PII: S0958-3947(02)00142-5

EFFECT OF BITE TRAY IMPRESSION TECHNIQUE ON RELOCATIONACCURACY IN FRAMELESS STEREOTACTIC RADIOTHERAPY

CLARE E. HERBERT B.A., MARTIN A. EBERT, PH.D., D. BARCLAY, DAVID S. WHITTALL,DAVID J. JOSEPH, M.B.B.S., F.R.A.N.Z.C.R.,

CHRIS S. HARPER, M.B.B.S., F.R.A.N.Z.C.R, andNIGEL A. SPRY, M.B.B.S., F.R.C.P., F.R.C.R., F.R.A.N.Z.C.R., F.A.C.P.M.

Department of Radiation Oncology and Dental Clinic, Sir Charles Gairdner Hospital, Nedlands, Western Australia;Department of Physics, University of Western Australia, Western Australia; and Perth Radiation Oncology Clinic,

Wembley, Western Australia

( Accepted 10 March 2002)

Abstract—A previously developed method for achieving patient relocation in fractionated stereotactic radio-therapy (attachment of an infrared fiducial system to a bite tray) relies on the integrity of a bite tray system thatincorporates moulding to the patient’s upper dentition. Reproducible and accurate patient positioning requiresstability of the bite tray and mould during the full treatment process, both during the time the bite tray is insertedin the patient’s mouth, and between separate bite tray insertions. The optimum construction method for a stablereproducible tray has not been sufficiently investigated. We undertook a study to identify factors which mightinfluence the integrity of the hard palate bite tray system. Reprosil Fast Set Putty was used to construct 3impression conditions; teeth only; teeth and alveolar sulcus; and teeth, alveolar sulcus, and the hard palate.Reproducibility was assessed by volunteers inserting the impressions multiple times and recording the locationsof 8 standard reference points. Our results showed the optimal impression technique (i.e., the one that led to thesmallest ranges in positional and rotational errors) was that which incorporated the teeth, alveolar sulcus, andhard palate. © 2003 American Association of Medical Dosimetrists.

Key Words: Fractionated stereotactic radiotherapy, Bite tray stability, Frameless.

INTRODUCTION

Stereotactic radiosurgery (SRS) involves the delivery ofmultiple coplanar arcs directed to a single common iso-center or multiple isocenters, allowing high-dose confor-mality. The precision and accuracy available with SRS(of the order of 1 to 2 mm) allows the administration ofhigh radiation doses to well defined treatment areas.

More recently, fractionated stereotactic radiother-apy (SRT) has been developed as a viable option to treatlesions close to the optic apparatus or large lesions wheresurgery is deemed unsuitable. Traditionally, SRT hasinvolved a number of cumbersome relocation techniqueswith reported inaccuracies of between 1.0 to 3.0 mm.1–6

Frameless repeat fixation has aided the development ofthe high accuracy technique, allowing fast, efficient, andaccurate treatments. The Medtronic Sofamor Danek(MSD, Memphis, TN) frameless system utilizes an im-pression on a bite tray connected to an array of fiducials(This system is currently owned and marketed by ZmedInc., San Diego CA). The array of fiducials is detected byan infrared localizing device, enabling the patient’s po-sition to be monitored in real time, altering the patient’sposition until a match exists with the prerecorded daily

treatment position. This is distinct from procedures thatimmobilize the patient prior to the commencement of thetreatment and assume no movement throughout the treat-ment process. This separation of immobilization from local-ization can deliver high-precision stereotactic radiotherapy.7,8

The advantages of the frameless system are the easeof relocation, real time patient movement assessment,and a greater reported relocation accuracy.9 The contin-ued development of frameless repeat fixation has aidedin refinement of the level of accuracy in this technique.

The relocation of the impression for subsequenttreatment fractions is the link between the patient and thelocation of the isocenter/s, and requires consideration foressential repeated accuracy.

During commissioning the relocation system, weinvestigated potential positioning errors and found that asignificant contribution to the total error in relocationaccuracy was due to some bite tray relocations.10 Due tothe limited key defining factors for impression technique,we developed an evaluation system to define key areas ofinfluence in relation to the manufacture of more stableand reproducible impressions.

