Article Abstract

Clinical trends and effects on quality metrics for surgical gastroesophageal cancer care

Authors: Roderich E. Schwarz


Background: Surgical therapy of mid-stage gastric cancer (GC) and other neoplastic conditions requiring gastric resection remains at the center of curative outcomes, while epidemiologic changes and multimodality treatment options have evolved rapidly. Putative quality metrics for gastrectomy such as R0 rate, total lymph node (LN) count or postoperative morbidity may depend partly on changing disease and treatment patterns, and deserve evaluation under various practice conditions.
Methods: Data within a U.S.-based single surgical oncologist’s practice over 15 years were prospectively recorded and retrospectively analyzed for clinicopathologic factors, operative treatment aspects and outcomes. Trends and spectrum changes over three time intervals were analyzed with contingency analysis and continuous data comparative statistics.
Results: Of 179 patients undergoing gastric resection, 119 were male and 60 female, with a median age of 63 years (range, 24–98 years). Resections included 56 total, 56 subtotal/distal, 30 proximal and 37 segmental gastrectomies. Diagnoses included 96 GCs, 31 gastroesophageal (GE) junction (GEJ) cancers, 21 GI stromal tumors (GISTs), and 31 other conditions. Significant trends from first towards last time interval were observed for resection type (16% to 32% proximal, 9% to 30% segmental, P=0.0003), curative intent (76% to 98%, P=0.002), diagnosis (5% to 42% GEJ cancer, P<0.0001) and preoperative therapy use (0% to 58%, P<0.0001), among others. Intraoperative aspects showed significantly reduced blood loss (median: 500 to 150 mL) and transfusion requirements (39% to 4%), and an increased use of minimally invasive techniques over time (all at P<0.001). Among patients undergoing curative intent GC resection with LN dissection, total LN counts remained steady (mean: 26), while the number of involved LNs decreased (9.0 to 3.7, P=0.0003) and the R0 resection rate increased from 74% to 85% (P=0.05). The number of specimens with >15 LNs examined increased from 69.0% to 92.5% (P=0.022). At the same time, spleen preservation rate (91% overall) and major morbidity (16%) remained unchanged throughout. Postoperative length of stay decreased from a median of 12 to 8 days (P<0.0001).
Conclusions: This experience represents some variable practice patterns within a clinicopathologic spectrum of GE diseases. Postoperative or oncologic quality metrics have been sustained or did improve, which would support their utility for various practice settings; they compare favorably to other published U.S. experiences during the same time period.