Development and validation of nomogram for predicting early recurrence after radical gastrectomy of gastric cancer

Background: After radical surgery, early detection of recurrence and metastasis is a crucial factor in enhancing the prognosis and survival of patients with gastric cancer (GC). Therefore, assessing the risk of recurrence in gastric cancer patients and determining the timing for postoperative recurrence is crucial.

Methods: The clinicopathological data of 521 patients with recurrent gastric cancer, who underwent radical gastrectomy at Zhejiang Cancer Hospital between January 2010 and January 2017, were retrospectively analyzed. These patients were randomly divided into two groups: a training group (n = 365) and a validation group (n = 156). In the training set, patients were further categorized into early recurrence (n = 263) and late recurrence (n = 102) groups based on a 2-year boundary. Comparative analyses of clinicopathological features and prognoses were conducted between these two groups. Subsequently, a nomogram for predicting early recurrence was developed and validated.

Results: In this study, the developed nomogram incorporated age, serous infiltration, lymph node metastasis, recurrence mode, and the tumour marker CA19-9. In the training cohort, the area under the curve (AUC value) was 0.739 (95% CI, 0.682-0.798), with a corresponding C-index of 0.739. This nomogram was subsequently validated in an independent validation cohort, yielding an AUC of 0.743 (95% CI, 0.652-0.833) and a C-index of 0.743. Furthermore, independent risk factors for prognosis were identified, including age, absence of postoperative chemotherapy, early recurrence, lymph node metastasis, abdominal metastasis, and vascular cancer embolus.

Conclusion: Independent risk factors for gastric cancer recurrence following radical surgery were utilized to construct a nomogram for predicting early relapse. This nomogram effectively assesses the risk of recurrence, aids in treatment decision-making and follow-up planning in clinical settings, and demonstrated strong performance in the validation cohort.