Population-Based Validation of the MIA and MSKCC Tools for Predicting Sentinel Lymph Node Status

Roger Olofsson Bagge ^1,2,3, Rasmus Mikiver ⁴, Michael A Marchetti ⁵

¹Sahlgrenska Center for Cancer Research, Institute of Clinical Sciences, Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden.
²Wallenberg Centre for Molecular and Translational Medicine, University of Gothenburg, Gothenburg, Sweden.
³Department of Surgery, Sahlgrenska University Hospital, Gothenburg, Sweden.
⁴Regional Cancer Center Southeast Sweden and Department of Clinical and Experimental Medicine, Linköping University, Linköping, Sweden.
⁵Dermatology, Skagit Regional Health, Mt Vernon, Washington.
⁶Melanoma Institute Australia, The University of Sydney, Sydney, New South Wales, Australia.
⁷Faculty of Medicine and Health, The University of Sydney, Sydney, New South Wales, Australia.
⁸Gastric and Mixed Tumor Service, Department of Surgery, Memorial Sloan Kettering Cancer Center, New York, New York.
⁹Department of Clinical Sciences, Surgery, Lund University, Lund, Sweden.
¹⁰Department of Surgery, Kristianstad Hospital, Kristianstad, Sweden.
¹¹Tissue Pathology and Diagnostic Oncology, Royal Prince Alfred Hospital and NSW Health Pathology, Sydney, New South Wales, Australia.
¹²Charles Perkins Centre, The University of Sydney, Sydney, New South Wales, Australia.
¹³Department of Plastic Surgery, Westmead Hospital, Sydney, New South Wales, Australia.
¹⁴Department of Surgery, Dartmouth-Hitchcock Medical Center, Lebanon, New Hampshire.
¹⁵Department of Health, Medicine and Caring Sciences, Linköping University, Linköping, Sweden.

⁰Sahlgrenska Center for Cancer Research, Institute of Clinical Sciences, Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden.

Jama Surgery +

|

January 10, 2024

View abstract on PubMed

Summary

Predictive models for sentinel lymph node biopsy (SLNB) in melanoma patients showed similar performance. Clinical utility was demonstrated at higher risk thresholds (≥10%), particularly for T2 melanomas, reducing unnecessary biopsies.

Area Of Science

Oncology
Surgical Oncology
Dermatology

Background

Sentinel lymph node biopsy (SLNB) is crucial for melanoma staging, but patient selection based on positivity risk is key.
Predictive models from Memorial Sloan Kettering Cancer Center (MSKCC) and Melanoma Institute Australia (MIA) aim to improve SLNB selection.
The clinical utility of these predictive models requires validation in diverse patient populations.

Purpose Of The Study

To evaluate the clinical utility of the MSKCC and MIA predictive models for SLNB in cutaneous melanoma.
To compare the performance of these models against a strategy of performing SLNB on all patients.

Main Methods

A population-based study analyzed 10,089 melanoma patients undergoing SLNB from the Swedish Melanoma Registry (2007-2021).
MSKCC and a limited MIA model (mitotic rate, no lymphovascular invasion) were used to predict SLN positivity.
Model performance was assessed using operating characteristics and decision curve analysis.

Main Results

Both models demonstrated good calibration and similar predictive accuracy (AUC: MSKCC 70.8%, MIA 69.7%).
No added benefit was observed at a 5% risk threshold compared to universal SLNB.
Clinical utility, indicated by added net benefit, was found at risk thresholds of 10% or higher.
For T2 melanomas at a 10% threshold, models reduced avoidable SLNBs by 7-8 per 100 patients.

Conclusions

The MSKCC and limited MIA models confirm statistical performance in a large, national dataset.
Model utility for improving SLNB selection over universal biopsy is evident only at risk thresholds of 7% or higher.
Caution is advised when applying these nomograms for SLNB selection at lower risk thresholds, especially for T2 melanomas.