The inclusion criteria for neoadjuvant therapy were: (1) adenocarcinoma located in the head or uncinate process of the pancreas or in the ampullary region, with a diameter of less than 5 cm and without evidence of local or metastatic spread of the disease, (2) WHO performance status less than 2, (3) good co-operation, and (4) no other malignancy or serious illness. The radiotherapy dose was constant whereas gemcitabine was given as a 30-min infusion twice weekly before irradiation at three dose levels, which were 20, 50 and 100 mg/m². The targeted irradiation volume included the tumor, possible surrounding oedema, and 1 cm margin. The tumor radiation dose in planning target volume was 50.4 Gy (ICRU). The therapy was given in 28 fractions of 1.8 Gy per day, five days per week. The median number of gemcitabine-cycles was 10 and everyone received the whole planned radiotherapy. During the study the maximal tolerated dose of gemcitabine was found to be 50 mg/m². Dose reduction of gemcitabine was needed in 65% of patients. Toxicity and tolerance of concomitant gemcitabine and radiotherapy treatment in these patients has been reported previously[12].

Following CRT, patients were given a 4-wk break for recovery of blood counts and nutrition, and new staging CT and MRI scans were performed.

The CT or MRI criteria for unresecability of tumor were metastatic disease or tumor encircling the superior mesenteric artery. Portal vein resection and reconstruction due to tumor invasion was performed when needed.

Patients underwent pancreaticoduodenectomy without neoadjuvant therapy. Preoperative CT and MRI scans were performed, and a biliary stent was inserted if the patient was jaundiced.

In both groups, the operation was begun with a diagnostic laparoscopy to rule out metastatic or locally advanced disease. Radical pancreaticoduodenectomy was performed with extended lymphadenectomy and removal of retroperitoneal tissue. In the case of tumor infiltration into the portal vein, resection of the vein and reconstruction with autologous venal graft was performed. All operations were performed by one of two surgeons, however, the vast majority of the patients were operated on by surgeon T.K.

Frozen sections of the resection margins of the pancreas and common hepatic duct were collected during operations. All specimens were delivered as fresh, the resection margins were stained and the specimen was fixed in formalin for 1-2 d. The entire tumor area was sectioned in 2-3 mm slides. One macro slide was taken from the largest macroscopically identified tumor area. Lymph nodes and resection margins were also sectioned. All tissues were embedded in paraffin and stained according to the Herovici-van Gieson and Alcian blue-PAS methods. One pathologist (P.K.) examined all specimens. The tumor stage was defined according to the AJCC 2002[14].

After radical surgery, the patients were followed using a routine follow-up program. Liver tests and CA19-9 were measured. Radiological examinations were performed only if the patients developed symptoms, such as pain or ascites, or elevation of the CA19-9 value was observed. In the case of recurrence of the disease, an oncologist was consulted and the patient was offered chemotherapy.

Fishers exact test (F) was used to test differences between categorical variables and Student’s t-test or the Wilcoxon-Mann-Whitney test (W) was used to assess differences between means of continuous variables. Normality of continuous variables was tested with the Kolmogorov-Smirnov test. Kaplan-Meier analyses and log rank tests were used to assess survival. This study was designed to detect 25 mo increase in median survival time, assuming 20 patients in each group, power of 0.80, significance level of 0.05 and median survival time of 15 mo in the surgically treated only group. Accrual time was 48 mo and follow-up time after the end of the recruitment was 48 mo. P < 0.05 was considered statistically significant (NS: not significant).

Of the 47 patients enrolled in the study, 22 (9 females 13 males) belonged to the CRR group. Their mean age was 65 (49-83) years. 25 patients (13 females and 12 males) belonged to the RO group. Their mean age was 63 (43-76) years. There were no differences between the groups regarding age and sex, status of lymph nodes, tumor stage or tumor size. After chemoradiotherapy, 8 patients were judged as inoperable for laparotomy: 4 patients had metastases to the liver, 2 patients had a poor general condition and 1 patient had a fractured femur and died from a pulmonary embolism before the operation. In addition, one patient refused the operation. All these patients were excluded from the series.

