Magnetic Resonance Parkinsonism Index: diagnostic accuracy of a fully automated algorithm in comparison with the manual measurement in a large Italian multicentre study in patients with progressive supranuclear palsy

Salvatore Nigro, Gennarina Arabia, Angelo Antonini, Luca Weis, Andrea Marcante, Alessandro Tessitore, Mario Cirillo, Gioacchino Tedeschi, Stefano Zanigni, Giovanna Calandra-Buonaura, Caterina Tonon, Gianni Pezzoli, Roberto Cilia, Mario Zappia, Alessandra Nicoletti, Calogero Edoardo Cicero, Michele Tinazzi, Pierluigi Tocco, Nicolò Cardobi, Aldo Quattrone

Research output: Contribution to journalArticle

Abstract

Objectives: To investigate the reliability of a new in-house automatic algorithm for calculating the Magnetic Resonance Parkinsonism Index (MRPI), in a large multicentre study population of patients affected by progressive supranuclear palsy (PSP) or Parkinson’s disease (PD), and healthy controls (HC), and to compare the diagnostic accuracy of the automatic and manual MRPI values. Methods: The study included 88 PSP patients, 234 PD patients and 117 controls. MRI was performed using both 3T and 1.5T scanners. Automatic and manual MRPI values were evaluated, and accuracy of both methods in distinguishing PSP from PD and controls was calculated. Results: No statistical differences were found between automated and manual MRPI values in all groups. The automatic MRPI values differentiated PSP from PD with an accuracy of 95 % (manual MRPI accuracy 96 %) and 97 % (manual MRPI accuracy 100 %) for 1.5T and 3T scanners, respectively. Conclusion: Our study showed that the new in-house automated method for MRPI calculation was highly accurate in distinguishing PSP from PD. Our automatic approach allows a widespread use of MRPI in clinical practice and in longitudinal research studies. Key Points: • A new automatic method for calculating the MRPI is presented.• Automatic MRPI values are in good agreement with manual values.• Automatic MRPI can distinguish patients with PSP from patients with PD.• The automatic method overcomes MRPI application limitations in routine practice.• The automatic method may allow a more widespread use of MRPI.

Original languageEnglish
Pages (from-to)1-11
Number of pages11
JournalEuropean Radiology
DOIs
Publication statusAccepted/In press - Oct 19 2016

Keywords

  • Automatic segmentation
  • Magnetic Resonance Parkinsonism Index
  • Neurodegenerative disease
  • Parkinson’s disease
  • Progressive supranuclear palsy

ASJC Scopus subject areas

  • Radiology Nuclear Medicine and imaging

Fingerprint Dive into the research topics of 'Magnetic Resonance Parkinsonism Index: diagnostic accuracy of a fully automated algorithm in comparison with the manual measurement in a large Italian multicentre study in patients with progressive supranuclear palsy'. Together they form a unique fingerprint.

  • Cite this

    Nigro, S., Arabia, G., Antonini, A., Weis, L., Marcante, A., Tessitore, A., Cirillo, M., Tedeschi, G., Zanigni, S., Calandra-Buonaura, G., Tonon, C., Pezzoli, G., Cilia, R., Zappia, M., Nicoletti, A., Cicero, C. E., Tinazzi, M., Tocco, P., Cardobi, N., & Quattrone, A. (Accepted/In press). Magnetic Resonance Parkinsonism Index: diagnostic accuracy of a fully automated algorithm in comparison with the manual measurement in a large Italian multicentre study in patients with progressive supranuclear palsy. European Radiology, 1-11. https://doi.org/10.1007/s00330-016-4622-x