Abstract
Background: Periampullary carcinomas generally confer a poor outcome. Choosing the most effective treatment regimen for each sub-entity proves challenging and is usually based on experience from pancreatic adenocarcinoma (PDAC). Patients and Methods: The long-term follow-up is presented of 472 patients with periampullary tumors [PDAC, distal cholangiocarcinoma (dCC) and ampullary carcinomas (AC)] who underwent radical resection considering clinical characteristics, paraclinical findings and histopathological features in order to define factors of prognostic relevance. Results: Patients with PDACs presented with larger tumor sizes, more frequent R1 resection, higher rate of nodal and perineural invasion, higher tumor stage according to the classification of tumors of the Union Internationale contre le Cancer when compared to those with dCCs and ACs. In a multivariate analysis, age >65 years, postoperative complications and higher grading of the tumor proved to be independent prognostic factors for survival. Conclusion: Patients suffering from PDAC have the worst prognosis and greatest benefit from radical resection of all patients with periampullary tumors. More detailed studies are warranted to better distinguish between the different entities.
The family of periampullary carcinomas comprises adenocarcinomas originating from the pancreatic head (PDAC), the distal bile duct (dCC), the periampullary duodenum and carcinomas of the ampulla of Vater (AC; Figure 1) (1). Due to their close anatomic vicinity and molecular similarities, the final diagnosis sometimes remains undetermined (2). Therefore, choosing the most effective therapeutic regimen for each given entity proves problematic. However, the clinical significance of histopathological patterns such as lymph-node (LN) and perineural or lymphatic vessel involvement may differ among periampullary tumor entities and can notably influence the clinical course.
The most frequent carcinomas of the periampullary region are PDACs, accounting for up to 70% of all tumors, whilst dCC and AC make up 10-20% and duodenal carcinomas 3-7% (3, 4). While PDACs are associated with the poorest outcome, with 5-year survival rates of 15-20% after resection with curative intent (3-6), rates of dCC and AC are considerably better at 20-30% and 40%, respectively (4, 7-9).
The most relevant histopathological subclassification of periampullary carcinomas distinguishes between an intestinal type and a more aggressive pancreaticobiliary type (10, 11). Sequencing attempts further evaluated distinct genomic signatures in (peri-) ampullary cancer and revealed various differences, at least between the intestinal and pancreatobiliary types (12, 13). However, molecular signatures which distinguish between periampullary entities have not yet been evaluated. Therefore, we still have to rely on pathological assessments and their anatomical relationship with the biliopancreatic duct system.
In contrast to the adjuvant treatment of PDACs, no standard regimen for AC and dCC has been established due to a lack of data (14). The only curative treatment of all tumor sub-entities is based on radical resection of the entire tumor burden (R0). Recently, neoadjuvant therapy approaches have been implemented into clinical practice for PDAC. Neoadjuvant chemotherapy can be taken into account for patients presenting with borderline resectable tumors or for patients with a resectable tumor but additional high-risk features (15). Current regimens include FOLFIRINOX (folinic acid, fluorouracil, irinotecan hydrochloride and oxaliplatin), and the combination of gemcitabine and nab-paclitaxel (16). For dCC and AC, data on the efficacy of neoadjuvant treatment are lacking.
Anatomical location and structure of the periampullary region including the distal bile duct (#1), the pancreatic duct (#2) and the periampullary region (#3).
In this retrospective study, the data of 472 patients suffering from periampullary carcinomas who underwent radical resection were analyzed considering clinical characteristics, paraclinical findings and histopathological features in order to define factors of prognostic relevance. The primary purpose of this study was to analyze variables influencing perioperative morbidity and long-term survival of the different study cohorts in periampullary carcinoma. Defining patients with the greatest benefit from surgery would have substantial impact on the quality of treatment.
Patients and Methods
Patients and data acquisition. Patients who underwent resection of a tumor in close proximity (up to 2 cm) to the ampulla of Vater were included in the study. Clinical and paraclinical data of patient, treatment and tumor characteristics were collected in a database. Procedures were in accordance with the ethical standards of the responsible committee on human experimentation, following approval from the Institutional Review Board (EA1/292, EA2/035) and in accordance with the Helsinki Declaration of 1975.
