INTRODUCTION: Due to technical innovations in sectional diagram methods, whole-body imaging has increased in importance for clinical radiology, particularly for the diagnosis of systemic tumor disease. Large numbers of images have to be evaluated in increasingly shorter time periods. The aim was to create and evaluate a new software tool to assist and automate the process of diagnosing whole-body datasets. MATERIAL AND METHODS: Thirteen whole-body datasets were evaluated by 3 readers using the conventional system and the new software tool. The times for loading the datasets, examining 5 different regions (head, neck, thorax, abdomen and pelvis/skeletal system) and retrieving a relevant finding for demonstration were acquired. Additionally a Student T-Test was performed. For qualitative analysis the 3 readers used a scale from 0 - 4 (0 = bad, 4 = very good) to assess dataset loading convenience, lesion location assistance, and ease of use. Additionally a kappa value was calculated. RESULTS: The average loading time was 39.7 s (+/- 5.5) with the conventional system and 6.5 s (+/- 1.4) (p < 0.01) with the new software tool. For the different regions (conventional system/new software tool), the time reduction for readers 1, 2, and 3 were as follows: in the head region 35.9 % (p < 0.01)/49.9 % (p < 0.01)/54.3 % (p < 0,01), in the neck region 48.5 % (p < 0.01)/52.6 % (p < 0.01)/59.4 % (p < 0.05), in the thorax region 59.1 % (p < 0.01)/56.2 % (p < 0.05)/62.1 % (p < 0.05), in the abdominal region 61.9 % (p < 0.01)/62.7 % (p < 0.05)/47.9 % (p < 0.01) and in the pelvis region 73.1 % (p < 0.01)/63.7 % (p < 0.05)/55 % (p < 0.01), respectively. 148.2 s (+/- 94.8) compared to 2.5 s (+/- 0.5) were required to retrieve a previously described finding (p < 0.01). With and without the new software tool the same number of metastases was found (p < 0.01, k > 0.9). The qualitative analysis showed a significant advantage with respect to convenience (p < 0.01, k > 0.9). CONCLUSION: Use of the new software can achieve a significant time savings when working with whole-body datasets with a constant quality of findings and a significant advantage with respect to convenience. As a result, the problem of evaluating examinations with thousands of images can be approached systematically.
INTRODUCTION: Due to technical innovations in sectional diagram methods, whole-body imaging has increased in importance for clinical radiology, particularly for the diagnosis of systemic tumor disease. Large numbers of images have to be evaluated in increasingly shorter time periods. The aim was to create and evaluate a new software tool to assist and automate the process of diagnosing whole-body datasets. MATERIAL AND METHODS: Thirteen whole-body datasets were evaluated by 3 readers using the conventional system and the new software tool. The times for loading the datasets, examining 5 different regions (head, neck, thorax, abdomen and pelvis/skeletal system) and retrieving a relevant finding for demonstration were acquired. Additionally a Student T-Test was performed. For qualitative analysis the 3 readers used a scale from 0 - 4 (0 = bad, 4 = very good) to assess dataset loading convenience, lesion location assistance, and ease of use. Additionally a kappa value was calculated. RESULTS: The average loading time was 39.7 s (+/- 5.5) with the conventional system and 6.5 s (+/- 1.4) (p < 0.01) with the new software tool. For the different regions (conventional system/new software tool), the time reduction for readers 1, 2, and 3 were as follows: in the head region 35.9 % (p < 0.01)/49.9 % (p < 0.01)/54.3 % (p < 0,01), in the neck region 48.5 % (p < 0.01)/52.6 % (p < 0.01)/59.4 % (p < 0.05), in the thorax region 59.1 % (p < 0.01)/56.2 % (p < 0.05)/62.1 % (p < 0.05), in the abdominal region 61.9 % (p < 0.01)/62.7 % (p < 0.05)/47.9 % (p < 0.01) and in the pelvis region 73.1 % (p < 0.01)/63.7 % (p < 0.05)/55 % (p < 0.01), respectively. 148.2 s (+/- 94.8) compared to 2.5 s (+/- 0.5) were required to retrieve a previously described finding (p < 0.01). With and without the new software tool the same number of metastases was found (p < 0.01, k > 0.9). The qualitative analysis showed a significant advantage with respect to convenience (p < 0.01, k > 0.9). CONCLUSION: Use of the new software can achieve a significant time savings when working with whole-body datasets with a constant quality of findings and a significant advantage with respect to convenience. As a result, the problem of evaluating examinations with thousands of images can be approached systematically.
Authors: C Plathow; M Walz; M P Lichy; P Aschoff; C Pfannenberg; H Bock; S M Eschmann; C D Claussen; H P Schlemmer Journal: Radiologe Date: 2008-04 Impact factor: 0.635
Authors: Oleh Dzyubachyk; Jorik Blaas; Charl P Botha; Marius Staring; Monique Reijnierse; Johan L Bloem; Rob J van der Geest; Boudewijn P F Lelieveldt Journal: Int J Comput Assist Radiol Surg Date: 2013-06-01 Impact factor: 2.924