Académique Documents
Professionnel Documents
Culture Documents
ISSN 1533-0346
Volume 3, Number 3, June (2004)
©Adenine Press (2004)
Key words: High dose rate brachytherapy; Surface molds; Skin lesions; Dose optimization.
Introduction
Surface mold therapy was introduced in the beginning of the 20th century to treat
skin lesions following the discovery of radium (1). Radioactive molds became
one of the established radiation modalities to treat skin cancer along with super-
ficial and orthovoltage X-rays. The surface mold applicators were typically
loaded with gamma emitters such as Ra-226, Rn-222, Au-198, Co-60 or Cs-137
and they were widely used for the treatment of skin lesions of the extremities.
This technique was described in a report by Ashby et al. in 1989 (2) where they
used radon sources in wax and paraffin molds.
259
260 Sabbas et al.
mization points are within the 15-20% range relative to the pre- We then compare ∫Adt to the one derived based on the activ-
scription dose with a standard deviation of about 10%. ity of the HDR source multiplied by the sum of all the dwell
times of the HDR plan.
Of significant clinical importance when treating surface
lesions is the dose to the skin. In Figure 8 we plot the skin In Figure 9 we plot the ratio of the ∫Adt as derived from the
dose as a function of treatment area for a number of rectan- HDR brachytherapy calculations and that based on the inde-
gular planar surface molds. The skin dose is expressed as a pendent dosimetry verification, as outlined above, for various
percentage of the prescription dose and both the average and sizes of planar surface molds. Each surface mold is desig-
maximum dose to the selected skin points are shown. The nated by its size in terms of number of catheters times num-
skin points are positioned opposite to the source dwell posi- ber of positions (C × P) with 1 cm spacing between dwell
tions at a distance of 5 mm. Most of the calculations are positions. The prescription distance for all the molds was set
based on the geometric optimization, which is what is used to 1 cm (5 mm depth in tissue) except for the 11 × 9 mold
mostly in the clinic. For the 5 × 5 and 9 × 9 molds we also where the distance was set at 0.8 cm (3 mm depth in tissue).
show the calculated skin doses based on the dose point opti- For plans based on the dose point optimization algorithm the
mization which results in larger dose inhomogeneities and agreement is within 3%. For plans based on the geometric
maximum doses compared to the geometric optimization, optimization algorithm the agreement is acceptable, mostly
240% versus 175%, at the skin for the 5 × 5 mold. within 8%, though not as good as the dose point optimization
plans that seem to replicate the Paterson Parker rules better.
In general, the skin dose decreases with size of the surface
area treated since the dose fall-off is less with larger surface Comparison with Analytical Model
molds. This decrease is more pronounced when no opti-
mization is used and is least noticeable when the dose point We studied the exact dependence of the depth dose from an
optimization is used, since the optimization compensates for HDR surface mold on the treatment area by considering an
the finite surface area by increasing the peripheral dwell analytical model of the surface mold. For 192Ir the radial
times. For no optimization the maximum skin dose occurs dose function g(d) exhibits less than 1% variation over the
near the center of the treatment area. The dose variation at first 5 cm distance from the source (21). The variation of
the skin for dose-point optimized plans is the largest. The dose with distance d along the central axis of the mold is
center dose is close to the minimum and the maximum therefore mainly due to the geometry factor G(d).
occurs at the edge of the treatment area where the optimized Considering a circular mold of radius R, total activity A and
dwell times are the largest. We have observed skin reactions uniform activity density α, the geometry factor can be cal-
at the treatment edges for patients treated with plans opti- culated by integrating over the area of the applicator.
mized on dose points and this is a clinical confirmation of
[ ( )]
the substantial surface hot spot at the periphery. The least 2π R ρdρdθ 2
α∫ α∫ 2 2
0 ρ + d 1 R
variation in the skin dose and the smallest hot spots are G (d ) = 0
= ln 1 + . [1]
A R2 d
exhibited by plans calculated using geometric optimization.
