In recent years, DM has gradually been regarded as a manifestation of pancreatic cancer. However, more importantly, DM appears typically 2–3 years earlier than PC, making DM important in early PC detection.
Diagnosis of PC
Numerous international guidelines recommend computed tomography (CT) as an initial measure for diagnosing suspected PC.70,71 Because of its low cost, CT is preferred over magnetic resonance imaging (MRI) as the first-line modality. In the CT set, the tumor detection rates were about 90%, and MRI shared a similar sensitivity.72 The classic CT feature of PC includes a hypoattenuating pancreatic mass, pancreatic duct dilatation, and atrophy of the upstream pancreas. Therefore, CT is also applied to screen patients at high risk for PC.73 In the early stage of PC, inhomogeneous parenchyma, interruption of the pancreatic duct, and loss of fatty marbling can be seen in the CT set, which was helpful for early detection of PC.74,75 Other measures for the diagnosis of PC include magnetic resonance cholangiopancreatography (MRCP) and endoscopic ultrasound (EUS).
Early detection of PC in DM patients
Given that the prevalence of PDAC in the general population is relatively low, population CT screening is cost-inefficient and challenging to achieve. Therefore, the identification of high-risk populations for pancreatic cancer is essential.
The key to early PC detection in DM patients is differentiating pancreatic cancer-related diabetes (PCRD) from common DM. New onset diabetes in PDAC often differs from T2DM in significant ways. For example, PCRD is usually concomitant with weight loss, while T2DM is commonly associated with weight gain.76 Besides, diabetes can be better controlled by weight loss in common DM, which is barely seen in PCRD. A cohort study using multivariate logistic regression analysis reported that BMI, the age of onset of diabetes, HBV infection, TBIL, ALT, Cr, APO-A1, and WBC are factors that could differentiate PC + DM from common DM.77
NOD is vital in the early diagnosis of pancreatic cancer. Sharma et al. developed a model called enriching new-onset diabetes for pancreatic cancer (END-PAC) to evaluate the risk of pancreatic cancer in patients with new-onset diabetes.78 The model contains the change of blood glucose, change of weight, and age at glycemically-defined new-onset diabetes. More significant changes in blood glucose, lower change of weight, and higher age lead to a higher risk of PC in patients with new-onset diabetes. Several studies have validated the END-PAC model. For example, a study by Boursi et al. focused on 5,408 patients with NOD, and according to their research comparing the high-risk group compared with the low-risk group, the sensitivity, specificity, positive predictive value, and negative predictive value of the model were 54.2%, 76.98%, 2.57%, and 99.4%, respectively.79 Similarly, a study of 13,947 NOD patients in a healthcare setting validated the END-PAC model. At the 3+ threshold, the sensitivity, specificity, PPV, and NPV were 62.6%, 78.5%, 2.0%, and 99.7%, respectively.80 In conclusion, current studies supported the robustness, generalizability, and clinical applicability of the END-PAC model. However, the original investigation and validation studies have some limitations. For instance, the definition of NOD was restricted and may exclude certain patients who truly had NOD. Besides, all the studies were retrospective. Therefore, more efforts are needed to validate the screening strategies for patients with NOD in real-world settings.
In recent years, many other metabolites have been shown to play a role in the discrimination of DM from PC+DM. With the help of 1H mass spectrometry and nuclear magnetic resonance (NMR) spectroscopy, alterations of many plasma metabolite concentrations can discriminate T2DM and PC, suggesting that some proposed plasma metabolite biomarkers can be included in the model to evaluate the risk of pancreatic cancer in patients with new-onset diabetes.81 Similarly, He et al. analyzed their data using principal components analysis (PCA) and orthogonal projection to latent structures (OPLS). They identified 62 differential metabolites that may function as individual biomarkers of PC.82 The droplet digital PCR (ddPCR) confirmed that miR-20b-5p, a kind of miRNA, showed a higher level in PC patients with new-onset diabetes, suggesting that miR-20b-5p achieved higher diagnostic accuracy than PC with new-onset diabetes.83
A study evaluated the cost-effectiveness of PDAC early detection strategy targeting high-risk new-onset diabetes patients using MRCP and positive MRI underwent EUS/fine-needle aspiration (FNA). They found out that considering a willingness to pay (WTP) threshold between $100,000 and $150,000 per quality-adjusted life-year, a minimum predicted three-year PDAC risk of 1.0% to 2.0% may be cost-effective.84 Advances in the early detection of PDAC in NOD were concluded in Table 2.78–80,84
Table 2Advances in the early detection of PDAC in NOD
Topic | Study | Conclusion |
---|
END-PAC | Sharma et al78 | END-PAC* can evaluate the risk of pancreatic cancer in patients with new-onset diabetes according to change in blood glucose, change of weight, and age at glycemically-defined new-onset diabetes |
| Boursi et al79 | The sensitivity, specificity, positive predictive value, and negative predictive value of the END-PAC model were 54.2%, 76.98%, 2.57%, and 99.4%, respectively. |
| Chen W. et al80 | At the 3+ threshold, the sensitivity, specificity, PPV, and NPV of the END-PAC model were 62.6%, 78.5%, 2.0%, and 99.7%. |
Cost-effectiveness | Wang L., et al84 | Considering a WTP* threshold between $100,000 and $150,000 per quality-adjusted life-year, a minimum predicted 3-year PDAC risk of 1.0% to 2.0% may be cost-effective |