PGL is the most common extranodal NHL and ranks as second most common gastric malignant neoplasm[1–3]. The PGL is a relatively rare cancer and easily misdiagnosed due to its unspecific symptoms[15–17]. There are many treatment options include gastrectomy, radiotherapy, chemotherapy, immunotherapy, and observation, while gastrectomy remains controversial due to considerably favourable prognosis versus quality of life[18–24]. A previous study by Wang has shown that advanced staging and malignant pathological type are significantly associated with poor overall survival[25]. Most studies have demonstrated that female gender, low-grade histology, good PS, and surgical resection were associated with better overall survival[2, 17, 20]. Our study is consistent with these previous reports except for gastrectomy. The balance between efficacy of gastrectomy and quality of life needs further studies.
Although Ann-Arbor staging system is widely used and recognized for determining the prognosis in PGL, it neglects some significant risk factors such as age, race and marital status[26, 27]. It is necessary to establish a model to predict the risk of PGL, and to make better therapeutic strategies for individual patients. As far as we know, nomograms have been applied to predict the survival status of various cancers[28]. In the present study, we constructed more comprehensive models based on a combination of various risk factors to better predict prognosis in PGL patients. Although the clinical treatment of inert lymphoma such as MALT and invasive lymphoma is different, the nomogram can well distinguish PGL according to pathological type, and better guide the treatment and predict prognosis. These nomograms were capable of making more accurate assessments and predictions in PGL patients both in the training and validation cohorts. The results showed that C-index and calibration curves were great in the validation cohort, indicating that the models were reproducible and reliable. To the best of our knowledge, this is the comprehensive and intensive large-population study to construct nomograms for patients of PGL, and the web-based dynamic nomogram can directly help clinicians quantify the probability. CSS is an epidemiological statistical method, which can exclude deaths caused other than tumor. It is an ideal prediction model, which is different from the actual situation, however, the index can better predict the mortality attributable to the cancer, so as to better design appropriate therapeutic strategies for individual patients.
Several limitations should be acknowledged in our study. First, this was a large sample retrospective study based on the SEER database, which might have some inherent biases. Second, some potential important parameters and specific information related to prognoses, such as the family history, surgical details, the surgical margin status, vascular invasion, molecular pathologic characteristics and Helicobacter pylori infection status were not included in the SEER database, which might improve predictive ability if incorporated. Third, the majority race of SEER cohort was white and all patients in the Chinese cohort were yellow, therefore the study could be with potential race heterogeneity. Forth, we excluded patients from the study because of missing data, which could have resulted in a potential selection bias. Finally, our nomogram was externally validated in a single cancer center, which requires to be validated by multi-cencer and large samples.