An improved risk prediction tool for melanoma spread will reduce the number of patients undergoing unnecessary sentinel node biopsies, its Australian developers say.
The Melanoma Institute of Australia (MIA) has created a 6-item risk prediction tool for sentinel node positivity to help guide clinician-patient discussions around the likely spread of primary melanoma and need for sentinel node biopsy (SNB).
The new SNB risk calculator – freely available online – builds on and improves the accuracy of the 15-year old risk calculator from Memorial Sloan Kettering Cancer Center (MSKCC).
The items in the new model are patient age, tumour thickness, ulceration, mitotic rates, melanoma subtype and lymphovascular invasion.
Patient age, tumour thickness, and ulceration were retained from the MSKCC nomogram while body site and Clark level were removed.
The two prediction tools were compared in an Australian population of 3,500 patients from the research database at the Melanoma Institute Australia (MIA).
The analysis, published in the Journal of Clinical Oncology, found the new MIA model was more specific and more sensitive than the original calculator from MSKCC.
“This improvement was substantial, with a 9.2% increase in accuracy,” the study authors said.
The MIA risk calculator was subsequently externally validated in another US data set and has been trialed at sites including Royal Prince Alfred, Royal North Shore and Westmead Hospitals.
“Importantly, the number of patients undergoing unnecessary sentinel node biopsy would be significantly reduced compared with use of the MSKCC nomogram or the NCCN or ASCO/SSO guidelines, without losing sensitivity,” the study authors concluded.
Dr Alexander Varey, a staff specialist in plastic surgery at Westmead and with affiliations at MIA and the University of Sydney, told the limbic the calculator complements existing guidelines.
“Most of these guidelines recommend trying to do biopsies on patients with ≥10% risk of a positive node. What this prediction tool does is give a more accurate quantification of the patient’s individual risk rather than an average risk dependent on melanoma thickness.”
“This gives a much more nuanced estimate,” he said.
“In this context I think giving the right people the SNB is the key. Don’t do them unnecessarily and make sure you do them when it is necessary.”
He said some patients will still want the extra procedure even with a low probability of finding positive nodes.
While SNB was usually conducted at the same time as wide local excision, it increased the complexity of the dual procedures, the level of anaesthesia required and the potential risk of fairly minor complications.
Dr Varey said nodes could sometimes be tricky to access and there can be multiple sentinel nodes – for example, where a melanoma in the middle of the back could drain to either armpit, the neck and the groin.
“There are costs to the patient and the health service so where we can avoid doing procedures that are unlikely to be of benefit to the patients, we avoid them.”
He said clinicians in Australia and in the US where the risk calculator had been validated would be the first to adopt it.
An accompanying editorial in JCO by Dr Mark Faries, from the Cedars-Sinai Medical Center in the US, said the tool has a good chance of being useful for many populations throughout the world.
“The authors have created an online resource, which is free to use and can bring the risk calculator to smartphones everywhere,” he said.
“Although this nomogram may move a small number of patients to SLN biopsy recommendation who would have previously had their nodes observed, it seems that the most apparent effect of this improved accuracy is in sparing the procedure for patients who have a low probability (<5%) of finding metastases.”
It concluded that the tool “raises the bar that any new algorithm, including gene expression profiling, will have to surpass to demonstrate added value.”