ABSTRACT
Background
Ampullary adenocarcinoma (AAC) is a rare malignancy with heterogeneous biological behavior. Although surgical resection is the only potentially curative option, the prognostic role of histopathologic features and adjuvant chemotherapy (AC) remains unclear.
Methods
Consecutive patients undergoing curative‐intent pancreatoduodenectomy for AAC at the University of Colorado Hospital between January 2013 and May 2025 were retrospectively reviewed. Clinicopathologic variables associated with overall (OS) and recurrence‐free survival (RFS) were analyzed using multivariable Cox regression.
Results
Ninety‐seven patients were included. Median OS was 93.9 months with 1‐, 3‐, 5‐, and 10‐year OS rates of 89%, 67%, 59%, and 46%. Older age (HR, 3.79; 95% CI, 1.71–8.41), lymph node (LN) involvement (HR, 4.92; 95% CI, 1.87–12.93), and R1 margin (HR, 4.33; 95% CI, 1.09–17.25) independently predicted poorer OS. The pancreatobiliary subtype (pbAAC) showed worse OS than the intestinal subtype (iAAC) (5‐year OS, 48% vs. 75%; p = 0.012). In pbAAC, LN metastasis (HR, 4.16; 95% CI, 1.31–13.18) and R1 margin (HR, 5.47; 95% CI, 1.10–27.20) predicted worse OS, whereas AC was associated with improved survival (HR, 0.404; 95% CI, 0.18–0.91). No AC benefit was observed in iAAC (p = 0.323).
Conclusions
Histopathologic subtype is a key prognostic factor in AAC, with AC associated with better survival in pbAAC.
Keywords: adjuvant chemotherapy, ampulla of Vater, ampullary adenocarcinoma, pancreatoduodenectomy, survival
1. Introduction
Ampullary adenocarcinoma (AAC) is a rare malignancy, accounting for less than 1% of all gastrointestinal cancers [1] and approximately 6% of periampullary tumors [2]. AAC originates from any portion of the ampulla of Vater, including the duodenal mucosa covering the papilla and the intra‐Oddian segments of the common bile duct and main pancreatic duct [3]. Owing to their anatomical location, these tumors often present with early symptoms such as biliary obstruction [4], and are therefore more frequently resectable at diagnosis compared with other periampullary cancers, with reported resection rates around 50% [5]. Consequently, complete oncological resection remains the only potentially curative treatment, yielding 5‐year overall survival (OS) rates ranging from 30% to 70% [6]. This wide variability in survival likely reflects the biological heterogeneity of AAC, which comprises intestinal (iAAC), pancreatobiliary (pbAAC), and mixed (mAAC) subtypes with distinct pathological features and prognostic implications [2, 3]. The intestinal subtype shares morphological and clinical characteristics with duodenal adenocarcinoma, whereas the pancreatobiliary subtype more closely resembles pancreatic cancer, being associated with a higher incidence of lymph node (LN) metastases, perineural invasion, and more advanced T stages [7, 8]. Accordingly, histopathological subtype, together with TNM stage and patient age, plays a key role in determining prognosis [7]. However, given the rarity of AAC, data on survival outcomes and prognostic factors across different histological subtypes remain limited.
Likewise, the role of adjuvant chemotherapy (AC) is still a matter of debate. While several studies have reported a survival benefit associated with adjuvant treatment following curative resection [9, 10], others have failed to confirm such an advantage [11, 12]. This inconsistency may reflect the limited evidence specifically addressing the effectiveness of AC according to histopathologic subtype. Although current NCCN guidelines recommend considering adjuvant treatment based on histologic differentiation, the prognostic and therapeutic implications of this distinction have not been fully clarified [2].
The aim of this study is to investigate histopathological predictors of survival and recurrence, with particular attention to differences among the intestinal and pancreatobiliary subtypes in a large single‐center cohort of patients undergoing pancreatoduodenectomy for AAC.
2. Methods
This single‐center, retrospective study was approved by the Colorado Multiple Institutional Review Board (COMIRB #18‐2839) and is reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines [13].
2.1. Patients
All consecutive adult patients (≥ 18 years) who underwent curative‐intent pancreatoduodenectomy for AAC at the University of Colorado Hospital between January 2013 and May 2025 were identified from a prospectively maintained database. Exclusion criteria included local excision of AAC, palliative procedures, R2 resections, metastatic disease at baseline, or concomitant resection for distant metastases. To ensure a more accurate assessment of long‐term outcomes and to better evaluate the role of AC, patients who received neoadjuvant chemotherapy, died within 90 days after surgery, or had a follow‐up shorter than 90 days were also excluded.
2.2. Perioperative Assessment and Management
All patients underwent standard preoperative evaluation, including clinical examination, laboratory testing (with measurement of CEA and CA 19‐9 levels), and cross‐sectional imaging with contrast‐enhanced computed tomography (CT) or magnetic resonance imaging (MRI) of the abdomen. Endoscopic retrograde cholangiopancreatography (ERCP) or endoscopic ultrasound (EUS) was performed when clinically indicated to obtain tissue diagnosis and/or relieve biliary obstruction. In patients presenting with jaundice, preoperative biliary drainage was selectively performed in cases of cholangitis, severe hyperbilirubinemia, or anticipated delay in surgery.
All patients underwent open or minimally invasive (laparoscopic or fully robotic) pancreatoduodenectomy with curative intent, according to surgeon preference and tumor characteristics. Standard lymphadenectomy was performed in all cases, including the peripancreatic, hepatoduodenal ligament, and superior mesenteric artery nodal stations. The choice between pylorus‐preserving and classical Whipple procedure was based on intraoperative findings and surgeon discretion. Resection margins were assessed intraoperatively by frozen section, and all specimens were examined by dedicated gastrointestinal pathologists. Resections were classified as R0 when no microscopic residual disease was present at the surgical margin. Tumor staging was based on the 7th edition of the American Joint Committee on Cancer (AJCC) classification [14] until 2018, and on the 8th edition thereafter [15]. For analytical purposes, all cases treated before 2018 were retrospectively reviewed and reclassified according to the latest AJCC staging system. Histopathologic subtypes were determined using a combination of morphological assessment and immunohistochemical staining, with markers CDX2, MUC1, CK20, CK7, and MUC2 to distinguish between pancreatobiliary and intestinal subtypes. Mixed tumors were classified based on the dominant histologic and immunohistochemical features. Given their biological behavior, which closely resembles that of pbAAC, these cases were grouped with the pancreatobiliary category for all analyses. Previous studies have shown that mAAC tend to be more aggressive, resistant to treatment, and associated with higher recurrence rates and poorer survival compared to intestinal tumors [2, 7].
