Pancreato-duodenal cancers in Pakistan: Patterns of presentation, disease burden and outcomes at a tertiary cancer facility

Patient flow at SKMCHRC

All patients who register at our hospital’s walk-in clinic are reviewed in our gastrointestinal specialist clinic, where a detailed assessment is undertaken and further investigations are performed (Figure 1). Patients undergo staging radiological examinations (Computed Tomography/Magnetic Resonance Imaging), as well as Endoscopic Retrograde Cholangio-Pancreatography (ERCP) or Endoscopic Ultrasound (EUS), and Positron Emission Tomography (PET), where indicated. All cases are then discussed at the weekly Multi-Disciplinary Team (MDT) meeting, comprising trained hepatobiliary surgeons, gastroenterologists, diagnostic and interventional radiologists, pathologists, and medical and radiation oncologists. The MDT records the treatment plan and TNM (Tumor, Node, Metastasis) stage, as defined by the Union for International Cancer Control (UICC) classification system. During the study period, staging pathways evolved as pancreatic protocol CT imaging, MRI, and EUS-guided biopsy became more widely used. Multidisciplinary tumor board discussions and subspecialized hepatopancreatobiliary surgery and neoadjuvant chemotherapy protocols for borderline resectable disease were standardized over time.

Figure 1.

Those patients with operable cancer undergo Pancreatico-Duodenectomy (PD) or distal pancreatectomy depending on the location of the tumor. Patients with borderline resectable cancer, as defined by The National Comprehensive Cancer Network (NCCN), undergo neo-adjuvant chemotherapy first, followed by surgery if no progression (15). Unresectable cancer patients are offered palliative intent chemotherapy. The patients who are not suitable for both surgery and chemotherapy due to their general condition or co-existing disease receive palliative therapy only. Surgical resections were performed by a dedicated hepatopancreatobiliary surgical team comprising of 2 consultant surgeons with subspecialty expertise.

Adjuvant treatment for most patients at our institution is typically 24 weeks of gemcitabine and capecitabine; some other combinations are also used depending on the final histopathology. The standardized approach for borderline resectable pancreatic cancer at our institution includes the FOLFIRINOX (Leucovorin, Fluorouracil, Irinotecan, Oxaliplatin) regimen. Patients with good performance status who get recurrent disease are offered systemic salvage therapy. Patients are followed for surveillance and imaged per NCCN guidelines, unless clinically indicated.

Data recording

General information and Clinicopathological data of patients, who were registered with Shaukat Khanum Memorial Cancer Hospital and Research Centre (SKMCH&RC) Pakistan, with a radiological or pathological diagnosis of duodenal, peri-ampullary, pancreatic and distal bile duct tumour was retrieved retrospectively from the prospectively maintained Hospital Health Information system (HIS) using International Classification of Diseases (ICD) code ranges for duodenum (C17.0), ampullary tumours (C24.1 to 24.9) and pancreas (C25.0 to C25.9). Patients with radiological diagnosis were also included, as we felt this was important to understand the true state of pancreatic-duodenal cancers in real-world clinical practice, since these patients usually undergo invasive or palliative procedures from the outset.

All patients were followed for 5 years from registration to allow for survival analysis. Survival time was measured in months. Patients who died of pancreatic cancer during the follow-up period were considered to have complete data. Those who survived beyond the study period or were lost to follow-up at the end of the follow-up period were treated as having censored data. Patients who refused treatment or were lost to follow-up before treatment initiation were excluded from the survival analysis. Patients under 18 years of age and those who underwent emergency surgery were also excluded. Patients with non-cancer diagnoses or pre-malignant tumors were also excluded to avoid bias. To have any meaningful disease-free survival (DFS) or overall survival (OS) analysis, the surgical cohort was limited to 31st Dember 2020 to allow a 5-year DFS and OS analysis (Figure 2).

Figure 2.

Demographic data recorded included age, gender, co-morbidities, radiological investigations, endoscopic procedures, American Society of Anaesthesiologists (ASA) score, pre-operative histopathology, clinical staging, and MDT decisions (Table 1). Neuroendocrine tumors were defined according to the WHO classification, and data were recorded into three grades: Grade I, Grade 2, and Grade 3, based on Ki-67 levels.