Comparison with a Classical Brachytherapy System For treatment distances much greater than the lateral extent of
the applicator, R<<d, G(d)=1/d2 which is the familiar inverse
The optimization methods achieve a dose uniformity at the pre- square dependence of the depth dose from a point source. In
scription dose plane by yielding a set of unequal dwell times. Figure 10 we plot the dose fall-off, as predicted from Eq. [1] for
This is in close analogy to the Paterson Parker rules of the surface molds of uniform activity of radii 2, 3, 4, and 5 cm. For
Manchester LDR brachytherapy system (17), which is a set of comparison the 1/d2 curve of a point source and 1/d curve from
rules for arranging Radium or other types of low activity an infinitely long linear source are also plotted. The curves are
sources in order to achieve uniform dose to a plane or volume. normalized to a distance of 1 cm from the source plane or to a
The total activity multiplied by the duration of the implant ∫Adt, depth of 0.5 cm in tissue. As seen from Figure 10 the surface
is a good indicator of the dose delivered by such an implant. dose varies from 176% for the smallest 4 cm diameter applica-
tor to142% for the largest 10 cm diameter applicator. The depth
The Manchester tables (20) for planar implants list the dose is also steeper for the smaller size applicators.
mg·hr to deliver 10 Gy to tissue for different surface areas
and treatment distances from the source plane (RA). We In Figure 11 we compare the depth dose of the analytical
have used these tables to independently verify ∫Adt for com- model and the computerized HDR brachytherapy calcula-
puterized HDR plans. Specifically, we calculate ∫Adt using tions for a 6x6 cm2 mold (R=3.4 cm). Both the unoptimized
RA, the prescription dose D and a conversion factor of (equal dwell times) and the dose-point optimized calculations
6.3734 (Ci·s/mg·hr). The RA is based on a surface area are plotted. All three curves are normalized to a depth of 0.5
Figure 11: Depth doses from a 6 × 6 surface applicator based on the ana-
Figure 8: Skin dose expressed as a percentage of the treatment dose as a lytical model (continuous curve) and the HDR stepping source non-opti-
function of the applicator size (# of catheters x # of dwell positions). There mized and optimized dose calculations.
is a 0.5 cm bolus between the plane of the catheters and the skin and the
dose is prescribed to a depth of 0.5 cm in tissue. Both the average and max-
imum skin doses are plotted for the geometric optimization (all sizes) and
for the dose point optimization (two sizes).
Figure 12: Comparison of the depth doses for different modalities: Co-60
80 SSD, 6 MeV electrons 100 SSD, 100kVp superficial X-rays 15 SSD and
a 6 × 6 surface mold with a 0.5 cm bolus between the skin and the plane of
the catheters. The HDR calculations are optimized. The dose is normalized
Figure 9: Agreement between the HDR contact therapy calculation and to 100% at the skin for the superficial and HDR modalities and dmax for the
the Paterson Parker tables of the integrated activity for various surface Co-60 and the electrons.
applicator sizes.
Figure 10: Depth dose calculations for various radii of surface applicators
based on the analytical model of the applicator. For comparison the inverse
square dose fall-off from a point source and the inverse distance dose fall- Figure 13: Comparison of the dosimetry on a curved surface from A. 6 ×
off from an infinitely long linear source are also shown. The dose is nor- 6 HDR surface applicator with 0.5 cm bolus and B. 6MeV electron beam
malized to a depth of 0.5 cm in tissue. with a 6 × 6 cone and 1 cm bolus.
cm in tissue. The analytical model predicts well, to within a In conclusion, some of the key features of HDR contact ther-
few percent, the depth dose from the stepping source unopti- apy are summarized below:
mized plan. Upon optimization the depth dose becomes less
rapid by as much as 7% compared to the unoptimized plan (I) Surface molds can conform easily to the curva-
since peripheral dwell positions are weighted heavier. ture of the treatment surface. On the other hand,
there are uneven air gaps between the tip of the
Discussion electron cone and the patient’s surface, especial-
ly when treating curved surfaces, which con-
In Figure 12 we plot the percent depth dose for three external tribute to the dose inhomogeneity at depth with
beam modalities along with that from an optimized 6 × 6 cm2 electron beam therapy.
HDR surface mold used clinically. The high dose region for 6
MeV electrons ( > 80% of the maximum dose) extends over (II) Optimization algorithms are routinely incorporated
1.8 cm from the surface and is then followed by a precipitous in the HDR brachytherapy calculations and con-
drop in the radiation dose. In contrast the depth doses of the tribute further to the dose homogeneity at depth.
