Demographics
A total of 4,079 patients were included in this study. The median age of the patients was 66 years (IQR 59–72), with 49.7% male patients. Most patients received NAC (n = 2,909, 71.3%), while 28.7% received NAC + RT (n = 1,170). The median total dosage of radiation was 50 Gy (IQR 3-50.4) The median duration of radiation was 25 days (IQR 10–39), and 18.1% of patients received radiation to draining lymph nodes in addition to the pancreas.
Patient characteristics were compared between the NAC and NAC + RT groups. The NAC group had a higher median age (66 years vs. 65, P < 0.001), and a higher proportion of patients with government-issued insurance (63% vs. 57.9% for NAC and NAC + RT, respectively, P = 0.001) (Table 1). As expected, the NAC + RT group had a higher clinical T stage, with a higher percentage of patients having T3-T4 tumors compared to the NAC group (26.9% vs. 42.4% for NAC + RT and NAC, respectively, P < 0.001). The NAC + RT group also had a higher rate of R0 margin compared to the NAC group (82.5% vs. 87.2% for NAC and NAC + RT, respectively, P = 0.003).
Table 1
Patient demographic and clinicopathologic stratified by type of progression
| Total | NAC | NAC + RT | P |
n | 4,079 | 2,909 | 1,170 | |
Age (years), median (IQR) | 66 (59–72) | 66 (60–72) | 65 (59–71) | < 0.001 |
Male sex, N (%) | 2,028 (49.7) | 1,438 (49.4) | 590 (50.4) | 0.57 |
Race, N (%) (n = 4,029) | | | | |
White non-Hispanic | 3,294 (81.8) | 2,342 (81.4) | 952 (82.7) | 0.37 |
Black non-Hispanic | 364 (9.0) | 257 (8.9) | 107 (9.3) | |
Others non-Hispanic | 148 (3.7) | 109 (3.8) | 39 (3.4) | |
Hispanic | 223 (5.5) | 170 (5.9) | 53 (4.6) | |
Insurance (n = 4,053) | | | | |
Private | 1,495 (36.9) | 1,018 (35.2) | 477 (41.1) | 0.001 |
Medicare/Medicaid | 2,495 (61.6) | 1,822 (63.0) | 673 (57.9) | |
No insurance | 63 (1.5) | 51 (1.8) | 12 (1.0) | |
Facility Type (n = 4,045) | | | | |
Community cancer program | 83 (2.0) | 59 (2.0) | 24 (2.1) | < 0.001 |
Comprehensive cancer program | 904 (22.4) | 724 (25.0) | 180 (15.6) | |
Academic program | 2,183 (54.0) | 1,503 (52.0) | 680 (58.9) | |
Integrated network cancer program | 875 (21.6) | 605 (20.9) | 270 (23.4) | |
Median income (n = 3,429) | | | | |
< $40,227 | 498 (14.5) | 374 (15.6) | 124 (12.1) | 0.001 |
$40,227 - $50,353 | 683 (19.9) | 502 (20.9) | 181 (17.7) | |
$50,354 - $63,332 | 814 (23.7) | 566 (23.5) | 248 (24.2) | |
≥ $63,333 | 1,434 (41.8) | 962 (40.0) | 472 (46.0) | |
Charlson comorbidity index | | | | |
0 | 2,578 (63.2) | 1,804 (62.0) | 774 (66.2) | 0.005 |
1 | 1,045 (25.6) | 752 (25.9) | 293 (25.0) | |
≥ 2 | 456 (11.2) | 353 (12.1) | 103 (8.8) | |
Type of surgery | | | | |
Whipple | 2,838 (69.6) | 2,046 (70.3) | 792 (67.7) | 0.019 |
