Keywords
3D reconstruction - DICOM software - imaging technologies - intracranial lesions - keyhole techniques - minimally invasive surgery - neuronavigation - preoperative planning - resource-limited settings - surgical efficacy - surgical outcomes
Introduction
Neurosurgical procedures in their initial years used to be primarily exploratory in nature; hence, considerably large skin and craniotomy flaps were the recommended technical standards. With advances in preoperative imaging technologies, accurate localization of intracranial lesions became possible; thus, smaller skin and craniotomy flaps became the practice trend. The technological advancements to assist in the preoperative planning of neurosurgical procedures include sophisticated neuronavigation systems that can be integrated with other operative adjuncts, including microscopes and endoscopes, dedicated software for presurgical planning using correct and detailed anatomical images, and modern simulation technologies.[1]
[2] High-fidelity three-dimensional (3D) visualization of anatomy is more critical for minimally invasive neurosurgical approaches.[3]
[4] The significance of minimally invasive neurosurgical procedures is increasing as these sophisticated techniques are meticulously designed with the primary objective of lesser surgical trauma, quicker recovery, and reducing the overall risk for complications.
The fundamental concepts in the keyhole approaches include minimal safe skin incision and craniotomy, avoiding skin flaps to linear or curvilinear incisions, use of gravity assist with minimal use of self-retaining retractors, use of shortest route to the lesion, and keeping the long axis of the tumor perpendicular to the floor. Navigation has revolutionized these minimally invasive approaches by allowing for the exact marking of the skin incision, precise craniotomy, a direct attack on the lesion, and retractor-less surgery with the correct positioning.
Options without navigation include craniomapper,[5] fiducial markers,[6] hand tracking navigation,[7] and caliper navigation.[8] Challenges to keyhole without navigation include deciding entry and trajectory for deep-seated lesions and surface marking for superficial lesions.
Objective
This study aims to present the experience of using open-source Digital Imaging and Communications in Medicine (DICOM) software as a preoperative tool in planning keyhole neurosurgical procedures. The objective was to demonstrate that these desktop- and laptop-based softwares can be used to plan keyhole surgical procedures and techniques. This would help to compensate for the absence of neuronavigation and the advantage of keyhole surgery can be given to the patients. The primary objective of the present study was to assess the usefulness and efficacy of this novel method compared with traditional methods with respect to the following parameters: incision length, accuracy of lesion identification, operative time, blood loss, bone loss, size of craniotomy, the extent of resection, recovery rate, and complications associated with the novel design method for intracranial tumors.
Materials and Methods
Imaging Requirements
The DICOM images should be volumetric (thin slice from chin to vertex). The images should be free of movement artifacts, and 3D acquisition of images should be performed for all the sequences (T1-weighted [T1-w], T2-w, fluid-attenuated inversion recovery, magnetization prepared rapid gradient echo, post-gadolinium-T1-w) to be used in the presurgical planning. These images are similar to neuronavigation protocol, which includes the vertex of the head to the chin, the tip of the nose, and the most prominent part of the occiput included in the imaging. This is important for accurate 3D volume rendering (VR) and multiplanar reconstructions and further image analysis.
DICOM Software Requirements
The DICOM software should be capable for generating interactive 3D VR imaging with variable windowing capability. This variable windowing is used for visualizing the area of interest in relation to the patient positioning preoperatively as well as intraoperatively. These DICOM softwares should be optimized for use on handheld devices like tablets and laptops.
Technique
We have created an image-based (magnetic resonance imaging [MRI] ± computed tomography [CT]) algorithm where in preoperative CT and MRI, DICOM images are loaded on a laptop or tablet using freely available DICOM viewers (Radiant for Windows and Horos for Mac/PadOS). Then, specific image sets are created using predefined protocols and volume reconstructions that depict the lesion morphology of the cranium and the brain. We used iPad Pro intraoperatively to accurately position the patient as per angles obtained in three axes from the reconstructed images. The case examples illustrated below describe the involved steps.
This setup is used as an alternative to the navigation guidance system for:
-
Incision and craniotomy planning
-
Patient positioning and head fixation
-
Intraoperative referencing
Preoperative planning of incision, craniotomy, and trajectory follow easily identifiable landmarks ([Fig. 1]). Then, optimal patient positioning and head fixation in the skull clamp are done using the available portable devices (tablet PC or a laptop) ([Fig. 2]). Finally, an easy-to-recreate intraoperative image-viewing setup is arranged so the surgeon can control the DICOM software intraoperatively from the sterile environment (as described in the illustrative examples). This allows the surgeon to execute the keyhole plan and enables the surgeon to use intraoperative guidance using the DICOM software ([Fig. 3]).
Fig. 1 Using variable windowing (A) on three-dimensional (3D) volume reconstructed an image, the lesion projection is done on the skin (B) and then the incision is planned (C).
Fig. 2 Positioning of the patient is done under guidance of the three-dimensional (3D) image projected on the handheld device (A). The skin incision and flap are marked based upon the lesion projection on the scalp using variable windowing (B and C).
