1- I anhdd (CW1214): 283-289. 1969.

Evaluation of Neutron Irradiation of Pancreatic Cancer

Results of a Randomized Radiation Therapy Oncology Group Clinical Trial

Frank J. Thomas,M.D., John Krall, Ph.D., Frank Hendrickson, M.D., Thomas w.Griffin, M.D., Jerrold P. Saxton, M.D., -RobertG. Parker, M.D., and Lawrence W. Davis, M.D.

Baween 1980-84, the Radiation Therapy Oncology Group Carcinoma of the pancreas annually accounts for conducted a trial in patients with untreated, unrrsectable 10- over 24,000 deaths per year in the United States, where carcinomas of the pancreas. Patients were randomly alized it is the leading cause of cancer deaths (I). Only chosen to rcaive either 6,400 cGy with photons, the cquiv- fifth dent dose with a combination of photons and neutrons about 10-1546 of patients are candidates for radical (mixed-beam irradiation), or neutrons alone. A total of 49 excision (2), and even among these patients, the 5-year cases were evaluable, of which 23 wen treated with photons, crude survival rate is 1546, ‘with local recurrences in up 1 I with mixed-beam therapy, and 15 with neutrons alone. to 50% (3). Overall, the prognosis is poor, with only The median survival time was 5.6 months with neutrons, 7.8 months with mixed-beam radiation, and 8.3 months with 3-5% surviving 5 years (I). photons. The median local control time was 6.7 months with For the casc that is medically or technically inop neutrons, 6.5 months with mixed-beam radiation, and 2.6 erable, radiation can palliate the local symptoms of months with photons. These differences are not statistically advanced disease with relatively low doses of 3,000- significant. Evidence of moderate-to-life-thrnin~ gas- 4,000 cGy (4). With higher doses of 5,000-6,500 cGy trointestinal or hepatic injury was present in three patients (with or without chemotherapy), the symptom-free in- treated with neutrons and one patient treated with photons. The causes of this apparent difference are discussed. This terval is extended (3,and the median survival time is study demonstrates there is no evidence to suggest that neu- typically 7-13 months (1.5-8). At the time of death, tron irradiation, either alone or in combination with photon locally persistent disease is frequently Seen (9,lO). Thus, irradiation, produces better local control or survival rates improvement in the local control rate should improve than photon irradiation. the overall survival rate in patients with this disease. Key Words: Mixed-beam radiation-Neutron therapy- Pancreatic cancer-Photon therapy. As an alternative to increasing the physical dose, high linear energy transfer (LET) radiation attempts to deliver a biologically more effective dose and thus OOlb303 improve the therapeutic ratio. The potential advantage From the Department of Radiation Oncology, Cleveland Clinic of neutron therapy is a reflection of the biological con- Foundation, Cleveland, Ohio (FJ.T.,J.P.S.); the Dividon of Statistics. sequences of high LET in tissue (1 1). The dense ion- Radiation Therapy Oncology Group. Philadelphia Pennsylvania ’K);the Neutron Therapy Department Femi National Acalerator ization produced by neutrons leads to more effective bratory. Batavia, Illinois (F.H.); the Department of Radiation killing of cells protected by virtue of hypoxia. The ox- .icology, University of Washington Hospital. Seattle. Washington ygen enhancement ratio with neutrons is ap (T.W.G.);the Department of Radiation Oncology. University of Cal- (OER) ifornia-los Angela, ksAngeles. California (R.G.P.):and the De- proximately I .6, compared to an OER of 2.5-3.0 with partment of Radiation Therapy. Albert Einstein Medical College, high-energy photons. The repair of both sublethal and Montcfiore Medical Center, Bronx. New York (L.W.D.). potentially lethal damage is diminished with high Address correspondence and reprint requests to Dr. F. J. Thomas. LET Case Western University School of Medicine. 2074 Abington Road. radiation as compared to photon radiation. Finally, Cleveland, OH 44 106. U.S.A. there is less variation in radiation wmitivitv as a fiinc- 284 F. 1. THOMAS ET AL.

