Pergamon

Int. J. Radiation Oncology Biol. Phys., Vol. 31, No. 4, pp. 967 972, 1995 Copyright © 1995 Elsevier Science Ltd Printed in the USA. All rights reserved

0360-3016/95 $9.50 + .00

0360 -3016(94)00485 -4

• Technical Innovations and Notes

HIGH D O S E - R A T E M I C R O S E L E C T R O N M O L D S IN THE T R E A T M E N T OF SKIN T U M O R S

VLAD1MIR H . J. SVOBODA, M.D., JOSEF KOVARIK, M . D . AND FORBES MORRIS , M . S c .

Department of Radiotherapy and Oncology, Saint Mary's Hospital, Portsmouth PO3 6AD, UK

Purpose: The feasibility of high dose rate ir idium afterloaded molds in the treatment of skin tumors. Methods and Materials: Expanded silicone rubber and bronchial applicator tubes were used in the construc- tion of various molds. The number of tubes used and the separation between them depends on the size and area to be treated. Source dwell position and time are set to follow P a t e r s o n - P a r k e r rules. Single plane expanded silicone rubber molds between 15 x 15 m m and 60 × 70 m m were used in 120 different sites, and of the others, seven were perspex double molds, two were cylindrical applicators, and one was a part ial ly shielded intranasal applicator. One hundred and thirty lesions of various site and histology were treated in all. After exclusion of treatments that were combined with either external beam or chemotherapy, 106 lesions (76 patients) were evaluated. Results: Full response was obtained in all but four basal cell carcinomas. The acute reaction ranged from moist desquamation (27 sites) to erythema only. Follow-up at 5 or more months revealed no changes whatever in 47 sites; 53 showed an excellent cosmetic outcome, although slight changes in pigmentation or minimal atrophy was demonstrable and 6 sites became noticeably atrophic with patchy pigmentation. For 9.6 months of average follow-up time, no recurrences have been observed. Conclusion: The high-dose-rate iridium-loaded skin applicators offer the possibility of improved therapeutic ratio in the treatment of superficial skin tumors. With the availability of a high-dose-rate afterloader this technique is simple and straightforward.

Skin molds and applicators, High-dose-rate afterloader.

I N T R O D U C T I O N M E T H O D S A N D M A T E R I A L S

Primary nonmelanoma skin cancers are best treated by surgery or radiotherapy. However , convent ional superfi- cial radiotherapy often leads to cosmet ic damage and is poor ly tolerated in some areas such as the lower leg or dorsum of hands. The molds used t radi t ional ly for treat- ment of lesions on the shin were cumbersome, and re- quired hospital izat ion. We have modif ied this technique, by fixing MicroSelec t ron ~ intrabronchial af ter loading tubes to the surface of 7 .5-mm thick expanded si l icone rubber, and used convent ional fract ionation on an outpa- tient basis. The technique was simple, and popular with staff and patients. As the early skin react ions were mild and tumor responses good, it became the standard method of t reatment for superficial skin mal ignancies situated more than 7 cm from the eyes.

One hundred and thirty lesions on 87 patients were treated. Eleven patients were excluded from the analysis because of con temporary chemotherapy or a combinat ion with external beam treatment. The average age was 72 years (range 4 6 - 9 2 years) and patients compr ised 45 men and 31 women. The average fol low-up was 9.6 months (range 5 - 2 2 months). Eight sites received treatment for chest skin metastases from breast carc inoma and two for intradermal non-Hodgk in ' s l ymphoma (one patient). Ninety-s ix sites were treated for pr imary skin neoplasms; of these, 9 were Bow e n ' s disease, I 1 squamous cell carci- nomas ( including penis and lower lip), and 76 basal cell carc inomas where the cl inical diagnosis was confirmed his tological ly in 31 sites. Nine lesions were situated on the pinna, 28 on the head and neck, 35 on the trunk, 16 on the hand, 5 on the arm, and 13 on the lower limbs.