Referral for AC was based on multidisciplinary team (MDT) consensus. In accordance with NCCN guidelines [2], patients considered eligible for AC received postoperative systemic therapy for a planned duration of approximately 6 months, with treatment duration adapted according to individual tolerance and clinical conditions. The choice of chemotherapy regimen was left to the discretion of the treating medical oncologist. AC was initiated within 10–12 weeks after surgery, following postoperative restaging with CT scan and serum CA 19‐9 measurement. After completion of adjuvant therapy, patients underwent surveillance with cross‐sectional imaging and CA 19‐9 measurement every 6 months. Patients not receiving AC were followed according to the same radiologic and laboratory surveillance schedule. Standard follow‐up was performed every 6 months during the first 5 years after surgery. Beyond 5 years, follow‐up was individualized based on patient characteristics, fitness for further treatment, and patient preference, with either annual surveillance or symptom‐driven follow‐up for those unfit for treatment.
3. Statistical Analysis
Continuous variables are reported as medians with interquartile ranges (IQRs) and group comparisons were performed using the Mann–Whitney U‐test. Categorical variables were analyzed using Chi‐square or Fisher's exact test, as appropriate. OS was defined as the time from surgery to death from any cause or last follow‐up, whichever occurred first. Recurrence‐free survival (RFS) was defined as the time from surgery to the first documented recurrence or last imaging without evidence of disease. Survival curves were estimated using the Kaplan–Meier method, and differences between groups were assessed using the log‐rank test. OS and RFS rates at 1‐, 3‐, and 5‐years after resection were derived from life tables. All statistical tests were two‐sided, and p‐values < 0.05 were considered statistically significant. Prognostic factors for OS and RFS were evaluated using Cox proportional hazard models. Variables with a p‐value < 0.100 in univariable analysis were included in multivariable models. Continuous variables were dichotomized at their median value for inclusion in the Cox regression. All statistical analyses were conducted using IBM SPSS Statistics, version 31.0 (IBM Corp., Armonk, NY, USA).
4. Results
4.1. Patient Characteristics
A total of 106 patients underwent curative‐intent pancreatoduodenectomy for AAC during the study period. Of these, three patients (2.8%) died within 90 days from postoperative complications, three (2.8%) had a follow‐up shorter than 90 days, and three (2.8%) had received neoadjuvant chemotherapy; all were excluded from the analysis, leaving 97 patients in the final cohort.
Baseline demographics and tumor characteristics are summarized in Table 1. The median age was 67 years (IQR, 56–77), with 54 patients (55.7%) being male. Postoperative complications of Clavien‐Dindo (CD) grade ≥ 3 occurred in 35.1% of patients. Clinically relevant POPF was the most frequent complication (35.1%), followed by DGE (19.6%), PPH (16.5%), chyle leak (8.2%), and bile leak (3.1%). The median hospital stay was 11 days (IQR, 8–15). Histopathologic subtype was documented in 95 cases: 38 (39.2%) intestinal, 49 (50.5%) pancreatobiliary, and 8 (8.2%) mixed. Tumors were well differentiated in 13 patients (13.4%), moderately differentiated in 63 (64.9%), and poorly differentiated in 21 (21.6%). No T4 tumors were observed; 55 cases (56.7%) were staged as T3, 24 (24.7%) as T2, and 18 (18.6%) as T1. LN involvement was present in 53 patients (54.6%), lymphovascular invasion in 57 (58.8%), and perineural invasion in 46 (47.4%). R1 resection margins were identified in four patients (4.1%). A comparison of clinicopathological factors between iAAC and pbAAC is provided in Table S1.
TABLE 1.
Patient characteristics in the entire cohort.
| Variables | Complete cohort (n = 97; 100%) | Adjuvant chemotherapy (n = 69; 71.1%) | No adjuvant chemotherapy (n = 28; 28.9%) | p |
|---|---|---|---|---|
| Age, years, median (IQR) | 67 (56–77) | 67 (55–72) | 72 (61–81) | 0.128 |
| Sex, male, n (%) | 54 (55.7) | 40 (58.0) | 14 (50.0) | 0.474 |
| BMI, kg/m2, median (IQR) | 26.2 (22.7–30.5) | 26.4 (22.7–30.8) | 26.6 (22.8–30.8) | 0.994 |
| ASA, n (%) | ||||
| 2 | 30 (30.9) | 24 (34.8) | 6 (21.4) | 0.435 |
| 3 | 64 (66.0) | 43 (62.3) | 21 (75.0) | |
| 4 | 3 (3.1) | 2 (2.9) | 1 (3.6) | |
| CCI, median (IQR) | 3 (1–4) | 3 (1–4) | 3 (2–5) | 0.065 |
| CA 19‐9, n (%) | ||||
| < 5 UI/mL | 10 (10.3) | 9 (13.0) | 1 (3.6) | 0.339 |
