Shaib Yasser, Davila Jessica, Naumann Chris, El-Serag Hashem
Division of Gastroenterology, Michael E. DeBakey VA Medical Center and Baylor College of Medicine, Houston, Texas 77030, USA.
Am J Gastroenterol. 2007 Jul;102(7):1377-82. doi: 10.1111/j.1572-0241.2007.01202.x. Epub 2007 Mar 31.
Pancreatic cancer is the fourth leading cause of cancer death in the United States. Curative intent surgery (CIS) is presumed to be the only curative option in the treatment of pancreatic cancer. The determinants and the outcomes of surgical resection are not clear at the population level.
Using data from nine registries of the Surveillance Epidemiology and End Results (SEER) program, we examined the predictors of receiving CIS as well as the survival of patients receiving CIS between 1987 and 2001. The definition of CIS in SEER is accurate and has been previously validated and found to be highly accurate. Cox proportional hazard model was used to examine the effect of potential determinants on survival.
We identified 32,348 cases of pancreatic cancer. Of those, 3,545 (10.9%) received CIS. The proportion of patients receiving CIS decreased significantly with age (P < 0.0001), was similar across all racial groups (10.8% in whites, 11.4% in blacks, 11.5% in Asians, and 11.2 % in Hispanics, P= 0.5), was slightly higher in men (11.3%vs 10.5%, P= 0.02), decreased with advanced disease stage (P < 0.0001), progressively increased over time (7.5% in 1987-89, 9.1% in 1990-92, 10.4% in 1993-95, 12.4% in 1996-98, and 13.4% in 1999-2001, P < 0.0001), and differed significantly across different SEER registries (9.7% in San Francisco, 11.8% in Connecticut, 12.5% in Detroit, 11.7% in Hawaii, 9.3% in Iowa, 10.05% in New Mexico, 9.5% in Seattle, 13.2% in Utah, 11.4% in Atlanta, P < 0.0001). In the multivariable logistic regression analysis, more recent time periods, younger age, early disease stage, and geographic location, but not race or gender, were independent predictors of receiving CIS. In the Cox survival analysis, younger age, early disease stage, more recent time period, geographic location, and receipt of CIS were independent predictors of improved survival and black race was an independent predictor of shorter survival.
Early stage disease, female gender, younger age, geographic location, and more recent time periods are predictors of receiving CIS. Early disease stage and receipt of CIS are the strongest predictors of improved survival among patients with pancreatic cancer. Black race is an independent predictor of shorter survival.
胰腺癌是美国癌症死亡的第四大主要原因。根治性手术（CIS）被认为是治疗胰腺癌的唯一治愈选择。在人群层面，手术切除的决定因素和结果尚不清楚。
利用监测、流行病学和最终结果（SEER）计划的九个登记处的数据，我们研究了1987年至2001年间接受CIS的预测因素以及接受CIS患者的生存率。SEER中CIS的定义准确，先前已得到验证且被发现高度准确。采用Cox比例风险模型来研究潜在决定因素对生存的影响。
我们识别出32348例胰腺癌病例。其中，3545例（10.9%）接受了CIS。接受CIS的患者比例随年龄显著下降（P<0.0001），在所有种族群体中相似（白人10.8%，黑人11.4%，亚洲人11.5%，西班牙裔11.2%，P = 0.5），男性略高（11.3%对10.5%，P = 0.02），随疾病晚期阶段下降（P<0.0001），随时间逐渐增加（1987 - 1989年为7.5%，1990 - 1992年为9.1%，1993 - 1995年为10.4%，1996 - 1998年为12.4%，1999 - 2001年为13.4%，P<0.0001），并且在不同的SEER登记处有显著差异（旧金山为9.7%，康涅狄格为11.8%，底特律为12.5%，夏威夷为11.7%，爱荷华为9.3%，新墨西哥为10.05%，西雅图为9.5%，犹他为13.2%，亚特兰大为11.4%，P<0.0001）。在多变量逻辑回归分析中，更近的时间段、更年轻的年龄、疾病早期阶段和地理位置，但不是种族或性别，是接受CIS的独立预测因素。在Cox生存分析中，更年轻的年龄、疾病早期阶段、更近的时间段、地理位置和接受CIS是生存改善的独立预测因素，而黑人种族是生存缩短的独立预测因素。
早期疾病、女性、更年轻的年龄、地理位置和更近的时间段是接受CIS的预测因素。早期疾病阶段和接受CIS是胰腺癌患者生存改善的最强预测因素。黑人种族是生存缩短的独立预测因素。
