Patient and tumor characteristics
Between 2000 and 2017, 914 patients were diagnosed with PASC. Of these, 815 patients were included in our OS and descriptive analyses, whereas 773 patients were included in our DSS analysis (Figure.1). We divided the study population into four subgroups based on tumor site (head of pancreas, body of pancreas, tail of pancreas, and others) (Table 1). In the study population, the median age at diagnosis was 69 years old (interquartile range, [IQR]: 62–76 years old). There was a similar proportion of males (n = 417, 51.2%) and females (n = 398, 48.8%) in the overall cohort. Caucasians comprised the vast majority of the patients (n = 666, 81.7%). Based on SEER stage, most patients had regional (n = 355, 43.6%) and distant PASC (n = 375, 46.0%). Localized tumors occurred in a relatively small number of patients (n = 68, 8.3%). Based on the degree of differentiation, most patients had poorly differentiated PASC (n = 311, 38.2%), followed by moderately differentiated disease (n = 115, 14.1%).
Table 1
Trends In Baseline Demographic And Pathological Characteristics Of The Study Population(2004–2017)
Variable | Total | Head of pancreas | Body of pancreas | Tail of pancreas | Other* |
No. of patients (n) | 815 | 368 | 119 | 175 | 153 |
Median age (years) | 69(62,76) | 70(61.5,76) | 69(60,74) | 69(62,77) | 69(62,75) |
Gender, n(%) | | | | | |
Female | 398(51.2) | 175(47.6) | 65(54.6) | 86(49.1) | 72(47.1) |
Male | 417(48.8) | 193(52.4) | 54(45.4) | 89(50.9) | 81(52.9) |
Race, n (%) | | | | | |
White | 666(81.7) | 309(84.0) | 97(81.5) | 135(77.1) | 125(81.7) |
Black | 87(10.7) | 35(9.5) | 14(11.8) | 26(14.9) | 12(7.8) |
Other | 62(7.6) | 24(6.5) | 8(6.7) | 14 (8) | 16(10.5) |
SEER historic stage, n(%) | | | | | |
Localized | 68(8.3) | 34(9.2) | 12(10.1) | 13(7.4) | 9(5.9) |
Regional | 355(43.6) | 196(53.3) | 47(39.5) | 64(36.6) | 48(31.4) |
Distant | 375(46.0) | 130(35.3) | 59(49.6) | 97(55.4) | 89(58.2) |
Unstaged | 17(2.1) | 8(2.2) | 1(0.8) | 1(0.6) | 7(4.6) |
Grade, n (%) | | | | | |
Well differentiated | 4(0.5) | 1(0.3) | 0(0) | 1(0.6) | 2(1.3) |
Moderately differentiated | 115(14.1) | 40(10.9) | 21(17.6) | 35(20.0) | 19(12.4) |
Poorly differentiated | 311(38.2) | 152(41.3) | 41(34.5) | 72(41.1) | 46(30.1) |
Undifferentiated | 17(2.1) | 11(3.0) | 3(2.5) | 3(1.7) | 0(0) |
Unknown | 368(45.1) | 164(44.5) | 54(45.4) | 64(36.6) | 86(56.2) |
* other group include:Pancreatic duct: 4 individuals, Overlapping lesion of pancreas:78 individuals, Other specified parts of pancreas:8 individual, Pancreas:63 individuals. |
As shown in Table 1, the median age of diagnosis in patients with PASC of the head was 70 years old (IQR: 61.5–76 years old). The median age of patients with PASC in the body and tail of the pancreas was the same at the time of diagnosis, both 69 years old (IQR, body: 60–74 years old; tail: 62–77 years old). Similar to the general study population, the proportions of males and females with tumors at different sites were approximately equal. Interestingly, the rate of regional disease was significantly higher in patients with PASC of the head than in those with PASC of the body (p = 0.009) and tail (p < 0.001) of the pancreas. The rate of distant disease was significantly lower in patients with PASC of the head than in those with PASC of the body (p = 0.006) and tail (p < 0.001) of the pancreas. However, the proportion of moderately differentiated disease was significantly higher in patients with PASC of the head than in patients with PASC of the tail (p = 0.004).
