Neoadjuvant SBRT and intraoperative electron radiotherapy in pancreatic cancer resection Authors: Gianella Cornejo, Alon Pikarsky, Ayala Hubert, Marc Wygoda, Antoni Skripai, Aron Popovtzer, Jon Feldman, Yair Hillman, Liat Appelbaum, Mark Temper, Abed Khalaileh, Ashraf Imam, Gideon Zamir, Philip Blumenfeld Published: 2025 ABSTRACT Background/objectivesPancreatic cancer surgery frequently results in positive margins and local recurrence despite multimodal treatment. This study evaluated whether combining neoadjuvant stereotactic body radiotherapy (SBRT) with intraoperative electron radiotherapy (IOeRT) during resection could improve local control and surgical outcomes.MethodsA retrospective analysis was performed on 15 patients with resectable or borderline resectable pancreatic adenocarcinoma treated between 2021 and 2023. All patients received image-guided, motion-managed SBRT (35–40 Gy/5 fractions to PTV_high; 25 Gy/5 fractions to PTV_low) followed by surgical resection and IOeRT (median 10 Gy; 12 Gy when margins were at risk). Toxicities were graded by CTCAE v5.0 and postoperative complications by Clavien–Dindo criteria. Follow-up included imaging and CA 19–9 every 3 months. Survival was estimated using Kaplan–Meier analysis.ResultsMean patient age was 66 years; 60% had tumors in the pancreatic body and 40% in the head. Two-thirds were borderline resectable and received neoadjuvant chemotherapy. Margin-negative resection was achieved in 86.7%, including two complete pathologic responses in BRCA2-mutated tumors. Median overall and progression-free survival were 30 and 16 months, respectively. One patient (6.7%) developed isolated local recurrence, while distant metastases occurred in over half. Toxicities were mainly grade 1–2 fatigue, nausea, or pain; surgical complications were grade 1–2 in 53%, grade 3 in 7%, and grade 5 in 7%.ConclusionsNeoadjuvant SBRT with IOeRT during pancreatic cancer resection is feasible, achieves high rates of negative margins, and provides promising local control. Distant progression remains the dominant mode of failure. Introduction Pancreatic ductal adenocarcinoma (PDAC) remains one of the deadliest solid tumors, with a 5-year survival rate of less than 12% despite advancements in systemic and local therapies (Park et al. 2021; Mizrahi et al. 2020). While complete surgical resection offers the only potential cure, fewer than 20% of patients present with resectable disease at diagnosis (Słodkowski et al. 2023). Even among those undergoing curative-intent resection, recurrence is common, and local failure continues to contribute significantly to morbidity and mortality (Groot et al. 2018). Multimodality treatment has therefore become essential, especially for patients with borderline resectable or locally advanced tumors. Neoadjuvant therapy, particularly chemotherapy followed by radiation, has shown promising improvements in surgical and survival outcomes. The PREOPANC randomized trial demonstrated improved overall survival with preoperative chemoradiotherapy compared to upfront surgery (Versteijne et al. 2022). However, 28% of patients still had margin-positive resections, underscoring the limited local efficacy of conventional radiation (36 Gy in 15 fractions; Biologic Effective Dose (BED) ~ 45 Gy) (Versteijne et al. 2022). Several retrospective and prospective studies now suggest that dose escalation to a BED > 70–100 Gy is necessary for durable local control in PDAC (Zaorsky et al. 2019; Krishnan et al. 2016; Balaban et al. 2016). While SBRT has enabled high-dose, conformal treatment over short courses, the proximity of gastrointestinal organs at risk restricts further dose escalation using CT-based external beam techniques alone (Balaban et al. 2016). Intra-operative electron radiation therapy (IOeRT) offers a promising solution to this challenge by enabling focal dose escalation directly to the tumor bed during surgery. Multiple retrospective series and task force guidelines support IOeRT for patients with high-risk resection margins after external beam radiation therapy (EBRT) (Harrison et al. 2020; Calvo et al. 2020). However, no prior study has explored the integration of neoadjuvant SBRT with IOeRT, a combination that could maximize ablative dosing while preserving surrounding tissues. Moreover, SBRT's ability to treat the triangle volume, a common site of margin-positive resections, adds further rationale for its use in