The effectiveness of traditional craniotomy in the treatment of intracerebral hemorrhage remains controversial. Minimally invasive surgery, specially, image-guided hematoma aspiration, proves to be effective and may have some advantages compared with craniotomy. This multicenter randomized controlled trial aims to evaluate and compare the clinical efficacy of two minimally invasive treatment strategies for patients with spontaneous supratentorial intracerebral hemorrhage (ICH) with moderate hematoma volume (20-50 mL): (1) AI-assisted, navigation-guided hematoma aspiration, and (2) targeted pharmacological therapy. This study is designed to address the current lack of prospective comparative evidence between advanced image-guided surgical intervention and medical management in this specific patient population. By focusing on functional recovery, hematoma resolution, and safety outcomes, this trial seeks to provide high-quality evidence to guide treatment decision-making and optimize individualized care for patients with spontaneous ICH.

Background and Rationale Spontaneous intracerebral hemorrhage (ICH) is a devastating form of stroke with high morbidity and mortality, especially in low- and middle-income populations. Traditional open craniotomy for hematoma evacuation remains controversial due to its invasiveness and the associated risk of secondary brain injury. In recent years, minimally invasive approaches have gained traction, particularly in patients with moderate hematoma volume. Among them, AI-assisted navigation-guided hematoma aspiration offers the promise of precise, real-time trajectory planning and minimized parenchymal disruption. In parallel, pharmacological approaches that promote hematoma resolution and neuroprotection-termed "targeted pharmacological therapy"-are emerging as non-surgical alternatives. However, no large-scale, prospective randomized controlled trials have directly compared these two strategies in patients with moderate-volume ICH (defined as 20-50 mL), a group that often poses management dilemmas in clinical practice.

Study Objective The primary aim of this study is to compare the efficacy and safety of AI-assisted navigated hematoma aspiration versusconservative therapy in patients with spontaneous supratentorial ICH of moderate volume (20-50 mL). We hypothesize that AI-assisted aspiration will yield superior functional outcomes compared to pharmacological treatment, without increasing the risk of adverse events.

Study Design This is a prospective, multicenter, randomized controlled trial involving 680 patients recruited from multiple high-volume stroke or neurosurgical centers. After obtaining informed consent, eligible participants will be randomly assigned in a 1:1 ratio to receive either AI-guided navigated hematoma aspiration or conservative therapy. Due to the nature of the interventions, this is an open-label trial; however, outcome assessment will be performed by independent evaluators blinded to the treatment allocation.

Randomization will be centralized and stratified by participating center using a computer-generated allocation sequence. All participating centers will receive standardized training to ensure protocol adherence, including surgical techniques, drug regimens, and outcome assessment procedures. The anticipated recruitment period is 24 months, with each patient followed up for a total of 6 months post-treatment.

Eligibility Criteria Participants aged between 18 and 80 years who present with spontaneous supratentorial ICH confirmed by non-contrast CT scan will be considered for inclusion. Eligible patients must have a hematoma volume between 20 and 50 mL as meassured with the 3D Slicer software, and a Glasgow Coma Scale (GCS) score of at least 8 upon admission. Patients must be randomized within 24 hours of symptom onset. Exclusion criteria include infratentorial hemorrhage, secondary causes of ICH (e.g., trauma, vascular malformations, tumors), severe comorbidities, coagulopathy that cannot be reversed, and pregnancy.

Intervention Details Patients in the intervention group will undergo AI-assisted, navigation-guided hematoma aspiration. This procedure involves real-time planning of the optimal puncture path using AI-enhanced image processing software integrated with electromagnetic or stereotactic neuronavigation systems. The aspiration will be performed via a minimally invasive burr-hole approach, under local or general anesthesia, depending on patient condition and institutional protocol. After surgery patients will receive best pharmacological therapy as patients in the control group will receive. Postoperative imaging will be conducted within 24 hours to confirm the extent of hematoma evacuation.

In the control group, patients will receive best pharmacological therapy aimed at promoting hematoma resolution and neuroprotection. This may include agents such as low-dose urokinase administered via lumbar or intraventricular routes, iron chelators such as deferoxamine, or other neuroprotective agents based on standardized protocols agreed upon by the study group. Supportive neurocritical care will be uniformly applied in both groups, including blood pressure control, ICP management, and seizure prophylaxis as appropriate.

Outcome Measures The primary outcome of the trial is the functional outcome at 6 months, assessed by the modified Rankin Scale (mRS). A favorable outcome is defined as an mRS score of 0 to 3.

Secondary outcomes include: Mortality at 30 days and 6 months; Hematoma volume reduction at 72 hours post-intervention; Incidence of complications, including rebleeding, infection, seizures, and hydrocephalus; Neurological status at discharge and day 90 (assessed by NIHSS and GCS); Health-related quality of life assessed by the EQ-5D instrument at 6 months; Duration of ICU and total hospital stay; Economic evaluation comparing the cost-effectiveness of both treatment strategies.

Sample Size and Statistical Analysis Based on prior data and assuming a favorable outcome rate of 32% in the surgery group and 22% in the conservative group, a total sample of 620 participants (310 per group) would provide 80% power to detect a statistically significant difference at a two-sided alpha level of 0.05. Allowing for up to 10% loss to follow-up, the total sample size has been inflated to 680 patients.

All analyses will follow the intention-to-treat principle, with additional per-protocol analyses conducted for sensitivity. Categorical variables will be compared using chi-square or Fisher's exact tests, and continuous variables using t-tests or non-parametric equivalents. Logistic regression will be used to adjust for baseline covariates and perform subgroup analyses, including by age, hematoma location, and initial GCS.

An interim analysis is planned after 50% of patients have completed the 6-month follow-up, under the supervision of an independent Data and Safety Monitoring Board (DSMB).

Inclusion Criteria:

• Confirmed supratentorial hypertensive intracerebral hemorrhage on brain CT scan
• Hematoma volume 20-50mL
• Patients with with GCS score ≥8
• Admitted within 24h of ictus

Exclusion Criteria:

• Intracerebral hemorrhage caused by tumor, coagulopathy, aneurysm, or arteriovenous malformation
• Concurrent head injury or history of head injury
• Multiple intracerebral hemorrhage
• Known advanced demential or disability before
• Severe concomitant diseases that affect life expectancy
• With severe intraventricular hemorrhage
• Pregnant women

Patients' personal data are made anonymous and numbered by the system. A list of patient names and study numbers is kept in a separate file to ensure that patients' confidentiality is maintained.