tion of cell cycle position with neutrons. In clinical Treatment plans were designed to deliver doses t practice, these potential advantages of neutron ther- equivalent to 6,400 cGy at 180-200 cGy per fraction c€ apy have been demonstrated in the treatment of ad- to the tumor volume in the photon or mixed-beam D enocarcinomas of the prestate (12.13) and salivary arms but not to exceed 5,000 &y to the bowel outside ai slaad (14). the target volume, 2,000 d3y to the kidneys, 4,000 This study (Radiation Therapy Oncology Gfoup, or cGyto the spinal cord, or 3,500 cGy to the liver. When P; RTOG, 79-21) was designed to detedne whether patients were randomly assigned to dveneutrons r; neutron radiation alone or combined with conven- alone, the Same equivalent doses were employed, but I< tional photon radiation (mixed-beam radiation) was the treatment was delivered in IS fkidons over 7 ii superior to conventional photon irradiation in the weeks. The target volume was specified as coveringthe n treatment of lacally advanced pancreatic adenocarci- gross disease with a 2cm margin. The recommended n nomas. left lateral margin was 3 cm to encompass a greater ! H volume of the pancreas, as pancreatic cancers may i tl spread along the pancreas in an occult After f; MATERIAIS AND METHODS fashion. I the equivalent of 4,500 cGy, the fields could be reduced f! Forty-nine patients were entered in this study be- to a margin of 1-2 cm around the gross disease. i' C tween February 1980 and October 1984. Mandatory i premndomization evaluation included a complete his- patient i ta TABLE 1. characteristics ! tory and physical examination, laparotomy with biopsy ! L and marking of the gross tumor margins and biliary Therapy. no. of Pa- ! t r or gastrointestinal bypass procedures when indicated, Photons Neutrorrr Mixedbeam chest roentgenogram, upper gastrointestinal series, i.v. '! Total no. ocpatiec~ts 23 15 11 pyelogram, computed tomography scan, complete I blood count, and blood chemistry studies including Sex i male 14 6 5 i carcinoembryonic antigen, alkaline phosphatase, blood female 9 9 6 urea nitrogen, and blood glucose level determinations. 4 Histology I To be eligible for randomization, patients had to be 23 14 9 I ambulatory and, although locally advanced, tumors isbetcel- 0 1 2 I had to be limited to a target volume no greater than Mececltiation 1,700 well 5 4 2 cm'. Patients were excluded from this study if I they had a prior malignancy except skin cancer, unless moderate 11 6 5 poor 4 3 4 they had been disease free for 5 years; prior chemo- unknown 3 2 0 therapy or radiation therapy; chronic renal disease; Location overt metastases; or acute intercurrent postoperative head 16 10 4 complications that would preclude radiation. body 1 1 1 The study contained three treatment arms (photons, tail 0 1 0 body and tad 1 1 1 neutrons, and mixed-beam irradiation). Facilities par- -andbody 5 1 5 ticipating in this study (Fermilab, the Cleveland Clinic diffuse 0 1 0 Foundation, and the University of Washington) could Kamofsky value -70 4 1 3 elect to randomly assign patients to either two or three 80-90 14 14 6 of the treatment arms but were required to participate 100 5 0 2 in the photon control arm. Fermilab chose neutrons Tunor size, cm2 alone, while the Cleveland Clinic and the University 1-40 15 10 8 of Washington chose mixed beam. Patients were ran- 41-80 6 3 3 80-160 1 1 0 domly assigned through the RTOG operational ofice unknown 1 1 0 to receive either photon (control) radiation therapy, neutrons, or mixed-beam radiation therapy. Neutron 0 0 1 doses were scaled according to the relative biologic ef- 7 4 2 fectiveness calculations for the various institu- 4 3 4 (RBE) 12 8 4 tions: 3.3 for the University of Washington and the Cleveland Clinic Foundation and 3.0 for Fennilab. The 8 9 6 i gamma contaminate was included in the neutron dose. 9 4 3 Mixed-beam irradiation is a mixture of 40% neutrons 2 0 1 and 60% photons. 4 2 1

Am J <%n Onrvl (CC'l). t'id 12. No 4. IY80 001b3011 i NEUTRON IRRADIATION OF PANCREATIC CANCER 285

