In/ J Rudmron Oncolo~v BIO/. Ph?, Vol. Printed in the U.S.A. All rights reserved.

19, pp. 65 I-660 Copynght

0360.3016/90 $3.00 + .OO 0 1990 Pergamon Press plc

??Original Contribution

TUMOR INDUCTION FOLLOWING LATE RESULTS OF THE NATIONAL

INTRAOPERATIVE RADIOTHERAPY: CANCER INSTITUTE CANINE TRIALS

M. BARNES, M.D.,’ P. DURAY, M.D.,’ A. DELUCA, M.S.,2 W. ANDERSON, PH.D.,~ W. SINDELAR, M.D., PH.D.~ AND T. KINSELLA, M.D.4 ‘Fox Chase Cancer Center,

Philadelphia,PA; ‘National Cancer Institute, Bethesda, MD; ‘Indiana University School of Medicine,

Terre-Haute,

IN; and 4University of Wisconsin Clinical Cancer Center, Madison, WI

Intraoperativeradiotherapyhas been employedin human cancer research for over a decade.Since 1979, trials to assess the acute and late toxicity of IORT have been carried out at the National Cancer Institute in an adult dog model in an attempt to establish dose tolerance guidelines for a variety of organs. Of the 170 animals entered on 12 studies with a minimum follow-up of 2 years, 148 dogs received IORT, 22 control animals received only surgery. Animals were sacrificed at designated intervals following IORT, usually at 1, 6, 12, 24, and 60 month intervals. 102 of 148 irradiated dogs were sacrificed ~24 months; 46 dogs were followed 224 months after IORT. To date, 34 of the 46 animals have been sacrificed; the 12 remaining animals are to be followed to 5 years. These 12 animals have minimum follow-up of 30 months. In the irradiated group followed for 224 months, 10 tumors have arisen in 9 animals. One animal developed an incidental spontaneous breast carcinoma outside the IORT port, discovered only at scheduled post-mortem exam. The remaining nine tumors arose within IORT ports. Two tumors were benign neural tumors-a neuroma and a neurofibroma. One animal had a “collision” tumor comprised of grade I chondrosarcoma adjacent to grade III osteosarcoma arising in lumbar vertebrae. Two other grade III osteosarcomas, one grade III fibrosarcoma, and one grade III malignant fibrous histiocytoma arose in retroperitoneaI/paravertebraI sites. An embryonal rhabdomyosarcoma (sarcoma botryoides) arose within the irradiated urinary bladder of one animal. No sham irradiated controls nor IORT animals sacrificed 40

Left pneumonectomy

15

12

>40

Right thoracotomy mobilization of esophagus

13

12

>40

11. Nerve 1(12, 14)

Laparotomy, unilateral plexus

27

24

60

12. Nerve II (12)

See nerve I

12

12

170

148

9. Branch. (22)

(1)

stump

10. Esophagus

(34)

Total dogs

of

and cystotomy

exposure of Iumbosacral

of

>34

IORT-related tumors 0 M.

prescribed in that particular study design or as indicated by the veterinary staff for diagnostic evaluation of clinically evident toxicity. Animals were sacrificed at predetermined intervals from treatment in order to assess acute, subacute, and late effects of the IORT, or for compassionate reasons as indicated. Post-mortem examination consisted of full gross autopsy and histopathologic assessment of both irradiated and non-irradiated tissues. Special histologic staining techniques, as well as immunoperoxidase staining of tissues, were performed as necessary to clarify pathologic diagnoses.

Esperimrntul

unimuls

All experiments included in this analysis used adult male and female purebred American Foxhounds. The Foxhounds ranged in age from 6 to 18 months and weighed 20 to 35 kg. These large animals were necessary to allow for careful, adequate surgical exposure and placement of IORT cones for anatomical coverage comparable to that used in human IORT research. The animals were maintained in sheltered runs and were fed dog chow and water ad libitum. All animals were under the constant supervision of the National Institute of Health Veterinary Resources Branch.

Stlrgical prowdww Dogs were fasted for 12 hours prior to surgery. The animals were anesthetized with intravenous pentobarbital using 25 mg/kg, with supplementary doses as required to maintain anesthetic effect. Dogs had endotracheal intubation and spontaneously breathed room air, a technique which provided normal blood oxygen tensions (39). Ringer’s lactate intravenous infusion of approximately 25 ml/kg/hr was used during the surgeries. Twelve hours following the procedure the animals were allowed water and food as tolerated. The details of the various surgeries are described previously (1, 10, 13, 14, 22, 34-39, 42). A summary of the various surgeries performed and the numbers of dogs undergoing each procedure are presented in Table 1.

Rudiution treutment jields The treatment fields varied with each study. Standard IORT lucite cones of rectangular and circular field shapes were used. The field parameters are listed in Table 2. Each cone was sterilized for the experimental treatments, defined the treatment volume, and served as retractor to prevent the slippage of other tissues, not of interest in this particular study, into the IORT field. The treatment cone was coupled to the linear accelerator beam axis for the actual treatment. Sham-irradiated animals had the lucite cones placed in the surgical bed, and the cone was docked onto the linear accelerator and left in place for a period of time comparable to an actual IORT treatment. The IORT portals and the “sham” portals were then marked using steel surgical clips to allow for later localization.

