The aim of this study is to construct unimodal and multimodal prediction models for overall survival (OS), progression free survival (PFS), and treatment response (complete response+partial response) of advanced first-line chemotherapy based on the imaging, pathological, and genomic characteristics of patients, and evaluate their predictive efficacy.

Inclusion Criteria:

• 1. Patients with pancreatic ductal adenocarcinoma diagnosed by histology or cytology;
• 2. Male or female patients aged ≥ 18 years old;
• 3. In the past, except for first-line treatment, he has not received systematic treatment for unresectable local advanced or metastatic pancreatic cancer;
• 4. Have complete blood routine, biochemical, and tumor marker blood tests before first-line maintenance treatment;
• 5.According to RECIST 1.1 standard, there are measurable lesions at baseline;
• 6. ECOG physical fitness score is 0-1;
• 7. Expected lifespan of at least 3 months (assessed by researchers);
• 8. Female patients must be in a non pregnant and non lactating state.
• 9. Understand the significance of the purpose of this study and sign an informed consent form.

Exclusion Criteria:

• 1. Within 28 days prior to receiving first-line maintenance treatment, have undergone major surgery or trauma (such as major surgeries via abdomen, chest, etc.; excluding ultrasound-guided pancreatic fine needle aspiration biopsy [EUS-FNB], percutaneous liver biopsy, laparoscopic biopsy or subcutaneous mass resection biopsy, peripheral venous catheterization and biliary stent implantation, etc.).
• 2. Previously received allogeneic hematopoietic stem cell transplantation or organ transplantation.
• 3. Receive live vaccines (including attenuated live vaccines) within 28 days before receiving first-line maintenance treatment.
• 4. Those who have previously or currently suffered from interstitial pneumonia/lung disease and have been determined by the researcher to be inactive and do not require hormone therapy are excluded.
• 5. Past or current autoimmune diseases, including but not limited to Crohn's disease, ulcerative colitis, systemic lupus erythematosus, sarcoidosis, Wegener syndrome (granulomatous disease of polyangitis), Graves' disease, rheumatoid arthritis, pituitary inflammation, uveitis, autoimmune hepatitis, systemic sclerosis (scleroderma, etc.), Hashimoto's thyroiditis, autoimmune vasculitis, autoimmune neuropathy (Guillain Barre syndrome), etc. Except for the following cases: type I diabetes, hypothyroidism with stable hormone replacement therapy (including hypothyroidism caused by autoimmune thyroid disease), psoriasis or vitiligo without systemic treatment.
• 6. Within 14 days before receiving first-line maintenance therapy, systemic corticosteroids (>10mg/day prednisone or equivalent other corticosteroids, continuous treatment for ≥ 7 days) or immunosuppressive therapy are required. Excluding inhalation or local application of hormones; Or receiving physiological replacement doses of hormone therapy due to adrenal insufficiency; Allow short-term (<7 days) corticosteroids to be used for prevention (such as contrast agent allergies) or treatment of non autoimmune disorders.
• 7. Other malignant tumors were combined within 5 years before receiving first-line maintenance treatment, except cured skin squamous cell carcinoma, basal cell carcinoma, non basic invasive bladder cancer, and prostate/cervical/breast cancer in situ.
• 8. It is known that there is active central nervous system (CNS) metastasis and/or malignant meningitis. Note: If a patient has previously received treatment for brain metastases and the lesion has been clinically or radiologically stable for at least 14 days prior to receiving first-line maintenance treatment (confirmed by repeated imaging studies at least 4 weeks apart and conducted during the screening period), they may be considered for inclusion.
• 9. Uncontrolled cancer pain, where the anesthetic analgesics did not reach a stable dose at the time of enrollment.
• 10. Compression fractures of the spine that have not been treated with surgery and/or radiation therapy; The treated spinal compression fracture requires at least 2 weeks of disease stabilization before receiving first-line maintenance treatment.
• 11. Presence of malabsorption syndrome, inability to take oral medication, or severe gastrointestinal diseases such as poorly controlled gastrointestinal inflammatory lesions (active Crohn's disease or ulcerative colitis), or high-risk gastrointestinal bleeding or abdominal bleeding.
• 12. There is clinical or imaging evidence of intestinal obstruction, or the etiology of recurrent intestinal obstruction has not been ruled out.
• 13. There is active infection with poor systemic treatment control.
• 14. Patients with known human immunodeficiency virus (HIV) infection, active hepatitis B virus (HBV) or hepatitis C virus (HCV) infection, or active syphilis and tuberculosis. Note: Subjects who are HbsAg positive and/or HCV antibody positive during the screening period must undergo HBV DNA or HCV RNA testing. If the test result is negative, they can be enrolled. HbsAg positive subjects must be monitored for HBV DNA during the treatment process.
• 15. Patients who are known to be allergic to drugs such as gemcitabine, albumin paclitaxel, capecitabine, 5-FU, or irinotecan, as well as their drug components, may be included in this study if they choose to receive maintenance treatment with oral or intravenous chemotherapy preparations other than the aforementioned allergic chemotherapy drugs.
• 16. It is known that there are mental illnesses or substance abuse conditions that can interfere with patients' compliance to participate in the study.
• 17. According to the researcher's determination, there are any medical history, treatment, laboratory test abnormalities, or other situations that may confuse the research results, interfere with patient compliance, or harm patient interests.

1. The retrospective training set consists of 300 cases from Renji Hospital affiliated to Shanghai Jiao Tong University School of Medicine;
2. The retrospective test set includes 50-100 cases from Zhongshan Hospital Affiliated to Fudan University, 50-100 cases from Third Hospital Affiliated to Zhongshan University, 50-100 cases from Sir Run Run Shaw Hospital Affiliated to Zhejiang University, 50-100 cases from Nanjing Drum Tower Hospital, and 178 cases from CPTAC-PDA database;
3. The prospective test set consists of 100 cases from Renji Hospital Affiliated to Shanghai Jiao Tong University School of Medicine.

IPD contains sensitive data of patients, and even after past identification processing, there is still a potential risk of privacy leakage. In addition, some informed consent forms did not explicitly include the clause of "IPD external sharing" when signed. If they are shared without the patient's supplementary authorization in the future, it will directly violate ethical approval requirements.

2. There is a lack of clear standards for the ownership and division of responsibilities of the results after data use; Due to improper use of data leading to adverse consequences, it is difficult to clearly define the rights and responsibilities of researchers and data users, and researchers may face unnecessary legal disputes.

In summary, the cautious attitude of researchers towards IPD sharing is essentially the result of seeking a balance between scientific value, ethical responsibility, practical costs, and rights protection.