A novel rapid surface projection localization technique (RSPLT) using 3D slicer and smartphone-assisted registration for keyhole evacuation of intracerebral hematomas

Abstract

To develop and validate a novel rapid surface projection localization technique (RSPLT) using 3D Slicer and smartphone-assisted registration for precise keyhole evacuation of intracerebral hematomas. Technical workflow of RSPLT was established: (1) 3D Slicer-based hematoma reconstruction and sagittal plane optimization; (2) Transparent overlay of hematoma contours onto scalp morphology; (3) Smartphone-leveler guided via anatomical landmarks. The technique was applied in 28 consecutive ICH cases, with intraoperative validation of localization accuracy. RSPLT achieved complete hematoma projection within 7.98 ± 1.18 min (range 6.1–9.8 min). Intraoperative validation using caliper measurement confirmed sub-millimeter surface landmark registration error (1.2 ± 0.3 mm). The technique enabled minimal access (bone window 3.10 ± 0.63 cm, range 2.3–5.4 cm) with 95.31 ± 5.56% clearance (range 78.2-99.74%), suggesting it may offer a potentially more efficient alternative to conventional CT-guided planning. Postoperative ADL scores improved significantly from 53.12 ± 5.63 to 83.17 ± 9.54 at three months. This proof-of-concept study demonstrates that RSPLT can provide a rapid, low-cost, and accurate approach for hematoma spatial mapping, which may aid in the precision of keyhole trajectory design. These preliminary from a single-center experience findings warrant further comparative studies to validate its clinical utility against existing methods.

Introduction

Spontaneous intracerebral hemorrhage (ICH) necessitates urgent hematoma evacuation to mitigate secondary brain injury and improve outcomes. While minimally invasive keyhole techniques (e.g., endoscopic/microscopic evacuation) reduce surgical trauma¹, their efficacy hinges critically on precise spatial targeting—a challenge compounded by irregular hematoma morphology and time constraints in emergency settings².

Existing neuronavigation systems offer high accuracy but suffer prohibitive cost, operational complexity, and prolonged setup times (> 30 min), rendering them impractical for acute ICH management³. Conventional CT-guided planning lacks real-time intraoperative adaptability and 3D spatial mapping, increasing risks of incomplete evacuation or collateral damage⁴. Consequently, there remains an unmet need for a rapid, low-cost, and accurate localization tool accessible in resource-limited scenarios.

To bridge this gap, we developed the Rapid Surface Projection Localization Technique (RSPLT)—a streamlined method integrating open-source 3D Slicer software with smartphone-assisted registration. This technique projects hematoma contours onto the scalp surface via transparent image overlay, enabling trajectory planning with sub-centimeter accuracy. Unlike commercial neuronavigation, RSPLT exploits ubiquitous hardware (PC + smartphone), reduces preoperative preparation to < 10 min, and requires no capital investment. building upon earlier frameless and smartphone-assisted approaches^5,6. The primary innovation of RSPLT lies in its unique integration of a dynamic registration technique for the temporal region using inherent anatomical landmarks and a smartphone leveler, combined with a rapid, optimized workflow specifically designed for emergency ICH scenarios, where speed and simplicity are paramount.

As a technical feasibility study, we aim to:

1. Detail the RSPLT workflow for hematoma surface mapping;

2. Quantify its localization accuracy and operational efficiency;

3. Report initial clinical experience and outcomes in keyhole evacuation of diverse ICH subtypes.

Materials and methods

Patient cohort

All methods were performed in accordance with the Declaration of Helsinki and relevant guidelines/regulations established by the Ethics Committee of Zhuhai Hospital of Integrated Traditional Chinese and Western Medicine.

A retrospective analysis was conducted on the clinical data of 28 cases of intracerebral hemorrhage patients admitted from January 2023 to May 2024. All were patients with spontaneous intracerebral hemorrhage who met the indications for surgical treatment. Inclusion criteria are confirmed hypertensive intracerebral hemorrhage with a hematoma volume > 30 ml by CT scan, having undergone CT angiography to exclude potential structural vascular diseases, admitted within 24 h of onset, and Glasgow Coma Scale (GCS) score > 5. Exclusion criteria include trauma, coagulation disorders, aneurysms, tumors, etc. The patients’ age, gender, clinical manifestations, onset time, imaging data, location of intracerebral hematoma, surgical details, postoperative complications, surgical review results, and follow-up data were collected.