653

BARNES et al.

Radiation parameters Linear accelerators were used to deliver the electron beam irradiation. The electron energies used varied with the study and ranged from 9 MeV to I3 MeV. The doses delivered varied with each study and are listed in Table 2. The radiation dose was quoted at the 90 to 100% isodose line. The dose rate of IORT varied from 500-l 400 cGy/min.

RESULTS A total of 170 animals were entered on the 12 trials with minimal follow-up of 224 months (Fig. 1). One hundred forty-eight animals have actually received IORT, and the remaining 22 animals were control animals that received only sham IORT.

Control animals Eighteen control animals have been sacrificed, with four still alive. None of these animals have any clinical or pathologic evidence of tumor development.

Irradiated animuls One hundred two of 148 irradiated dogs have been sacrificed at ~24 months of follow-up at either the prescribed follow-up interval or when sacrifice was indicated for compassionate reasons. No tumor was found in any animal sacrificed at less than 24 months. Of the 46 animals followed for 224 months, 34 have been sacrificed and 12 remain alive with planned sacrifice to be done at 60 months of follow-up. The 12 living dogs have each been followed for a minimum of 30 months and none have clinical evidence of tumor development. In the remaining group of 34 irradiated and sacrificed dogs, 25 animals have had no evidence of tumor on postmortem examination. The remaining 9 animals were found to have tumors. The range of times at which the IORT-related tumors became evident was from 24 to 58 months, with a median time to diagnosis of tumor of 40 months. Since all sham portals and actual radiation portals were marked with stainless steel clips, determination of the relationship of site of tumor origin to the IORT portal was easily accomplished. One 7-year-old female dog (#3F92) had an incidental 1 cm well-differentiated infiltrating adenocarcinoma of a breast discovered post-mortem. This animal had received IORT to the bladder trigone 60 months previously. This breast tumor is a spontaneous malignancy, unrelated to the IORT. The remaining eight dogs with tumor had tumor arising from within the clipped IORT portal (Table 3). These IORT-related tumors included two benign neural masses, three soft tissue sarcomas, and four bone sarcomas. A benign neuromu arose from a segment of irradiated lumbosacral plexus (#F2690). The neuroma has been defined as a proliferative, non-neoplastic mass formed at the site of either traumatic transection or crush of a nerve

1. J.

654

Radiation Oncology 0 Biology0 Physics

September 1990, Volume 19, Number 3

Table 2. IORT radiation parameters Study no.

Target tissues irradiated

IORT field size/shape/ electron energy

Paravertebral soft tissues, aorta, vena cava, one ureter, lower pole one kidney

4 X I5 cm./Rectangle/ MeV

Paravertebral soft tissues, aorta, vena cava, blind end loop of jejunum

3.5 X 15 cm./Rectangle/l MeV

Abdominal aorta, vean cava, one ureter

3.5 X 15 cm./Rectangle/l MeV

Retropetitoneal soft tissues, blind end loop of jejunum 5

I

Doses delivered, cGy (no. of dogs treated)

11

Bone in Port

0 (3) 2000 (4), 3000 (4) 4000 (4) 5000 (4)

Yes

1

0 (l), 2000 (1) 3000 (1) 4500 (1)

Yes

1

0 (1) 2000 (2) 3000 (2) 4500 (2)

Yes

3.5 X 15 cm./Rectangle/ll MeV

0 (l), 2000 (2) 3000 (2) 4500 (2)

No

Extra-hepatic bile duct

5 cm. dia./Circle/l

1 MeV

0 (I), 2000 (2) 3000 (2) 4500 (2)

No

6

Extra-hepatic bile duct with anastomosis to jejunum

5 cm. dia./Circle/l

1 MeV

0 (I), 2000 (2) 3000 (2) 4500 (2)

No

7

Trigone of bladder (through cystotomy)

5 cm. dia./Circle/12

0 (3) 2000 (3) 2500 (3) 3000 (3) 3500 (3) 4000 (3)*

No

8

A. Upper lobe of rt. lung B. Mediastinal soft tissues (right atrium large vessels, phrenic nerve, bronchi)

5 cm. dia./Circle/9

0 (3) 2000 (6) 3000 (6) 4000 (6)+

No

9

Left bronchial stump, pulm. artery and vein esophagus, aorta, pericardium, segment of left atrium and ventricle

5 cm. dia./Circle/l3

0 (3) 2000 (4) 3000 (4) 4000 (4)

No

10

Esophagus

6 cm. dia./Circle/9

MeV

0 (1) 2000 (7) 3000 (5)

No

11

Lumbosacral nerve plexus (L4-S5)

9 cm. dia./Circle/l

1 MeV

0 (3) 2000 (4) 2500 (4), 3000 (3) 3500 (3), 4000 (4) 5000 (l), 5400 (2). 7000 (2)$

Yes

12

Lumbosacral nerve plexus L4-S5

9 cm. dia./Circle/9

MeV

1000 (4) 1500 (4), 2000 (4)

Yes

MeV

MeV

MeV

* Two additional animals from Study #7 died in the immediate post-operative period. + Three additional animals from Study #8 were inappropriately treated with steroids in the post-operative period, so were excluded from the final analysis. * Three additional animals from Study #l 1 died in the immediate post-operative period.