The CRR group included 15 ductal adenocarcinomas, 4 ampullary adenocarcinomas, 2 cystadenocarcinomas and one invasive malignant intraductal papillary mucinous tumor (IPMT) of the pancreas. Thirteen patients with ductal adenocarcinoma were treated by pancreticoduodenectomy. For two patients, total pancreatectomy was performed due to positive resection margins after partial resection. One portal resection was performed due to tumor invasion to the portal vein. The two patients with cystadenocarcinoma were also treated with pancreaticoduodenectomy. In the patient with IPMT, total pancreaticoduodenectomy was performed. Ampullary tumors were treated with pancreaticoduodenectomy.

The RO group included 14 patients with ductal adenocarcinomas, 4 with ampullary adenocarcinomas, 4 with cystadenocarcinomas and 3 with invasive malignant IPMTs. All patients with ductal adenocarcinomas and ampullary tumors were treated with pancreaticoduodenectomy. One portal resection was performed due to tumor invasion to the portal vein. One patient with cystadenocarcinoma was treated with pancreaticoduodenectomy, and three patients with total pancreatectomy. One patient with IPMT underwent pancreaticoduodenectomy. However, after investigation of the paraffin specimens, an intraluminal noninvasive component was found close to the resection border and the operation was completed with total pancreatectomy six months later. The two remaining patients with IPMT were treated with pancreaticoduodenectomy. Table 1 gives the distribution of the stages of tumors for both groups.

One patient in the CRR group died of multiorgan failure on the 43^rd postoperative day (overall hospital mortality rate 2%). In this patient, duodenal perforation with abscess formation due to a biliary stent was found in the primary operation. The patient was reoperated on the same day because of bleeding, and again on the 10^th postoperative day because of leakage of the gastrojejunal anastomosis and peritonitis.

The overall morbidity rate was 38%. In the CRR group, ten patients (45%) had complications. There were five cases of intra-abdominal infections, which were treated conservatively, and one case of pneumonia. Intra-abdominal infection was recorded if the CRP had increased and the broad spectrum antibiotic had been changed after the 5^th postoperative day and pneumonia was excluded. Two patients suffered from ileus. The central venous catheter of one patient was accidentally inserted into the pleural space. In the RO group, eight patients (32%) had complications. Two patients were reoperated because of bleeding with gastrointestinal anastomosis. Three patients had minor wound infections and one had pneumonia. In one patient, rupture of the intima of the common hepatic artery was suspected due to lack of pulse and immediate reconstruction with venous graft was performed. One patient suffered a stroke during the postoperative convalescence period. Information on complications is presented in Table 2.

The length of hospital stay was 15 ± 1.6 d (mean +SE) in the CRR group and 11.8 ± 0.9 d (NS) in the RO group. Blood loss during operation was 2200 ± 300 mL (range 700-4300 mL) in the CRR group and 2400 ± 400 mL (range 300-8500 mL; NS) in the RO group.

Survival rates were calculated for all types of pancreatic cancers and for the subgroups of ductal adenocarcinoma. The median survival time was 30.2 mo (95% CI 25.46-34.94) in the CRR group (n = 22), and 35.9 mo (95% CI 10.51-61.29) in the RO group (n = 25). No statistical differences were found by log Rank analyses. 1-y survival was 81% ± 8.6 (SE) in the CRR group and 72% ± 9.0 in the RO group, 3-year survival rates were 33 ± 10 and 47% ± 10 and 5 year survival rates were 33% ± 10.3 and 23% ± 9.0, respectively (Figure 1A).

Open in New Tab   Full Size Figure   Download Figure

Figure 1 A: Survival rates of all pancreatic cancers; B: Survival rates of ductal adenocarcinomas. CRR: chemoradiation and resection; RO: resection only.

