2286 - Induction Chemotherapy Exposure as a Determinant of Local Control Independent of Radiation Modality in Pancreatic Cancer

Purpose/Objective(s):

Over half of patients with pancreatic ductal adenocarcinoma (PDAC) present with borderline resectable (BR) or locally advanced (LA) disease. Standard management includes induction chemotherapy (IC) followed by consolidative radiation therapy (RT), yet optimal IC duration and intensity remain undefined relative to tolerance and clinical outcomes. We hypothesized that IC tolerance would predict chemoradiation tolerance and inform RT modality selection.

Materials/Methods:

We evaluated 124 patients with BR/LA PDAC treated at a single institution between 2014-2023. IC (FOLFIRINOX, FOLFOX, or gemcitabine/nab-paclitaxel) was followed by SBRT (33 Gy/5 fractions) or IMRT with concurrent chemotherapy (50.4 Gy/28 fractions with capecitabine or gemcitabine). Chemotherapy exposure was characterized by completed induction cycles and modifications (breaks, dose reductions, regimen changes). Local control (LC) and overall survival (OS) were defined from induction initiation. Cox and logistic regression analyses evaluated associations between induction intensity, RT selection, treatment tolerance, and clinical outcomes.

Results:

IC intensity was significantly associated with improved LC; each additional completed cycle reduced local failure risk by 13% (HR 0.87; 95% CI 0.81–0.96; p<0.001). Patients receiving =7 cycles had superior 12-month LC (87% vs 63%) and 24-month LC (72% vs 31%) compared to <7 cycles (p=0.0002). On multivariable analysis adjusting for stage, RT modality, and delivered BED, IC intensity remained independently associated with LC (HR 0.87, p=0.001) whereas RT modality and dose were not. Greater IC exposure predicted IMRT receipt (OR 1.15; 95% CI 1.04–1.28; p=0.008).

85% of patients required =1 induction modification: dose reduction (61%), treatment break (22%), or regimen switch (20%), most commonly due to diarrhea. Hematologic and hepatic toxicities were more frequent during concurrent chemoradiation than induction, whereas infection rates were similar. IC tolerance did not predict concurrent chemoradiation intolerance, and RT modality did not modify the IC intensity-LC association.

Conclusion:

Despite frequent modifications, IC intolerance did not predict chemoradiation tolerance. However, induction intensity (i.e. number of cycles) remained independently associated with improved LC, suggesting that systemic therapy exposure may define the biologic ceiling for local tumor control. As such, inadequate induction due to intolerance may attenuate measurable RT benefits irrespective of modality.

	SBRT (n=54)	IMRT (n=70)	p-value
Cycles Completed, median (range)	8 (1-13)	9 (2-30)	0.009
Completed Induction	22 (40.7%)	45 (64.3%)	0.015
IC Dose Reduction	30 (55.6%)	46 (65.7%)	0.334
IC Regimen Switch	11 (20.4%)	14 (20.0%)	1.000
IC Treatment Break	12 (22.2%)	15 (21.4%)	1.000
Concurrent Dose Reduction	N/A	14 (20.3%)	-
Concurrent Break	N/A	20 (29.0%)	-