hr J. Radmron Oncology Biol. Phys Vol. 21. pp Prmted in the U.S.A All rights reserved
1137-I 143 Copynght
0360.3016/91 $3.00 + .OO 0 1991 Pergamon Press plc
??Clinical Original Contribution
ADJUVANT
THERAPY
OF RESECTED
RICHARD WHITTINGTON M.D.,*
ADENOCARCINOMA
MARK P. BRYER M.D.,*
LAWRENCE J. SOLIN M.D.*
OF THE PANCREAS
DANIEL G. HALLER M.D.,+
AND ERNEST F. ROSATO M.D.*
University of Pennsylvania School of Medicine, Philadelphia, PA Seventy-two patients underwent resections of pancreatic carcinomas between 1981 and 1989 at the Hospital of the University of Pennsylvania and were evaluable for follow-up. There were three treatment groups as treatment policies evolved. Initially, patients were observed after surgery without adjuvant treatment (Group l-33 patients). Beginning in 1984, patients were offered adjuvant radiation therapy postoperatively (Group 2-19 patients) and eight of these patients also received 5-FU as an ZV bolus on the first 3 days of the first and fifth weeks of treatment. Twenty patients were treated with chemosensitized radiation therapy following surgery using 96-hour 5-FU infusions during the first and fifth weeks of treatment. There were four postoperative deaths, which are excluded from the analysis, and sites of failure could not be determined for five other patients. Among evaluable patients, local recurrences occurred in 85% of the patients in group 1, 55% of the patients in group 2, and 25% of the patients in group 3. The 2-year survival was 35% in group 1, 30% in group 2, and 43% in group 3. Patients with involved surgical margins had a poor survival; only 2 of these 16 patients survived longer than 18 months. Among patients with negative margins, the 2-year survival is 41% in group 1, 33% in group 2, and 59% in group 3. Although the number of patients is smaller, the 3-year survival is 22% in group 1, 11% in group 2, and 47% in group 3. Chemosensitized irradiation is well tolerated in these patients. The major challenge in this group of patients is nutritional maintenance. There was no other significant toxicity. The trend in these observations suggests that survival following pancreatic resection is substantially improved with the addition of adjuvant chemosensitized radiation therapy. Adenocarcinoma of the pancreas, Adjuvant therapy.
Whippie
procedure,
Pancreatectomy,
as well
as a high
incidence
radiation
therapy,
Given the relative infrequency of curative resections, it is difficult to accumulate an experience in large series of patients from a single institution. Some authors have added preoperative or postoperative irradiation to the resection to improve the results, but significant toxicity occurred and there was little survival benefit (4, 7). Chemotherapy has been noted to have limited efficacy in advanced metastatic disease, although drugs such as 5-Fluorouracil (5-FU) may have synergistic activity with radiation (6).
INTRODUCTION
In the United States, pancreatic adenocarcinomas are not only the fifth most common cause of cancer death, but they present particularly difficult management problems. In one recent review by Michelassi ef al. the rate of resectability reported for the period 1981-1989 was 10.5%, compared to 14.5% for the period 1971-1979 (5). A review of 37,000 cases by Gudjonsson demonstrated that only 10% of the patients were able to undergo curative resection (2). The 5-year survival among all patients was 0.4%, which was increased to just 3.8% among resected patients. This result is substantially better than the survival among unresected patients of 0.02%. These results suggest that although surgical resection is critical for successful treatment, it is not by itself sufficient. In evaluating the causes for failure following surgery, Tepper and Nardi reported a 50% incidence of local recurrence, distant failure (9).
Chemosensitized
Building on prior experience in patients with locally advanced pancreatic cancer, the Gastrointestinal Tumor Study Group (GITSG) treated a series of patients with combined chemotherapy and radiation therapy following curative resection of the tumor (3). Patients were randomized to observation after surgery or adjuvant postoperative irradiation and bolus 5-FU chemotherapy. In this study, the survival was significantly improved by adjuvant treatment. The results were confirmed in a subsequent non-randomized se-
of
$Department of Surgery. Reprint requests to: Richard Whittington M.D., Department of Radiation Oncology, Hospital of the University of Pennsylvania, 34th and Spruce Sts., Philadelphia PA 19104. Accepted for publication 29 March 1991.
Presented at the 32nd Meeting of the American Society of Therapeutic Radiology and Oncology, Miami FL, 1.5-19 October 1990. *Department of Radiation Oncology. tDepartment of Internal Medicine-Division of Medical Oncology. 1137
1138
I. J. Radiation
Oncology
0 Biology 0 Physics
ries of patients treated with 5-FU and irradiation (1). Subsets of patients showing the greatest improvement in 3-year survival were those with tumors confined to the pancreas (40% vs 6%) and better performance status. Based on these observations, a series of patients was treated at the University of Pennsylvania after curative resection. METHODS
AND MATERIALS
Seventy-five patients underwent curative pancreatic resection at the Hospital of the University of Pennsylvania between January 1981 and September 1989. Three patients were excluded from further analysis because of inadequate follow-up information, and the remaining 72 patients serve as the basis for this report. Between January 1981 and June 1984, patients were not offered postoperative therapy (group 1). Beginning in July 1984, patients were offered adjuvant radiation therapy to reduce the risk of local recurrence (group 2a). Following presentation of the results of the GITSG study, patients were also treated with bolus 5-FU (group 2b). In an effort to intensify therapy further to improve survival, patients undergoing resection after May 1986 were offered chemosensitized radiation therapy (group 3). Some patients or their physicians declined some or all adjuvant therapy and are grouped to reflect the treatment delivered. Adjuvant treatment was not withheld for any patient because of their clinical status. Thirty-three patients are in group 1, 11 patients are in group 2a, 8 patients are in group 2b, and 20 patients are in group 3. Surgery consisted of Whipple resection in 59 patients, distal pancreatectomy in 8, and total pancreatectomy in 5. The procedure was completed even if positive lymph nodes were identified in association with an otherwise resectable lesion. Formal lymph node dissection was not routinely carried out, but an attempt was made to remove all palpable lymph nodes. Since 1984, all patients had a feeding jejunostomy placed at the time of surgery that was left in place through the first 2 weeks of radiation. If a patient was maintaining adequate oral intake the jejunostomy was removed; otherwise it was left in until the patient was nutritionally repleted. Surgical procedures did not change appreciably during the course of this review, and 43 of the 72 resections were performed by one surgeon (EFR). Postoperative irradiation began 4 to 6 weeks following surgery, although treatment was delayed if necessary to allow full recovery from surgery. Patients were treated using CT-based treatment planning. The target volume included the initial tumor volume with a 3 cm lateral and craniocaudal margin. The anastamosis of the pancreatic remnant to the intestine was included with a margin of 2 to 3 cm. Lymph nodes treated included the celiac and paraaortic lymph nodes to the bottom of L, as well as the periductal nodes to the level of the cystic duct and the splenic vein nodes. On the lateral fields the target volume included the nodes posterolateral to the aorta and the vena cava.
