Contributions
Abstract: PB2240
Type: Publication Only
Background
Investigations for skeletal disease is essential in the diagnostic work-up of multiple myeloma. According to international guidelines, low dose CT, MRI or PET-CT are the preferred choices of modality.
Aims
To compare the ability of MRI and CT to detect of skeletal disease in multiple myeloma. To investigate the inter-observer agreement between two radiologists in the interpretation of the images.
Methods
A retrospective, comparative study of multiple myeloma patients that that had taken MRI and CT within four months between 2007 and 2015. We included only patients where the entire vertebral column was portrayed and ended up with 12 patients. The radiologists looked at the images independently blinded for the colleague’s’ previous descriptions. The images were assessed for malignant fracture, osteoporotic fracture, malignant infiltration and osteoporosis. Later they sat down together, to make up a gold standard, using all pictures and information available to them.
Results
Radiologist 1 diagnosed 20 malignant fractures in the vertebral column on CT, and 26 on MRI, while Radiologist 2 diagnosed 12 fractures on CT and 22 on MRI. The gold standard, however, showed 10 fractures. The two radiologists agreed that 1 patient had no malignant fractures on CT, while the gold standard showed that 6 of 12 patients did not have malignant fractures. The radiologists had a total of 24 different diagnostic assessments of malignant fractures on CT and 18 on MRI. Radiologist 1 diagnosed 9 osteoporotic fractures on CT and 23 on MRI, and Radiologist 2 diagnosed 9 fractures on CT and 24 on MRI. The gold standard showed 14 osteoporotic fractures. The sensitivity for radiologist 1 to discover malignant fractures on CT was 0.83, specificity 0.5. For radiologist 2, the sensitivity for malignant fracture on CT was 0.5 and the specificity 0.67. On MRI, the sensitivity of Radiologist 1 for detecting malignant fractures was 0.83, specificity 0.17. Radiologist 2 had sensitivity of 1 on MRI, but specificity of 0.17. The inter-observer agreement for detection of malignant fractures showed a Cohen’s kappa of 0.42.
Conclusion
Both radiologists diagnosed malignant and osteoporotic fractures when the gold standard showed that there was none. MRI was not better than CT. Furthermore, the agreement between the radiologists was unsatisfactory. Our study shows that the current guidelines for skeleton imaging in multiple myeloma may result in overdiagnosis of skeletal disease.
Session topic: 14. Myeloma and other monoclonal gammopathies - Clinical
Keyword(s): Bone disease, MRI, Myeloma
Abstract: PB2240
Type: Publication Only
Background
Investigations for skeletal disease is essential in the diagnostic work-up of multiple myeloma. According to international guidelines, low dose CT, MRI or PET-CT are the preferred choices of modality.
Aims
To compare the ability of MRI and CT to detect of skeletal disease in multiple myeloma. To investigate the inter-observer agreement between two radiologists in the interpretation of the images.
Methods
A retrospective, comparative study of multiple myeloma patients that that had taken MRI and CT within four months between 2007 and 2015. We included only patients where the entire vertebral column was portrayed and ended up with 12 patients. The radiologists looked at the images independently blinded for the colleague’s’ previous descriptions. The images were assessed for malignant fracture, osteoporotic fracture, malignant infiltration and osteoporosis. Later they sat down together, to make up a gold standard, using all pictures and information available to them.
Results
Radiologist 1 diagnosed 20 malignant fractures in the vertebral column on CT, and 26 on MRI, while Radiologist 2 diagnosed 12 fractures on CT and 22 on MRI. The gold standard, however, showed 10 fractures. The two radiologists agreed that 1 patient had no malignant fractures on CT, while the gold standard showed that 6 of 12 patients did not have malignant fractures. The radiologists had a total of 24 different diagnostic assessments of malignant fractures on CT and 18 on MRI. Radiologist 1 diagnosed 9 osteoporotic fractures on CT and 23 on MRI, and Radiologist 2 diagnosed 9 fractures on CT and 24 on MRI. The gold standard showed 14 osteoporotic fractures. The sensitivity for radiologist 1 to discover malignant fractures on CT was 0.83, specificity 0.5. For radiologist 2, the sensitivity for malignant fracture on CT was 0.5 and the specificity 0.67. On MRI, the sensitivity of Radiologist 1 for detecting malignant fractures was 0.83, specificity 0.17. Radiologist 2 had sensitivity of 1 on MRI, but specificity of 0.17. The inter-observer agreement for detection of malignant fractures showed a Cohen’s kappa of 0.42.
Conclusion
Both radiologists diagnosed malignant and osteoporotic fractures when the gold standard showed that there was none. MRI was not better than CT. Furthermore, the agreement between the radiologists was unsatisfactory. Our study shows that the current guidelines for skeleton imaging in multiple myeloma may result in overdiagnosis of skeletal disease.
Session topic: 14. Myeloma and other monoclonal gammopathies - Clinical
Keyword(s): Bone disease, MRI, Myeloma