Preoperative assessment. Preoperative evaluation included laboratory values, medical history, imaging results, previous treatment, cardiac and pulmonary function tests and anesthesia evaluation. Diagnosis and preoperative staging of periampullary cancer was based on computed tomography (CT) or magnetic resonance imaging (MRI), sometimes accompanied by endoscopic retrograde cholangiopancreatography, diagnostic laparoscopy and fluorodeoxyglucose positron-emission tomography. Treatment was decided by a multidisciplinary tumor board. None of the patients received neoadjuvant treatment.
Surgical technique and postoperative treatment. All resections were performed as pylorus-preserving pancreaticoduodenectomy (PPPD) or Kausch-Whipple procedure as described elsewhere (17, 18). The anastomosis of the pancreatic remnant was performed either as pancreatogastrostomy or pancreatojejunostomy. The biliodigestive continuity was reconstructed by an anastomosis of the common bile duct and a jejunal loop with a Roux-en-Y gastrojejunostomy. Adjuvant treatment was decided by multidisciplinary institutional tumor board and initiated within 6 weeks following operation.
Postoperative assessment. Postoperative survival was defined as starting from the day of resection of the tumor to the day of last follow-up, and postoperative mortality as death within 30 days after the operation. Postoperative morbidity was defined as occurrence of complications (including anastomotic leak, pancreatic fistula, bleeding, intra-abdominal infection, and organ failure) within 30 days after the operation. Postoperative fistulas were classified according to Bassi et al. (19).
Histological evaluation. Histopathological assessment was carried out by a senior physician specialized in pathology of the pancreas and bile duct. Tumors were classified according to the TNM classification of the Union Internationale contre le Cancer (UICC) classification of tumors (20).
Statistical analysis. For statistical analysis, SPSS (version 16 and 22; IBM Corp., Armonk, New York, USA) was used. The distribution of variables was arithmetically averaged. In the case of missing values, the relative frequency referred to the data available unless indicated otherwise. The D'Agostino and Pearson omnibus normality test was used to test for normal distribution of data. When two variables (numeric and categorical or numeric and numeric) between two groups were compared, Student's t-test was applied for normally distributed data and the Mann–Whitney U-test in the case of non-normally distributed data. For comparison of two categorical variables, the Chi-square test was applied. When comparing more than two groups, analysis of variance (ANOVA) was implemented for normally distributed data and Kruskal–Wallis for not normally distributed data. To describe cumulative survival, the Kaplan–Meier method was used. Paired comparison between the tumor entities was calculated using Bonferroni corrected log-rank (Mantel-Cox) test. Cut-off levels were identified using the Youden Index, where the maximum value of the index was used to identify the ideal cut-off point based on receiver operating characteristics curve (ROC). Sensitivity and specificity were calculated with a confusion matrix of binary classification. A logistic regression model was applied to evaluate the effect of positive LNs on survival. A p-value of 0.05 or less was considered to be significant. Variables showing a significant effect on survival were included in multivariate analysis.
Operation characteristics of patients suffering from adenocarcinomas originating from the pancreatic head (PDAC), the distal bile duct (dCC), and carcinomas of the ampulla of Vater (AC).
Results
Surgical characteristics. Patients underwent either a PPPD, or a Kausch-Whipple procedure (284 vs. 176; p=0.015). The duration of surgery was longest in patients suffering from PDAC compared to those with AC or dCC (p<0.001 and p=0.022, respectively). There were no significant differences regards blood loss, intraoperative or postoperative complications, perioperative mortality, need for surgical revision, intensive care unit and hospital stay between the three tumor entities (Table I).
Patient characteristics. In 53.2% of the patients (251/472), a PDAC was diagnosed, whereas 29.9% (141/472) of the patients had an AC and 16.9% (80/472) were diagnosed with dCC. There were no significant differences with regard to gender, age, pre-existing illness, alcohol or nicotine abuse or family history of malignant diseases. Body mass index was significantly higher in patients suffering from PDAC (p=0.004; Table II).
Tumor characteristics. Comparing disease extent of periampullary entities, significant differences were found between the groups regarding T (tumor size) and N (LN) classification, and UICC staging (p<0.001). dCCs were found to be smaller in size as well as having a lower T-stage when compared to ACs and PDACs. Tumors in patients suffering from PDAC were of a significantly higher UICC stage than in both patients with dCC (p<0.001) and those with AC (p=0.001). When comparing AC and dCC, patients suffering from dCC presented with higher UICC stage at surgery than those with AC (p<0.001). Higher-staged disease resulted in a higher frequency of positive resection margins. There were no differences between the groups regards tumor grading and metastasis (Table III).