Partial pancreatectomy | 531 (13.0) | 387 (13.3) | 144 (12.3) | |
Total | 492 (12.1) | 338 (11.6) | 154 (13.2) | |
Extended pancreatectomy | 218 (5.3) | 138 (4.7) | 80 (6.8) | |
Grade post treatment | | | | |
Well differentiated | 309 (7.6) | 229 (7.9) | 80 (6.8) | < 0.001 |
Moderately differentiated | 2,077 (50.9) | 1,490 (51.2) | 587 (50.2) | |
Poorly/undifferentiated | 935 (22.9) | 694 (23.9) | 241 (20.6) | |
Unknown | 758 (18.6) | 496 (17.0) | 262 (22.4) | |
Tumor size (cm), median (IQR) | 2.9 (2.2–3.7) | 2.8 (2.1–3.6) | 3.0 (2.3–3.9) | < 0.001 |
AJCC clinical T stage | | | | |
T1 | 524 (12.9) | 427 (14.7) | 97 (8.3) | < 0.001 |
T2 | 2,262 (55.5) | 1,697 (58.3) | 565 (48.3) | |
T3 | 738 (18.1) | 522 (17.9) | 216 (18.4) | |
T4 | 555 (13.6) | 263 (9.0) | 292 (24.0) | |
AJCC clinical N stage | | | | |
N0 | 3,114 (76.3) | 2,206 (75.8) | 908 (77.6) | 0.243 |
N1 | 898 (22.0) | 650 (22.3) | 248 (21.2) | |
N2 | 67 (1.7) | 53 (1.8) | 14 (1.2) | |
AJCC pathologic T stage post therapy | | | | |
0 | 179 (4.4) | 105 (3.6) | 74 (6.3) | < 0.001 |
T1 | 1,253 (30.7) | 846 (29.1) | 407 (34.8) | |
T2 | 1,986 (48.7) | 1,486 (51.1) | 500 (42.7) | |
T3 | 578 (14.2) | 423 (14.5) | 155 (13.3) | |
T4 | 83 (2.0) | 49 (1.7) | 34 (2.9) | |
AJCC pathologic N stage post therapy | | | | |
0 | 1,999 (49.0) | 1,217 (41.8) | 782 (66.8) | < 0.001 |
1 | 1,431 (35.1) | 1,123 (38.6) | 308 (26.3) | |
2 | 649 (15.9) | 569 (19.6) | 80 (6.8) | |
AJCC 8th edition post therapy | | | | |
Complete response | 165 (4.1) | 94 (3.2) | 71 (6.1)) | < 0.001 |
Stage 1A | 846 (20.7) | 525 (18.1) | 321 (27.4) | |
Stage 1B | 770 (18.9) | 473 (16.3) | 297 (25.4) | |
Stage IIA | 199 (4.9) | 123 (4.2) | 76 (6.5) | |
Stage IIB | 1,383 (33.9) | 1,091 (37.5) | 292 (24.9) | |
Stage III | 716 (17.6) | 603 (20.7) | 113 (9.7) | |
Lymph vascular invasion (n = 2,939) | 1,175 (40.0) | 955 (44.1) | 220 (28.4) | < 0.001 |
Margin statues (n = 4,026) | | | | |
R0 | 3,376 (83.8) | 2,369 (82.5) | 1,007 (87.2) | 0.003 |
R1 | 403 (10.0) | 313 (10.9) | 90 (7.8) | |
R2 | 18 (0.5) | 15 (0.5) | 3 (0.3) | |
Any positive | 229 (5.7) | 174 (6.0) | 55 (4.8) | |
Adjuvant chemotherapy | 1,606 (39.4) | 1,306 (44.9) | 300 (25.6) | < 0.001 |
Adjuvant radiotherapy | 433 (10.6) | 425 (14.6) | 8 (0.7) | < 0.001 |
NAC neoadjuvant chemotherapy; NAC + RT neoadjuvant chemotherapy with radiation; IQR Interquartile range; AJCC American Joint Committee on Cancer
Among the 1,170 patients who underwent NAC + RT, 98.0% completed their intended radiation course. The reason for early termination of radiation in 2% of patients was listed as follows; 0.8% (n = 9) stopped due to toxicity, 0.1% (n = 1) due to contraindications related to other patients' risk factors, 0.3% (n = 4) due to patient’s decision, and 0.9% (n = 10) for unknown reasons.
