Outcomes in Locally Advanced Pancreatic Cancer After Induction Ablative Radiation Therapy and Resection

Abstract

BACKGROUND:

Ablative dose radiotherapy (A-RT) may result in durable local control and encouraging survival in patients with locally advanced pancreatic cancer (LAPC). A subset of patients with LAPC are eligible for exploration after completion of induction chemotherapy and A-RT. Outcomes in this subset of patients are yet to be described.

METHODS:

This was a single-institution retrospective analysis of patients with LAPC treated with induction chemotherapy, A-RT (≥98Gy biologically effective dose using 15–25 fractions in 3–4.5 Gy/fraction) and subsequently underwent surgical exploration.

RESULTS:

Over a 6-year period, 34 patients with LAPC underwent exploration after induction chemotherapy and A-RT. Chemotherapy was given prior to A-RT in all patients, with the majority receiving FOLFIRINOX (94%). The median time to exploration after completion of A-RT was 75 days. Pancreaticoduodenectomy was the most frequent procedure (n=18), followed by distal pancreatectomy (n=7) and resection aborted in 9 patients (26%) following discovery of distant (n=6) or locally unresectable disease (n=3). Vascular resection or divestment was required in 56% of patients. There were no post operative pancreatic fistulae, however clinically significant ascites was observed in 36% of resected patients. A major pathologic response was observed in 17% of resected specimens upon final pathologic review. Post operative mortality within 90 days occurred in 2 patients (5.9%). Median overall survival for the entire cohort was 31 months from date of diagnosis and 24 months from completion of A-RT.

CONCLUSIONS:

Resection of LAPC is feasible in a select cohort of patients after A-RT with encouraging 2-year overall survival.

INTRODUCTION:

Pancreas cancer is a leading cause of cancer-related mortality with over 50,000 projected annual deaths in the United States for 2024¹. Effective treatment options remain limited, even for patients with localized disease. While surgical resection of the primary tumor combined with multimodality therapy improves survival, when major vascular structures are involved a margin-negative resection with an upfront surgical approach is unlikely^2,3. Chemoradiotherapy following induction systemic therapy has an increased role in this cohort of pancreatic cancer patients⁴.

Radiation has been widely utilized in both the induction setting and for destination locoregional control or palliation in unresectable, locally advanced pancreatic cancer (LAPC). Nevertheless, there has not been definitive data demonstrating a survival benefit with conventionally fractionated radiation therapy (CRT) nor hypofractionated stereotactic body radiation therapy (SBRT) compared to systemic chemotherapy alone^5–10. Historically, these studies have used a biologically effective dose (assuming standard alpha/beta ratio for tumor of 10, BED₁₀) of 59.47–64.8Gy. Prior experiences of ablative dose radiation treatments for hepatocellular cancer and non-small cell lung cancer suggest that a BED₁₀ of 100 Gy may be necessary to maximize the therapeutic response and obtain durable local control^11,12. At Memorial Sloan Kettering Cancer Center (MSKCC), high-dose ablative radiation (A-RT) of ≥98Gy BED₁₀ is utilized for the treatment of LAPC. A-RT in conjunction with respiratory gating, Conebeam CT (CBCT) image guidance, and adaptive re-planning to protect luminal organs can be delivered safely to patients with promising locoregional disease control^13–15. A subset of these patients with persistent localized disease may be eligible for an attempt at resection following A-RT and have undergone pancreatectomy, initially on a prospective single institutional protocol (NCT03523312) that provided an initial safety signal in 13 participants. However, the full impact of A-RT on the safety of subsequent operative exploration is not yet defined.

The main objectives of this cohort study are to evaluate the post-treatment survival and safety of pancreatectomy following A-RT treatment for locally advanced pancreatic disease. Additionally, we report the clinicopathological variables and pathologic response of resected pancreatic specimen following neoadjuvant chemotherapy and A-RT.

METHODS:

Patient Cohort

This study is a retrospective analysis of a prospectively maintained institutional database of pancreatectomy patients and was approved by the institutional review board. Patients who received A-RT to the pancreas and subsequent exploration at MSKCC between 2018 and 2023 were included in this study. All patients received neoadjuvant chemotherapy in addition to A-RT. Chemotherapy regimens were selected at the discretion of the treating medical oncologist. All patients included in this study underwent review at multidisciplinary tumor board. Patients who were determined to have locally advanced unresectable tumors consistent with Americas Hepato-Pancreato-Biliary Association (AHPBA)/Society of Surgical Oncology (SSO)/Society for Surgery of the Alimentary tract (SSAT) consensus definitions were recommended induction A-RT.

