Abstract
Background:
Few studies have evaluated outcomes of total neoadjuvant therapy (TNT) compared to single modality neoadjuvant therapy (SMNT) or surgery first (SF) for pancreatic ductal adenocarcinoma (PDAC).
Methods:
A single institution retrospective review of PDAC patients who underwent pancreatectomy was conducted (1993–2019). Overall survival (OS) estimates from diagnosis and from surgery were determined using Kaplan-Meier methods; Cox proportional hazards models adjusted for covariates.
Results:
Surgery was performed upfront (SF) in 168 (46.9%), while 111 (31.0%) had chemotherapy or chemoradiation prior to resection (SMNT), and 79 (22.1%) underwent TNT (chemotherapy and chemoradiation). Resection margins were more frequently R0 in the TNT group (86.1%) compared to SMNT (64.0%) and SF (72%)(p<0.001). Complete pathologic response was more common in the TNT group (10.1%) compared to SMNT (3.6%) or SF (0.6%)(p=0.001), resulting in prolonged survival (median OS = 100.2 months). TNT patients demonstrated longer median OS from surgery (33.6 months) compared to SF (19.1 months) and SMNT (17.4 months) (p=0.010), which persisted after controlling for covariates.
Conclusions:
TNT is associated with more frequent complete pathologic response, a higher rate of margin negative resection, and prolonged OS as compared to SF or SMNT. Additional studies to identify subgroups who derive the greatest benefit are warranted.
Keywords: Pancreas, cancer, neoadjuvant, outcomes, survival
INTRODUCTION
The incidence of pancreatic ductal adenocarcinoma (PDAC) continues to rise, with the American Cancer Society predicting 60,430 new cases of PDAC and 48,220 deaths in the United States in 20211. Unfortunately, only 10–15% of PDAC patients are eligible for potentially curative surgical resection at diagnosis2,3. Overall survival (OS) remains low for this aggressive disease owing to frequent distant metastasis, highlighting the critical importance of systemic therapy in the multimodality treatment paradigm of PDAC. There is increasing interest in the role of neoadjuvant therapy for patients with PDAC, particularly owing to the observation that up to 40% of patients do not receive postoperative chemotherapy and an additional 20% do not complete the intended course due to postoperative complications, diminished performance status, or early disease progression4. This phenomenon was reaffirmed by the recent Dutch PREOPANC Trial, wherein 76% of patients in the neoadjuvant chemoradiation group completed adjuvant therapy (intention to treat analysis) as compared to 51% of the surgery first group receiving postoperative chemotherapy5. In a disease where the failure of therapy is often systemic, data such as this provide a strong basis to consider delivering adjuvant therapy preoperatively to optimize completion of the intended course.
Total neoadjuvant therapy (TNT) is comprised of full course systemic chemotherapy preceding or following the delivery of chemoradiation prior to surgery. Our institution started employing TNT for borderline resectable (BR) and locally advanced (LA) PDAC in the mid-1990s after anecdotally observing profound responses in patients with initially unresectable PDAC after implementation of this regimen in a “palliative” manner, some of which were converted to resectable disease. The rationale for TNT as a neoadjuvant regimen includes delivery of systemic therapy to eradicate micrometastatic disease, facilitation of margin-negative (R0) resection via sterilization of at-risk margins by radiation to improve local control, potentially downsizing of the primary tumor to render the operation easier, and to gauge tumor response to therapy as a marker of disease biology. To this latter point, TNT (and neoadjuvant therapy in general) may allow for time to identify aggressive disease prior to surgery and selection of patients who would not benefit from surgery owing to high likelihood of rapid systemic failure6,7. Owing to these potential benefits, interest in applying neoadjuvant approaches to resectable disease have arisen in recent trials such as SWOGS15058.
Although consensus statements support the use of neoadjuvant therapy for BR PDAC, the optimal strategy to accomplish this is not yet defined9. The aim of this study was to evaluate and compare OS and clinicopathologic outcomes in PDAC patients who underwent surgery first (SF), single modality neoadjuvant therapy (SMNT), and TNT. We hypothesized that TNT would yield more frequent R0 resection, greater pathologic responses, and longer OS as compared to SMNT or SF.
RESULTS
Demographic and Tumor Characteristics of the Analytic Cohorts
During the study period, 918 patients (34/year) with potentially resectable, clinical stage I-III PDAC were evaluated for resection, of which a total of 358 patients eventually underwent surgery and were included in this study (Table 1). The distribution of patients by treatment strategy is illustrated in Supplementary Figure 1. The median age at diagnosis was 68 years, 51.4% were male, and 91.9% were Caucasian. The ECOG performance status scores did not differ between the groups, nor did the prevalence of most of the analyzed comorbidities; there were differences between the groups, however, in the proportion of patients having coronary artery disease (CAD) and hyperlipidemia (HLD), with these conditions being more prevalent in the TNT group than the SMNT or SF groups (43.0% vs 21.5% vs 31.0%, p=0.007 and 50.0% vs 32.7% vs 26.3%, p<0.001 for CAD and HLD, respectively).
Table 1. Demographic, pre-operative, and clinicopathologic characteristics by treatment group.
