Factors Driving Pancreatic Cancer Survival Rates

Abstract

Objectives:

To analyze trends driving the increase in the 5-year survival rate of pancreatic cancer over the last decade with a focus across disease stages.

Methods:

Pancreatic cancer survival data was analyzed by disease stage, age, and race/ethnicity using the most recent data from the Surveillance, Epidemiology, and End Results (SEER) program. Stage distribution and survival rates were combined to evaluate the contribution of each stage to the overall survival rate.

Results:

The most recent SEER 17 data registry reports an overall 5-year survival rate for pancreatic cancer of 12%, with rates varying significantly across stages and age groups. The overall 5-year survival, irrespective of stage at the time of diagnosis, increased from 6% for patients diagnosed in 2004 to 12% for those diagnosed in 2015. During this same period, the incidence rate of distant disease decreased from 60% to 53%; however, the 5-year survival rate only increased from 2% to 3%, suggesting minimal advancements in treatment outcomes for individuals identified with distant disease at the time of diagnosis. In contrast, the 5-year survival rate for localized disease rose from 24% to 46%, while the stage distribution only slightly increased from 11% to 14% during the same period. Survival rates for distant disease at 4-, 3-, 2-year did not increase, however, the 1-year survival increased from 14% to 22% (2004–2019). These trends suggest progress in short-term outcomes, aligning with the availability of new FDA-approved treatments for advanced or metastatic disease with a median overall survival of <1 year.

Conclusions:

This research confirms that the overall 5-year survival rate of 12% primarily reflects improved survival among those diagnosed with localized disease. Importantly, since only 14% of patients are diagnosed at this stage, the survival rate for most patients remains unchanged. This work also emphasizes the need for more research around variables that influence the overall 5-year survival rate, along with greater investment in early detection strategies to improve patient survival.

Pancreatic cancer is the third leading cause of cancer death in the United States¹ and is expected to overtake colorectal cancer as the second leading cause of cancer-related mortality by 2030.² In 2024, it is projected that ∼66,440 individuals in the United States will be diagnosed with pancreas cancer, with an estimated 51,750 deaths expected from the disease.¹ Over the past 2 decades, there has been a steady rise in the incidence of pancreas cancer. From 2001 to 2018, the incidence rate increased by ∼1% per year, while death rate from pancreas cancer increased by 0.2% annually between 2001 and 2019.³ The rise in both incidence and mortality underscores pancreatic cancer as one of the most lethal cancers with an overall 5-year survival rate of just 13%.^1,3 The 13% overall 5-year survival rate is based on analysis using the Surveillance, Epidemiology, and End Results (SEER) 22 registry. Importantly, the present study uses the SEER 17 registry, which reports a 12% survival rate. SEER 17 was selected for use in this research because it offers more variables compared with other groupings. The difference in overall survival 5-year survival rates observed between SEER 17 and SEER 22 is likely due to differences with data collection and reporting methods between the 2 registries.

Despite the diseases’ high morbidity and mortality, pancreatic ductal adenocarcinoma (PDAC), the most common type of pancreas cancer, often goes undetected due to delays in symptom identification.⁴ Early symptoms, including recent changes to blood sugar and weight, abdominal pain, or jaundice, are easily misdiagnosed. These symptoms warrant attention in the presence of certain risk factors, including a history of tobacco use, type 2 diabetes, chronic pancreatitis, a personal or family history of certain types of cancer, and other known genetic risk factors. The absence of early disease detection typical of PDAC is compounded by critical barriers to preventative health care among historically excluded communities, including a lack of culturally responsive primary care and inadequate health communication.⁵ In addition, certain socioeconomic factors may contribute to a delayed diagnosis.^6,7 Without the application of evidence-based risk stratification and early detection initiatives, the majority of PDAC patients continue to receive late-stage diagnosis, when therapies are less effective.

The most effective PDAC treatments integrate surgical resection and chemotherapy. Unfortunately, only 12% of PDAC patients are eligible for this regimen due to the advanced stage of disease at the time of diagnosis, and only 10% are candidates for potentially curative surgery.^8–10 Currently, up to 80% of PDAC patients receive their diagnoses after the disease has advanced to a stage where treatment is often limited to palliative care, contributing to a 5-year survival rate of 3%.^8,11 Staging at time of diagnosis plays a major role in predicting survival rates, making this variable essential for understanding where improvements are needed. For patients with localized disease, the cancer is confined to the pancreas and can often be surgically removed. In these situations, the 5-year survival rate is 44%–46%.^1,12 For individuals diagnosed with regional disease, where cancer has spread to nearby structures or lymph nodes, the 5-year survival rate drops to 16%.¹ Finally, the 5-year survival rate drops to 3% for patients with distant disease, when the cancer has metastasized to distant organs.¹ This trend confirms that early diagnosis is key to increasing treatment options and improving overall prognosis.^13,14

