Abstract
Introduction:
Optimal cut-offs for CA19-9 response after neoadjuvant therapy (NT) for pancreatic adenocarcinoma (PDAC) are not well characterized. This study aims to analyze the relationship of serum CA19-9 with other markers of response and identify thresholds correlating to outcomes.
Methods:
Retrospective review of resected PDAC patients from 2010 to 2017 at an academic tertiary referral center.
Results:
Two hundred and fifty subjects entered the analysis. Normalization and multiple cut-off points for CA19-9 response were assessed. Normalization was not associated with improved survival (35.17 vs 29.43 months; p=0.173). While a response ≥45% was associated with longer survival (35 vs 20 months; p=0.018), a response ≥85% was optimal (55.7 vs 25.97 months; p<0.0001). Response ≥85% remained a strong independent predictor of survival (HR: 0.47; p=0.007). Subjects with ≥85% response had received more NT cycles (3 [2–6] vs 3 [2–4]; p=0.006) and less adjuvant cycles (4 [3–6] vs 5 [3–6]; p=0.027). Patients with an R1 resection and <85% CA19-9 response had a greater risk of recurrence (HR=2.64; p=0.0001) than patients with an R1 resection who had ≥85% CA19-9 response (HR=1.77; p=0.136). T-size reduction correlated with drop in Ca19-9 and size reduction ≥25% (56.97 vs 28.17 months, p=0.016) improved survival. Serum CA19-9 response ≥85% was a strong independent predictor of T-size reduction ≥25% (HR=2.40; p=0.007).
Conclusion:
CA19-9 response ≥85% is the optimal threshold for predicting survival. It is predictive of T-size reduction and may mitigate the impact of R1 resections on recurrence. Future NT trials should incorporate CA19-9 response as an endpoint.
Keywords: Serum CA19-9, Neoadjuvant Therapy, Survival, Pancreatic Cancer
INTRODUCTION
Pancreatic Adenocarcinoma (PDAC) has poor survival despite surgical resection1. A surgery first approach followed by adjuvant chemotherapy (AT) had been the optimal approach prior to more effective chemotherapy regimens being introduced in the metastatic setting 2–5. Receipt of AT is limited by high post-operative morbidity after surgery which impacts survival6–8. In addition, patients experience high rates of early loco-regional recurrence and distant metastasis, and experts frequently view PDAC as a systemic disease from diagnosis requiring neoadjuvant chemotherapy (NT)9,10.
NT has been shown to improve the rates of therapy completion, lymph node sterilization, negative margin resection, and perioperative morbidity7,11–14. The optimal length of NT and time to proceed to resection remains unclear. Multiple trials have shown that in patients undergoing AT post-resection, a six months span of treatment was best for survival15,16. Our group has shown in a cohort of patients who underwent resection with or without NT, a total of 6 months of treatment (NT + AT) was optimal for survival17. The number of cycles to give up-front remains unknown, but may be a function of institutional protocols or vascular involvement to achieve negative margin resection18.
Imaging has been shown to be an unreliable indicator of response, with no association with negative margin resection or survival19,20. Utilizing serum CA19-9 to maximize biologic response has shown promise19–24 and may serve as a potential guide for length of NT administration and time to resection. This study aims to assess the optimal cut-off for serum CA19-9 response to NT, its association with tumor size reduction and negative margin resection.
METHODS
Study Design and Patient Population
After institutional review board approval (PRO16040384), the records of patients diagnosed with PDAC receiving NT followed by resection between February 2010 and June 2017 at the University of Pittsburgh were retrospectively identified from a prospectively maintained surgical database. Serum CA19-9 levels were retrospectively collected for pre-NT and post-NT time-points. Patients with normal serum CA19-9 pre-NT (<33 IU/mL) were excluded. A normal total bilirubin (<2 mg/dL) was required to assess CA19-9 values. Pre-NT Tumor size (T-size) was measured by EUS when available and supplemented by CT when unavailable. Post-NT T-size was defined as the pathologic T-size. Percentage change was then calculated.
Chemotherapeutic regimens were classified into either gemcitabine-based regimens or 5-FU based regimens. Cycles were retrospectively counted from the number of treatments received. For patients receiving 5-FU based regimens, two doses were considered one cycle. For patients receiving gemcitabine-based regimens, two doses in the form of two weeks of treatment and two weeks with no treatment, or three doses in the form of three weeks of treatment and one week with no treatment were considered one cycle. Tumor anatomic type was decided upon by a multidisciplinary team in the course of the multidisciplinary clinic at the hospital and was determined according to SSO/AHPBA/SSAT/NCCN criteria. Pathologic staging was according to the American Joint Commission for Cancer version 7.