(12). The formation of the neuroma may reflect more the trauma of placement of the IORT cone and docking than an actual complication of ionizing radiation. A histologically benign but grossly invasive extradural neurojibroma arose, too, from a segment of irradiated lumbosacral plexus (#F269 1). Of the three soft tissue sarcomas, a grade ZZZjbrosarcoma arose in irradiated retroperitoneal soft tissues (#W4 16 1). Similarly, a grade IZZ malignant fibrous his-

tiocytoma arose from irradiated retropetitoneal connective tissue (#F3040). A sarcoma botryoides (embryonal rhabdomyosarcoma) developed in the bladder trigone of a dog

(#W2462) whose bladder base was irradiated through a cystotomy opening (Fig. 2). Finally, four bone sarcomas were found. One animal (#W4 162) had a tumor mass arising from lumbar vertebral bodies which were partially included in an IORT port. This tumor was composed of two distinct histologies, a

IORT-related tumors 0

I ND TUMOR

\

/

I

102 DEAD < 24 MONTHS

ND TUMOR

49 FOLLOWED ~24 MONTHS

34 SACRIFICED

NO TUMOR

\

/

25 ANIMALS

9 ANIMALS I

I

10 TUMORS

NO TUMOR

Fig. 1. Flow chart of 170 American Foxhounds entered into 12 different IORT-normal tissue toxicity studies at the National Cancer Institute. Note that only 148 dogs actually received IORT and the other 22 animals received sham IORT.

grade I chondrosarcoma

and a grade III osteosarcoma. staining techniques supported the diagnoses of two separate tumors rather than a high grade osteosarcoma with chondromatous differentiation. Two additional animals (#W4088 and #W204 1) developed grade III osteosarcomas arising in partially irradiated lumbar and sacral vertebral bodies. In 6 of the 9 animals that developed tumors, the tumors were associated with necrotic bone. The IORT ports in Studies #I, 2, 3, Ii, and 12 from Table 1 contained segments of bones. Vertebral bodies were in the fields in all Special

M.

BARNES

655

e/ a/.

of these five studies. Studies # 11 and 12 also had the upper aspect of one sacroiliac junction in the portals used to treat the lumbosacral plexus. A total of 61 animals received IORT on these five studies. All 61 dogs had bone from within the portal studied histologically. Occasionally marrow fibrosis, fibrous dysplasia adjacent to the endosteum, and depletion of the marrow elements were noted microscopically, but bone necrosis was not found in any specimen where tumor was not evident as well. Dog #F2691 had bone included in the IORT portal, but no evidence of bone necrosis was noted on histologic exam of the irradiated bone, despite the finding of the benign neurofibroma in the animal. The remaining two animals in which tumors were discovered, but no bone necrosis was diagnosed, were #3F92, the animal with breast cancer, and #W2462, the dog with embryonal rhabdomyosarcoma of the urinary bladder. Both of these animals were treated on the Bladder Protocol (Study #7, Table 1) in which no bone was included in the IORT portal. Each of the treatment protocols had regularly scheduled radiologic assessment to screen animals for toxicity, and are described in detail elsewhere (I, 10, 13, 14, 22, 3439,42). Diagnosis of the presence of tumor was made by early detection in screening studies or on investigative evaluation of new clinical symptoms. Figure 3 is typical of such an investigative study, a Tl weighted MRI evaluation of dog W4 162. The study reveals a tumor mass arising from a distal lumbar vertebral body. No attempt has been made to calculate specifically the dose absorbed at depth in tissues in any of these animals.

Table 3. Dog no.

Study protocol

(no.)

Tumor

histology

Dose (~GY)

Time (mo.) to diagnosis

Location tumor

of

Bone in portal

Associated bone necrosis

W4161

Peripheral Nerve I(1 1)

Fibrosarcoma Grade III

2500

58

In field

+

Yes

W4162

Peripheral Nerve I (I 1)

Chondrosarcoma Grade I, Osteosarcoma Grade III

2000

42

In field

+

Yes

F3040

Peripheral NerveI

Malignant fibrous histiocytoma

3500

39

In field

+

Yes

F2690

Peripheral NerveI(l1)

Neuroma

3000

24

In field

+

Yes

F269 1

Peripheral Nerve I (I 1)

Neurofibroma

3500

36

In field

+

No

W4088

Peripheral Nerve I (11)

Osteosarcoma Grade III

3000

30

In field

+

Yes

W2462

Bladder (7)

Embryonal Rhabdomyosarcoma

3000

48

In field

_

No

3F92

Bladder (7)

Adenocarcinoma breast

2500

60

Outside

_

No

W204 1

Retroperitoneum

3000

40

In field

+

Yes

(1)

Osteosarcoma

of

field

I. J. Radiation Oncology 0 Biology 0 Physics

September 1990, Volume 19, Number 3

ticular the fields containing bone, the assumption can be made that the underlying bone actually received a dose either equal to or within a 10% variation of the quoted IORT dose. The IORT dose range associated with tumor development was 2000 cGy to 3500 cGy with a median dose of 3000 cGy.