The overall 5-year survival rate for pancreatic cancer has increased from 6% to 12% over the last decade.¹² This improvement is often conflated with better treatment options and improved outcomes across all stages of the disease. However, this change may more accurately reflect enhanced survival at specific disease stages. To better understand these changes and ensure that pancreatic cancer outcomes are improving for all patients, it is important to closely evaluate disease epidemiology data, with consideration of staging at time of diagnosis and adjoining survival rates. In this study, we utilize SEER data to investigate survival rates across various stages of the disease and provide an in-depth analysis to look at the underlying drivers that contribute to the observed change in the overall survival rate.

METHODS

Data on pancreatic cancer stage, age at diagnosis, sex, and race/ethnicity were obtained from the SEER 17 registry.¹² The SEER program collects and analyzes data on demographics, cancer statistics, and survival outcomes from population-based cancer registries covering about 35% of the US population. For this study, survival rates were calculated for patients diagnosed with pancreatic cancer between 2004 and 2019, stratified by stage, sex, and age with race and ethnicity combined. To assess survival rates at different time points, we analyzed 1- (2004–2019), 2- (2004–2018), 3- (2004–2017), 4- (2004–2016), and 5-year (2004–2015) survival rates. Each survival period included patients diagnosed within the specified range and followed until the end of the period (eg, patients diagnosed in 2004 were followed until 2009 for 5-year survival analysis).

Contribution of survival by stage was calculated by multiplying the survival rate by the stage distribution for 5-, 4-, 3- 2- and 1-year survival data from 2004 to the most recent available data. Recognizing the potential for imbalances in the distribution of patients across different stages at diagnosis, a weighted analysis strategy was employed.¹⁵ This approach was chosen over alternative methods, such as stratification or matching because it allows for the creation of pseudo-populations with equal representation of each stage. By weighting each patient’s contribution to the analysis based on the inverse probability of their observed stage, we effectively balanced the groups and mitigated the potential for confounding due to stage imbalances. This ensured a more robust comparison of survival rates across different demographic and clinical characteristics. In addition, weighting is a well-established and accepted method for analyzing SEER data, particularly in survival analysis. It is often recommended by SEER documentation and has been used extensively in published research.^16,17

The impact of a hypothetical shift in stage distribution on overall 5-year survival was evaluated using stage-specific survival rates and stage distribution. The initial stage distribution included 13% localized, 29% regional, 51% distant, and 7% unknown or unstaged cases, with corresponding survival rates of 46%, 15%, 3%, and 8%, respectively. A hypothetical stage shift reflecting improvements in early detection, increased localized cases to 70%, reduced regional and distant cases to 9% and 14%, respectively, and left unknown or unstaged cases unchanged at 7%. Overall survival for the shifted scenario was calculated as the sum of the weighted survivals by stage.

RESULTS

Survival Rates and Stage Distribution

The overall 5-year survival increased from 5% to 11% for males and from 7% to 12% for females diagnosed in 2004 and 2015, respectively, where the year reflects the date of diagnosis (Fig. 1A). During this same period, the stage distribution of distant disease decreased from 60% in 2004 to 53% in 2015; however, the 5-year survival rate for distant disease only increased by 1%, from 2% in 2004 to 3% in 2015 (Fig. 1B). In contrast, the 5-year survival rate for localized disease rose from 24% to 46% during the same period, while the stage distribution only increased from 11% to 14%. Similarly, the 5-year survival rate for regional disease increased from 9% to 15% during the same period, while the stage distribution decreased from 31% to 29% (Fig. 1B).

FIGURE 1.

Survival and stage distribution. Pancreatic cancer 5-year survival rates for males and females diagnosed between 2004 and 2015 (A), stage distribution of localized, regional and distant disease compared with 5-year survival (B). Survival by stage and age group. 5-year relative survival rates of localized, regional, and distant disease by age. Green represents localized disease, blue represents regional disease, and purple represents distant disease (C).