| 5–37 UI/mL | 38 (39.2) | 25 (36.2) | 13 (46.4) | |
| > 37 UI/mL | 46 (47.5) | 34 (49.3) | 12 (42.9) | |
| Missing | 3 (3.1) | 1 (1.4) | 2 (7.1) | |
| Vascular resection, n (%) | 1 (1.1) | 1 (1.5) | 0 (0.0) | 0.520 |
| EBL, mL, median (IQR) | 300 (200–450) | 300 (200–400) | 350 (150–650) | 0.239 |
| DOS, min, median (IQR) | 356 (321–403) | 352 (314–402) | 364 (329–412) | 0.679 |
| POPF grade B/C, n (%) | 34 (35.1) | 22 (31.9) | 12 (42.9) | 0.305 |
| PPH grade B/C, n (%) | 16 (16.5) | 12 (17.4) | 4 (14.3) | 0.709 |
| DGE grade B/C, n (%) | 19 (19.6) | 10 (14.5) | 9 (32.1) | 0.047 |
| Bile leak grade B/C, n (%) | 3 (3.1) | 1 (1.4) | 2 (7.1) | 0.199 |
| Chyle leak grade B/C, n (%) | 8 (8.2) | 6 (8.7) | 2 (7.1) | 0.978 |
| LOS, days, median (IQR) | 11 (8–15) | 10 (8–14) | 13 (9–22) | 0.040 |
| CD≥IIIa, n (%) | 34 (35.1) | 21 (30.4) | 13 (46.4) | 0.135 |
| Histologic subtype, n (%) | ||||
| Intestinal | 38 (39.2) | 28 (40.6) | 10 (35.7) | 0.585 |
| Pancreatobiliary | 49 (50.5) | 33 (47.8) | 16 (57.1) | |
| Mixed | 8 (8.2) | 7 (10.1) | 1 (3.6) | |
| Missing | 2 (2.1) | 1 (1.4) | 1 (3.6) | |
| Grade of differentiation, n (%) | ||||
| Well | 13 (13.4) | 8 (11.6) | 5 (17.9) | 0.327 |
| Moderate | 63 (64.9) | 48 (69.6) | 15 (53.6) | |
| Poor | 21 (21.6) | 13 (18.8) | 8 (28.6) | |
| T stage, n (%) | ||||
| T1 | 18 (18.6) | 6 (8.7) | 12 (42.9) | < 0.001 |
| T2 | 24 (24.7) | 19 (27.5) | 5 (17.9) | |
| T3 | 55 (56.7) | 44 (63.8) | 11 (39.3) | |
| N stage, n (%) | ||||
| N0 | 43 (44.3) | 28 (40.6) | 15 (53.6) | 0.500 |
| N1 | 26 (26.8) | 20 (29.0) | 6 (21.4) | |
| N2 | 28 (28.9) | 21 (30.4) | 7 (25.0) | |
| R1 margin, n (%) | 4 (4.1) | 3 (4.3) | 1 (3.6) | 0.673 |
| Lymphovascular invasion, n (%) | 61 (62.9) | 49 (71.0) | 12 (42.9) | 0.009 |
| Perineural invasion, n (%) | 47 (48.5) | 38 (55.1) | 9 (32.1) | 0.122 |
Note: Bold numbers indicates p values < 0.05.
Abbreviations: CA19‐9, carbohydrate antigen 19‐9; CCI, Charlson Comorbidity Index; CD, Clavien‐Dindo classification; DGE, delayed gastric emptying; DOS, duration of surgery; EBL, estimated blood loss; LOS, length of stay; POPF, post‐operative pancreatic fistula; PPH, post‐pancreatectomy hemorrhage.
After oncologic resection, 69 patients (71.1%) received AC. Although not statistically significant, these patients tended to have a lower comorbidity burden (Charlson Comorbidity Index, p = 0.065). They also experienced a lower rate of clinically relevant DGE (32.1% vs. 14.5%, p = 0.047) and a shorter median hospital stay (13 vs. 10 days, p = 0.040), without significant differences in severe postoperative morbidity (CD ≥ 3: 46.4% vs. 30.4%, p = 0.135). Tumors in the AC group were more frequently advanced (T3 stage: 63.8% vs. 39.3%, p < 0.001) and more often associated with lymphovascular invasion (71.0% vs. 42.9%, p = 0.009). No significant differences were observed in histopathologic subtype (p = 0.585), nodal status (p = 0.500), rate of R1 margin (p = 0.673), or perineural invasion (p = 0.122).
Details of AC regimens according to histopathologic subtype are reported in Table S2. In the overall cohort, gemcitabine‐based regimens were the most commonly administered (50.7%) and were used more frequently in pbAAC than in iAAC (60.0% vs. 37.9%, p = 0.070), although this difference did not reach statistical significance. Conversely, 5‐FU‐based regimens, particularly FOLFOX, were preferred in the intestinal subtype (10.0% vs. 34.5%, p = 0.017).
5. Overall Survival
At a median follow‐up of 39.1 months (IQR 15.0–76.4), 55 patients (56.7%) were alive. The median OS was 93.9 months with 1‐, 3‐, 5‐, and 10‐year OS rates of 89%, 67%, 59%, and 46%, respectively. Kaplan–Meier survival curves for the entire cohort and stratified by histopathologic subtypes are shown in Figure 1.
FIGURE 1.
Kaplan–Meier curves for (a) Overall Survival and (b) Recurrence‐Free Survival according to histopathologic subtypes.
Univariable and multivariable analyses of prognostic factors are summarized in Table 2. On Kaplan–Meier analysis, survival was significantly lower in patients with pancreatobiliary subtype (p = 0.012), poor tumor differentiation (p = 0.030), T3 stage (p < 0.001), LN involvement (p < 0.001), R1 margin (p < 0.001), lymphovascular invasion (p = 0.003), and perineural invasion (p = 0.007) (Supplementary Figure S1). In the multivariable Cox model, only age (HR, 4.16; 95% CI, 1.71–8.41; p = 0.001), LN involvement (HR, 4.92; 95% CI, 1.87–12.93; p = 0.001), and R1 margin (HR, 4.33; 95% CI, 1.09–17.25; p = 0.038) remained independent predictors of decreased OS.
TABLE 2.