DISCUSSION Radiation-induced tumor development has long been recognized as a relatively rare complication of therapeutic and environmental radiation. Tumors of connective tissues, both soft tissues (2 1) and osseous ( 16,3 1), have been reported to be caused by radiation since early in this century. The use of a variety of animal species to study the

Fig. 2. Sarcoma botryoides (embryonal rhabdomyosarcoma) extending through fresh cystotomy performed for diagnosis and resection under genera1 anesthesia. The tumor arose from the trigone of the bladder which had received 3000 cGy of IORT 48 months earlier (#W2462).

Doses, in general, were quoted at the 90 or 100% isodose depth using electron energies of 9 through 13 MeV. Therefore, the quoted doses were generally delivered to tissues at 0.75 cm to 3.0 cm. In the fields treated, in par-

incidence and pathogenesis of radiation-induced tumors has proven to be a very effective research tool. Generally, because of the cost involved in housing and caring for large animals, most early animal carcinogenesis data had been collected using small mammals such as the mouse, rat, and rabbit. When a group of radioluminescent watch dial painters began to develop bone malignancies years after ingesting minute bits of the radium used to paint the dials, the need for refining research techniques and developing a more useful animal model became more urgent. Many elegant experiments were carried out on small laboratory animals in order to measure the carcinogenic risk not just to the Roentgen ray, but to radioisotopes such as fission products

Fig. 3. This T 1-weighted MRI image reveals a tumor mass arising from a distal lumbar vertebral body. This mass, a collision tumor composed of both low grade chondrosarcoma and high grade osteosarcoma arose from tissues which had received 2000 cGy IORT 42 months earlier (#W4 162).

IORT-related tumors 0 M. BARNESet al.

and heavy elements. The vast majority of the radioisotopes studied, regardless of means of entry in to the experimental animals’ systems, localized in the skeleton, ultimately producing bone cancers. Finkel and Biskis (5) in a summary of many such mouse experiments carried out at the Argonne National Laboratory, proposed a formula (A) using radium as the standard for toxicity in the mouse and as the common denominator for extrapolating information from mouse to man:

effect of x in mouse (A)

effect of 226Ra in mouse

ZI

effect of x in man effect of 226Ra in man

x = new radioisotope

This formula, while appealing in its simplicity, actually contributed little to the understanding of interspecies response to a particular radiation insult, that is. x. The metabolism of any one specific radioisotope, the physiology of the tissues involved, the specific dosimetry of each radioisotope in a given pathophysiologic situation, and the baseline predilection of any one species to developing certain tumors within those target tissues must factor into creating such a model. The primary experimental endpoint in most of the early mouse studies was the determination of simple tumor incidence, often using as the experimental end-point of tumor death in the study population over the study interval following administration of the radioisotope. The disadvantages of such a study design are several. but most markedly, the incidence curves are influenced by extraneous deaths in the population (5). Small laboratory animal colonies can be severely affected by infection and by other non-tumor death, and thus the size of the study population can be drastically reduced, especially when the latency period to time of tumor development can be several hundred days. Other drawbacks to such small animal studies are that much useful information about that latency period, the actual pathologic progression of toxicity and the radiologic changes, particularly when studying bone tumors, cannot be well appreciated. For many of these reasons, the need became evident for larger animal models to understand better not just the incidence but the dosimetry and the radiologic, histopathologic, and physiologic nature of radiation carcinogenesis. The beagle dog has been used now for several decades in the study of cancer development following the ingestion, inhalation, or other exposure to a variety of different radioisotopes such as radium, plutonium, radon gas, etc. The beagle was felt to be a good choice for these studies for several reasons: the species has a relatively low incidence of spontaneous osteogenic sarcoma, the primary tumor found in the human radium watch dial painters; the amount of radioisotope ingested by the dogs could be more readily monitored and quantified compared to the smaller animals; the pattern of distribution of the radium

