KEY POINTS
- The investigator-initiated PACK study (NCT04194268) was a prospective, single-center phase II trial conducted at Heidelberg University Hospital and the Heidelberg Ion Beam Therapy Center (HIT). Between December 2019 and July 2024, 25 patients with nonmetastatic, inoperable pancreatic cancer were enrolled, and 23 were evaluable, including 11 with locally advanced pancreatic cancer (LAPC) and 12 with locally recurrent disease after surgery (LRPC). Two patients were excluded before treatment because of excessive respiratory motion or newly detected distant progression.
- Carbon-ion radiotherapy delivered 48 Gy(RBE) in 12 fractions over two weeks, using active raster scanning, posterior oblique beams, and biological dose optimization with LEM I. Twenty patients (87%) received induction chemotherapy, predominantly FOLFIRINOX, while 22/23 (96%) received concurrent gemcitabine 300 mg/m² weekly. All patients underwent baseline FAPI-PET/CT, with additional CT or MRI for target definition. Elective nodal irradiation was omitted, and treatment planning incorporated individualized margins, gastrointestinal PRVs, and priority for OAR constraints.
- The prespecified primary endpoint of one-year overall survival was met, with an observed rate of 78% (95% CI 63–97%), compared with a predefined historical benchmark of 36%. Two-year OS was 59% (95% CI 41–84%), and median OS was 26.5 months from study inclusion. One-year OS was numerically higher in locally recurrent than locally advanced disease (92% vs 64%), although the small subgroups and different prior disease trajectories prevent reliable comparative interpretation.
- Progression-free survival remained substantially shorter than overall survival. Median PFS was 8.1 months, with one- and two-year PFS rates of 26% and 10%, respectively. Median local progression-free survival was 16.9 months, whereas local control at one and two years was 87% and 77%. At two years, local control was numerically higher in LAPC than in locally recurrent disease (91% vs 66%). Distant progression occurred predominantly in the liver (seven patients) and lungs (six patients), underscoring the persistent systemic failure risk.
- Gastrointestinal organ sparing frequently required compromising target coverage. Although all predefined OAR constraints were achieved, mean GTV V95% was 92.5% (SD 7.4%), and mean ITV V95% was only 72.1% (SD 10.1%). Five patients required adaptive replanning, four because of gastrointestinal anatomical changes and one to improve robustness. Only one patient (4%) achieved a partial radiological response, and no patient became eligible for surgical resection following treatment.
- Three treatment-related grade 3 adverse events occurred in two patients, comprising abdominal pain, upper gastrointestinal hemorrhage, and ureteric stricture. The hemorrhage and stricture occurred in the same patient, who also developed peritoneal carcinomatosis, complicating attribution to irradiation versus progressive malignancy. There were no reported treatment-related grade 4–5 events, and body weight and performance status were generally maintained. However, the study was too small to reliably exclude uncommon severe gastrointestinal complications.
- The findings are limited by the small, nonrandomized cohort and substantial heterogeneity between LAPC and postoperative local recurrence, with 52% of patients belonging to the latter group. Highly selected eligibility, FAPI-PET/CT staging, exclusion of tumors inseparable from the gastrointestinal tract, extensive previous chemotherapy, and variable post-progression treatment may all have influenced survival. Furthermore, OS was measured from trial inclusion rather than the beginning of induction chemotherapy, and comparisons with historical photon chemoradiotherapy studies cannot establish carbon-ion superiority.
CLINICAL TAKEAWAY
The PACK study demonstrates that hypofractionated carbon-ion chemoradiotherapy combined with gemcitabine is feasible in selected patients with nonmetastatic, inoperable pancreatic cancer, with encouraging survival and relatively limited observed severe toxicity. However, frequent distant progression, compromised target coverage, and the absence of a randomized comparator mean that superiority over contemporary photon-based treatment or systemic therapy alone remains unproven.