Gauthier Dodin1, Julia Salleron2, Salma Jendoubi3, Waled Abou Arab3, François Sirveaux4, Alain Blum3, Pedro Augusto Gondim Teixeira3. 1. Service d'imagerie Guilloz, Hôpital Central, Centre Hospitalier Régional Universitaire de Nancy, 29 avenue du Maréchal de Lattre de Tassigny, 54035, Nancy cedex, France. gauthier-dodin@orange.fr. 2. Département de Biostatistique, Institut de Cancérologie de Lorraine, 6 avenue de Bourgogne, F-54519, Vandœuvre-lès-Nancy cedex, France. 3. Service d'imagerie Guilloz, Hôpital Central, Centre Hospitalier Régional Universitaire de Nancy, 29 avenue du Maréchal de Lattre de Tassigny, 54035, Nancy cedex, France. 4. Centre Chirurgical Emile-Gallé, Centre Hospitalier Régional Universitaire de Nancy, 49, rue Hermite, 54000, Nancy cedex, France.
Abstract
OBJECTIVES: To evaluate the added value of DWI, qualitative proton MR spectroscopy (H-MRS) and dynamic contrast-enhanced perfusion (DCE-P) to conventional MRI in differentiating benign and malignant non-fatty soft tissue tumors (NFSTT). METHODS: From November 2009 to August 2017, 288 patients with NFSTT that underwent conventional and advanced MRI were prospectively evaluated. The study was approved by the local ethics committee. All patients signed an informed consent. A musculoskeletal (R1) and a general (R2) radiologist classified all tumors as benign, malignant, or indeterminate according to morphologic MRI features. Then, DWI, H-MRS, and DCE-P data of indeterminate tumors were analyzed by two additional radiologists (R3 and R4). Advanced techniques were considered individually and in combination for tumor benign-malignant differentiation using histology as the gold standard. RESULTS: There were 104 (36.1%) malignant and 184 (63.9%) benign tumors. Conventional MRI analysis classified 99 tumors for R1 and 135 for R2 as benign or malignant, an accuracy for the identification of malignancy of 87.9% for R1 and 83.7% for R2, respectively. There were 189 indeterminate tumors for R1. For these tumors, the combination of DWI and H-MRS yielded the best accuracy for malignancy identification (77.4%). DWI alone provided the best sensitivity (91.8%) while the combination of DCE-P, DWI, and H-MRS yielded the best specificity (100%). The reproducibility of the advanced imaging parameters was considered good to excellent (Kappa and ICC > 0.86). An advanced MRI evidence-based evaluation algorithm was proposed allowing to characterize 28.1 to 30.1% of indeterminate non-myxoid tumors. CONCLUSION: The prioritized use of advanced MRI techniques allowed to decrease by about 30% the number of non-myxoid NFSTT deemed indeterminate after conventional MRI analysis alone. KEY POINTS: • When morphological characterization of non-fatty soft tissue tumors is possible, the diagnostic performance is high and there is no need for advanced imaging techniques. • Following morphologic analysis, advanced MRI techniques reduced by about 30% the number of non-myxoid indeterminate tumors. • DWI is the keystone of advanced imaging techniques yielding the best sensitivity (91.8%). Optimal specificity (> 90%) is obtained by a combination of advanced techniques.
OBJECTIVES: To evaluate the added value of DWI, qualitative proton MR spectroscopy (H-MRS) and dynamic contrast-enhanced perfusion (DCE-P) to conventional MRI in differentiating benign and malignant non-fatty soft tissue tumors (NFSTT). METHODS: From November 2009 to August 2017, 288 patients with NFSTT that underwent conventional and advanced MRI were prospectively evaluated. The study was approved by the local ethics committee. All patients signed an informed consent. A musculoskeletal (R1) and a general (R2) radiologist classified all tumors as benign, malignant, or indeterminate according to morphologic MRI features. Then, DWI, H-MRS, and DCE-P data of indeterminate tumors were analyzed by two additional radiologists (R3 and R4). Advanced techniques were considered individually and in combination for tumor benign-malignant differentiation using histology as the gold standard. RESULTS: There were 104 (36.1%) malignant and 184 (63.9%) benign tumors. Conventional MRI analysis classified 99 tumors for R1 and 135 for R2 as benign or malignant, an accuracy for the identification of malignancy of 87.9% for R1 and 83.7% for R2, respectively. There were 189 indeterminate tumors for R1. For these tumors, the combination of DWI and H-MRS yielded the best accuracy for malignancy identification (77.4%). DWI alone provided the best sensitivity (91.8%) while the combination of DCE-P, DWI, and H-MRS yielded the best specificity (100%). The reproducibility of the advanced imaging parameters was considered good to excellent (Kappa and ICC > 0.86). An advanced MRI evidence-based evaluation algorithm was proposed allowing to characterize 28.1 to 30.1% of indeterminate non-myxoid tumors. CONCLUSION: The prioritized use of advanced MRI techniques allowed to decrease by about 30% the number of non-myxoid NFSTT deemed indeterminate after conventional MRI analysis alone. KEY POINTS: • When morphological characterization of non-fatty soft tissue tumors is possible, the diagnostic performance is high and there is no need for advanced imaging techniques. • Following morphologic analysis, advanced MRI techniques reduced by about 30% the number of non-myxoid indeterminate tumors. • DWI is the keystone of advanced imaging techniques yielding the best sensitivity (91.8%). Optimal specificity (> 90%) is obtained by a combination of advanced techniques.
Entities:
Keywords:
Diffusion; Magnetic resonance imaging; Perfusion; Soft tissue neoplasm; Spectroscopy
Authors: Derek F Papp; A Jay Khanna; Edward F McCarthy; John A Carrino; Adam J Farber; Frank J Frassica Journal: J Bone Joint Surg Am Date: 2007-10 Impact factor: 5.284