SF, surgery first; SMNT, single modality neoadjuvant therapy; TNT, total neoadjuvant therapy; BMI, body mass index.
| Overall cohort (N=358) | SF (N=168) | SMNT (N=111) | TNT (N=79) | ||
|---|---|---|---|---|---|
| N (%) | N (%) | N (%) | N (%) | p-value | |
| Gender | |||||
| Female | 174 (48.6) | 85 (50.6) | 56 (50.5) | 33 (41.8) | 0.40 |
| Male | 184 (51.4) | 83 (49.4) | 55 (49.5) | 46 (58.2) | |
| Race | |||||
| White | 329 (91.9) | 161 (95.8) | 100 (90.1) | 68 (86.1) | 0.053 |
| Black | 18 (5.0) | 4 (2.4) | 6 (5.4) | 8 (10.1) | |
| Other | 11 (3.1) | 3 (1.8) | 5 (4.5) | 3 (3.8) | |
| Age at diagnosis, years | |||||
| Median [IQR] * | 68 [61, 75] | 70 [61.5, 76] | 68 [61, 75] | 67 [59, 73] | 0.079 |
| ECOG | |||||
| 0 | 262 (78.2) | 125 (82.2) | 82 (78.1) | 55 (70.5) | 0.32 |
| 1 | 67 (20.0) | 25 (16.5) | 21 (20.0) | 21 (26.9) | |
| 2 | 6 (1.8) | 2 (1.3) | 2 (1.9) | 2 (2.6) | |
| Missing | 23 | 16 | 6 | 1 | |
| Hypertension | |||||
| Yes | 201 (56.8) | 89 (53.0) | 62 (57.9) | 50 (63.3) | 0.31 |
| No | 153 (43.2) | 79 (47.0) | 45 (42.1) | 29 (36.7) | |
| Missing | 4 | 0 | 4 | 0 | |
| Smoker | |||||
| Yes | 146 (41.4) | 62 (36.9) | 50 (47.2) | 34 (43.0) | 0.23 |
| No | 207 (58.6) | 106 (63.1) | 56 (52.8) | 45 (57.0) | |
| Missing | 5 | 0 | 5 | 0 | |
| Diabetes Mellitus | |||||
| Yes | 123 (34.8) | 62 (36.9) | 36 (33.6) | 25 (31.6) | 0.70 |
| No | 231 (65.2) | 106 (63.1) | 71 (66.4) | 54 (68.4) | |
| Missing | 4 | 0 | 4 | 0 | |
| Coronary.Artery Disease | |||||
| Yes | 109 (30.8) | 52 (31.0) | 23 (21.5) | 34 (43.0) | 0.007 |
| No | 245 (69.2) | 116 (69.0) | 84 (78.5) | 45 (57.0) | |
| Missing | 4 | 0 | 4 | 0 | |
| Hyperlipidemia | |||||
| Yes | 110 (33.5) | 41 (26.3) | 32 (32.7) | 37 (50.0) | 0.002 |
| No | 218 (66.5) | 115 (73.7) | 66 (67.4) | 37 (50.0) | |
| Missing | 30 | 12 | 13 | 5 | |
| BMI ≥ 25 | |||||
| Yes | 188 (66.2) | 92 (71.3) | 44 (55.7) | 52 (68.4) | 0.065 |
| No | 96 (33.8) | 37 (28.7) | 35 (44.3) | 24 (31.6) | |
| Missing | 74 | 39 | 32 | 3 | |
| NCCN Resectability Classification | |||||
| Resectable | 155 (46.8) | 121 (79.1) | 22 (21.6) | 12 (15.8) | <0.001 |
| Borderline | 144 (43.5) | 29 (19.0) | 61 (59.8) | 54 (71.1) | |
| Locally advanced | 32 (9.7) | 3 (2.0) | 19 (18.6) | 10 (13.2) | |
| Missing | 27 | 15 | 9 | 3 | |
| Clinical Stage | |||||
| 0 | 5 (1.5) | 5 (3.4) | 0 (0.0) | 0 (0.0) | <0.001 |
| IA | 40 (12.4) | 32 (21.6) | 5 (5.0) | 3 (4.0) | |
| IB | 133 (41.1) | 66 (44.6) | 39 (38.6) | 28 (37.3) | |
| IIA | 31 (9.6) | 13 (8.8) | 14 (13.9) | 4 (5.3) | |
| IIB | 59 (18.2) | 25 (16.9) | 19 (18.8) | 15 (20.0) | |
| III | 56 (17.3) | 7 (4.7) | 24 (23.8) | 25 (33.3) | |
| Missing | 34 | 20 | 10 | 4 | |
| Initial CA19–9 | |||||
| N, median | 309190 | 151194 | 82194 | 76188 | 0.93 |
| IQR * | [50.1, 672] | [61.9, 545] | [52.1, 707] | [33, 1062.5] | |
| Post Neoadjuvant CA 19–9 | |||||
| N, median | 7039.9 | 7537 | 0.35 | ||
| IQR * | [11, 178] | [12.5, 68] |
Interquartile range given as [25th percentile,75th percentile]
A larger proportion of clinical stage IA tumors were present in the SF group compared with the SMNT and TNT groups (21.6% vs 5.0% vs 4.0%, respectively, Table 1) and a smaller proportion of clinical stage III tumors were present in the SF group compared with the SMNT and TNT groups (4.7% vs 23.8% vs 33.3%, respectively, p<0.001). Differences in resectability classification also existed between the groups, with a higher proportion of patients in the SF group having resectable disease and a higher proportion of patients in the TNT group having BR disease (p<0.001). The median initial CA19–9 level did not differ between the three groups, and post-neoadjuvant CA19–9 values did not differ between the SMNT and TNT groups.