Demographics and Clinicopathological Characteristics

Demographic and clinical variables were extracted from the electronic medical record (EMR) and a prospective institutional database. Outcomes of clinical interest include overall survival (OS), pathologic response and post operative complications. Follow-up data were retrieved from medical records and an institutional database. Date of diagnosis was defined as the date of pathologic confirmation through tissue biopsy. The date of death was obtained from the EMR or local obituary. OS was calculated from both the date of diagnosis and the completion of radiation to the date of death from any cause, or censored at last follow up. Recurrence Free Survival (RFS) was calculated from date of surgery to time of documented recurrence as defined by surveillance imaging. Biopsy was not required for recurrence confirmation. International Study Group of Pancreatic Surgery (ISGPS) guidelines for differentiating main pancreatic duct diameter were utilized based on most recent preoperative imaging dichotomizing between ≤3mm and >3mm at area of greatest duct dimension. R0 resection was defined by the absence of malignant cells within 1mm of the surgical margin. The radial (anterior/posterior), pancreatic neck, superior mesenteric artery vascular groove and bile duct margins are routinely assessed at our institution. All patients in the study received neoadjuvant chemotherapy in addition to A-RT. Chemotherapy regimens were selected at the discretion of the treating medical oncologist. Following induction chemotherapy, imaging was again reviewed at a multidisciplinary tumor board. Patients deemed appropriate for resection were explored, while those with persistent locally advanced disease were referred for A-RT. Timing of surgery after A-RT was determined on an individual basis considering patient performance status, CA19–9 and disease extent, stability or improvement via imaging. All patients received diagnostic laparoscopy prior to laparotomy, aborting surgery should occult metastatic disease be confirmed. When intraoperative frozen sections were negative and periarterial divestment was feasible, an arterial sparing surgical technique was performed to extirpate the tumor without en-bloc arterial resection as previously described¹⁶. Resection of hepatic artery with formal reconstruction with autologous saphenous vein was performed in scenarios that were not amenable to periarterial divestment. Superior mesenteric artery resection was not conducted in any patients in this cohort. Ligation of a replaced hepatic artery or splenic artery was not considered a vascular resection. The College of American Pathologist (CAP) scoring criteria was utilized to grade the degree of treatment response on final pathologic assessment of resected tumor specimen with a grade 0–1 considered a major pathologic response¹⁷.

Post operative complications were defined by the Clavien-Dindo clinical scale and were documented at 30-days from the date of operation¹⁸. Additional complications up to 6 months from operation were also abstracted from medical records. Interventions for ascites included image guided paracentesis and/or placement of an indwelling drainage catheter. Post operative pancreatic fistula (POPF) included ISGPS grade B/C only¹⁹.

Ablative Radiation Technique

The background, rationale, and techniques for the planning and delivery of A-RT in the treatment of locally advanced pancreatic disease at MSKCC have been described in previous reports^13,15. In brief, one or two ablative hypofractionated regimens, 75Gy in 25 fractions (BED10 = 97.5 Gy) for tumors less than 1cm from stomach or intestines, or 67.5 Gy in 15 fractions (BED10 = 97.88 Gy) for tumors of 1cm or further, are delivered with daily CBCT image guidance and respiratory motion management. Patients with >5cm length of luminal GI abutment by the pancreatic tumor are not eligible for A-RT modality. The BED is calculated using the linear quadratic formula and is based upon the dose per fraction, total dose, and the response characteristics of the tumor to radiation (defined by the α/β ratio; BED10 assumes an α/β ratio of 10). Patients enrolled in the MAIBE (Maximal Ablative Irradiation Because of Encasement) trial (NCT03523312) were included in this analysis.

Statistical Analysis

Statistical analyses were performed using Stata/MP 14.0 (Stata Corp, College station, TX). Categorical variables were expressed as percentages and were compared using a χ² or a Fisher exact test. Continuous variables were presented as median and interquartile range (IQR),and were compared using a Kruskal-Wallis test. Survival analyses were performed using Kaplan-Meier survival estimates and calculated from both the time of diagnosis and the end of A-RT. A two-tailed P value < 0.05 was considered statistically significant.

