A report card on outcomes for surgically treated gastrointestinal cancers: are we improving?

Longitudinal analyses of cancer registries provide an opportunity for population-based explanations of epidemiology and survival-related outcomes. This study used two population-based data sets to report on nine surgery-related cancers over the past three decades.

Using the SEER cancer database (1973-1999), all patients (>18 years old) with adenocarcinoma of esophagus, gastric, biliary system, pancreas, small bowel, colon, rectum; esophageal squamous cell carcinoma (ESC), or hepatocellular (HCC) carcinoma (n = 379,640) were analyzed. Changes in incidence rates, stage at diagnosis, and 5-year cancer and stage-specific survivals were determined. A separate database, the California inpatient database (1990-2000), was concurrently used to evaluate inpatient mortality after surgical resection (n = 34,057).

Incidence rates increased for three cancers (esophageal, HCC, small bowel); decreased for three (rectal, gastric, ESC); and stayed constant for three (biliary, pancreatic, colon). More patients presented with local/regional disease in the 1990s versus 1970s for eight tumors (except small bowel, P < 0.05). Five-year overall survival improved for all but small bowel (P < 0.05); and local stage survival was improved for all except small bowel and biliary (P < 0.05). Finally, inpatient mortality rates improved significantly for liver, esophageal, pancreatic, and gastric resections (P < 0.05) over the past decade.

For these nine surgically treated cancers, we are detecting disease at earlier and therefore more treatable stages, and surgical care and outcomes also appear to have improved. Continued reexamination of longitudinal trends of surgically relevant outcomes is important for future improvement of surgical care.