Circulating heat shock protein 90 alpha (Hsp90α) as a biomarker for cancer cachexia.

12042 Background: Cancer cachexia is defined clinically by a consensus definition of ≥5% unintentional weight loss in cancer patients over previous 6 months. There is a need to develop blood-based biomarkers for more accurately identifying patients with cancer cachexia. Tumor-cell derived Hsp90α has been shown to be a key driver for muscle wasting and systemic inflammation in pre-clinical models of cancer cachexia. Methods: We conducted a prospective case control study between April 2020 and February 2025 to identify biomarkers of cancer cachexia. Cases were newly diagnosed patients with stage III/IV cancers with ≥5% unintentional weight loss and serum albumin of < 3.5g/dl (Cachexia group). Controls were newly diagnosed patients with stage I-III cancers without weight loss and serum albumin of ≥3.5g/dl (non- cachexia group). We collected 5ml of blood at first visit and analyzed plasma Hsp90α levels using Enzyme linked immunosorbent assay (ELISA). We performed univariable comparisons of geometric means of Hsp90α between the groups using two sample t-test . To evaluate the classification performance of Hsp90α for identifying cancer cachexia, we used receiver operating characteristic (ROC) analysis with area under the curve (AUC). Results: A total of 200 patients were consented and 171 were included in final analysis. There were 110 cases and 61 controls. Mean age of cases was 68 years and of controls 62 years (p=0.003). 64% of cases were males as opposed to 30% in controls (p <0.001). There was no significant difference in race between the groups. Cancer types included lung, gastro-intestinal, breast and other cancers in both cases and controls. Mean plasma Hsp90α at diagnosis was significantly higher in cases 40.60 ng/mL than in controls 32.47 ng/mL (p < 0.001). On up to one year follow up (6 visits total) the mean Hsp90α remained significantly elevated in cases as compared to controls (p <0.001). Mean Hsp90α levels among cases was highest in patients with GI cancers (49.98 ng/mL) followed by Lung cancer (40.67 ng/mL) and other cancer types (27.92 ng/mL), p <0.001. Based on ROC analysis, Hsp90α level of ≥35 ng/mL has AUC of 94% for differentiating cancer patient with cachexia from those without cachexia. When stratified by sex, Hsp90α ≥35 ng/mL remained significantly higher in cases than in controls for both men and women (p<0.001). Conclusions: Hsp90α is a blood-based biomarker that is significantly elevated in cancer patients with cachexia as opposed to those without cachexia. Plasma Hsp90α ≥35 ng/mL accurately identifies cancer patients with cachexia and represents a promising biomarker for cancer cachexia both in clinical care and research. Distribution of Hsp90α amongst case-control based on optimal cutoff (≥35 ng/mL). Variable Study group Cutoff category Number of patients Hsp90 Case < 35 20 Hsp90 Case ≥ 35 90 Hsp90 Control < 35 61 Hsp90 Control ≥ 35 0