657

throughout the dog skeleton was similar to that found in the humans; the average lifespan of the beagle extends to almost two decades, so tumors with long latency periods could be studied, and the interim pathologic changes in the high risk tissues could be studied radiographically as well as histologically and dosimetrically. We now have a vast collection of information on radiation carcinogenesis in the canine model. Regrettably, most of the beagle data deal with carcinogenesis following the administration of a radioisotope. The above discussion is intended to aid in developing a better understanding as to why such radioisotope information may have application in understanding the data presented here. The American Foxhound was chosen for our studies because the breed is without predilection for spontaneous tumor development, a serious problem with many larger dog breeds. We were concerned about losing animals to spontaneous tumor death, thus potentially losing valuable late toxicity data. Most importantly, we chose this breed over the smaller beagle because many of the more technically demanding surgeries such as biliary-jejunal anastomosis and aortic transection with anastomosis would be more easily accomplished, with fewer surgical complications expected in the larger animal. Also, placement of the IORT cones into the smaller pelvis or open urinary bladder of the beagle might have been technically more complicated, if not impossible. Finally, carcinogenesis was not a real consideration at the formulation of our studies; the data presented here are an unexpected result. Therefore, the necessity of having specific background data available dealing with the rate of radiation oncogenesis in the American Foxhound was not initially recognized. We are not aware of any such pre-existing data for the Foxhound, nor is there any data for this breed using IORT specifically as the carcinogen. have Powers et al. at the Colorado State University recently reported a cohort of beagle dogs which were studied for development of acute and late complications of IORT (24). Not only was the breed of dog different, but the experimental design used by Powers and her co-workers varied significantly from ours. The Colorado study documented the response of vertebral bone, that is, bone necrosis, to either electron beam IORT alone or IORT plus external photon beam irradiation (EBRT). In this study, the beagles were given 15 to 50 Gy IORT alone, 10 to 47.5 Gy IORT with 50 Gy EBRT, or 50 to 80 Gy EBRT alone. EBRT fractionation was 2 Gy/fraction. At 4 to 5 years following irradiation, eight of the dogs developed osteosarcomas in or around the irradiated vertebrae. Seven of the eight dogs developing tumors had 50 Gy of EBRT plus IORT ranging in dose from 25 Gy to 47.5 Gy. One of the eight tumor-bearing dogs had received 47.5 Gy of IORT only. These authors conclude from their data that IORT doses to vertebral bone of greater than 20 Gy should be used with caution when combined with EBRT doses of 50 Gy, as there appears to be significant risk of bone necrosis and a smaller risk of oncogenesis.

658

1. J. Radiation Oncology 0 Biology 0 Physics

Our data confirm that bone necrosis is found following IORT alone. However, we found bone necrosis only in animals in which tumor developed. This difference from the data of Powers et al. (24) can be explained by the small volumes of bone irradiated using IORT alone as in our studies versus larger, more homogenous volumes of irradiated bone as would be treated in an EBRT field. The size of the vertebral bone of the beagle, a breed almost half the size of the American Foxhound, would be significantly smaller as well. Thus, in our animals, with larger bone size receiving only IORT, more vertebral bone volume would have been spared from high doses of radiation compared to the Colorado beagles. Pool et al. studied the pathogenesis of radium-induced intracortical bone lesions in human radium dial workers and compared them to findings in a group of 38 radiuminjected beagles (23). The two species shared the following skeletal responses to radium: (a) dead bone tissue with delayed resolution, (b) a chronic disturbance in the remodeling mechanism of bone tissue, and (c) radiationinduced bone sarcomas. These authors coined the term “radiation osteodystrophy” to define a spectrum of proliferative and degenerative changes, starting with initial vascular injury and cell death, progressing eventually to malignant degeneration. While the physical technique of internal emitter irradiation in such radium studies is quite different than that used in our studies and in the Colorado study, the pathologic findings reported here and by Powers suggest a similar pathologic process occurring in the two IORT dog models. In the Colorado study, osteogenic sarcoma was the only tumor reported to have developed. In our animals, the predominant tumor was osteogenic sarcoma, occurring in 3 of the 9 animals that had IORT-related tumors. Three other sarcomas arose within IORT fields adjacent to portions of necrotic bone, one grade III fibrosarcoma, one grade III malignant fibrous histiocytoma, and one grade I chondrosarcoma. The concept of radiation osteodystrophy as proposed by Pool et al. might explain the development of the osteogenic as well as the non-osteogenic sarcomas. Pool et al. described in his radium-treated bone specimens a proliferative fibro-osseous response, or dysplasia, that contained nonosseous fibroblast-like cells with sarcomatous features (23). He could not determine whether the sarcomas that eventually developed arose by neoplastic transformation of such cells in these proliferative lesions of radiation osteodystrophy, but proposed that such might be the case. Atypical fibrous dysplasia, whether or not related to radiation, has been reported by other authors as well to be the site of origin of both osteogenic sarcoma and fibrosarcoma in humans (3, 30,4 1). Many of our irradiated specimens did reveal such atypical fibrous dysplasia, both within and adjacent to cortical bone. Also, each of the tumors in the spectrum of sarcomata found in our dogs has been reported by authors cataloging tumor development as a complication of therapeutic bone irradiation in man (9, 18, 30). Thus, the