Statistical Analysis
Continuous data was summarized as mean±standard deviation (SD) or median with interquartile range (IQR). Categorical variables were reported as frequency and percentages. For continuous variables, comparisons were made using Student’s t-test when normally distributed and with Wilcoxon rank-sum test otherwise. For categorical variables, comparisons were made using Fisher’s exact test. CA19-9 response was stratified into normalization and 25–85% reduction, and survival was assessed. The associations with T-size reduction and positive margins were also assessed as secondary outcomes. Survival was calculated using the Kaplan-Meier method and statistical significance was assessed using the log-rank test. Hazard ratios (HRs) were determined using Cox proportional hazard regression modeling and reported with 95% confidence intervals (CI). Statistical analysis was performed using Stata/SE 13.0 (StataCorp LLC, College Station, Texas, USA).
RESULTS
Patient Demographics and Outcomes
Six hundred and seventy-five patients underwent resection: 369 patients received NT. Patients with normal pre-NT CA19-9, abnormal TB (>2mg/dL), or missing CA19-9 values were excluded: 250 patients entered the analysis. The average patient age was 65.70 ± 9.48 and 49.6% were females. Most tumors were borderline resectable 132 (53%), most patients received gemcitabine-based regimens 157 (73%), and only 46 (18%) included radiation following chemotherapy (Table 1).
TABLE 1.
Pre-operative patient characteristics and postoperative outcomes
| Variables | All Cohort N= 250 | CA19-9 Response <85% N=159 | CA19-9 Response ≥85% N=91 | P-Value |
|---|---|---|---|---|
| Preoperative Variables | ||||
| Age (Years) | 65.70 ± 9.48 | 65.94 ± 9.96 | 65.32 ± 10.36 | 0.621 |
| Sex (Females) | 124 (49.6) | 77 (48.4) | 47 (51.7) | 0.624 |
| CCI Age Unadjusted | 2.61 ± 0.88 | 2.68 ± 0.92 | 2.50 ± 0.79 | 0.111 |
| CCI Age Adjusted | 4.71 ±1.47 | 4.80 ± 1.47 | 4.55 ± 1.46 | 0.197 |
| BMI (Kg/m2) | 26.80 ± 5.24 | 26.78 ± 4.94 | 26.85 ± 5.77 | 0.912 |
| Pre-operative Albumin (g/dL) | 3.55 ± 0.49 | 3.51 ± 0.50 | 3.63 ± 0.45 | 0.045 |
| ASA Class | 0.993 | |||
| II | 30 (12) | 19 (12) | 11 (12) | |
| III | 202 (81) | 129(82) | 73 (81) | |
| IV | 16 (7) | 10 (6) | 6 (7) | |
| Prior Surgery | 132 (54.5) | 77 (48.4) | 47 (51.7) | 0.624 |
| CT Size (cm) | 3.04 ± 1.27 | 3.01 ± 1.31 | 3.03 ± 1.20 | 0.924 |
| EUS Size (cm) | 2.91 ± 0.96 | 2.83 ± 0.96 | 3.03 ± 0.96 | 0.150 |
| Anatomic Type | 0.435 | |||
| Resectable | 87 (35) | 60 (38) | 27(30) | |
| Borderline Resectable | 132 (53) | 80 (50) | 52(57) | |
| Locally Advanced | 31 (12) | 19 (12) | 12 (13) | |
| Type of NT | ||||
| Gem (vs. 5FU) | 157 (73) | 109 (77) | 48 (66) | 0.158 |
| Type of Radiation | 0.079 | |||
| None | 204 (82) | 128 (80) | 76 (84) | |
| SBRT | 26 (10) | 14 (9) | 12 (13) | |
| Other | 20 (8) | 17 (11) | 3 (3) | |
| NT Cycle Number | 3 (2 – 4) | 3 (2 – 4) | 3 (2 – 6) | 0.006 |
| Total Chemotherapy (AT+NT) | 8 (6–9) | 8(7–9) | 8 (6–10) | 0.639 |
| Pre-NT CA19-9 (IU/mL) | 317 (110, 1088) | 189 (84, 605) | 626 (314, 2271) | 0.0001 |
| Post-NT CA19-9 (IU/mL) | 55 (27, 167) | 71 (35, 279) | 37 (19, 81) | 0.0001 |
| Postoperative Outcomes | ||||
| Vein Resection | 80 (39) | 55 (41) | 25 (34) | 0.358 |
| T-Size (cm) | 2.78 ± 1.22 | 2.97 ± 1.12 | 2.44 ± 1.31 | 0.001 |
| Margin Status (R1) | 45 (18) | 30 (19) | 15 (17) | 0.637 |