Reprint requests to: Dr. Vladimir H. J. Svoboda, Department of Radiotherapy and Oncology, Saint Mary's Hospital, Milton Road, Portsmouth PO3 6AD, UK Acknowledgments--We are most grateful for the expertise and enthusiasm of Mr. M. Lanighan and Mr. E. Saunders who pro- duced our molds promptly and efficiently. The authors also

wish to thank Mrs. B. Boyer for her invaluable help with prepa- ration of the material and text.

Accepted for publication 26 August 1994. High-Dose-Rate ~92Ir Brachytherapy Unit, Nucletron Inter-

national, The Netherlands.

967

968 I.J. Radiation Oncology • Biology • Physics

Table 1. Size of the tumors treated by single molds

T size(cm 2) < 0.5 0.51-0.8 0.81 3.0 3.1-6.0 > 6.1 No. of tumors 15 24 35 12 10

Volume 3 I, Number 4, 1995

Table 2. Active areas off the single molds

Active area(cm 2) < 4 4.1 8.0 8.1-12.0 > 12.1 No. of molds 39 31 15 1 I

The size of tumors treated by a single mold is shown in Table I and active areas of the single molds are shown in Table 2.

Mould construction and physical details The molds are constructed of expanded silicone rubber

of 0.75 cm thickness. MicroSelectron bronchial applicator tubes, of 2 mm outer diameter and 100 cm length, are sandwiched between two layers of this material so that the mold can be applied either way up. This gives a treatment distance of 0.85 cm to the center of the applica- tor tube. The number of tubes used and the separation between them depends on the size of the area to be treated. Source dwell positions and dwell times are set to follow Pa te rson-Parker rules as closely as possible, the area being divided into strips of width approximately equal to twice the "d is tance ." Dwell times for the inner applicator tubes are either one-half, or one-third, that of the outer tubes, depending on the size of the mold. To simulate "cross ing the active ends ," the dwell times at the limits of each line are twice that of the other dwell times along the outer applicator tubes.

Figure 1 is an example of dwell weights and positions for a 3 × 3 cm mold with 1.5 cm between adjacent lines and 0.5 cm between each dwell position. There are 21 dwell positions, with a total weight of 49 units.

The average dose is calculated over the treated area, taking the source to be an isotropic point source, absorp- tion and scattering within the medium being taken into account using the method of Sakelliou et al. (1), as recom-

mended in the Report of a Joint Working Party of the BIR and the IPSM (2). The dwell weights are considered to be equivalent to the dwell times in seconds. In the above example, for a 3 × 3 cm square, the dose at 0.85 cm from a l0 Ci source is 1.98 2 Gy in 49 s. The actual dwell times for, say, 3.75 Gy on the surface from a 10 Ci source, will be as follows:

7.6 7.6 7.6 3.8 1.9 3.8 3.8 1.9 3.8 3.8 1.9 3.8

3.8 1.9 3.8 3.8 1.9 3.8 7.6 7.6 7.6

i.e., total time = 9.31 s.

Such calculations are performed using the Excel spread- sheet 2 and checked independently using inhouse software. A number of standard molds are now available in this department, which may be used to treat any area, square or rectangle, from 1 x 1 cm to 6 x 6 cm, in 0.5 cm steps (Fig. 2). Other sizes, or molds for uneven surfaces, are custom made, while a sandwich mold with 1 cm between the two planes is used to treat some sites, such as lesions on the pinna (Fig. 3). Cylindrical surface moulds have also been used to treat the penis and thumb (Fig. 4). With this arrangement of source dwell positions and weights for the surface molds, the dose rate over the treated area is acceptably uniform. For instance, any point on the sur- face of a 4 x 4 cm mold is within 5% of the average

Source dwell w ~ t

Expanded nlbbe r silicone ~ II[" ~ r tube

MicroSelectron i

So~-~e ~ i ~,

< 1.5 ::~ c l l ' t

Fig. 1. Example of dwell weights and positions for a 3 × 3 cm mold.

