Abstract
This population-based study investigates the survival and causes of death in patients with pancreatic adenocarcinoma to 21 years after diagnosis.
Most patients who undergo resection of pancreatic ductal adenocarcinoma (PDAC) die within 5 years.1,2,3,4,5 Five-year survivors remain at risk for PDAC-related death,1,2,4,5 and recurrences more than 15 years after diagnosis are reported.4 A study reporting conditional survival (CS) extended only 6.5 years after diagnosis.3 Reported 10-year survival rates range from 3.9% in a National Cancer Database (NCDB) study to 12% to 13% in single-institution studies.1,5,6 In the NCDB study, the risk of death decreased for 6 years and then remained fairly constant at 10% per year in years 7 to 15, which was double that of an age- and sex-matched general population.6
Limited prognostic information exists for 5- to 15-year survivors. We used the Surveillance, Epidemiology, and End Results (SEER) database (http://www.seer.cancer.gov) to quantify yearly risks of death due to PDAC vs other causes and calculate CS estimates extending 21 years after diagnosis.
Methods
The 1988-2009 SEER database was queried for patients aged 18 to 89 years with nonmetastatic PDAC with no previous cancers, at least 1 month of follow-up, and known stage. Data analysis was conducted from January 10 to February 15, 2018. Tumors were classified as T1-T2N0M0, T3-T4N0M0, and T1-T4N1M0 owing to inconsistent available stage codes. Deaths were classified as due to PDAC or other causes. The University of Utah institutional review board considers SEER studies exempt from review.
Overall survival (OS) and cancer-specific survival (CSS) were evaluated using the Kaplan-Meier method. Survival was right-censored at 21 years because the last PDAC-related death occurred at 20.5 years. Smoothed yearly hazards of death due to PDAC, other causes, and any cause were evaluated graphically. Conditional survival was calculated as the probability of survival for x additional years (CSx) given y years of accumulated survival (CSx = S(x + y)/S(y)). Analyses were conducted with Stata, version 14.0 (StataCorp).
Results
Of 10 988 patients, 5437 (49.5%) were women, the mean (SD) age was 64.8 (10.8) years, and 8206 (74.7%) were of non-Hispanic white race/ethnicity. Tumors were in the pancreatic head in 8591 patients (78.2%). Stage was T1-T2N0M0 in 1588 patients (14.5%), T3-T4N0M0 in 2951 patients (26.9%), and T1-T4N1M0 in 6449 patients (58.7%)
Observed CSS and OS are shown in the Figure, A. Hazards of overall and PDAC-related death peaked during year 3 (Figure, B). Hazards of death due to other causes exceeded that of PDAC-related death by approximately 8.75 years after diagnosis. Hazards of PDAC-related mortality decreased until approximately year 13 and then remained less than 3% per year. Five PDAC-related deaths occurred after 15 years, and 2 occurred after 20 years. The CSS- and OS-based estimates of future survival increased with longer accrued survival (Table).
Figure. Survival and Hazards of Death in Patients With Pancreatic Ductal Adenocarcinoma (PDAC).
Cancer-specific and overall survival (A) and yearly hazards of death due to PDAC, other causes, and any cause (B). The display of the yearly hazard of death due to PDAC stops at 20.5 years because that is when the last death due to PDAC occurred.
Table. Cancer-Specific and Overall CS Probabilitiesa.