| Number of Positive Nodes | 1 (0–4) | 2 (0–5) | 1 (0–3) | 0.003 |
| Positive Nodes (N1) | 163 (65) | 116 (73) | 47 (52) | 0.001 |
| LV | 171(73) | 120 (80) | 51 (61) | 0.001 |
| PNI | 196 (80) | 136 (87) | 60 (67) | 0.001 |
| Pancreatic Leak | 32 (13) | 19 (12) | 13 (14) | 0.662 |
| Clavien >2 | 38 (59) | 21 (55) | 17 (63) | 0.535 |
| LOS (days) | 8 (6–11) | 8 (6–11) | 7 (6–10) | 0.146 |
| 90 Day Readmission | 71 (29) | 44 (28) | 27 (30) | 0.678 |
| 30 Day Mortality | 3 (1.2) | 3 (1.9) | 0 (0) | 0.556 |
| 90 Day Mortality | 7 (3.2) | 7 (4.4) | 1 (1.1) | 0.265 |
| Adjuvant Chemotherapy | 172 (72.3) | 113 (74.3) | 59 (68.6) | 0.342 |
| Adjuvant Cycles | 4 (3–6) | 5 (3–6) | 4 (3–6) | 0.027 |
| Adjuvant Radiation | 18 (7.2) | 14 (8.8) | 4 (4.4) | 0.308 |
| Overall Survival | 31.5 (27.6, 39.6) | 26.0 (20.5, 31.7) | 55.7 (32.0, 58.3) | 0.0008 |
| Disease-free Survival | 12.4 (10.6, 14.6) | 10.5 (8.5, 12.4) | 16.5 (12.7, 26.8) | 0.011 |
| Median Follow-Up | 34.7 (31.9, 41.1) | 41.1 (31.2, 48.3) | 32.8 (23.4, 37.4) | 0.071 |
Abbreviations: BMI: body mass index (in Kg/m2); CCI: Charlson comorbidity index; ASA: American society of anesthesiology class; NT: neoadjuvant chemotherapy; Gem: Gemcitabine based chemotherapy; 5-FU: 5-Fluorouracil based chemotherapy; R1: positive margin; LV: lymphovascular invasion; PNI: perineural invasion; LOS: length of stay; Overall survival, Disease-free survival and median follow-up in months.
Serum CA19-9 Response Analysis
Multiple cut-off points were investigated to determine the relationship between Ca19-9 response and survival (Table 2). These cut-offs included normalization, and 25–85% response (Supplemental Figure 1). Normalization occurred in 98 (39.2%) of patients and was not associated with improved survival (35.17 vs 29.43 months; p=0.173). While response ≥45% was predictive of improved survival (35 vs 20 months, p=0.018), a response ≥85% was found to be optimal (55.7 vs 25.97 months; p<0.0001). Furthermore, grouping subjects by CA19-9 response into three groups, survival is significantly improved from a median of 20 months for <65%, 28.17 months for 65–84%, and 55.7 months for response ≥85%. A serum CA19-9 response ≥85% remained a strong independent predictor of survival (HR:0.47; p=0.007) on multivariate analysis (Table 3). Other independent predictors of better survival were receipt of adjuvant chemotherapy (HR:0.57, p=0.007), and higher pre-operative albumin (HR:0.66, p=0.04). Higher pathologic grade (HR:2, p<0.001) and vein resection (HR:1.81, p=0.004) were independently predictive of worse survival.
TABLE 2.
Serum CA19-9 response threshold and survival
| CA19-9 Cut-off | OS Below | OS Above | Difference | P-value |
|---|---|---|---|---|
| Normalization | 29.4 | 35.2 | 5.8 | 0.173 |
| 15% | 28 | 31.5 | 3.5 | 0.620 |
| 25% | 28 | 31.5 | 3.5 | 0.459 |
| 35% | 20 | 31.7 | 11.7 | 0.103 |
| 40% | 20.1 | 32 | 11.9 | 0.076 |
| 45% | 20 | 35 | 15 | 0.019 |
| 50% | 19.7 | 35 | 15.3 | 0.016 |
| 55% | 20 | 35.1 | 15.1 | 0.012 |
| 65% | 19.7 | 39.6 | 19.9 | 0.001 |
| 75% | 21 | 42.6 | 21.6 | 0.001 |
| 80% | 22.6 | 45 | 22.4 | 0.001 |
| 85% | 26 | 55.7 | 29.7 | 0.001 |
| 90% | 28 | 55.7 | 27.7 | 0.011 |
Abbreviations: OS below: overall survival in months of patients with CA19-9 response below threshold; OS above: overall survival in months of patients with CA19-9 response above threshold
TABLE 3.
