AI-Driven Histological Detection of Cerebral Microhemorrhages: Assessing the Therapeutic Potential of the Mitochondria-Targeted Antioxidant SS-31

Objective: Cerebral microhemorrhages (CMHs) contribute to vascular cognitive impairment and dementia (VCID), with aging and hypertension being key risk factors. Mitochondrial oxidative stress is a hallmark of cerebrovascular aging, leading to endothelial dysfunction. We tested whether mitochondrial oxidative stress increases CMH susceptibility and evaluated the mitochondria-targeted antioxidant SS-31 in an aged, hypertensive mouse model. Hypothesis: We hypothesized that SS-31 treatment would reduce CMH burden by reducing mitochondrial oxidative stress and enhancing cerebrovascular resilience. Methods: Aged hypertensive C57BL/6 mice (n=3) were treated with SS-31 (10 mg/kg/day, i.p. ) for 10 days before and during hypertension induction via angiotensin II and L-NAME administration. CMHs were detected in diaminobenzidine (DAB)-stained brain sections using a newly developed, high-throughput, semi-automatic and machine-learning (ML)-based image analysis pipeline. The semi-automated pipeline integrated color space transformation while the ML-based used a supervised random forest algorithm to generate 3D reconstructions. Results: SS-31 treatment did not significantly reduce hypertension-induced CMH burden. Normalized CMH area was similar between SS-31-treated (0.13%±0.03%) and untreated (0.13%±0.08%) aged hypertensive mice (p>0.99). Similarly, CMH density remained comparable (SS-31: 13.39±2.42 vs. untreated: 11.79±2.50; p=89). Our novel methods have been benchmarked against manual counting and color deconvolution-based approaches. The ML-based classifier significantly differed in CMH area (p<0.001) but not in CMH number compared to manual counting (p>0.99), suggesting its strength lies in CMH count estimation rather than area quantification. The color space transformation showed no significant difference from manual counting in either CMH area (p>0.99) or CMH number (p>0.99), confirming it as the most reliable and accurate alternative for CMH quantification. Conclusions: These findings indicate that targeting mitochondrial oxidative stress alone may not be sufficient to prevent CMH formation in aging. However, our imaging pipeline provides a scalable and unbiased method for CMH detection, improving accuracy and efficiency. Future studies should explore multi-targeted interventions for cerebrovascular protection. Ethical Statement: All animal procedures were conducted in accordance with the APS Guiding Principles for Research Involving Animals and approved by the Institutional Animal Care and Use Committee (IACUC) at OUHSC. Conflict of Interest: The authors declare no competing financial or personal interests. Supported by NIA (RF1AG072295, R01AG055395, R01AG068295, R01AG070915), NINDS (R01NS100782), NCI (R01CA255840) and AHA (AHA.24CDA1276505, AHA.25PRE1378958). This abstract was presented at the American Physiology Summit 2025 and is only available in HTML format. There is no downloadable file or PDF version. The Physiology editorial board was not involved in the peer review process.