Poor Staging Compromises Survival in Gastric Cancer

TORONTO, Sept. 25 -- More than two-thirds of gastric cancer patients received inadequate lymph node assessment, jeopardizing their treatment and survival, researchers reported after an analysis of nearly 11,000 cases.

Median survival in a U.S. database region with the highest level of lymph-node assessments was 33 months compared with only 17 months for the region with the lowest assessments, according a study published online by the Nov. 1 issue of Cancer.

In 1997, the staging guidelines set by the American Joint Commission on Cancer and the Union Internationale Contre le Cancer were revised to standardize largely lax staging techniques. The new guidelines require examination of at least 15 resected lymph nodes.

However, compliance with the new guidelines is poor, said Natalie Coburn, M.D., of the University of Toronto here, and colleagues. They found that after 1997 only 29% of cases had 15 or more lymph nodes examined. The median number of lymph nodes assessed increased from nine to only 10 (P < .0001).

These findings came from a 2005 analysis of 10,807 resected, non-metastatic gastric cancers in the NCI's Surveillance, Epidemiology and End Results (SEER) database, 1988-2002. Kaplan-Meier survival curves and Cox proportional hazards were used for the analysis.

Although only 29% of gastric cancer patients in this study had at least 15 nodes resected, regional differences affected the rates.

Prior to the new guidelines, 27% of patients had adequate assessment; after the changes, 33.1% had adequate assessment (P< .0001). Moreover, in recent years the median number of nodes assessed reached a plateau, with no trend toward further improvement, Dr. Coburn said.

Factors that predicted adequate node assessment were higher stage, worse grade, age younger than 74, later year of diagnosis, non-white race, more extensive surgery, female sex, and SEER region.

Differences in the rate of node assessment, ranged from 19.7 % to 53% (P<.0001), the researchers found. Of T1 N0 patients, only 19% were adequately assessed.

Because of inadequate nodal assessment, up to 11% of the total population failed to meet the threshold needed to qualify for potentially beneficial adjuvant therapy, the researchers said.

Survival was significantly associated with the region in which the patient underwent resection (P<.0001). After adjustment for potential confounding factors, median overall survival was 33 months in the region with the best node assessment rate and 17 months in the region with the lowest rate.

Improved survival was predicted by earlier stage, lower grade, marital status, Asian race, younger age, T-stage, female sex, SEER region, and adequate lymph node assessment.

Better survival with node assessment was most apparent in stages I and II, but maintained statistical significance through stages III and IV. On the other hand, inadequate lymph node assessment led to poorer survival at every stage (P < .001).

The notable three-fold regional difference in node examination rates may have been due to the region's socioeconomic status, surgery rates at tertiary care centers, hospital or surgeon volumes, or regional variation in practice patterns. There may also be key management differences in surgery and pathology departments, Dr. Coburn suggested.

Among the study's limitations, the researchers mentioned the inability to control all possible confounding variables, including comorbidities. Chemotherapy is not recorded in the SEER database, so that its effect could not be assessed on the likelihood of receiving treatment.

The type of institution and provider volume were also not accessible through the SEER database, and there was no way to tell whether the surgeon performed the resection for palliative or curative purposes. In addition, the majority of cases did not differentiate histopathology by type (intestinal or diffuse).

Despite these limitations, the researchers said, the SEER database offered a good representation of actual practice patterns in the U.S. and eliminated referral bias that occurs in single institution studies.

For an overwhelming majority of patients, lymph node assessment in gastric cancer is inadequate, with significant regional variations that compromise patient care, the researchers wrote. Reasons for noncompliance with the guidelines need to be further assessed.

Education for pathologists, surgeons, and medical oncologists should improve lymph node assessment, Dr. Coburn wrote, and "by proxy, improve the care received by patients with gastric cancer and their overall survival."