METHODS

The MSD bite trays are supplied in 3 sizes. Impres-sion material is used to make each bite tray patient

Reprint requests to: Clare E. Herbert, Department of RadiationOncology, Sir Charles Gairdner Hospital, Hospital Avenue, Nedlands,Western Australia 6009. E-mail: Clare.Herbert@health.wa.gov.au

Medical Dosimetry, Vol. 28, No. 1, pp 27–30, 2003Copyright © 2003 American Association of Medical Dosimetrists

Printed in the USA. All rights reserved0958-3947/03/$–see front matter

27

specific. They are then attached to reference frames usingsteel screws. (Fig. 1) The reference frames consist of anarray of 4 reflecting fiducials, which are imaged by aninfrared CCD camera system (“radiocamera” system).The radiocamera system localizes and reports the posi-tion and orientation in space of the reference frame.Given the known relationship between the treatmentmachine (linac) and the reference frame, and assuming afixed relationship between the reference frame and thepatient’s cranial anatomy, the positional relationship be-tween the linac and the patient’s cranial anatomy can beascertained and followed in real time.

Fiducial registration wandThe fiducial registration wand (Fig. 2) is a feature of

the MSD system that allows the calculation of the geo-metric relationship between the divot on the referenceframe and the wand tip. This facility is conventionallyused at patient setup to verify the localization system. Inthis study, it was used to investigate bite tray stability.

Alteration of standard impression equipmentPatient cooperation and comfort is essential to en-

able accurate repositioning of the impression over sub-

sequent relocations. The supplied impression trays weremodified to enable greater patient comfort, adhesion ofthe impression material to the trays, and to include anacrylic hard palate.

The supplied impression material (Reprosil HeavyBody Vinyl Polysiloxane Impression Material Type 1,very high viscosity) was replaced with Reprosil EasyMix, fast-set putty. The fast-set putty has several advan-tages, such as: ease of moulding the material onto thetray and specific areas of dentition, reduced setting timerequired (2.5 minutes compared to 4.5 minutes—time forimpression to set from time mixed.), and lower cost.(AUS$4.72 compared to AUS$11.20—cost per impres-sion).

Fast-setting impression material decreases theamount of time the patient is uncomfortable, as well asminimizes movement within the impression.

Three different impression techniques per volunteerwere constructed to investigate accuracy of repositioning(Fig. 3):

1. Teeth only—to determine if a shallow impressionof the teeth only would suffice.

2. Teeth and alveolar sulcus—to determine the rela-tionship of the alveolar sulcus to suction of theimpression to the dentition.

3. Teeth, alveolar sulcus, and hard palate—to deter-mine if the addition of a hard palate to the impres-sion creates greater suction by increasing the sur-face area of the impression.

We employed 8 volunteers to determine the effectof impression technique on the subsequent reproducibil-ity of a bite tray inserted several times. Fast-settingimpression material was distributed evenly within theimpression tray, and all dentition impressions were cen-trally placed inside the ridges of the trays. Each volunteerwas instructed to bite with even pressure. Standard fidu-cial markers (MSD) were applied to the volunteer’s skinas close as feasible to 8 standard reference points (rightouter canthi, left outer canthi, right ala naisi, left alanaisi, right tragus, left tragus, right forehead, and leftforehead). The positions chosen were relatively immo-bile, least influenced by skin movement, and far enough

Fig. 1. Impression bite tray attached to reference frame. (A)Impression tray, (B) reference frame, (C) fiducials, (D) divot.

Fig. 2. Fiducial registration wand. (A) Fiducials, (B) wand tip.

Fig. 3. Impression techniques. (A) teeth only, (B) teeth �alveolar sulcus, (C) teeth � alveolar sulcus � hard palate.

Medical Dosimetry Volume 28, Number 1, 200328

away from delicate eye areas to be independent of in-voluntary eye reflex movements.