October
1991, Volume 21, Number 5
Since 1984, all patients underwent treatment planning CT scans to allow for beam’s-eye-view projection of target volumes and critical structures and the calculation of dosevolume histograms. Treatment was constrained to limit the kidney volume treated to a dose of 1800 cGy to less than 50% of the total volume and to limit the volume of liver exceeding 3000 cGy to less than 50% of the liver. Patients were treated with 3 or 4 fields at a dose of 180 cGy/fx treating 5 fx/wk. The prescribed dose was 4500 to 4860 cGy for all patients with negative surgical margins. Those patients with positive margins or residual disease were treated to a dose of 5400 to 6300 cGy. There were two chemotherapy regimens used during this period. Eight patients in group 2b were treated with adjuvant irradiation and received adjuvant chemotherapy with bolus 5-FU (500 mg/m*) on the first 3 days of the first and fifth week of irradiation. Only two of these patients received maintenance chemotherapy with weekly bolus 5-FU at a dose of 500 mg/m2. Those patients treated in group 3 with chemosensitized irradiation were treated with 5-FU infusions (1000 mg/m2/d) for 96 hr during the first and fifth week of irradiation and a single dose of mitomycin-C (10 mg/m2) on the first day of treatment. Patients receiving chemosensitized radiation therapy were not offered maintenance chemotherapy. Follow-up studies included computed tomography every 4 months in the first 2 years and every 6 months for the next 2 years, then annually. All sites of recurrence were recorded for each patient and are included in the analysis of the patterns of failure. Survival was measured from the date of surgery and survival curves were calculated by the Berkson-Gage method. In comparing the results between these non-randomized groups of patients, statistical tests are applied only to suggest trends in the observations.
RESULTS The characteristics of the patients in each of the treatment groups are shown in Table 1. Because the survival and the patterns of failure among those patients receiving adjuvant irradiation with or without bolus 5-FU are similar, they are considered as a single group in the analysis (group 2). Morbidity related to treatment is shown in Table 2. There were four postoperative deaths due to bleeding (3) and infection due to an anastamotic leak (l), and six incidents requiring reoperation, two episodes of bleeding one anastamotic leak and one liver abcess. Less severe problems were seen in 14 patients including less severe bleeding (2), superficial wound infection (9), and protracted ileus (3). Morbidity related to irradiation and chemotherapy was infrequent. One patient developed cholangitis and a second developed a duodenal ulcer following radiation. These were the only two episodes requiring hospitalization. Less severe problems requiring treatment interruption for less than 1 week included nausea and vomiting (3) and weight loss (1). Treatment assignment was not changed in any
Adjuvant
therapy of pancreatic
carcinoma
??R.
WHITTINGTON er al.
1139
Table 1. Characteristics of patients undergoing resection of pancreatic adenocarcinoma Group I Treatment Chemosensitized RT Number of patients Age (med) (range) Sex (M/F) Location Head Body and tail Post-op deaths Tumor vol. Mean Range Nodal involvement (%) Residual tumor (%) Positive margin Gross residual Note: Characteristics
IIa
Surgery
Surgery + RT
33
11
59 45-15 14/19 27
54 31-72 615 9
Bolus 5-FU 8 61 51-73 414 6
Surgery + 20 58 23-69 12/8 16
2 0
2 0
4 0
61
69
38
78
l-384 45
8-125 54
8-105 15
2-514 70
21
21
25
15
0
9
0
10
of the patient included in each treatment group.
related morbidity Group
Bleeding Anastamotic leak Infection Protracted ileus Chylous ascites Jaundice
III
6 4
patient because of postoperative clinical course, and radiation prescription was not altered in any patient because of treatment tolerance. One patient refused the second infusion of 5-FU due to mucositis during the first course, and this was the only chemotherapy dose adjustment. Although many patients had anorexia and were limited in their oral intake during treatment, malnutrition was an infrequent problem as 33 of the 39 patients in groups 2 and 3 had jejunostomies at the beginning of treatment. More than onethird of the patients had feeding jejunostomies removed during the course of treatment. Patterns of recurrence are listed in Table 3. The four patients in group 1 dying in the early postoperative period were excluded from this analysis. Physicians signing death certificates listing pancreatic carcinoma as a cause of death were contacted to determine the extent of disease at the
Table 2. Treatment
IIIb Surgery + RT
I
II
III
5 (3) 2 (1) 3 2 1 0
1 0 1 2 0 1
0 0 3 0 0 0
Note: Treatment related morbidity of the patients in each treatment group.
time of death. For three patients in group 1 and two patients in group 2, more specific information could not be obtained. These patients were also excluded from the analysis of local control. Among the 26 evaluable patients in group 1, 22 of the 23 patients recurring had a local component of tumor recurrence. Postoperative irradiation was associated with a reduced risk of local recurrence (53%). There was no significant effect of adjuvant bolus 5-FU on the risk of local recurrence (3 of 8 in group 2b vs 6 of 9 in group 2a). Patients treated with chemosensitized irradiation had the lowest risk of local recurrence (25%). In the group of patients treated with chemosensitized irradiation, one of the recurrences occurred in a patient with gross residual tumor following surgery, and three recurrences occurred in the common bile duct at the margin of the field. Sites of metastatic disease are also listed in Table 3. The distribution of metastatic sites was similar in all treatment groups Q-not significant for all comparisons). The survival of patients is shown in Figure 1. Median survival was similar in all three groups (group 1-15 mo, group 2-15 mo, group 3-16 mo), although there was an improvement in 3-year survival with chemosensitized irradiation (34% in group 3 vs 8% and 5% in groups 1 and 2). Eight of the 20 patients in group 3 are alive and diseasefree following chemosensitized irradiation with follow-up ranging from 13 to 41 months (median 27 months). Three patients in group 1 are alive and disease free at 10, 13, and 30 months, and three patients are alive in group 2 at 11, 21, and 40 months. The latest recurrence yet observed was 30 months following diagnosis.
