BACKGROUND: The aim of the current study was to identify the minimum number and the optimal range of lymph nodes (LNs) to be examined among patients with gallbladder cancer (GBC). METHODS: Between January 1, 2004, and December 31, 2015, patients with a diagnosis of GBC were identified in the National Cancer Database. A machine-based learning approach was used to identify the minimum number and range of LNs to evaluate relative to long-term outcomes. RESULTS: Among 6531 patients with GBC, median number of LNs evaluated was 2 (IQR:1-5); only 21.1% (n = 1376) of patients had 6 or more LNs evaluated. The median number of metastatic LNs was 0 (IQR: 0-1). On multivariable analysis, evaluation of < 4 LNs was associated with a higher hazard of death (referent 4-7 LNs: < 4 LNs, HR = 1.27, 95% CI, 1.16-1.40; P < 0.001), whereas, patients who had 4 to 7 LNs and > 7 LNs evaluated had comparable long-term mortality risk (HR = 1.10, 95%CI, 0.98-1.24; P = 0.11). There was no difference in the proportion of patients who had at least one metastatic LN identified per T category based on total number of nodes resected (all P > 0.05). CONCLUSION: The overwhelming majority of patients did not have the American Joint Committee on Cancer (AJCC) recommended 6 total LN count . A machine-based learning approach identified evaluation of 4 to 7 LNs as the LN number associated with optimal staging and survival. While obtaining 6 LNs may be challenging, evaluation of at least 4 LNs may be a more appropriate threshold as this cut-off value was associated with optimal patient outcomes and staging.
BACKGROUND: The aim of the current study was to identify the minimum number and the optimal range of lymph nodes (LNs) to be examined among patients with gallbladder cancer (GBC). METHODS: Between January 1, 2004, and December 31, 2015, patients with a diagnosis of GBC were identified in the National Cancer Database. A machine-based learning approach was used to identify the minimum number and range of LNs to evaluate relative to long-term outcomes. RESULTS: Among 6531 patients with GBC, median number of LNs evaluated was 2 (IQR:1-5); only 21.1% (n = 1376) of patients had 6 or more LNs evaluated. The median number of metastatic LNs was 0 (IQR: 0-1). On multivariable analysis, evaluation of < 4 LNs was associated with a higher hazard of death (referent 4-7 LNs: < 4 LNs, HR = 1.27, 95% CI, 1.16-1.40; P < 0.001), whereas, patients who had 4 to 7 LNs and > 7 LNs evaluated had comparable long-term mortality risk (HR = 1.10, 95%CI, 0.98-1.24; P = 0.11). There was no difference in the proportion of patients who had at least one metastatic LN identified per T category based on total number of nodes resected (all P > 0.05). CONCLUSION: The overwhelming majority of patients did not have the American Joint Committee on Cancer (AJCC) recommended 6 total LN count . A machine-based learning approach identified evaluation of 4 to 7 LNs as the LN number associated with optimal staging and survival. While obtaining 6 LNs may be challenging, evaluation of at least 4 LNs may be a more appropriate threshold as this cut-off value was associated with optimal patient outcomes and staging.
Authors: Kota Sahara; Diamantis I Tsilimigras; Shishir K Maithel; Daniel E Abbott; George A Poultsides; Ioannis Hatzaras; Ryan C Fields; Matthew Weiss; Charles Scoggins; Chelsea A Isom; Kamran Idrees; Perry Shen; Itaru Endo; Timothy M Pawlik Journal: J Surg Oncol Date: 2020-01-06 Impact factor: 2.885