Treatment Characteristics Amongst SF, SMNT, and TNT
In the TNT group, a median of 13.1 weeks of chemotherapy was given (IQR 9.6–20.7 weeks, range 5.6–66 weeks), most commonly with a gemcitabine-based regimen. Thirty-two patients in the TNT group (40.5%) received chemotherapy first, followed by radiation therapy (Table 2); the majority of patients underwent radiation therapy first followed by systemic chemotherapy (59.5%). In the SMNT group, 11 patients received neoadjuvant chemotherapy without radiation and 100 patients had chemoradiation prior to surgery. The chemotherapy regimens did not differ significantly between TNT and SMNT groups, however a larger proportion of TNT patients received 5-FU as a radiosensitizer with chemoradiation.
Table 2. Neoadjuvant and adjuvant chemotherapy and chemoradiotherapy treatment use and regimens for three treatment groups.
Values are listed as N (percentage of total in treatment group). SF, surgery first; SMNT, single modality neoadjuvant therapy; TNT, total neoadjuvant therapy.
| Treatment | ||||
|---|---|---|---|---|
| SF (N=168) | SMNT (N=111) | TNT (N=79) | ||
| N (%) | N (%) | N (%) | p-value | |
| Neoadjuvant chemo regimen: | 0.10 | |||
| Gemcitabine based | 4 (36.4) | 55 (69.6) | ||
| FOLFIRINOX | 6 (54.5) | 19 (24.1) | ||
| Other 5-FU based | 1 (9.1) | 4 (5.1) | ||
| Unspecified | 0 (0.0) | 1 (1.3) | ||
| Neoadjuvant radiosensitizer: | 0.027 | |||
| Gemcitabine based | 76 (76.0) | 49 (62.0) | ||
| Other 5-FU based | 20 (20.0) | 27 (34.2) | ||
| Unspecified | 4 (4.0) | 1 (1.3) | ||
| None(SBRT) | 0 (0.0) | 2 (2.5) | ||
| Received Adjuvant Chemo? | <0.001 | |||
| Yes | 119 (70.8) | 62 (55.9) | 31 (39.2) | |
| No | 38 (22.6) | 40 (36.0) | 45 (57.0) | |
| Unknown | 11 (6.5) | 9 (8.1) | 3 (3.8) | |
| Adjuvant chemo regimen: | 0.75 | |||
| Gemcitabine based | 105 (88.2) | 54 (87.1) | 27 (87.1) | |
| FOLFIRINOX | 3 (2.5) | 1 (1.6) | 1 (3.2) | |
| Other 5-FU based | 5 (4.2) | 4 (6.5) | 0 (0.0) | |
| Unspecified | 6 (5.0) | 3 (4.8) | 3 (9.7) | |
| Received Adjuvant Radiation? | N/A | |||
| Yes | 89 (53.0) | |||
| No | 65 (38.7) | |||
| Unknown | 14 (8.3) | |||
| Adjuvant radiosensitizer: | N/A | |||
| Gemcitabine based | 17 (19.1) | |||
| Other 5-FU based | 54 (60.7) | |||
| Unspecified | 17 (19.1) | |||
| None(SBRT) | 1 (1.1) | |||
The median length of time between diagnosis and resection differed significantly between the groups: 0.7 months (IQR 0.2–1.1) for SF, 3.9 months (IQR 3.4–4.9) for SMNT, and 7.7 months (IQR 6.3–8.8) for TNT (p<0.001). The majority of patients underwent pancreatoduodenectomy (96.1%); 4 (2.4%) in the SF group, 5 (4.5%) in the SMNT group, and 5 (6.3%) in the TNT group required total pancreatectomy (p=0.27). Vascular resection was undertaken more frequently in the TNT group (38.0%) than the other groups (31.5% SMNT vs 17.3% SF, p<0.001).
The groups differed significantly with respect to postoperative therapy received, with many fewer TNT patients undergoing any adjuvant therapy compared to the SMNT and SF groups (39.2% TNT vs 55.9% SMNT vs 76.8% SF, p<0.001, Table 2). TNT patients received a median of 11 weeks of postoperative treatment (IQR 6.7–18.1 weeks, range 2–107.1 weeks) while those in the SMNT group received a median of 15.6 weeks (IQR 10.6–20.9 weeks, range 2–62.4 weeks) and those in the SF group a median of 21.6 weeks (IQR 12.7–27 weeks, range 1–74.3 weeks) (p=0.13).
Oncologic outcomes based on type of neoadjuvant treatment
Resection margins were more commonly R0 in the TNT group compared to SMNT and SF (86.1% vs 64.0% vs 72%, respectively, p<0.001) (Table 3). There was no difference in the incidence of postoperative complications between the three treatment groups. Downstaging on final pathology was more frequent with TNT (65.3%) as compared to SMNT (41.4%) and SF (7.4%)(p<0.001). A greater proportion of patients in the TNT group had a pCR (10.1%) compared to the SMNT group (3.6%) or the SF group (0.6%) (p<0.001) (Table 3). Fibrosis estimates on final pathology were available for 87% of the cohort; median percentage of fibrosis was highest in the TNT group (80%), followed by 70.5% in the SMNT group and 0% in the SF group (p<0.001). TNT also had the highest rate of downstaging from clinical to final pathologic stage groupings (65% TNT vs. 41% SMNT vs. 7% SF, p<0.001).