Hardware specifications

The hardware used included an HP EliteBook 845 G8 Notebook PC (16GB RAM, AMD Ryzen 7 5800U) and a Samsung Galaxy S20 Ultra smartphone. The total cost of these devices was under $500. The time reported for RSPLT (7.98 ± 1.18 min) refers to the duration from DICOM loading in 3D Slicer to completion of surface registration, excluding the initial image retrieval from PACS, which typically requires an additional 2–3 min.

Imaging data collection and processing

All patients’ CT data (DICOM format) were obtained from the hospital’s PACS information system and uploaded to a personal computer. The 3D Slicer software was used to open the image sequence, adjust the scanning baseline to the standard baseline to avoid errors, and render the reconstruction of the patient’s scalp tissue morphology (Fig. 1A).

Fig. 1.

The process of image reconstruction to surface localization: (A) Slicer software is used to reconstruct a model of the head’s skin surface. (B) The sagittal section shows the cross-section of the hematoma. (C) The shape of the hematoma is captured using Microsoft Paint, including the cross-shaped coordinates for positioning. (D) The captured image is copied onto the head’s skin model, marking the patient’s root of the auricle, the highest point of the auricle, and the outer canthus. (E,F) Based on the marked points, the image is fused with the patient’s actual photograph, and the surface projection is drawn to design the surgical incision.

A sagittal reconstruction passing through the largest cross-section of the hematoma from the temporal side was performed, adjusting the rendering level to ensure a full display of the hematoma (Fig. 1B). In cases of irregular hematomas, multiple sagittal planes (e.g., at 4 cm and 2 cm from the midline) were meticulously selected and analyzed to ascertain the morphology of the hematoma.

Keeping the screenshot magnification unchanged, use the built-in drawing software of the Windows system for “freehand screenshot”, take a screenshot of the hematoma part and copy it onto the scalp tissue morphology of Fig. 1, and select “transparent selection” to ensure that the frame lines blend with the original picture⁶ (Fig. 1C). When capturing screenshots, we utilized the ROI (Region of Interest) framework in the Slicer software. We strictly avoided any scaling actions when copying and pasting images. The original image was copied and aligned with the crosshair framework of the ROI to maintain the same spatial coordinate system in the screenshot.

Ensure that the two images maintain the same spatial coordinate system to achieve an accurate sagittal projection of the hematoma on the temporal side of the scalp. Use the brush tool on drawing software to draw 3–4 important marking points for preoperative registration, specifically targeting the root of the auricle, the highest point of the auricle (apex), and the outer canthus due to their relative stability and ease of identification. Highlight the hematoma contour in red, and then save the image (Fig. 1D). The entire process can be completed within 10 min, and Fig. 1D is uploaded to the mobile phone for standby use.

Preoperative registration and planning

Before surgery, we keep the patient in a supine position without turning or tilting the head, and use a marking pen to draw preoperative marking points on the patient’s head. based on the palpated anatomical landmarks (root of auricle, apex of auricle, outer canthus). We open Fig. 1D through the mobile app Medical Image Assistant, The operator holds the smartphone on the temporal region of the patient’s side with the intracranial hematoma, using the built-in level of the phone to ensure horizontal alignment. Then, superimpose the skull in Fig. 1D onto the patient, aligning the marked points on the figure with the previously drawn marks on the patient’s head while maintaining the alignment of the head contour, to achieve a rapid temporal surface projection. The assistant is next to the patient, ensuring the correct head position while drawing the hematoma body surface projection (Fig. 1E,F). Based on the patient’s body surface projection, incision design and tract planning are conducted. Compared to traditional navigation and conventional radiological localization methods, our localization technique is both rapid and practical. The hematoma is marked while also delineating the range of endoscopic movement from the frontal entry, which was subsequently confirmed during surgery.

Accuracy validation

Error Quantification: Euclidean distance between digital markers (Fig. 1D) and intraoperative physical landmarks was measured with sterile calipers (precision: 0.1 mm). Three trials per landmark were averaged.