The effect of swallowing on repositioning was nottaken into account, as previous indications showed onlya minor influence.11

The bite tray position was registered prior to record-ing positions of each data set by placing the wand tip intothe reference frame divot. The center of each fiducial wasthen located by positioning the wand tip into its recessedcenter.

All measurements were conducted by the principleauthor to reduce interobserver technique bias. The im-pression techniques were registered in random order toreduce bias from volunteer familiarity for subsequentimpression registrations.

All volunteers removed and replaced each of the 3bite trays 5 separate times (n � 960 measurements).

Relative differences in facial fiducial points be-tween bite tray reinsertions resulted from movement ofthe bite tray relative to each volunteer’s anatomy. Thismovement can be represented by rotation and translationof the bite tray in 3D space. To determine the variation inlocation and orientation of the origin of the bite traycoordinate system between bite tray refittings, the meanpositions of the 8 reference points across the 5 trials weredetermined. The translation and rotation of the 8 pointsfor each refitting required to determine the optimal align-ment with the mean positions were then determinedusing a constrained minimization procedure (the IDL12

implementation of the generalised reduced gradient al-gorithm.13).

A least-squares comparison was used as an objec-tive function in the minimization to determine the re-quired x, y, z translations and rotations about the 3coordinate axes. The result of this is an indication of thetranslation and rotation of the fiducial reference systembetween bite tray refittings.

The required translations, the geometric sum ofthese (R � �x2 � y2 � z2), and the required rotationswere tabulated into histogram form for each bite traytype (Figs. 4 and 5). These rotational and translationalerrors indicate the stability of each bite tray (i.e., anideally stable tray would have no rotation and transla-tional errors).

RESULTS

The influences of bite tray construction on reloca-tion accuracy are demonstrated by frequency histogramsshown in Figs. 4 and 5.

DISCUSSION

For frameless repeat localization in stereotactictreatment techniques, the fiducial reference system andthe patient’s skull define the coordinate system to whichthe target is referenced. The relationship between thefiducial array and the patient’s skull must therefore be

maintained for subsequent treatment fractions (bite trayinsertions) and the duration of a treatment fraction (typ-ically 20 minutes10).

Our results support the contention that constructionfactors do contribute to bite tray repositioning accuracy.Mean translational motion and rotation of the fiducialreference system showed increased accuracy of reposi-tioning when the impression covers the teeth, alveolarsulcus, and the hard palate. The likely explanation forthis observation is the increased surface area of theimpression provided by the addition of the acrylic hardpalate.

Bite tray 3 impression demonstrated a decrease inmean radial motion value by � 0.16 mm, and maximumvalue by � 1.74 mm, and a reduced frequency of valuesexperienced in the range of 1.5 to 3 mm.

The rotation of the fiducial reference system be-tween the bite tray refittings demonstrated that bite tray

Fig. 4. Mean translational motion of fiducial reference systembetween bite tray refittings, where bite tray 1 � impression ofteeth only; bite tray 2 � impression of teeth � alveolar sulcus;bite tray 3 � impression of teeth � alveolar sulcus � hard

palate.

Fig. 5. Rotation of fiducial reference system between bite trayrefittings, where bite tray 1 � impression of teeth only; bite tray2 � impression of teeth � alveolar sulcus; bite tray 3 �

impression of teeth � alveolar sulcus � hard palate.

Relocation in frameless stereotactic radiotherapy ● C. E. HERBERT et al. 29

3 displayed the least total angle differences, with maxi-mum values of � 0.7° as compared to � 1.2° for im-pression techniques of the teeth and teeth and alveolarsulcus.

Although only minor improvements resulted, thereduction of these values was beneficial when assessingthe resultant total error. The effect of impression tech-niques on the reproducibility of bite tray repositioning ishelpful to indicate factors that contribute to accuracy.The presence of a bite tray that is accurate and repro-ducible from day 1 to day X, and to know which are thefactors that allow precision in relocation, creates confi-dence in the current fractionated system, and assurancein the move toward a totally frameless system.