Patients who underwent surgery had a higher proportion of regional PASC than those who did not (77.8% vs. 22.5%, P < 0.001) (Figure. S1a). In contrast, the proportion of distant PASC in patients who received surgical treatment was lower than that in patients without surgical treatment (17.5% vs. 70.0%, P < 0.001). In addition, we studied the relationship between tumor size and SEER stage (Figure. S1b). Regional PASC was more common in patients with tumors ≤ 3.5 cm (58.9% vs. 43.9%, P = 0.003), and distant PASC was more common in patients with tumors > 3.5 cm (49.4% vs. 27.0%, P < 0.001).
Overall Incidence and IB Mortality Trends
The incidence of PASC steadily increased during the study period (Figure.2a). The incidence of PASC increased from 0.33 cases per 1,000,000 individuals in 2000 to 0.79 per 1,000,000 individuals in 2017, with an APC of 3.9% (95% CI: 2.2–5.7%, P < 0.05) (i.e., the slope or extent of the increase in incidence). We also analyzed the incidence of PDAC and PSCC for comparison. The incidence of PSCC also showed a consistent upward trend during the study period (Figure.2b). The incidence of PSCC was 0.11 cases per 1,000,000 individuals in 2000, and 0.33 cases per 1,000,000 individuals in 2017. The APC was 5.6% (95% CI: 2.9–8.4%, P < 0.05). Furthermore, the incidence of PDAC increased faster earlier in the study, from 62.14 cases per 1,000,000 individuals in 2000 to 72.00 cases per 1,000,000 individuals in 2002, with an APC of 2.9% (95% CI: 1.6–4.2%, P < 0.05) (Figure.2c). Since 2008, however, the increase in PDAC rates has been moderate, with an APC of 0.7% (95% CI: 0.3–1.0%, P < 0.05).
During the research period, IB mortality of PASC increased continuously (Figure.2d). The IB mortality rate of PASC increased from 0.17 cases per 1,000,000 individuals in 2000 to 0.74 per 1,000,000 individuals in 2017, with an APC of 5.0% (95% CI: 2.5–7.6%, P < 0.05). Similarly, the IB mortality of PSCC also showed a continuous upward trend, with an APC of 6.3% (95% CI: 3.7–9.0%, P < 0.05) (Figure.2e). The IB mortality of PDAC rapidly increased early in the study period, but the upward trend gradually slowed (Figure.2f). The APC declined from 2.1% (95% CI: 1.2–2.9%, P < 0.05) between 2002 and 2009 to 0.9% (95% CI: 0.3–1.4%, P < 0.05) between 2009 and 2017.
Incidence and IB Mortality Trends by Sex
We split the study population by sex for additional research and found that the incidence of PASC increased steadily in both males and females, with APCs of 3.4% (95% CI: 0.8–6.1%, P < 0.05) and 4.8% (95% CI: 2.6–7.1%, P < 0.05), respectively (Figure.S2a). In addition, the incidence was higher in males than in females. The IB mortality of PASC in both males and females also displayed a consistent rising trend over the study period, with APCs of 5.4% (95% CI: 1.3–9.6%, P < 0.05) in males and 5.5% (95% CI: 2.6–8.5%, P < 0.05) in females (Figure.S2b).
Incidence and IB Mortality Trends by Stage
We then stratified the study cohort according to the SEER stage. Overall, the incidence of distant and regional PASC was considerably higher than that of localized PASC (Figure.S3a). Moreover, the increasing trends in the incidence of PASC were similar at both distant and regional stages, with APCs of 4.1% (95% CI: 1.0–7.4%, P < 0.05) in distant PASC and 4.3% (95% CI: 1.4–7.3%, P < 0.05) in regional PASC. As with incidence, the IB mortality of localized PASC remained the lowest of the three subgroups (Figure.S3b). Furthermore, the increasing trends in IB mortality of PASC were also similar at both distant and regional stages, with APCs of 5.0% (95% CI: 1.4–8.7%, P < 0.05) and 5.8% (95% CI: 3.1–8.7%, P < 0.05) in distant and regional PASC, respectively.
Incidence and IB Mortality Trends by Tumor Site
Next, we grouped the study cohort according to the tumor site. Overall, the incidence of PASC was highest in the head of the pancreas, followed by the tail of the pancreas, and lowest in the body of the pancreas; both showed a continuous increase over the study period, with APCs of 4.1% (95% CI: 1.9–6.3%, P < 0.05), 5.7% (95% CI: 0.4–11.3%, P < 0.05), and 6.0% (95% CI: 1.7–10.5%, P < 0.05) in the head, tail, and body of the pancreas, respectively (Figure.S4a). The IB mortality of PASC at the head of the pancreas was still the highest and showed continuously increasing trend, with an APC of 5.7% (95% CI: 2.4–9.1%, P < 0.05) (Figure.S4b). The increasing trend in IB mortality of the pancreatic tail and body was similar, with APCs of 6.1% (95% CI: 1.8–10.6%, P < 0.05) and 5.9% (95% CI:1.4–10.6%, P < 0.05), respectively.