The quality of the radiotheraw was ensured by a antral review of localization films and munent plans Data were reviewed at a central office by data managers The majox end points of this study are local/eonal and the study chairman. control, survival, and &-free survival rates. Com- The efficacy of the treatment was evaluated by com- plication rates and tolerance of the inadiated tissue are paring the duration of local control, the overall survival secondary end points. rate, and the disease-free survival rate. In each instance, Table 1 shows that, of the major patient character- Kaplan-Meier plots were generated and compared us- istics, there were no significant differences in the dis- ing the Mantel-Haenszel test. The survival rate was tribution of pretreatment factors among the three measured as absolute survival from the start of treat- treatment groups. ment. For the evaluation of local control. patients in Figure 1 shows the fractional survival rate by treat- whom there was a persistence of disease at the time of ment arm as a function of time. The median survival tbe fix% follow-up visit were considered to have local duration of the entire group (n = 49) is 7.1 months, failures from day 1. For the calculation of the dise;lsc- with no patient alive at 2 years. The median survival free survival rate, death was not considered failure, but duration by treatment arm is 5.6 months with neutrons cases were censored at the time of death. (n = 15); 7.8 months with mixed-beam radiation (n The acute reactions during treatment and the long- = 1 l), 2nd 8.3 months with photons (n = 23). A Man- term serious complications were recorded prospectively tel-Haensz4 test was performed to compare these thing the RTOG acute morbidity scoring criteria and curves; th*p value is 0.1. the RTOG late radiation morbidity scoring scheme, The median duration of local control of all patients respectively. in this study is 4.3 months (n = 49). The local control

100

75

I

S U R FIG. 1. RTOG 7921 Survival rate by V 50 treatment (n = 49). Photons (-), n I = 23. Mixed beam (- - -), n = 1 1. Neb Y trons (- - -), n = 15. P = 0.10. A I

15

0 001b305 I I I r 0 6 12 18 24

MONTHS FROM START Of TREATMENT. - ._ 286 F. J. THOMAS ET AL.

rate is shown in Rg. 2. Persistent disease at the time Four patients (one photon, two mixed beam, and one of the first follow-up eoaluation was more common neutron) had severe ~lilll~ea,and vomiting; in the neu- among patients treated with photons (43%) than with tron patien4 this was accompanied by severe diarrhea either neutrons ( 13%) or mixed-beam radiation ( 18%), One patient in the photon arm had Sfe-threatening but the three curves converge by about7'months. This nausea and vomiting. observation is reflected in the median duration of local The frequency of late complications of treatment is control: 6.7 months with neutrons (n = IS), 6.5 months difficult to evaluate given the few patients at risk beyond with mixed-beam therapy (n = ll), and 2.6 months 1 year. Nevertheless, it does appear that the risk was i with photons (n = 23). However, a comparison of these higher in patients treated with neutrons (Table 3). One i curves employing a Mantel-Haensztl test revealed no patient treated with neutrons experienced life-threat- significant difference among the treatment arms. ening liver and small bowel complications. Another 1, The disease-free survival rate is shown in Fs. 3. For patient experienced moderate stomach, small bowel, I the entire group (n = 49). the median disease-free sur- and large bowel complications A third patient expe- i vival time is 3.7 months. Using the Mantel-Haenszel rienced moderate Liver and severe stomach reactions. f test, $ere are no significant differences among the Only one patient treated with photons experienced a i treatment anns, with disease-fiee survival times of 3.7, moderate small bowel don,and none had a severe i 3.4, and 3.7 months for patients treated with neutron, late effect. nus, the frequency of moderate or worse I mixed-beam, and photon therapy, rtspeCtively. reactions was 4% in the photon-treated patients and The short-term reactions to therapy were similar in 20% in the neutron-treated patients. No patient treated i all arms of the study (Table 2). Minor reactions were '4th mixed-beam radiation had worse than a minor somewhat more cOmmon in the experimental arins. late reaction.

FIG. 2. RTOG 7921 Local control rate by treatment (n = 49). Gmptotals same as m Figure 1. P = N.S.