Multivariate Analysis for Serum CA19-9 response and survival
| Variables | Survival | ||
|---|---|---|---|
| Serum CA19-9 Response Groups | |||
| Response <65% | 1 | ||
| Response 65–84% | 0.726 | 0.458, 1.151 | 0.174 |
| Response ≥85% | 0.474 | 0.278, 0.812 | 0.007 |
| Pre-operative Albumin (g/dL) | 0.661 | 0.445, 0.981 | 0.040 |
| Vein Resection | 1.807 | 1.206, 2.705 | 0.004 |
| Margin Status (R1) | 1.540 | 0.971, 2.443 | 0.067 |
| Pathologic Grade | 1.999 | 1.361, 2.937 | <0.0001 |
| Positive Nodes (N1) | 1.415 | 0.913, 2.192 | 0.120 |
| Adjuvant Chemotherapy | 0.573 | 0.382, 0.861 | 0.007 |
Abbreviations: R1: positive margin, N1: any positive nodes.
Patient Characteristics by CA19-9 Response
Patients with ≥85% CA19-9 response had on average higher pre-NT CA19-9 (626 vs 189; p=0.0001) and lower post-NT CA19-9 (37 vs 71; p=0.0001). This high response group spent more time in NT therapy (3[2–6] vs 3[2–4] cycles; p=0.0003), but less time in AT (4[3–6] vs 5[3–6] cycles; p=0.027). However total cycles between the two groups were not significantly different (8[6–10] vs 8[7–9] cycles; p=0.639). Despite no difference in pre-operative clinical tumor size, the ≥85% CA19-9 response group’s pathologic T-size was smaller (2.41 vs 2.97; p=0.001) and less lymphovascular invasion (p=0.001) and perineural invasion (p=0.001) was seen (Table 1).
Serum CA19-9 Response and T-size Reduction
There was a linear relationship between CA19-9 reduction and T-size reduction (p=0.002; Supplemental Figure 2). The relationship between different levels of T-size reduction and survival was also analyzed and revealed that a 25% reduction in T-size had improved survival (56.97 vs 28.17 months; p=0.016).On multivariate analysis, serum CA19-9 response ≥85% was a strong independent predictor of 25% T-size reduction (HR:2.40; p=0.007) (Table 4). No other factors including NT cycles or resectability predicted T-size reduction.
TABLE 4.
Predictors of T-Size reduction
| Variables | Univariate Analysis | Multivariate Analysis | ||||
|---|---|---|---|---|---|---|
| Age (Years) | 0.993 | 0.963, 1.023 | 0.631 | |||
| Sex (Female) | 1.804 | 1.000, 3.255 | 0.050 | 1.887 | 0.998, 3.566 | 0.051 |
| ASA Class | 1.410 | 0.717, 2.773 | 0.319 | |||
| BMI (Kg/m2) | 1.012 | 0.958, 1.068 | 0.674 | |||
| Prior Abdominal Surgery | 1.384 | 0.770, 2.489 | 0.278 | |||
| Pre-operative Albumin (g/dL) | 1.072 | 0.591, 1.945 | 0.819 | |||
| CCI Age Unadjusted | 1.035 | 0.743, 1.442 | 0.839 | |||
| NT Cycle Number | 1.207 | 1.046, 1.393 | 0.010 | 1.137 | 0.973, 1.330 | 0.107 |
| Type of NT (Gem vs 5FU) | 1.672 | 0.839, 3.333 | 0.144 | |||
| Pre-NT CA19-9 (IU/mL) | 1.0001 | 0.9999, 1.0002 | 0.110 | |||
| Post-NT CA19-9 (IU/mL) | 0.9998 | 0.999, 1.0004 | 0.456 | |||
| Percent Change in CA19-9 | 0.187 | 0.064, 0.544 | 0.002 | |||
| CA19-9 Response ≥85% | 2.760 | 1.515, 5.027 | 0.001 | 2.401 | 1.268, 4.545 | 0.007 |
| Anatomic Type | ||||||
| Resectable | 1 | 1 | ||||
| Borderline Resectable | 2.148 | 1.104, 4.177 | 0.024 | 1.964 | 0.934, 4.129 | 0.075 |
| Locally Advanced | 0.938 | 0.306, 2.871 | 0.910 | 1.015 | 0.300, 3.432 | 0.981 |
Abbreviations: BMI: body mass index (in Kg/m2); CCI: Charlson comorbidity index; ASA: American society of anesthesiology class; NT: neoadjuvant chemotherapy; Gem: Gemcitabine based chemotherapy; 5-FU: 5-Fluorouracil based chemotherapy.
Serum CA19-9 Response and Margin
Patients were grouped based on serum CA19-9 response and margin status (Supplemental Figure 3). Patients with ≥85% CA19-9 response and negative margin had the best survival whereas who had <85% CA19-9 response and positive margin had the worst (56.97 vs 17.73 months; p<0.0001). This relationship held true for disease-free survival, as well (23.57 vs 12.07 months; p<0.0001).