Characteristics of patients suffering from adenocarcinomas originating from the pancreatic head (PDAC), the distal bile duct (dCC), and carcinomas of the ampulla of Vater (AC).
LN metastasis. Patients with PDAC were more often diagnosed with nodal-positive disease compared to patients with AC or dCC (p<0.001 and p=0.003; Table I). Comparing the total number of positive LNs, LNs removed, and the resulting LN ratio (LNR), no significant differences were found between the tumor entities. Focusing on the correlation between T-stage and LNR, for all entities, a strong correlation between a higher T-stage and a higher LNR (p=0.006) was found.
Perineural infiltration. PDAC had the highest rate of perineural invasion (91.4%) compared to dCC (88.1%) and AC (61.3%) considering all cases with available perineural assessment (p=0.032). As we had very few cases with histologically-proven negative perineural invasion, the analysis was repeated comparing positive cases with all other cases and still found PDAC to have the highest rate of perineural infiltration (59.4% vs. 54.8% vs. 14.5%, respectively).
Tumor markers carcinoembryonic antigen (CEA) and carbohydrate antigen (CA 19-9). Measurement of tumor markers CA 19-9 and CEA before surgery showed significant differences between the three entities. PDAC displayed the highest levels of both markers, followed by dCC and AC. Differences were also significant when compared pairwise (Table II).
CEA discriminated between dCC and PDAC (p=0.041) with a threshold of 5 μg/l difference (specificity 94%, sensitivity 25%). For CA 19-9, specificity (40%) and sensitivity (80%) in regard to discrimination between PDAC and AC was low. Both CEA and CA 19-9 differentiated reliably between PDAC and AC (CEA: p<0.001; CA19-9: p=0.025), while a high specificity of 92% with a sensitivity of 75% was achieved when both markers were combined.
Survival analysis. The median survival of all 472 patients was 21.2 months, the 1-year survival rate was 66.2%. The overall 5-year survival rate was 26.3%. Patients with PDAC had the poorest median survival (17.2 months) followed by dCC (19.9 months; Figure 2). Patients with AC exhibited the most favorable prognosis (31.9 months) and survived significantly longer than patients with PDAC (p<0.001). However, there was no significant difference in survival rates between patients with dCC and AC. The T-stage (p=0.003) as well as the N-stage (p<0.001) and metastasis (M) stage (p=0.001) had a significant effect on overall survival. Tumor grading (p<0.001) and the status of the resection margin (R0 vs. R1) also significant influenced survival. In general, patients with a higher UICC stage had a significantly poorer survival rate (p<0.001; Table IV).
Looking into technical aspects, longer duration of surgery (p=0.001), greater blood loss (p<0.001) and longer stay in the intensive care unit (p=0.010) resulted in a diminished overall survival. Whereas postoperative complications and the need for operative revisions (p=0.001) had a negative effect on survival (p=0.020), intraoperative complications and duration of hospital stay did not have any significant effect (Table IV).
Tumor characteristics of adenocarcinomas originating from the pancreatic head (PDAC), the distal bile duct (dCC), and carcinomas of the ampulla of Vater (AC).
Only in dCC were we able to find an impact of perineural invasion on survival. This difference was significant when cases with positive perineural infiltration were compared to all cases with dCC (p=0.009). In PDAC and AC, perineural invasion showed no correlation to patient survival.
We further analyzed the variables LN status, number of positive LNs and LNR separately and discovered an effect on overall survival. In a logistic regression analysis of positive LNs, with each increase of positive LN number, the chance of survival decreased by 25% (odds ratio=0.075, p=0.002). With an increase of the LNR by 10%, the chance of survival decreased by 31% (odds ratio=69, p=0.001).
Both preoperative CA 19-9 (p<0.001) and CEA were (p<0.001) predictive for overall survival when including all three entities in the analysis (Table IV). An increase of CEA of 1 μg/l resulted in a decrease of overall survival of 33%.
In multivariate analysis, age older than 65 years (hazard ratio=1.031, 95% confidence interval=1.009-1.053, p=0.005), postoperative complications (hazard ratio=1.477, 95% confidence interval=1.049-7.172, p=0.04) and higher grading of the tumor (hazard ratio=2.743, 95% confidence interval=0.049-7.172, p=0.004) proved to be independent prognostic factors for survival (Figure 2).
