The Christie score for post-treatment response assessment PET/CT in patients with head and neck squamous cell carcinoma: a safe and simple scoring system

Background: Radiotherapy with or without concurrent chemotherapy is a standard of care treatment for patients with head and neck squamous cell carcinoma (HNSCC). Upon completion, patients are referred for a post-treatment F-18-FDG PET/CT (Fluorine-18 fluorodeoxyglucose positron emission tomography/computed tomography) scan to help guide ongoing management by assessing for the presence or absence of residual or recurrent disease and differentiating this from post-treatment inflammation. To improve objective reporting of response, we developed the Christie score. The study aims to assess the validity of the Christie score as a response assessment tool in patients with HNSCC and to compare its performance against the widely used Hopkins score. Methods and materials: All newly diagnosed head and neck cancers between July 2018 and July 2020 were retrospectively reviewed. In total, 291 patients (224 men and 67 women) were included in the study. Patients with squamous cell carcinoma of the nasopharynx, oropharynx or oral cavity, hypopharynx or larynx were included. All other cell lineages or anatomical locations were excluded. Hopkins and Christie scores were applied to post-treatment PET/CT, and sensitivity, specificity, positive predictive value, negative predictive value, and likelihood ratio assessed for each score. Fisher's exact tests and receiver-operating characteristic (ROC) curves were used to determine the ability of the Hopkins and Christie scores to differentiate residual or recurrent disease from treatment response. p values < 0.05 were considered to indicate statistical significance. Results: 39 patients (13%) were confirmed to have residual or recurrent disease. This was significantly more likely in patients with positive Hopkins (p < 0.0001) and Christie (p < 0.0001) scores. The Christie score has a higher sensitivity (92% vs. 85%) and negative predictive value (99% vs. 97%) compared to Hopkins, though the differences were not statistically significant. Comparison of the ROC curves for the Hopkins and Christie score revealed no significant difference between the two scores' ability to discriminate patients with residual or recurrent disease from cases where disease is absent (p = 0.382). 'Subjectivity rates' of the 291 patients are as follows. Six patients (2.1%; 95% CI 0.76-4.5%) were assigned borderline scores on the Hopkins criteria, compared to only a single patient (0.3%; 95% CI 0-1.9%) on the Christie criteria. The 'subjectivity rate' difference is 0.017 (95% CI - 0.06 to 3.5%; p = 0.06) and the ratio is 6.0 (95% CI 0.73-276; p = 0.07). Conclusion: Our study identifies three clear results: (a) the Christie score is an excellent treatment follow-up assessment tool; (b) it is comparable with current gold standard methodology showing no statistically significant differences in performance when compared with the Hopkins score; and (c) there was a lower rate of observer variation when using the Christie score, which is trending towards significance.