September 1990, Volume 19, Number 3

finding of fibrosarcoma and other soft tissue sarcomas in addition to the usual osteogenic sarcoma might be expected in a large sample of animals studied following bone irradiation. The remaining tumors found in our IORT irradiated dogs were much less typical. Two lesions of neural origin were noted. One histologically appeared to be a benign traumatic neuroma (8). We have included it in this report because the lesion was particularly exuberant and associated with necrotic bone. The suggestion is that the trauma of the lucite IORT cone on the lumbar plexus caused the development of the mass, but perhaps the additional trauma from the electron irradiation caused an atypical proliferation to occur. The second neural lesion was an extradural neurofibroma. This lesion appeared entirely benign histologically. However, the tumor mass was a grossly aggressive tumor, encasing the surrounding neural and vascular structures. Laskin et al. have reported the development of two malignant Schwannoma following therapeutic irradiation in the human (17). The histology of the neural tumor arising in dog #F2691 was not that of a malignant Schwannoma, but we may be seeing an early phase of the spectrum of peripheral nerve oncogenesis (8). Our final IORT-related tumor was an embryonal rhabdomyosarcoma which arose grossly and histologically as a sarcoma botryoides within the bladder of an animal whose trigone had been irradiated through a surgical cystotomy. Rhabdomyosarcomas have rarely been reported as tumors caused by prior radiation in humans, particularly in children (9, 18). We know of no similar animal data. The critical issue when confronting data such as is presented here is its relevance to human cancer therapy. To date, there have been no IORT-related secondary tumors reported in the human literature. Several factors are to be considered as an explanation for this. First, and most importantly, IORT remains an experimental modality, used primarily in patients with advanced or recurrent tumors or tumors considered to be fairly resistant to standard radiation. Consequently, most patients receiving IORT do not survive the 5 to 10 years of latency that are necessary for the expression of radiation carcinogenesis in humans. Second, most centers outside of Japan have been involved in human IORT treatment for less than 10 years, so even early survivors may have as yet inadequate follow-up time to express tumor development. A few research centers working in human IORT therapy do have patients followed for a minimum of 5 to 10 years (7, 15). Even at these centers, the numbers of human patients at risk for secondary tumor development at this time remains very small. As the clinical application of IORT expands, the determination of the risk of carcinogenesis will become more critical. The scant data presented here and the data presented by Powers suggest that single fraction, high dose electron irradiation to bone, as well as to other connective

IORT-related tumors 0 M. BARNESet al.

tissues, has a relatively high carcinogenic potential in the dog. Using the Finkel and Biskis equation (A) (supra vide), it would suggest a very frightening possibility that a very large percentage of IORT-treated humans might progress on to develop aggressive secondary tumors. Following the work of Finkel and Biskis, more sophisticated, and we hope, more accurate attempts at interspecies scaling of such data to predict the risk of human carcinogenesis have been published. Raabe et al.. using beagles and radium as the carcinogen, have very elegantly analyzed radium-induced oncogenesis in beagles and compared that to data from humans and from mice (2527). Using log-normal dose response relationships for bone tumor deaths from 226Rafor the three species, Raabe and co-workers calculated a relative biologic sensitivity (RBS) for the imprecise human and mouse data relative to the dosimetrically precise dog data. The mathematical derivation of this factor is too cumbersome to reproduce here. These workers have better accomplished what Finkel and Biskis failed to do years earlier: that is, to create a mathematical concept, based on clinical and laboratory data, to compare radiation oncogenesis in two different species more directly. Raabe concludes from his data that humans have an RBS to 226Ra of 0.28, or humans are only 0.28 times as sensitive to radiation induced bone cancer from “‘Ra than dogs, and the RBS of humans compared to mice is 0.094. Therefore, humans are roughly only onetenth as sensitive to radium-induced bone cancer as mice. Finally, continuing the line of argument that the dog as a species has a higher carcinogenic response to irradiation, at least in bone, are the conclusions from Pool et al.

659

(23) also based on beagle data. Using the histopathology of radium-induced bone lesions, these authors conclude that the beagle “appears to have a greater capacity for maintenance of a bone vascular bed and to the greater viability and reactivity of bone mesenchyme in the dog skeleton in the face of continuing radionuclide toxicity” (23). Thus, the canine bone tissue, because of its greater viability, could maintain radiation-transformed cells long enough to allow for progression of clinically detectable tumors. The less viable human bone, conversely, had consistently lower populations of viable osteoblasts and osteoclasts and essentially no capacity to heal or provide a sustaining milieu for radiation transformed cells to progress to tumors as frequently. Can we use the data, arguments (no matter how eloquent), and conclusions from radionuclide studies to explain our findings, and more importantly, to aid to extrapolating our dog data to the human IORT clinical setting? At this time this question remains unanswered. Conclusions such as those of Raabe and colleagues (2527) that the human as a species is much less susceptible to radiation-induced sarcomas are optimistic and very soothing to the clinician engaged in IORT research when confronting our data and the Colorado data (24). Certainly the possibility exists that single fraction, high dose, and high dose-rate electron beam therapy may be far more carcinogenic in the human that other forms of low LET therapeutic radiation. Therefore, we suggest that the clinical delivery of IORT be carried out in well-controlled research settings, and that all acute and late toxicities be reported in a timely fashion in the radiotherapy literature.

