the histopathology of the lesions (biopsy and in a sub-cohort 41 radical prostatectomy specimen) was assessed.
Results: The quantitative parameters showed a significant difference between non-cancerous (benign) and cancerous lesions (Gleason score≥3+3) in the prostate gland. The cut-off values for prostate cancer differentiation were: Ktrans (0.205 min-1) and kep (0.665 min-1) in the extended Tofts model (ET) and Ktrans(0.205 min-1 and kep (0.63 min-1) in the Lawrence and Lee delay (LD) models respectively. The mean Ktrans value also showed a difference between low-grade cancer (Gleason score=3+3) and high-grade cancer (Gleason score ≥ 3+4). With the addition of DCE-MRI quantitative parameters, the sensitivity of the PIRAD scoring system was increased from 56.6% to 92.1% (Ktrans_ET), 93.1% (kep_ET), 91.0%, (Ktrans_LD) and 89.4% (kep_LD).
Conclusion: Quantitative DCE-MRI parameters improved the diagnostic performance of conventional MRI in distinguishing normal and prostate cancers, including characterization of grade of cancers. The ET and LD models in post-image processing analysis provided better cut-off values for prostate cancer differentiation than the other quantitative DCE-MRI parameters.
All site content, except where otherwise noted, is licensed under a Creative Commons Attribution 3.0 License.