0 12 18 11

001b30b MONTHS FROM STAR1 OF TREATYEMI NEUTRON IRRADIATION OF PANCREATIC CANCER 28 7 .

ffi. 3. RTOG 7921 Dise-free sur- vival rate by treatment (n = 49). Group totals same as in Fgure 1. P = N.S.

0 6 12 18 2+

MONTHS FROU START OF TREATMENT

TABLE 2. Acute reactions by grade DISCUSSION

Grade.. no. Of The treatment of patients with locally advanced patients pancreatic Cancer continues to present a significant Total no. ol Reaction 123 4 5 pStientsW) challenge. Without treatment (1 5) or with only con- servative surgery (3), the median survival time is less Photohrapy (n = 23) 60000 6 (26) than 4 months. GP (nausealvomting) 3 5 1 1 0, lO(43) GI (dianhea) 7 4 0 0 0 ll(48) Dobelbower and colleagues (5,6,16) have reported hematobgii 20000 2 (9) a considerable improvement in survival with highdose Mixed-beam therapy radiation therapy. The 1-year survival rate was 49% (n = 11) among patients receiving 5,900-7,OOO cGy in 7-9 skh 40000 4 (36) weeks with or without adjuvant chemotherapy. Similar GI (nausea/vOmiting) 4 4 2 0 0 lO(91) GI (diarrha) 33000 6 (55) results were reporled by Komaki et al. (10). although hematologic 00000 0 (0) in other series with highdose radiation, the median -m*aPY survival time is closer to 7.0 months (2,8,17). Both (n = 15) adjuvant chemotherapy and techniques to increase the skin 8 4 0 0 0 12(80) radiation dose to the tumor have been employed to GI (nausealvomiting) 14100 6 (40) GI (dw) 7 2 1 0 0 lO(67) further improve the results of therapy. The addition of hematologic 10000 1 (8) chemotherapy improved the median survival time in a Gastrointestinal Tumor Study Group trial (9). Pa- ' 1. minor; 2, moderate; 3, severe; 4, life threatening; end 5. death. tients randomly chosen to receive 6.000 cGy had a 'GI. gastrointestinal. 001b301 median survival time of23 weeks. With 6,000 cGy and 288 F. J. THOMAS ET AL.