| Total Survival Time, y | Already Accrued Survival Time, y | |||||
|---|---|---|---|---|---|---|
| 3 | 6 | 9 | 12 | 15 | 18 | |
| Cancer-Specific CS Probabilities, All Patients (T1-T2N0MO, T3-T4N0M0, T1-T4N1M0), % | ||||||
| 6 | 61.7 (77.3, 63.4, 51.2) | |||||
| 9 | 51.4 (67.7, 53.9, 40.0) | 83.4 (87.6, 85.1, 78.2) | ||||
| 12 | 46.4 (63.5, 47.6, 35.3) | 75.2 (82.2, 75.1, 69.0) | 90.2 (93.8, 88.3, 88.2) | |||
| 15 | 43.1 (58.8, 47.0, 30.9) | 69.9 (76.1, 74.1, 60.5) | 83.9 (86.9, 87.1, 77.3) | 93.0 (92.6, 98.7, 87.6) | ||
| 18 | 42.2 (57.3, 45.5, 30.9) | 68.3 (74.1, 71.8, 60.5) | 82.0 (84.6, 84.3, 77.3) | 90.9 (90.1, 95.5, 87.6) | 97.8 (97.3, 96.8, 100) | |
| 21 | 38.1 (57.3, 42.6, 22.1) | 61.8 (74.1, 67.3, 43.2) | 74.1 (84.6, 79.1, 55.2) | 82.2 (90.1, 89.6, 62.6) | 88.4 (97.3, 90.7, 71.5) | 90.4 (100, 93.7, 71.5) |
| Overall CS Probabilities, All Patients (T1-T2N0MO, T3-T4N0M0, T1-T4N1M0), % | ||||||
| 6 | 56.1 (70.5, 58.2, 46.1) | |||||
| 9 | 40.5 (53.9, 42.1, 31.6) | 72.2 (76.5, 72.3, 68.5) | ||||
| 12 | 31.4 (44.1, 31.5, 23.9) | 56.0 (62.6, 54.1, 51.9) | 77.5 (81.8, 74.9, 75.7) | |||
| 15 | 23.2 (32.3, 24.3, 17.3) | 41.3 (45.9, 41.7, 37.4) | 57.1 (60.0, 57.7, 54.6) | 73.7 (73.3, 77.0, 72.1) | ||
| 18 | 18.3 (23.4, 19.7, 14.9) | 32.6 (33.2, 33.8, 32.4) | 45.2 (43.4, 46.8, 47.3) | 58.3 (53.1, 62.5, 62.4) | 79.1 (72.4, 81.1, 86.6) | |
| 21 | 13.6 (18.8, 15.3, 8.9) | 24.2 (26.6, 26.3, 19.2) | 33.4 (34.8, 36.4, 28.1) | 43.1 (42.6, 48.6, 37.1) | 58.5 (58.1, 63.1, 51.4) | 74.0 (80.2, 77.8, 59.4) |
Abbreviation: CS, conditional survival.
For example, for a patient with a T1-T4N1M0 tumor who had survived for 6 years after diagnosis, the probability of surviving until 15 years after diagnosis without dying of pancreatic adenocarcinoma was 60.5% and the probability of surviving until 15 years without dying of any cause was 37.4%.
Discussion
In this study, patients with resected PDAC were more likely to die of causes other than PDAC starting 8.75 years after diagnosis, but PDAC-related deaths occurred as late as 20.5 years after diagnosis. In a 2008 study, PDAC-related death occurred in 16% of 5-year survivors, but not after 7.8 years.1 A more recent study reported recurrences in 15-year survivors.4 The approximately 10% per year risk of death between years 9 and 15 reported here is similar to that in an NCDB study.6 PDAC-related deaths are unlikely among 13- to 15-year survivors. Our CS estimates expand on previous studies by demonstrating that prognosis continues to improve with accrued survival beyond 5 years.3
Limitations of the study include the possibility of misclassification of cause of death, although very late PDAC-related deaths were previously reported,4 and the fact that the survival probabilities herein are means during a 22-year period of improving outcomes. These findings may be useful to physicians responsible for advising patients with resected PDAC about their prognosis.
References
- 1.Schnelldorfer T, Ware AL, Sarr MG, et al. Long-term survival after pancreatoduodenectomy for pancreatic adenocarcinoma: is cure possible? Ann Surg. 2008;247(3):456-462. doi: 10.1097/SLA.0b013e3181613142 [DOI] [PubMed] [Google Scholar]
- 2.Ferrone CR, Pieretti-Vanmarcke R, Bloom JP, et al. Pancreatic ductal adenocarcinoma: long-term survival does not equal cure. Surgery. 2012;152(3)(suppl 1):S43-S49. doi: 10.1016/j.surg.2012.05.020 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Katz MH, Hu CY, Fleming JB, Pisters PW, Lee JE, Chang GJ. Clinical calculator of conditional survival estimates for resected and unresected survivors of pancreatic cancer. Arch Surg. 2012;147(6):513-519. doi: 10.1001/archsurg.2011.2281 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Dusch N, Weiss C, Ströbel P, Kienle P, Post S, Niedergethmann M. Factors predicting long-term survival following pancreatic resection for ductal adenocarcinoma of the pancreas: 40 years of experience. J Gastrointest Surg. 2014;18(4):674-681. doi: 10.1007/s11605-013-2408-x [DOI] [PubMed] [Google Scholar]
- 5.Stark AP, Sacks GD, Rochefort MM, et al. Long-term survival in patients with pancreatic ductal adenocarcinoma. Surgery. 2016;159(6):1520-1527. doi: 10.1016/j.surg.2015.12.024 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Paniccia A, Hosokawa P, Henderson W, et al. Characteristics of 10-year survivors of pancreatic ductal adenocarcinoma. JAMA Surg. 2015;150(8):701-710. doi: 10.1001/jamasurg.2015.0668 [DOI] [PubMed] [Google Scholar]
