By William B. Ershler, MD, Editor
Synopsis: In a retrospective analysis of second-line or greater treatment of advanced pancreatic cancer, the change in CA19-9 level was found to predict progression-free survival and overall survival. In this palliative setting, such data may support early discontinuation of potentially toxic treatments.
Source: Nakai Y, et al. A retrospective analysis of early CA19-9 change in salvage chemotherapy for refractory pancreatic cancer. Cancer Chemother Pharmacol 2013;72:1291-1297.
Advances in the clinical management of advanced pancreatic cancer have been slow in coming since the original demonstration of improved overall survival (OS) with gemcitabine.1 However, two recent randomized trials have described improvements in clinical parameters including OS with either the multi-agent FOLFIRINOX regimen2 or, as presented last year at ASCO and more recently in print, gemcitabine with nab-paclitaxel3 (also reviewed in this issue of Clinical Oncology Alert). Yet, despite these advances, most patients experience disease progression in 3-6 months and second-line chemotherapy regimens have proven only marginally effective.4,5
Acknowledging that quality of remaining life remains high priority, early assessment of whether salvage chemotherapy is going to be of benefit is of practical importance. Accordingly, Nakai and colleagues at the University of Tokyo conducted a retrospective analysis of outcomes of those patients who received salvage chemotherapy or chemoradiation for advanced or recurrent pancreatic cancer. At their institution CA19-9 was routinely measured prior to each course of chemotherapy. Patients whose serum CA19-9 was below 1 U/mL (likely due to Lewis blood group antigen phenotype6 or was not measured after the first course of salvage chemotherapy) were excluded from the analysis.
A total of 239 salvage regimens were given in 167 patients. The treatment was second line in 150 patients and third or more in the remainder. Median follow-up was 5.9 months. There were no complete remissions (by RECIST criteria) but partial remissions were evident in 18 and stable disease was observed in 109 patients. Median progression-free survival (PFS) was 2.7 months (95% confidence interval [CI], 2.4-3.1) and median OS was 6.1 months (95% CI, 5.3-7.2).
For all evaluable treatment cycles, the median pretreatment CA19-9 was 2362 U/mL, and median CA19-9 change after the first course was 17.8% increase. Of note, the CA19-9 change was associated with tumor response. PFS was 1.7 vs 3.5 months and OS was 3.9 and 8.6 months in patients with ≥ 50% vs < 50% increase. By multivariate analysis, CA19-9 increase ≥ 50 % was prognostic of both PFS and OS (hazard ratio [HR]. 2.28 and 2.50; P < 0.001, respectively).
The measurement of CA19-9 after the first cycle of salvage chemotherapy was useful in identifying those individuals who were likely to benefit from additional therapy. This group had previously demonstrated the efficacy of early CA19-9 measurement in predicting initial treatment response with gemcitabine monotherapy in chemotherapy-naïve pancreatic cancer.7 In that study, both pretreatment CA19-9 and CA19-9 after the first course were predictive of PFS and OS. The current study extends this observation to the assessment of second- or third-line treatment. Once again, demonstration of a significant (> 50%) rise in CA19-9 after the first cycle was shown to predict poor PFS and OS. Although no standard approach to second-line treatment of pancreatic cancer is established, oncologists frequently evaluate treatment response by using imaging and other response criteria after two full cycles of therapy. The current report would suggest an earlier evaluation of response (i.e., by the post-cycle 1 CA19-9 level) may allow the avoidance of unnecessary adverse events from additional cycles of ineffective chemotherapy and, when appropriate, the administration of an alternative regimen.
However, as the authors themselves indicate, some reservation in adopting this strategy should be kept in mind in light of the retrospective nature of the review and its origin from a single institution. Hopefully, as multi-institution, randomized trials are developed for second-line treatment of pancreatic cancer, CA19-9 levels will be included both pre- and post-first cycle treatment and responses compared with the established RECIST criteria.