Table 3. Pathologic resection findings after pancreatectomy for three treatment groups.
Values are listed as N (percentage of total in treatment group). R0 resection is defined as microscopically negative resection margins. SF, surgery first; SMNT, single modality neoadjuvant therapy; TNT, total neoadjuvant therapy; IQR, interquartile range.
| Treatment | ||||
|---|---|---|---|---|
| SF (N=168) | SMNT (N=111) | TNT (N=79) | ||
| N (%) | N (%) | N (%) | p-value | |
| pStage | ||||
| 0 | 1 (0.6) | 4 (3.6) | 8 (10.1) | <0.001 |
| IA | 16 (9.5) | 20 (18.0) | 26 (32.9) | |
| IB | 17 (10.1) | 24 (21.6) | 12 (15.2) | |
| IIA | 9 (5.4) | 4 (3.6) | 12 (15.2) | |
| IIB | 63 (37.5) | 31 (27.9) | 13 (16.5) | |
| III | 53 (31.5) | 19 (17.1) | 4 (5.1) | |
| IV | 9 (5.4) | 5 (4.5) | 4 (5.1) | |
| unknown | 0 (0.0) | 4 (3.6) | 0 (0.0) | |
| Downstaged * | ||||
| Yes | 11 (7.4) | 41 (41.4) | 49 (65.3) | <0.001 |
| No | 137 (92.6) | 58 (58.6) | 26 (34.7) | |
| Change in CA19–9 post-neoadjuvant, median (IQR) | −134 (−619 to −13) | −179 (−900 to −13) | 0.37** | |
| Margins | ||||
| R0 | 121 (72.0) | 71 (64.0) | 68 (86.1) | <0.001 |
| R1/R2 | 47 (28.0) | 39 (35.1) | 10 (12.7) | |
| unknown | 0 (0.0) | 1 (0.9) | 1 (1.3) | |
excludes 36 patients missing cStage or pStage
Wilcoxon rank sum test
There were 288 deaths during the follow-up period; of those who had not died, the median length of follow-up from diagnosis was 37.2 months, with the longest follow-up being 17.8 years, 12.0 years, and 14.5 years for SF, SMNT, and TNT groups, respectively. Median OS from diagnosis was highest in the TNT group at 42.0 months (95%CI=31.0–53.9), followed by and 23.1 months (95%CI=18.8–27.4) in the SMNT group and 20.1 months (95% CI=17.8–23.7) in the SF group (Figure 2A). To alleviate the lead time bias imposed by neoadjuvant therapy, survival analyses were performed from the time of surgery as well (Figure 2B). The median OS from the time of operation was 33.6 months (95%CI 22.2–41.7) in the TNT group, 17.4 months (95%CI 14.1–21.5) in the SMNT group, and 19.1 months (95%CI 17.1–23.0) in the SF group (p=0.011, Figure 2B). There was no significant difference in median recurrence free survival (RFS) amongst the three groups (SF = 13.9 months, SMNT = 11.5 months, TNT = 16.6 months, p=0.133, Supplementary Figure 2). The median OS from surgery for those with a pCR was 100.2 months (95%CI 44.4-undetermined), compared to median OS of 31.8 months (95%CI 19.1–37.2) in the pathologic stage IA, 24.8 months (95%CI 15.1–39.0) in the IB group, 23.4 months (95%CI 12.1–41.7)in the IIA group, 18.8 months (95%CI 15.1–22.2) in the IIB group, and 17.9 months (95%CI 14.3–21.7) in the stage III group, and 10.0 months (95%CI 6.9–17.4) in the stage IV group (p<0.001). Amongst patients with a pCR, 5/13 (38.5%) developed a recurrence.
Figure 2.
Kaplan Meier survival plots for the three treatment groups from A) date of diagnosis and B) date of operation. SF = surgery first; SMNT = single modality neoadjuvant therapy; TNT = total neoadjuvant therapy.
Variables Associated with Survival
On univariate analysis, performing an R1 resection (HR 1.99, p<0.001) was significantly associated with OS. On multivariable analysis (Table 4), after controlling for vascular resection, positive margin status, age at diagnosis, node positivity on final pathology, and adjuvant chemotherapy as a stratification variable, patients in the SF or SMNT groups had significantly greater mortality than those in the TNT group (HR 1.45 (95% CI 1.02–2.05) and HR 1.55 (95% CI 1.9–2.22), respectively, p=0.046).
Table 4. Multivariable analysis of factors associated with overall mortality.