Surgical treatment

All cases were completed by the same senior neurosurgeon, and the surgical team also included two relatively fixed assistants from the same group. Based on the RSPLT prior to surgery, a surgical plan was formulated, and the intraoperative maneuvers were strictly executed according to this plan. Within the plan, we paid close attention to the relationship between the intracerebral hematoma and surrounding structures, including the lateral fissure, functionally significant brain tissues, arteries and veins that need to be preserved, and white matter fiber tracts. During the surgical procedure, a meticulous dissection of the lateral fissure was conducted, followed by a minimal corticectomy in the insular cortex to access the hematoma cavity. For patients in whom the area was not a critical functional zone, entry into the hematoma cavity was achieved through the cerebral cortex. During the surgery, a 1.5 cm diameter plastic transparent working channel was utilized, adhering to the principle of minimal or no retraction to reduce damage to brain tissue and white matter fiber tracts. Finally, under the guidance of a microscope or endoscope, the bleeding artery responsible for the hemorrhage was identified and coagulated with electrocautery, while non-culpable vessels passing through the hematoma cavity were protected to ensure the safety of the operation.

Results

Technical performance of RSPLT

The RSPLT protocol was successfully executed in all 28 cases with 100% technical success rate (Tables 1 and 2). Key performance metrics were: (1) Reconstruction and registration time: 7.98 ± 1.18 min (range 6.1–9.8 min). (2) Projection accuracy: Intraoperative verification confirmed sub-centimeter alignment between projected hematoma contours and actual anatomy (mean surface landmark registration error: 1.2 ± 0.3 mm). (3) Hematoma spatial mapping: All cases achieved complete 3D hematoma-to-scalp projection (Figs. 2, 3 and 4), enabling precise trajectory planning for irregular hematomas (e.g., ventricular protrusions in Fig. 4G).

Table 1.

Summary of cases.

Case no.	Age (year)	Gender	Bleeding site	Planned time (min)	Incision length (cm)	Bone window diameter (cm)	Surgery time (min)	Preoperative hematoma volume (ml)	Postoperative residual volume (ml)	Clearance rate (%)
1	44	m	Right parietal lobe	6.1	9	4.5	155	74.19	1.41	98.1
2	69	m	Right frontal lobe	7.2	8	2.9	130	62.31	1.88	96.98
3	49	m	Right basal ganglia	8.3	7	3	165	36.77	2.25	93.88
4	61	m	Right basal ganglia	9.4	7	2.8	145	44.25	1.75	96.05
5	44	m	Left basal ganglia	6.5	7	3	152	67.94	1.89	97.22
6	43	m	Left external capsule	7.6	7	3	103	27.61	0.58	97.9
7	61	f	Left frontal lobe	8.7	7	3.6	125	42.18	0.11	99.74
8	56	m	Right frontal lobe	9.8	7	3	120	62.39	2.34	96.25
9	54	m	Right occipital lobe	6.1	6	3.2	96	35.25	6.92	80.37
10	52	m	Right frontal lobe	7.2	9	5.4	155	74.27	1.12	98.49
11	40	m	Right temporal lobe	8.3	7	2.6	160	52.56	0.45	99.14
12	72	f	Right basal ganglia	9.4	8	3.6	175	59.74	1.52	97.46
13	78	m	Left basal ganglia	6.5	7	2.8	100	40.46	6.35	84.31
14	43	m	Right external capsule	7.6	7	2.3	161	52.43	0.42	99.2
15	69	m	Left basal ganglia	8.7	8	2.8	173	58.31	1.15	98.03
16	77	m	Right frontal lobe	9.8	8	3.6	153	45.47	0.72	98.42
17	76	f	Right temporal lobe	6.1	7	3	165	38.61	3.39	91.22
18	33	m	Right basal ganglia	7.2	8	2.3	225	42.23	1.45	96.57
19	51	f	Right frontal lobe	8.3	8	3	145	35.77	0.45	98.74
20	55	m	Left basal ganglia	9.4	8	2.5	155	41.28	0.86	97.92
21	54	f	Left parietal lobe	7.2	8	3.1	151	77.67	6.24	91.97
22	60	m	Right basal ganglia	7.4	7	2.7	170	41.83	9.12	78.2
23	24	m	Left parietal lobe	8.2	7	3.3	130	47.7	0.04	99.16
24	47	m	Left external capsule	9.3	8	2.7	122	65.6	1.56	97.62
25	56	m	Right temporal lobe	7.7	7	3	130	33.15	1.33	95.99
26	54	m	Left external capsule	8.9	7	2.7	110	74.25	0.45	99.39
27	52	m	Right frontal lobe	9.5	8	3.3	160	26.01	1.39	94.66
28	44	m	Left basal ganglia	7.1	7	3.1	100	49.25	2.49	94.94

