Why this matters
After induction chemotherapy for pancreatic adenocarcinoma, selecting a preoperative radiation strategy involves more than tumor response. Treatment duration, concurrent chemotherapy, readiness for surgery, and patient-reported function all matter.
SOFT Preop directly compared five-fraction SBRT with conventionally fractionated chemoradiotherapy. The investigators hypothesized that the longer chemoradiotherapy regimen would produce fewer pathologically positive lymph nodes at surgery.
That hypothesis was not supported. The clearest favorable finding for SBRT was better preservation of patient-reported quality of life, a secondary endpoint. The study did not establish that the two strategies provide equivalent cancer control.
Study design
SOFT Preop was a prospective randomized phase II trial in patients with resectable, borderline-resectable, or protocol-defined locally advanced type A pancreatic ductal adenocarcinoma. Patients had ECOG performance status 0-2 and had already received more than one month of neoadjuvant chemotherapy before enrollment.
Patients were randomized between:
- Conventional chemoradiotherapy: 50.4 Gy in 28 fractions with concurrent gemcitabine or capecitabine.
- SBRT: five fractions without concurrent chemotherapy, delivering 25 Gy to elective/high-risk nodal regions and a higher dose to the primary tumor. The presentation schema specifies 33-35 Gy to the primary, while the abstract reports 33-40 Gy.
Both strategies treated high-risk nodal regions. This was not a comparison between comprehensive chemoradiotherapy and primary-tumor-only SBRT. Most patients receiving SBRT were treated using adaptive 1.5 T MRI guidance.
The primary endpoint was pathological nodal positivity. The design sought to detect a reduction from an assumed 35% with SBRT to 15% with chemoradiotherapy. Secondary endpoints included survival, patient-reported quality of life using PROMIS, treatment toxicity, surgical complications, and pathological outcomes.
Of 102 enrolled patients, four withdrew. The presentation included 98 patients: 50 receiving SBRT and 48 receiving chemoradiotherapy. Most had received FOLFIRINOX before enrollment.
Median follow-up was 3.4 years.
Key results
The primary endpoint did not favor conventional chemoradiotherapy.
Eighty-three patients underwent resection: 41 after SBRT and 42 after chemoradiotherapy. Pathologically positive lymph nodes were found in:
- 17/41 patients after SBRT: 41%
- 16/42 after chemoradiotherapy: 38%
- p=0.754
The anticipated improvement in nodal clearance with conventional chemoradiotherapy was not demonstrated. Importantly, this pathological comparison concerns patients who reached surgery, not the entire enrolled population.
Patient-reported quality of life favored SBRT.
The investigators reported better preservation of physical function, global physical health, and fatigue with SBRT. The presentation’s secondary-endpoint summary reported p<0.01 for these comparisons.
The displayed PROMIS trajectories covered the period from after induction chemotherapy through six months after surgery. However, numerical between-arm effect sizes, questionnaire completion rates, and full longitudinal modeling details were not supplied. The result supports better reported preservation of these domains, not a claim that SBRT produced better absolute scores at every assessment.
Survival differences were not statistically significant.
| Outcome | SBRT | Conventional chemoradiotherapy | p-value |
|---|---|---|---|
| Median overall survival | 26.0 months | 43.0 months | 0.639 |
| Three-year overall survival | 42.6% | 57.5% | 0.198 |
| Median progression-free survival | 14.3 months | 21.6 months | 0.96 |
Overall survival was measured from enrollment after induction chemotherapy, not from diagnosis. The presentation identified the comparison of three-year survival estimates as post hoc.
The median OS confidence intervals were wide: approximately 17.9-59.3 months with SBRT and 18.7-54.7 months with chemoradiotherapy. The numerical differences favored chemoradiotherapy, but neither the overall survival comparison nor the PFS comparison reached statistical significance.
These findings do not demonstrate a survival advantage for either strategy. They also do not exclude clinically important differences.
Surgical and locoregional outcomes require more nuance than “no difference.”
Overall, 83/98 patients, or 84.7%, completed treatment through surgical resection.
The surgical-complication table reported at least one complication in 6/41 patients after SBRT, approximately 15%, versus 1/42 after chemoradiotherapy, 2.4%, p=0.057. The difference was not statistically significant, but the small event count does not establish equivalent surgical safety.
Eight local or regional recurrences were reported: six after SBRT and two after chemoradiotherapy. No comparative significance test for these counts was provided.
The presentation reported no significant differences in the displayed pathological tumor-stage or closest-margin comparisons.
Interpretation
SOFT Preop provides a direct randomized comparison of two preoperative radiation strategies after systemic chemotherapy. Its practical contribution is a patient-reported advantage with the shorter approach, without demonstrating the hypothesized nodal-clearance benefit of conventional chemoradiotherapy.
The quality-of-life finding matters. Five radiation visits without concurrent chemotherapy represent a substantially different treatment experience from 28 fractions with chemotherapy.
However, this is a comparison of treatment packages, not fractionation alone. Concurrent chemotherapy differed between arms, and most SBRT patients received adaptive MRI-guided treatment. The study cannot determine how much of the patient-reported difference arose from course length, chemotherapy exposure, image guidance, or their combination.
The oncologic interpretation must remain separate. Failure to demonstrate superiority of chemoradiotherapy for nodal positivity does not prove that SBRT is noninferior for survival, recurrence, or surgical outcomes. The trial was designed around a relatively large pathological difference, not around excluding a clinically important survival disadvantage.
The numerical survival estimates, locoregional recurrence counts, and surgical-complication rates should therefore remain visible rather than being compressed into a blanket claim of equivalent efficacy and safety. None establishes that SBRT is inferior, but collectively they reinforce the need for adequately powered comparisons and longer follow-up.
The treatment volumes also matter. Both arms included high-risk nodal regions. These findings should not automatically be applied to primary-only SBRT, different nodal coverage, or substantially different SBRT prescriptions.
Finally, there was no chemotherapy-alone arm. SOFT Preop addresses the choice of radiation strategy once preoperative RT is planned; it does not determine whether adding radiotherapy improves outcomes over chemotherapy and surgery alone.
Limitations
The study included 98 patients after withdrawals, and only 83 contributed surgical pathology. Its primary endpoint therefore depended on reaching resection, which can introduce selection into the pathological comparison.
Survival, quality of life, and surgical outcomes were secondary endpoints. The small sample and wide survival confidence intervals limit conclusions about oncologic equivalence.
The presentation also showed a baseline performance-status imbalance: ECOG 0 was more frequent in the conventional chemoradiotherapy arm. This is relevant to interpreting unadjusted survival and patient-reported outcomes, although the supplied results do not establish its effect on the comparisons.
The quality-of-life report lacks detailed effect sizes and questionnaire attrition data. It should not be expanded into a general claim that SBRT reduces all treatment toxicity.
Enrollment occurred after patients had already tolerated induction chemotherapy. The outcomes therefore describe a selected preoperative population and should not be compared directly with survival measured from diagnosis in an unselected pancreatic cancer cohort.