An additional multivariate model was performed using pre-operative and pathologic variables to evaluate the relationship between CA19-9 response and R0/R1 resection (Table 5). Patients with an R1 resection and <85% CA19-9 response had a greater risk of recurrence (HR=2.64; p<0.0001) than patients with an R1 resection who had ≥85% CA19-9 response (HR=1.77; p=0.136) which was not significantly different from the reference. Additionally, in patients with an R0 resection, having a <85% response had worse survival (HR=1.52; p=0.042) than patients with ≥85% CA19-9 response. This association did not hold true for overall survival.
TABLE 5.
Multivariate analysis for serum CA19-9 response and Margin
| Variables | Disease-free Survival | Survival | ||||
|---|---|---|---|---|---|---|
| CA19-9 Response ≥85% & R0 | 1 | 1 | ||||
| CA19-9 Response <85% & R0 | 1.518 | 1.014, 2.271 | 0.042 | 1.690 | 0.993, 2.878 | 0.053 |
| CA19-9 Response ≥85% & R1 | 1.775 | 0.835, 3.771 | 0.136 | 2.047 | 1.010, 4.148 | 0.047 |
| CA19-9 Response <85% & R1 | 2.638 | 1.537, 4.531 | <0.0001 | 2.619 | 1.032, 6.649 | 0.043 |
| BMI (Kg/m2) | 0.955 | 0.916, 0.996 | 0.032 | |||
| Pre-operative Albumin (g/dL) | 0.649 | 0.464, 0.908 | 0.012 | 0.581 | 0.401, 0.841 | 0.004 |
| Vein resection | 1.467 | 1.046, 2.058 | 0.026 | 1.329 | 0.899, 1.965 | 0.153 |
| T-size (cm) | 1.252 | 1.009, 1.554 | 0.041 | |||
| Positive Nodes (N1) | 1.616 | 1.124, 2.325 | 0.010 | 1.137 | 0.973, 1.330 | 0.107 |
| Number of Positive Nodes | 1.078 | 1.036, 1.122 | <0.0001 | |||
| Adjuvant Chemotherapy | 0.549 | 0.363, 0.830 | 0.004 | |||
Abbreviations: BMI: body mass index (in Kg/m2); T-size: pathologic tumor size; N1: any positive lymph nodes.
Cohort with Markedly Elevated Serum CA19-9
Sixty six patients had markedly elevated [>1000] serum CA19-9 pre-NT: 39 (59.1%) had a CA19-9 response ≥85%. Median post-treatment serum CA19-9 levels were 83 IU/dL for those with response ≥85% and 617 IU/dL for the cohort with response <85%. While there were no significant differences in baseline patient characteristics between the two groups or in terms of pathologic or short-term post-operative outcomes, patients with a response >85% had significantly longer survival (42.6 vs 20 months, p=0.010). Evaluating multiple serum CA19-9 response cut-offs for this subgroup revealed improved survival with higher cut-offs [data not shown]. This did not hold true for DFS.
DISCUSSION
This study has shown that increasing magnitude of CA19-9 response is associated with better survival, with a response ≥85% to NT as optimal. The data also shows that serum CA19-9 response predicted T-size reduction, but the impact of T-size reduction on survival, independently or with CA19-9 response, remains unclear. It is well established that patients with R1 resections have worse survival; however, it was unknown whether CA19-9 response could mitigate this variable. This data indicates lower risk of recurrence in patients with ≥85% CA19-9 response compared to patients with R1 resections in addition to R0 resections. While this did not hold true for survival, it may be due to sample size limitations. The implications of this data can guide clinical decision making for determining neoadjuvant regimens and duration prior to surgical resection.
Only a few studies have investigated the role of serum CA19-9 response to NT and survival. Tzeng et al. investigated the role of CA19-9 in 141 patients with borderline resectable disease undergoing NT24. Serum CA19-9 normalization post-NT was independently associated with survival, which is also consistent with another study by Williams et al. on 109 patients with borderline resectable and locally advanced tumors undergoing NT25. Longer pre-operative treatment duration was independently associated with survival. However, patients who normalized did not appear to benefit from additional months of treatment. In these two studies, 42% and 62% of patients that proceeded to resection had normalized their serum CA19-9, respectively. Notably, these studies did not take into account the magnitude of change reflected by this normalized cohort. A more recent study by Tsai et al. on serum CA19-9 response to NT for patients with resectable and borderline resectable pancreas cancer included 131 patients26. Patients with pre-treatment serum CA19-9 <80 IU/dL were analyzed separately (es. The cohort with serum CA19-9 ≥80 IU/dL was stratified into normalization and four groups based on the magnitude of response. Among patients with serum CA19-9 ≥80 IU/dL, normalization post-NT predicted improved survival. Differing magnitudes of response to NT were not associated with any survival benefit. These three studies suggested that serum CA19-9 normalization post-NT should prompt proceeding to resection. In the present study, 98 (39%) patients normalized their CA19-9. Serum CA19-9 normalization post-NT did not lead to improved survival. It showed that the higher the magnitude of CA19-9 response to NT the better the survival. A previous study by our group showed that the optimal serum CA19-9 cut-off for survival was ≥50%, with a trend at ≥90%22. With a larger sample size, the optimal cut-off shifted to ≥85%. Furthermore, despite patients with markedly elevated serum CA19-9 still having elevated serum CA19-9 post-treatment, survival improved with the magnitude of response.