TABLE 3. Comp/karions by grade without chemotherapy. Among the patients treated with neutron therapy, 40% were alive at 1 year. By 'Qnbded. no. of comparison, historical con&ols treated with photons patbnts Totd no. of qiBAu by implant had 1-year survival rates of 23 and Rsactlon 1 2 3 4 5..(%) 32%, mpectively. The Mid-Atlantic Neutron Restarch Center (24) employed 1,700-1,75OcGy neutrons (equivalent to approximately 5,200 CGy of photons) with or without fluorouracil. The median survival time of these patients was 6 months, with 47% (nine of 19) having failure within the treatment field. Cohen, Kaul, and their colleagues (25,26) also reported a 6-month median survival time in 77 patients treated with neu- tron doses of 1,500-2,500 cGy. In a small series of patients treated with mixed-beam radiation to a 6,000- cGy equivalent and sucptomtocin, fluorouracil, and mitomycin< at the Cleveland Clinic (27), the median survival time was IO months. In none of the prior stud- .. (n = 15) ies of neutron therapy for pancrtatic cancer were con- StaMCh 02100 3 (20) current controls employed. ova 11010 3 (a) This study designed to determine if neutron ra- SmaaboWeI 31010 5 (3) was lase- 12000 3 (20) diation alone or as mixed-beam radiation was superior to conventional radiation in the treatment of unre- '1. 2. moderate; 3. severs; 4. life threatening; and 5. sectable pancreatic carcinomas. death.- The acute toxicity encountered in the experimental arms of this study was similar to that observed with fluorouracil or 4,000 cGy and fluorouracil, the median photons. The moderate or worse late complications survival time inmased to 40 and 42 weeks, respec- were more frequent and more severe in the neutron tively. To further incrcase the radiation dose, both in- arm as compared with either the photon or mixed- traoperative radiation therapy (IORT) and '=I implants beam arms. Among patients treated with neutrons, the have been employed in selected cases with or without frequency of severe and life-threatening reactions was adjuvant therapy. With IORT, the median survival 13%; 20% of the patients experienced a moderate or time is in the range of 10-16.5 months (7,18-20) but worse reaction. A variety of factors may have contnb has not been compared with that of conventional ther- uted to this effect. In comparison with high-energy apy in a randomized trial. With '*'I implants, median photon beams, the neutron beams have a wider pen- survival times of 1 1-14 months have been reported umbra and poorer depth dose characteristics. Some (2 1,22). patients who received neutrons were treated in a stand- Despite highdose radiation, locally penistent disease ing position. It may have allowed a greater volume of is common. Komaki et al. (IO) noted that 80% of the the stomach and intestine to fall into the treatment patients with progressive disease had local failure with volume than might have occurred with the conven- or without distant disease. Even with doses > 5,000 tional supine treatment position. cGy by I9'Au implant or IORT,Rich (7) reported 77% The median survival time and local control rate in symptomatic local failure. The addition of chemo- this series is similar to those reported in other series therapy does little to alter this pattern of failure (9). using photons (7,s. I7,28), neutrons (23-27). or heavy Given this high local failure rate, it is reasonable to charged particles (29). As judged by the local control, assume that improvements in local control will impact overall survival, or disease-free survival rates, no sta- on the survival rate. tistically significant difference could be appreciated h The use of neutrons in the treatment of pancreatic among the treatment arms. The further escalation of u carcinomas is predicated on the hope that the biological the neutron dose or, in the case of mixed-beam radia- -.. advantages of neutrons would improve the local control tion, increasing the proportion of the dose delivered a- and survival rates. In a retrospective study, both neu- with neutrons does not seem warranted based on the W trons alone and mixed-beam radiation have been em- complication rate experienced in the neutron arm. Q ployed in the treatment of pancreatic cancer. In an Acknowledgment: This work was supported in part by the 03 early study, AI-Abdulla et al. (23) employed mixed- National Cancer Institute under grants CA-23 1 13. CA-77 133. beam doses equivalent to 4,920-6,040 cGy with or CA- I244 I. CA- 1 7465. and CA- 1808 I.

