Statistical approach to mediastinal staging in NSCLC with M.E.S.S.i.a. software

Area of Science:

• Oncology
• Medical Informatics
• Diagnostic Imaging

Background:

• Accurate staging of mediastinal lymph nodes is crucial for non-small cell lung cancer (NSCLC) prognosis and treatment decisions.
• Current guidelines may not fully address the probabilistic assessment of lymph node involvement.
• The M.E.S.S.i.a. software aims to provide a more refined probability of mediastinal metastasis.

Purpose of the Study:

• To evaluate the accuracy of the M.E.S.S.i.a. software in predicting mediastinal lymph node involvement in NSCLC patients.
• To compare the software's calculated probabilities with pathological staging results.
• To assess the potential of M.E.S.S.i.a. to reduce invasive staging procedures.

Main Methods:

• Retrospective analysis of 108 NSCLC patients.
• Comparison of pathological staging with M.E.S.S.i.a. software's calculated post-test probabilities.
• Receiver Operating Characteristic (ROC) curve analysis to determine accuracy (Area Under Curve - AUC).

Main Results:

• M.E.S.S.i.a. showed high accuracy with an AUC of 0.773 (p<0.001).
• In 39% of cases with calculated probability <8%, no N2/N3 metastasis was found (NPV 100%).
• The software helped avoid invasive procedures in 12 out of 41 eligible cases.

Conclusions:

• Preliminary data suggest M.E.S.S.i.a. software is highly accurate in excluding N2/N3 lymph node involvement in NSCLC.
• The software shows potential for improving clinical decision-making and reducing unnecessary invasive procedures.
• A prospective multicenter study is underway to validate the calculator's performance across different probability thresholds.