this context (Hill et al. 2022). The goal of the current study is to report the clinical outcomes of patients with resectable or borderline resectable PDAC treated with neoadjuvant SBRT followed by pancreatectomy and IOeRT at a single institution. Methods Study design and patient eligibility We conducted a retrospective review of consecutive patients with localized non-metastatic pancreatic adenocarcinoma who underwent pre-operative SBRT followed by Whipple or Distal pancreatectomy for adenocarcinoma with IORT between January 2021 to December 2024. All cases were discussed in multi-disciplinary tumor board with Surgical Oncologist, Medical Oncologist and Radiation Oncologist. Patients deemed eligible for this treatment included patients who were resectable or borderline resectable disease in whom according to surgeon assessment were at risk not to achieve satisfactory resection margins by surgery alone. Data collection Patients’ information was extracted from electronic medical record including tumor characteristics. A detailed, systematic evaluation of patient's information was collected including demographics, clinical and treatment characteristic. Resectability was determined using National Comprehensive Cancer Network guidelines. All tumor characteristics and staging were determined through use of acceptable diagnostic imaging including triphasic Computed Tomography (CT), Magnetic Resonance Imaging, and/or Positron Emission Tomography–CT. Treatment Neoadjuvant chemotherapy Select patients received neoadjuvant chemotherapy prior to radiotherapy at the discretion of the treating medical oncologist. Options included FOLFIRINOX, FOLFOX, or Gemcitabine. SBRT planning and delivery Stereotactic body radiation therapy was performed as follows: Patients were immobilized per standard institutional protocol and underwent CT Simulation scan using 2–3 mm slice thickness. 4-dimensional CT data was used to create internal target volume (ITV). Gross tumor volume (GTV) was delineated using diagnostic triphasic CT, PET-CT and/or MRI, ITV was created after 4D-CT was reviewed. An additional 3 mm margin was added to create planning target volume (PTV) named PTV_high. A clinical target volume (CTV) encompassing the at risk nodes and triangle volume (celiac axis, superior mesenteric artery, portal vein and adjacent vasculature) was delineated when clinically indicated. An additional 5 mm margin was added to create PTV_low. Treatment plan was generated using volumetric modulated arc therapy and patients treated on a Varian Truebeam system. Daily CBCT with soft-tissue matching ensured image-guided delivery. Dose fractionation included 35 – 40 Gy in 5 fractions to PTV_high and 25 Gy in 5 fractions to PTV_low. The highest isodose line of the prescribed dose was recommended to cover > 95% of the PTV and 99% of the GTV/ITV. Organs-at-risk (OAR) constraints for 5-fraction SBRT followed institutional standards adapted for upper-abdominal stereotactic radiotherapy. For the stomach, duodenum, and small bowel, the maximum dose was limited to 35 Gy, with mean dose < 20 Gy, V30 Gy < 5 cc, and V35 Gy < 1 cc. Liver irradiation was restricted to V30 Gy < 10% of total liver volume. Renal constraints required both kidneys to have a mean dose < 10 Gy. The spinal cord maximum dose was limited to 22 Gy. Figure 1 shows an example of pre-operative SBRT.Fig. 1Example of pre-operative SBRT borderline resectable case with high dose to the gross disease and coverage of the Triangle Volume with CTV Example of pre-operative SBRT borderline resectable case with high dose to the gross disease and coverage of the Triangle Volume with CTV Surgery and IOeRT Following SBRT, patients underwent restaging imaging in order to confirm localized/resectable disease. If deemed suitable, patients underwent Whipple or Distal pancreatomy 3 to 8 weeks from completion of SBRT. IOeRT was delivered as follows: after surgical removal of the tumor the mobile IOeRT accelerator was mobilized. A 40-80 mm applicator was chosen which encompassed the tumor bed focused on areas of concern for recurrence. The applicator was placed flush on the post-operative bed and organs at risk (stomach, small bowel, kidney) were confirmed to be out of field. The surgeon dictated the depth of coverage needed and a virtual simulation with physics quality assurance was performed. After ensuring adequate coverage, 10 Gy was delivered using 6-10 MeV electrons