Introduction

The majority of studies on the prognostic factors of pancreatic cancer focus on resected cancers, which represent a minority (1). Indeed, nearly 30 % of patients do not undergo surgery, and among those who do, 55 to 70 % do not benefit from tumor resection (2). Without cancer excision, patient survival is very short. (1) Understanding prognostic factors and developing scores is particularly valuable if they have a decisional impact. If the diagnosis is made and survival probabilities are identified, it helps identify patients who will potentially benefit from surgical intervention (2). Furthermore, given the limited effectiveness of chemotherapy, it seems advisable to offer it only to patients with a sufficient life expectancy, by enrolling them in prospective trials. Few data are reported in the literature regarding the choice of adjuvant treatments (chemotherapy, radiotherapy, or a combination of both), which should consider the expected survival time of patients (5). No studies have been conducted to date on the Algerian population regarding pancreatic cancer and the identification of prognostic factors (3). The rationale for identifying such prognostic factors is to improve survival, which is particularly short (4). The aim of this study is to conduct a descriptive prognostic study with prospective data collection to identify prognostic factors in patients with metastatic pancreatic cancer.

1. Material and Methods

A descriptive, observational, prognostic, and prospective study was conducted in the Medical Oncology Department of the University Hospital of Oran from November 1, 1954, to December 2021.

1.1 Study Population

The target population consisted of patients with metastatic pancreatic cancer treated in the Medical Oncology Department of EHU Oran.

1.2 Inclusion Criteria

Inclusion criteria for the study population included patients with metastatic pancreatic cancer confirmed by histological and/or cytological or biological evidence, with at least one lesion measurable or evaluable by CT angiography and hepatic MRI according to RECIST Stage IV criteria, Age> 18, WHO performance status 1, 2, 3, 4, and written and signed consent.

1.3 Exclusion Criteria

Exclusion criteria comprised pregnancy or breastfeeding, treatment with other experimental antimitotics, progressive intercurrent infection, clinical follow-up impossible due to psychological, family, social, or geographical reasons, loss of sight, and refusal of treatment.

1.4 Physical Examination

A comprehensive physical examination was conducted to assess the general condition of the patients according to WHO criteria. This included measuring temperature, calculating body surface area based on weight and height, measuring blood pressure, evaluating for jaundice, hepatomegaly, ascites, lymphadenopathy, edema of the lower limbs, and other pathological conditions. The visual analogue scale was utilized to measure the intensity of pain.

1.5 Diagnostic Procedures

Diagnostic procedures included pancreatic ultrasound, thoraco-abdomino-pelvic CT scan with systematic injection (helical or multi-bar), or CT angiography to assess resectability. Retrograde cholangiopancreatography (ERCP) was performed in cases of biliary strictures, and PET scan was conducted to assess operability in the event of lymph node involvement. Liver MRI was performed to search for liver metastases. Endoscopic ultrasound was conducted for tumors without histological evidence with normal tumor markers. Electrocardiogram (ECG) and echocardiography with assessment of ejection fraction were conducted with cardiology consultation. Bone scintigraphy was performed for cases with suspicion of warning signs and bone lesions on CT scan, and cerebral CT scan was conducted in the event of abnormal complete blood count (CBC), prothrombin time (PT), activated partial thromboplastin time (TCA), erythrocyte sedimentation rate (ESR), C-reactive protein (CRP), albumin level, creatinine level, protein level, bilirubin level, alkaline phosphatase level, transaminases, renal assessment, magnesium level, phosphorus level, potassium level, and calcium level.

1.6 Other Parameters Investigated

• Tumor markers: ACE, CA 19.9, CA 125, LDH.

• COVID, HBV, HCV, HIV serology.

• A pregnancy test in patients of childbearing age.

The patients in the study were presented to the multidisciplinary team (MDT) to discuss therapeutic modalities. The objective of this procedure was to propose the best therapeutic strategy for patients and thus improve results in terms of survival and quality of life. Patients received systematic LMWH treatment.

1.7 Proposed Treatment

• FOLFIRINOX: Oxaliplatin 85 mg/m2 over 2 h in 250 ml of G5%, then folinic acid 400 mg/m2 (or lfolinic 200 mg/m2) over 2 h in 250 ml G5%, then after 30 min start in Irinotecan 180 mg/m2 in 90 minutes, then 5-FU bolus 400 mg/m2 followed by 5-FU 2400 mg/m2 in continuous infusion for 46 hours in G5% using a portable infuser, pump, or portable syringe pump. Treatment repeated every 14 days. Adjuvant Gemcitabine for metastatic cases.