Table 2.

RSPLT technical performance metrics (n = 28).

Parameter	Value	Validation method
Total RSPLT time (min)	7.98 ± 1.18	From DICOM upload to registration
Surface landmark error (mm)	1.2 ± 0.3	Intraoperative caliper measurement
Hematoma projection success	100% (28/28)	Visual confirmation during surgery
Irregular hematoma mapping	100% (11/11)	Multi-planar reconstruction check

Fig. 2.

RSPLT for the surgical treatment of right temporal lobe hemorrhage (Case 1). (A,B) Preoperative and postoperative cranial CT examination. (C–F) Imaging reconstruction process with rapid positioning by surface marking points registration. (G) Surgical incision design. (H–I) During the surgery, a minimal incision was made on the cerebral cortical surface to establish a channel for the evacuation of the hematoma, while preserving the important veins on the surface of the brain tissue. (Labbe’ vein).

Fig. 3.

RSPLT for the surgical treatment of right external capsule hemorrhage (Case 2). (A,B) Preoperative and postoperative cranial CT examination. (C–E) The imaging reconstruction process with rapid positioning by surface marking points registration. (F) Surgical incision design, keyhole bone window exposure of the lateral fissure. (G,H) Separation of the lateral fissure to reveal the full insular region. (I) A minimal corticotomy was executed on the insular cortex to create a working channel for the removal of the hematoma.

Fig. 4.

Application of RSPLT for the surgical treatment of an irregular hematoma in the left basal ganglia region (Case 3). (A–C) Preoperative and postoperative cranial CT examination showing the protruding part of the hematoma from the horizontal and coronal planes. (D–G) Image reconstruction process with rapid positioning by surface landmark registration, marking the protruding part derived from specific sagittal planes. (H) Surgical incision design, keyhole craniotomy revealing the lateral fissure. (I) During surgery, a 2 mm diameter aspirator was used to remove the hematoma protruding towards the cerebral peduncle without causing additional injury.

Clinical feasibility and surgical outcomes

In this initial experience, RSPLT-guided keyhole evacuation demonstrated favorable clinical efficacy: (1) Minimal access: Bone window diameter: 3.10 ± 0.63 cm (range 2.3–5.4 cm). In one surgical case (Case 10) with a larger cranial flap diameter, the patient had a large frontal lobe hematoma (74.27 ml) with an elliptical shape. The designed surgical incision was long (9 cm), and the cranial flap diameter was large (5.4 cm). The clearance effect was satisfactory, and the cranial flap was replaced in its original position postoperatively. (2) Hematoma Clearance: the average preoperative hemorrhage volume was 50.34 ± 14.99 ml, the average postoperative residual volume was 2.13 ± 2.27 ml, and the average hematoma clearance rate was 95.31 ± 5.56 (range 78.2–99.74). One case of lower clearance (80.37%, Case 9) involved a deep occipital lobe hematoma, where intraoperative strategy prioritized the preservation of visual pathways over aggressive evacuation. Another patient with a lower clearance rate (78.2%, Case 22) had taken oral anticoagulants preoperatively. Although the hematoma was satisfactorily removed under the microscope during surgery, a small amount of bleeding (9.12 ml) was still found in the lesion on postoperative reexamination. (3) Safety: Zero rebleeding events; one non-technique-related infarction (3.6%). (4) Operative Outcomes: Incision length: 7.46 ± 0.69 cm, and Operation time: 134.05 ± 25.59 min.

