Split-course versus continuous pelvis irradiation in carcinoma of the uterine cervix: A prospective randomized clinical trial of the radiation therapy oncology group

Inr. 1. Radiation Oncology Biol. Phys.. Vol. Printed in the U.S.A. All rights rescrvcd.

036lk3016/83/040431Alf,$03.00/0 Copyright 0 1983 Pcrgamon Press Ltd.

9. pp. 431436

??Original Contribution SPLIT-COURSE VERSUS CONTINUOUS PELVIS IRRADIATION IN CARCINOMA OF THE UTERINE CERVIX: A PROSPECTIVE RANDOMIZED CLINICAL TRIAL OF THE RADIATION THERAPY ONCOLOGY GROUP VICTOR Professor

and Chairman,

Radiation

A.

MARCIAL,

M.D.

Oncology Division, University San Juan, PR

DAVID A. AMATO, Division of Biostatistics

and Epidemiology,

Professor

of Radiology,

Radiation

PH.D.

Sidney Farber Cancer

RICHARD Therapy

D.

of Puerto Rico School of Medicine,

MARKS,

Institute,

44 Binney St., Boston, MA 02115

M.D.

Department, Medical University SC 29403

of South Carolina,

Charleston,

MARVIN ROTMAN, M.D. Professor

and Chairman,

Radiation

Oncology

Department, S.U.N.Y. NY 11203

Ellis Fischel Cancer

Instructor

in Radiation

Oncology,

F.

Radiation

JAMES Division of Biostatistics

Hospital,

Columbia,

FIGUEROA-VALLES,

Brooklyn,

MI

M.D.

Oncology Division, University San Juan, PR 00935 A.

Medical Center,

CANOY, M.D.

NESTOR

NAYDA

Downstate

HANLEY,

and Epidemiology,

of Puerto Rico School of Medicine,

PH.D.* Sidney Farber Cancer Institute

In August 1980, the Radiation Therapy Oncology Group (RTOG) completed a prospective randomized clinical trial for the comparison of a split-course versus a standard continuous course of pelvic irradiation for carcinoma of the uterine cervix Stages II-B, III-A, III-B, and IV-A. The split-course consisted of 10 fractions of 250 rad each, 5 times a week, up to 2500 rad followed by a rest period of approximately 2 weeks and then another 2500 rad was given (250 x IO).The continuous course consisted of 30 fractions of 170 rad each, 5 times per week, for a total of 5100 rad. In both groups the external pelvis irradiation was followed by intracavitary brachytherapy in the uterus and vagina, with tandem-colpostat or tandem only, for a dose of 3000 rad at point A for the former, or at 2 cm from the center of the linear source for the latter. In cases where brachytherapy was not possible, a boost of external irradiation with reduced field, with a dose of 1600 rad (200 x 8) was advised. Three hundred and one patients were registered, of which

*Present address: Department of Epidemiology and Health, McGill University, 3775 University St., Montreal, Canada H 3A2B4. Presented at the American Society of Therapeutic Radiologists Annual Meeting, October 1980; data updated April 1982. This investigation was supported in part by the National Cancer Institute. (NCI) Grant No. RlO CA-12258-08 and CA-25287.

Reprint requests to Dr. Victor A. Marcial, Professor and Chairman, Radiation Oncology Division, Center for Energy and Environment Research, Caparra Heights Station, San Juan, Puerto Rico 00935. Acknowledgement-The authors want to thank Ms. Raquel Santos for her help in typing the manuscript. Accepted for publication 23 November 1982.

431

432

Radiation

Oncology

0 Biology 0 Physics

April 1983, Volume 9, Number

4

287 are currently evaluable. No differences between the treatments were detected for the following study end-points: treatment tolerance in terms of acute normal tissue reactions and completion of therapy, tumor control in the pelvis, severe late normal tissue reactions, and survival. In the entire study population the estimated tumor control in the pelvis at two years after initiation of therapy was: 81% for Stage II-B, 67 % for III-A, 53 % for Stage III-B, and 32 % for Stage IV-A. The estimated two-year survival was: 70 % for Stage II-B, 58 % for III-A, 46 % for III-B, and 23 % for IV-A.