Adjuvant chemotherapy was included as a stratification variable with 3 levels (Yes, No, Unknown/missing) due to severe violation of the proportional hazards assumption.
| Variable | Hazard Ratio | 95% CI | p-value |
|---|---|---|---|
| Type of treatment | |||
| Total neoadjuvant therapy | Ref. | -- | 0.046 |
| Single modality neoadjuvant therapy | 1.55 | 1.09 – 2.22 | |
| Surgery first | 1.45 | 1.02 – 2.05 | |
| Age | |||
| 1 year difference | 1.00 | 0.99 – 1.02 | 0.49 |
| Vascular resection | |||
| No/unknown | Ref | -- | 0.16 |
| Yes | 1.22 | 0.92 – 1.62 | |
| Margins | |||
| R0 | Ref | -- | <0.001 |
| R1/R2 | 1.88 | 1.43 – 2.48 | |
| Lymph node positivity | |||
| No/Missing | Ref | -- | 0.15 |
| Yes | 1.22 | 0.93 – 1.59 |
DISCUSSION
Herein we have described one of the longest institutional experiences with TNT for PDAC, building upon recently published data to suggest its potential therapeutic role. As compared to surgery first or SMNT, TNT was associated with more frequent tumor downstaging and pathologic complete response, greater rates of R0 resection, and longer OS. On multivariable analysis, receipt of TNT was one of the predictors of improved survival in conjunction with known variables such as age, nodal positivity, and adjuvant chemotherapy. Given the growing interest in utilizing neoadjuvant strategies, even amongst patient who would qualify for up-front surgical resection by historical standards, these data suggest that TNT may provide an opportunity to improve oncologic outcomes and provide additional time to select patients who would be biologically appropriate for surgery.
Our data join that of additional reports of institutional TNT utilization and yield similar improvements in oncologic outcome. Kim et al. reported on a single institutional experience of 541 consecutive PDAC patients from 2009–2019, with 89 (16%) patients having received TNT and single-agent neoadjuvant therapy (similar to our SMNT group) in the remaining 452 (84%)10. Similar to our analysis, it is one of the only other series to compare TNT to regimens with single components (i.e. chemotherapy alone or chemoradiation alone). The patients who received TNT in that series were more likely to have a pCR (8% vs. 4%, p<0.01). Although the OS comparison did not reveal a difference in TNT vs. single-agent neoadjuvant treatment as in our study, the TNT cohort had not yet reached median OS and longer follow up is needed to truly understand if any long-term survival advantage exists. Unlike our study, R0 resection was similar amongst single-agent therapy and TNT (85% vs. 89%, respectively, p=0.44). Importantly, patients who received TNT made it to surgery with similar frequency to those with single-agent neoadjuvant (72% vs. 71%, respectively). Truty et al. also recently published their experience with 194 patients who underwent resection following TNT11. Major pathologic response (defined as College of American Pathology Score 0 or 1) was achieved in 38.7%, of which 10.3% were complete responses. Major pathologic response was associated with improved recurrence free survival and OS in their series. Further, on multivariable analysis extended duration of chemotherapy, major pathologic response, and optimal post-neoadjuvant CA19–9 were the only factors associated with improved survival in their TNT cohort.
Our data add to these by demonstrating a significant survival benefit to TNT compared to SMNT. The mechanism for this is likely multifactorial. We did observe an improvement in R0 resection with TNT, which is not consistent with the series by Kim and colleagues10. It must be acknowledged that our rate of R0 resection in the SMNT group was lower than expected, but that subgroup did contain a slightly higher proportion of locally advanced tumors and the TNT group did undergo vascular resection more frequently. Another mechanism is that TNT provides a longer lead-time prior to surgery, such that patients who may potentially fail surgical therapy by metastasizing early postoperatively do so prior to resection. We can’t specifically comment on that phenomenon in our series as our dataset did not capture all patients diagnosed with PDAC during the study period (i.e. providing the true denominator), and so we do not know the fraction of patients who never made it to surgery. Third, we did observe a higher rate of pCR with TNT. Complete pathologic responses have been shown be associated with improved survival in a multitude of studies11–15 and as such represent a strong surrogate for efficacy of an adjuvant treatment strategy in PDAC. We observed a profoundly long median OS with pCR (100.2 months), which is very encouraging. Given how infrequent pCR occurs, studying this subgroup of patients to understand the underlying biology of these robust responses may reveal novel strategies for treatment. Further, analyses to understand which patients achieve pCR may also aid in selecting which cohorts benefit most from TNT.
What remains to be determined for TNT is the optimal regimen, sequencing, and settings which optimize responses when it is utilized. Our cohort was comprised of many patients who underwent chemoradiation first, followed by consolidative chemotherapy. This is unusual compared to most TNT regimens described in the literature10,11,16,17 and is a reflection of our early adoption of TNT, which occurred in an era when our institution favored neoadjuvant chemoradiation in most BR or LA patients. Extrapolating from the recent OPRA trial in rectal cancer, there may be benefit to applying radiation first and providing consolidative chemotherapy18. Although FOLFIRINOX has demonstrated clear superiority in the postoperative setting since the PRODIGE study19, this has yet to be replicated in the neoadjuvant and TNT settings. Unfortunately, current high-quality data assessing neoadjuvant regimens are heterogeneous, not only in chemotherapy and radiation applied but resectability of primary tumors included. This makes definitive conclusions difficult to draw. The recent Alliance A021501 data provided encouraging neoadjuvant data for BR disease, with a median OS of 31.0 months in the modified FOLFIRNOX arm20. Results of the accruing Alliance A021806 trial (NCT04340141) will certainly bring more definitive evidence to the debate over how beneficial neoadjuvant therapy may be in the setting of resectable disease.
