Authors' reply to Rayter.

Rayter raises concerns about our editorial on intraoperative assessment of axillary nodes.1 2 He believes that the meta-analysis of one step nucleic acid amplification (OSNA) is seriously flawed, that it contains a mathematical error, and that the positive predictive value (PPV) should be 0.86 not 0.79.3 The PPV of 0.79 is statistically correct. When calculating the PPV of a relatively rare event its prevalence must be taken into account.4 He has not done this. The PPV of OSNA was even lower in a paper by Rayter’s group. Of 30 nodes classified by OSNA as positive, just 13 had histological evidence of macrometastases.5 Rayter argues that the meta-analysis used the wrong mathematical model to translate tumour volume into a diameter,3 but the authors confirmed that they modelled using both a sphere and a cube.

Rayter argues that only 26 UK centres use OSNA, so it may not be the most widely used technique. A survey of surgeons at the 2014 Association of Breast Surgeons meeting identified OSNA as the only technique in routine use. Rayter asserts that we quoted 4700 copy numbers as the lower limit for the diagnosis of macrometastasis. We did not; we said that it varied between 4700 and 140 000. Rayter is correct that the National Institute for Health and Care Excellence (NICE) approved OSNA in 2013 not 2011: our reference was dated 2013.2

Rayter justifies the 21% false positive rate of OSNA by citing the 7% false negative rate for sentinel node biopsy. In a study of 2001 women with a negative sentinel node biopsy, only nine (0.45%) had developed axillary recurrence by 10 years.6 These and other results fully justify the incorporation of sentinel node biopsy into routine clinical practice. By contrast, OSNA exposes up to 21% of women with axillary node involvement to surgery that may not be beneficial, a 50% rate of damage to the intercostobrachial nerve, and a 40% rate of lymphoedema at one year.7

Rayter accepts that routine axillary clearance can be omitted safely in patients with a low tumour burden. Randomised trials show that patients with micrometastasis do not benefit from axillary dissection,6 8 but OSNA cannot accurately differentiate micrometastasis from macrometastasis.3 4 As the invited clinical lead for the NICE guidelines, Rayter should encourage NICE to re-evaluate its guidelines on OSNA urgently. None of Rayter’s criticisms makes us doubt our conclusion that routine intraoperative sentinel lymph node assessment should be abandoned forthwith.