- . .--- .. 9 .. .. , .nnn NEUTRON IRRADIATION OF PANCREATIC CANCER 289

REFERENCES 16. Borgelt BB, Dobelbowcr RR, StruMa KA. Betatron th~~for unrrseaable pa~crutkanar. Am JSug 1978;135:7680. 1. American Canccr Society. Gmerfms andltgurcs-I 987. Nm Ycuk American clnm soday, 1987:8. 17. HIIslam J& CaMlnaugh PJ, Straup SL Rdiotioo tberpw io the 2 Hamann RE, Coopmnan AV. Cancer of the panmas. Nal matmcnt 0fimSsaaMe adenocarcinoma oftbe paacrran Cancer J Mcd 197930 I :482-5. 1973;32: 134 1-5. 3. Teppa E, Nardi GISuit H. Carcinoma of the panmas. IS. Guodmon 4 Martin JK. Kvok WC. d rl. Intraoperative and dMGH cxpaiencC from 1963 to 1973. 1976;37:1519-- external beam irradiation 2 5-FLJ for locally advanad -tic Cancer 24. caom. Int J Radiat Oncd Bid Phys 1987;13:3 19-29. 4. GFscn N. Beron E, Melbyc RW, Gcorgc W.Corcioomr of the 19. Tcppr JE, Shipley WU, Nan5 GL, Wood WC. Orlow EL The pmcrtas-palliative radiotherapy. Am I Roenfgd 1973;Ill: deof misonidazole combined with intraoperative radiation 62&2. therapy in the treatment of pancreatic carcinoma J Clin Oncd 5. Doklbomr RR. The radiotherapy of pancreatic canar. Semin 1987;5579-84. Oncd 1979;6:378-89. 20. Shipley WU. Wood WC, Teppcr JE, et al. Intraoperative dcctron 6. Dobtlbomr RR Borgclt BB, Strubla KA. Kutcher GJ, Sun- beam inadiation for patients with uorcscctablc pancreatic car- thvzlingham N. Wonmdiothcrapy for marof the pan- cinoma. Ann Surg 1984;200:289-96. Q~PTtechnique pad results IN J UiatOncd Bid Phys 198&6: 21. Shipky WU, Nudi GL, Cohen AM, Ling m. Iodiae- 125 imphot 1127-33. and uttcmal bearn irradiation in patients with localized #c 7. Rich TA. -ti00 thmpy for pancreatic canar dewn-year carcinoma Cancer I980;45:7O9- 14. cxpcricoa at the JCRT. IN J Radior OndBid Phys 1985; I I: 22. Syed AMN. Puthnwala M, Ncblm DL Intedtial iodine-I25 75943. implant in the management of unrescctablc panmatic carci- 8. Sdch MT, Parka RG. Results in the management of locally mma. Cancer 1983;52808-13. unmcaabk pancreatic carcinoma Am J Clin Oncd 19869: 23. AI-AWulla ASM, Hussey DH, Olson MH,Wright AE. Expri- 139-45. ena with fast neutron therapy for unrCSeFtabe carcioomr of 9. Ganrointestinal Tumor Study Group. Radiation therapy com- the pancreas. Int J Radiat Biol Phys 1981;7:165-72. bined with Adriamycin or 5-fluorouracil for the treatment of 24. Smith FP, Schein PS, MacDonald JS, Woolcy PV, Ornitz R, loally unrescctaMe panmatic carcinoma Cancer l985;56:2563- Rogrrs C. Fast neutron irradiation for locally advanad pancreatic 8. canar. In! J Rad& Oncd Biol Phys 198 1;7: 1527-3 I. ' I 0. Komaki Wilson JF, Cox JD, Kline RW. Carcinoma of the R, 25. Cohen Woodruff KH, Hcndrickson et al. Raponsc of pmcrras: of irradiation for unrsectable lesions. In! JRa- I., FR, results canar to h@ Bid Phys 1980;6:2O9- I 2. pancreatic local inadiition with energy oeutrons. diat oncd CMCer 1985;56:123541. Hall EJ. Radiobidogv for the radidogiis. Hagc~~~wr~,Maryland: H- & ROW, 1978: 295-303. 26. Kaul R, Cohen L, Hcndrickson F, Amhalorn M. Hrcjsa AF. 12 LPramorc GE, Krill JM, Thomas FJ, Griffin TW,Maor MH, Roscnbcrg I. Pancreatic carcinoma: results with fast neutron Hendrickson FR. Fas~neutron radiotherapy for locally advanad therapy. 1nt JMiat Oncol Bid Phys 1981;7:173-8. prostate mar mlts of an RTOG randomized study. 1n1 J 27. Bukowski RM,Gahbauer R, Rodriguez-Antuna A, Hcrmann Mia OncdBiol Phys 1985;11:1621-7. R. Mixed beam radiotherapy and combination chemothuapy 13. Russell KI, Lammore GE, Krall JM, et al. Eight-year expcriena in localized panmatic adenocarcinoma-pdiminary results. Int with neutron radiotherapy in the treatment of stages C and D J Radiat Oncd Biol Phys 1982;8:1231-3. prostate canar updated results of the RTOG 7704 randomized 28. Moertel CG. FwkS, Hahn RG. et al. Therapy of loollly un- clinical trial. Prosrare 1987;11:183-93. resectable pancreatic carcinoma: a randomized comparison of 14. Laramore GE. Fast neutron therapy for inoperable salivary gland high dose (6OOO rad) radiation alone, moderate dmradiation tumors: is it the treatment of choice? Int J Radiat Oncol Biol (4ooo rad + 5-fluorouracil). and high dose radiation + 54110- Phys 1987;13:142 1-3. rouracil. Cancer 1981;48:1705-10. 15. Moertel CG, Reitemcier RJ. Advanced gastrointestinal canar. 29. Castro JR, Saundm WM, Quivey JM. et al. Clinical problems In: Clinical management andchemorherapy. New York: Harper in radiotherapy of carcinoma of the pancreas. Am J CIin Oncol and Row, 1969:3-14. 1982;5:579-87.