The accumulation of peritoneal ascites is a hallmark of epithelial ovarian cancer (EOC), which remains the deadliest gynaecological malignancy. Despite the availability of several treatments, there is currently no test to assess which patients will respond to first-line platinum-based chemotherapy prior to administration. The contributions of multicellular ascitic spheroids to chemoresistance and disease progression have prompted the use of three-dimensional (3D) models of EOC to better understand treatment response. While artificial spheroids can be generated from primary cells, they are slow to generate and lack the cellular diversity of native spheroids that can be collected directly from ascites prior to cytoreductive surgery or through routine paracentesis. Furthermore, few proteomics studies have been performed on primary tumour samples to investigate potential markers of chemoresponse. This preliminary study compares the ex vivo carboplatin responses of artificial spheroids to those of chemotherapy-naïve native spheroids and demonstrates that native spheroids are a representative and rapidly-available 3D model for ex vivo treatment response assessments. Through mass spectrometry-based proteomic analyses of artificial and native spheroids, we have identified over 6600 protein groups that reveal vast differences in key cellular processes. We show that only the global proteomes of patient-matched native spheroids correlate with their ex vivo treatment responses and contain molecular markers that are predictive of first-line treatment response. This proteomics method could form the basis of future screening strategies for the personalisation of EOC treatment in a clinically relevant timeframe.