RESULTS:

Patient demographics and survival

From 2018 to 2023, a total of 34 patients underwent surgical exploration following induction chemotherapy and A-RT at MSKCC and were included for analysis. Preoperative patient characteristics of the cohort can be found in Table 1. Extensive vascular involvement was appreciated on imaging with arterial abutment or encasement in 91% and venous encasement or occlusion in 38%. Figure 1 displays representative cross-sectional imaging of vascular involvement in a patient resected in this study. A median of 8 cycles of chemotherapy were given. CA19–9 peak at diagnosis was 232 (IQR: 32.5, 692) which improved or was stable in all patients prior to exploration with a post A-RT median of 47 (IQR: 21, 91). The median time to exploration after completion of A-RT was 75 days ranging from 39 to 488 days. Median overall survival from the date of diagnosis for the entire cohort was 31.2 months (95% CI: 20.0 to 44.7) (Figure 2). The median overall survival calculated from the date of A-RT completion was 23.5 months (95% CI: 13.3 to 34.8). The 2-year OS from diagnosis was 68% (95% CI: 0.28–0.64) for the whole cohort, and 74% (95% CI: 0.48–0.81) for those that underwent resection. Recurrence occurred in 14 patients with a median RFS from time of surgery of 11.3 months (95% CI: 5.4 to 15.3) in the resected cohort. Among patients that recurred, the most common disease site was carcinomatosis or multiple sites (n=7, 50%), followed by liver only (n=3, 21%), local (n=3, 21%) and lung only (n=1, 7%). Among patients explored, 9 cases were aborted with 6 due to the discovery of distant metastatic disease and 3 due to unresectable local vascular involvement (proper hepatic artery n=1, SMV n=1, SMA n=1). Demographic, clinical, and pathological information for the 25 patients who underwent curative intent extirpation are described in Table 2. The majority (64%) of resected patients were male, with a median age of 66 years (IQR 60–73) at the time of surgery. Induction chemotherapy was used in all patients and regimens included 5-fluorouracil, irinotecan and oxaliplatin (FOLFIRINOX) (n=22), gemcitabine and nab-paclitaxel (n=1) or multiple therapies (n=2).

Table 1.

Patient and tumor characteristics of the treatment cohort

Variable	n=34
Age, median years [IQR]	68	[61, 73]
Male, n (%)	19	56%
Neoadjuvant regimen
FOLFIRINOX	30	88%
Gemcitabine and Nab-Paclitaxel	1	4%
Multiple/Other	3	8%
Cycles of chemotherapy, median [IQR]	8	[7, 9]
ECOG
0–1	29	94%
2	2	6%
Pre-ablative venous* involvement
None	4	12%
Abutment <180	17	50%
Encasement	10	29%
Occlusion	3	9%
Pre-ablative arterial# involvement
None	3	9%
Abutment <180	16	47%
Encasement	15	44%
Post-ablative venous* involvement
None	5	15%
Abutment <180	19	56%
Encasement	8	24%
Occlusion	2	6%
Post-ablative arterial# involvement
None	4	12%
Abutment <180	14	41%
Encasement	16	47%
Pancreatic fistula grade B, C	0	0%

Figure 1.

Representative cross-sectional imaging of locally advanced pancreatic cancer with long segment vascular involvement in a patient resected in this study following ablative radiation therapy.

Figure 2.

(A) Kaplan Meier overall survival estimates of patients with pancreatic cancer that underwent exploration following induction A-RT from the date of diagnosis. (B) Kaplan Meier overall survival estimates of patients with pancreatic cancer that underwent exploration following induction A-RT from date of A-RT completion.

Table 2.

