Accuracy of automated patient positioning in CT using a 3D camera for body contour detection

Abstract

Objective: To assess the accuracy of a 3D camera for body contour detection and patient positioning in CT compared to routine manual positioning by radiographers.

Methods and materials: Four hundred twenty-three patients that underwent CT of the head, thorax, and/or abdomen on a scanner with manual table height selection and 254 patients on a scanner with table height suggestion by a 3D camera were retrospectively included. Within the camera group, table height suggestion was based on infrared body contour detection and fitting of a scalable patient model to the 3D data. Proper positioning was defined as the ideal table height at which the scanner isocenter coincides with the patient's isocenter. Patient isocenter was computed by automatic skin contour extraction in each axial image and averaged over all images. Table heights suggested by the camera and selected by the radiographer were compared with the ideal height.

Results: Median (interquartile range) absolute table height deviation in millimeter was 12.0 (21.6) for abdomen, 12.2 (12.0) for head, 13.4 (17.6) for thorax-abdomen, and 14.7 (17.3) for thorax CT scans positioned by radiographers. The deviation was significantly less (p < 0.01) for the 3D camera at 6.3 (6.9) for abdomen, 9.5 (6.8) for head, 6.0 (6.1) for thorax-abdomen, and 5.4 (6.4) mm for thorax.

Conclusion: A 3D camera for body contour detection allows for accurate patient positioning, thereby outperforming manual positioning done by radiographers, resulting in significantly smaller deviations from the ideal table height. However, radiographers remain indispensable when the system fails or in challenging cases.

Key points: • A 3D camera for body contour detection allows for accurate patient positioning. • A 3D camera outperformed radiographers in patient positioning in CT. • Deviation from ideal table height was more extreme for patients positioned by radiographers for all body parts.

Keywords: Diagnostic imaging; Health physics; Radiation dosage; Tomography, x-ray computed.

Conflict of interest statement

Guarantor

The scientific guarantor of this publication is Prof. Gabriel P. Krestin.

Conflict of interest

The authors of this manuscript declare relationships with the following companies:

R. Booij: Research collaboration, Siemens Healthineers

R.P.J. Budde: None

M.L. Dijkshoorn: Clinical training consultant: Siemens Healthineers

M. van Straten: Research collaboration, Siemens Healthineers

Our department has a Master Research Agreement with Siemens Healthineers. No funding or financial support was received for preparation of this article.

Statistics and biometry

One of the authors has significant statistical expertise.

Informed consent

Written informed consent was waived by the Institutional Review Board.

Ethical approval

Institutional Review Board approval was obtained.

Methodology

• retrospective

• experimental

• performed at one institution

Figures

Fig. 1

a Measured depth values. b Depth surface after perspective correction. c User interface of touch panel on CT scanner: 1, table position and table height; 2, selectable body region; 3, adjustable scan range; 4, taking planning image; 5, automatic position of the patient on base of selected scan range. d Virtual patient avatar. e Patient positioning accuracy: Red horizontal line: average patient isocenter, blue horizontal line: scanner isocenter, green horizontal line: average patient isocenter estimated by camera

Fig. 2

3D region growing to extract patient isocenter. a Axial view of region growing to extract patient isocenter for each slice, defined as the midpoint between the highest and lowest points (red dashed lines) of the extracted patient skin contour. b Sagittal MIP is created to demonstrate vertical height measurement of consecutive axial images and demonstrate scanner isocenter does not coincide with patient isocenter (blue line: scanner isocenter, dashed red line: patient isocenter)

Fig. 3

a Box-and-whisker plots of patient positioning performance of all different body parts separately for the radiographers and the 3D camera. b Box-and-whisker plots of patient positioning performance of all different body parts combined for the radiographers and the 3D camera. The median (horizontal line within box), interquartile range (box), and nonoutlier range (whiskers). The largest deviations from the scanner isocenter are plotted as open dots and represent values outside the nonoutlier range of the IQR; the latter computed as 1.5 times interquartile range (i.e., 25–75%)

Fig. 4

a Distribution of patients positioned higher (negative numbers) and lower (positive numbers) from the ideal table height with table height suggestion by the 3D camera. b Distribution of patients positioned higher (negative numbers) and lower (positive numbers) from the ideal table height with manual positioning done by radiographers