I. .I. Radiation Oncology 0 Biology 0 Physics
1140
October 1991, Volume 21, Number 5
Table 3. Patterns on recurrence Group
Evaluable patients Sites of recurrence Pancreatic bed Liver Peritoneal seeding GI tract Not specified Abdominal wall/ incision Other
I
II
III
29 22
19 9
20 5
6 6 4 3 3
8 4 0 2 1
5 3 0 0
Bone- 1 Supraclav LN- 1
Lung- 1 Axilla- I
4
4
No failure
I Lung- 1
Bone-2 8
Sites of tumor recurrence following curative treatment of pancreatic carcinoma.
Patients with complete resections and negative surgical margins were examined seperately, and their survival is plotted in Figure 2. Survival is substantially better in this group of patients with the greatest effect in those patients receiving adjuvant chemosensitized radiation. Median survival in group 3 was 29 months compared to 21 months in group 1 and 16 months in group 2. Median survival was longer in patients receiving postoperative treatment even if they had positive margins (9.5 mos-group 1, 14.5 mosgroup 2, 14 months-group 3), but there are no patients currently alive in this group, and only two patients survived longer than 18 months. Since patients were not randomized to treatment groups, strict criteria to test for statistical significance cannot be fulfilled. Statistical calculations are reported only to suggest trends in the results. A log rank comparison showed a modest improvement in late survival with chemosensitized radiation when all patients were analyzed (group 3 vs group 1 p > 0.1, group 3 vs group 2 p > 0.09). A comparison
90 L:
z
80
5
70
;
Cd 50
$
40
2
30
$
20 10 0
0
4
8
I2
16
20
2.1
2x
32
36
40
MONTHS
Fig. 1. Survival of all patients following carcinoma.
resection
of pancreatic
Fig. 2. Suvival of patients following creatic carcinoma.
complete
resection of pan-
of patients undergoing complete resection with fewer patients showed a similar trend (group 1 vs group 3 p > 0.06, group 2 vs group 3 p > 0.14).
DISCUSSION Pancreatic carcinoma presents a therapeutic challenge to the surgeon. The location of the pancreas will preclude resection of many small tumors because of local extension to critical vascular structures such as portal vein, celiac vessels, and superior mesenteric artery. When the tumor is resected it is equally difficult to obtain adequate radial margins because limited extrapancreatic extension to vascular structures as well as the pancreatic remnant, liver, and the biliary tree are difficult to mobilize and dissect. Similarly, these structures complicate efforts to perform regional node dissections. All of these factors raise the risk of local and regional recurrence in this disease. Efforts to extend the operative procedure to obtain wider margins and to improve the survival, thus far, have failed to improve survival. Sensitive tissues such as kidney and liver may also lie in the volume traversed by an external radiation field, creating pressure to limit the size of the target volume. Small bowel will fill the volume previously occupied by the pancreatic tumor, and it is not feasible to perform procedures to displace small bowel as in the pelvis; therefore, it is also necessary to limit total dose. Previous attempts to improve the results following surgery include efforts directed at preoperative radiation as described by Kopelson (4) or postoperative treatment described by Rich (8) and Nguyen et al. (7). While these studies demonstrate the feasibility of moderate dose postoperative irradiation, they fail to demonstrate a significant survival benefit. Kopelson treated five patients preoperatively with 4045 Gy directed at the tumor. Four of the five patients had prior surgeries and one had an endoscopic biopsy. Three of the five patients had significant regression noted at the sec-
Adjuvant therapy of pancreatic carcinoma 0 R. WHIITNGTONet al.
@I
64
Fig. 3. Typical treatment portal of patient in group 3 demonstrating (a_ante& field, b: lateral field). _
ond surgery and all were resected following irradiation. Nodes were negative in all patients. One patient had a marginal recurrence 7 years following treatment and two patients are alive, one at 15 months and the other more than 8 years. Nguyen et al. treated nine patients following resection to a dose of 60 Gy using small fields. Maximum field dimensions were 15 cm length, 15 cm anterior width, and 10 cm lateral width. Median survival was 13 months. Patients treated at the University of Pennsylvania had tumors that were too large to be treated with fields that small. The median volume was 67 cm3 in group 1, 55 cm3 in group 2, and 78 cm3 in group 3. Rich treated 10 patients with a high risk of recurrence following resection. Two patients had gross residual tumor, two had microscopic residual tumor, two had perineural invasion, and three had positive nodes. Local recurrence occurred in only 1 of 10 patients following doses of 40 to 45 Gy. The two patients with gross residual also received intraoperative orthovoltage boost of 12.5 Gy One patients survived more than 8 years and two are alive without disease less than 5 years. There are two rationales for delivering chemotherapy and irradiation concurrently. The first is to apply early treatment to both the tumor bed and to micrometastases when the efficacy of each is increased. This is unlikely to affect survival since there are no data to support the use of adjuvant chemotherapy alone following pancreatic resection. A second rationale is that the chemotherapy-radiation
1141
target volume and blocking of critical structures
interaction renders residual tumor cells more sensitive to the effects of treatment. Vokes and Weichselbaum have reviewed data suggesting greater cytotoxicity of radiation in the presence of 5-FU and further enhancement with protracted exposure to chemotherapy (10). They also review data suggesting enhanced activation of mitomycin in hypoxic cells where radiation effects are reduced. The effects of mitomycin may be additive with radiation. CT-based treatment planning has allowed for the treatment of much larger volumes with acceptable morbidity. An example of the treatment fields used in one patient are demonstrated in Figure 3. The volume includes the primary tumor volume as well as the adjacent nodal and vascular structures. Because of the variable position of the kidneys and liver, it is necessary to use CT scans to create beam’s-eye-view projections to facilitate field shaping. It is frequently necessary to change the field weights based on the location of the target. Patients with tumors in the body and tail region were generally treated with 60% of the dose delivered through the lateral fields, whereas those patients with smaller lesions in the head of the pancreas were treated with 60% of the dose delivered through the anterior and posterior fields. In comparing the results of this population with the previously reported series from the GITSG, the patients treated in this series had more factors associated with a poorer prognosis. Twenty-eight percent of the patients in the GITSG trial had positive nodes compared to 45% in
1142
I. J. Radiation
Oncology
0 Biology 0 Physics
group 1, 63% of patients in group 2, and 70% of patients in group 3 in this series. Four percent of the patients in the GITSG trial had lesions located in the body and tail compared with 12% of patients in group 1 and 20% of patients in groups 2 and 3. Postoperative treatment appeared to reduce the risk of local recurrence significantly among patients following surgery. The observed incidence of local recurrence among patients in group 1 is substantially higher than was observed in patients on the GITSG trial (85% vs 33%) but it is not clear whether this represents different surgical results, closer surveillance for local recurrence, or more advanced lesions. Recurrence rates among patients in group 2 and 3 are significantly lower than observed in group 1, 53% in group 2, and 25% in group 3 (p < .OOOl group 1 vs 2 vs 3). While there is a trend toward a lower risk of local recurrence in group 3, it does not reach significance (group 2 vs group 3-p < 0.1). There is no significant difference in the frequency or site of distant metastases in the three treatment groups (p > 0.2 for all comparisons). One could therefore infer that the improved survival observed in these patients is related to the enhanced local control that can be achieved with chemosensitized radiation therapy. Nutritional support of patients during the course of treatment was best accomplished with an enteral feeding tube placed at the time of surgery. Although quantitative data were not prospectively obtained, it was a subjective observation of all observers that nutritional maintenance was substantially improved with the availability of a feeding tube until the patient’s tolerance of the GI effects of radiation could be determined.