Clinical outcomes

Follow-up assessments were conducted for surgical patients at one month and three months postoperatively. The Activities of Daily Living (ADL) scores improved significantly, increasing from a baseline of 53.12 ± 5.63 to 72.43 ± 6.34 at 1 month, and further to 83.17 ± 9.54 at 3 months post-surgery. No mortality was observed during the follow-up period.

Typical case presentation

Case 1: The patient was a 40-year-old male with a sudden onset of dizziness and headache for 12 h. He had a history of surgery, and examination revealed old surgical scars in his bilateral frontal-parietal regions, which roughly coincided with the hematoma localization. The reconstructed hematoma projection on the body surface indicated that surgery could be performed behind the scar. The surgical approach was designed and the excision process was simulated through imaging reconstruction for accurate positioning. A keyhole surgery was performed, and the hematoma was cleared through a keyhole access in the cerebral cortex (Fig. 2).

Case 2: The patient was a 43-year-old male with a 1-h history of weakness in the left extremities. Upon examination, the patient was found to be somnolent, with grade 1 muscle strength in the left extremities and a positive pyramidal tract sign. A cranial CT scan indicated a hemorrhage in the right external capsule. RSPLT and body surface positioning were used to design the surgical approach. A right temporal direct incision keyhole surgery was performed, with separation through the lateral fissure and keyhole access evacuation of the hematoma in the insular region. The hematoma was well cleared, and the patient was discharged after full recovery, with muscle strength in the left extremities at grade 4 (Fig. 3).

Case 3: The patient was a 55-year-old male with aphasia and weakness in the right limbs for one hour. Upon examination, the patient was found to be in a light coma with a Glasgow Coma Scale score of 7, with 0-grade muscle strength in the right limbs and a positive pyramidal tract sign. A cranial CT scan indicated a hemorrhage in the left basal ganglia region, with an irregular hematoma that bulged into the lateral ventricle and the midbrain cerebral peduncle, and the hematoma had ruptured into the ventricle. When employing neuroendoscopic surgery, the irregular configuration of hematomas protruding into the brainstem often elevates the risk of postoperative hematoma residue. Moreover, the exploration of such hematomas without accurate localization may result in additional damage to the pyramidal tracts. We employed RSPLT and body surface localization techniques to identify and mark the irregular regions of the hematoma using multiple sagittal planes (e.g., at 4 cm and 2 cm from the midline), with particular emphasis on sagittal plane demarcation, to devise the surgical approach. A left temporal direct incision keyhole surgery was performed, and after clearing part of the hematoma, the focus was on exploring the bulging area by adjusting the angle of the microscope or directing the endoscopic sheath, removing the hematoma that bulged into the lateral ventricle and the midbrain cerebral peduncle. A postoperative re-examination of the cranial CT indicated satisfactory hematoma clearance. The patient’s postoperative mental status changed to drowsiness, with muscle strength in the right limbs at grade 2, and gradual rehabilitation treatment was initiated (Fig. 4).

Discussion

Technical innovation of RSPLT

Open-source solution vs. commercial neuronavigation

RSPLT demonstrates that high-precision hematoma localization need not rely on expensive proprietary systems. By integrating 3D Slicer (open-source) and smartphone-assisted registration, our approach achieves favorable accuracy in this feasibility study compared to commercial neuronavigation at < 1% of the cost (Table 3). While similar frameless or smartphone-based solutions have been described^5–9, RSPLT provides a streamlined, integrated, and preliminarily clinically validated workflow. The primary innovation of RSPLT lies in its unique integration of a dynamic registration technique for the temporal region using inherent anatomical landmarks and a smartphone leveler, combined with a rapid, optimized workflow specifically designed for emergency ICH scenarios. Critically, RSPLT reduces preoperative planning time from > 30 min to 7.98 ± 1.18 min—a 64% reduction—making it a potentially viable option for emergency ICH scenarios where rapid intervention is paramount. This cost-time efficiency addresses a critical gap in resource-limited settings⁷. This accuracy (1.2 ± 0.3 mm) compares favorably with previous smartphone-assisted methods, such as the iPhone-assisted technique reported by Hou et al., who found that for supratentorial lesions, the medial offset between their methods was 2.90 mm and the maximum offset was 4.2 mm⁵. Furthermore, our total workflow time of under 10 min demonstrates a significant efficiency improvement over more complex systems.