• Gemcitabine (adjuvant): 1000 mg/m² by infusion over 30 min on D1. Treatment repeated every week for 3 weeks if WHO SUP A 2. Metastatic PNET : PNET G1 G2 in metastatic situation :

• 1st Line: Somatuline LP, deep SC: 120 mg/month

• 2nd line: Sunitinib PO: 37.5 mg/day continuously. TEMCAP: Temozolomide PO: 200 mg/m2 x1/d from D10 to D14 (Precautions for the first course: Start temozolomide at 150 mg/m2 x1/d from D10 to D14 to assess tolerance) – Capecitabine PO: 600 mg/m2 x2/d from D1 to D14.

1.8 Supportive Care

Patients received systematic LMWH treatment.

• Treatment of obstructive symptoms: Biliary drainage was necessary in case of bilirubin> 1.5 times normal.

• The retrograde endoscopic route was recommended as first intention by placing a short metallic stent rather than a plastic stent.

• Duodenal stenosis: place a duodenal stent rather than a surgical gastrojejunostomy was indicated. Symptomatic treatment was introduced for all side effects.

We will evaluate the treatment’s effectiveness on survival, primarily focusing on PFS and OS. We will also evaluate the therapeutic response according to the RECIST criteria, defining a complete response as the disappearance of all target lesions, and a partial response or progression based on predetermined criteria. Additionally, we will assess the treatment regimen’s toxicity intensity using the criteria of NCI-CTCAE v 4.0, graded from 0 to 5, where 0 indicates the absence of toxicity and 5 denotes death secondary to a toxic effect.

Following treatment completion, patients will be regularly monitored through scheduled consultations as per the outlined calendar.

2. Results and discussion

During the inclusion period, one hundred and seventy-six (n =176) patients with pancreatic cancer were enrolled (Table 1).

.		Characteristics	Patients
	Total number	176	100.0
	Mean age (years)	64.51 ± 12.02 (25–86)
	Median age (years)	67
Sexes	Male	89
	Female	87
Performance status (IP-WHO)	0	16	9
	1	88	50
	2	61	35
	3	11	6
	4	0	0
	Time to diagnosis (months)	8.2 ± 0.17
	Medical Oncology Treatment Time (Days)	58 ± 8.9
Histology	Adenocarcinoma	143	81.3
	Tumor neuroendocrine	20	11.3
	Others	13	7.4
Tumor	T1	7	4
	T2	38	21.6
	T3	99	56.3
	T4	34	18.2
Lymph node involvement	N0	34	19.3
	N1	135	76.7
	N2	7	4
Metastases	M0	69	39
	M1	107	61
Stage	I	5	2.3
	II	11	5.7
	III	53	29.5
	IV	107	61

2.1. Demographic Characteristics and Symptoms

Our study population consisted of 89 men (50.6%) and 87 women (49.4%), resulting in a sex ratio of 1.02. The mean age at diagnosis was 64.51 ± 12.02 years. In our cohort, all patients had a history of diabetes, and 16 patients (9%) had a family history of pancreatic cancer. The combination of alcohol and tobacco use was observed in 95.6% of cases (P=0.0001). The symptoms varied depending on the anatomical location of the pancreas, with weight loss being the most common at 57%, followed by anorexia at 48%, asthenia at 43%, jaundice at 40%, pain at 34.4%, and ascites at 17%. Hormonal secretion was present in 1% of our population.

2.2. Tumor Characteristics, Biomarkers, Staging, and Metastasis

2.2.1. Tumor Characteristics and Biomarkers

We found that 48% of patients had tumors in the pancreas body, 38% in the head, and 14% in the tail. Only two patients had a double location, with one having tumors in the head and tail, and the other having tumors in the head and body.

Furthermore, 54% of patients had an initial CEA level above normal. The difference in CEA distribution and tumor location was statistically insignificant (p=0.0004). We observed that 67.6% of patients had an initial CA 19.9 level above normal.

The difference in CA 19.9 distribution and tumor location was statistically insignificant (p=0.0004). However, 65% of patients with COVID-19 had elevated CA19.9 markers.

