Re: Radio-Guided Lymph Node Mapping in Bladder Cancer Using SPECT/CT and Intraoperative γ-Probe Methods
We have read with great interest the article by Połom et al that was recently published in the Clinical Nuclear Medicine. They investigated the concept of sentinel node mapping in 38 patients with bladder cancer. Although sentinel node mapping has been successfully applied to urological and gynecological cancers,1–5 it has only been evaluated for bladder cancer in a few studies. So, the authors should be really commended as their study is meticulously designed and thus far is one of the largest studies on sentinel node mapping in bladder cancer.6
However, Połom et al had several inaccuracies in reporting the results, which is not uncommon in sentinel node mapping diagnostic reports.7,8
Sentinel node mapping was successful in 35 patients. Lymph node metastases were found in 15 patients. The authors reported two false-negative cases among these 15 patients. The correct definition of false-negative rate, sensitivity, negative predictive value, and accuracy in sentinel node mapping are as follows8:
So, the following results can be calculated for the Połom et al study: false-negative rate = 2/15 (13.3% not 8% as reported by the authors), sensitivity = 86.7% not 92.3% as reported by the authors.
In addition, positive predictive value is irrelevant in sentinel node mapping studies as it is always 100% (no false-positive case is possible in sentinel node mapping).8 So, 35% positive predictive value as reported by authors is not correct. Actually, the negative predictive value should be calculated for the Połom et al study. Negative predictive value = 20/22 = 90.9%. Finally accuracy = (13 + 20)/35 = 94.2% not 34% as reported by the authors.
In our opinion, Połom et al report is a seminal study in the field of sentinel node mapping in bladder cancer and their meticulous design should be commended. However, the authors should have been more careful in reporting their results as the numbers they provided are somehow misleading.