A total of 39.4% (n = 1,606) of patients received adjuvant chemotherapy. A lower percentage of patients who achieved pCR received adjuvant chemotherapy compared to the group that did not achieve pCR (16.6% vs 40.3%, P < 0.01). The NAC + RT group received less adjuvant chemotherapy compared to those in the NAC group (25.6% vs. 44.9%, respectively; P < 0.001). Interestingly, in a subgroup analysis of patients who did not have pCR, the NAC + RT group still received less adjuvant chemotherapy comparedta to the NAC group alone. (26.1% vs. 45.9%, P < 0.001).
Complete pathologic response
Out of the total 4,079 patients, 165 (4.1%) achieved pCR, with a higher rate among patients who received NAC + RT (n = 71, 6.1%) compared to those who received only NAC (n = 94, 3.2%) (P < 0.001). Other factors associated with pCR were age (median age, pCR 63 years vs. No pCR 66 years, P = 0.001) and higher clinical T stage (T3-T4; pCR 38.8 vs. no pCR 31.4; P = 0.001). After adjusting for all competing risk factors, NAC + RT was found to be associated with a higher rate of pCR compared to NAC (OR 1.78, 95%CI 1.28–2.47; P = 0.001) (Table 3). Details of radiation administration are summarized in Table 2. Patients who experienced pCR had a higher total radiation dose, with a median of 50 Gy compared to 45 Gy in the no pCR group (P = 0.01). A total of 223 out of 1170 patients in the NAC + RT group received SBRT (19.1%). Patients who received SBRT had a higher pCR compared to patients who received NAC only (NAT + SBRT: 7.2% vs NAT: 3.2%, P = 0.002). No difference was observed between the standard radiation and SBRT groups in terms of pCR (SBRT, 7.2% vs standard radiation 5.8%, P = 0.44).
Table 2
Detail for radiation among patients underwent neoadjuvant chemoradiation
| Complete pathologic response | AJCC down stage |
| Yes | No | P | Yes | No | P |
Total Dose (GY) | 50 (36-50.4) | 46.8 (36-50.4) | 0.06 | 50(36-50.4) | 45 (35-50.4) | 0.057 |
Number of Phases of Radiation Treatment | | | | | | |
1 Phase (n, %) | 66 (93.0) | 963 (87.6) | 0.70 | 549 (87.6) | 414 (87.7) | 0.37 |
2 phases (n, %) | 5 (7.0) | 119 (10.8) | | 69 (11.0) | 50 (10.65 | |
3 phases (n, %) | - | 2 (0.2) | | 2 (0.3) | 0 | |
> 4 phases (n, %) | - | 2 (0.2) | | 0 | 2 (0.4) | |
Unknown (n, %) | - | 13 (1.2) | | 7 (1.1) | 6 (1.3) | |
Duration of radiation (days) | 21 (7–37) | 25 (10–39) | 0.40 | 29 (11–39) | 21 (10–38) | 0.24 |
Phase I Radiation to Draining Lymph Nodes (n, %) | 10 (14.5) | 191 (18.4) | 0.42 | 121 (20.4) | 70 (15.6) | 0.047 |
Phase I Radiation Treatment Modality | | | | | | |
External beam (n, %) | 71 (100) | 1091 (99.5) | 0.57 | 620 (99.4) | 471 (99.8) | 0.29 |
Brachytherapy (n, %) | - | 5 (0.5) | | 4 (0.6) | 1 (0.2) | |
Dose per Fraction (Gy, median, IQR) | 2 (1.8–6.6) | 2.1 (1.8–6.6) | 0.44 | 2.1 (1.8-6) | 2.14 (1.8–6.6) | 0.89 |
Number of Fractions (median, IQR) | 25 (5–28) | 24 (5–28) | 0.35 | 25 (5–28) | 18 (5–28) | 0.68 |
Total Dose (Gy, median, IQR) | 50 (38-50.4) | 45 (35-50.4) | 0.01 | 45 (36-50.4) | 45 (33.3–50.4) | 0.12 |
IQR Interquartile range |
Interestingly in the NAC + RT group, interval > 11 weeks between surgery and radiation was associated with higher pCR (≤ 11 weeks, 4.8% vs > 11 weeks 7.8%; P = 0.038). After adjusting for other factors associated with pCR such as age and clinical T stage, the time between radiation and surgery > 11 weeks was associated with a higher rate of pCR (OR 1.68, 95%CI 1.03–2.75; P = 0.04).