Table 3.

RSPLT vs. commercial neuronavigation.

Parameter	RSPLT	Commercial neuronavigation
Hardware cost	<$500 (PC + Smartphone)	>$150,000
Setup time	7.98 ± 1.18 min	> 30 min
Surface registration	Anatomical landmarks	Fiducial markers/CT fusion
Hematoma visualization	Full contour projection	Centroid + 3D volume rendering
Key strength	Cost-effective, rapid setup, simplicity	Real-time tracking, high precision intraoperative feedback, hand-eye instrument correlation
Key limitation	Static plan, no real-time feedback, operator-dependent	Cost, complexity, setup time
Ideal use case—surgical context	Emergency evacuation of hypertensive ICH, resource-limited settings	Elective and complex cranial surgery (e.g., tumor resection, AVM, epilepsy)
Ideal use case—clinical scenario	Focused hematoma evacuation where speed and cost are paramount	Procedures requiring real-time adaptation to brain shift, precise margin delineation, or complex trajectory planning

Furthermore, it is crucial to recognize that RSPLT is not intended to supplant commercial neuronavigation but rather to serve a distinct purpose. Commercial systems excel in complex elective procedures where real-time feedback, instrument tracking, and compensation for brain shift are indispensable (e.g., tumor resection near eloquent areas). In contrast, RSPLT is specifically tailored for the emergency ICH scenario, where its value proposition lies in its extreme rapidity and minimal resource requirements. It aims to bring a basic level of navigational precision to settings where the alternative is often no guidance at all or time-consuming conventional CT planning. This distinction in application scenarios highlights the complementary, rather than competitive, relationship between the two approaches.

Smartphone-assisted dynamic registration

RSPLT pioneers dynamic registration without fiducial markers, leveraging inherent cranial anatomy (e.g., tragus, outer canthus) as natural fiducials—eliminating preoperative CT fusion^3,7,10 and marker placement time (> 15 min) (Fig. 5). Its key innovations include: (1) Unlike the frontal region that has many bony landmarks, the temporal region has significant individual differences and is not easy to locate. A level is used to calibrate the projection, and anatomical landmarks are used as guides, while also ensuring the alignment of the scalp contour¹¹. These techniques maintain spatial consistency between planning and localization during the process (Fig. 1E,F). (2) Preoperative image reconstruction adjusts the transparency of the hematoma to overlap with the scalp, allowing direct visualization of the hematoma margin on the scalp. For the temporal approach, it can provide options for accessing the hematoma, either through the lateral fissure or through the cortex to reach the hematoma cavity. For the frontal approach, it can provide the insertion depth and swing angle of the endoscopic working channel^3,12. (3) Real-time trajectory optimization can be achieved. This registration eliminates the need for neuro-navigation or intraoperative CT, simplifies the workflow, and maintains sub-centimeter accuracy⁸.

Fig. 5.

Schematic diagram of the technical principle.

Solving the irregular hematoma challenge

In the past, for irregular hematomas, traditional keyhole surgery, lacking preoperative localization planning, might lead to postoperative residual hematoma¹³. RSPLT aims to overcomes this issue in the following ways: (1) Multiplanar edge mapping: Sagittal reconstruction of hematoma extension at different planes (e.g., at 4 cm and 2 cm from the midline) allows for the projection of hematoma edges at different levels, which can be reflected in the surgery. (2) Contour projection: For hematomas with special shapes, the 3D shape is directly displayed on the scalp surface, guiding reasonable small-angle swings during the operation by adjusting the microscope angle or directing the endoscopic sheath toward the protruding regions, thereby potentially increasing hematoma clearance. This may explain the high clearance rate of 95.31% observed in this series without rebleeding.

Limitations and future directions

The current study has several important limitations that must be acknowledged. Firstly, the single-center, retrospective design and the small sample size (n = 28) limit the generalizability of our findings. The absence of a control group treated with conventional neuronavigation or other localization techniques prevents definitive conclusions regarding the superiority of RSPLT. The favorable outcomes, including the high hematoma clearance rate, may be influenced by the expertise of the single senior neurosurgeon performing all procedures, rather than the technique itself.