Some caution must be exercised in interpreting ourresults. Reproducibility was assessed with healthy vol-unteers and all had good complete dentition, not allowingus to test for less desirable dentition conditions. Markerswere placed in positions with limited skin mobility, andpressure on these points was limited; however, possibil-ity of slight skin movement cannot be ignored. Due to thesmall number of fractionated SRT patients currently re-ferred for treatment, trials were conducted on volunteerstaff and therefore could not incorporate the use of theBRW frame as an absolute positioning reference. Somediscrepancies may exist between the position pointed towith the fiducial array wand and the wand calibration(0.67 mm).

Construction factors play a major role in reliabilityof a reproducible bite tray system. Modifying the impres-sion technique to include the teeth, alveolar sulcus andhard palate resulted in greater relocation accuracy. Werecommend the following: (1) that jaw fixation systemsemploy impression techniques that extend to cover thehard palate, (2) ensure an adequate amount of fast settingimpression material is placed in the tray, (3) instruct thepatient to bite with even pressure, and (4) apply carefulattention to detail.

Acknowledgments—The authors thank Russell Taylor and Ray Hicks(manufacture of the articulator device); and Stephen Alfrich, ColinCochrane, and Sofamor Danek (technical and financial support); andthe volunteers and helpful staff at the Radiation Oncology Departmentof SCGH.

REFERENCES

1. Gill, S.S.; Thomas, D.G.T.; Warrington, A.P.; Brada, M. Relocat-able frame for stereotactic external beam radiotherapy. Int. J.Radiat. Oncol. Biol. Phys. 20:599–603; 1991.

2. Rosenberg, I.; Alheit, H.; Beardmore, C.; et al. Patient positionreproducibility in fractionated stereotactic radiotherapy: An updateafter changing dental impression material. Radiother. Oncol. 50:239–40; 1999.

3. Graham, J.D.; Warrington, A.P.; Gill, S.S.; Brada, M. A non-invasive, relocatable stereotactic frame for fractionated radio-therapy and multiple imaging. Radiother. Oncol. 21:60 –2;1991.

4. Kooy, H.M. Adaptation and verification of the relocatable Gill-Thomas-Cosman frame in stereotactic radiotherapy. Int. J. Radiat.Oncol. Biol. Phys. 30:685–91; 1994.

5. Warrington, A.P.; Laing, R.W.; Brada, M. Quality assurance infractionated stereotactic radiotherapy. Radiother. Oncol. 30:239–46; 1994.

6. Scott, T.W.; Beach, J.L.; Mendiondo, O.A. A precision repeatlocalisation head frame for fractionated stereotactic radiotherapy.Med. Dosim. 22:5–8; 1997.

7. Bova, F.J.; Buatti, J.M.; Friedman, W.A.; et al. The University ofFlorida frameless high-precision stereotactic radiotherapy system.Int. J. Radiat. Oncol. Biol. Phys. 38:875–82; 1997.

8. Buatti, J.M.; Bova, F.J.; Friedman, W.A.; et al. Preliminary expe-rience with frameless stereotactic Radiotherapy. Int. J. Radiat.Oncol. Biol. Phys. 42:591–9; 1998.

9. Bova, F.J.; Meeks, S.L.; Friedman, W.A.; Buatti, J.M. Optic-guided stereotactic radiotherapy. Med. Dosim. 23:221–8; 1998.

10. Ebert, M.A.; Perry, A.M.; Harper, C.S. Implementation of a ste-reotactic radiotherapy system. Aust. Phys. Eng. Sci. Med. 22:145–52; 2000.

11. Menke, M.; Hirschfeld, F.; Mack, T.; et al. Photogrammetricaccuracy measurements of head holder systems used for fraction-ated radiotherapy. Int. J. Radiat. Oncol. Biol. Phys. 29:1147–55;1994.

12. IDL Research Systems Inc., Boulder, CO; 1998.13. Lasdon, L.S.; Waren, A.D.; Jain, A.; Ratner, M. Design and testing

of a generalised reduced gradient code for nonlinear programming.ACM Trans. Math. Software 4:34–50; 1978.

Medical Dosimetry Volume 28, Number 1, 200330