High-sensitivity multicancer detection of stage 1 cancer in dogs.

Abstract

OBJECTIVE: To evaluate an orthogonal test that uses antibodies in a small blood sample to sensitively detect stage 1 tumors, when standard treatment can save lives. METHODS: Sera were studied from 283 dogs with stage 1 hemangiosarcoma (30), lymphoma (34), mast cell tumors (60), osteosarcoma (41), and soft tissue sarcoma (49) and dogs established to be cancer-free (69). Samples were applied to microarrays containing peptides synthesized on silicon wafers using photolithography and tert-butoxycarbonyl chemistry. Serum immunoglobulin G binding was measured. RESULTS: Peptides with antibody-binding activities associated with 1 of the 5 cancers or the cancer-free group were identified. Selections were used to build 2 multiclass models. Test performance was verified by peptide resampling or sample holdouts. A simple model detected the 5 different stage 1 tumors at sensitivities from 68% to 98%; the complex model provided stage 1 sensitivities from 60% to 88%, both at high specificities. CONCLUSIONS: Antibody activities to stage 1 tumors can be sensitively detected on peptide microarrays. Two divergent classifier-building approaches yielded similar test results. A field study is needed to validate findings. CLINICAL RELEVANCE: This blood-based test may address the challenges in detecting stage 1 canine cancers, creating opportunities for improved treatment outcomes.