Peri-operative data included margin status, which was further divided into 3 groups: negative margin (R0), positive margin (including microscopic [R1] and macroscopic [R2] as defined by the Royal College of Pathologists. Other data included reoperations/readmissions within 30 days, mortality within 90 days, Surgical Site Infections (SSIs), and length of stay (LOS). Post Operative Pancreatic fistula (POPF) and Post-Pancreatectomy Hemorrhage (PPH) were recorded as defined by the International Study Group on Pancreatic Surgery (ISGPS) (16,17). Morbidity was recorded according to the Clavien-Dindo (CD) classification (18).

Textbook outcomes were defined as the absence of major postoperative complications, no perioperative mortality, no prolonged hospital stay, no readmission, negative resection margins, and adequate lymph node retrieval according to previously published criteria (19).

Recurrence was defined as radiological or biopsy-proven recurrent disease. Disease-free survival (DFS) was measured from the date of surgery to the first documented recurrence or death. Overall survival (OS) was calculated from the date of surgery to the date of death or last follow-up. Death from any cause was considered an event for overall survival analysis, while patients alive at the last follow-up were censored.

The study was exempted from informed consent requirements by the Institutional Review Board (IRB) under Exemption Number 04-12-25-04.

Statistical analysis

SPSS 27.0 statistical software was used in the statistical analysis. Continuous variables were reported as medians and categorical variables as percentages. The Kaplan–Meier method was used for survival analysis, and the log-rank test was used to compare groups. Variables with significance on univariate analysis were entered into multivariable Cox regression models to identify independent predictors of overall survival and disease-free survival. Statistical significance was set at p < 0.05.

Patient demographics

A total of 1653 patients were registered with Shaukat Khanum Memorial Cancer Hospital and Research Centre, with a diagnosis of either duodenal cancer, peri-ampullary cancer, pancreatic cancer, or distal bile duct cancer. Of this cohort, 861 patients were eligible for the 5-year survival analysis (Figure 2). 68% of this patient cohort were male patients, and 32% were female. The age at presentation was 53 (Interquartile Range 19–90) years with a mean BMI score of 23 +/− 4.5. Of this cohort, 27% of patients were aged 60 years or older. The majority of patients presented with an Eastern Cooperative Oncology Group (ECOG) performance status of 1 (54.6%) and 0 (39%). The most common presentation was for jaundice (64.6%), followed by abdominal pain (26.7%). Co-morbidities commonly reported included diabetes (16.1%), hypertension (17.4%), and smoking (18.7%), being the most common (Table 1).

Pre-operative work-up

The ASA (American Society of Anesthesia) score was most frequently labeled as ASA 2 (72%) and ASA 3 (20%) in patients. Interestingly, 31 patients were marked as ASA 4; these patients underwent surgical intervention after optimization. There was a trend toward increased morbidity; however, mortality was not significantly higher, though the numbers are too small for conclusive interpretation (Table 2). ERCP with brush cytology or stenting was undertaken in 82% of the cohort. The main indication for performing primary ERCP at our institute was for diagnosis or symptom relief in jaundiced patients. The majority of the patients (60%) already had an ERCP with stent placement before presenting at our institution, and ERCP was only repeated at our organization for stent-related issues or diagnostic uncertainty. The diagnostic yield of brush cytology at our institute was 58%, with a sensitivity of 65%.

There is a higher use of EUS in our study, as it is performed for both staging and diagnostic purposes in patients with pancreaticoduodenal cancers, per our institutional policy, which mandated biopsy confirmation of malignancy during the study period. Many patients presented after external imaging with incomplete tissue diagnosis, and EUS was used to help differentiate between resectable and borderline tumors.

The area of the pancreas targeted for Endoscopic Ultrasound-guided Fine Needle Aspiration (EUS-FNA) included the pancreatic head (n= 362, 68.5%), pancreatic tail (n= 157, 29.8%), and pancreatic uncinate process (n= 35, 6.6%). The diagnostic yield of the cohort was 92.6%. In patients in whom EUS did not yield a diagnostic tissue sample, a repeat EUS was performed in the majority of cases, with some patients referred for CT-guided biopsy. In a very small sample, where repeat attempts at diagnosis failed, and there were high-risk features, upfront surgery was advised. The median tumor size was 25 mm (interquartile range 18–35 mm). The sensitivity and specificity of EUS-FNA were 93% and 97%, respectively. The most common diagnosis was adenocarcinoma (80%), followed by neuroendocrine tumor (7%). The most common site of disease was the peri-ampullary region (46.2%), followed by the duodenal region (30.8%) and then pancreatic cancer (17.9%) (Table 2).

Multi-disciplinary team discussion

All patients (100%) underwent discussion during the weekly MDT, with input from radiologists, HPB surgeons, gastroenterologists, pathologists, and oncologists. The most common TNM stage at presentation was Stage 3 (18%).