using the 80–95% isodose line under the supervision of Radiation Oncologist. In cases of concern for positive margin at time of resection, 12 Gy was delivered. Figure 2 shows an example of delivery of IoeRT during whipple surgery.Fig. 2Example IoERT being delivered at time of pancreatic resection (Whipple) with coverage of high risk regions Example IoERT being delivered at time of pancreatic resection (Whipple) with coverage of high risk regions Toxicity Treatment-related toxicity after SBRT was graded according to Common Terminology Criteria for Adverse Events (CTCAE) criteria (v5.0). Post-operative complications were divided between short-term (< 90 days) and long-term (> 90 days). The Clavien Dindo Classification was used to rank the severity of short-term post-operative complications. Follow-up Patients were monitored in the oncology clinic through routine physical examinations, clinical assessments, laboratory testing (including tumor markers), and imaging studies (CT or PET-CT) every three months. Disease status, including progression and type of progression (local versus metastatic), overall survival was documented throughout follow-up. Statistical analysis Descriptive statistics were used to summarize patient demographics, tumor characteristics, and treatment variables, including age, sex, tumor location, neoadjuvant therapy, and surgical approach. Kaplan–Meier survival analyses were performed to estimate overall and progression-free survival. Results The study cohort included 15 patients with a mean age of 66.1 years (range: 44–81). Eleven patients were male (73.3%) and four were female (26.7%). Tumors were located in the body of the pancreas in 9 patients and 6 patients in the head. Ten patients (66.7%) were classified as borderline resectable and received neoadjuvant chemotherapy prior to stereotactic body radiotherapy (SBRT), while five patients (36%) were resectable at diagnosis and received upfront SBRT. Genomic analysis revealed BRCA2 mutations in 3 patients (20%), while 12 patients (80%) had no BRCA mutations. Neoadjuvant chemotherapy regimens included FOLFIRINOX in 7 patients (46%), FOLFOX in 2 (13%), and gemcitabine in 1 (6%). Preoperative SBRT was delivered at a dose of 35 Gy in 5 fractions in 11 patients (73%) and 40 Gy in 5 fractions in 4 patients (27%). Surgical approaches included distal pancreatectomy in 9 patients (60%) and pancreaticoduodenectomy (Whipple procedure) in 6 patients (40%). Intraoperative radiotherapy (IORT) was administered at 10 Gy in 14 patients (93%) and 12 Gy in 1 patient (7%) (Table 1).Table 1Demographic, clinical and treatment characteristicsCharacteristicsN = 15GenderMale11 (73%)Female4 (27%)Race/EthnicityAshkenazi Jewish Descent9 (60%)Sephadi Jewish Descent5 (33%)Arab Descent1 (7%)Genomic analysisBRCA 23 (20%)No BRCA12 (80%)Neoadjuvant chemotherapyFOLFIRINOX7 (46%)FOLFOX2 (13%)Gemcitabine1 (6%)Pre-op SBRT dose35 Gy in 5 fractions11 (73%)40 Gy in 5 fractions4 (27%)IOeRT Dose10 Gy14 (93%)12 Gy1 (7%)Surgery TypeWhipple (pancreaticoduodenectomy)6 (40%)Distal Pancreatectomy9 (60%) Table 1Demographic, clinical and treatment characteristicsCharacteristicsN = 15GenderMale11 (73%)Female4 (27%)Race/EthnicityAshkenazi Jewish Descent9 (60%)Sephadi Jewish Descent5 (33%)Arab Descent1 (7%)Genomic analysisBRCA 23 (20%)No BRCA12 (80%)Neoadjuvant chemotherapyFOLFIRINOX7 (46%)FOLFOX2 (13%)Gemcitabine1 (6%)Pre-op SBRT dose35 Gy in 5 fractions11 (73%)40 Gy in 5 fractions4 (27%)IOeRT Dose10 Gy14 (93%)12 Gy1 (7%)Surgery TypeWhipple (pancreaticoduodenectomy)6 (40%)Distal Pancreatectomy9 (60%) Demographic, clinical and treatment characteristics Oncologic outcomes Thirteen of fifteen patients (86.7%) attained negative margins. Notably, among these, two patients harboring BRCA gene mutations demonstrated complete pathological responses following neoadjuvant chemotherapy followed by pre-operative SBRT. Median progression-free survival and overall survival was 16 months (CI 95% 7–25 months) and 30 months (CI 95% 23–36 months), respectively (See Figs. 3, 4).Fig. 3Kaplan Meir curve of progression free survivalFig. 4Kaplan Meir curve of overall survival Kaplan Meir curve of progression free survival Kaplan Meir curve of overall survival Within this study group, 1 patient (6.7%) exhibited isolated local recurrence while 8 patients (53.3%) presented with metastatic disease. 