2.2.2. Tumor Staging and Metastasis

CT angiography identified 90% of non-resectable tumors with a size>3cm (51% borderline and 44% locally advanced). Liver MRI detected liver metastases in 60% of tumors with a size <2cm, which were not detected on CT scans.

Using the UICC staging system, we found that more than half of the tumors were classified as stage IV (107 patients), while the other half were predominantly stage III (29.5%). The most common secondary site was the liver in 63% of cases, followed by peritoneal metastases in 30%, pulmonary metastases in 32%, and less commonly, ovarian and bone metastases. Ductal adenocarcinoma was the most frequently observed histological type (81.3%), while 11.3% had neuroendocrine tumors. In our study, 54% of patients received palliative chemotherapy, with 75% receiving first-line treatment, 23.6% receiving second-line treatment, and only 1.38% receiving third-line treatment. Among neuroendocrine tumors treated with first-line somatostatin analogues in 3% of the population, a response rate of 33.33% and a resection rate of 40% were achieved, with 50% experiencing stability, allowing primary tumor and liver metastasis resection in 28% of cases after treatment.

2.3. Prognostic Factors and Survival Analysis

In our series (Table 2), we restricted the age to 70 years old. The median survival of patients over 70 years old was 7 months. Overall survival for female subjects was better than males (13 months for females vs. 9 months for males), with estimated 50% and 45% survival rates at one year, respectively. For ethnic race, the median survival was 12 months for white individuals and 11 months for black individuals. The median survival of malnourished patients (BMI <18.5) was shorter than patients with a BMI> 18.5 (5 months vs. 14 months). We observed better survival for patients with normal CA19.9 levels compared to patients with high CA19.9 levels (14 months vs. 11 months). Conversely, the difference was significant (p= 0.006) compared to patients with high ACE levels, with survival of 8 months vs. 17 months in patients with normal ACE levels. Pathological levels of PA and LDH were associated with mediocre survival, around 13 months. We observed that overall survival was better in patients with T2 tumors, with a median of 23 months. Conversely, patients with T4 stage tumors had an overall survival of 7 months. According to our analysis, patients who started with monotherapy had a longer duration and were eligible for second-line treatment compared to patients who received first-line dual or triple therapy. We achieved better survival for neuroendocrine neoplasms (NNE) with estimated one- and two-year survival rates of 56% and 42% respectively, compared to adenocarcinomas, where the survival estimate was 44%. The median survival was better in neuroendocrine tumors, at around 20 months compared to 11 months for adenocarcinomas. Absence of lymph node involvement was correlated with better survival (27 months), while median survival in patients with N2 lymph node involvement was mediocre (2 months). This study enrolled 176 pancreatic cancer patients to identify prognostic factors for survival.

2.3.1. Age and Sex

Given the limitations of the data, some patients are not administered chemotherapy based on the doctor’s decision or the wishes of the patient’s family. This highlights the reluctance of clinicians to offer chemotherapy to older patients with advanced pancreatic cancer, as discussed by Kuroda et al (8). Tas et al also reported poorer median overall survival in elderly patients (>60 years) with metastatic disease compared to younger subjects (<60 years) (148 vs 198 days, P= 0.039), with 1-year survival rates of 3% and 10%, respectively (7).

The median survival of patients over 70 years old is 7 months compared to 16 months in those under 70 years old (P=0.02). We observed 1-year survival rates of 33% and 50% respectively.

We observed favorable survival outcomes compared to the literature, possibly due to the treatment provided and the opportunity given to the patient to receive combination therapy. Thus, we confirm that age is a prognostic factor. In Li et al’s study, receiving palliative chemotherapy was associated with better survival in patients aged ≥ 75 years (survival of 7.9 months), compared to untreated patients (survival of 2.3 months).

Recent studies based on the SEER registry also report poorer median overall survival in women than in men (9). Similarly, we found better overall survival in women compared to men (13 months for women and 9 months for men), with estimated 50% and 45% survival rates at one year, respectively. This difference is significant, confirming the hypothesis that gender influences patient survival.

2.3.2. BMI and Albuminemia

The nutritional status of the patient at diagnosis, as indicated by albumin levels, has emerged as a predictive factor for response to chemotherapy. Several studies have identified serum albumin as a prognostic factor in advanced pancreatic cancer (10). For instance, in a retrospective study, Wang et al. demonstrated that albumin levels below the normal range are associated with a poor overall survival prognosis (11).