2 Microsoft Corporation.

MicroSelectron skin molds • V. H. J. SVOBODA et al. 969

Fig. 2. Standard molds.

dose rate over 90% of the treated area. There is, however, a great variation with treatment distance, 1 mm difference at 0.85 cm treatment distance giving rise to an 8% change in dose.

To improve the depth dose for thicker lesions a double thickness mold is used, having a treatment distance of 1.6 cm. For example, the dose at 0.5 cm deep with a treatment distance of 1.6 cm is about 10% greater than

Fig. 3. Sandwich mold.

970 1. J. Radiation Oncology• Biology • Physics Volume 3 I, Number 4, 1995

Fig. 4. Cylindrical mold,

with a t reatment distance of 0.85 cm. The dose variat ion with depth is s l ightly less when a double thickness mold is used, being about 6.5% per mi l l imeter at the surface.

Figure 5 shows the compar ison of percentage depth doses of a 4 × 4 cm square mold on the MicrSelectron, t reatment distance 0.85 cm, with a 4 × 4 cm square field at 15 cm FSD and 80 kV, using either a 0.8 m m A1 filter or 2.0 m m A1 filter, on the Therapax Series 3 Superficial x-ray unit. 3 These two filters provide beams of HVL 1.2 mm AI and 2.25 m m AI, respect ively.

t~2Ir has an average gamma- ray energy of approxi- mately 360 keV, the main intensit ies fall ing between 296 keV and 468 keV. The eye dose is a lways calcula ted for molds appl ied to the head or neck, and treatment only proceeds if this dose is sat isfactori ly low. The HVL in lead is approximate ly 3 m m for ~92Ir, and lead sheet of 4 m m thickness is used to reduce the dose to the eyes when treating the face or ear, the radiat ion passing obl iquely through the lead.

The duration of the t reatment is easi ly tolerated by the

100 90 l --1 80 ~ ] 70 ? -~> ' - . I 60 " ~ < ' " SIV'L 1 ~ 2 r ~

D -dOo' - "-" % 30 20 10 0 i

0 0.5 1 1.5 2 2.5 3 3.5

Depth in cm

Fig. 5. Comparison of % depth doses, 4 × 4 cm field.

4 4.5 5

3 PANTAK Limited, UK.

MicroSelectron skin molds • V. H. J. SVOBODA et al. 971

Table 3. Dose, fractionation, volume

No. of Total dose No. of sites fractions (Gy) Fractionation

54 1 18-22 single exposure 17 3 27- 30 weekly 27 l 0 40 daily

2 10 40 twice a day 2 5 30 daily 1 (48 ml cylinder) 15 50 daily 1 (37 mi cylinder) 10 46 daily 2 (NHL) 1 12 single exposure

patient. To achieve a dose of 4 G y on the surface (SSD = 0.85 cm) of a 3 × 4 cm mold it takes approximate ly 2 min with a new source of 10 Ci activity. Using this type of mold the t reatment t ime never exceeded 5 min, but using the same mold to achieve a dose of 20 Gy would take between 10 and 20 rain.

Dose and f rac t iona t ion Doses and fract ionations used are shown in Table 3.

Doses were prescr ibed at the surface of the appl icator in all cases.

R E S U L T S

Comple te tumor regression was observed in 102 sites while four basal cell carc inomas remained residual after 4, 4, 11 and 12 months, respect ively. Within the fol low- up per iod we observed no local recurrences. Some small tumors s imply d isappeared within 6 weeks or less without any side effects not iced by the patient; some took up to 12 weeks to clear complete ly .

Six sandwich molds were used for the treatment of basal cell carc inoma of the pinna. Act ive areas were be- tween 15 and 30 m m square. Comple te response and an excel lent cosmet ic outcome was achieved in all six cases. Moist desquamat ion was observed in one patient (30 x 20 mm, 27 Gy in 3 fractions); otherwise, the only side effect was slight or modera te erythema.

One patient was treated for squamous cell carc inoma of the lower lip by a cus tom-made sandwich mold with shielding of the poster ior aspect to protect the tongue and buccal mucosa. Mois t desquamat ion deve loped but the tumor responded comple te ly and the final cosmetic result was excellent.