Survival by Demographics

Across all stages, survival rates declined with increasing age. For patients with localized disease, the 5-year survival rate was 90% for those aged 40–44, 73% for ages 45–49, 72% for ages 50–54, 61% for ages 55–59, decreasing further to 33% for ages 75–79 and 15% for ages 80–84 (Fig. 1C). The 5-year survival rate for patients diagnosed with regional disease was 38% for those aged 40–44, 27% for ages 45–49, 18% for ages 50–54, 16% for ages 55–59, 18% for ages 60–64, dropping to 6% for ages 75–79. Similarly, the 5-year survival rate for patients diagnosed with distant disease was 15% for ages 40–44, 4% for ages 45–49, 5% for ages 50–54, and continued to decrease with advancing age (Fig. 1C). Analyzing survival rates by sex across different stages revealed consistent trends. For both males and females, survival rates decrease with advancing age. While there are variations in survival rates between males and females in certain age groups, the overall pattern of survival remains similar when the sexes are combined (Fig. 1).

Survival Trends by Stage

The 5-year survival for localized disease doubled from 24% in 2004 to 46% in 2015 (Fig. 2A). The 4-, 3- and 2-year survival also doubled from 2004 to 2016, 2004 to 2017, and 2004 to 2018, respectively (Fig. 2B and eFigure 1, Supplemental Digital Content 1, http://links.lww.com/MPA/B360). The 1-year survival for localized disease increased from 42% in 2004 to 66% in 2019 (Fig. 2C).

FIGURE 2.

The 5- (A), 3- (B), and 1-year (C) survival rates over time. The year of diagnosis indicates the year patients were diagnosed, followed by the specific time period (eg, for 5-year survival, patients diagnosed in 2015 were followed until 2020). Dark blue represents localized disease, orange represents regional disease, green represents distant disease, and light blue represents combined stages.

Increases were also observed for individuals with regional disease at the time of diagnosis. The 5-year survival rate for regional disease increased from 9% in 2004 to 15% in 2015 (year of diagnosis) (Fig. 2A). The 4-year survival rate increased from 11% in 2004 to 18% in 2016 (eFig. 1A, Supplemental Digital Content 1, http://links.lww.com/MPA/B360), while the 3-year survival rate increased from 14% in 2004 to 24% in 2017 (Fig. 2B). The 2-year survival rate increased from 21% in 2004 to 35% in 2018 (eFig. 1B, Supplemental Digital Content 1, http://links.lww.com/MPA/B360), and the 1-year survival rate increased from 41% in 2004 to 61% in 2019 (Fig. 2C).

Survival rates for distant disease showed only slight increases over time. The 5-year survival rate rose from 1.9% in 2004 to 2.6% in 2015 (Fig. 2A). The 4-year survival rate increased from 2.5% in 2004 to 3.5% in 2016 (eFig. 1A, Supplemental Digital Content 1, http://links.lww.com/MPA/B360), while the 3-year survival increased from 3.2% in 2004 to 5.2% in 2017 (Fig. 2B). 2-year survival increased from 5.2% in 2004 to 9% in 2018 (eFig. 1B, Supplemental Digital Content 1, http://links.lww.com/MPA/B360). The 1-year survival for distant disease increased from 14% in 2004 to 22% in 2019 (Fig. 2C).

Survival Weighted by Stage

The 5-year survival rate increased from 6% in 2004 to 12% in 2015 (Fig. 3 and eFig. 2, Supplemental Digital Content 1, http://links.lww.com/MPA/B360). During this period, the contribution of distant disease to the 5-year survival rate (% contribution overall survival) increased slightly from 1.1% to 1.4% (Fig. 3A). The contribution from regional disease increased from 2.8% to 4.8%, while localized disease increased from 2.4% in 2004 to 6.3% in 2015 (Fig. 3A). Similar trends were observed for the 3-year survival rate (Fig. 3B). The overall 3-year survival rate increased from 8.4% in 2004 to 17% in 2017. Contributions to the 3-year survival rate from distant disease increased from 1.9% to 2.8%. Meanwhile, the contribution from regional disease increased from 4.4% to 7.6%, and localized disease contribution increased from 2.5% to 7.2% (Fig. 3B).

FIGURE 3.

Weighted survival by stage. A weighted strategy was adopted for comparative analysis for 5- (A), 3- (B), and 1-year (C) survival rates combined with stage. The calculated weighted survival rates for each stage do not add up to the combined staged rate due to rounding and because the stage categories do not sum to 100%. This is because not all cases in the SEER database have sufficient information for staging. The year of diagnosis indicates the year patients were diagnosed, followed by the specified period (eg, for 5-year survival, patients diagnosed in 2015 were followed until 2020). Dark blue represents localized disease, orange represents regional disease, green represents distant disease, and light blue represents combined stages.