In our series, the median survival of malnourished patients (BMI <18.5) was less than that of patients with a BMI> 18.5, at around 5 months versus 14 months (p=0.002). Regarding patients with albumin levels <30, survival was 23 months compared to 24 months in patients with albumin levels> 30. We conclude that undernutrition is a prognostic factor for survival.

2.3.3. Tumor Markers

Maisey et al. demonstrated significantly shorter overall survival in patients with elevated CA19.9 and CEA (13). In another study, Bergers et al. did not find any significant difference between metastatic patients and those with locally advanced disease in overall survival or response to chemotherapy (12).

In our study, we found better survival for patients with normal CA19.9 compared to those with high CA19.9 levels: 14 months versus 11 months, respectively. Conversely, the difference was highly significant compared to patients with high CEA levels (p=0.006); survival was 8 months versus 17 months in patients with normal CEA levels. It appears that CEA levels influence survival.

A high level of lactate dehydrogenase (LDH) and/or alkaline phosphatase (ALP) has been correlated with poor prognosis in several studies (13). In our study, elevated LDH and ALP levels were associated with poor survival, approximately 13 months.

2.3.4. Classification T, N, M

An important factor influencing survival is the absence of lymph node invasion, with 5-year survival rates of 36% and 14% respectively in the absence or presence of invaded lymph nodes (14). The prognosis becomes unfavorable when, in addition to lymph node invasion, there is positivity of the resection margins, undifferentiated nature of the tumor, and a size greater than 2 cm (15).

In our cohort, overall survival in patients with T2 tumors is better, with a median of 23 months. Conversely, patients with T4 tumors have an overall survival of 7 months. Comparison of these data reveals a significant difference (p=0.004). The absence of lymph node involvement is correlated with better survival, approximately 27 months. However, median survival in patients with N2 lymph node involvement is mediocre (2 months). The comparison of these data shows a significant difference (p=0.01). Absence of metastases is also correlated with better survival (24 months) compared to patients with metastases (7 months). The comparison of these curves demonstrates a significant difference (p <0.0001). Our results align with those of B Trigui et al, where survival in patients with lymph node involvement and the presence of metastases is 4 months and 3 months, respectively. Regarding tumor size, better survival was observed compared to the same series when the tumor size was <4 cm (survival was 5 months). In conclusion, T, N, M influence survival.

2.3.5. Ascites

The best overall survival was observed in patients without ascites, with a mediocre survival of around 6 months and an estimated survival of 24% at one year. However, comparison of the data reveals a significant difference (p=0.04).

Our results are statistically stronger than those of a prospective, multicenter study conducted by 25 French hospital teams (16,17), which demonstrated a survival rate of only 2 months. Therefore, we can conclude that ascites is a factor influencing survival.

2.3.6. Nausea and Vomiting

Our data analysis was comparable to that of the original MPACT study (Von Ho DD et al, NEJM 2013). The quality of life score at baseline was correlated with tumor volume and comorbidities. It appears to be a significant indicator in multivariate physical function and vomiting (18,19).

In our study, survival was influenced by the onset of nausea and vomiting G2, and the estimated one-year survival rates were 35% and 30% with a median survival of 7 and 2 months (p= 0.004).

We seem to confirm that chemotherapy-induced nausea and vomiting influence survival. However, in our study, we could not confirm this hypothesis since survival was the same in both populations (12 months vs. 13 months). This is probably due to the preventive treatment in all patients (20).

The median overall survival was estimated at 12.8 months, and progression-free survival was estimated at 10 months. Our results are similar to those of the large phase III studies PRODIGE 4 and MPACT, where survival was respectively 11.1, 6.8, 6.7, and 8.5 months. The progression-free survival was 6.4, 3.3, 3.7, and 5.5 months.

Conclusion

Eleven prognostic factors have been identified : age, sex, malnutrition, ACE, T, N, M, and stage, ascites, nausea, and vomiting. Developing a score using these prognostic factors should facilitate providing elements for therapeutic decision-making regarding the indication for surgery, chemotherapy, radiotherapy, and/or palliative care. It is essential to identify, during consultation, patients at high risk of serious complications. This identification should extend until the proposed treatment, firstly to select the right candidates for surgical intervention or systemic treatment, and secondly, to provide them with clear and impartial information that is ethically essential for obtaining consent. a computer application is necessary and established in court