Univariable and multivariable analysis for risk factors associated with overall survival in the overall cohort.
| Univariate analysis | Multivariate analysis | |||||
|---|---|---|---|---|---|---|
| HR | 95% CI | p | HR | 95% CI | p | |
| Age | 3.993 | 1.987–8.026 | < 0.001 | 3.788 | 1.707–8.407 | 0.001 |
| Sex | ||||||
| Male | 0.480 | 0.257–0.894 | 0.021 | 1.050 | 0.505–2.180 | 0.897 |
| CA 19‐9 | ||||||
| 5–37 UI/mL | Ref | Ref | Ref | |||
| > 37 UI/mL | 1.348 | 0.696–2.610 | 0.376 | |||
| < 5 UI/mL | 1.160 | 0.385–3.492 | 0.792 | |||
| Histologic subtype | ||||||
| Intestinal | Ref | Ref | Ref | Ref | Ref | Ref |
| Mixed/pancreatobiliary | 2.371 | 1.186–4.740 | 0.015 | 1.453 | 0.631–3.346 | 0.379 |
| Grade of differentiation | ||||||
| Well‐moderate | Ref | Ref | Ref | Ref | Ref | Ref |
| Poor | 2.036 | 1.056–3.927 | 0.034 | 1.926 | 0.926–4.006 | 0.079 |
| T stage | ||||||
| T1–T2 | Ref | Ref | Ref | Ref | Ref | Ref |
| T3 | 3.227 | 1.628–6.398 | < 0.001 | 1.640 | 0.615–4.373 | 0.323 |
| N stage | ||||||
| N0 | Ref | Ref | Ref | Ref | Ref | Ref |
| N 1–2 | 6.865 | 3.121–15.105 | < 0.001 | 4.915 | 1.868–12.928 | 0.001 |
| R1 margin | 10.328 | 3.356–31.788 | < 0.001 | 4.331 | 1.087–17.251 | 0.038 |
| Lymphovascular invasion | 2.672 | 1.354–5.273 | 0.005 | 0.720 | 0.297–1.747 | 0.468 |
| Perineural invasion | 2.376 | 1.241–4.551 | 0.009 | 1.077 | 0.389–2.977 | 0.887 |
| Adjuvant chemotherapy | 0.728 | 0.383–1.384 | 0.332 | |||
Note: Bold numbers indicates p values < 0.05.
Abbreviation: CA19‐9, carbohydrate antigen 19‐9.
6. Recurrence Free Survival
For RFS analysis, six patients (6.2%) who, after MDT discussion, were followed symptomatically only were excluded. The remaining 91 cases had a median follow‐up of 34.8 months, during which 31 patients (34.1%) developed recurrence. The 1‐, 3‐, 5‐, and 10‐year RFS rates were 81%, 68%, 61%, and 56%, respectively. Kaplan–Meier RFS curves for the entire cohort and stratified by histopathologic subtype are shown in Figure 1, and detailed survival data are provided in Table S3.
Univariable and multivariable analyses of factors associated with RFS are presented in Table 3. In the multivariable Cox model, only age (HR, 2.91; 95% CI, 1.23–6.88; p = 0.015), LN involvement (HR, 3.97; 95% CI, 1.29–12.20; p = 0.016), and R1 margin (HR, 26.04; 95% CI, 3.95–171.81; p < 0.001) remained independent predictors of shorter RFS.
TABLE 3.
Univariable and multivariable analysis for risk factors associated with recurrence free survival in the overall cohort.
| Univariate analysis | Multivariate analysis | |||||
|---|---|---|---|---|---|---|
| HR | 95% CI | p | HR | 95% CI | p | |
| Age | 2.411 | 1.157–5.025 | 0.019 | 2.911 | 1.232–6.876 | 0.015 |
| Sex | ||||||
| Male | 0.568 | 0.280–1.152 | 0.117 | |||
| CA 19‐9 | ||||||
| 5–37 UI/mL | Ref | Ref | Ref | |||
| > 37 UI/mL | 1.274 | 0.587–2.763 | 0.540 | |||
| < 5 UI/mL | 1.071 | 0.302–3.802 | 0.915 | |||
| Histologic subtype | ||||||
| Intestinal | Ref | Ref | Ref | Ref | Ref | Ref |
| Mixed/pancreatobiliary | 2.291 | 1.012–5.189 | 0.047 | 1.377 | 0.520–3.644 | 0.519 |
| Grade of Differentiation | ||||||
| Well‐moderate | Ref | Ref | Ref | Ref | Ref | Ref |
| Poor | 2.288 | 1.092–4.793 | 0.028 | 2.159 | 0.944–4.940 | 0.068 |
| T stage | ||||||
| T1–T2 | Ref | Ref | Ref | Ref | Ref | Ref |
| T3 | 1.947 | 0.926–4.095 | 0.079 | 0.669 | 0.223–2.006 | 0.473 |
| N stage | ||||||
| N0 | Ref | Ref | Ref | Ref | Ref | Ref |
| N 1–2 | 3.808 | 1.684–8.609 | 0.001 | 3.968 | 1.291–12.199 | 0.016 |
| R1 margin | 35.108 | 7.690–160.276 | < 0.001 | 26.043 | 3.948–171.805 | < 0.001 |
| Lymphovascular invasion | 1.888 | 0.884–4.031 | 0.100 | 0.606 | 0.213–1.726 | 0.348 |
| Perineural invasion | 2.081 | 0.970–4.463 | 0.060 | 1.743 | 0.532–5.711 | 0.359 |
| Adjuvant chemotherapy | 0.867 | 0.408–1.844 | 0.712 | |||
Note: Bold numbers indicates p values < 0.05.
Abbreviation: CA19‐9, carbohydrate antigen 19‐9.
7. Overall Survival According to Histopathologic Subtype
Differences in survival outcomes were observed across histopathologic subtypes. Median OS was 53.0 months in pbAAC and not reached in iAAC (log‐rank p = 0.012). The corresponding 5‐year OS rates were 48% and 75%, respectively.
Tables 4 and 5 summarize the univariable and multivariable analyses of factors associated with OS in the pbAAC and in iAAC, respectively. On multivariable Cox regression, LN involvement (HR, 4.16; 95% CI, 1.31–13.18; p = 0.016) and R1 margin (HR, 5.47; 95% CI, 1.10–27.20; p = 0.038) were independent predictors of poorer OS in pbAAC, whereas only age (HR, 5.43; 95% CI, 1.28–22.98; p = 0.022) predicted survival in iAAC. Notably, AC was independently associated with improved OS in pbAAC (HR, 0.40; 95% CI, 0.18–0.91; p = 0.029) but showed no benefit in iAAC (p = 0.323). Kaplan–Meier curves comparing OS according to AC use in the overall cohort and by histopathologic subtype are shown in Figure 2.
TABLE 4.