REFERENCES 1. Barnes, M.: Pass, H.; DeLuca, A.; Tochner, Z.; Potter, D.; Terrill, R.; Sindelar, W.; Kinsella, T. Response of the mediastinal and thoracic viscera of the dog to intraoperative radiation therapy (IORT). Int. J. Radiat. Oncol. Biol. Phys. 13:371-378; 1987. 2. Caldwell, G.; Kelley, D.; Heath, C. Leukemia among participants in military maneuvers at a nuclear bomb test: a preliminary report. J. Am. Med. Assn. 244: 1575-1578; 1980. 3. Dehner, L. Fibro-osseous lesions of bone. In: Ackerman, L., Spjut, H., Abell, M., eds. Bones and joints. Baltimore, MD: Williams & Wilkins; 1976:209-235. 4. devilliers, A.; Windish, J.; de N. Brent, F.; Hollywood, B.; Walsh, C.; Fisher, J.; Parson, W. Mortality experience of the community and the fluorospar mining employees at St. Lawrence, Newfoundland. Occup. Health. Rev. 22: I- 15; 1971. 5. Finkel, M.; Biskis, B. Experimental induction of osteosarcomas. Prog. Exp. Tumor Res. 10:72- 11I ; 1968. 6. Folley, J.; Borges, W.; Yomawaki, T. Incidence ofleukemia in survivors of the atomic bomb in Hiroshima and Nagasaki, Japan. Am. J. Med. l3:3l l-321; 1952. 7. Gunderson, L.; Martin, J.; Beart, R.; Nagorney, D.; Fieck, J.; Wieand, H.; Martinez, A.; O’Connell, M.; Martenson, J.: Mclllrath, D. External beam and intraoperative electron irradiation for colorectal cancer. Ann. Surg. 20752-60; 1988.

8. Harkin, J.; Reed, R. Tumors of the peripheral nervous system, Fascicle 3, Second Series. Washington, DC: Armed Forces Institute of Pathology; 1968: 19-23. 9. Hatcher, C. The development of sarcoma in bone subjected to roentgen or radium irradiation. J. Bone Joint Surg. 27: 179-195; 1945. IO. Hutchinson, G. B. Carcinogenic effects of medical irradiation. In: Hiatt, H. H., Watson, J. D., Winsten, J. A., eds. Origins of human cancer, Book A. New York: Cold Spring Harbor Laboratory; 1977:501-509. 1 1. Jablon. S.; Kato, H. Studies of the mortality of A-bomb survivors. 5. Radiation dose and mortality. 1950- 1970. Radiat. Res. 50:649-698; 1972. 12. Kinsella, T.; Sindelar, W.; DeLuca, A. Threshold dose for peripheral nerve injury following intraoperative radiotherapy (IORT) in a large animal model (Abstr.). Int. J. Radiat. Oncol. Biol. Phys. lS(Supp1. 1):205; 1988. 13. Kinsella, T.; Sindelar, W.; DeLuca, A.; Barnes, M.; Tochner, Z.; Mixon, A.; Glatstein, E. Tolerance ofthe canine bladder to intraoperative radiation therapy: an experimental study. Int. J. Radiat. Oncol. Biol. Phys. 14:939-946; 1988. 14. Kinsella, T.; Sindelar, W.; DeLuca, A.; Pezeshpour, G.; Smith, R.; Maher, M.; Terrill, R.; Miller, R.; Mixon, A.; Harwell, J.; Rosenberg, S.; Glatstein, E. Tolerance of peripheral nerve to intraoperative radiotherapy (IORT): clinical and experimental studies. Int. J. Radiat. Oncol. Biol. Phys. I l:l579-1585; 1985.

