REGULAR CONTENT
Final ID
453
Type
Original Scientific Research-Oral or Pos
Authors
R Farhat1, T French1, D Bechhofer1, A Kranz2, J Hoffmann3
Institutions
1Winthrop-University Hospital, Mineola, NY, 2Winthrop-University Hospital, Old Westbury, NY, 3WInthrop-University Hospital, Mineola, NY
Purpose
To describe the development of a novel tool to quantify the relative strength of abnormality on nuclear medicine GI bleeding scans, and report the potential impact of this tool on interventional radiology (IR) practice.
Materials & Methods
A novel technique was developed at a single institution, allowing for calculation of percent increase of activity in the region of interest/ROI (area of positive bleeding) and activity in ROIs in the aorta and liver (controls). A nuclear medicine database search was performed to find all GI bleeding scans between January 1, 2013 and December 31, 2015. Reports from all studies were reviewed, and imaging from all positive studies was reviewed. The new technique was used to analyze all of the positive bleeding scans. The IR database was queried to determine which patients with positive bleeding scans underwent angiography, and which had positive angiographic findings. Data was analyzed by two radiology attendings and two residents to determine median percent increase in ROI in patients with positive scintigraphy and positive angiography, versus those with positive scintigraphy and negative angiography.
Results
Of 194 GI bleeding scans performed in the study period, 71 were positive for active lower GI bleed. 37 of these had conventional angiograms, and 9 had positive angiography (24%). All positive angiograms did correlate with the area of bleeding on scintigraphy. Using ROI percent change quartiles, we observed a statistically significant association between increasing ROI and the probability of having a positive angiogram (Cochran-Armitage trend test, p-value=0.01), such that there are no positive angiogram cases when ROI change was <20% and a majority of the cases (6/9) in the highest quartile. Thus, a percent increase of less than 20% on a positive GI bleed study in this series yielded no positive angiograms. This can impact the overall number of angiograms performed in IR for lower GI bleed, while also increasing the percentage of procedures where active bleeding is found and intervention can be performed.
Conclusions
This novel quantitative tool for evaluating GI bleeding scans has the potential to improve patient selection for IR angiography and intervention.
Final ID
453
Type
Original Scientific Research-Oral or Pos
Authors
R Farhat1, T French1, D Bechhofer1, A Kranz2, J Hoffmann3
Institutions
1Winthrop-University Hospital, Mineola, NY, 2Winthrop-University Hospital, Old Westbury, NY, 3WInthrop-University Hospital, Mineola, NY
Purpose
To describe the development of a novel tool to quantify the relative strength of abnormality on nuclear medicine GI bleeding scans, and report the potential impact of this tool on interventional radiology (IR) practice.
Materials & Methods
A novel technique was developed at a single institution, allowing for calculation of percent increase of activity in the region of interest/ROI (area of positive bleeding) and activity in ROIs in the aorta and liver (controls). A nuclear medicine database search was performed to find all GI bleeding scans between January 1, 2013 and December 31, 2015. Reports from all studies were reviewed, and imaging from all positive studies was reviewed. The new technique was used to analyze all of the positive bleeding scans. The IR database was queried to determine which patients with positive bleeding scans underwent angiography, and which had positive angiographic findings. Data was analyzed by two radiology attendings and two residents to determine median percent increase in ROI in patients with positive scintigraphy and positive angiography, versus those with positive scintigraphy and negative angiography.
Results
Of 194 GI bleeding scans performed in the study period, 71 were positive for active lower GI bleed. 37 of these had conventional angiograms, and 9 had positive angiography (24%). All positive angiograms did correlate with the area of bleeding on scintigraphy. Using ROI percent change quartiles, we observed a statistically significant association between increasing ROI and the probability of having a positive angiogram (Cochran-Armitage trend test, p-value=0.01), such that there are no positive angiogram cases when ROI change was <20% and a majority of the cases (6/9) in the highest quartile. Thus, a percent increase of less than 20% on a positive GI bleed study in this series yielded no positive angiograms. This can impact the overall number of angiograms performed in IR for lower GI bleed, while also increasing the percentage of procedures where active bleeding is found and intervention can be performed.
Conclusions
This novel quantitative tool for evaluating GI bleeding scans has the potential to improve patient selection for IR angiography and intervention.