In a subgroup analysis, excluding patients who had AJCC downstaging, age, clinical T stage, clinical N stage, and receipt of NAC + RT compared to NAC (OR 1.89, 95%CI 1.32–2.71; P = 0.001) were associated with an increased rate of pCR (Supplemental Table 2).
Table 3
Factors associated with complete pathologic response
| Univariable | Multivariable |
| OR (95% CI) | P | OR (95% CI) | P |
Age | 0.97 (0.96–0.98) | 0.001 | 0.97 (0.96–0.99) | 0.001 |
Female sex | 0.80 (0.58–1.09) | 0.16 | - | |
Race | | | | |
White non-Hispanic | Ref | | | |
Black non-Hispanic | 1.06 (0.62–1.83) | 0.16 | - | |
Others non-Hispanic | 1.60 (0.80–3.22) | | - | |
Hispanic | 1.53 (0.85–2.75) | | - | |
Insurance | | | | |
Private | Ref | | | |
Medicare/Medicaid | 0.75 (0.54–1.03) | 0.08 | - | |
No insurance | 1.36 (0.48–3.85) | 0.56 | - | |
Facility Type | | | | |
Community cancer program | Ref | | | |
Comprehensive cancer program | 3.30 (0.45–24.42) | 0.24 | - | |
Academic program | 3.61 (0.50-26.21) | 0.21 | - | |
Integrated network cancer program | 3.32 (0.45–24.53) | 0.24 | - | |
Median income | | | | |
< $40,227 | Ref | | | |
$40,227 - $50,353 | 1.37 (0.78–2.39) | 0.27 | - | |
$50,354 - $63,332 | 0.79 (0.44–1.43) | 0.44 | - | |
≥ $63,333 | 1.06 (0.63–1.78) | 0.82 | - | |
Charlson comorbidity index | | | | |
0 | Ref | | | |
1 | 0.73 (0.50–1.07) | 0.11 | - | |
≥ 2 | 0.62 (0.34–1.10) | 0.10 | - | |
Type of surgery | | | | |
Whipple | Ref | | | |
Partial pancreatectomy | 1.9 (0.75–1.86) | 0.46 | - | |
Total | 1.45 (0.94–2.24) | 0.09 | - | |
Extended pancreatectomy | 0.59 (0.24–1.46) | 0.25 | - | |
AJCC clinical T stage | | | | |
T1 | Ref | | | |
T2 | 0.59 (0.38–0.92) | 0.02 | 0.56 (0.36–0.89) | 0.01 |
T3 | 0.67 (0.39–1.16) | 0.15 | 0.61 (0.35–1.05) | 0.08 |
T4 | 1.27 (0.76–2.10) | 0.36 | 1.00 (0.59–1.70) | 0.99 |
AJCC clinical N stage | | | | |
N0 | Ref | | | |
N1 | 1.02 (0.70–1.48) | 0.92 | - | |
N2 | 0.73 (0.18–3.01) | 0.66 | - | |
Type of Neoadjuvant treatment | | | | |
NAC | Ref | | | |
NAC + RT | 1.93 (1.41–2.65) | < 0.001 | 1.78 (1.28–2.47) | 0.001 |
NAC neoadjuvant chemotherapy; NAC + RT neoadjuvant chemotherapy with radiation; AJCC American Joint Committee on Cancer |
AJCC Downstaging
A total of 1,774 (43.5%) of patients had downstaging. Among these patients 1,423 (80.2%) had primary tumor downstaging, 136 (7.7%) had nodal downstaging, and 215 (12.1%) had both primary tumor and nodal downstaging. Patients who received NAC + RT had a higher AJCC downstaging rate (n = 627, 57.1%) compared to those who received NAC (n = 1,147, 40.8%) (P < 0.001). Other factors associated with downstaging in univariable analysis were age, insurance type, histologic grade, clinical T stage, and clinical N stage (Table 4). In multivariable analysis, receipt of NAC + RT was associated with downstaging (OR 1.52, 95%CI 1.28–1.81; P < 0.001). Of note, SBRT was associated with an increased rate of AJCC downstaging compared to NAC (SBRT 54.1% vs NAC 40.8%, P < 0.001). No difference was observed between the standard radiation and SBRT groups in terms of AJCC downstaging (SBRT, 54.1% vs standard radiation 57.7%, P = 0.34).