October
1991, Volume 21, Number 5
Adjuvant radiation therapy can be delivered safely to patients as part of a planned combined modality approach to resected pancreatic carcinoma. In this group of patients the addition of chemosensitization enhanced the results, and that patient tolerance was acceptable. Among patients with incomplete resections, the mediam survival was prolonged, although there were no long term survivors with postoperative treatment. Among patients with complete resections, there is a substantial improvement in 2-year and 3-year survival with the addition of chemosensitized radiation. With the improvement in local control observed in these patients with the adjuvant treatment, the focus for future treatment efforts must concentrate on efforts to reduce the risk of systemic metastases. Because of the relative low frequency of complete resections and the limited experience of single institutions, this disease may not be amenable to further randomized studies. In an effort to increase the potential for resection, a current research effort at the Fox Chase Cancer Center delivers chemosensitized irradiation prior to attempted resection (11). Previous reports have suggested that more extensive manipulation of the tumor may increase the risk of distant metastases, and preoperative irradiation may also reduce the risk of distant failure (12). An NCI-sponsored research protocol attempted to reduce the risk of metastases with the further intensification of the GITSG regimen by delivering one cycle of streptozotocin, mitomycin, and 5-FU (SMF) prior to irradiation and a second cycle following irradiation. That trial was terminated because of poor accrual. Future efforts to address subclinical micrometastatic disease will be needed to further improve the survival of these patients.
REFERENCES 1. Gastrointestinal Tumor Study Group. Further evidence of effective adjuvant combined radiation and chemotherapy following curative resection of pancreatic cancer. Cancer 59: 20062010; 1987. 2. Gudjonsson, B. Cancer of the pancreas: 50 years of surgery. Cancer 60:2284-2403; 1987. 3. Kalser, M. H.; Ellenberg S. S. Pancreatic cancer: adjuvant combined radiation and chemotherapy following curative resection. Arch. Surg. 120899-903; 1985. G. Curative surgery for adenocarcinoma of the 4. Kopelson, pancreas/ampulla of vater: the role of adjuvant pre- or postoperative radiation therapy. Int. J. Radiat. Oncol. Biol. Phys. 9:91 l-915; 1983. 5. Michelassi, F.; Erroi, F.; Dawson, P. J.; Pietrabissa, A.; Noda, S.; Handcock, M.; Block, G. E. Experience with 647 consecutive tumors of the duodenum, ampulla, head of the pancreas, and distal common bile duct. Ann. Surg. 210:544553; 1989. 6. Moertel, C. G.; Frytak, S.; Hahn, R. G.; O’Connell, M. J.; Reitmeier, R. J.; Rubin, J.; Schutt, A. J.; Weiland, L. H.; Childs, D. S.; Holbrook, M. A.; Lavin, P. T.; Livstone, E.; Spiro, H.; Knowlton, A.; Kalser, M.; Barkin, J.; Lessner,
H.; Mann-Kaplan, R.; Ramming, K.; Douglas, H. 0.; Thomas, P.; Nave, H.; Bateman, J.; Lokich, J.; Brooks, J.; Chaffey, J.; Corson, J. M.; Zamcheck, N.; Novak, J. W. Therapy of locally unresectable pancreatic carcinoma: a randomized comparison of high dose (6000 rads) radiation alone, moderate dose radiation (4000 rads + 5_fluorouracil), and high dose radiation + 5-fluorouracil. Cancer 48:17051710; 1981.
7. Nguyen, T. D.; Bugat, R.; Combes, P. F. Postoperative irradiation of carcinoma of the head of the pancreas area: short-time tolerance and results to precision high dose technique in 18 patients. Cancer 50:53-56; 1982. 8. Rich, T. A. Radiation therapy for pancreatic cancer: elevenyear experience at the JCRT. Int. J. Radiat. Oncol. Biol. Phys. 11:759-763; 1985. 9. Tepper, J.; Nardi, G.; Suit, H. Carcinoma of the pancreas: Review of the MGH experience from 1963 to 1973. Cancer 37:1519-1524; 1976. 10. Vokes, E. E.; Weichselbaum R. R. Concomitent chemoradiotherapy: rationale and clinical experience in patients with solid tumors. J. Clin. Oncol. 8:911-934; 1990.
Adjuvant therapy of pancreatic carcinoma 0 R. 11. Weese, J. L.; Nussbaum, M. L.; Paul A. R.; Solin L. J.; Kowalyshyn, M. J.; Hoffamn J. P. Enhanced resectability of pancreatic and periampullary cancer after neoadjuvant chemoradiotherapy. Int. J. Pancreatology 7: 177-185; 1990.
WHITTINGTONet al.
1143
12. Whittington, R.; Solin, L.; Mohiuddin, M.; Cantor, R. I.; Rosato, F. E.; Biermann, W. A.; Weiss, S. M.; Pajak, T. F. Multimodality therapy of localized unresectable pancreatic adenocarcinoma. Cancer 54: 1991-1998; 1984.
