Fadila Mihoubi Bouvier1, Vincent Thomas De Montpréville2, Benjamin Besse3, Gilles Missenard4, Charles Court4, Mickael Tordjman5, Cécile Le Pechoux6, François Leroy Ladurie7, Corinne Balleyguier8, Elie Fadel9, Caroline Caramella8. 1. Department of Radiology, Gustave Roussy Cancer Campus, 114 Rue Edouard Vaillant, Villejuif, 94800, Paris, France. fadila.mihoubi@live.fr. 2. Pathology Department, Marie Lannelongue Hospital, University of Paris Sud, Le Plessis Robinson, Paris, France. 3. Department of Medical Oncology, Gustave Roussy Cancer Campus, University of Paris Sud, Villejuif, Paris, France. 4. Orthopaedic Department, Tumor and Spine Unit, Bicêtre University Hospital, AP-HP Paris, University of Paris Sud, Le Kremlin-Bicêtre, Paris, France. 5. Department of Osteoarticular Radiology, Cochin Hospital, AP-HP Paris, Paris, France. 6. Department of Radiation Oncology, Gustave Roussy Cancer Campus, University of Paris Sud, Villejuif, Paris, France. 7. Department of Pneumology, Marie Lannelongue Hospital University of Paris Sud Le Plessis Robinson, Paris, France. 8. Department of Radiology, Gustave Roussy Cancer Campus, 114 Rue Edouard Vaillant, Villejuif, 94800, Paris, France. 9. Department of Thoracic and Vascular Surgery and Heart-Lung Transplantation, Marie Lannelongue Hospital, University of Paris Sud, Le Plessis Robinson, Paris, France.
Abstract
OBJECTIVES: Vertebral invasion is a key prognostic factor and a critical aspect of surgical planning for superior sulcus tumors. This study aims to further evaluate MRI features of vertebral invasion in order to distinguish it from reactive inflammatory changes. METHODS: Between 2000 and 2016, a retrospective study was performed at a single institution. All patients with superior sulcus tumors undergoing surgery, including at least two partial vertebrectomies, were included. An expert radiologist evaluated qualitative and quantitative MRI signal intensity characteristics (contrast-to-noise ratio [CNR]) of suspected involved and non-involved vertebrae. A comparison of CNR of invaded and sane vertebrae was performed using non-parametric tests. Imaging data were correlated with pathological findings. RESULTS: A total of 92 surgical samples of vertebrectomy were analyzed. The most specific sequences for invasion were T1 and T2 weighted (92% and 97%, respectively). The most sensitive sequences were contrast enhanced T1 weighted fat suppressed and T2 weighted fat suppressed (100% and 80%). Loss of extrapleural paravertebral fat on the T1-weighted sequence was highly sensitive (100%) but not specific (63%). Using quantitative analysis, the optimum cut-off (p < 0.05) to distinguish invasion from reactive inflammatory changes was CNR > 11 for the T2-weighted fat-sat sequence (sensitivity 100%), CNR > 9 for contrast-enhanced T1-weighted fat-suppressed sequence (sensitivity 100%), and CNR < - 30 for the T1-weighted sequence (specificity 97%). Combining these criteria, 23 partial vertebrectomies could have been avoided in our cohort. CONCLUSION: Qualitative and quantitative MRI analyses are useful to discriminate vertebral invasion from reactive inflammatory changes. KEY POINTS: • Abnormal signal intensity in a vertebral body adjacent to a superior sulcus tumor may be secondary to direct invasion or reactive inflammatory changes. • Accurate differentiation between invasion and reactive inflammatory changes significantly impacts surgical planning. T1w and T2w are the best sequences to differentiate malignant versus benign bone marrow changes. The use of quantitative analysis improves MRI specificity. • Using contrast media improves the sensitivity for the detection of tumor invasion.
OBJECTIVES: Vertebral invasion is a key prognostic factor and a critical aspect of surgical planning for superior sulcus tumors. This study aims to further evaluate MRI features of vertebral invasion in order to distinguish it from reactive inflammatory changes. METHODS: Between 2000 and 2016, a retrospective study was performed at a single institution. All patients with superior sulcus tumors undergoing surgery, including at least two partial vertebrectomies, were included. An expert radiologist evaluated qualitative and quantitative MRI signal intensity characteristics (contrast-to-noise ratio [CNR]) of suspected involved and non-involved vertebrae. A comparison of CNR of invaded and sane vertebrae was performed using non-parametric tests. Imaging data were correlated with pathological findings. RESULTS: A total of 92 surgical samples of vertebrectomy were analyzed. The most specific sequences for invasion were T1 and T2 weighted (92% and 97%, respectively). The most sensitive sequences were contrast enhanced T1 weighted fat suppressed and T2 weighted fat suppressed (100% and 80%). Loss of extrapleural paravertebral fat on the T1-weighted sequence was highly sensitive (100%) but not specific (63%). Using quantitative analysis, the optimum cut-off (p < 0.05) to distinguish invasion from reactive inflammatory changes was CNR > 11 for the T2-weighted fat-sat sequence (sensitivity 100%), CNR > 9 for contrast-enhanced T1-weighted fat-suppressed sequence (sensitivity 100%), and CNR < - 30 for the T1-weighted sequence (specificity 97%). Combining these criteria, 23 partial vertebrectomies could have been avoided in our cohort. CONCLUSION: Qualitative and quantitative MRI analyses are useful to discriminate vertebral invasion from reactive inflammatory changes. KEY POINTS: • Abnormal signal intensity in a vertebral body adjacent to a superior sulcus tumor may be secondary to direct invasion or reactive inflammatory changes. • Accurate differentiation between invasion and reactive inflammatory changes significantly impacts surgical planning. T1w and T2w are the best sequences to differentiate malignant versus benign bone marrow changes. The use of quantitative analysis improves MRI specificity. • Using contrast media improves the sensitivity for the detection of tumor invasion.
Entities:
Keywords:
Bone neoplasms; Lung neoplasms; Magnetic resonance imaging; Pancoast syndrome