Table 4
Factors associated with AJCC downstage
| Univariable | Multivariable |
| OR (95% CI) | P | OR (95% CI) | P |
Age | 0.99 (0.98–0.99) | 0.03 | 0.99 (0.98–1.01) | 0.34 |
Female sex | 1.06 (0.94-.1.21) | 0.38 | - | |
Race | | | | |
White non-Hispanic | Ref | | | |
Black non-Hispanic | 1.05 (0.84–1.31) | 0.68 | - | |
Others non-Hispanic | 0.92 (0.66–1.31) | 0.62 | - | |
Hispanic | 0.96 (0.72–1.26) | 0.78 | - | |
Insurance | | | | |
Private | Ref | | Ref | |
Medicare/Medicaid | 0.86 (0.75–0.98) | 0.03 | 0.97 (0.80–1.18) | 0.79 |
No insurance | 0.64 (0.38–1.12) | 0.11 | 0.73 (0.38–1.40) | 0.35 |
Facility Type | | | | |
Community cancer program | Ref | | - | |
Comprehensive cancer program | 0.85 (0.54–1.35) | 0.51 | - | |
Academic program | 0.87 (0.56–1.36) | 0.55 | - | |
Integrated network cancer program | 0.85 (0.54–1.34) | 0.49 | - | |
Median income | | | | |
< $40,227 | Ref | | - | |
$40,227 - $50,353 | 1.12 (0.87–1.40) | 0.34 | - | |
$50,354 - $63,332 | 1.06 (0.84–1.32) | 0.65 | - | |
≥ $63,333 | 1.12 (0.91–1.38) | 0.28 | - | |
Charlson comorbidity index | | | | |
0 | Ref | | - | |
1 | 0.89 (0.77–1.04) | 0.11 | - | |
≥ 2 | 0.92 (0.75–1.13) | 0.42 | - | |
Type of surgery | | | | |
Whipple | Ref | | - | |
Partial pancreatectomy | 1.02 (0.84–1.23) | 0.84 | - | |
Total | 1.02 (0.84–1.25) | 0.84 | - | |
Extended pancreatectomy | 1.10 (0.84–1.45) | 0.51 | - | |
Grade post treatment | | | | |
Well differentiated | Ref | | Ref | |
Moderately differentiated | 0.85 (0.67–1.08) | 0.19 | 0.79 (0.58–1.06) | 0.11 |
Poorly/undifferentiated | 0.67 (0.52–0.87) | 0.002 | 0.54 (0.40–0.75) | < 0.001 |
AJCC clinical T stage | | | | |
T1 | Ref | | Ref | |
T2 | 9.05 (6.08–13.47) | < 0.001 | 11.65 (7.73–17.53) | < 0.001 |
T3-T4 | 70.12 (46.42-105.92) | < 0.001 | 148.56 (94.40-233.81) | < 0.001 |
AJCC clinical N stage | | | | |
N0 | Ref | | Ref | |
N1 | 2.19 (1.87–2.55) | < 0.001 | 2.25 (1.88–2.72) | < 0.001 |
N2 | 3.29(1.93–5.60) | < 0.001 | 5.40 (2.83–10.52) | < 0.001 |
Type of Neoadjuvant treatment | | | | |
NAC | Ref | | Ref | |
NAC + RT | 1.93 (1.68–2.22) | < 0.001 | 1.52 (1.28–1.81) | < 0.001 |
NAC neoadjuvant chemotherapy; NAC + RT Neoadjuvant chemotherapy with radiation; AJCC American Joint Committee on Cancer
We also evaluated the association between downstaging and chemotherapy on margin status. The NAC + RT group also had a higher rate of R0 margin compared to the NAC group (82.5% vs. 87.2% for NAC and NAC + RT, respectively, P = 0.003). Patients who received SBRT had a higher rate of R0 resection compared to NAC (88.6% vs 82.5%, P = 0.02) but no differences were seen between SBRT and standard radiation