Two patients were treated using cyl indr ical surface molds, one with a large (60 ml) squamous cell carc inoma of the thumb, replacing the nail. The inner d iameter of the cyl inder was 28 m m and length was 50 mm. A dose of 46 Gy was given in 10 fract ions over 12 days and the only side effect was dry desquamat ion and a crust on the surface of the tumor. Af ter 16 weeks o f the heal ing pro- cess a very good cosmet ic result was achieved. The other patient, with a large verrucous carc inoma of the glans penis, was also successful ly treated using a cyl indr ical

mold (inner d iameter 35 m m and active length 50 mm). Two small (10 m m in diameter) cutaneous nodules of low grade non-Hodgk in ' s l ymphoma were treated using single molds 20 x 20 mm, giving 12 Gy in 1 fraction. Comple te response was recorded within 2 weeks and the final cosmet ic outcome was excellent.

Five patients were treated for metastat ic skin deposi ts from breast carc inoma on previously irradiated chest walls. The active areas were between 2 x 2 cm and 6 x 4 cm, and either dai ly or three t imes weekly sessions were used. Al l eight lesions regressed fully within 6 - 1 2 weeks. In four sites e ry thema was observed and in three, moist desquamation. One area showed no sign of radio- therapy at all. All except one of the eight lesions healed with good cosmet ic result; for the one only-fa i r cosmet ic result, a 3 X 6 cm area after 40 Gy in 12 days healed with patchy atrophic pigmentat ion.

Eleven sites were treated by molds with an active area of more than 12 c m 2. In all except one, complete response was achieved with acceptable side effects (moist desqua- mation was recorded in five sites, and in three slight e ry thema was observed); in the remaining patient only partial response of his large basal cell carc inoma on the dorsum was achieved and, therefore, the patient was sub- sequently treated by superficial x-ray. In the major i ty , the final cosmetic result was very good or excellent.

For those four his tological ly verified basal cell carcino- mas where we failed to achieve comple te response, all were large lesions, three being larger than 2 cm in d iame- ter. Al l tumors were ulcerated and raised an est imated 3 mm (minimum) in thickness.

Tolerance

Treatment was well tolerated. It never became neces- sary to interrupt any course because of side effects. Mois t reaction was observed in 26 sites, usual ly appearing dur- ing the second week after comple t ion of t reatment and healing within 2 or 3 weeks. In 32 sites minimal side effects (slight e ry thema or dry desquamat ion) were re- corded, lasting usual ly between 2 and 4 weeks. In 48 sites no side effects were descr ibed; the majori ty o f small tumors s imply disappeared. There were no substantial differences in the incidence and severi ty of side effects and the durat ion of the heal ing process among the various sites. Treatment was well tolerated in areas of known poor radiat ion tolerance such as the dorsum of the hand (12 tumors) or shin (7 tumors).

D I S C U S S I O N

Surface appl icators were introduced in the first quarter of this century to deal with skin tumors involving irregular surfaces or those sitting on the cart i lage, which tolerated low vol tage x-ray therapy badly. Treatment by these ap- pl icators loaded with gamma ray sources (Ra, Co, Au, Rn, Cs) was also better tolerated by the poor ly vascu- larized skin on the leg and dorsum of the hand. With the

972 I.J. Radiation Oncology • Biology • Physics Volume 31, Number 4, 1995

introduction of higher energy x-rays and electron beams, molds became less popular. The last major report about the use of this technique was published by Ashby e t al.

in 1989 (3). These authors treated 642 patients for non- melanoma skin malignancies in areas of poor radiation tolerance, not deeper than 4 ram. They used wax molds loaded with radon sources.

Brock e t al. of the University of Leipzig designed a special surface applicator for use with an HDR brachy- therapy unit (4, 5). Its advantage is in an improved depth dose and the possibility of treating close to the eye; how- ever, the largest applicator is only 3 cm in diameter. In 1992 they reported on 120 patients; in 108 of them, com- plete remission was obtained. The tumor dose was calcu- lated at 8 mm and between 30 and 40 Gy was delivered in 4 - 8 fractions. Neither the overall time nor the exposure times were given, and cosmetic results were not men- tioned.

