Konrad Leszczynski

The investigation and rectification of field placement errors in the delivery of complex head and neck fields

PURPOSE To evaluate a novel technique for resolving field placement errors into their components and to quantify the improvement in accuracy potentially achievable by translation and rotation of the radiation beam. METHODS AND MATERIALS One hundred and eighty-five films (both simulator and portal) from seventeen patients receiving radiotherapy to the head and neck region were analyzed in pairs. The computer based comparisons of complex fields with curved edges employed the intersections of perpendiculars from two reference points with the field periphery to define field match points. Field placement errors were resolved into those due to patient motion within the immobilization shell and those due to incorrect beam position, orientation, or shape. RESULTS The median and the 95 percentile of the distribution of differences between prescribed (simulator) fields and treated (portal) fields referenced to the patients anatomy were 4.4 mm and 8.9 mm, respectively. The analysis suggests that with appropriate translation and rotation of the beam with respect to the immobilization shell these figures could be reduced to 3.1 mm and 8.2 mm, respectively, confirming the large contribution of patient motion within the shell to field placement accuracy. Comparisons between treated fields indicated smaller variability during treatment than between simulation and treatment. CONCLUSION The perpendicular intersection method described here was found appropriate for the identification of field match points. The distributions of field placement errors were similar to those in a published study of straight edged fields. Translation and rotation of the applied field with respect to the immobilization shell would generally result in only a small improvement in field placement accuracy.

Optimization of metal/phosphor screens for on‐line portal imaging

Studies were conducted to determine the optimal metal/phosphor screen for on-line video verification of radiation treatment portals. Screens were evaluated for luminance and spatial resolution as a function of composition and thickness at 6- and 23-MV x-ray energies. A new video technique was used to determine modulation transfer functions. Gd2O2S was found to be the most efficient (brightest) phosphor for this application. Luminance was found to vary linearly with phosphor thickness up to a thickness of 500 mg/cm2. Metal plates made of iron, brass, copper, lead, and sintered tungsten of various thicknesses were also tested for luminance and resolution with Gd2O2S phosphor. Brightness peaked at about 2-mm thickness for most metals. Significant contributions to the brightness were found to come from x rays interacting with the phosphor itself.

On few‐view tomographic reconstruction with megavoltage photon beams

Currently portal imaging devices are used to obtain information on patient localization during radiation therapy treatments. Such obtained information is two dimensional in nature, limited to the plane of the captured image. It has been proposed that megavoltage computed tomography images be reconstructed to overcome this limitation. This study explores the feasibility of reconstructing tomographic images from fan-beam projection data acquired with a commercial portal imaging device on a standard radiotherapy linear accelerator. Several CT reconstruction algorithms are examined as to their performance and suitability for applications in radiation therapy verification. The results show that it is possible, using some of the iterative reconstruction techniques, to obtain an image useful for patient localization from only several (< or =10) projection views.

A test tool for the visual verification of light and radiation fields using film or an electronic portal imaging device.

We describe the design and evaluation of a simple test tool which can be used in conjunction with either film or an electronic portal imaging device (EPID) to verify light and radiation fields and their congruence. The precision of the technique is better than 0.5 mm under all conditions tested. When used with film the accuracy or offset of the technique (the difference between test tool observations and a scanned conventional film) is better than 0.5 mm but, with an EPID as the image receptor, the accuracy dropped to, in one trial, 0.86 mm. The offset may be due to a systematic observer bias in determining the 50% O.D. level on the image, compounded, in the case of EPID measurements, by image acquisition and display parameters. Thus, when used with an EPID, calibration of the system will be required if absolute field dimensions are required. When used with film, the test tool method described here is of sufficient accuracy and precision to confirm the compliance of light and radiation field parameters with currently accepted quality control protocols.

Computer-aided radiation therapy simulation: image intensifier spatial distortion correction for large field of view digital fluoroscopy.