This study includes resectable, borderline resectable, and locally advanced tumors that all underwent resection. Unique to this analysis was grouping patients into four groups by their serum CA19-9 response and their margin status. The patients with both a serum CA19-9 ≥85% response and a negative margin resection had the best survival. This group was set as a reference and patients with <85% serum CA19-9 response and positive margin were shown to have the highest risk of recurrence. In patients with R1 resections, those with ≥85% CA19-9 response had better survival. Although shown previously for post-surgery follow-up serum CA19-9 levels and margin27, this study is the first to show that serum CA19-9 response to NT may have mitigated the impact of positive margin resection on recurrence.
Determining therapeutic response to NT is an ongoing challenge. Ferrone et al. studied the role of radiologic response in predicting resectability in 188 patients undergoing resection for PDAC, where 40 patients received neoadjuvant FOLFIRINOX23. It was shown that radiology did not predict resectability after NT therapy. In addition, Reni et al. assessed the role of pre-operative factors (including CA19-9 response ≥50%) in predicting survival and resection in 223 patients with borderline or locally advanced disease undergoing NT28. In resected patients, pre-operative factors such as tumor anatomy or radiologic response were not associated with survival; however, serum CA19-9 response ≥50% was predictive. Tumor radiologic response was not assessed in this study since previous studies have shown this to be an inadequate marker for treatment response due to the inability to distinguish tumor from fibrosis on imaging. However, the change in T-size was estimated by subtracting clinical T-size by EUS from pathologic T-size after resection in this study. This is the first study to perform this analysis. T-size reduction and CA19-9 reduction after NT were shown to be associated. T-size reduction ≥25% was shown to predict improved survival and the group with both CA19-9 response ≥ 85% and T-size reduction ≥25% had the best. Hopefully, future liquid or solid biomarkers will help guide decision-making for NT, but the current data remains limited.
This study has several limitations. It is a retrospective study and may suffer from selection bias. The cohort of the patients were treated on and off available protocols, and there was no standardized method on when to operate. Serum CA19-9 levels were not necessarily used to determine the length of NT before surgical resection. Instead, it was influenced by the oncologist’s and surgeon’s preferences. In addition, clinical trial enrollment and patient performance can also influence this decision. Additionally, there is no perfect way to assess T-size reduction using clinical staging. Pre-operative pathology would have offered a gold-standard, but this is not technically feasible. Thus, EUS is the best estimation when available. An additional systematic bias would be the fact that EUS and CT T-size estimates have been shown to be potentially inaccurate29. While Evans grade is known to be associated with serum CA19-9 response, this was not available for the majority of patients (only ones on clinical trials). T-size reduction between size at diagnosis and on pathology was employed as a surrogate pathologic outcome. There were some missing CA19-9 levels pre-NT and post-NT which lowered our sample size. This was owing to concomitant total bilirubin elevation and due to some surgeons electing not to take CA19-9 levels. Lastly, the American Joint Commission for Cancer switched to version 8 at the end of this study and differences in classification of margin could have different results. Despite these limitations, this study offers a large sample size, reflects the clinical situation faced by providers at a large volume referral center and encompasses a breadth of chemotherapeutic regimens currently in clinical practice.
CONCLUSION
Serum CA19-9 response to NT has been shown to be a useful marker in predicting patient survival after NT and resection but the optimal reduction has not been determined. This study advocates for ≥85% CA19-9 response to double survival but needs to be validated externally. Additionally, T-Size reduction on final pathology correlates with CA19-9 response and improved survival. This study is also the first to show that CA19-9 response may mitigate inferior survival in R1 resections. Currently, surgeons and oncologists or clinical trial participation decide on when to proceed to resection based on multifactorial characteristics. The role of serum CA19-9 as a potential endpoint instead of cycle number or treatment duration may allow for decision-making that is better tailored to the patient’s biologic response and offer improved outcomes.
