Standardization of MRI and scintigraphic scores for assessing the severity of bone marrow involvement in adult patients with type 1 gaucher disease

Giuliano Mariani, Marzio Perri, Fabrizio Minichilli, Simona Ortori, Silvia Linari, Fiorina Giona, Maja Di Rocco, Maria Domenica Cappellini, Federica Guidoccio, Paola A. Erba

Research output: Contribution to journalArticle

Abstract

OBJECTIVE. MRI and 99mTc-sestamibi scintigraphy are used to estimate bone marrow infiltration in patients with Gaucher disease (GD), but comparison of data obtained at different institutions is difficult because different scores are employed for semiquantitative assessment. We developed normalized scores for comparing data both within a single method (MRI) and between different methods (MRI versus scintigraphy). MATERIALS AND METHODS. We evaluated 51 patients with type 1 GD (26 women, 25 men; mean age ± SD, 36.3 ± 10.9 years old). T1- and T2-weighted turbo spin-echo sequences at 1.5 T served to derive the bone marrow burden score (0-16), the vertebra-disk ratio (VDR), the Terk score (0-3), and the Spanish-MRI score (S-MRI, 0-24). Scintigraphy was scored between 0 and 8. Each score was normalized into four categories: 0 = normal, 1 = mild, 2 = intermediate, 3 = severe involvement. Interobserver and intraobserver agreements were evaluated by kappa statistics; nonparametric statistics with Bonferroni correction assessed correlations among the various original and normalized scores. RESULTS. Interobserver agreement was excellent for the original scores (κ = 0.730- 0.843) and even more so for the normalized scores (κ = 0.775-0.940). Intraobserver agreement kappa values ranged from 0.753 to 0.937 for the original scores and 0.851 to 1.000 for the normalized scores. Highly significant correlations were found among the various original scores (r = 0.42-0.86, p values between 0.0296 and < 0.0001), except for VDR versus S-MRI and Terk. Normalization generally induced marginal reductions in statistical significance, whereas S-MRI versus VDR reached statistical significance with the normalized scores. CONCLUSION. Our data indicate no significant loss of statistical information is caused by the normalization we employed. Our approach therefore facilitates comparison of different scores obtained in different institutions with different imaging modalities.

Original languageEnglish
Pages (from-to)1245-1252
Number of pages8
JournalAmerican Journal of Roentgenology
Volume206
Issue number6
DOIs
Publication statusPublished - Jun 1 2016

Fingerprint

Gaucher Disease
Radionuclide Imaging
Spine
Bone Marrow
Technetium Tc 99m Sestamibi
Nonparametric Statistics

Keywords

  • Bone marrow
  • Gaucher disease
  • MRI
  • Scintigraphy

ASJC Scopus subject areas

  • Radiology Nuclear Medicine and imaging

Cite this

Standardization of MRI and scintigraphic scores for assessing the severity of bone marrow involvement in adult patients with type 1 gaucher disease. / Mariani, Giuliano; Perri, Marzio; Minichilli, Fabrizio; Ortori, Simona; Linari, Silvia; Giona, Fiorina; Di Rocco, Maja; Cappellini, Maria Domenica; Guidoccio, Federica; Erba, Paola A.

In: American Journal of Roentgenology, Vol. 206, No. 6, 01.06.2016, p. 1245-1252.

Research output: Contribution to journalArticle

Mariani, Giuliano ; Perri, Marzio ; Minichilli, Fabrizio ; Ortori, Simona ; Linari, Silvia ; Giona, Fiorina ; Di Rocco, Maja ; Cappellini, Maria Domenica ; Guidoccio, Federica ; Erba, Paola A. / Standardization of MRI and scintigraphic scores for assessing the severity of bone marrow involvement in adult patients with type 1 gaucher disease. In: American Journal of Roentgenology. 2016 ; Vol. 206, No. 6. pp. 1245-1252.
@article{139b2495d17a49dfa6386232fefd5182,
title = "Standardization of MRI and scintigraphic scores for assessing the severity of bone marrow involvement in adult patients with type 1 gaucher disease",
abstract = "OBJECTIVE. MRI and 99mTc-sestamibi scintigraphy are used to estimate bone marrow infiltration in patients with Gaucher disease (GD), but comparison of data obtained at different institutions is difficult because different scores are employed for semiquantitative assessment. We developed normalized scores for comparing data both within a single method (MRI) and between different methods (MRI versus scintigraphy). MATERIALS AND METHODS. We evaluated 51 patients with type 1 GD (26 women, 25 men; mean age ± SD, 36.3 ± 10.9 years old). T1- and T2-weighted turbo spin-echo sequences at 1.5 T served to derive the bone marrow burden score (0-16), the vertebra-disk ratio (VDR), the Terk score (0-3), and the Spanish-MRI score (S-MRI, 0-24). Scintigraphy was scored between 0 and 8. Each score was normalized into four categories: 0 = normal, 1 = mild, 2 = intermediate, 3 = severe involvement. Interobserver and intraobserver agreements were evaluated by kappa statistics; nonparametric statistics with Bonferroni correction assessed correlations among the various original and normalized scores. RESULTS. Interobserver agreement was excellent for the original scores (κ = 0.730- 0.843) and even more so for the normalized scores (κ = 0.775-0.940). Intraobserver agreement kappa values ranged from 0.753 to 0.937 for the original scores and 0.851 to 1.000 for the normalized scores. Highly significant correlations were found among the various original scores (r = 0.42-0.86, p values between 0.0296 and < 0.0001), except for VDR versus S-MRI and Terk. Normalization generally induced marginal reductions in statistical significance, whereas S-MRI versus VDR reached statistical significance with the normalized scores. CONCLUSION. Our data indicate no significant loss of statistical information is caused by the normalization we employed. Our approach therefore facilitates comparison of different scores obtained in different institutions with different imaging modalities.",
keywords = "Bone marrow, Gaucher disease, MRI, Scintigraphy",
author = "Giuliano Mariani and Marzio Perri and Fabrizio Minichilli and Simona Ortori and Silvia Linari and Fiorina Giona and {Di Rocco}, Maja and Cappellini, {Maria Domenica} and Federica Guidoccio and Erba, {Paola A.}",
year = "2016",
month = "6",
day = "1",
doi = "10.2214/AJR.15.15294",
language = "English",
volume = "206",
pages = "1245--1252",
journal = "American Journal of Roentgenology",
issn = "0361-803X",
publisher = "American Roentgen Ray Society",
number = "6",