660

I. J. Radiation Oncology 0 Biology 0 Physics

15. Kinsella, T.; Sindelar, W.; Lack, E.; Glatstein, E.; Rosenberg, S. Preliminary results of a randomized study of adjuvant radiation therapy in resectable adult retroperitoneal soft tissue sarcomas. J. Clin. Oncol. 6: 18-25; 1988. 16. Lacassagne, A.; Vinzent, R. Action des rayons x sur un foyer infectieux local, provoque chez le lapin par l’injection de streptobacillus caviae. Compt. Rend. Spc. de Biol. 100:247249; 1929. 17. Laskin, W.; Silverman, T.; Enzinger, F. Postradiation soft tissue sarcoma. Cancer 62:2330-2340; 1988. 18. Li, F. Second malignant tumors after cancer in childhood. Cancer 40: 1899-1902; 1977. 19. Lyon, J.; Klauber, M.; Gardner, J.; Udall, K. Childhood leukemias associated with fallout from nuclear testing. N. Eng. J. Med. 300:397-402; 1979. 20. Major, I.; Mole, R. Myeloid leukemia in x-ray irradiated CBA mice. Nature 272:455-456; 1978. 21. Marie, P.; Clunet, J.; Raulot-Lapointe, G. Contribution a l’etude du developpement des tumeurs malignes sur les ulceres de Roentgen. Bull. Assoc. Francaise p L’Etude du Cancer 3:404-426; 19 10. 22. Pass, H.; Sindelar, W.; Kinsella, T.; DeLuca, A.; Barnes, M.; Kurtzman, S.; Hoekstra, H.; Tochner, Z.; Roth, J.; Glatstein, E. Delivery of intraoperative radiation therapy after pneumonectomy: experimental observations and early clinical results. Ann. Thorac. Surg. 44: 14-20; 1987. 23. Pool, R.; Morgan, J.; Parks, N.; Farnham, J.; Littman, M. Comparative pathogenesis of radium-induced intracortical bone lesions in humans and beagles. Health Phys. 44(Suppl. 1):155-177; 1983. 24. Powers, B. E.; Gilette, E. L.; McChesney, S. L.; Withrow. S. J.; LeCouteur, R. A. Bone necrosis and tumor induction following experimental intraoperative irradiation. Int. J. Radiat. Oncol. Biol. Phys. 17559-567; 1989. 25. Raabe. 0.; Book, S.; Parks, N. Bone cancer from radium: canine dose response explains data for mice and humans. Science 208:6 l-64; 1980. 26. Raabe, 0.; Book, S.; Parks, N. Lifetime bone cancer doseresponse relationships in beagles and people from skeletal burdens of 226Ra and 90Sr. Health Phys. 44(Suppl. 1):3348; 1983. 27. Raabe, 0.; Parks, N.; Book, S. Dose-response relationships for bone tumors in beagles exposed to 226Ra and 90Sr. Health Phys. 40:863-880; 198 1. 28. Ron, E.; Modan, B. Benign and malignant thyroid neoplasms after childhood irradiation for tinea capitis. J. Natl. Cant. Inst. 65:7-l 1; 1980. rela29. Rowland, R.; Stehney, A.; Lucas, H. Dose-response

September 1990, Volume 19, Number 3

30. 31.

32.

33.

34.

35.

36.

37.

38. 39.

40. 41. 42.

43.

tionships for female radium dial workers. Radiat. Res. 76: 368-383; 1978. Sabanas, A.; Dahlin, D.; Childs, D.; Ivins, J. Postradiation sarcoma of bone. Cancer 9528-542; 1956. Sabin, F.; Doan, C.; Forkner, C. The production of osteogenic sarcomata and the effects on lymph nodes and bone marrow of intravenous injections of radium chloride and mesothorium in rabbits. J. Exp. Med. 56:267-289; 1932. Schwarz, M.; Burgess, P.; Fee, W.; Donaldson, S. Postirradiation sarcoma in retinoblastoma. Arch. Otol. Head Neck Surg. 114:640-644; 1988. Shellabarger, C.; Chmelevsky, D.; Kellener, A. Induction of mammary neoplasms in the Sprague-Dawley rat by 430 KeV neutrons and x-rays. J. Natl. Cant. Inst. 64:82 l-833: 1980. Sindelar, W.; Hoekstra, H.; Kinsella, T.; Barnes, M.; DeLuca, A.; Tochner, Z.; Pass, H.; Kranda, K.; Terrill, R. Response of canine esophagus to intraoperative electron beam radiotherapy. Int. J. Radiat. Oncol. Biol. Phys. 15: 663-669; 1988. Sindelar, W.; Kinsella, T.; Barnes, M.; DeLuca, A.; Matthews, D.; Anderson, W. Late effects of intraoperative radiation therapy on retroperitoneal tissues, intestine, and bile duct. Int. J. Radiat. Oncol. Biol. Phys. (In press); 1990. Sindelar, W.; Kinsella. T.; Tepper, J.; Travis, E.; Rosenberg, S.; Glatstein, E. Experimental and clinical studies with intraoperative radiotherapy. Surg. Gyn. Obstet. 157:205-2 19; 1983. Sindelar, W.; Morrow. B.: Travis, B.; Tepper, J.: Merkel, A.: Kranda, K.: Terrill. R. Effects of intraoperative electron irradiation in the dog on cell turnover in intact and surgically anastomosed aorta and intestine. Int. J. Radiat. Oncol. Biol. Phys. 9:523-532; 1983. Sindelar. W.; Tepper, J.; Travis, E. Tolerance of bile duct to intraoperative irradiation. Surgery 92:533-540; 1982. Sindelar, W.; Teppar, J.; Travis, E.; Terrill. R. Tolerance of retroperitoneal structures to intraoperative radiation. Ann. Surg. 196:60 I-608; 1982. Smith, P.; Doll, R. Mortality from cancer and all causes among British radiologists. Br. J. Radio]. 54: 187- 194; I98 1. Taconis, W. Osteosarcoma in fibrous dysplasia. Skeletal Radiol. 17: 163- 170; 1988. Tepper, J.; Sindelar, W.; Travis, E.; Terrill, R.; Padikal, T. Tolerance of canine anastomoses to intraoperative radiation therapy. Int. J. Radiat. Oncol. Biol. Phys. 9:987-992; 1983. Upton, A.; Randolph, M.; Conklin, J. Late effects of fast neutrons and gamma-rays in mice as influenced by dose rate of irradiation: induction of neoplasia. Radiat. Res. 4 I : 467-491; 1970.