Uterine cervix, Radiotherapy, Fractionation, Split-course radiation, Carcinoma. INTRODUCTION

In the search for optimal radiotherapy of cancer, the Radiation Therapy Oncology Group has conducted a number of Phase II and Phase III clinical studies for tumors of various sites. Among these, the split-course study compared teletherapy in a standard continuous course versus treatment with a rest period mid-way during radiotherapy. An associated increase in the fraction dose and a decrease in the fractionation were introduced in this split-course regimen to compensate for the possible reduction in effectiveness resulting from the rest period. The radiobiological basis and the details for this study have been published.2 The following tumor sites were included: nasopharynx, tonsillar fossa, base of tongue, urinary bladder Stages C & D,, and uterine cervix Stages III-A, III-B, and IV-A initially (later in the study Stage II-B patients were also included). This report presents the results observed in the uterine cervix part of the study. Preliminary reports have been presented for base of tongue,4 tonsillar fossa,’ nasopharynx,5 and urinary bladder.’ The nasopharynx experience has been published’and the urinary bladder part has been submitted for publication. Manuscripts on the base of tongue and tonsillar fossa parts of the study are in preparation. METHODS

AND

MATERIALS

Case accession to all sites of the split-course study started in 197 1; Stage I I-B uterine cervix patient registration was initiated in 1975. Registration for the cervix part was closed in August 1980. Each RTOG institution had a copy of the protocol containing details and requirements for the study. Patients with the diagnosis of carcinoma of the cervix with the stated stages, eligible for curative radiotherapy, qualified for the study provided that the diagnosis was non-glandular carcinoma and there were no reasons for exclusion. Patients were excluded for the following reasons: age over 80, diagnosis of adenocarcinoma, other diagnosis of cancer (previous or present) except skin carcinoma, distant metastases, carcinoma during pregof the cervical stump, when the nancy, carcinoma patient’s medical condition made treatment completion unlikely, and when the prospects of follow-up appeared to be unpromising. The pre-treatment studies included complete history and physical examination, diagram of the lesion on a standard form, chest PA and lateral radiographs, CBC, urinalysis, fasting blood sugar, urea nitro-

gen, alkaline phosphatase, IV pyelogram, cystoscopy, and proctoscopy and barium enema study of the colon when rectal invasion was suspected. The staging used was as follows: Stage II-B-The carcinoma extended beyond the cervix, but had not extended into the pelvic side wall. The carcinoma could invade the vagina, but not the lower third, with obvious parametrial involvement. Stage III-A: Tumor involvement of the distal third of the vagina or extension to one pelvic wall with fixation of the mass to that structure. Stage III-B: Extension of the tumor to both pelvic walls and/or extension to one pelvic wall with fixation and to the distal third of the vagina. Stage IV-A: Invasion of the bladder and/or rectum on clinical grounds with no evidence of extrapelvic disease. In the absence of tumor induration of the anterior vaginal wall bullous edema could not be taken as representing bladder invasion. When an eligible patient was adequately staged and had the pretreatment studies according to the requirements of the protocol, the collaborating center called the RTOG central office in Philadelphia for random allocation of the patient to one of the two treatment categories. The stratification at the RTOG center was by institution and stage. Patients with urinary obstruction and rising blood urea nitrogen were to be submitted to urinary divertion by surgery prior to irradiation. Patients with symptomatic rectal obstruction were to be submitted to temporary colostomy prior to irradiation. Patients with hemoglobin levels below 10 gm were to have one or more blood transfusions to raise this level to at least 10 grams prior to irradiation. The irradiation in this study was administered with teletherapy energy of 1 MEV or higher with a minimal source skin distance of 75 cm. The collaborating centers were free to use their own treatment techniques regarding the choice and size of fields, provided that the corresponding treatment plan achieved an adequate dose distribution throughout the entire pelvis to encompass the primary tumor, the direct extensions, and potential metastatic foci in the lymph nodes of the pelvis. Stationary or rotating fields could be used at the discretion of the collaborating center. The combination of anterior and posterior, oblique and lateral fields could be used. The fields were to cover the treatment volume which would extend to a point at least 2 cm lateral to the pelvic brim, would include the lowermost visible or palpable part of the tumor and the