Supplementary Material
SUPPLEMENTAL FIGURE 3. Serum CA19-9 response and margin
SUPPLEMENTAL FIGURE 2. Serum CA19-9 response and T-size reduction
SUPPLEMENTAL FIGURE 1. Serum CA19-9 response thresholds and survival
SYNOPSIS.
This manuscript investigates different thresholds of serum CA19-9 response to neoadjuvant chemotherapy in patients with pancreatic adenocarcinoma. It ultimately shows that a response ≥85% is optimal in terms of survival, predictive of T-size reduction, and may mitigate the impact of R1 resections on recurrence.
Acknowledgments
Disclosures:
MH received funding from the VA in the way of salary support and has education grants from SAGES Society and Intuitive Surgical.
REFERENCES
- 1.Kuhlmann KF, de Castro SM, Wesseling JG, et al. Surgical treatment of pancreatic adenocarcinoma; actual survival and prognostic factors in 343 patients. Eur J Cancer. March 2004;40(4):549–558. [DOI] [PubMed] [Google Scholar]
- 2.Neoptolemos JP, Moore MJ, Cox TF, et al. Effect of adjuvant chemotherapy with fluorouracil plus folinic acid or gemcitabine vs observation on survival in patients with resected periampullary adenocarcinoma: the ESPAC-3 periampullary cancer randomized trial. JAMA. July 11 2012;308(2):147–156. [DOI] [PubMed] [Google Scholar]
- 3.Oettle H, Post S, Neuhaus P, et al. Adjuvant chemotherapy with gemcitabine vs observation in patients undergoing curative-intent resection of pancreatic cancer: a randomized controlled trial. JAMA. January 17 2007;297(3):267–277. [DOI] [PubMed] [Google Scholar]
- 4.Conroy T, Desseigne F, Ychou M, et al. FOLFIRINOX versus gemcitabine for metastatic pancreatic cancer. N Engl J Med. May 12 2011;364(19):1817–1825. [DOI] [PubMed] [Google Scholar]
- 5.Von Hoff DD, Ervin T, Arena FP, et al. Increased survival in pancreatic cancer with nab-paclitaxel plus gemcitabine. N Engl J Med. October 31 2013;369(18):1691–1703. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Merkow RP, Bilimoria KY, Tomlinson JS, et al. Postoperative complications reduce adjuvant chemotherapy use in resectable pancreatic cancer. Ann Surg. August 2014;260(2):372–377. [DOI] [PubMed] [Google Scholar]
- 7.Tzeng CW, Tran Cao HS, Lee JE, et al. Treatment sequencing for resectable pancreatic cancer: influence of early metastases and surgical complications on multimodality therapy completion and survival. J Gastrointest Surg. January 2014;18(1):16–24; discussion 24–15. [DOI] [PubMed] [Google Scholar]
- 8.Wu W, He J, Cameron JL, et al. The impact of postoperative complications on the administration of adjuvant therapy following pancreaticoduodenectomy for adenocarcinoma. Ann Surg Oncol. September 2014;21(9):2873–2881. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Rhim AD, Mirek ET, Aiello NM, et al. EMT and dissemination precede pancreatic tumor formation. Cell. January 20 2012;148(1–2):349–361. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Yachida S, Jones S, Bozic I, et al. Distant metastasis occurs late during the genetic evolution of pancreatic cancer. Nature. October 28 2010;467(7319):1114–1117. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Artinyan A, Anaya DA, McKenzie S, Ellenhorn JD, Kim J. Neoadjuvant therapy is associated with improved survival in resectable pancreatic adenocarcinoma. Cancer. May 15 2011;117(10):2044–2049. [DOI] [PubMed] [Google Scholar]
- 12.Cheng TY, Sheth K, White RR, et al. Effect of neoadjuvant chemoradiation on operative mortality and morbidity for pancreaticoduodenectomy. Ann Surg Oncol. January 2006;13(1):66–74. [DOI] [PubMed] [Google Scholar]
- 13.de Geus SW, Eskander MF, Bliss LA, et al. Neoadjuvant therapy versus upfront surgery for resected pancreatic adenocarcinoma: A nationwide propensity score matched analysis. Surgery. March 2017;161(3):592–601. [DOI] [PubMed] [Google Scholar]
- 14.Lutfi W, Talamonti MS, Kantor O, et al. Perioperative chemotherapy is associated with a survival advantage in early stage adenocarcinoma of the pancreatic head. Surgery. September 2016;160(3):714–724. [DOI] [PubMed] [Google Scholar]
