Diagnostic performance of gadofosveset-enhanced axillary MRI for nodal (re)staging in breast cancer patients: results of a validation study

Highlights

• GDF-MRI can differentiate between benign and malignant axillary lymph nodes.

• A learning curve for reading GDF-MRI was observed for one reader.

• After the assessment of 176 nodes, NPV of GDF-MRI increased to 92%.

• Prevalence of lymph node metastases was 28.5% in this cohort.

Aim

To evaluate diagnostic performance of gadofosveset (GDF)-enhanced magnetic resonance imaging (MRI) in addition to T2-weighted (T2W) MRI for nodal (re)staging in newly diagnosed breast cancer patients.

Materials and methods

Ninety patients underwent axillary T2W- and GDF-MRI. Two radiologists independently scored each lymph node; first on T2W-MRI, subsequently adjusting their score on GDF-MRI. Diagnostic performance parameters were calculated on node-by-node and patient-by-patient validation with histopathology as the reference standard. Furthermore, learning curve analysis for reading GDF-MRI was performed.

Results

In patient-by-patient validation, overall reader performances for T2W- and GDF-MRI were similar with area under the receiver operating characteristic curves (AUC) of 0.75 and 0.77 (p=0.731) for reader 1 and 0.79 and 0.72 (p=0.156) for reader 2. For node-by-node validation, AUC values of T2W- and GDF-MRI were 0.76 and 0.82 (p=0.018) and 0.77 and 0.77 (p=0.998) for reader 1 and 2. The AUC for reader 1 was 0.71 for first one-third of nodes evaluated, improving to 0.80 and 0.95 for the next and last one-third, respectively. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) improved from 38%, 89%, 56%, and 79% to 60%, 93%, 64%, and 92%. The AUC of reader 2 improved from 0.69 to 0.79.

Conclusion

The present study confirmed that GDF-MRI, in addition to T2W-MRI, has potential as a non-invasive method for nodal (re)staging in breast cancer.

Introduction

Breast cancer is one of the most frequently diagnosed cancers among women, with a worldwide incidence rate of 1.7 million.¹ Survival rates have increased in the last decades due to improved diagnostic techniques and treatment regimens. Five-year survival rates are up to 98% for early-stage, lymph-node-negative breast cancer.² Consequently, more attention is afforded to maintaining quality of life by limiting overtreatment and its associated lifetime morbidity.

For a long time, axillary lymph node dissection (ALND) was routinely performed to assess nodal status in breast cancer patients. This procedure is associated with significant short- and long-term morbidity. Seroma, lymphoedema, nerve injury, and reduced shoulder function are reported in up to 49% of the patients after 3 years of follow-up.3, 4, 5 About 15 years ago, sentinel lymph node biopsy (SLNB) became the standard procedure in clinically node-negative patients. It was followed by a completion ALND in cases of positive sentinel lymph node(s).⁶ Nevertheless, SLNB remains an invasive procedure with short-term side effects in 25% of patients and long-term morbidity (for example, lymphoedema) in up to 6% of the patients.4, 7, 8

As 74% of the sentinel lymph nodes show no metastasis, a non-invasive imaging technique able to identify these node-negative patients would be a step forward in personalised treatment.⁹ It would result in a significant reduction of morbidity and increase the patient's quality of life, while maintaining the high survival rates achieved so far.

An accurate non-invasive nodal staging tool could also guide treatment in clinically node-positive breast cancer patients. In cases of neoadjuvant systemic therapy, pathological complete response of axillary lymph node metastases is achieved in approximately 37%.¹⁰ Identifying these patients using a non-invasive imaging technique, and thereby, avoiding ALND, would further reduce breast cancer treatment-related morbidity.

In rectal cancer, the use of magnetic resonance imaging (MRI) with a blood-pool contrast agent, gadofosveset (GDF), improved diagnostic performance for lymph node staging.11, 12 Consequently, a feasibility study of Schipper et al. on the use of GDF-enhanced MRI for axillary nodal staging in 10 breast cancer patients showed a sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of 86%, 94%, 75% and 97%, respectively. They concluded that GDF-enhanced MRI is a promising tool to accurately detect node-negative breast cancer patients.¹³

The aim of this study was to prospectively assess the diagnostic performance of GDF-enhanced MRI for nodal (re)staging in a larger population of newly diagnosed breast cancer patients using histopathology as reference standard.