group (SBRT 88.6% vs 86.8%, P = 0.47). After adjusting for all competing risk factors, including sex, facility type, comorbidities, T stage, node positivity, and pathologic grade, AJCC downstaging (OR 0.55, 95%CI 0.43–0.71, P < 0.001) and NAC + RT (OR 0.65, 95%CI 0.47–0.90, P = 0.01) were associated with a lower risk of a positive margin. The model was also adjusted for the interaction between NAC + RT and AJCC downstaging (Supplementary Table 1).
Subgroup analysis among patients with available data on lymphovascular invasion
A total of 2,939 patients had available data on LVI, which accounted for 72% of the total population. Patients with LVI had a lower AJCC downstaging rate (LVI 33.2% vs. no LVI 50.7%, P < 0.001). Receipt of NAC + RT improved AJCC downstaging (NAC + RT 44.1% vs. NAC 30.6%; P < 0.001). In a multivariable analysis of factors associated with downstaging, receipt of NAC + RT was associated with a higher rate of downstaging (OR = 1.28, 95% CI 1.03–1.59; P = 0.028), and the presence of LVI was associated with a lower rate of downstaging (OR 0.41, 95%CI 0.33–0.50; P < 0.001).
Overall survival
The median OS of the entire cohort was 32 months, which was longer for patients who received NAC + RT (36.5 months) compared to those who received NAC (30.9 months) (P = 0.001) (Fig. 2). Furthermore, patients who had AJCC downstaging had a longer median OS than those who did not (33.5 vs. 28.8 months, P < 0.001). Subgroup analysis revealed that downstaging with NAC had an association with OS (33.4 vs. 27.9 months, P < 0.001), but not with NAC + RT (36.5 vs. 32.5 months, P = 0.53) (Fig. 3a). On the other hand, patients with pCR demonstrated longer median OS regardless of the type of neoadjuvant therapy (with pCR median not reached, without pCR 31.5 months, P < 0.001) (Fig. 3b, Supplementary Table 4).
In multivariable analysis after adjusting for variables such as age, insurance type, comorbidities, type of surgery, histologic grade, post-treatment pathologic staging, margin, and adjuvant chemotherapy, patients with pCR had a 50% reduction in the risk of death (HR 0.50, 95% CI 0.27–0.93; P = 0.03). However, AJCC downstaging and type of neoadjuvant therapy did not remain significant after adjusting for other factors (Supplementary Table 3).
In terms of post-operative treatment, patients with pCR who received adjuvant chemotherapy (n = 27) did not demonstrate a significant difference in OS compared to the group who did not receive adjuvant chemotherapy (P = 0.41). However, among patients without pCR, adjuvant chemotherapy was associated with longer OS regardless of AJCC downstaging (P = 0.03).