Spanish 
Chinese 
Russian 
German 
French 
Japanese 
Portuguese 
Hindi Research Article, J Clin Exp Oncol Vol: 4 Issue: 2
Evaluation of Automatic and Manual Contouring of Cranial Structures
| Vadim Y Kuperman1*, Kouskoulas TN2, Battle JA2, Brad A Factor2, Charles D Hechtman2, Oscar F Carbonell3, Miles DT3, Lubich LM4, Mendoza AS2 and Page DB5 | |
| 1Department of Radiation Oncology, Florida Hospital, Tampa, FL 33613, USA | |
| 2Department of Radiation Oncology, Halifax Health, Daytona Beach, FL 32114, USA | |
| 3Department of Radiology, Halifax Health, Daytona Beach, FL 32114, USA | |
| 4Medical Physics Support, Tampa, FL 33634, USA | |
| 5Florida State University, College of Medicine, Tallahassee, FL 32304, USA | |
| Corresponding author : Vadim Y Kuperman, PhD Department of Radiation Oncology, Florida Hospital 3100 E. Fletcher Avenue, Tampa, FL, 33613, USA Tel: (813) 300-0514; Fax: (813) 615-7409 E-mail: vadimkuperman@yahoo.com |
|
| Received April 07, 2015 Accepted August 06, 2015 Published August 11, 2015 | |
| Citation: Kuperman VY, Kouskoulas TN, Battle JA, Factor BA, Hechtman CD, et al. (2015) Evaluation of Automatic and Manual Contouring of Cranial Structures. J Clin Exp Oncol 4:3. doi:10.4172/2324-9110.1000141 |
Abstract
Evaluation of Automatic and Manual Contouring of Cranial Structures
Objective:
This study evaluates automatic and manual segmentations of cranial structures in iPlan treatment planning system (Brainlab, Munich, Germany).
Methods:
For the analysis of the automatic segmentation algorithm in iPlan, delineations of cranial structures including brainstem, optic nerves, optic chiasm, eyes and eye lenses were analyzed for ten patients with brain metastasis. Five participating radiation oncologists contoured the same structures manually on axial slices of T1-weighted MR scans of the head. The automatic and manual delineations of the cranial structures were compared by using the following quantitative measures: Dice similarity coefficient, differences in the structures’ volumes and positions of the center of mass.
Results:
The comparison of the automatically generated contours and those produced by clinicians, and inter-comparison between clinicians’ delineations indicate relatively good agreement for the brainstem and eyes but unsatisfactory agreement for other structures.
Conclusions:
The obtained results indicate the need for careful verification of the automatically segmented structures in the iPlan treatment planning system. The observed variability between the delineations by different clinicians is of concern. One possible remedy is to utilize expansion margins for the critical cranial structures during treatment planning.