Discussion
Periampullary carcinomas are comprised of different tumor entities which are often difficult to discriminate in clinical practice. Nevertheless, most studies about prognostic factors focus on one tumor entity only and fail to compare the different entities occurring in the periampullary region.
In our cohort, patients suffering from PDAC, dCC and AC were all comparable as regards patient characteristics. In most previous studies, age and gender did not affect survival (3, 21, 22). In our cohort, age (>65 years) had a negative effect on long-term survival in multivariate analysis. This confirms the results of Futagawa et al. who identified age older than 75 years as a negative predictive factor for survival in 340 cases (23). However, older age was not correlated with an increased rate of postoperative complications in our cohort. On the contrary, a higher body mass index had a significant impact on survival in univariate analysis and was correlated with postoperative complications. Postoperative complications proved to have an effect on survival in the univariate and multivariate analyses.
Kaplan–Meier curves of survival according to variables found to be significant in the multivariate analysis: Age, tumor grade and postoperative complications.
The tumor size was significantly larger in patients suffering from PDAC. This might be due to the earlier onset of symptoms in patients with dCC and AC as jaundice usually occurs earlier in those patients (24). Patients with a tumor size smaller than 2 cm had a significantly longer survival in our cohort. This finding is consistent with current literature (24-27).
Comparing the three tumor entities, PDAC had the worst prognosis when compared to dCC and AC. Patient survival was consistent with that reported in other studies (24, 26-30). It is questionable if the impaired prognosis of the PDAC cohort was due to differences in tumor biology or based on delayed diagnosis. Schmidt et al. showed that the tumor entity is an independent factor influencing survival in a multivariate analysis of 516 cases, with poorest survival for patients suffering from PDAC (29). In addition, we found that patients with PDAC had a higher UICC stage at the time of operation. Patients with more advanced tumor stages are characterized by a significantly higher rate of LN metastasis in PDAC. In other words, more than two-thirds of patients with PDAC were nodal-positive, representing more aggressive tumor biological behavior.
Interestingly, the number of LN removed and the LNR did not differ between the entities but did have an effect on overall survival. In colon cancer, the LNR has been proven to be a relevant prognostic factor in representative patient cohorts (31). For this, attempts to include this variable in the classification of colon cancer have already been made (31). Corresponding findings have been described for PDAC (26). A study from Baltimore described LNR as one of the most powerful predictors of survival in a cohort of over 900 patients (24). Similar results have been found in periampullary malignancies and in particular in AC where the LNR was found to be an independent prognostic factor after resection (27, 32, 33). In our cohort, an increasing number of positive LNs as well as the LNR had a negative effect on overall survival, an increase of the LNR by 10% resulted in a survival decrease of 31%. This highlights the relevance of LN dissection for accurate diagnostic staging.
Variables showing a significant effect on the survival of patients in the univariate analysis.
Larger tumor size, a higher number of R1 resections, a higher rate of nodal and perineural invasion, a higher UICC stage and an overall worse prognosis of PDAC compared to dCC and AC raises the question if the one-size-fits-all treatment of periampullary cancer with resection up-front and adjuvant chemotherapy adequately address this aggressive tumor type. One important aspect that will shape this discussion is the increasing relevance of neoadjuvant chemotherapy. Whereas neoadjuvant treatment seems to be beneficial for locally advanced PDAC, evidence is still lacking for primary resectable PDAC, dCC and AC (30, 34). Different neoadjuvant regimens are usually applied in pancreatic cancer (30, 35-37). However, no randomized controlled trial has yet been completed. It is postulated that in the neoadjuvant setting, patients profit from early systemic control, which also helps discover those who will benefit from radical surgery and specific adjuvant regimens (36).
Reliable data for neoadjuvant treatment of periampullary cancer other than PDAC are scarce and less convincing (38, 39). To clarify whether patients with dCC or AC could nevertheless also benefit from this approach, prospective and treatment studies with exact discrimination between the tumor entities are needed.
Acknowledgements
The Authors thank all patients for their participation in the study. Moreover, we thank all nurses and physicians involved in the care of the patients.
Footnotes
Conflicts of Interest and Funding
None of the Authors has any conflict of interests to declare. Internal funding of the surgical clinic was used only. Dr. Rosa Schmuck is participant in the BIH-Charité Clinician Scientist Program funded by the Charité – Universitaetsmedizin Berlin and the Berlin Institute of Health.
- Received July 12, 2018.
- Revision received August 15, 2018.
- Accepted August 23, 2018.
- Copyright© 2018, International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.