The median survival time of LN negative patients in the CRR group (n = 15) was 30.2 mo (95% CI 17.98-42.42), and in the RO group (n = 14) 37.4 mo (95% CI 7.79-67). The respective values of LN positive patients in the CRR group (n = 7) were 27.0 mo (95% CI 0.0-55.2), and in LN positive patients in the RO group (n = 11) 20.7 mo (95% CI 6.75-34.65). The 1-year survival rate for LN negative patients was 79% ± 11 in the CRR group and 79% ± 11.0 in the RO group, 3-year survival rates were 43% ± 13.2 and 57% ± 13.2, and 5-year survival rates were 43% ± 13.2 and 41% ± 13.6, respectively. In LN positive patients, the 1-year survival rate was 86% ± 13.2 in the CRR group and 64% ± 14.5 in the RO group, 3-year survival rates were 14% ± 13.2 and 33% ± 15 and 5-year survival rates were and 14% ± 13.2 and 0%, respectively. There were no significant differences between the groups (Figure 2). Median survival times for different cancers are given in Table 3.

Open in New Tab   Full Size Figure   Download Figure

Figure 2 Survival rates of pancreatic cancers. A: with nodal involvement; B: without nodal involvement. CRR: chemoradiation and resection; RO: resection only; LN: lymph node.

When the CRR and RO groups were divided according T classes (T1, T2, and T3), no statistical differences were revealed between the groups (data not shown).

The survival calculations were also made according to tumor size, tumor stage and perineural invasion, but no statistically significant differences between the CRR and RO groups were found in these subgroups.

The median survival time was 27.00 mo (95% CI 5.55-48.45) in the CRR group (n = 15) and 20.20 (95% CI 0-52) in the RO group (n = 14, Table 3). The 1-year survival rate was 78.6% ± 11.0 in the CCR group and 64.3 % ± 12.8 in the RO group. The respective 3-year survival rates were 21.4% ± 11.0 and 50.0% ± 13.4 and 5-year survival rates were 21.4% ± 11 and 14.3 % ± 9.4. No statistical differences were found between the groups. (Figure 1B).

No difference was found between the CRR and RO groups regarding lymph node status. The median survival time of LN negative patients in the CRR group (n = 8) was 30.2 mo (95% CI 9.94-50.46) and in the RO group (n = 9) 20.2 mo (95% CI 19.62-20.78) (Figure 2B). The median survival time of LN positive patients in the CRR group (n = 7) was 27.0 mo (95% CI 0.00-55.23) and in the RO group (n = 5) the median survival time was 38.8 mo (95% CI 0.00-110.51). The 1-year survival rates in LN negative patients were 71% in the CRR group and 67% in the RO group, 3-year survival rates were 29% and 44% and 5-year survival rates were 29% and 22%, respectively. In LN positive patients, the rates were 86% vs 60%, 14% vs 60% and 14% vs 0%, respectively. There were no statistical differences between the groups (Figure 3).

Open in New Tab   Full Size Figure   Download Figure

Figure 3 Survival rates of ductal adenocarcinomas. A: with nodal involvement; B: without nodal involvement. CRR: chemoradiation and resection; RO: resection only; LN: lymph node.

Survival calculations were also made according to tumor size, perineural invasion and tumor stage, but no statistical differences were found between the CRR and RO groups for any subgroup.

At the end of the follow up, two patients with ductal adenocarcinoma in the CRR group were disease free. Their follow up times were 71 and 57 mo and tumor stages were T1N0M0 and T3N0M0, respectively. One patient (T3N1M0) was followed 71 mo and recurrence of disease was detected two months before the end of follow-up. Two of the 14 patients with ductal adenocarcinoma in the RO group were alive, but only one was disease free (follow up 53 mo, T2N0M0). The other (T3N0M0) had local recurrence at 46 mo with the total follow up time being 61 mo.

Patients with adenocarcinoma originating from the ampullary region had a median survival time of 39.4 mo (95% CI 5.95-72.85). In this group, the one year survival rate was 88%, the three year survival rate was 60% and the five year survival rate was 45%. In patients with cystadenocarcinoma (n = 6), the median survival time was 20.70 (95% CI 0.0-48.2), and the one year survival rate was 83%, the three year survival rate was 33%, and the five year survival rate was 33%.