Premium
A statistical tool for risk assessment as a function of the number of lymph nodes retrieved from rectal cancer patients
Author(s) -
Wu Z.,
Qin G.,
Zhao N.,
Jia H.,
Zheng X.
Publication year - 2018
Publication title -
colorectal disease
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.029
H-Index - 89
eISSN - 1463-1318
pISSN - 1462-8910
DOI - 10.1111/codi.14264
Subject(s) - medicine , colorectal cancer , stage (stratigraphy) , lymph node , confidence interval , occult , t stage , cancer , surgery , pathology , paleontology , alternative medicine , biology
Aim Although a minimum of 12 lymph nodes (LNs) has been recommended for examination in colorectal cancer patients there remains considerable debate with regard to rectal cancer. Inadequacy of examined LNs could lead to understaging and inappropriate treatment as a consequence. We describe a statistical tool that allows an estimate of the probability of false‐negative nodes. Method A total of 26 778 patients diagnosed between 2004 and 2013 with rectal adenocarcinoma [tumour stage (T stage) 1–3] who did not receive neoadjuvant therapies and had at least one histologically assessed LN were extracted from the Surveillance, Epidemiology and End Results (SEER) database. A statistical tool using beta‐binomial distribution was developed to estimate the probability of missing a positive node as a function of the total number of LNs examined and T stage. Results The probability of falsely identifying a patient as node‐negative decreased with increasing number of nodes examined for each stage. It was estimated to be 72%, 66% and 52% for T1, T2 and T3 patients, respectively, with a single node examined. To confirm an occult nodal disease with 90% confidence, 5, 9 and 29 nodes need to be examined for patients from stages T1, T2 and T3, respectively. Conclusion The false‐negative rate of the examined LNs in rectal cancer was verified to be dependent preoperatively on the clinical T stage. A more accurate nodal staging score was developed to recommend a threshold for the minimum number of examined nodes with regard to the favoured level of confidence.