Our technique, which uses soft expanded silicone rub- ber applicators loaded with an HDR iridium source from a MicroSelectron, allows treatment of any flat areas away from the eyes by a single mold up to approximately 24 cm square. Specially constructed double molds or cylindrical applicators can be loaded in a similar way. The early tumor responses have been very satisfactory and the skin tolerance excellent, indicating the possibility of a thera- peutic ratio superior to conventional superficial irradia- tion.

Reactions, although generally mild, were not entirely predictable. This could be a result of occasional difficul- ties with proper fixation of the molds to the skin surface, and a short treatment distance with consequent low pene- tration of the radiation.

Tumors thicker than 3 mm require an increased source- skin distance. This, however, considerably increases the exposure times and makes the technique impractical for more deeply infiltrating tumors. Before each session, any superficial crust must be removed. In some sites we had difficulty with adequate fixation of the silicone applica- tors, and their perfect adherence to the skin and tumor surface. Perspex masks, elastic bandages, and dressings can help. For irregular surfaces a traditional applicator can easily be constructed. The silicone rubber mold could also be sealed to the skin around the lesion by a suitable adhesive. We are, at present, investigating the use o f a more pliable tissue equivalent material like Superflab. 4 We expect this to have better adhesive qualities and so improve overall dose distribution.

As the average gamma ray energy is approximately 360 keV and we used doses comparable to conventional superficial radiotherapy, there was some risk of underdos- ing the lesions because of a lower RBE. The relative effect of, say, a standard 100 kV x-ray beam, and a fast moving small Ir source with a variable dose rate, cannot easily be predicted. Only experience and dose escalation study will tell us the optimal fractionation and total dose.

The lower RBE and poor penetration o f the beam may at least partly account for the good cosmetic results we observed. However, early tumor control was so satisfac- tory that we have now changed our policy in the treatment of the majority of skin carcinomas. With more experience, we hope gradually to improve our technique and rational- ize further our fractionation.

The superficial dose distribution and good cosmetic results make this method potentially suitable for the treat- ment of non-neoplastic skin conditions such as eczemas, pustular psoriasis, or Ha i ley-Hai ley disease.

R E F E R E N C E S

1. Sakelliou, L.; Sakellariou, K.; Sarigiannis, K.; Ange- lopouios, A.; Pelvis, A.; Zarris, G. Dose rate distributions around 6°Co, 137Cs, J98Au, JgZIr, 241Am, J25I (models 6702 and 6711) brachytherapy sources and the nuclide ~gTcm. Phys. Med. Biol. 37:1859-1872; 1992.

2. Aird, E. G. A.; Jones, C. H.; Joslin, C. F. A.; Klevenhagen, S. C.; Rossiter, M. J.; Welsh, A. D.; Wilkinson, J. M.; Woods, M. J.; Wright, S. J. Recommendations for brachy- therapy dosimetry. Report of a Joint BIR/IPSM Working Party. The British Institute of Radiology; 1993.

3. Ashby, M. A.; Pacella, J. A.; de Groot, R.; Ainslie, J. Use of radon mould technique for skin cancer: Results from the Peter MacCallum Cancer Institute (1975-1984). Br. J. Radiol. 62:608-612; 1989.

4. Brock, A.; Pohlmann, S.; Prager, W. Surface applicators for HDR brachytherapy in the head and neck region. Selectron Brachytherapy J. Suppl. 3:22-25; 1992.

5. Brock, A.; Prager, W.; Pohlmann, S. Methodik der Kontakt- therapie mit Hilfe des Afterloading-Verfahrens im Kopf- Hals-Bereich. Radiobiol. Radiother. 29:609-615; 1988.

4 SUPERFLAB Bolus Material, Nuclear Associates, Division of Victoreen, Inc., New York.