Behavioral phenotyping of cancer pain in domesticated cats with naturally occurring squamous cell carcinoma of the tongue: initial validation studies provide evidence for regional and widespread algoplasticity

Abstract

Feline oral squamous cell carcinoma (FOSCC) is a common and naturally occurring condition that recapitulates many features of human head and neck cancer (HNC). In both species, there is need for improved strategies to reduce pain caused by HNC and its treatment. Research to benefit both species could be conducted using pet cats as a comparative model, but this prospect is limited by lack of validated methods for quantifying FOSCC-associated pain. A prospective non-randomized pilot study was performed for initial validation of: (1) a pet owner administered quality of life questionnaire and visual assessment scoring tool (FORQ/CLIENT); (2) a clinician assessment questionnaire (UFEPS/VET); (3) electronic von Frey testing [EVF]; and (4) Cochet-Bonnet (COBO) aesthesiometry. To assess intra-rater reliability, discriminatory ability, and responsiveness of each assay, 6 cats with sublingual SCC and 16 healthy control cats were enrolled. The intra-rater reliability was moderate-to-good for the clinical metrology instruments and EVF (intraclass correlation coefficient [ICC] ≥ 0.68), but poor for COBO (ICC = 0.21). FORQ/CLIENT scores were higher (worse quality of life) in FOSCC cats vs healthy controls. The internal reliability of FORQ/CLIENT scoring was high (Cronbach α = 0.92); sensitivity and specificity were excellent (100% when using cut-offs determined using receiver operating characteristic [ROC] curves). For the FORQ/CLIENT, there was strong and inverse correlation between scores from the questions and visual assessment (r =  − 0.77, r2 = 0.6, P < 0.0001). For the UFEPS/VET, Cronbach’s α was 0.74 (high reliability). Sensitivity and specificity were 100% and 94%, respectively, when using a cut-off score (3.5) based on ROC curves (Youden index of 0.94). Total UFEPS/VET scores were positively correlated with FORQ/CLIENT scores (r2 = 0.72, P < 0.0001). Sensitivity of EVF and COBO ranged from 83 to 100% and specificity ranged from 56 to 94%. Cats with cancer were more sensitive around the face (lower response thresholds) and on the cornea (longer filament lengths) than control animals (P < 0.03). Reduced pressure response thresholds were also observed at a distant site (P = 0.0002) in cancer cats. After giving buprenorphine, EVF pressure response thresholds increased (P = 0.04) near the mandible of cats with OSCC; the length of filament required to elicit a response in the COBO assay also improved (shortened; P = 0.017). Based on these preliminary assessments, the assays described herein had reasonable inter-rater reliability, and they were able to both discriminate between cats with and without oral cancer, and respond in a predictable manner to analgesic therapy. In cats with tongue cancer, there was evidence for regional peripheral sensitization, and widespread somatosensory sensitization. These results provide a basis for multi-dimensional assessments of pain and sensitivity in cats with oral SCC.