This study has several strengths and limitations. Although patients were not studied on a prospective protocol, there was strict adherence to high-quality vascular imaging for preoperative and post-treatment staging by board-certified abdominal radiologists and treatment decisions were based on consensus from multidisciplinary tumor board at a National Cancer Institute-designated Comprehensive Cancer Center. One limitation of this study is that it only included patients who underwent surgery after successful completion of SMNT or TNT regimens. Thus, patients who were able to complete multimodality treatment likely had favorable tumor biology compared to those who did not make it to surgery. Additionally, chemotherapy regimens were heterogenous in the early years and primarily consisted of gemcitabine-based therapy. We look forward to re-evaluating our long-term survival outcomes with FOLFIRINOX as the backbone of our TNT regimen given its encouraging results in the adjuvant19 and neoadjuvant20 settings. The sequencing of chemotherapy and chemoradiation during TNT was not homogenous as well, and therefore ascertainment of which approach may be more effective requires further investigation.
In conclusion, improved survival was achieved after TNT compared to SF or SMNT amongst surgically resected patients with PDAC. TNT therefore offers favorable short- and long-term outcomes, as well as the benefit of optimally selecting patients for surgery based on fitness for TNT and tumor biology by providing additional time on systemic therapy before surgery is undertaken. There is much work to be done to identify which patients derive the greatest benefit from a TNT approach and which may forego this for either SF or SMNT. Our data add to recent literature and provide compelling evidence that a prospective, multicenter trial design evaluating TNT is needed.
MATERIALS AND METHODS
Study Cohort
The Fox Chase Cancer Center Internal Review Board approved the conduct of this study prior to initiation. Data was acquired via retrospective chart review. All adult (age ≥18) patients who underwent definitive surgical resection for PDAC with or without neoadjuvant therapy from February 1993 to October 2019 at a single NCI-designated institution were included in this study. All patients in our cohort had a primary head, uncinate, or neck mass for which pancreatoduodenectomy (PD) with or without vascular resection was the planned operation for curative intent. Patients with distant metastases were excluded, as were patients with disease progression who did not undergo definitive surgical resection. This yielded a total of 358 eligible patients with PDAC. These patients were subdivided into three main analytic cohorts comprised of SF (N=168), SMNT (N=111), and TNT (N=79) (Figure 1). Patients in the SMNT cohort received either chemotherapy or chemoradiation alone prior to definitive surgical resection. Patients in the TNT cohort received both chemotherapy and chemoradiation prior to surgery. Following surgical resection, they received postoperative chemotherapy at the discretion of the treating medical oncologist. To provide a rough estimate of the denominator (i.e. how many total patients with potentially resectable PDAC were evaluated over the same time period), we queried separate institutional data for the total number of patients with potentially resectable PDAC (stages I-III). From 1993–2019, 918 new patients with stage I-III PDAC were evaluated (34/year).
Figure 1.
Flow diagram illustrating construction of the analytic cohorts. SF = surgery first; SMNT = single modality neoadjuvant treatment; TNT = total neoadjuvant therapy; neoadj = neoadjuvant; CRT = chemoradiotherapy; chemo = chemotherapy.
Workup and Treatment
A pancreas protocol triple-phase computed tomography scan of the abdomen/pelvis (CTAP) and chest computed tomography (CT) were obtained for all patients to determine stage group and resectability, allergies permitting, prior to the initiation of therapy. CA19–9 levels were obtained at diagnosis and at the discretion of the treating team. Review of all diagnostic tissue (including outside materials) to confirm PDAC was performed by the treating institution prior to initiation of treatment. Resectability and the decision to proceed with SF, SMNT, or TNT was determined by consensus decision at an institutional multidisciplinary tumor board prior to the initiation of any treatment. It must be noted that early in the experience with TNT many patients were assigned to SMNT (namely, neoadjuvant chemoradiation) and went on to get TNT instead to optimize tumor response prior to surgery. Systemic chemotherapy consisted of either gemcitabine-based therapy (primarily gemcitabine/nab-paclitaxel) or oxaliplatin-based (primarily FOLFIRINOX [5-FU, oxaliplatin, irinotecan]) at the discretion of the medical oncologist. Changes in chemotherapy due to poor tolerance or progression were made at the discretion of the primary medical oncologist. Chemoradiation consisted of 5.5 weeks of intensity-modulated radiation therapy (IMRT) to a total dose of 50.4 cGy with a radiosensitizer (most commonly gemcitabine or 5-FU) or stereotactic body radiation therapy (SBRT).
Independent variables
Demographic variables included age at diagnosis, gender, race, Eastern Cooperative Oncology Group (ECOG) performance status, and comorbidities (hypertension, hyperlipidemia, type 2 diabetes, coronary artery disease, obesity, smoking status). Clinical stage group was defined based on the AJCC edition at the time of the patient’s initial diagnosis. Tumor resectability classification was based on the NCCN definition9. Chemotherapy regimen was broadly categorized into gemcitabine-based, FOLFORINOX, other 5-FU based, and other. The necessity for vascular resection was collected as a dichotomous variable (yes/no).
