Baseline CA19-9 at diagnosis stratifies prognosis in patients with resected pancreatic adenocarcinoma treated with neoadjuvant therapy or surgery-first sequencing

Presenter: James Griffin MD | The University of Texas MD Anderson Cancer Center Background: Carbohydrate antigen (CA) 19-9 is often used as a biomarker for pancreatic adenocarcinoma (PDAC) disease burden and treatment response. However, the treatment-sequencing prognostic utility of CA19-9 at diagnosis is undefined. The aim of this study was to evaluate the ability of baseline CA19-9 to stratify oncologic outcomes of patients undergoing neoadjuvant therapy (NT) versus surgery-first (SF) sequencing. Methods: Data for consecutively resected PDAC patients (2/2010-10/2018) were obtained from a prospective database. CA19-9 nonproducers ( < 1 U/ml) and patients with missing baseline CA19-9 and corresponding total bilirubin ≤2 mg/dl were excluded. Normal CA19-9 was defined as ≤40 U/ml. Patients were divided into 4 groups according to treatment sequencing and baseline CA19-9: NT/Normal, NT/Elevated, SF/Normal, and SF/Elevated. Results: Of 302 patients, 75% received NT. Postoperative therapy was given to 67% of NT and 89% of SF patients. Nearly all SF patients were potentially resectable (99%), whereas NT patients tended to be of more advanced clinical stage (borderline 32%, locally advanced 8%, p<0.001) and with higher baseline CA19-9 (median 56 vs. 123 U/ml, p<0.001). NT patients had lower rates of final pathology node positivity (54% vs. 76%, p=0.001), lymphovascular invasion (49% vs. 76%, p<0.001), and perineural invasion (78% vs. 97%, p<0.001). Median follow-up among survivors was 45 months (range 2-101 m). OS for NT/Normal was the most favorable (Fig. 1A: median not reached) and significantly better than SF/Normal (median 35mo, p=0.04). Despite starting with presumably lower disease burden as reflected in the normal baseline CA19-9, SF/Normal patients had OS comparable to NT/Elevated (median 38 mo, p=0.8). OS was worst for SF/Elevated (median 29 mo). For NT/Elevated, 45% of patients normalized their CA19-9 prior to resection, with a median OS 52mo vs. 30mo for non-normalization (Fig. 1B; p=0.04). After adjusting for clinical classification, treatment sequencing, tumor size, and margin status, Cox regression identified elevated baseline CA19-9 as an independent predictor of OS (HR 1.9, p=0.007) in addition to non-receipt of postoperative therapy (HR 1.6, p=0.03), node-positivity (HR 1.9, p=0.001), and perineural invasion (HR 2.5, p=0.004). Conclusion: Elevated baseline CA19-9 is an independent predictor of worse OS in patients with resected PDAC. This worse prognosis may be mitigated in part by the use of NT to down-stage the disease burden through normalization of CA19-9. Even in patients with putative “early-stage” disease as defined by normal CA19-9, there seems to be benefit to choosing NT over SF.