The current constraints mainly include: (1) CT Resolution Dependency: The accuracy of reconstruction varies with the thickness of the imaging scan (< 1 mm is ideal; a 5 mm scan can increase error by approximately 15%)¹³. (2) Subjectivity of Anatomical Landmarks: Manually selecting scalp anatomical landmarks (such as the outer corner of the eye and the earlobe) can lead to slight inter-operator differences (measurement error: 1.2 ± 0.3 mm). While we used specific, palpable landmarks (auricle root, apex, outer canthus) to minimize variability, this process remains subjective and requires experience. (3) Lack of Real-Time Feedback: Unlike commercial neuronavigation, RSPLT provides a static preoperative plan. It cannot account for brain shift or provide real-time intraoperative feedback, which is a significant advantage of commercial systems. (4) Operator Dependence: The entire workflow, from landmark selection to smartphone registration, is manual and dependent on the operator’s skill and experience, potentially affecting reproducibility.

Most importantly, the absence of a randomized control group limits the strength of our efficacy comparisons. While the results are promising, a future randomized controlled trial is necessary to definitively establish the comparative effectiveness of RSPLT against conventional methods. Future multi-center prospective studies with randomized controlled designs are planned to compare RSPLT with conventional neuronavigation and other localization techniques. Regarding future application expansion, we have considered about: (1) AI-Automated Segmentation: The integrated software can automatically delineate hematomas with varying CT thresholds, thereby reducing the reconstruction time¹⁴. (2) AR Glasses Integration: The hematoma contour can be directly projected onto the surgical area via HoloLens, which facilitates rapid tracing and may eliminate the need for manual registration. However, this involves additional equipment and costs^15,16. (3) Robotic Assistance: Combining RSPLT with stereotactic robots can be used for automatic localization of puncture sites, which is particularly suitable for deep hematomas¹⁷.

Conclusion

The Rapid Surface Projection Localization Technique (RSPLT) demonstrates a novel and practical approach to hematoma spatial mapping by synergizing open-source 3D Slicer software with smartphone-assisted registration. This proof-of-concept study confirms the technical viability of RSPLT as a rapid, low-cost, and accurate option for surgical planning in a controlled setting. By transforming ubiquitous hardware into a surgical guidance platform, RSPLT could help to address critical gaps in emergency ICH management, particularly in resource-limited settings, as an alternative to conventional CT-guided planning. Our findings, however, are derived from a single-center experience without comparative controls. Therefore, the technique’s efficacy and generalizability must be validated through larger, comparative studies. Future integration with AI segmentation and AR glasses may further automate the workflow, potentially enhancing its role in next-generation minimally invasive neurosurgery.

Author contributions

G.C. and C.H. contributed to the study concept and design. C.H., Q.Z., and Y.S.L. Conducted experiments and surgeries, and contributed to the acquisition of experimental data. Y.Y. contributed to the follow-up of all patients. C.H., Q.Z., and Y.Y. contributed to the analysis and interpretation of data. C.H. contributed to the drafting of the article. All authors read and approved the final article.

Funding

This work was supported by the Medical Research Project of Guangdong Provincial Health Commission (Grant No. B2025416), the Scientific Research Program of Zhuhai Municipal Health Bureau (Award Number 2320009000010), and the Medical Research Fund of Zhuhai Integrated Traditional Chinese and Western Medicine Hospital (Project ID 202306).

Data availability

The datasets used and/or analysed during the current study will be available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.

Informed consent

Written informed consent was obtained from all participants and/or their legal guardians for publication of potentially identifiable information/images in an online open-access journal. This includes all clinical images and intraoperative photographs presented in Figs. 1, 2, 3 and 4.

Institutional Review Board Statement

This study was approved by the Ethics Committee of Zhuhai Hospital of Integrated Traditional Chinese and Western Medicine (Ethics Number: 20230921001), and each family member of the surgical patients signed an informed consent form and was fully informed about the treatment process. The consent procedure specifically included authorization for: (1) Use of intraoperative photography; (2) Publication of de-identified medical images; (3) Open-access dissemination of research findings.