Clinicopathologic characteristics resected patients

Variable	n=25
Age, median years [IQR]	66	[60, 73]
Male, n (%)	16	64%
Neoadjuvant regimen
FOLFIRINOX	22	88%
Gemcitabine and Nab-Paclitaxel	1	4%
Multiple/Other	2	8%
Prior exploration	3	12%
Surgery type
Pancreaticoduodenectomy	18	72%
Distal pancreatectomy	7	28%
Vascular resection	14	56%
Venous resection	10	40%
Arterial resection or divestment	9	36%
Both arterial and venous	5	20%
ISGPS duct size
≤3mm	4	22%
>3mm	14	78%
T stage
ypT1	4	18%
ypT2	13	59%
ypT3	3	14%
ypT4	2	9%
Nodal stage
N0	17	71%
N1	6	25%
N2	1	4%
Perineural invasion	15	63%
Lymphovascular invasion	13	57%
R0 resection	17	71%
CAP Regression score
0–1	4	17%
2	11	46%
3	9	37%
Pancreatic fistula grade B, C	0	0%

Clinicopathologic characteristics

Patients underwent pancreaticoduodenectomy most frequently (72%) followed by distal pancreatectomy (28%). An arterial and/or venous resection was required in 14 (56%) resections (Table 3). Among these, a venous resection was most common, including either the portal vein (n=9) or superior mesenteric vein (n=1). Hepatic artery resection (n=2) or periarterial divestment (n=7) were also frequently performed. Venous reconstruction was most frequently addressed with an end-to-end approach (n=9), followed by autologous jugular vein interposition (n=1), and cryopreserved saphenous vein allograft (n=1). Autologous saphenous vein bypass was performed for hepatic artery reconstructions (n=2).

Table 3.

Vascular intervention details during pancreatectomy

Operative intervention for PV/SMV	n (10)
End-to-end primary reconstruction	6
Saphenous cryograft	1
Autologous graft (IJV)	1
Patch	1
Other*	1
Operative intervention for SMA/CHA	n (9)
SMA +/− CHA divestment	7
CHA resection, autologous reconstruction (SV)	2

CHA= common hepatic artery, IJV= internal jugular vein, PV= portal vein, SMA= superior mesenteric artery, SMV= superior mesenteric vein, SV= saphenous vein

^*(SMV resection with maintenance of PV collaterals)

On final pathologic review of the resected specimen, nodal metastases were identified in 29% of patients (n=7), of whom one was classified as N2 disease (Table 2). Evidence of perineural and lymphovascular invasion were reported in 63% and 57% respectively (n=15, n=13). An R0 resection was achieved in the majority of patients (n=17, 71%). When positive, the uncinate/SMA vascular groove was the most common site of persistent microscopic disease (n=4), followed by the distal pancreatic neck (n=2), PV (n=1) and radial margin (n=1). Of note, an arterial divestment was performed in 3 of 4 positive vascular groove cases. CAP Regression scoring of the resected specimen revealed a major pathologic response with no cancer cells or only small groups of single cells (CAP score 0–1) in 17% (n=4), one of which was a pathologic complete response. A partial response (CAP 2) was observed in 46% (n=11), while the remainder had extensive residual cancer (CAP 3) (n=9).

Post operative complications

Post operative complications are documented in Table 4. Any perioperative complication within 30 days of surgical resection was experienced in 44% of resected patients. Complications requiring surgical, endoscopic or radiological intervention (Clavien-Dindo ≥3a¹⁸) were observed in 36% of resected patients. Re-exploration for post operative bleeding was required in one patient with the identification of nonspecific retroperitoneal oozing with intact, patent hepatic artery reconstruction. Two mortalities occurred within 90-days of operation: one secondary to myocardial infarction the other due to early PV thrombosis and acute post operative liver failure in the setting of complete CHA divestment and long segment PV resection. There were no clinically relevant POPF (ISGPS grade B,C¹⁹) observed. Failure to thrive requiring feeding tube placement was documented in 20% of patients (n=5). While typically multifactorial, persistent diarrhea after SMA divestment was often a driving factor. Clinically significant ascites requiring intervention beyond diuretics with either paracentesis or drain was relatively common, seen in 16% of patients within 30-days of extirpation, and 36% of patients within 6 months of resection. Chylous ascites was appreciated in 1 patient. The median time to intervention for post operative ascites was 58 days from date of operation (range of 8–162 days). An associated partial venous thrombus was appreciated in 3 of the 9 cases of clinically significant ascites.

Table 4.