473 (55%) of patients in the cohort were not candidates for curative treatment due to the advanced nature of the disease or the patient’s frailty, and this cohort was offered palliative treatment. 51 patients were referred for Neo-adjuvant therapy given the locally advanced nature of the disease on staging investigations and based on NCCN criteria. 8 patients out of this cohort underwent an attempt at curative resection after completion of neo-adjuvant therapy. 6 patients underwent an R0 resection while 2 patients had R1 resection. All 8 patients went on to have adjuvant chemotherapy.

Surgery group

388 patients (45%) of the cohort were considered as candidates for upfront surgery, 74 patients (19%) ended up with a palliative bypass at the time of surgery, as they were found to have either metastasis or had irresectable disease due to the locally advanced nature of the disease. The majority of these patients had occult metastatic disease, borderline vascular involvement, or locally advanced tumors not fully appreciated on preoperative imaging, particularly during the earlier years of the study period (2000–2015). High-resolution pancreatic protocol imaging and advanced staging modalities were not uniformly available. Additionally, a proportion of patients had staging imaging performed externally prior to referral, with variability in imaging quality and pancreatic protocol CT acquisition. A PET scan is not routinely performed at our setup; we only order one if a triphasic CT performed within 4–6 weeks of surgery shows suspicious features.

The total number of patients who underwent curative-intent resection surgery was 314. This number increased from 2–6 patients per year to 48 patients per year between 2000 and 2020. The main improvement in numbers occurred after 2014, when an HPB team was established, and 2 surgeons with a specialist interest in HPB developed the service. Pancreatico-duodenectomy was performed in 254 patients, distal pancreatectomy was done in 56 patients, and Total pancreatectomy in 4 patients (Figure 3).

Figure 3.

Figure 4.

Figure 5.

Postoperatively, the median length of stay in HDU was 1.3 +/− 1.9 days. The overall length of stay for the operated cohort was 9 +/− 5.8 days. The reoperation and readmission rates within 30 days were 4.7% (17 patients) and 10% (34 patients), respectively. Table 2 shows morbidity and mortality graded according to the Clavien-Dindo classification 18. Severe complication (i.e., CD>3a) occurred in 72 patients (21.1%), with a reduced frequency of other grades of complication. 55 (16.2%) patients developed a Post-Operative Pancreatic Fistula (POPF) grade A or biochemical leak, whereas 33 patients (9.8%) had POPF grade B. Only 15 patients (4.5%) developed a POPF type C, and 1 patient each had a hepaticojejunostomy or gastrojejunostomy site leak requiring re-look laparotomy.

Overall, 60% of patients undergoing surgery achieved Textbook Outcomes, defined as the absence of major postoperative complications, no perioperative mortality, no prolonged hospital stays, no readmission, negative resection margins, and adequate lymph node retrieval.

Overall survival was analyzed according to tumor type, location, stage, and postoperative complications. Patients with duodenal cancers demonstrated the longest survival compared to those with pancreatic cancers.

Adjuvant treatment

65% of patients went on to have adjuvant chemotherapy, and the majority of the patients were able to start chemotherapy within 4–6 weeks of surgery, with a small number starting between 8–12 weeks due to multi-factorial reason but mostly due to delayed recovery from surgery. 24% of patients opted not to have adjuvant chemotherapy, and 30% of this cohort, who did not undergo adjuvant chemotherapy, experienced surgical complications, which may have influenced decision-making. The majority of patients completed 6–12 weeks of chemotherapy, with only 20% requiring a dose reduction or a therapy break. OS and DFS of those patients who completed chemotherapy showed better outcomes than the cohort that did not complete chemotherapy.

Palliative chemotherapy

Out of the 473 patients not deemed suitable for curative attempt, 347 patients (73%) were offered palliative chemotherapy. In the cohort of patients who started palliative chemotherapy, 159 patients (45%) tolerated only 3 or fewer cycles due to multifactorial reasons, 123 patients (35%) tolerated 4–7 cycles, and only 61 patients (18%) tolerated 8–12 cycles. This cohort understandably had the longest overall survival.

The majority of these patients received Gemcitabine-based chemotherapy (n=99, 28%), with lower rates with other regimens. Interestingly, FOLFIRINOX or its modifications were used in 52 patients (15%) in this cohort, with the hope of controlling the disease. Only 18 patients receiving this regimen tolerated 6 or more cycles.