Cervix split-course radiotherapy 0 V. A. with a minimal 4 cm margin (unless the disease reached the introitus in which case a 2 cm margin was adequate), and extended superiorly into the abdomen at least to the midpart of the sacro-iliac joint. The usual size of the fields for the AP and PA portals was 16 cm wide and 14 cm or more along the axis of the abdomen. In the AP and PA verification fields, one had to identify the upper half of the obturator fossa and the lower half of the sacro-iliac joint, with a 2 cm margin lateral to the pelvic brim. Verification films were to be taken of all fields of treatment in the study. The stated tumor doses referred to central axis dose. The split-course patients were treated with 10 fractions of 250 rad each, 5 times a week, up to a dose of 2500 rad in 2 weeks; after a rest period of approximately 2 weeks the irradiation was repeated identically to the first part (10 fractions of 250 rad each, 5 times a week, up to 2500 rad in 2 weeks). The continuous irradiation patients received the standard continuous course of irradiation with 30 fractions, 5 times a week, 170 rad each for a total of 5 100 rad in approximately 6 weeks. Rest periods (week-ends, etc.) were not to exceed 3 days each. Both the test cases and the controls were to receive intracavitary curietherapy whenever possible by means of radioactive sources in the uterine and vaginal canals. This was to be done with a combination of intra-uterine tandem and colpostat, or a long linear source as a second choice. The dose to be delivered by this technique was calculated at point A when the tandem-colpostat combination was used, or at 2 cm distance from the center of the axis of the linear source when a tandem only was employed. Point A was defined as a point 2 cm lateral to the midline of the uterus and 2 cm superior to the plane of the external OS of the cervix, which at the time of curietherapy was to be marked with a metal clip. Intracavitary curietherapy was to follow external irradiation with a time interval between external and internal therapy of approximately 2 weeks; it was advised not to exceed 3 weeks. The intracavitary brachytherapy dose was the same for all patients. Initially the required dose was 4000 rad at the stated point, but later in the study it was lowered to 3000 rad. Patients not suitable for intracavitary curietherapy were to receive additional irradiation of 1600 rad by rotational technique with a 15 x 15 cm field with a center of rotation in the mid-line of the pelvis and with fractions of 200 rad 5 times per week. No additional radiotherapy was permissible besides the stated treatment program mentioned in the protocol unless the patient was considered a failure of protocol therapy. Personnel from the Radiological Physics Center at the M. D. Anderson Hospital in Houston calibrated each of the teletherapy units used for these patients. Patients were to be examined regularly during and at the completion of therapy, and at regular intervals after treatment for evaluation of normal tissue reactions and tumor status. The split-course patients were to be evaluated at the end of the first treatment portion, at the end of

MARCIAL et al.

433

Table 1. Patient registration

cervix

No. of patients

Institution University of Puerto Rico University of South Carolina Ellis Fischel Cancer Hospital New York Medical College University of Pennsylvania Loyola University Allegheny General Hospital New York University Medical Center Vanderbilt University Boston University Other institutions (less than 4 patients each)

93 40 36 35 22 20 I6 I 5 5 22

Total

301

the rest period, and at the end of the second treatment portion. The continuous irradiation patients were to be evaluated when they attained a dose of 3000 rad and at the completion of teletherapy. The following study forms were to be completed on each patient and forwarded to the RTOG Headquarters: initial evaluation, treatment data, normal tissue reactions and tumor response, follow-up, and death report. The statistical analysis for this study was made by the RTOG Statistical Center. The control of tumor in the pelvis and the survival figures have been estimated using the method of Kaplan and Meier.’ The following end-points were studied: treatment tolerance in terms of frequency of acute normal tissue reactions and completeness of therapy, control of tumor in the pelvis, late normal tissue reactions, and survival.

