All patients underwent distal pancreatectomy for adenocarcinoma of the body or the tail of the pancreas. shows baseline patient, surgical, and pathologic tumor characteristics for the entire cohort and stratified by treatment group. The median age at the time of surgery was 68 (range, 26–92) years. Forty-six percent of patients were men, and 87% were white. Severe comorbid disease was present in 28% of patients at the time of presentation. Ninety-eight percent underwent splenectomy and 16% underwent bowel resection. The median length of hospital stay was 7 (range, 5–110) days. Those who did and did not receive adjuvant therapy were similar with respect to known poor prognostic factors. The only statistically significant difference between the adjuvant CRT and no CRT group was that more patients in the adjuvant CRT group were white (p = 0.02). Tumor size was >3 cm in 76% of patients. Positive margins were identified in 29% of patients and 55% had lymph node metastases. Because the American Joint Committee on Cancer (AJCC) staging manual was updated in 2002, staging before this date was updated for consistency.
Baseline patient, surgical, and pathologic tumor characteristics for the entire cohort stratified by treatment group
Overall median survival for the entire cohort was 16.2 (95% CI, 13.1–18.9) months. One-year overall survival was 62% and 2-year overall survival was 31%. The influence of patient- and treatment-specific factors on survival is presented in . Age and race did not significantly affect survival in our analysis. Female gender was associated with a significantly better overall survival (p = 0.05) as was severe comorbid disease (p = 0.04).
Univariate analyses of the associations between overall survival and patient characteristics
There was no statistically significant difference in median, 1-, or 2-year overall survival for patients treated with adjuvant CRT compared with surgery alone (16.7 vs. 12.1 months; 66% vs. 50%, 29% vs. 32%, p = 0.23), respectively (). After adjusting for age, gender, race, node, and margin status, there remained no significant difference in overall survival between patients who received adjuvant CRT and those treated with surgery alone (hazards ratio (HR) = 0.61; 95% CI, 0.34–1.08; p = 0.09).
Kaplan–Meier curves comparing overall survival in patients who received adjuvant radiation therapy versus distal pancreatectomy alone. A total of five patients in the adjuvant chemoradiation therapy group were censored beyond 85 months
To determine whether our results varied based on location of adjuvant therapy (JHH vs. elsewhere), we performed two subgroup analyses. First, we compared the characteristics and survival for patients who received adjuvant therapy at JHH and elsewhere. Some differences in patient characteristics were observed between those treated with adjuvant therapy at JHH versus elsewhere (node positive: 46% vs. 54%, p = 0.61; margin positive: 12% vs. 39%, p = 0.02; and age >65 years: 41% vs. 64%, p = 0.09). On univariate analysis, there was a significant survival benefit for patients treated at JHH versus elsewhere (21.5 vs. 14.5 months; HR, 0.46; 95% CI, 0.26–0.82; p < 0.01); however, this benefit was not significant after adjusting for age, gender, race, and nodal and margin status (HR, 0.52; 95% CI, 0.26–1.02; p = 0.06). Second, we attempted to replicate our primary results using only individuals who received adjuvant CRT at JHH (excluded elsewhere CRT). In this case, there was a significant increase in survival with adjuvant CRT at JHH vs. observation (21.5 vs. 12.1 months; HR, 0.51; 95% CI, 0.27–0.97; p = 0.04); however, this association was not significant after adjusting for age, gender, race, and nodal and margin status (HR, 0.59; 95% CI, 0.28–1.23; p = 0.16). The effects on survival observed for age, gender, race, and margin status and the interaction between nodal status and treatment were similar to those obtained using the entire cohort with a slight loss of power as would be expected due to the reduction in sample size. Therefore, we thought it was reasonable to combine patients who received adjuvant therapy at JHH and elsewhere in the final analysis.
Although the study was not designed to detect interactions between therapy type and risk factors, exploratory analyses were performed. displays the median overall survival for patients receiving adjuvant CRT versus surgery alone by risk factors. There is a significant interaction between node status and treatment (p < 0.01) indicating that patients respond differently to therapy depending upon nodal status. shows the overall survival for patients in the CRT versus surgery alone arms for node-positive individuals. Both alone and after adjusting for age, gender and margin status, there was a significant improvement in overall survival for node-positive patients treated with adjuvant CRT versus surgery alone (HR, 0.28; 95% CI, 0.14–0.58; p < 0.01; and HR, 0.23; 95% CI, 0.1–0.52; p < 0.01, respectively). For node-negative individuals, the sample size was extremely small (N = 8 for the surgery alone and N = 31 adjuvant chemoradiation therapy). In fact, only three deaths were observed from eight subjects in the surgery alone group. Nevertheless, univariate analysis indicated that there might be a survival benefit for the surgery alone arm (HR, 3.0; 95% CI, 0.9–10; p = 0.06). The results above were nearly identical when the analysis was restricted to JHH patients (results not shown).
Comparison of the overall survival between treatment groups stratified by patient characteristics
Kaplan–Meier curves comparing treatment groups in node-positive patients. A total of three patients with node-positive disease in the adjuvant chemoradiation therapy group were censored at time points beyond 85 months
The data were missing for a number of individuals for several key variables (e.g., 16 for tumor diameter, 7 for node and margin status). Multiple imputation with ten replicates was used to determine the impact of the missing data on the estimated effects. These analyses produced estimates that were nearly identical to those found using the observed data (results not shown).
Data on patterns of failure were available for 35 patients (10 surgery and 24 adjuvant therapy). There was no significant difference in time to relapse between the surgery alone and CRT groups, respectively (median, 4.8 vs. 9.8 months; p
= 0.08). Of all patients with recorded recurrences, 12 were local, 21 were distant, and 1 patient had both a local and distant recurrence. Of patients with distant recurrences the majority were liver,10
one was at an unknown site, and ten were defined as other, including peritoneal involvement, omental seeding, or lung metastasis. Stratifying by CRT, 6 of 10 (60%) failures occurred in the surgery only group versus 6 of 24 (25%) for the adjuvant CRT group.