2 Clarke Drive
Suite 100
Cranbury, NJ 08512
© 2024 MJH Life Sciences™ and OncLive - Clinical Oncology News, Cancer Expert Insights. All rights reserved.
The Melanoma Institute Australia nomogram was found to accurately estimate the risk of sentinel node metastasis positivity in patients with melanoma and its use could help to reduce the rate of unnecessary invasive biopsies without sacrificing sensitivity.
The Melanoma Institute Australia (MIA) nomogram was found to accurately estimate the risk of sentinel node metastasis positivity in patients with melanoma and its use could help to reduce the rate of unnecessary invasive biopsies without sacrificing sensitivity, according to results from a study published in the Journal of Clinical Oncology.1
While the Memorial Sloan Kettering Cancer Center (MSKCC) online nomogram model had a predictive accuracy of 67.7% (95% confidence interval [CI], 65.3%-70.0%), the MIA model had a higher rate of 73.9% (95% CI, 71.9%-75.9%); this translated to a 9.2% increase in accuracy comparted with the MSKCC model (P < 0.001).
Of the patients with sentinel node negativity (n = 2748), a biopsy would not have been offered to 22.1% (n = 608) based on the MIA model, 13.4% (n = 359) based on the MSKCC model, and 12.4% (n = 332) based on the National Comprehensive Cancer Network (NCCN) or ASCO/ Society of Surgical Oncology (SSO; T2+) criteria. External validation created a C-statistic of 75.0% (95% CI, 73.2%-76.7%).
“It is notable that the higher accuracy of this nomogram was true in both the Australian and the American cohorts, even though those 2 cohorts had substantially different clinical and pathologic characteristics,” wrote Mark B. Faries, MD, a specialist in Cutaneous (Skin) Medical Oncology, General Surgery, Surgical Oncology at The Angeles Clinic & Research Institute, in an accompanying editorial.2 “This suggests that the tool has a good chance of being useful for many populations throughout the world.”
Data were collected from the MIA research database on patients with primary cutaneous melanomas who had previously received a sentinel node biopsy between January 2003 and December 2014. Collected information included age at diagnosis, site of primary tumor, tumor thickness, presence of ulceration, mitotic rate, Clark level, disease subtype, and sentinel node status.
Investigators only included data from patients who were 18 years of age or older with a tumor thickness of 10.0 mm or less. Several disease subtypes were included, such as superficial spreading melanoma, nodular melanoma, desmoplastic melanoma, lentigo maligna melanoma, and acral melanoma. In total, 3,477 patients fulfilled the study criteria and were included in the final analyses.
“Having applied the existing MSKCC nomogram to the MIA data set using the MSKCC online prediction tool, a new nomogram was developed based on a variable selection procedure in which the same parameters as the MSKCC model and additional parameters were considered and sequentially removed,” the study authors wrote.
The full set of parameters examined included age, tumor thickness, ulceration, Clark level, primary tumor anatomic site, sex, mitotic rate, melanoma subtype, and lymphovascular invasion (LVI). Moreover, covariate candidates for model selection comprised all parameters in the MSKCC nomogram as well as mitotic rate, disease subtype, and LVI, as these variables were considered to potentially be predictive of sentinel node positivity, according to the authors.
Notable statistical differences were observed between the MIA and MD Anderson datasets. Results showed a lower rate of sentinel node positivity (21.0%; n = 729) in the MIA dataset versus the MD Anderson dataset (28.44%; n = 990; P < .001).
Additionally, patients with sentinel node positivity were found to have a greater tumor thickness, have increased likelihood of an ulceration, and have a higher mitotic rate than those with node negativity. Moreover, disease subtype was also found to statistically significantly differ between those who were sentinel node–positive and –negative. Those with sentinel node positivity were also observed to be younger, although this difference was only found to be statistically significant in the MIA dataset.
“The corresponding co-variance matrix of the coefficients was also provided to compute 95% Cis for the estimated risks,” the study authors added.
Predictive risk in the MIA dataset ranged between 0.3% to 91.7%, while the risk in the MD Anderson dataset ranged from 0.3% to 94.9%. Notably, 17.5% of patients in the MIA cohort and 23.7% of those in the MD Anderson cohort had a risk between 5% and 10%.
“Although this nomogram may move a small number of patients to sentinel lymph node 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 of less than 5% of finding metastases,” explained Faries. “This includes not only a substantial number of patients with T1 melanomas, but also 5%-7% of patients with T2 and approximately 2%-3% of patients with T3 primary melanomas.”
Although the MIA sentinel node prediction model showed superior accuracy of status prediction compared to several other models—as was confirmed through both internal and external validation—several caveats must be recognized.
The analysis was limited to patients who had already been selected for a biopsy, although that may not be a critical issue for melanomas that are T2 or greater, according to Faries. Additionally, interobserver variability could prove problematic in terms of the parameters that were added to strengthen earlier models, he added. Although the determination of histologic subtype might be straightforward in the majority of cases, differentiation can prove more challenging in others such as desmoplastic melanomas, according to Faries.
Lastly, although the model demonstrated positive consideration of the entire tumor thickness spectrum, the CIs it yields within thin melanomas may not be enough to resolve any clinical uncertainties in that subset.
“Although a definite step forward, this model is unlikely to be the final word on predicting sentinel lymph node metastases,” Faries concluded. “The genomic revolution appears to be coming to staging in melanoma with the advent of gene expression profiling. While the currently available data are not sufficiently granular to know where they will fit in guiding the management of patients, there is clearly a possibility that they will prove useful.”