RESULTS A total of 22 institutions contributed patients to this study, but only 7 registered more than 10 patients each (Table 1). The patient population is presented in Table 2. Three hundred and one patients were registered, of whom 287 are available for analysis. Of these, 280 have the initial study form, 272 have treatment data, 268 have acute normal tissue reaction data and 270 have some follow-up data. The age distribution of the patients is shown in Table 3. There was a tendency to have younger patients in the Table 2. Patient population

Patients

registered

Cancelled or withdrawn Ineligible Patients available for analysis On study form Treatment data Normal tissue reaction data Follow-up data

Cont.

Split

Both

147 3 3

154 5 3

301 8 6

141 137 132 130 129

146 143 140 138 141

287 280 272 268 270

434

Radiation Oncology 0 Biology 0 Physics

Table 3. Patients characteristics

April 1983, Volume 9, Number 4

age distribution-percentage

Cont.

Split

Table 5. Severe normal tissue reaction

Both Reaction

I

> 30 30-39 40-49 5&59 60-69 70-80

I 21 29 32 9

4 10 24 33 I8 12

3 9 23 31 25 IO

Median

56.0

53.5

55.0

2

I

3

Patients with unknown age

group. The tumor stage distribution (Table 4) reveals that there was a fairly even distribution of stages in the two treatment groups. The predominant stage is III-B with 43% of patients, followed by II-B with 32%, III-A with 17% and IV-A with 9%. Ninety-six percent of patients were ambulatory with 4% being in bed part or all of the time. Previous abdominal surgery had been done in 33% of patients, with pelvic surgery history in 5%. There was a fairly even distribution of performance status, abdominal and pelvic surgery history, and histological pattern in the two groups. split

Cont. (% of I30 pts.)

Split (% of 138 pts.)

27 9 8 7 I 5 3 3

34 6 8 9 6 IO 2 6

Diarrhea Frequency of urination Soreness on urination Rectal tenesmus Urgency of urination Pain in irradiated area Bladder tenesmus Painful defecation

had previously unplanned surgery in 1 continuous and 1 split case; and other causes in 2 continuous cases, for a total of 8 patients in the continuous and 5 in the split groups. Almost all patients who did not complete therapy had very advanced tumors. (Two were Stage III-A, 4 were III-B, and 3 were IV-A.) One patient dropped out of the study before receiving half of the therapy and 1 was a psychiatric patient who developed a psychotic episode during teletherapy. The 3 patients who died during therapy were bedridden and in poor condition before initiating radiotherapy.

Tumor control in the pelvis Treatment tolerance The percentage frequency of severe normal tissue reactions is shown in Table 5. A severe normal tissue reaction was one that required medication. The most frequently reported reaction was diarrhea in close to 30% of patients, followed by frequency of urination, soreness of urination, rectal tenesmus, and urgency of urination in slightly below 10% for each complaint. No significant differences were noted in the frequencies of acute normal tissue reactions in the two treatment groups. The completeness of external pelvic therapy can be seen in Table 6, based on 272 patients with information on this parameter. Ninety-four percent of continuous and 95% of split therapy patients completed therapy. There is a tendency to have treatment completed as planned more often in the split group. (p = .04 one-sided exact test). There were more treatment interruptions in the continuous group (22% vs. I I %). The reasons for not completing therapy in 6% of continuous and 4% of split cases were: death in 2 continuous and I split patients; distant metastases in I patient in the continuous irradiation group; excessive reactions in 2 continuous and 3 split, the patient

Table 4. Stage distribution

Stage

Cont. w

Split %

Both %

IIB

32

IIIA IIIB TVA

I5 45 7

32 I8 41 IO

32 I7 43 9

The estimated proportion of treated patients diseasefree in the pelvis can be seen in Table 7; this refers to 272 patients. Two years after beginning therapy 62% of all patients were free of tumor in the pelvis; 59% were in the continuous and 65% in the split group. The pelvic tumor control rate at 2 years for Stage II-B was 81%, with 84% continuous and 78% for split. For Stage III-A the tumor control rate was 67%, 57% for continuous and 74% for the split group. For Stage III-B it was 53%, 48% for the continuous and 59% for the split group. The number of Stage IV-A patients eligible for tumor control evaluation is too low (8 in the continuous and 14 in the split group) to make any meaningful comparison of the two techniques, but 32% of patients with this stage were free of tumor in the pelvis by the end of two years. Table 6. Completeness of therapy (272 patients) Cont. %