9 patients in the follow up period died with causes of mortality including distant metastases (6 patients), as well as early post-operative sepsis (1), biliary sepsis (1), and complications related to coronavirus infection (1). Toxicity Toxicity and postoperative complications are summarized in Tables 2 and 3. The most common SBRT-related adverse events were grade 1 fatigue and nausea, each occurring in 6 patients (40%), followed by grade 2 abdominal pain in 4 patients (26.7%) and grade 1 abdominal discomfort in 3 patients (20%). Postoperative complications, classified by Clavien–Dindo, included grade 1 events in 4 patients (ileus or superficial surgical-site infection) and grade 2 events in 4 patients (hospital-acquired pneumonia, bacteremia, deep-vein thrombosis, or urinary-tract infection), all requiring medical management. One patient developed a grade 3 complication—wound dehiscence requiring re-operation. One patient experienced a grade 5 early postoperative event due to acute-on-chronic renal failure and pneumonia not related to radiotherapy. An additional patient died from late postoperative complications (> 30 days), specifically hepatic artery occlusion with biloma and biliary sepsis, deemed possibly related to radiotherapy.Table 2SBRT acute toxicity (CTCAE v.5)ToxicityGradeNumber of patients (%)FatigueGrade 17 (46.7%)Grade 20NauseaGrade 13 (20.0%)Grade 23 (20.0%)Abdominal PainGrade 14 (26.7%)Grade 24 (26.7%)Table 3Post-operative complications (30 Days)Clavien-Dindo gradeNumber of patientsComplicationsTreatment requiredGrade 14 (26.7%)Ileus, surgical site infectionConservative managementGrade 24 (26.7%)Hospital-acquired pneumonia (HAP), bacteremia, DVT, UTIAnticoagulants, antibioticsGrade 31 (6.7%)Dehiscence requiring repeated surgical interventionSurgical re-interventionGrade 51 (6.7%)Early post-operative complications (acute on chronic renal failure, Hospital Acquired Pneumonia)Fatal outcome, post-operative complications Table 2SBRT acute toxicity (CTCAE v.5)ToxicityGradeNumber of patients (%)FatigueGrade 17 (46.7%)Grade 20NauseaGrade 13 (20.0%)Grade 23 (20.0%)Abdominal PainGrade 14 (26.7%)Grade 24 (26.7%) SBRT acute toxicity (CTCAE v.5) Table 3Post-operative complications (30 Days)Clavien-Dindo gradeNumber of patientsComplicationsTreatment requiredGrade 14 (26.7%)Ileus, surgical site infectionConservative managementGrade 24 (26.7%)Hospital-acquired pneumonia (HAP), bacteremia, DVT, UTIAnticoagulants, antibioticsGrade 31 (6.7%)Dehiscence requiring repeated surgical interventionSurgical re-interventionGrade 51 (6.7%)Early post-operative complications (acute on chronic renal failure, Hospital Acquired Pneumonia)Fatal outcome, post-operative complications Post-operative complications (30 Days) Discussion Achieving a margin-negative (R0) resection remains a cornerstone of curative treatment in pancreatic ductal adenocarcinoma (PDAC), yet R1 rates remain unacceptably high, ranging from 40 to 60% in large randomized trials such as ESPAC-4 and PRODIGE-24 (Cunningham et al. 2024; Ghaneh et al. 2024). The PREOPANC trial demonstrated that neoadjuvant chemoradiotherapy can improve resection margins and overall survival compared to upfront surgery, but even in that setting, 28% of patients underwent surgery with positive margins and 21% had a locoregional recurrence as a first event, likely due to the modest radiation dose (36 Gy in 15 fractions; BED ~ 45 Gy) (Versteijne et al. 2022). Our study sought to address this therapeutic gap by combining stereotactic body radiation therapy (SBRT) with intraoperative electron radiation therapy (IOeRT) to safely escalate the biological dose in the neoadjuvant setting and at the time of surgery. This approach resulted in an R0 resection rate of 86%, a low isolated local failure rate of 6.7%, and two complete pathological responses among BRCA2 mutation carriers, outcomes that compare favorably to historical controls. These findings suggest that combining SBRT with IOeRT may improve local control while maintaining acceptable toxicity, particularly in patients with high-risk tumor biology. Dose escalation has been increasingly recognized as critical for improving outcomes in PDAC. The SMART trial demonstrated that delivering 50 Gy in five fractions via SBRT under MRI guidance was feasible and associated with encouraging local control rates in both patients who underwent surgery after ablative radiation therapy and those who did not (Herman et al. 2024). Other retrospective series have similarly reported