Univariable and multivariable analysis for risk factors associated with overall survival in pbAAC.
| Univariate analysis | Multivariate analysis | |||||
|---|---|---|---|---|---|---|
| HR | 95% CI | p | HR | 95% CI | p | |
| Age | 3.007 | 1.328–6.812 | 0.008 | 2.338 | 0.981–5.572 | 0.055 |
| Sex | ||||||
| Male | 0.690 | 0.325–1.465 | 0.334 | |||
| CA 19‐9 | ||||||
| 5–37 UI/mL | Ref | Ref | Ref | |||
| > 37 UI/mL | 1.299 | 0.582–2.897 | 0.523 | |||
| < 5 UI/mL | 0.744 | 0.163–3.402 | 0.703 | |||
| Grade of differentiation | ||||||
| Well‐moderate | Ref | Ref | Ref | Ref | Ref | Ref |
| Poor | 2.116 | 0.997–4.493 | 0.051 | 1.669 | 0.743–3.750 | 0.215 |
| T stage | ||||||
| T1–T2 | Ref | Ref | Ref | Ref | Ref | Ref |
| T3 | 3.234 | 1.424–7.344 | 0.005 | 1.863 | 0.709–4.898 | 0.207 |
| N stage | ||||||
| N0 | Ref | Ref | Ref | Ref | Ref | Ref |
| N 1–2 | 6.094 | 2.104–17.651 | < 0.001 | 4.156 | 1.311–13.177 | 0.016 |
| R1 margin | 6.033 | 1.328–27.404 | 0.020 | 5.469 | 1.100–27.199 | 0.038 |
| Lymphovascular invasion | 1.810 | 0.803–4.082 | 0.153 | |||
| Perineural invasion | 1.685 | 0.768–3.698 | 0.193 | |||
| Adjuvant chemotherapy | 0.491 | 0.231–1.042 | 0.064 | 0.404 | 0.179–0.910 | 0.029 |
Note: Bold numbers indicates p < 0.05.
Abbreviation: CA19‐9, carbohydrate antigen 19‐9.
TABLE 5.
Univariable and Multivariable Analysis for Risk Factors Associated with Overall Survival in iAAC.
| Univariate analysis | Multivariate analysis | |||||
|---|---|---|---|---|---|---|
| HR | 95% CI | p | HR | 95% CI | p | |
| Age | 5.194 | 1.385–19.473 | 0.015 | 5.431 | 1.283–22.983 | 0.022 |
| Sex | ||||||
| Male | 0.306 | 0.094–0.989 | 0.480 | |||
| CA 19‐9 | ||||||
| 5–37 UI/mL | Ref | Ref | Ref | |||
| > 37 UI/mL | 2.629 | 0.482–14.357 | 0.264 | |||
| < 5 UI/mL | 1.258 | 0.363–4.356 | 0.717 | |||
| Grade of differentiation | ||||||
| Well‐moderate | Ref | Ref | Ref | |||
| Poor | 1.364 | 0.298–6.247 | 0.689 | |||
| T stage | ||||||
| T1–T2 | Ref | Ref | Ref | Ref | Ref | Ref |
| T3 | 3.210 | 0.905–11.392 | 0.071 | 1.844 | 0.347–9.793 | 0.473 |
| N stage | ||||||
| N0 | Ref | Ref | Ref | Ref | Ref | Ref |
| N 1–2 | 6.322 | 1.776–22.504 | 0.004 | 2.512 | 0.552–11.430 | 0.233 |
| R1 margin | 25.258 | 3.474–183.658 | 0.001 | 6.719 | 0.877–51.511 | 0.067 |
| Lymphovascular invasion | 4.252 | 1.173–15.414 | 0.028 | 2.711 | 0.354–20.755 | 0.337 |
| Perineural invasion | 2.055 | 0.563–7.500 | 0.276 | |||
| Adjuvant chemotherapy | 2.161 | 0.468–9.964 | 0.323 | |||
Note: Bold numbers indicates p values < 0.05.
Abbreviation: CA19‐9, carbohydrate antigen 19‐9.
FIGURE 2.
Kaplan–Meier overall survival curves according to adjuvant chemotherapy use in (a) the overall AAC cohort, (b) patients with pbAAC, and (c) patients with iAAC.
8. Discussion
This study analyzed oncologic outcomes after curative‐intent pancreatoduodenectomy for AAC, focusing on the prognostic significance of tumor histopathological features. In the overall cohort, age, LN involvement, and R1 resection margin independently predicted both OS and RFS. When stratified by histopathologic subtypes, patients with pbAAC exhibited a more aggressive disease course, with a 5‐year OS rate of 48%. In this subgroup, LN involvement and R1 margin remained independent predictors of poorer survival, whereas AC was associated with improved outcomes. Conversely, patients with iAAC showed a more indolent disease pattern, achieving a 5‐year OS rate of 75%. In this group, only age was independently associated with survival, and AC did not appear to affect long‐term outcomes. Altogether, these findings highlight that although AAC is anatomically defined as a single entity, it encompasses biologically distinct subtypes with distinct prognostic and therapeutic implications.
Several previous studies have attempted to identify histopathologic predictors of survival in resected AAC. In a single‐center series of 72 patients, Lazaryan et al. [11] identified perineural invasion and LN involvement as independent predictors of OS. Similar results were reported by Qiao et al. [16] in a cohort of 102 patients. However, these earlier studies included patients treated in the late 20th century, when both biological understanding of the disease and diagnostic and therapeutic strategies were limited. Consequently, perioperative mortality rates reached 5%–10%, compared with 2%–5% in more recent series [17, 18]—consistent with the 2.8% rate observed in our cohort. Moreover, several key histopathologic features, including tumor subtype, were not evaluated. Carter et al. [19] were the first to highlight the prognostic relevance of histologic subtype, reporting poorer outcomes in pbAAC compared with iAAC. However, their cohort included patients who underwent transduodenal local excision, potentially affecting nodal staging accuracy. Similarly, Chang et al. [7] correlated histopathologic features and protein expression (CDX2 and MUC1) with outcomes, confirming that prognosis is closely related to both pancreatobiliary phenotype and LN involvement. Moekotte et al. [17], in a multicenter study including 887 patients, further reported that although the pancreatobiliary subtype is associated with inferior survival (5‐year OS: 47% vs. 60%), histologic subtype alone is not an independent prognostic factor once LN involvement and perineural invasion are considered.