Split %

Both %

Completed therapy As planned With interruption With dose modification With interruption and modification

94 72 22 0

95 81 II I

94 77 I6 I

0

I

I

Did not complete therapy

6

4

5

Unknown

0

I

I

Reasons treatment not completed Patient died Excessive treatment reactions Other

2 2 4

I 3 I

3 5 5

Cervix split-course

radiotherapy

Table 7. Tumor control in pelvis by stage Time

Stage

Treatment

IIB

Both Cont. Split

87 89 84

84 84 84

81 84 78

17189 7144 lo/45

IllA

Both Cont. Split

75 70 79

72 64 79

67 57 74

15144 8120 7124

IIlB

Both Cont. Split

62 60 65

59 56 61

53 48 59

56/l 17 33160 23157

IVA

All

1 yr. %

2 yr. %

No. failed/ total

Initial %

Both

32

32

32

15122

Cont. Split

25 36

25 36

25 36

6/8 9114

Both Cont. Split

69 68 71

67 63 69

62 59 65

1031272 54/132 491140

Table 8. Severe late tissue reactions

Radiation proctitis Radiation cystitis Soft tissue necrosis Bowel stenosis Small bowel damage Bowel obstruction Bone fracture Edema Other

Cont. (% of II 3 pts.)

Split (% of 129 pts.)

Both (% of 242 pts.)

2 2 3 0 0 1 1

5 2 2 2 3 2 2

4 2 2 I 2 2 2

I

I

I

3

6

4

0 V. A. MARCIAL et al.

435

There is a consistent tendency towards better tumor control in the pelvis among the more advanced stages treated with the split technique; however, the observed differences are not statistically significant. Metastases The incidence of distant metastases was 8% (continuous 7% split 10%). Of the observed 23 patients who have developed distant metastases, 15 (7 continuous, 8 split) had this as the first sign of failure (with pelvic disease controlled). Severe late tissue reactions The percentage of reported severe late tissue treatment induced reactions are shown in Table 8. A severe reaction required hospitalization or medical management. The most frequently reported severe tissue reaction was proctitis which occurred in 2% in the continuous and 5% in the split radiation group. This was followed by cystitis and soft tissue necrosis in 2%. A tendency is observed towards more reactions in the split group although the difference is not statistically significant. Fibrosis was excluded from the severe late tissue reactions. Some degree of treatment induced fibrosis (transient or permanent) was reported in 5% of continuous and 8% of split patients. Survival The estimated survival for all patients by stage is shown in Table 9. The estimated survival rate at 1 year was 7 1% in both treatment groups; this falls to 54% at 2 years. The median survival is 27.6 months for the continuous and 24.0 months for the split therapies. When we analyze the estimated survival at 1 and 2 years, and the median. survival by stage, we find no statistically significant difference in the two treatment groups. Time disease free is shown in Table 10.

Table 9. Survival by stage Time Median (mo.)

No. dead/total

70 76 64

Undefined Undefined 47.8

26189 lOj44 16145

13126 5110 8116

58 63 54

32.6 28.3 32.6

19144 8120 11124

13/19

Split

76 72 78

IllB

Both Cont. Split

68 70 65

46 48 43

22.1 23.6 18.0

80/l I7 39160 41157

52180 28139 24141

IVA

Both Cont. Split

38 I4 50

23 I4 27

7.9 4.2 9.5

17122

13/17

618

516

Both Cont. Split

71 71 71

54 58 50

26.1 28.3 24.9

Stage

Treatment

I yr. (%)

2 yr. (%)

I!B

Both Cont. Split

83 83 82

IllA

Both Cont.

All

Disease present at death

618 7111

Ill14

8/ll

1421272 631132 791140

911142 44163 47179

436

Radiation

Oncology

0 Biology 0 Physics

Table 10. Time disease free (months) Patient subgroup All IIB IllA IIIB IVA

cont. Undefined Undefined 20.4 17.3 o.o*

Split Undefined Undefined Undefined Undefined o.o*

Both Undefined Undefined Undefined 25.9 o.o*

Undefined-Have not reached median yet. *More than half of these patients failed, so they were never disease free.