Complication details

Variable	n=25
LOS, median [IQR]	6	[6,9]
Any 30-day complication	11	44%
30-day complications by highest score
5*	2	8%
4	1	4%
3b	1	4%
3a	5	20%
2	2	8%
Pancreatic fistula grade B,C	0	0%
Biliary anastomotic leak	1	4%
Respiratory failure	1	4%
Reoperation for bleeding	1	4%
Endoscopy or Feeding tube placement	3	12%
Vascular thrombus	3	12%
Infection requiring antibiotics	5	20%
Diarrhea requiring hydration	2	8%
30-day ascites requiring paracentesis or drain	4	16%
Additional complications <6 months
Feeding tube placement	2	8%
Ascites requiring paracentesis or drain	5	20%
Diarrhea requiring hydration	3	12%
Late PV thrombus	1	4%
Total post operative ascites requiring procedure	9	36%
Total post operative diarrhea requiring hydration	5	20%
Total feeding tube placement	4	16%

^*(POD0 myocardial infarction, POD9 liver failure)

DISCUSSION:

In the current study, we report outcomes on a series of 34 patients with LAPC treated with induction chemotherapy, high dose A-RT and subsequent exploration. Complete resection after A-RT is feasible and can be performed safely with encouraging survival outcomes. Median OS for this cohort with advanced local disease was encouraging at 24 months, with a 2-year OS of 47%. The complication profile after A-RT appears to be different than typically described in the upfront resectable cohort of pancreatic disease without any observed POPF, but with a significant rate of clinically significant ascites (36%) requiring paracentesis or drain. A major pathologic response was observed in 17% of patients’ tumors, however viable cancer persisted in most resected specimens suggesting resection should still be pursued in appropriately selected patients.

Resection of LAPC is technically challenging with a subsequent higher risk of complications than the upfront resectable cohort based on historical data. Rates of severe post operative complications requiring invasive intervention within 30-days in this cohort was 36%, a rate consistent with other published reports of surgical programs performing high volumes of LAPC surgery ranging from 31–54%^2,20–22. POPF is one of the most dreaded complications and defining management challenges after pancreatic surgery and have been reported to occur in 12–15% of pancreaticoduodenectomy and 18–20% of distal pancreatectomies^23,24. No POPF were observed in the present post A-RT treated cohort. Other groups have similarly reported reduced rates of POPF development following induction chemoradiation treatment, hypothesizing an effect of radiation induced fibrosis, as compared to an untreated, soft, friable pancreatic gland^25–28. It has been hypothesized that preoperative RT may increase the risk of postpancreatectomy hemorrhage (PPH) or ischemic complications²⁹, however, a signal of increased bleeding complications was not observed in this cohort. In fact, there were no arterial pseudoaneurysms in this series, possibly influenced by the lack of POPF. Other groups similarly failed to find a significant increase in PPH for PDAC patients having received induction radiation as compared to patients who did not receive radiation^25,30,31. One patient in the present series, required post operative exploration for hemorrhage, however, was ultimately found to have nonspecific retroperitoneal bleeding not associated with their vascular reconstruction. While not necessarily a direct consequence of A-RT, this highlights that even in experienced hands at high volume centers, these cases are incredibly complex and should not be taken lightly.

Ascites was a common complication reported in this study. The presence of a PV/SMV thrombus may play a role in postoperative ascites and was observed in 33% (n=3) of these patients. Of note, 2 of these 3 patients with thrombus and concomitant ascites had undergone PV reconstruction, of whom one had end-to-end and the other saphenous cryograft reconstruction. A technical component of vascular reconstruction or choice of graft could play a role in thrombus and ascites development³², although not yet verified by prospective data, it is conceivable the fibrotic changes induced by A-RT may be contributing to alterations in perivascular lymphatic flow, local venous stasis or injury attributing to the increased rates of post operative ascites unique to this series. The potential risk of clinically relevant ascites, as well as an overall higher risk of morbidity and mortality, should be included in the preoperative patient discussion and informed consent process.

The median overall survival for this highly selected study population with LAPC was 31 months from diagnosis and 24 months from A-RT. Retrospective series have reported a similar median overall survival of ~30–35 months for well selected patients with LAPC receiving neoadjuvant treatment and subsequent resection^2,3,33,34. These rates are encouraging in the context of other contemporary reports of chemoradiation followed by resection including 17 months in the FOLFIRINOX+RT arm of ALLIANCE A021501 and 16 months in the gemcitabine and CRT arm of PREOPANC^10,35. Formal hypothesis testing to potentially identify further survival associations was not possible for this current cohort due to the small overall size.