Dependent variables
The primary outcome of interest was OS, calculated from the date of diagnosis, and in separate analyses, from the date of surgery, to the date of death (event) or last follow-up visit (censored). Follow-up data for survival calculations was based on information available as of March 29, 2020. Secondary outcomes included final pathologic margin status (R0 vs. R1/R2), final pathologic stage, percentage of tumor fibrosis in the surgical specimen (as communicated by the pathologist’s report), and complete pathologic response (defined as no appreciable residual tumor cells in the resected specimen). Downstaging was defined as a decrease in stage group from the clinical stage assignment to the pathologic stage assignment. Lymph node yield was collected and presented as means.
Statistical Analysis
Cohort characteristics were summarized using counts with percentages or medians with interquartile range (IQR), as appropriate, and compared by treatment group with Chi-square or Fisher’s Exact test for categorical variables and Kruskal-Wallis test for continuous variables. Differences were considered statistically significant when p<0.05. Downstaging was assessed in those with complete data for clinical and pathologic stage. Overall survival (OS) was estimated using Kaplan-Meier methods from the date of histologic diagnosis to the date of death; patients still alive were censored at the date of last follow-up. To reduce immortal survival bias for those receiving neoadjuvant therapy, a secondary analysis of OS was also performed from the date of surgery. The log-rank test was used to compare survival distributions by treatment group. OS differences by treatment were also examined separately, for those who received adjuvant chemotherapy and those who did not, as a way to control for increased survival due to adjuvant treatment. Types of chemotherapy or radiosensitizer agents were compared by treatment group among those who received these regimens using Fisher’s exact tests. Similarly, pathologic resection findings after surgery were compared by treatment group using Chi-square tests. Cox proportional hazards regression was used to examine the association of treatment group with time from surgery to death from all causes with adjustment for covariates, including age at diagnosis, gender, race, clinical stage, resectability classification, vascular involvement, surgical margin status, lymph node positivity, and adjuvant chemotherapy treatment; adjuvant chemotherapy was included as a stratification variable due to a severe violation of the proportional hazards assumption. Age was included in the multivariable model a priori, and other covariates were included if the univariate association with overall mortality was significant at p<0.10. Statistical analysis was performed using SAS software (version 9.4, Cary NC).
Supplementary Material
Supplementary Figure 1. Distribution of patients included in the analysis of the study period (1993–2019) by treatment strategy received. SF = surgery first; SMNT = single modality neoadjuvant treatment; TNT = total neoadjuvant therapy.
Supplementary Figure 2. Recurrence free survival plots for the three treatment groups. SF = surgery first; SMNT = single modality neoadjuvant therapy; TNT = total neoadjuvant therapy.
SYNOPSIS.
Total neoadjuvant therapy for pancreatic cancer is associated with more frequent complete pathologic responses, a higher rate of margin negative resection, and prolonged overall survival as compared to surgery first or single-modality neoadjuvant approaches. Patients with pathologic complete response had median overall survival of >100 months and may be the best chance for cure in this aggressive disease.
ACKNOWLEDGEMENTS
We are grateful for the data curation performed by Dr. Tamsin Board and Dr. Lori Rhoades (Temple University Hospital General Surgery, Philadelphia, PA).
Funding:
Research reported in this publication was supported by the National Cancer Institute of the National Institutes of Health under Award Number P30CA006927. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
ABBREVIATIONS LIST
- PDAC
pancreatic ductal adenocarcinoma
- OS
overall survival
- TNT
total neoadjuvant therapy
- BR
borderline resectable
- LA
locally advanced
- SF
surgery first
- SMNT
single modality neoadjuvant therapy
- CAD
coronary artery disease
- HLD
hyperlipidemia
- IQR
interquartile range
- HR
hazard ratio
- CI
confidence interval
- CR
complete response
Footnotes
Meeting Information: Presented at the 2020 Society of Surgical Oncology Meeting
Conflicts of Interest: The authors have no disclosures or conflicts of interest.
Data Sharing:
The data that support the findings of this study are available from the corresponding author upon reasonable request.
REFERENCES
- 1.Siegel RL, Miller KD, Fuchs HE, et al. Cancer Statistics, 2021. CA Cancer J Clin. 2021;71(1):7–33. doi: 10.3322/caac.21654 [DOI] [PubMed] [Google Scholar]
- 2.Ferrone CR, Brennan MF, Gonen M, et al. Pancreatic adenocarcinoma: the actual 5-year survivors. J Gastrointest Surg 2008;12(4):701–706. doi: 10.1007/s11605-007-0384-8 [DOI] [PubMed] [Google Scholar]
- 3.Ferrone CR, Pieretti-Vanmarcke R, et al. Pancreatic ductal adenocarcinoma: long-term survival does not equal cure. Surgery. 2012;152(3 Suppl 1):S43–9. doi: 10.1016/j.surg.2012.05.020 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Bakens MJ, van der Geest LG, van Putten M, et al. The use of adjuvant chemotherapy for pancreatic cancer varies widely between hospitals: a nationwide population-based analysis. Cancer Med. 2016;5(10):2825–2831. doi: 10.1002/cam4.921 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Versteijne E, Suker M, Groothuis K, et al. Preoperative Chemoradiotherapy Versus Immediate Surgery for Resectable and Borderline Resectable Pancreatic Cancer: Results of the Dutch Randomized Phase III PREOPANC Trial. J Clin Oncol. 2020;38(16):1763–1773. doi: 10.1200/JCO.19.02274 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Abrams RA, Lowy AM, O’Reilly EM, et al. Combined modality treatment of resectable and borderline resectable pancreas cancer: expert consensus statement. Ann Surg Oncol. 2009;16(7):1751–1756. doi: 10.1245/s10434-009-0413-9 [DOI] [PubMed] [Google Scholar]
- 7.Katz MHG, Marsh R, Herman JM, et al. Borderline resectable pancreatic cancer: need for standardization and methods for optimal clinical trial design. Ann Surg Oncol. 2013;20(8):2787–2795. doi: 10.1245/s10434-013-2886-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Sohal DPS, Duong M, Ahmad SA, et al. Efficacy of Perioperative Chemotherapy for Resectable Pancreatic Adenocarcinoma: A Phase 2 Randomized Clinical Trial. JAMA Oncol. 2021;7(3):421–427. doi: 10.1001/jamaoncol.2020.7328 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Tempero, Margaret A, Malafa, et al. NCCN Guidelines for Pancreatic Adenocarcinoma Version 1.2020. 2020.