HDR 192Ir and superficial X-rays do not exhibit any skin spar-
ing and fall off exponentially with depth. For the HDR mold (III) Through adjustment of dwell positions and dwell
the dose is already down to 67% and 52% of the skin dose at times manually or as part of the optimization
depths of 5 mm and 1 cm respectively. HDR contact therapy process the radiation field can be customized in
treats more superficially than electrons. HDR contact therapy terms of its shape and intensity profile. No custom
also exhibits faster dose fall-off than the superficial beam. cutouts are needed to shape the radiation field.
The HDR mold, due to its flexibility, is able to conform well (IV) The dose fall-off is faster than Superficial
to the skin area and the dose can be delivered uniformly at Therapy and is thus better suited for the treatment
depth when treating curved surfaces as well. The electron of more superficial lesions.
cones on the other hand are rigid and as a result there can be
appreciable variations in the treatment SSD across the treat- (V) It is easy to shield selected areas of the treatment
ment area. Obliquity effects for electrons (22) when incident field using Pb sheets. The HVL for Ir-192 is 3
on curved surfaces can further contribute to the dose inho- mm of lead.
mogeneity at depth. We plot in Figure 13 the isodose distri-
butions on a curved surface (5 cm radius of curvature) from (VI) HDR contact therapy can be used to treat struc-
a standard 6cm × 6cm HDR surface applicator and from a 6 tures with excessive curvature like the septum of
MeV electron beam, 6 × 6 cone, 100 SSD and 1 cm bolus. the nose. It can be also used to treat cavities not
The thickness of the bolus used with electrons was chosen so easily accessible by electron applicators like the
that the prescription level of 100% occurs at approximately buccal mucosa.
7 mm from the surface at the center of the treatment area for
both modalities. For the HDR mold, the prescription dose is (VII) Radiation safety issues are minimal for the HDR
delivered uniformly at depth across the whole width of the remote afterloader since there is no radiation
treatment area. For electrons, on the other hand, the extent exposure to personnel.
of the prescription dose falls short of the intended treatment
width by as much as 2 cm. (VIII) The treatment times are of the order of a few min-
utes. Hence the dose is delivered quickly and
As mentioned earlier, a flat planar geometry is almost always effectively as an out-patient procedure.
assumed during dosimetry calculations when we use stan-
dard rectangular applicators to treat skin lesions of the (IX) Scheduling of HDR contact treatments does not
extremities and the torso. The dwell times for the 6cm × 6 impact on the busy schedule of the linac. In addi-
cm mold used to treat the curved surface shown in Figure tion, the HDR dose fractionation calls for fewer
13a were calculated based on a flat surface geometry to fractions or patient visits.
deliver 400 cGy at a treatment depth of 0.5 cm in the skin.
As seen from this figure, the curvature of the treatment area
causes the prescription isodose to appear deeper than 5 mm
by 1.7 mm on the concave side. Therefore, assuming a flat
geometry results in isodoses being shifted no deeper than 2
mm as long as the treatment surface curvature is not exces-
sive (more than 5 cm in radius).
References 12. Evans, M. D. C., Yassa, M., Podgorsak, E., et al. Surface Applicators
1. Meredith, W., Massey, J. Fundamental Physics of Radiology. Bristol, for High Dose Rate Brachytherapy in Aids-related Sarcoma. Int. J.
Wright (1977). Radiat. Oncol. Biol. Phys. 39, 769-774 (1997).
2. Ashby, M. A., Pacella, J.A., De Groot, R., et al. Use of Radon Mould 13. Svoboda, V. H. J., Kovarik, J., Morris, F. High Dose Rate Micro-
Technique for Skin Cancer: Results from the Peter MacCallum selectron Molds in the Treatment of Skin Tumors. Int. J. Radiat.
Cancer Institute. Br. J. Radiol. 62, 608-612 (1989). Oncol. Biol. Phys. 31, 967-972 (1995).
