Objective: To determine the predictive value of the symptoms of lumbar back pain for extrapancreatic neural plexus invasion in pancreatic cancer. Methods: The clinical and imaging data of 130 patients diagnosed as pancreatic cancer were retrospectively analyzed. Whether or not there was extrapancreatic neural plexus invasion was judged by CT in all patients, and the difference of the prediction of symptoms of lumbar back pain for extrapancreatic neural plexus invasion in pancreatic cancer and CT determination were analyzed. Results: Seventy-seven patients (59.2%) were judged to have extrapancreatic neural invasion by CT scan, wherein 38 cases (50.7%, 38/75) had cancer of the pancreatic head, 39 cases (70.9%, 39/55) had cancer of the pancreatic body and tail, and the percentage of the latter was significantly higher than that of the former (χ2=78.999, P=0.000); 87.0% of patients had tumor invasion around the celiac axis, which was significantly higher than those with tumor invasion in left celiac ganglion (57.1%) and right celiac ganglion (45.5%) (χ2=30.415, P=0.000). Of the 48 patients in lumbar back pain group, 47cases (97.9%) were diagnosed with extrapancreatic neural plexus invasion by CT scan. Compared with CT diagnosis, the sensitivity, specificity and positive predictive value of lumbar back pain for predicting extrapancreatic neural plexus invasion was 61%, 98.1% and 97.9%, respectively. Results of consistency check showed that there was a moderate consistency between the two methods (κ=0.545, P=0.000). Conclusion: There is a relationship between symptoms of lumbar back pain and extrapancreatic neural plexus invasion in pancreatic cancer, and lumbar back pain and CT image features have complementary roles in preoperative prediction of extrapancreatic neural plexus invasion in pancreatic cancer.