Validating the SumMean (18)F-FDG PET Textural Feature as a Prognostic Marker in an Independent Cohort of Locally Advanced Non-Small Cell Lung Cancer Patients Undergoing Concurrent Chemoradiotherapy

Lung Cancer

Pract Radiat Oncol. 2020 Oct 19:S1879-8500(20)30255-1. doi: 10.1016/j.prro.2020.10.004. Online ahead of print.


BACKGROUND: Analyses from the XXXX trial data identified the SumMean textural feature, calculated from 18F-FDG-PET for tumors with metabolic tumor volume (MTV) >93cm3, as a predictor of overall survival (OS) for patients with locally advanced non-small cell lung cancer (LA-NSCLC) receiving concurrent chemoradiotherapy. Here we validated that finding in a completely independent patient cohort from a single institution.

METHODS: We identified LA-NSCLC patients who underwent staging 18F-FDG PET and received definitive chemoradiotherapy at our institution between 2007 and 2018. Primary tumors were segmented semi-automatically, and SumMean score was calculated for each tumor and categorized according to the previously proposed cutoff of 0.018. In patients with MTV >93cm3, SumMean was evaluated as a predictor of progression-free survival (PFS) and OS using logrank and Cox proportional hazards testing.

RESULTS: 148 patients met inclusion criteria and 34 had large tumors (>93cm3). Twelve (35%) had high SumMean, and 22 (65%) had low SumMean. SumMean was not significantly associated with other clinical variables. Median PFS for patients with large tumors and low SumMean was 5.8 months, compared to 41.1 months for patients with large tumors and high SumMean (logrank p=0.022). Median OS for patients with large tumors and low SumMean was 15.0 months and not reached for patients with large tumors and high SumMean (logrank p=0.014). In multivariable analysis, high SumMean was an independent predictor of improved OS (HR=0.26, 95%CI: 0.07-0.94, p=0.041) and PFS (HR=0.30, 95%CI: 0.10-0.86, p=0.026).

CONCLUSIONS: We externally validated SumMean as a prognostic marker for LA-NSCLC patients treated with chemoradiotherapy in an independent patient cohort. Future studies will explore potential mechanisms for this association and how textural features may help guide treatment decisions.