For 5-year survival in 2015, distant disease contributed only 1% of the total 12% survival rate, while regional and localized diseases contributed 11% (Fig. 3A). In contrast, for 1-year survival, the contributions are more evenly distributed: distant disease contributed 11.3%, localized disease contributed 11.6%, and regional disease contributed 14.8% (Fig. 3C).

The 1-year survival rate increased from 24% in 2004 to 41% in 2019. Contribution from distant disease increased from 8.2% in 2004 to 11.3% in 2019, while the contribution from regional disease increased from 12.8% to 14.8% and localized disease increased from 4.3% to 11.6% in the same period.

Impact of Stage Shift on Overall Survival

Despite stage-specific survival rates remaining constant, a hypothetical shift in stage distribution led to a significant increase in overall survival, rising from 12% to 34% (Table 1). In the initial stage distribution, 13% of cases were localized, 29% regional, 51% distant, and 7% were unknown or unstaged, with an overall survival rate of 12%. After a shift in stage distribution—hypothetically reflecting improvements in early detection—the proportion of localized cases increased to 70%, regional cases decreased to 9%, and distant cases decreased to 14%, while the proportion of unknown or unstaged cases remained unchanged at 7%.

TABLE 1.

Impact of Stage Shift on Survival

Stage	Stage distribution (%)	Survival rate (5-y) (%)	Shifted stage distribution (%)	Survival rate (5-y) after stage shift (%)
Localized	13	46	70	46
Regional	29	15	9	15
Distant	51	3	14	3
Unknown/unstaged	7	8	7	8
Overall		12		34

DISCUSSION

This study provides a comprehensive analysis of survival trends for pancreatic cancer over the past decade. The overall 5-year survival rate increased from 6% in 2004 to 12% in 2015 (Fig. 1A), however, the improvement in survival rates varied considerably based on the stage of disease and age at diagnosis. Notably, the 5-year survival of localized disease increased from 24% in 2004 to 46% in 2015 (Fig. 1B), reflecting the progress made in the management of early-stage disease. The rise in overall 5-year survival, however, does not adequately reflect the 5-year survival rate for distant disease, which has not seen substantial improvement, only increasing from 2% in 2004 to 3% in 2015 (Fig. 2A). Importantly, distant disease represents the largest group of pancreatic cancer patients (51%). The increase in 1-year survival for distant disease from 14% in 2004 to 22% in 2019 is further indicative of progress in short-term outcomes, aligning with the development and FDA-approval of new treatments for advanced/metastatic disease, which have a median overall survival of less than one year.^18,19

We assessed the contribution of each stage to the overall increase in survival by combining stage distribution with survival rates. The increase in the overall 5-year survival appears to primarily be due to improvements in diagnosis and treatment of localized disease. Regional disease also contributed to the change in the overall 5-year survival, while the impact of distant disease remained relatively stable. When looking at the 1-year survival, the contributions between distant (11.3%), localized (11.6%), and regional (14.8%) diseases were more balanced (Fig. 3). Analysis of survival rates by age, across all disease stages, suggests that treatment is more challenging for older patients (Fig. 1C).

SEER 22 is frequently referenced when analyzing survival rates for pancreatic cancer; this analysis relies on the SEER 17 registry. Although SEER 22 represents a larger population, it offers fewer available variables compared with other groupings and is limited to certain statistics. As a result, there are slight differences in survival statistics reported from other sources that have utilized SEER 22^1,3 and this study, such as the 5-year survival rates of 13% (SEER 22) versus 12% (SEER 17). Further, while SEER is an invaluable resource for understanding cancer trends in the United States and is one of the only publicly available data registries of its kind, it has limitations. SEER mainly provides basic tumor information (site, size, stage), but lacks detail about specific pancreatic cancer histologic subtypes, molecular markers, patient comorbidities, and treatment details. These limitations make it difficult to conduct in-depth, generalizable research on pancreatic cancer, particularly in personalized medicine, and the impact of specific patient characteristics on treatment outcomes. It is crucial to be aware of these limitations and consider them when interpreting SEER-based findings.

Our analysis suggests that improvements in the overall 5-year survival rate may be due to changes in the treatment of localized disease and increased awareness of symptoms and risk factors, but further research is needed to better understand factors that contribute to early diagnosis. This includes socioeconomic disparities impacting access to health care, variations in diagnostic practices, and the influence of comorbidities on treatment decisions. Delineating these confounders will enhance the accuracy of SEER-based analyses and contribute to more effective strategies for early detection and improved outcomes in pancreatic cancer.

