Prospective imaging assessment of mortality risk after head-and-neck radiotherapy

Abstract

Purpose: The optimal roles for imaging-based biomarkers in the management of head-and-neck cancer remain undefined. Unresolved questions include whether functional or anatomic imaging might improve mortality risk assessment for this disease. We addressed these issues in a prospective institutional trial.

Methods and materials: Ninety-eight patients with locally advanced pharyngolaryngeal squamous cell cancer were enrolled. Each underwent pre- and post-chemoradiotherapy contrast-enhanced computed tomography (CT) and (18)F-fluorodeoxyglucose (FDG)-positron emission tomography (PET)/CT imaging. Imaging parameters were correlated with survival outcomes.

Results: Low post-radiation primary tumor FDG avidity correlated with improved survival on multivariate analysis; so too did complete primary tumor response by CT alone. Although both imaging modalities lacked sensitivity, each had high specificity and negative predictive value for disease-specific mortality risk assessment. Kaplan-Meier estimates confirmed that both CT and FDG-PET/CT stratify patients into distinct high- and low-probability survivorship groups on the basis of primary tumor response to radiotherapy. Subset analyses demonstrated that the prognostic value for each imaging modality was primarily derived from patients at high risk for local treatment failure (human papillomavirus [HPV]-negative disease, nonoropharyngeal primary disease, or tobacco use).

Conclusions: CT alone and FDG-PET/CT are potentially useful tools in head-and-neck cancer-specific mortality risk assessment after radiotherapy, particularly for selective use in cases of high-risk HPV-unrelated disease. Focus should be placed on corroboration and refinement of patient selection for imaging-based biomarkers in future studies.

Figure 1. Survival Characteristics

Kaplan-Meier plots for overall (grey) and disease-specific (black) survival are shown for the entire patient cohort. Ticks mark censored events.

Figure 2. Accuracy of Imaging Response as a Predictor of Disease-Specific Survival

ROC curves are shown for the accuracy of FDG-PET/CT imaging (lines) and CT imaging (symbols) in predicting disease-specific survival in (A) the entire cohort (n = 98) and in (B) the subset censored for DM as first site of failure (n = 92). The line of unity is displayed for comparison.

Figure 3. Imaging Biomarkers of Disease-Specific Survival

Kaplan-Meier curves for the entire cohort subdivided by post-radiation SUVmax threshold values for primary tumors (A) and nodes (C), and by radiographic CT response for primary tumors (B) and nodes (D). FDG-PET/CT and CT both discriminate survivorship groups by primary tumor biomarker (log rank, P < 0.005), but not by nodal tumor biomarker (log rank, P = NS).

Figure 4. Locoregional Failure Risk Grouping and Survival

Kaplan-Meier curves are shown for the entire cohort (A) stratified by locoregional radiotherapy failure risk grouping (high-risk [n = 32]: human papilloma virus (HPV)-negative disease, tobacco use, or non-oropharyngeal primary disease; low-risk [n = 66]: HPV-positive disease or those with oropharyngeal primary tumors and no tobacco use). The high-risk group was then sub-stratified by primary tumor imaging biomarker risk status – FDG-PET/CT-based (B) or CT-based (C). P values are derived from log-rank analyses.

Figure 5. Proposed Clinicoradiographic Risk-Stratification Algorithm

Clinical low-risk status (HPV-positive disease or those with oropharyngeal primary tumors and no tobacco use) predicts for excellent survival regardless of imaging findings. Post-radiotherapy imaging of primary disease treatment response improves stratification of mortality risk in patients with a high-risk clinical profile (HPV-negative disease, tobacco use, or non-oropharyngeal primary disease).