1137-I 143 Copynght
0360.3016/91 $3.00 + .OO 0 1991 Pergamon Press plc
??Clinical Original Contribution
ADJUVANT
THERAPY
OF RESECTED
RICHARD WHITTINGTON M.D.,*
ADENOCARCINOMA
MARK P. BRYER M.D.,*
LAWRENCE J. SOLIN M.D.*
OF THE PANCREAS
DANIEL G. HALLER M.D.,+
AND ERNEST F. ROSATO M.D.*
University of Pennsylvania School of Medicine, Philadelphia, PA Seventy-two patients underwent resections of pancreatic carcinomas between 1981 and 1989 at the Hospital of the University of Pennsylvania and were evaluable for follow-up. There were three treatment groups as treatment policies evolved. Initially, patients were observed after surgery without adjuvant treatment (Group l-33 patients). Beginning in 1984, patients were offered adjuvant radiation therapy postoperatively (Group 2-19 patients) and eight of these patients also received 5-FU as an ZV bolus on the first 3 days of the first and fifth weeks of treatment. Twenty patients were treated with chemosensitized radiation therapy following surgery using 96-hour 5-FU infusions during the first and fifth weeks of treatment. There were four postoperative deaths, which are excluded from the analysis, and sites of failure could not be determined for five other patients. Among evaluable patients, local recurrences occurred in 85% of the patients in group 1, 55% of the patients in group 2, and 25% of the patients in group 3. The 2-year survival was 35% in group 1, 30% in group 2, and 43% in group 3. Patients with involved surgical margins had a poor survival; only 2 of these 16 patients survived longer than 18 months. Among patients with negative margins, the 2-year survival is 41% in group 1, 33% in group 2, and 59% in group 3. Although the number of patients is smaller, the 3-year survival is 22% in group 1, 11% in group 2, and 47% in group 3. Chemosensitized irradiation is well tolerated in these patients. The major challenge in this group of patients is nutritional maintenance. There was no other significant toxicity. The trend in these observations suggests that survival following pancreatic resection is substantially improved with the addition of adjuvant chemosensitized radiation therapy. Adenocarcinoma of the pancreas, Adjuvant therapy.
Whippie
procedure,
Pancreatectomy,
as well
as a high
incidence
radiation
therapy,
Given the relative infrequency of curative resections, it is difficult to accumulate an experience in large series of patients from a single institution. Some authors have added preoperative or postoperative irradiation to the resection to improve the results, but significant toxicity occurred and there was little survival benefit (4, 7). Chemotherapy has been noted to have limited efficacy in advanced metastatic disease, although drugs such as 5-Fluorouracil (5-FU) may have synergistic activity with radiation (6).
INTRODUCTION
In the United States, pancreatic adenocarcinomas are not only the fifth most common cause of cancer death, but they present particularly difficult management problems. In one recent review by Michelassi ef al. the rate of resectability reported for the period 1981-1989 was 10.5%, compared to 14.5% for the period 1971-1979 (5). A review of 37,000 cases by Gudjonsson demonstrated that only 10% of the patients were able to undergo curative resection (2). The 5-year survival among all patients was 0.4%, which was increased to just 3.8% among resected patients. This result is substantially better than the survival among unresected patients of 0.02%. These results suggest that although surgical resection is critical for successful treatment, it is not by itself sufficient. In evaluating the causes for failure following surgery, Tepper and Nardi reported a 50% incidence of local recurrence, distant failure (9).
Chemosensitized
Building on prior experience in patients with locally advanced pancreatic cancer, the Gastrointestinal Tumor Study Group (GITSG) treated a series of patients with combined chemotherapy and radiation therapy following curative resection of the tumor (3). Patients were randomized to observation after surgery or adjuvant postoperative irradiation and bolus 5-FU chemotherapy. In this study, the survival was significantly improved by adjuvant treatment. The results were confirmed in a subsequent non-randomized se-
of
$Department of Surgery. Reprint requests to: Richard Whittington M.D., Department of Radiation Oncology, Hospital of the University of Pennsylvania, 34th and Spruce Sts., Philadelphia PA 19104. Accepted for publication 29 March 1991.
Presented at the 32nd Meeting of the American Society of Therapeutic Radiology and Oncology, Miami FL, 1.5-19 October 1990. *Department of Radiation Oncology. tDepartment of Internal Medicine-Division of Medical Oncology. 1137
1138
I. J. Radiation
Oncology
0 Biology 0 Physics
ries of patients treated with 5-FU and irradiation (1). Subsets of patients showing the greatest improvement in 3-year survival were those with tumors confined to the pancreas (40% vs 6%) and better performance status. Based on these observations, a series of patients was treated at the University of Pennsylvania after curative resection. METHODS
AND MATERIALS
Seventy-five patients underwent curative pancreatic resection at the Hospital of the University of Pennsylvania between January 1981 and September 1989. Three patients were excluded from further analysis because of inadequate follow-up information, and the remaining 72 patients serve as the basis for this report. Between January 1981 and June 1984, patients were not offered postoperative therapy (group 1). Beginning in July 1984, patients were offered adjuvant radiation therapy to reduce the risk of local recurrence (group 2a). Following presentation of the results of the GITSG study, patients were also treated with bolus 5-FU (group 2b). In an effort to intensify therapy further to improve survival, patients undergoing resection after May 1986 were offered chemosensitized radiation therapy (group 3). Some patients or their physicians declined some or all adjuvant therapy and are grouped to reflect the treatment delivered. Adjuvant treatment was not withheld for any patient because of their clinical status. Thirty-three patients are in group 1, 11 patients are in group 2a, 8 patients are in group 2b, and 20 patients are in group 3. Surgery consisted of Whipple resection in 59 patients, distal pancreatectomy in 8, and total pancreatectomy in 5. The procedure was completed even if positive lymph nodes were identified in association with an otherwise resectable lesion. Formal lymph node dissection was not routinely carried out, but an attempt was made to remove all palpable lymph nodes. Since 1984, all patients had a feeding jejunostomy placed at the time of surgery that was left in place through the first 2 weeks of radiation. If a patient was maintaining adequate oral intake the jejunostomy was removed; otherwise it was left in until the patient was nutritionally repleted. Surgical procedures did not change appreciably during the course of this review, and 43 of the 72 resections were performed by one surgeon (EFR). Postoperative irradiation began 4 to 6 weeks following surgery, although treatment was delayed if necessary to allow full recovery from surgery. Patients were treated using CT-based treatment planning. The target volume included the initial tumor volume with a 3 cm lateral and craniocaudal margin. The anastamosis of the pancreatic remnant to the intestine was included with a margin of 2 to 3 cm. Lymph nodes treated included the celiac and paraaortic lymph nodes to the bottom of L, as well as the periductal nodes to the level of the cystic duct and the splenic vein nodes. On the lateral fields the target volume included the nodes posterolateral to the aorta and the vena cava.