These observations are consistent with our results: LN involvement and R1 were the strongest predictors of poorer OS, whereas histologic subtype did not retain prognostic significance in multivariable analysis. Nevertheless, survival curves clearly show that iAAC and pbAAC—despite sharing a common anatomical origin—display distinct biological behaviors. Patients with iAAC achieved a 5‐year OS rate of 75%, influenced only by advanced patient age. Although this may partly reflect the limited number of events in this subgroup, the finding aligns with that of Finton et al. [18], who, in a series of 171 iAAC patients, identified ASA ≥ 3 and T stage ≥ 3 as the only independent predictors of survival. In contrast, the pbAAC subgroup in our cohort showed a higher rate of perineural and lymphovascular invasion, with a 5‐year OS rate of 48%, consistent with its more aggressive biology. Similar to pancreatic ductal adenocarcinoma (PDAC), pbAAC typically presents at a locally advanced stage and is characterized by LN metastasis, poor differentiation, positive margins, and frequent lymphovascular and perineural invasion [20]. Consistent with prior multicenter analyses by Bolm et al. [21] and Moekotte et al. [17], LN involvement and R1 margin were independent predictors of survival in our series, while AC was associated with improved OS specifically in this subgroup.
Regarding recurrence risk, the 5‐year RFS in our study was 61%, in line with previous reports (51%–58%) [17, 22, 23]. In the present study, older age, LN involvement, and R1 margin emerged as independent predictors of recurrence. These findings reinforce previous evidence identifying LN metastasis as the strongest independent risk factor for disease recurrence [12, 17, 22]. Other studies have also suggested an association between shorter RFS and pancreatobiliary subtype, higher T stage, and poor differentiation [12, 22, 23], although these trends did not reach statistical significance in our cohort.
Although AAC is a rare malignancy, it accounts for up to 20% of all pancreatoduodenectomies performed for cancer [24]. Understanding its biological behavior therefore has crucial prognostic and therapeutic implications. However, the role of AC remains controversial, as disease‐specific prospective studies are lacking and available evidence mainly derives from heterogeneous periampullary cohorts. The ASCOT trial [25] demonstrated that in patients with biliary tract cancer, adjuvant S‐1 therapy, compared with surgery alone, was associated with improved 3‐year OS (77.1% vs. 67.6%, p = 0.008) and 3‐year relapse‐free survival (62.4% vs. 50.9%, p = 0.088). However, this study did not specifically include AAC, and its findings may have limited applicability to western populations due to the high toxicity of S‐1. The ESPAC‐3 trial [26] investigated the role of AC in periampullary cancers and, although no OS benefit was observed in the overall cohort, a subgroup analysis limited to AAC showed improved survival among patients treated with gemcitabine compared with 5‐FU/leucovorin or observation (70.8 vs. 57.8 vs. 40.6 months). Nonetheless, AAC represented only a small subset of the study group, and disease‐specific evidence remains limited to retrospective studies with conflicting results. Lazaryan et al. [11] and Doepker et al. [12] found no survival benefit from AC compared with surgery alone, whereas Nassour et al. [9], analyzing 4190 cases from the National Cancer Database, demonstrated an advantage, particularly in patients with more advanced disease (higher T and N stage). Similarly, Duhn et al. [27], reported improved OS with AC, especially among patients with advanced T stage, LN involvement, lymphovascular invasion, or poor differentiation. In line with the increasing evidence suggesting a potential benefit, current NCCN guidelines [2] recommend adjuvant therapy for patients with LN involvement or T stage > 2. However, its role across histologic subtypes remains uncertain.
Our study is among the few to specifically examine the impact of AC according to histologic phenotype and found that AC was associated with better OS in pbAAC, whereas no survival benefit was observed in iAAC. These results are consistent with those reported by Bolm et al. [21] and Moekotte et al. [10], and are further supported by the large multicenter study by Uijterwijk et al. [8], which included 765 iAAC and 819 pbAAC patients. In that study, AC significantly improved OS in the pancreatobiliary subtype—regardless of the regimen—but provided no benefit in intestinal‐type tumors. The advantage observed in pbAAC may reflect the frequent use of gemcitabine‐based regimens (50.7% in our cohort), which are well established in pancreatic cancer [28], whereas their efficacy in intestinal‐type tumors remains uncertain [29]. Indeed, in vitro studies have shown reduced sensitivity to gemcitabine in iAAC and significant growth inhibition in pbAAC [30]. It is therefore plausible that patients with iAAC may benefit from adjuvant regimens used for duodenal adenocarcinoma, although robust evidence is still lacking.
Several limitations should be acknowledged when interpreting our findings. First, the retrospective design introduces potential information bias; however, as our analysis primarily focused on correlations between histopathologic and survival variables, such bias is expected to be limited. Furthermore, given the rarity of AAC and the lack of randomized controlled trials, well‐designed cohort studies based on prospectively maintained databases currently represent the most reliable approach to evaluate treatment strategies and oncologic outcomes for this disease. Second, the single‐center nature of the study and the relatively modest sample size may limit generalizability. Nevertheless, this design allowed for the collection of granular clinicopathologic data and ensured consistency in surgical management, pathological assessment, and follow‐up protocols. A potential selection bias must also be acknowledged, since patients who received AC generally had shorter postoperative courses and more advanced pathological features (higher T stage and more frequent lymphovascular and perineural invasion), suggesting that the allocation of AC may not have been random and could potentially affect the interpretation of treatment effects. In this regard, as previously mentioned, the difference in chemotherapy regimens and their efficacy between pbAAC and iAAC could represent an additional potential confounding factor in the interpretation of the observed subtype‐specific benefit.
In conclusion, age, LN involvement, and R1 margin were independent predictors of both OS and RFS, whereas AC was not associated with better survival or reduced recurrence risk in the overall cohort. Subgroup analysis revealed that patients with pbAAC had a more aggressive disease course but experienced improved OS when treated with AC. Future multicenter studies are warranted to better define the optimal adjuvant strategies, particularly for patients with iAAC.