A major pathologic response was observed in 17% of resected specimen following induction chemotherapy and A-RT with one pathologic complete response (pCR). Histopathological assessment of residual tumor burden after resection following induction therapy has been repeatedly shown to be an excellent prognostic factor predictive of long-term survival^36–39. Previously published experience from our group suggests that A-RT may offer excellent local control even without resection^13,14, however despite encouraging rates of major pathologic response, histopathologic review suggests that viable cancer remains in over 80% of specimen. This may be due to the timing of tissue sampling relative to the end of radiotherapy. Maximal radiation response may take months to manifest. In other cancers such as rectal cancer, the data suggests that a longer interval between chemoradiation therapy and surgery may lead to an increase in the pathologic complete response rate with ongoing response to radiation extending out 5 months⁴⁰. Accordingly, the duration of the treatment effect of A-RT is yet to be determined. It is possible that radiotherapy’s effect extends beyond the median 75 days prior to exploration observed in this study. Nevertheless, the presence of viable cancer in the majority of resected specimen post A-RT treatment reinforces that radiation alone is not curing these patients and thus surgical extirpation in a carefully selected cohort still has a role.

As multimodal therapy and radiation techniques continue to advance and limits are pushed with surgical technique, the concept of pancreatic cancer resectability is continuing to evolve. The classification of “locally advanced” disease is an immensely broad definition with many limitations leading to vast differences in patient selection for exploration, even amongst high volume institutions^3,41,42. Response to therapy and true vascular involvement is difficult to discern on imaging following chemoradiation adding an additional barrier to ideal patient selection^3,43. The challenges in determining optimal LAPC candidates for exploration persist following A-RT, nevertheless, within this cohort of advanced disease, there appears to be a localized phenotype without identifiable distant progression. The true value of surgery in these heavily selected patients is a fundamental question that remains unanswered. Although limited by selection bias, many groups have shown a similar robust median and 5yr OS when LAPC is successfully resected^2,44. Future molecular or immune profiling may be the key to helping predict patients in whom disease remains controlled with A-RT alone and surgery avoided. Nevertheless, the current paradigm at MSKCC is to offer exploration following induction chemotherapy and A-RT in healthy, fit patients, with stable localized-only disease, without superior mesenteric artery encasement and an informed preoperative understanding of the increased surgical complexity and potential post operative morbidity.

This study has limitations worthy of discussion. First, due to the nature of data abstraction from a prospective surgical database, the true denominator of patients with LAPC treated with chemotherapy and A-RT but were not explored is unclear. Completion and publication of two phase II trials investigating the role of induction chemotherapy and A-RT with intent for post treatment exploration may provide more clarity to this fundamental question (NCT03523312, NCT05851924). Although leveraging a prospectively collected database from a large tertiary referral center, the survival and post operative complication outcomes were collated in a retrospective nature permitting the associated biases of retrospective work. The sample size of the cohort of interest is small, and multivariable analyses to further ascertain variables associated with survival was not performed due to lack of power. For example, due to small overall number, no particular clinicopathologic differences were identified among patients who had a major pathologic response and those who did not. Better identification of which patients are more likely to respond and who might benefit most from surgery remains an unanswered question for future study. To the authors’ knowledge, this is the largest series of patients with surgical exploration following induction chemoradiotherapy with high dose A-RT. Lastly, the wide translatability of this study should be pursued cautiously: the techniques of achieving a BED₁₀ of nearly 100 Gy with A-RT as well as conducting the complex operations for LAPC mandates experienced radiation oncologists and surgeons, as well as substantial institutional resources to support post-operative recovery in order to optimize patient outcomes.

In conclusion, this single institution, cohort study seeks to define the postoperative outcomes and complication profile of surgical exploration and resection following high dose A-RT. Pancreatectomy is feasible in carefully selected patients with advanced pancreatic disease following induction chemotherapy and A-RT with encouraging overall survival outcomes. Viable cancer persists in the majority of resected specimen suggesting pancreatectomy continues to have a role in appropriately selected patients. This study helps to provide additional rationale for future prospective studies investigating surgery following high dose A-RT in patients with advanced pancreatic disease.

SYNOPSIS:

This retrospective series investigates patients with unresectable locally advanced pancreatic cancer who underwent exploration after induction chemotherapy and high ablative dose radiation therapy (A-RT). Complete resection after A-RT is feasible and safe with a median overall survival of 31 months.