Long-Term Survival Outcomes
The overall median survival of patients with PASC was 6 months (95% CI: 6–7 months), and the disease-specific median survival of patients with PASC was 6 months (95% CI: 5–7 months). The 1- and 2-year OS were 26.6% and 13.6%, respectively, and the corresponding DSS were 25.8% and 12.7%, respectively. Compared to patients with PASC, there was no statistical difference in the OS and DSS of patients with PDAC, but the OS (median survival 4 vs. 6 months, P < 0.001) and DSS (median survival 4 vs.6 months, P < 0.001) of patients with PSCC were significantly worse (Figure.3a, b). Patients older than 60 years had significantly worse OS (P < 0.001) and DSS (P < 0.001) [Figure.3c, d]. In contrast, patients with PASC who underwent surgery had significantly longer OS (median survival: 12 vs.4 months, P < 0.001) and DSS (median survival: 12 vs.4 months, P < 0.001) than those who did not (Figure.3e, f). Moreover, positive regional lymph nodes and tumors larger than 3.5 cm led to shorter OS (P < 0.001) and DSS (P < 0.001) (Figure.3g, h, I, j). According to SEER stage, the OS (median survival: 10 vs. 5 months, P < 0.001) and DSS (median survival 10 vs. 4 months, P < 0.001) of patients with localized PASC were better than those of patients with distant PASC (Figure.3k., l). However, there was no difference in OS and DSS between regional and localized PASC patients. There was also no significant difference in OS and DSS among patients with different sexes and different grades of PASC. (Figure.S5a, b, c, and d). Interestingly, OS (median survival 6 vs. 5 months, P = 0.017) and DSS (median survival 6 vs.5 months, P = 0.031) were longer in White patients than in Black patients (Figure.S5e, f). Furthermore, the location of the tumor along the pancreas did not affect OS and DSS (Figure.S5g, h).
We first used COX regression to conduct univariate analysis and found that age at diagnosis, SEER stage, treatment, regional lymph node involvement, and tumor size were significantly correlated with OS (P < 0. 1) and DSS (P < 0. 1) (TableS1). These factors were then incorporated into the multivariate COX regression analysis (Table 2). Here, age at diagnosis, treatment, regional lymph node involvement, and tumor size were identified as independent prognosticators of OS and DSS.
Table 2
Multivariate Cox’s Proportional Hazards Model Assessing Factors Associated With Mortality After Diagnosis Of Pancreatic Adenosquamous Carcinoma
| OS | 95%CI | | DSS | 95%CI | |
Risk Factor | HR* | Lower | Upper | P Value | HR* | Lower | Upper | P Value |
Age at diagnose(years) | | | | | | | | |
≦ 60 | Referent | | | | Referent | | | |
> 60 | 1.32 | 1.06 | 1.86 | 0.033 | 1.35 | 1.03 | 1.92 | 0.039 |
SEER stage | | | | | | | | |
Distant | Referent | | | | Referent | | | |
Localized | 0.66 | 0.35 | 1.20 | 0.201 | 0.68 | 0.35 | 1.35 | 0.25 |
Regional | 0.64 | 0.46 | 0.90 | 0.010 | 0.65 | 0.46 | 0.92 | 0.016 |
Treatment | | | | | | | | |
Surgery | Referent | | | | Referent | | | |
No surgery | 4.01 | 2.33 | 6.90 | < 0.001 | 3.88 | 6.77 | 3.69 | < 0.001 |
Regional lymph nodes | | | | | | | | |
Negative | Referent | | | | Referent | | | |
Positive | 1.51 | 1.13 | 2.00 | 0.006 | 1.53 | 1.13 | 2.08 | 0.020 |
Tumor size | | | | | | | | |
≦ 3.5cm | Referent | | | | Referent | | | |
༞3.5cm | 1.43 | 1.07 | 1.90 | 0.016 | 1.41 | 1.04 | 1.90 | 0.027 |
*HRs greater than 1.0 indicate a higher risk of death |