Funding
The authors have nothing to report.
Ethics Statement
Approval of the research protocol by an Institutional Reviewer Board: This study was approved by the Colorado Multiple Institutional Review Board (COMIRB #18‐2839).
Consent
The authors have nothing to report.
Conflicts of Interest
Marco Del Chiaro received an industry grant (Haemonetics Inc) to conduct a multicenter study. Marco Del Chiaro is the co‐principal investigator of an international multicenter study sponsored by Boston Scientific. Richard D. Schulick is a co‐inventor of a patent licensed to DynamiCure (managed by the University of Colorado) outside the submitted work. No other disclosures were reported from the other authors.
Supporting information
Figure S1: Kaplan–Meier survival curves in the overall cohort according to (a) histopathologic subtype, (b) grade of differentiation, (c) T stage, (d) N stage, (e) lymphovascular invasion, and (f) perineural invasion.
Table S1: Patient characteristics in the entire cohort and according to histopathologic subtype.
Table S2: Adjuvant chemotherapy regimens in the overall cohort and by histopathologic subtype.
Table S3: Recurrence free survival (RFS) rates at 1‐, 3‐, 5‐, and 10‐year according to histopathologic features in the overall cohort.
Acknowledgments
The authors have nothing to report.
Data Availability Statement
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.
References
- 1. de Jong E. J. M., Geurts S. M. E., van der Geest L. G., et al., “A Population‐Based Study on Incidence, Treatment, and Survival in Ampullary Cancer in The Netherlands,” European Journal of Surgical Oncology 47 (2021): 1742–1749, 10.1016/j.ejso.2021.02.028. [DOI] [PubMed] [Google Scholar]
- 2. Del Chiaro M., Cardin D. B., Christensen J. A., Chung V., Czito B., and Donahue T. R., “NCCN Guidelines Version 2.2025 Ampullary Adenocarcinoma NCCN Evidence Blocks TM,” (2025).
- 3. Abu Hilal M., Uijterwijk B. A., Lemmers D. H. L., et al., “The Brescia International Multidisciplinary Consensus Guidelines on the Optimal Pathology Assessment and Multidisciplinary Pathways of Non‐Pancreatic Neoplasms in and Around the Ampulla of Vater (PERIPAN),” United European Gastroenterology Journal 13 (2025): 1048–1068, 10.1002/UEG2.70074. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Ahn D. H. and Bekaii‐Saab T., “Ampullary Cancer: An Overview,” in American Society of Clinical Oncology Educational Book (ASCO Publications, 2014), 112–115, 10.14694/EDBOOK_AM.2014.34.112. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Rostain F., Hamza S., Drouillard A., Faivre J., Bouvier A. M., and Lepage C., “Trends in Incidence and Management of Cancer of the Ampulla of Vater,” World Journal of Gastroenterology 20 (2014): 10144, 10.3748/WJG.V20.I29.10144. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Moekotte A. L., van Roessel S., Malleo G., et al., “Development and External Validation of a Prediction Model for Survival in Patients With Resected Ampullary Adenocarcinoma,” European Journal of Surgical Oncology 46 (2020): 1717–1726, 10.1016/j.ejso.2020.04.011. [DOI] [PubMed] [Google Scholar]
- 7. Chang D. K., Jamieson N. B., Johns A. L., et al., “Histomolecular Phenotypes and Outcome in Adenocarcinoma of the Ampulla of Vater,” Journal of Clinical Oncology 31 (2013): 1348–1356, 10.1200/JCO.2012.46.8868. [DOI] [PubMed] [Google Scholar]
- 8. Uijterwijk B. A., Lemmers D. H., Ghidini M., et al., “The Road to Tailored Adjuvant Chemotherapy for All Four Non‐Pancreatic Periampullary Cancers: An International Multimethod Cohort Study,” British Journal of Cancer 131 (2024): 117–125, 10.1038/S41416-024-02692-W. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Nassour I., Hynan L. S., Christie A., et al., “Association of Adjuvant Therapy With Improved Survival in Ampullary Cancer: A National Cohort Study,” Journal of Gastrointestinal Surgery 22 (2018): 695–702, 10.1007/s11605-017-3624-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Moekotte A. L., Malleo G., van Roessel S., et al., “Gemcitabine‐Based Adjuvant Chemotherapy in Subtypes of Ampullary Adenocarcinoma: International Propensity Score‐Matched Cohort Study,” British Journal of Surgery 107 (2020): 1171–1182, 10.1002/BJS.11555. [DOI] [PubMed] [Google Scholar]
- 11. Lazaryan A., Kalmadi S., Almhanna K., Pelley R., and Kim R., “Predictors of Clinical Outcomes of Resected Ampullary Adenocarcinoma: A Single‐Institution Experience,” European Journal of Surgical Oncology 37 (2011): 791–797, 10.1016/j.ejso.2011.06.008. [DOI] [PubMed] [Google Scholar]
- 12. Doepker M. P., Thompson Z. J., Centeno B. A., Kim R. D., Wong J., and Hodul P. J., “Clinicopathologic and Survival Analysis of Resected Ampullary Adenocarcinoma,” Journal of Surgical Oncology 114 (2016): 170–175, 10.1002/JSO.24281. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. von Elm E., Altman D. G., Egger M., Pocock S. J., Gøtzsche P. C., and Vandenbroucke J. P., “Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement: Guidelines for Reporting Observational Studies,” BMJ [British Medical Journal] 335 (2007): 806, 10.1136/BMJ.39335.541782.AD. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Edge S. B. and Compton C. C., “The American Joint Committee on Cancer: The 7th Edition of the AJCC Cancer Staging Manual and the Future of TNM,” Annals of Surgical Oncology 17 (2010): 1471–1474, 10.1245/S10434-010-0985-4. [DOI] [PubMed] [Google Scholar]