that achieving BEDs above 70–100 Gy correlates with improved local outcomes (Teng et al. 2020). However, proximity to radiosensitive gastrointestinal structures often limits external beam dose escalation when daily MRI imaging is unavailable. Our study circumvents this limitation by combining high-dose SBRT with CT based treatment followed by IOeRT, allowing focused intraoperative boosts to areas at greatest risk for recurrence. This dual-modality strategy aligns with modern radiation oncology principles emphasizing the importance of ablative dosing for radioresistant malignancies. The use of IOeRT has long been explored in PDAC, and our findings are consistent with more recent data suggesting its evolving role in modern treatment paradigms. In a retrospective study by Keane et al., patients treated with intensive neoadjuvant therapy, including chemotherapy and conventional chemoradiotherapy, followed by IOeRT demonstrated favorable survival outcomes with acceptable toxicity (Keane et al. 2018). Their reported median survival for resected patients approached 33 months, with most local failures occurring in patients who had incomplete resections or did not receive IOeRT. Notably, Keane et al. highlighted that imaging after neoadjuvant therapy often overestimated unresectability, and IORT enabled local dose intensification while preserving surrounding tissue integrity. Our findings build on this by showing that preoperative SBRT may further complement IOeRT by sterilizing gross disease and treating at-risk regions (e.g., triangle volume) before surgery, facilitating earlier surgical exploration and maximizing the therapeutic index. Our local control outcomes compare favorably with other high-dose radiotherapy strategies reported in the literature. For instance, Krishnan et al. demonstrated in a large retrospective cohort that focal radiation dose escalation beyond 70 Gy (EQD2) following chemotherapy significantly improved overall survival and local control in locally advanced PDAC (Krishnan et al. 2016). Similarly, Herman et al. reported a local progression rate of 11% at 12 months using 33–40 Gy in 5 fractions with SBRT alone, though grade ≥ 2 GI toxicity approached 11%, reflecting the challenge of balancing efficacy and toxicity with external beam approaches (Herman et al. 2015). In our series, the combination of SBRT and IOeRT allowed us to maintain high conformality and dose intensification while minimizing gastrointestinal toxicity, with only one long-term toxicity-related death potentially attributable to radiation. Compared with historical IOeRT-only series, such as those from Calvo et al. or the MGH group, our R0 rate of 86% and 30-month median overall survival suggest that the combination of modern SBRT and IOeRT in a neoadjuvant protocol may further enhance local efficacy in selected patients (Calvo et al. 2020; Cai et al. 2015). The addition of preoperative SBRT may sterilize tumor margins prior to resection, allowing IOeRT to serve as a focused boost rather than a salvage measure, potentially explaining the low rate of isolated local recurrence in our study (6.7%) (Chuong et al. 2013). Importantly, we observed two complete pathological responses in BRCA2-mutated tumors, consistent with the emerging understanding that defects in homologous recombination may enhance radiosensitivity. Future studies should investigate molecular predictors of response, including germline or somatic mutations in DNA repair pathways, to guide intensified local therapy. Our study has limitations. The retrospective design introduces inherent selection biases, and the small sample size limits generalizability. For example, a notable feature of our cohort was the predominance of pancreatic body tumors (60%), which differs from typical surgical series. This likely reflects referral patterns as well as the anatomical suitability of body lesions for high-dose SBRT and IOeRT, potentially contributing to the favorable margin outcomes observed. Additionally, molecular profiling was limited to BRCA status, and although local control was excellent, distant metastases remained the dominant pattern of failure. As systemic therapy improves, the role of aggressive local control may become increasingly important, especially in biomarker-selected patients. Prospective randomized trials are warranted to confirm these findings and define the optimal integration of SBRT and IOeRT in PDAC management.