- 15.Neoptolemos JP, Palmer DH, Ghaneh P, et al. Comparison of adjuvant gemcitabine and capecitabine with gemcitabine monotherapy in patients with resected pancreatic cancer (ESPAC-4): a multicentre, open-label, randomised, phase 3 trial. Lancet. March 11 2017;389(10073):1011–1024. [DOI] [PubMed] [Google Scholar]
- 16.Oettle H, Neuhaus P, Hochhaus A, et al. Adjuvant chemotherapy with gemcitabine and long-term outcomes among patients with resected pancreatic cancer: the CONKO-001 randomized trial. JAMA. October 9 2013;310(14):1473–1481. [DOI] [PubMed] [Google Scholar]
- 17.Epelboym I, Zenati MS, Hamad A, et al. Analysis of Perioperative Chemotherapy in Resected Pancreatic Cancer: Identifying the Number and Sequence of Chemotherapy Cycles Needed to Optimize Survival. Ann Surg Oncol. September 2017;24(9):2744–2751. [DOI] [PubMed] [Google Scholar]
- 18.Bockhorn M, Uzunoglu FG, Adham M, et al. Borderline resectable pancreatic cancer: a consensus statement by the International Study Group of Pancreatic Surgery (ISGPS). Surgery. June 2014;155(6):977–988. [DOI] [PubMed] [Google Scholar]
- 19.Katz MH, Fleming JB, Bhosale P, et al. Response of borderline resectable pancreatic cancer to neoadjuvant therapy is not reflected by radiographic indicators. Cancer. December 1 2012;118(23):5749–5756. [DOI] [PubMed] [Google Scholar]
- 20.Xia BT, Fu B, Wang J, et al. Does radiologic response correlate to pathologic response in patients undergoing neoadjuvant therapy for borderline resectable pancreatic malignancy? J Surg Oncol. March 2017;115(4):376–383. [DOI] [PubMed] [Google Scholar]
- 21.Bergquist JR, Puig CA, Shubert CR, et al. Carbohydrate Antigen 19–9 Elevation in Anatomically Resectable, Early Stage Pancreatic Cancer Is Independently Associated with Decreased Overall Survival and an Indication for Neoadjuvant Therapy: A National Cancer Database Study. J Am Coll Surg. July 2016;223(1):52–65. [DOI] [PubMed] [Google Scholar]
- 22.Boone BA, Steve J, Zenati MS, et al. Serum CA 19–9 response to neoadjuvant therapy is associated with outcome in pancreatic adenocarcinoma. Ann Surg Oncol. December 2014;21(13):4351–4358. [DOI] [PubMed] [Google Scholar]
- 23.Ferrone CR, Marchegiani G, Hong TS, et al. Radiological and surgical implications of neoadjuvant treatment with FOLFIRINOX for locally advanced and borderline resectable pancreatic cancer. Ann Surg. January 2015;261(1):12–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Tzeng CW, Balachandran A, Ahmad M, et al. Serum carbohydrate antigen 19–9 represents a marker of response to neoadjuvant therapy in patients with borderline resectable pancreatic cancer. HPB (Oxford). May 2014;16(5):430–438. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Williams JL, Kadera BE, Nguyen AH, et al. CA19-9 Normalization During Pre-operative Treatment Predicts Longer Survival for Patients with Locally Progressed Pancreatic Cancer. J Gastrointest Surg. July 2016;20(7):1331–1342. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Tsai S, George B, Wittmann D, et al. Importance of Normalization of CA19-9 Levels Following Neoadjuvant Therapy in Patients With Localized Pancreatic Cancer. Ann Surg. October 11 2018. [DOI] [PubMed] [Google Scholar]
- 27.Kinsella TJ, Seo Y, Willis J, et al. The impact of resection margin status and postoperative CA19-9 levels on survival and patterns of recurrence after postoperative high-dose radiotherapy with 5-FU-based concurrent chemotherapy for resectable pancreatic cancer. Am J Clin Oncol. October 2008;31(5):446–453. [DOI] [PubMed] [Google Scholar]
- 28.Reni M, Zanon S, Balzano G, et al. Selecting patients for resection after primary chemotherapy for non-metastatic pancreatic adenocarcinoma. Ann Oncol. November 1 2017;28(11):2786–2792. [DOI] [PubMed] [Google Scholar]
- 29.Arvold ND, Niemierko A, Mamon HJ, Fernandez-del Castillo C, Hong TS. Pancreatic cancer tumor size on CT scan versus pathologic specimen: implications for radiation treatment planning. Int J Radiat Oncol Biol Phys. August 1 2011;80(5):1383–1390. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
SUPPLEMENTAL FIGURE 3. Serum CA19-9 response and margin
SUPPLEMENTAL FIGURE 2. Serum CA19-9 response and T-size reduction
SUPPLEMENTAL FIGURE 1. Serum CA19-9 response thresholds and survival