Published in:
01-07-2009 | Gastrointestinal Oncology
Prediction of Recurrence of Early Gastric Cancer After Curative Resection
Authors:
Ji Fu Lai, MD, Sungsoo Kim, MD, Kiyeol Kim, PhD, Chen Li, MD, Sung Jin Oh, MD, Woo Jin Hyung, MD, Sun Young Rha, MD, Hyun Cheol Chung, MD, Seung Ho Choi, MD, Lin Bo Wang, MD, Sung Hoon Noh, MD
Published in:
Annals of Surgical Oncology
|
Issue 7/2009
Login to get access
Abstract
Background
Recurrence of early gastric cancer (EGC) after curative resection is rare, and the types of EGC that may recur have not been well studied. We attempted to create a system for predicting recurrence of EGC after R0 resection.
Methods
From January 1987 to April 2005, 2,923 patients with EGC who underwent curative resection were retrospectively studied. Of them, 79 patients (2.7%) experienced recurrence. Logistic regression was performed to identify independent risk factors for overall recurrence and early recurrence (recurred within 24 months after resection) of EGC. A nomogram was developed on the basis of a Cox regression.
Results
Median time to recurrence was 20.5 months, and early recurrence accounted for 60.7% of instances. Presence of lymph node metastasis and elevated gross type were independent risk factors for overall recurrence; patients with both identified risk factors had a higher recurrence rate than average level (17.5% vs. 2.7%, P < 0.001). Meanwhile, male gender, elevated gross type, and presence of lymph node metastasis were significantly associated with early recurrence, and in patients with all of the aforementioned identified risk factors, the early recurrence rate was higher (12.2% vs. 1.6%, P < 0.001). A nomogram for predicting the disease-free survival after operation was constructed. Its c-index was 0.79 and it appeared to be accurate.
Conclusions
Recurrence of EGC after curative resection can be predicted by using common clinical characteristics. Patients at high risk of overall and early recurrence could be identified; individual disease-free survival was predictable by the internally validated nomogram.