DISCUSSION Split-course therapy is expected to be tolerated better than standard continuous irradiation, since the rest period should allow repair of the radiation-induced normal tissue damage. In addition, considerable tumor regression occurs by the end of the rest period; this should favor better oxygenation of the remaining viable tumor cells, which would be irradiated in the second part of treatment.* The preliminary results shown in this report do not reveal any difference between the two techniques. However, it is important to point out that we have limited power for detecting treatment differences. One can obtain a result of no significant difference by using sample sizes which are too small or having limited follow-up. For example, the observed difference in the estimated three-year survival rates was 7% (40 vs. 47%). Assuming that survival times are exponentially distributed and using a two-sided test at the .05 level of significance, the probability that we could have detected a difference of this magnitude was approximately 28%. To

April 1983, Volume 9, Number 4

have had 80% power, the difference would have had to have been twice as large (40 vs. 54%). Thus, given our accrual and follow-up it is unlikely that we could have found a meaningful difference. For example, we only had a 51% chance of detecting a 50% increase in the 5-year survival rate. The use of brachytherapy after external pelvic irradiation may represent an important factor that may reduce the chances of detecting differences in the two tested teletherapy techniques. Brachytherapy followed external pelvic irradiation in 79% of continuous versus 7 1% of the split patients. The reasons for this were: residue tumor (9 continuous, 14 split), unsuitable anatomy (5 continuous and 9 split), patient refusal (3 continuous and 1 split), and other or nonstated (8 continuous and 12 split). The fact that investigators had the option of boosting the external treatment with a rotational field may account for the high percentage of patients who did not receive brachytherapy. The influence of this management on tumor control and survival will be discussed in a future report. To evaluate late tissue damage in a reliable way, one would have to observe these patients for a minimum of two or three years. At the time of this report, we can state that 12% of patients in the continuous group and 16% in the split, developed some type of late tissue reactions. Details of the nature and seriousness of these complications will be the subject of a future report. If the rates of pelvic tumor control and survival remain the same in these two therapeutic regimens, and the late tissue reactions remain comparable, we would recommend the split technique which requires less fractionation (20 versus 30 fractions). This represents less social and economic burden to the patient, and less effort per case for the therapy facility.

REFERENCES Kaplan, E., Meier, P.: Non parametric incomplete 1958.

observations.

estimation from J. Am. Stat. Assoc. 53: 457-48 1,

Marcial, V.A.: Split-course radiation therapy project. Cuncer 29: 1463-1467, 1972. Marcial, V., Brady, L., Johnson, R., Tome, J.: Comparison of split-versus continuous radiotherapy in advanced carcinoma of the urinary bladder-An RTOG study. Proc. of ARS, (Abs. 30). Int. J. Radiat. Oncol. Biol. Phys. 6(l): 34-35, 1980. Marcial, V., Davis, L., Hendrickson, F., Hanley, J., Shalek, R., Ortiz, H.: Split-course versus continuous radiation therapy for carcinoma of the base of the tongue (National prospective trial), Proc. of ARS, (Abs. 25). Int. J. Radiat. Oncol. Biol. Phys. 2 (1): 40, 1977.

5. Marcial, V., Hanley, J., Chang, C., Davis, L.: Split-course radiation therapy of carcinoma of the nasopharynx: Results of a prospective national collaborative clinical trial of the RTOG. Proc. of ARS, (Abs. 17). Int. J. Radiat. Oncol. Biol. Phys. 5 (1): 36, 1979. 6. Marcial, V., Hanley, J., Chang, C., Davis, L., Moscol, J.: Split-course radiation therapy of carcinoma of the nasopharynx: Results of a national collaborative clinical trial of the Radiation Therapy Oncology Group. Int. J. Radiat. Oncol. Biol. Phys. 6: 40994 14, 1980. 7. Marcial, V., Hanley, J., Rotman, M.: Split-course radiation therapy of carcinoma of the tonsillar fossa-Results of a prospective national collaborative clinical trial of the RTOG, Proc. of ARS (Abs. 5). Int. J. Radiat. Oncol. Biol. Phys. 4 (I): 17-18, 1978.