October
1991, Volume 21, Number 5
Since 1984, all patients underwent treatment planning CT scans to allow for beam’s-eye-view projection of target volumes and critical structures and the calculation of dosevolume histograms. Treatment was constrained to limit the kidney volume treated to a dose of 1800 cGy to less than 50% of the total volume and to limit the volume of liver exceeding 3000 cGy to less than 50% of the liver. Patients were treated with 3 or 4 fields at a dose of 180 cGy/fx treating 5 fx/wk. The prescribed dose was 4500 to 4860 cGy for all patients with negative surgical margins. Those patients with positive margins or residual disease were treated to a dose of 5400 to 6300 cGy. There were two chemotherapy regimens used during this period. Eight patients in group 2b were treated with adjuvant irradiation and received adjuvant chemotherapy with bolus 5-FU (500 mg/m*) on the first 3 days of the first and fifth week of irradiation. Only two of these patients received maintenance chemotherapy with weekly bolus 5-FU at a dose of 500 mg/m2. Those patients treated in group 3 with chemosensitized irradiation were treated with 5-FU infusions (1000 mg/m2/d) for 96 hr during the first and fifth week of irradiation and a single dose of mitomycin-C (10 mg/m2) on the first day of treatment. Patients receiving chemosensitized radiation therapy were not offered maintenance chemotherapy. Follow-up studies included computed tomography every 4 months in the first 2 years and every 6 months for the next 2 years, then annually. All sites of recurrence were recorded for each patient and are included in the analysis of the patterns of failure. Survival was measured from the date of surgery and survival curves were calculated by the Berkson-Gage method. In comparing the results between these non-randomized groups of patients, statistical tests are applied only to suggest trends in the observations.
RESULTS The characteristics of the patients in each of the treatment groups are shown in Table 1. Because the survival and the patterns of failure among those patients receiving adjuvant irradiation with or without bolus 5-FU are similar, they are considered as a single group in the analysis (group 2). Morbidity related to treatment is shown in Table 2. There were four postoperative deaths due to bleeding (3) and infection due to an anastamotic leak (l), and six incidents requiring reoperation, two episodes of bleeding one anastamotic leak and one liver abcess. Less severe problems were seen in 14 patients including less severe bleeding (2), superficial wound infection (9), and protracted ileus (3). Morbidity related to irradiation and chemotherapy was infrequent. One patient developed cholangitis and a second developed a duodenal ulcer following radiation. These were the only two episodes requiring hospitalization. Less severe problems requiring treatment interruption for less than 1 week included nausea and vomiting (3) and weight loss (1). Treatment assignment was not changed in any
Adjuvant
therapy of pancreatic
carcinoma
??R.
WHITTINGTON er al.
1139
Table 1. Characteristics of patients undergoing resection of pancreatic adenocarcinoma Group I Treatment Chemosensitized RT Number of patients Age (med) (range) Sex (M/F) Location Head Body and tail Post-op deaths Tumor vol. Mean Range Nodal involvement (%) Residual tumor (%) Positive margin Gross residual Note: Characteristics
IIa
Surgery
Surgery + RT
33
11
59 45-15 14/19 27
54 31-72 615 9
Bolus 5-FU 8 61 51-73 414 6
Surgery + 20 58 23-69 12/8 16
2 0
2 0
4 0
61
69
38
78
l-384 45
8-125 54
8-105 15
2-514 70
21
21
25
15
0
9
0
10
of the patient included in each treatment group.
related morbidity Group
Bleeding Anastamotic leak Infection Protracted ileus Chylous ascites Jaundice
III
6 4
patient because of postoperative clinical course, and radiation prescription was not altered in any patient because of treatment tolerance. One patient refused the second infusion of 5-FU due to mucositis during the first course, and this was the only chemotherapy dose adjustment. Although many patients had anorexia and were limited in their oral intake during treatment, malnutrition was an infrequent problem as 33 of the 39 patients in groups 2 and 3 had jejunostomies at the beginning of treatment. More than onethird of the patients had feeding jejunostomies removed during the course of treatment. Patterns of recurrence are listed in Table 3. The four patients in group 1 dying in the early postoperative period were excluded from this analysis. Physicians signing death certificates listing pancreatic carcinoma as a cause of death were contacted to determine the extent of disease at the
Table 2. Treatment
IIIb Surgery + RT
I
II
III
5 (3) 2 (1) 3 2 1 0
1 0 1 2 0 1
0 0 3 0 0 0
Note: Treatment related morbidity of the patients in each treatment group.
time of death. For three patients in group 1 and two patients in group 2, more specific information could not be obtained. These patients were also excluded from the analysis of local control. Among the 26 evaluable patients in group 1, 22 of the 23 patients recurring had a local component of tumor recurrence. Postoperative irradiation was associated with a reduced risk of local recurrence (53%). There was no significant effect of adjuvant bolus 5-FU on the risk of local recurrence (3 of 8 in group 2b vs 6 of 9 in group 2a). Patients treated with chemosensitized irradiation had the lowest risk of local recurrence (25%). In the group of patients treated with chemosensitized irradiation, one of the recurrences occurred in a patient with gross residual tumor following surgery, and three recurrences occurred in the common bile duct at the margin of the field. Sites of metastatic disease are also listed in Table 3. The distribution of metastatic sites was similar in all treatment groups Q-not significant for all comparisons). The survival of patients is shown in Figure 1. Median survival was similar in all three groups (group 1-15 mo, group 2-15 mo, group 3-16 mo), although there was an improvement in 3-year survival with chemosensitized irradiation (34% in group 3 vs 8% and 5% in groups 1 and 2). Eight of the 20 patients in group 3 are alive and diseasefree following chemosensitized irradiation with follow-up ranging from 13 to 41 months (median 27 months). Three patients in group 1 are alive and disease free at 10, 13, and 30 months, and three patients are alive in group 2 at 11, 21, and 40 months. The latest recurrence yet observed was 30 months following diagnosis.