}

TY - JOUR

T1 - Standardization of MRI and scintigraphic scores for assessing the severity of bone marrow involvement in adult patients with type 1 gaucher disease

AU - Mariani, Giuliano

AU - Perri, Marzio

AU - Minichilli, Fabrizio

AU - Ortori, Simona

AU - Linari, Silvia

AU - Giona, Fiorina

AU - Di Rocco, Maja

AU - Cappellini, Maria Domenica

AU - Guidoccio, Federica

AU - Erba, Paola A.

PY - 2016/6/1

Y1 - 2016/6/1

N2 - OBJECTIVE. MRI and 99mTc-sestamibi scintigraphy are used to estimate bone marrow infiltration in patients with Gaucher disease (GD), but comparison of data obtained at different institutions is difficult because different scores are employed for semiquantitative assessment. We developed normalized scores for comparing data both within a single method (MRI) and between different methods (MRI versus scintigraphy). MATERIALS AND METHODS. We evaluated 51 patients with type 1 GD (26 women, 25 men; mean age ± SD, 36.3 ± 10.9 years old). T1- and T2-weighted turbo spin-echo sequences at 1.5 T served to derive the bone marrow burden score (0-16), the vertebra-disk ratio (VDR), the Terk score (0-3), and the Spanish-MRI score (S-MRI, 0-24). Scintigraphy was scored between 0 and 8. Each score was normalized into four categories: 0 = normal, 1 = mild, 2 = intermediate, 3 = severe involvement. Interobserver and intraobserver agreements were evaluated by kappa statistics; nonparametric statistics with Bonferroni correction assessed correlations among the various original and normalized scores. RESULTS. Interobserver agreement was excellent for the original scores (κ = 0.730- 0.843) and even more so for the normalized scores (κ = 0.775-0.940). Intraobserver agreement kappa values ranged from 0.753 to 0.937 for the original scores and 0.851 to 1.000 for the normalized scores. Highly significant correlations were found among the various original scores (r = 0.42-0.86, p values between 0.0296 and < 0.0001), except for VDR versus S-MRI and Terk. Normalization generally induced marginal reductions in statistical significance, whereas S-MRI versus VDR reached statistical significance with the normalized scores. CONCLUSION. Our data indicate no significant loss of statistical information is caused by the normalization we employed. Our approach therefore facilitates comparison of different scores obtained in different institutions with different imaging modalities.

AB - OBJECTIVE. MRI and 99mTc-sestamibi scintigraphy are used to estimate bone marrow infiltration in patients with Gaucher disease (GD), but comparison of data obtained at different institutions is difficult because different scores are employed for semiquantitative assessment. We developed normalized scores for comparing data both within a single method (MRI) and between different methods (MRI versus scintigraphy). MATERIALS AND METHODS. We evaluated 51 patients with type 1 GD (26 women, 25 men; mean age ± SD, 36.3 ± 10.9 years old). T1- and T2-weighted turbo spin-echo sequences at 1.5 T served to derive the bone marrow burden score (0-16), the vertebra-disk ratio (VDR), the Terk score (0-3), and the Spanish-MRI score (S-MRI, 0-24). Scintigraphy was scored between 0 and 8. Each score was normalized into four categories: 0 = normal, 1 = mild, 2 = intermediate, 3 = severe involvement. Interobserver and intraobserver agreements were evaluated by kappa statistics; nonparametric statistics with Bonferroni correction assessed correlations among the various original and normalized scores. RESULTS. Interobserver agreement was excellent for the original scores (κ = 0.730- 0.843) and even more so for the normalized scores (κ = 0.775-0.940). Intraobserver agreement kappa values ranged from 0.753 to 0.937 for the original scores and 0.851 to 1.000 for the normalized scores. Highly significant correlations were found among the various original scores (r = 0.42-0.86, p values between 0.0296 and < 0.0001), except for VDR versus S-MRI and Terk. Normalization generally induced marginal reductions in statistical significance, whereas S-MRI versus VDR reached statistical significance with the normalized scores. CONCLUSION. Our data indicate no significant loss of statistical information is caused by the normalization we employed. Our approach therefore facilitates comparison of different scores obtained in different institutions with different imaging modalities.

KW - Bone marrow

KW - Gaucher disease

KW - MRI

KW - Scintigraphy

UR - http://www.scopus.com/inward/record.url?scp=84999274188&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84999274188&partnerID=8YFLogxK

U2 - 10.2214/AJR.15.15294

DO - 10.2214/AJR.15.15294

M3 - Article

VL - 206

SP - 1245

EP - 1252

JO - American Journal of Roentgenology

JF - American Journal of Roentgenology

SN - 0361-803X

IS - 6

ER -