An accurate method of correcting spatial distortion in digital fluoroscopy images has been developed for generating fluoroscopy-based large field of view images for computer-aided radiation therapy simulation. This method is applicable to arbitrary gantry rotations and arbitrary shifts of the image intensifier relative to the central axis of the x-ray beam. It is therefore suitable for conventional radiation therapy simulation techniques that involve the arbitrary positioning of the image intensifier by the operator. Spatial distortion is modelled as two image intensifier orientation-dependent components, the first resulting from the projection of the x-ray image onto the curved surface of the image intensifier front end, and the second produced by the image intensifier electron optics, interactions with external magnetic fields and the video system. A geometrical model approximates the first component. The second component is modelled by a third-order polynomial transformation. A weighted mean approach is employed to achieve accurate distortion correction when the image intensifier is oriented differently from the calibration orientations. Mean and maximum residual errors (measured in the plane of the isocentre) of 0.4 mm and 1.0 mm respectively have been achieved with just 48 calibration orientations in four dimensions (gantry rotation and lateral, longitudinal and vertical shifts of the image intensifier).

Segmented chamfer matching for the registration of field borders in radiotherapy images

Optimum conformity between treated and prescribed radiotherapy fields is likely to be achieved when the full versatility of modern therapy equipment is reflected in the field registration method. To allow for independent jaws and custom shielding, chamfer matching has been used to register selected segments of field borders independently. In a study involving 50 clinical prescription-treatment field pairs it is shown that segmented chamfer matching is superior not only to conventional unsegmented chamfer matching but also to several other methods. An example of the clinical value of segmented chamfer matching applied to electronic portal images is given.

The Siemens Virtual Wedge

A Siemens Virtual Wedge has recently been installed and commissioned at the Northeastern Ontario Regional Cancer Centre. Measurements reported below show that 1) Virtual Wedge factors are within 1.5% of 1; 2) percentage depth doses down to 15 cm for open and virtually wedged fields are identical to within 0.7%; 3) relative cross beam profiles for 60 degrees virtual and physical wedges are very similar except at the toe end where a 5% difference in relative dose has been observed and 4) the peripheral dose from the 60 degrees Virtual Wedge is about half of that from the 60 degrees physical wedge. A clinical protocol requiring combined open and 60 degrees wedged fields has been developed and validated. This protocol, which does not impair the utility of the Virtual Wedge, facilitates the use of on-line portal imaging and significantly reduces the effort required to commission the system.

On the controversies surrounding the origins of radiation therapy

In this study the authors analyze the documentation regarding the earliest known X-ray treatments with the objective of identifying the true origin of radiation therapy. The four most often quoted events, including X-ray treatments allegedly performed in 1896 are analyzed in the light of available published reports. From this it is concluded that Despeignes of Lyon, who in July of 1896 irradiated a patient with cancer of the stomach, is in all likelihood the first person to perform documented radiation therapy treatments with a scientific and logical basis.

Independent corroboration of monitor unit calculations performed by a 3D computerized planning system

The checking of monitor unit calculations is recognized as a vital component of quality assurance in radiotherapy. Using straightforward but detailed computer‐based verification calculations it is possible to achieve a precision of 1% when compared with a three‐dimensional (3D) treatment planning system monitor unit calculation. The method is sufficiently sensitive to identify significant errors and is consistent with current recommendations on the magnitude of uncertainties in clinical dosimetry. Moreover, the approach is accurate in the sense of being highly consistent with the validated 3D treatment planning systems calculations. PACS number(s): 87.53.–j, 87.52.–g

Application of a fuzzy pattern classifier to decision making in portal verification of radiotherapy

With the large volume of electronic portal images acquired and stringent time constraints, it is no longer feasible to follow the convention whereby the radiation oncologist reviews and approves or rejects all portals. For that purpose we have developed a portal image classifier based on the fuzzy k-nearest neighbour (k-NN) algorithm. Each portal image is represented by a feature vector that consists of translational and rotational errors in the placement of radiation field borders that were measured in the portal image. Memberships in the acceptable portal class for the reference portal images within a training dataset were defined by a radiation oncologist expert. The fuzzy k-NN portal image classifier was trained and tested on a dataset of 328 portal images acquired during tangential irradiations of the breast. The memberships in the acceptable portal class produced by the fuzzy k-NN algorithm agreed very well with those defined by the expert. The linear correlation coefficient was equal to 0.89. Performance of the fuzzy k-NN classifier was also evaluated from the portal decision-making point of view using the measures of accuracy, sensitivity and specificity. The fuzzy k-NN portal classifier was capable of identifying almost all the truly unacceptable portals with an acceptably low false alarm rate.