I. .I. Radiation Oncology 0 Biology 0 Physics
1140
October 1991, Volume 21, Number 5
Table 3. Patterns on recurrence Group
Evaluable patients Sites of recurrence Pancreatic bed Liver Peritoneal seeding GI tract Not specified Abdominal wall/ incision Other
I
II
III
29 22
19 9
20 5
6 6 4 3 3
8 4 0 2 1
5 3 0 0
Bone- 1 Supraclav LN- 1
Lung- 1 Axilla- I
4
4
No failure
I Lung- 1
Bone-2 8
Sites of tumor recurrence following curative treatment of pancreatic carcinoma.
Patients with complete resections and negative surgical margins were examined seperately, and their survival is plotted in Figure 2. Survival is substantially better in this group of patients with the greatest effect in those patients receiving adjuvant chemosensitized radiation. Median survival in group 3 was 29 months compared to 21 months in group 1 and 16 months in group 2. Median survival was longer in patients receiving postoperative treatment even if they had positive margins (9.5 mos-group 1, 14.5 mosgroup 2, 14 months-group 3), but there are no patients currently alive in this group, and only two patients survived longer than 18 months. Since patients were not randomized to treatment groups, strict criteria to test for statistical significance cannot be fulfilled. Statistical calculations are reported only to suggest trends in the results. A log rank comparison showed a modest improvement in late survival with chemosensitized radiation when all patients were analyzed (group 3 vs group 1 p > 0.1, group 3 vs group 2 p > 0.09). A comparison
90 L:
z
80
5
70
;
Cd 50
$
40
2
30
$
20 10 0
0
4
8
I2
16
20
2.1
2x
32
36
40
MONTHS
Fig. 1. Survival of all patients following carcinoma.
resection
of pancreatic
Fig. 2. Suvival of patients following creatic carcinoma.
complete
resection of pan-
of patients undergoing complete resection with fewer patients showed a similar trend (group 1 vs group 3 p > 0.06, group 2 vs group 3 p > 0.14).
DISCUSSION Pancreatic carcinoma presents a therapeutic challenge to the surgeon. The location of the pancreas will preclude resection of many small tumors because of local extension to critical vascular structures such as portal vein, celiac vessels, and superior mesenteric artery. When the tumor is resected it is equally difficult to obtain adequate radial margins because limited extrapancreatic extension to vascular structures as well as the pancreatic remnant, liver, and the biliary tree are difficult to mobilize and dissect. Similarly, these structures complicate efforts to perform regional node dissections. All of these factors raise the risk of local and regional recurrence in this disease. Efforts to extend the operative procedure to obtain wider margins and to improve the survival, thus far, have failed to improve survival. Sensitive tissues such as kidney and liver may also lie in the volume traversed by an external radiation field, creating pressure to limit the size of the target volume. Small bowel will fill the volume previously occupied by the pancreatic tumor, and it is not feasible to perform procedures to displace small bowel as in the pelvis; therefore, it is also necessary to limit total dose. Previous attempts to improve the results following surgery include efforts directed at preoperative radiation as described by Kopelson (4) or postoperative treatment described by Rich (8) and Nguyen et al. (7). While these studies demonstrate the feasibility of moderate dose postoperative irradiation, they fail to demonstrate a significant survival benefit. Kopelson treated five patients preoperatively with 4045 Gy directed at the tumor. Four of the five patients had prior surgeries and one had an endoscopic biopsy. Three of the five patients had significant regression noted at the sec-
Adjuvant therapy of pancreatic carcinoma 0 R. WHIITNGTONet al.
@I
64
Fig. 3. Typical treatment portal of patient in group 3 demonstrating (a_ante& field, b: lateral field). _
ond surgery and all were resected following irradiation. Nodes were negative in all patients. One patient had a marginal recurrence 7 years following treatment and two patients are alive, one at 15 months and the other more than 8 years. Nguyen et al. treated nine patients following resection to a dose of 60 Gy using small fields. Maximum field dimensions were 15 cm length, 15 cm anterior width, and 10 cm lateral width. Median survival was 13 months. Patients treated at the University of Pennsylvania had tumors that were too large to be treated with fields that small. The median volume was 67 cm3 in group 1, 55 cm3 in group 2, and 78 cm3 in group 3. Rich treated 10 patients with a high risk of recurrence following resection. Two patients had gross residual tumor, two had microscopic residual tumor, two had perineural invasion, and three had positive nodes. Local recurrence occurred in only 1 of 10 patients following doses of 40 to 45 Gy. The two patients with gross residual also received intraoperative orthovoltage boost of 12.5 Gy One patients survived more than 8 years and two are alive without disease less than 5 years. There are two rationales for delivering chemotherapy and irradiation concurrently. The first is to apply early treatment to both the tumor bed and to micrometastases when the efficacy of each is increased. This is unlikely to affect survival since there are no data to support the use of adjuvant chemotherapy alone following pancreatic resection. A second rationale is that the chemotherapy-radiation
1141
target volume and blocking of critical structures
interaction renders residual tumor cells more sensitive to the effects of treatment. Vokes and Weichselbaum have reviewed data suggesting greater cytotoxicity of radiation in the presence of 5-FU and further enhancement with protracted exposure to chemotherapy (10). They also review data suggesting enhanced activation of mitomycin in hypoxic cells where radiation effects are reduced. The effects of mitomycin may be additive with radiation. CT-based treatment planning has allowed for the treatment of much larger volumes with acceptable morbidity. An example of the treatment fields used in one patient are demonstrated in Figure 3. The volume includes the primary tumor volume as well as the adjacent nodal and vascular structures. Because of the variable position of the kidneys and liver, it is necessary to use CT scans to create beam’s-eye-view projections to facilitate field shaping. It is frequently necessary to change the field weights based on the location of the target. Patients with tumors in the body and tail region were generally treated with 60% of the dose delivered through the lateral fields, whereas those patients with smaller lesions in the head of the pancreas were treated with 60% of the dose delivered through the anterior and posterior fields. In comparing the results of this population with the previously reported series from the GITSG, the patients treated in this series had more factors associated with a poorer prognosis. Twenty-eight percent of the patients in the GITSG trial had positive nodes compared to 45% in
1142
I. J. Radiation
Oncology
0 Biology 0 Physics
group 1, 63% of patients in group 2, and 70% of patients in group 3 in this series. Four percent of the patients in the GITSG trial had lesions located in the body and tail compared with 12% of patients in group 1 and 20% of patients in groups 2 and 3. Postoperative treatment appeared to reduce the risk of local recurrence significantly among patients following surgery. The observed incidence of local recurrence among patients in group 1 is substantially higher than was observed in patients on the GITSG trial (85% vs 33%) but it is not clear whether this represents different surgical results, closer surveillance for local recurrence, or more advanced lesions. Recurrence rates among patients in group 2 and 3 are significantly lower than observed in group 1, 53% in group 2, and 25% in group 3 (p < .OOOl group 1 vs 2 vs 3). While there is a trend toward a lower risk of local recurrence in group 3, it does not reach significance (group 2 vs group 3-p < 0.1). There is no significant difference in the frequency or site of distant metastases in the three treatment groups (p > 0.2 for all comparisons). One could therefore infer that the improved survival observed in these patients is related to the enhanced local control that can be achieved with chemosensitized radiation therapy. Nutritional support of patients during the course of treatment was best accomplished with an enteral feeding tube placed at the time of surgery. Although quantitative data were not prospectively obtained, it was a subjective observation of all observers that nutritional maintenance was substantially improved with the availability of a feeding tube until the patient’s tolerance of the GI effects of radiation could be determined.