While more robust data is needed to better understand factors contributing to improved outcomes, it is equally important to act on existing knowledge. Localized pancreatic cancer accounts for only 15% of cases, underscoring a significant disparity in outcomes based on the stage at diagnosis.¹ Complimenting this observation, if a stage shift occurred where localized disease increased from 13% to 70%, regional disease decreased to 9%, and metastatic disease to 14%, the overall survival would increase from 12% to 34% (Table 1). This projection highlights the need for enhanced early detection methods to shift the diagnosis toward earlier, more treatable stages.^11,20,21

The shift toward earlier-stage diagnosis has played a crucial role in improving survival outcomes in breast and colorectal cancers. Declines in breast cancer mortality are partly due to advances in screening, such as mammography, and treatment with endocrine therapy, which have increased the detection of localized tumors and improved survival rates.²² In colorectal cancer, increased screening has not only detected early-stage cancers but also enabled the removal of premalignant lesions, leading to a dramatic decrease in both late-stage and early-stage incidences.²³ These shifts emphasize the significant impact of early detection and intervention in improving cancer survival. Importantly, tools for early detection are available. Current early detection methods for pancreatic cancer include imaging techniques such as endoscopic ultrasound (EUS) and magnetic resonance imaging (MRI). EUS allows for detailed visualization of the pancreas and can facilitate biopsy procedures, while MRI offers a noninvasive way to detect early pancreatic lesions. Current research indicates that the use of high-resolution imaging and endoscopic ultrasound can assist in identifying early-stage pancreatic cancer.^21,24 A retrospective study of prediagnostic CT scans further supports this approach, finding that there were no cases of stage IV pancreatic cancer detected 6 months before diagnosis among patients.²⁵ This research indicates that risk-based imaging studies, such as a CT scan performed upon the new onset of diabetes, can increase the proportion of resectable tumors and improve outcomes.

Imaging strategies may be enhanced with the application of artificial intelligence (AI).^26,27 AI algorithms, particularly those utilizing deep learning, can analyze vast amounts of imaging data, identifying subtle patterns and anomalies that may be indicative of early-stage pancreatic cancer, often before they are detectable by human radiologists. AI applications can also help standardize the interpretation of imaging results, minimizing variability and ensuring that more patients receive timely and accurate diagnoses. Other early detection strategies, such as blood-based biomarkers are being investigated for their potential to identify pancreatic cancer at an earlier stage. Despite these advances, these methods are not yet sufficiently sensitive or specific to serve as widespread screening tools for the general population, and their use is limited to clinical trials for high-risk individuals.²⁶

Risk assessment is an important component of early detection. Current risk assessment guidelines for pancreatic cancer primarily focus on identifying individuals with a family history of the disease, genetic predispositions such as BRCA mutations, and certain lifestyle factors like smoking and obesity. However, these guidelines are not fully aligned with best practices for early detection and intervention. Many high-risk individuals remain undetected because the guidelines often fail to incorporate emerging biomarkers, advanced imaging techniques, or the use of genetic testing in risk assessment, all of which could facilitate earlier diagnosis. This is particularly problematic for historically excluded populations, who frequently experience delayed diagnoses and consequently face poorer health outcomes. For these communities and other high-risk populations, the lack of standardized screening protocols leads to inconsistencies in how patients are monitored and treated.²⁸

Improving awareness among the medical community about the risk factors and symptoms of pancreatic cancer is crucial for enhancing early detection and patient outcomes. Educating health care providers about common risk factors, such as family history, genetic mutations, chronic pancreatitis, diabetes, and lifestyle factors, can prompt more vigilant monitoring of at-risk individuals. In addition, increasing awareness about the subtle and often vague early symptoms, like jaundice, unexplained weight loss, abdominal and back pain, new-onset diabetes, loss of appetite, indigestion, unintended weight loss, changes in stool, fatigue, and mood fluctuations can lead to more timely investigations. It is equally important to inform medical professionals about the latest screening options and early detection methods, fostering a proactive approach to pancreatic cancer risk assessment and screening. The medical community can play a pivotal role in diagnosing the disease at earlier, more treatable stages, ultimately improving survival rates.

Moving forward, focused efforts on early diagnosis and intervention, driven by robust epidemiology data, are essential to address the current limitations and improve prognosis for a larger cohort of pancreatic cancer patients. Enhancing early detection methods, expanding their accessibility, and raising awareness are critical steps in shifting the landscape of pancreatic cancer diagnosis and treatment. Only through these concerted efforts can we hope to make significant strides in reducing the burden of pancreatic cancer and improving outcomes for those affected by this challenging disease.