- 10.Kim RY, Christians KK, Aldakkak M, et al. Total Neoadjuvant Therapy for Operable Pancreatic Cancer. Ann Surg Oncol. 2021;28(4):2246–2256. doi: 10.1245/s10434-020-09149-3 [DOI] [PubMed] [Google Scholar]
- 11.Truty MJ, Kendrick ML, Nagorney DM, et al. Factors Predicting Response, Perioperative Outcomes, and Survival Following Total Neoadjuvant Therapy for Borderline/Locally Advanced Pancreatic Cancer. Ann Surg. 2021;273(2):341–349. doi: 10.1097/SLA.0000000000003284 [DOI] [PubMed] [Google Scholar]
- 12.Macedo FI, Ryon E, Maithel SK, et al. Survival Outcomes Associated With Clinical and Pathological Response Following Neoadjuvant FOLFIRINOX or Gemcitabine/Nab-Paclitaxel Chemotherapy in Resected Pancreatic Cancer. Ann Surg. 2019;270(3):400–413. doi: 10.1097/SLA.0000000000003468 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Sell NM, Lee GC, Fernández-Del Castillo C, et al. Evaluation of Pathologic Response on Overall Survival After Neoadjuvant Therapy in Pancreatic Ductal Adenocarcinoma. Pancreas. 2020;49(7):897–903. doi: 10.1097/MPA.0000000000001590 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Cloyd JM, Ejaz A, Shen C, et al. Pathologic complete response following neoadjuvant therapy for pancreatic ductal adenocarcinoma: defining the incidence, predictors, and outcomes. HPB. 2020;22(11):1569–1576. doi: 10.1016/j.hpb.2020.01.013 [DOI] [PubMed] [Google Scholar]
- 15.He J, Blair AB, Groot VP, et al. Is a Pathological Complete Response Following Neoadjuvant Chemoradiation Associated With Prolonged Survival in Patients With Pancreatic Cancer? Ann Surg. 2018;268(1):1–8. doi: 10.1097/SLA.0000000000002672 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Murphy JE, Wo JY, Ryan DP, et al. Total Neoadjuvant Therapy With FOLFIRINOX Followed by Individualized Chemoradiotherapy for Borderline Resectable Pancreatic Adenocarcinoma: A Phase 2 Clinical Trial. JAMA Oncol. 2018;4(7):963–969. doi: 10.1001/jamaoncol.2018.0329 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Katz MHG, Shi Q, Ahmad SA, et al. Preoperative Modified FOLFIRINOX Treatment Followed by Capecitabine-Based Chemoradiation for Borderline Resectable Pancreatic Cancer: Alliance for Clinical Trials in Oncology Trial A021101. JAMA Surg. 2016;151(8):e161137. doi: 10.1001/jamasurg.2016.1137 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Garcia-Aguilar J, Patil S, Kim JK, et al. Preliminary results of the organ preservation of rectal adenocarcinoma (OPRA) trial. J Clin Oncol. 2020;38(15_suppl):4008. doi: 10.1200/JCO.2020.38.15_suppl.4008 [DOI] [Google Scholar]
- 19.Conroy T, Hammel P, Hebbar M, et al. FOLFIRINOX or Gemcitabine as Adjuvant Therapy for Pancreatic Cancer. N Engl J Med. 2018;379(25):2395–2406. doi: 10.1056/NEJMoa1809775 [DOI] [PubMed] [Google Scholar]
- 20.Katz MHG, Shi Q, Meyers JP, et al. Alliance A021501: Preoperative mFOLFIRINOX or mFOLFIRINOX plus hypofractionated radiation therapy (RT) for borderline resectable (BR) adenocarcinoma of the pancreas. J Clin Oncol. 2021;39(3_suppl):377. doi: 10.1200/JCO.2021.39.3_suppl.377 [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
Supplementary Figure 1. Distribution of patients included in the analysis of the study period (1993–2019) by treatment strategy received. SF = surgery first; SMNT = single modality neoadjuvant treatment; TNT = total neoadjuvant therapy.
Supplementary Figure 2. Recurrence free survival plots for the three treatment groups. SF = surgery first; SMNT = single modality neoadjuvant therapy; TNT = total neoadjuvant therapy.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.