Best supportive care

126 patients (27%) out of the total cohort of patients were thought not to be fit enough for palliative chemotherapy, or they opted for no treatment, and they were offered best supportive care. All the patients were offered psychological counseling and palliative care services. This cohort of patients was not included in the OS analysis because follow-up was inconsistent, and a large number were lost to follow-up after the initial consultation.

Palliative endoscopic procedures

Those patients undergoing palliative chemotherapy or best supportive care and who needed biliary or luminal stenting to relieve obstructive symptoms underwent Endoscopic Retrograde Cholangio-Pancreatography (ERCP) or Endoscopic Ultrasound (EUS) guided intervention for both achieving histological diagnosis or decompression. Of the 473 patients not undergoing curative treatment, ERCP and stenting were performed in 182 (38%), and EUS in 91 (19%).

Recurrence patterns

Recurrence was observed in 40% of patients, with the liver as the most common site; lung and local recurrences were less frequent. 85% of patients developed recurrence within the first 2 years. Lymph node positivity, vascular invasion, and stage at presentation were associated with early disease recurrence.

Survival analysis

The overall mean survival for the entire cohort that underwent curative-intent surgery, excluding the NET group, was 45.2 months (95% CI: 42.3–48.1). Figure 6 shows the median overall survival by anatomical location of the cancer. During the follow-up period, 70 deaths occurred (26.5%) for the entire cohort, while 194 patients (73.5%) were censored. The event rates were spread to varying degrees in each tumor type and included 3 deaths (18.7%) in the duodenal cancer group, 41 deaths (23.3%) in the peri-ampullary group, and 26 deaths (36.1%) in the pancreatic cancer group.

Figure 6.

The mean overall survival for the duodenal cancer group was 49.5 months (95% CI:38.9–60.1). After a modest decline in the first year, the survival curve stabilizes, remaining close to 75–78% survival at 5 years. The peri-ampullary cancer group showed a mean survival of 47.4 months (95% CI: 43.9–50.8) with a gradual decline over the years, particularly after 2 years, showing a moderate survival. The pancreatic cancer group carried a mean survival of 37.9 months, and in particular, the body and tail cancers carried a mean prognosis of 28.8 months (95% CI: 21.3–36.2). The survival curve drops sharply after the 1st year of surgery, with frequent events, reaching a 40% 5-year overall survival in the cohort. Survival differed significantly between groups (log-rank χ² = 7.50, df = 2, p = 0.024).

Univariate and multivariate analyses of the whole cohort (Tables 3 and 4) showed that tumor site, presence of POPF, and margin status were associated with improved OS. Interestingly, only the margin status was associated with DFS in the multivariate analysis. Age, ASA status, ECOG, and BMI did not correlate with DFS or OS. The association between POPF and improved survival likely reflects selection bias and survivorship bias, as patients with softer pancreatic texture and less aggressive tumor biology are more likely to develop fistulas while also demonstrating improved oncological outcomes.

Neuro-endocrine tumor

Neuroendocrine tumors were analyzed separately as a subgroup to avoid bias, since their prognosis differs from that of adenocarcinoma, and their inclusion could bias the survival analysis. A total of 134 patients with a diagnosis of Gastroentero-Pancreatic Neuro-Endocrine-Tumors (GEP-NET) were analyzed for this study. 60 patients were diagnosed with pancreatic NET (pNET). The mean age at presentation was 47 years, gender distribution was similar to that of the adenocarcinoma cohort. The majority of patients had Grade 1 disease (59%), followed by Grade 2 (23.9%). About 73.1% patients had localized disease at presentation, and 3 (2.2%) patients were found to have an associated with MEN I syndrome, with the rest being considered as sporadic presentations. 55.4% of patients underwent surgical resection with curative intent, whereas palliative surgery was performed in 44.6% of the cohort. 48.5% patients received systemic therapy in the form of chemotherapy, (5.9%) received Somatostatin analogues and (45.5%) did not receive any systemic therapy. The 5-year OS was 94% in the completely resected group, 79% in patients with incomplete resection, and 43% in patients with unresectable disease. Patients in the unresected group achieved a median overall survival of 23±3 months. Based on disease grade, the 5-year survival was 88% in patients with low-grade tumors and 57% in those with intermediate-grade tumors (Figure 6). High-grade tumor patients were not able to survive the 5-year window, and the median overall survival for this group was 8±4 months. DFS was similarly better in the low-grade NET group compared to the intermediate-grade NET group (Figure 7).

Figure 7.

Figure 8.