October
1991, Volume 21, Number 5
Adjuvant radiation therapy can be delivered safely to patients as part of a planned combined modality approach to resected pancreatic carcinoma. In this group of patients the addition of chemosensitization enhanced the results, and that patient tolerance was acceptable. Among patients with incomplete resections, the mediam survival was prolonged, although there were no long term survivors with postoperative treatment. Among patients with complete resections, there is a substantial improvement in 2-year and 3-year survival with the addition of chemosensitized radiation. With the improvement in local control observed in these patients with the adjuvant treatment, the focus for future treatment efforts must concentrate on efforts to reduce the risk of systemic metastases. Because of the relative low frequency of complete resections and the limited experience of single institutions, this disease may not be amenable to further randomized studies. In an effort to increase the potential for resection, a current research effort at the Fox Chase Cancer Center delivers chemosensitized irradiation prior to attempted resection (11). Previous reports have suggested that more extensive manipulation of the tumor may increase the risk of distant metastases, and preoperative irradiation may also reduce the risk of distant failure (12). An NCI-sponsored research protocol attempted to reduce the risk of metastases with the further intensification of the GITSG regimen by delivering one cycle of streptozotocin, mitomycin, and 5-FU (SMF) prior to irradiation and a second cycle following irradiation. That trial was terminated because of poor accrual. Future efforts to address subclinical micrometastatic disease will be needed to further improve the survival of these patients.
REFERENCES 1. Gastrointestinal Tumor Study Group. Further evidence of effective adjuvant combined radiation and chemotherapy following curative resection of pancreatic cancer. Cancer 59: 20062010; 1987. 2. Gudjonsson, B. Cancer of the pancreas: 50 years of surgery. Cancer 60:2284-2403; 1987. 3. Kalser, M. H.; Ellenberg S. S. Pancreatic cancer: adjuvant combined radiation and chemotherapy following curative resection. Arch. Surg. 120899-903; 1985. G. Curative surgery for adenocarcinoma of the 4. Kopelson, pancreas/ampulla of vater: the role of adjuvant pre- or postoperative radiation therapy. Int. J. Radiat. Oncol. Biol. Phys. 9:91 l-915; 1983. 5. Michelassi, F.; Erroi, F.; Dawson, P. J.; Pietrabissa, A.; Noda, S.; Handcock, M.; Block, G. E. Experience with 647 consecutive tumors of the duodenum, ampulla, head of the pancreas, and distal common bile duct. Ann. Surg. 210:544553; 1989. 6. Moertel, C. G.; Frytak, S.; Hahn, R. G.; O’Connell, M. J.; Reitmeier, R. J.; Rubin, J.; Schutt, A. J.; Weiland, L. H.; Childs, D. S.; Holbrook, M. A.; Lavin, P. T.; Livstone, E.; Spiro, H.; Knowlton, A.; Kalser, M.; Barkin, J.; Lessner,
H.; Mann-Kaplan, R.; Ramming, K.; Douglas, H. 0.; Thomas, P.; Nave, H.; Bateman, J.; Lokich, J.; Brooks, J.; Chaffey, J.; Corson, J. M.; Zamcheck, N.; Novak, J. W. Therapy of locally unresectable pancreatic carcinoma: a randomized comparison of high dose (6000 rads) radiation alone, moderate dose radiation (4000 rads + 5_fluorouracil), and high dose radiation + 5-fluorouracil. Cancer 48:17051710; 1981.
7. Nguyen, T. D.; Bugat, R.; Combes, P. F. Postoperative irradiation of carcinoma of the head of the pancreas area: short-time tolerance and results to precision high dose technique in 18 patients. Cancer 50:53-56; 1982. 8. Rich, T. A. Radiation therapy for pancreatic cancer: elevenyear experience at the JCRT. Int. J. Radiat. Oncol. Biol. Phys. 11:759-763; 1985. 9. Tepper, J.; Nardi, G.; Suit, H. Carcinoma of the pancreas: Review of the MGH experience from 1963 to 1973. Cancer 37:1519-1524; 1976. 10. Vokes, E. E.; Weichselbaum R. R. Concomitent chemoradiotherapy: rationale and clinical experience in patients with solid tumors. J. Clin. Oncol. 8:911-934; 1990.
Adjuvant therapy of pancreatic carcinoma 0 R. 11. Weese, J. L.; Nussbaum, M. L.; Paul A. R.; Solin L. J.; Kowalyshyn, M. J.; Hoffamn J. P. Enhanced resectability of pancreatic and periampullary cancer after neoadjuvant chemoradiotherapy. Int. J. Pancreatology 7: 177-185; 1990.
WHITTINGTONet al.
1143
12. Whittington, R.; Solin, L.; Mohiuddin, M.; Cantor, R. I.; Rosato, F. E.; Biermann, W. A.; Weiss, S. M.; Pajak, T. F. Multimodality therapy of localized unresectable pancreatic adenocarcinoma. Cancer 54: 1991-1998; 1984.