- 15. Amin M. B., Greene F. L., Edge S. B., et al., “The Eighth Edition AJCC Cancer Staging Manual: Continuing to Build a Bridge From a Population‐Based to a More “Personalized” Approach to Cancer Staging,” CA: A Cancer Journal for Clinicians 67 (2017): 93–99, 10.3322/CAAC.21388. [DOI] [PubMed] [Google Scholar]
- 16. Qiao Q. L., Zhao Y. G., Ye M. L., et al., “Carcinoma of the Ampulla of Vater: Factors Influencing Long‐Term Survival of 127 Patients With Resection,” World Journal of Surgery 31 (2007): 137–143, 10.1007/S00268-006-0213-3. [DOI] [PubMed] [Google Scholar]
- 17. Moekotte A. L., Lof S., Van Roessel S., et al., “Histopathologic Predictors of Survival and Recurrence in Resected Ampullary Adenocarcinoma: International Multicenter Cohort Study,” Annals of Surgery 272 (2020): 1086–1093, 10.1097/SLA.0000000000003177. [DOI] [PubMed] [Google Scholar]
- 18. Finton S., Bolm L., Nebbia M., et al., “The Role of Adjuvant Therapy in Duodenal Adenocarcinoma and Intestinal Subtype Ampullary Carcinoma After Curative Resection,” Annals of Surgery 280 (2024): 986–992, 10.1097/SLA.0000000000006129. [DOI] [PubMed] [Google Scholar]
- 19. Carter J. T., Grenert J. P., Rubenstein L., Stewart L., and Way L. W., “Tumors of the Ampulla of Vater: Histopathologic Classification and Predictors of Survival,” Journal of the American College of Surgeons 207 (2008): 210–218, 10.1016/J.JAMCOLLSURG.2008.01.028. [DOI] [PubMed] [Google Scholar]
- 20. Shin D. W., Kim S., Jung K., et al., “Impact of Histopathological Type on the Prognosis of Ampullary Carcinoma: A Systematic Review and Meta‐Analysis,” European Journal of Surgical Oncology 49 (2023): 306–315, 10.1016/j.ejso.2022.10.001. [DOI] [PubMed] [Google Scholar]
- 21. Bolm L., Ohrner K., Nappo G., et al., “Adjuvant Therapy Is Associated With Improved Overall Survival in Patients With Pancreatobiliary or Mixed Subtype Ampullary Cancer After Pancreatoduodenectomy—A Multicenter Cohort Study,” Pancreatology 20 (2020): 433–441, 10.1016/j.pan.2020.01.009. [DOI] [PubMed] [Google Scholar]
- 22. Kim W. S., Choi D. W., Choi S. H., Heo J. S., You D., and Lee H. G., “Clinical Significance of Pathologic Subtype in Curatively Resected Ampulla of Vater Cancer,” Journal of Surgical Oncology 105 (2012): 266–272, 10.1002/JSO.22090. [DOI] [PubMed] [Google Scholar]
- 23. Roth G., Pellat A., Piessen G., et al., “A Simple Prognostic Score to Predict Recurrence After Pancreaticoduodenectomy for Ampullary Carcinoma: Results From the French Prospective FFCD‐AC Cohort,” ESMO Open 9 (2024): 103988, 10.1016/J.ESMOOP.2024.103988. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Bassi C., Marchegiani G., Giuliani T., et al., “Pancreatoduodenectomy at the Verona Pancreas Institute: The Evolution of Indications, Surgical Techniques, and Outcomes: A Retrospective Analysis of 3000 Consecutive Cases,” Annals of Surgery 276 (2022): 1029–1038, 10.1097/SLA.0000000000004753. [DOI] [PubMed] [Google Scholar]
- 25. Nakachi K., Ikeda M., Konishi M., et al., “Adjuvant S‐1 Compared With Observation in Resected Biliary Tract Cancer (JCOG1202, ASCOT): A Multicentre, Open‐Label, Randomised, Controlled, Phase 3 Trial,” Lancet 401 (2023): 195–203, 10.1016/S0140-6736(22)02038-4. [DOI] [PubMed] [Google Scholar]
- 26. Neoptolemos J. P., Moore M. J., Cox T. F., et al., “Effect of Adjuvant Chemotherapy With Fluorouracil Plus Folinic Acid or Gemcitabine vs Observation on Survival in Patients With Resected Periampullary Adenocarcinoma: The ESPAC‐3 Periampullary Cancer Randomized Trial,” JAMA 308 (2012): 147–156, 10.1001/JAMA.2012.7352. [DOI] [PubMed] [Google Scholar]
- 27. Duhn J., Strässer J., von Fritsch L., et al., “Adjuvant Chemotherapy Is Associated With Improved Survival in Advanced Ampullary Adenocarcinoma—A Population‐Based Analysis by the German Cancer Registry Group,” Journal of Clinical Medicine 14 (2025): 14, 10.3390/JCM14113869. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Oettle H., Post S., Neuhaus P., et al., “Adjuvant Chemotherapy With Gemcitabine vs Observation in Patients Undergoing Curative‐Intent Resection of Pancreatic Cancer: A Randomized Controlled Trial,” JAMA 297 (2007): 267–277, 10.1001/JAMA.297.3.267. [DOI] [PubMed] [Google Scholar]
- 29. De Back T., Nijskens I., Schafrat P., et al., “Evaluation of Systemic Treatments of Small Intestinal Adenocarcinomas: A Systematic Review and Meta‐Analysis,” JAMA Network Open 6, no. 2 (2023): e230631, 10.1001/JAMANETWORKOPEN.2023.0631. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Lai Z. W., Bolm L., Fuellgraf H., et al., “Characterization of Various Cell Lines From Different Ampullary Cancer Subtypes and Cancer Associated Fibroblast‐Mediated Responses,” BMC Cancer 16 (2016): 195, 10.1186/S12885-016-2193-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Figure S1: Kaplan–Meier survival curves in the overall cohort according to (a) histopathologic subtype, (b) grade of differentiation, (c) T stage, (d) N stage, (e) lymphovascular invasion, and (f) perineural invasion.
Table S1: Patient characteristics in the entire cohort and according to histopathologic subtype.
Table S2: Adjuvant chemotherapy regimens in the overall cohort and by histopathologic subtype.
Table S3: Recurrence free survival (RFS) rates at 1‐, 3‐, 5‐, and 10‐year according to histopathologic features in the overall cohort.
Data Availability Statement
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.