No abstract

Background: The etiology of cancers of the small intestine is largely unknown. To gain insight into these rare malignancies, we evaluated contemporaneous incidence and survival patterns.Methods: Using small intestine cancer data from 12 population-based registries of the Surveillance, Epidemiology and End Results Program, we calculated age-adjusted and age-specific incidence rates (IRs), IR ratios, and relative survival (RS) rates.Results: In total, 10,945 small intestine cancers (IR = 2.10/100,000 person-years) were diagnosed during 1992 to 2006, including carcinomas (n = 3,412; IR = 0.66), neuroendocrine cancers (n = 4,315; IR = 0.83), sarcomas (n = 1,084; IR = 0.20), and lymphomas (n = 2,023, IR = 0.38). For all histologic groups, males had significantly higher IRs than females, and distinct age-specific gender patterns were limited to intermediate-/high-grade lymphomas. Neuroendocrine cancer rates varied significantly by race, with rates highest among blacks and lowest among Asians/Pacific Islanders. Carcinoma IRs were highest among blacks; sarcoma IRs were highest among Asians/Pacific Islanders; and lymphoma IRs were highest among whites. Age-specific IR patterns were similar across racial/ethnic groups. During 1992 to 2006, duodenal cancer IRs increased more markedly than those for other subsites. RS varied little by gender or race. Neuroendocrine cancers had the most favorable RS, and carcinomas had the least favorable. The greatest improvement in 5-year RS from 1992 to 1998 to 1999 to 2005 was observed for sarcomas and lymphomas.Conclusions: Distinct small intestine cancer IR patterns according to histologic subtype suggest different underlying etiologies and/or disease biology, with susceptibility varying by gender, racial/ethnic groups, and subsite. Temporal patterns support a possible role for diagnostic bias of duodenal cancers.Impact: Future epidemiologic studies of small intestine cancer should consider histologic subtype by gender, race/ethnicity, and subsite. Cancer Epidemiol Biomarkers Prev; 19(8); 1908-18. ©2010 AACR.

Because few studies have assessed the accuracy of lung cancer histologic diagnoses reported by state cancer registries, we examined whether the Iowa Surveillance, Epidemiology, and End Results Cancer Registry (i.e., the Iowa Cancer Registry)-reported lung cancer histologic diagnoses were reliable. We investigated agreement between lung cancer histologic types reported for 413 patients with lung cancer by the Iowa Cancer Registry and those obtained through an independent review of diagnostic slides. Among lung cancer histologic types, small-cell carcinoma had the highest sensitivity (94.1%, 95% confidence interval [CI] = 85.6% to 98.4%), positive predictive value (94.1%, 95% CI = 85.6% to 98.4%), negative predictive value (98.8%, 95% CI = 96.9% to 99.7%), and highest percent exact agreement (98.0%, 95% CI = 96.6% to 99.4%). The lowest sensitivity (21.9%, 95% CI = 9.3% to 40.0%) and positive predictive value (23.3%, 95% CI = 9.9% to 42.3%) were noted for large-cell carcinoma, probably because other more specific features of adenocarcinoma or squamous carcinoma were absent. Adenocarcinoma had the lowest specificity (84.4%, 95% CI = 79.0% to 88.9%), negative predictive value (85.2%, 95% CI = 79.9% to 89.6%), and percent exact agreement (82.9%, 95% CI = 79.2% to 86.6%). Samples collected by cytologic examination (odds ratio [OR] = 2.4, 95% CI = 1.1 to 5.2) or biopsy examination (OR = 2.2, 95% CI = 1.1 to 4.2) were more likely to be misclassified than samples obtained via resection. Thus, the histologic type obtained by the Iowa Cancer Registry is reasonably reliable, but independent slide review is needed for precise histologic typing of lung cancer.