3. Lovett, R. D., Perez C. A., Shapiro, S. J., et al. External Irradiation 14. Guix, B., Finestres, F., Tello, J. I., Palma, C., Martinez, A., Guix, J.
of Epithelial Skin Cancer. Int. J. Radiat. Oncol. Biol. Phys. 19, 235- R., Guix, R. Treatment of Skin Carcinomas of the Face by High-
242 (1990). dose-rate Brachytherapy and Custom-made Surface Molds. Int. J.
4. Joslin, C. A. F., Liversage, W. E., Ramsey, N. W. High Dose Rate Radiat. Oncol. Biol. Phys. 47, 95-102 (2000).
Treatment Mould by Afterloading Technique. Brit. J. Radiology 42, 15. Jolly, D. E., Nag, S. Technique for Construction of Dental Molds for
108-112 (1969). High-dose Rate Remote Brachytherapy. Spec. Care Dentist. 12, 219-
5. Joslin, C. A., Flynn, A. High Dose-rate Brachytherapy in the 224 (1992).
Treatment of Skin Tumors, in Principles and Practice of Brachy- 16. Harrison, L. B. HDR Intraoperative Brachytherapy, in International
therapy using Afterloading Systems, pp 393-399. Eds. C. A. Joslin, Brachytherapy: 8th International Brachytherapy Conference, pp 13-
A. Flynn, E. J.Hall. Oxford University Press, London (2001). 17. Nucletron-Oldelft, Veenendaal, The Netherlands (1995).
6. Sharma, S. C., Negi, P. S., Gupta, B. D. Selectron HDR Surface 17. Meredith, W. J. Radium Dosage, in The Manchester System. Ed., E.
Mould Therapy, in Brachytherapy 2, pp 572-575. Ed., R. F. Mould. S. Livingstone. Edinburgh and London (1967).
Nucletron International, Leersum, The Netherlands (1989). 18. Van der Laarse, R., De Boer, R. W. Computerized High Dose Rate
7. Porrazzo, M., Stabile, L., Ross, R., Moorthy, C., Tchelebi, A., Brachytherapy Treatment Planning, in Brachytherapy HDR and
Hilaris, B. S. HDR Remote Afterloading as an Alternative to LDR, pp 169-183. Eds., A. A. Martinez, C. G. Orton, R. F. Mould.
Electrons for Therapy of Superficial Tumours. Selectron Activity Nucletron International Publisher, Leersum, The Netherlands (1990).
Journal, Vol. 6, 11-13 (1992). 19. Edmundson, G. K. Geometry Based Optimization for Stepping
8. Kitchen, G., Dalton, A. E., Pope, B. P., Smith, P. D., Powner, M. Source Implants, in Brachytherapy HDR and LDR, pp 184-192.
Surface Applicator for Basal Cell Carcinoma of the Right Pinna: A Eds., A. A. Martinez, C. G. Orton, R. F. Mould. Nucletron
Case Report. Selectron Activity Journal, Vol. 5, 140 (1991). International Publisher, Leersum, The Netherlands (1990).
9. Kitchen, G., Dalton, A. E., Evans, M., Pope, B., Smith, P. D., 20. Johns, H., Cunningham, J. The Physics of Radiology. Springfield, C.
Selectron LDR Mould for Large Area Basal Cell Carcinoma. C. Thomas (1983).
Selectron Activity Journal, Vol. 4, 72-73 (1990). 21. Nath, R., Anderson, L. L., Luxton, G., Weaver, K. A., Williamson, J.
10. Brock, A., Prager, W., Pohlmann, S. Methodik der Kontakttherapie F., Meigooni, A. S. Dosimetry of Interstitial Brachytherapy Sources:
mit Hilfe des Afterloading-Verfahrens im Kopf-Hals-Bereich. Recommendations of the AAPM Radiation Therapy Committee Task
Radiobiol. Radiother. 29, 609-615 (1988). Group No. 43. Med. Phys. 22, 209-234 (1995).
11. Brock, A., Pohlmann, S., Prager, W. Surface Applicators for HDR 22. Biggs, P. J. The Effect of Beam Angulation on Central Axis Percent
Brachytherapy in the Head and Neck Region. Selectron Brachyther- Depth Dose for 4-29 MeV Electrons. Phys. Med. Biol. 29, 1089-
apy J. Suppl. 3, 22-25 (1992). 1096 (1984).