Fig. 3 Setup for intraoperative use of the technique using a mouse covered with double sterile sheet, which can be controlled without breach of sterility. Ipad connected with a mouse (A). Mouse covered with a sterile plastic sheath (B). Intraoperative use maintaining sterility (C).
After keyhole craniotomy, the surgery was conducted in standard neurosurgical manner using microneurosurgical techniques, initially with a microscope and then with endoscope assistance when needed based on the pathology. In lesions involving the skull base and deeper corridors, endoscope-assisted microsurgery was performed to reduce the amount of bone drilling and access corners hidden from the view of the microscope.
Data Collection and Statistical Analysis
The parameters of operative data were collected retrospectively using theater logbooks and operation theater notes, and recorded videos of all the cases were used to extract lengths of skin incisions, size of the bone flap, time from skin incision to durotomy, and beginning of dural closure to end of skin closure. Incidences of wound-related complications, including surgical site infections (SSIs), wound dehiscence, flap necrosis, or cerebrospinal fluid (CSF) leak were recorded. The novel method is compared with the surgical details of procedures done traditionally by the same surgeon. The matching is done to match the pathology, lesion location, size, age, and gender.
The statistical analysis was done using IBMS SPSS version 26; continuous variables were expressed using means and standard deviation when following normal distribution and median when nonnormal distribution was found. Pearson's correlation was obtained for correlation among the variables. Logistic regression was performed for quantitative variables, and multiple linear regression was performed for quantitative variables. A paired t-test was used to compare the means of two groups using the novel method and the traditional approach.
Illustrative Case Examples
Illustrative Case Examples
A video description of the technique along with illustrative case examples can be viewed at the following link: https://1drv.ms/v/s!An4f031g0vi6j-VTCbDg-HWfaNCU5Q?e=IIGw5O
Results
We operated on 176 patients using the present technique against a control group of 172 patients who were age-, gender-, and pathology-matched. There were no statistical differences between the two groups with respect to the tumor's location in the supratentorial or infratentorial compartment, pathology of the lesion, and surfacing or deep-seated lesion. The mean incision length in the test group was 50 ± 12 mm, significantly smaller than the control group, with a p-value of 0.001. The mean surface area of the craniotomy was 9 ± 2 cm2 in the test group and 120 ± 14 cm2 in the control group, with a p-value of 0.001. The mean duration of the surgery was significantly smaller in the test group (140 ± 28 minutes) than in the control group (345 ± 32 minutes), with a p-value of 0.002. Similarly, the test group had a significantly smaller incidence of wound-related complications. The wound-related complications included wound dehiscence, CSF leak, and superficial SSIs. There were no deep-seated infections in any of the patients or any mortality. A detailed description of the results is displayed in [Table 1].
Table 1
Comparison of novel technique with conventional cranial surgery (n = 176 vs. n = 172)
|
Variable
|
Present technique
(n = 176)
|
Conventional surgery
(n = 172)
|
p-Value
|
|
Supratentorial
|
120
|
132
|
0.82
|
|
Infratentorial
|
56
|
40
|
0.79
|
|
Age (mean ± SD; y)
|
42 ± 8
|
45 ± 9
|
0.53
|
|
Male
|
86
|
82
|
0.88
|
|
Female
|
90
|
90
|
0.90
|
|
Cortical (surfacing)
|
34
|
28
|
0.32
|
|
Subcortical
|
77
|
79
|
0.45
|
|
Skull base
|
65
|
65
|
0.87
|
|
Oncology
|
150
|
148
|
0.72
|
|
Vascular
|
26
|
24
|
0.66
|
|
Tumor size (mm)
|
38 ± 12
|
42 ± 13
|
0.35
|
|
Incision length (mm)
|
50 ± 12
|
200 ± 20
|
0.001
|
|
Surface area of craniotomy (cm2)
|
9 ± 2
|
120 ± 14
|
0.001
|
|
Surgery duration (min)
|
140 ± 28
|
345 ± 32
|
0.002
|
|
Extent of resection
|
92 ± 7
|
94 ± 5
|
0.12
|
|
Wound complications
|
3
|
21
|
0.001
|
|
Length of hospital stay
|
3 ± 2
|
7 ± 3
|
0.003
|
|
Major complications
|
12/176
|
15/172
|
0.25
|
|
CSF leak
|
4/176
|
8/172
|
0.45
|
Abbreviations: CSF, cerebrospinal fluid; SD, standard deviation.
Discussion
Presurgical planning has been an active area of research in neurosurgery, consisting of image registration and segmentation, often limited by poor resolution and inadequate integration of multimodal volumetric data sets.[9]
[10] This preoperative planning and intraoperative guidance in neurosurgery is done using a neuronavigation system, which is often cost-restrictive to most limited resource settings.
Typically, for a surfacing lesion, the extent of skin incision, craniotomy, and dural exposure is directly governed by the size and superficial extent of the lesion, so the margins of the normal brain are well exposed and meticulous dissection can be performed. However, for deeper seated lesions, minimally invasive approaches and keyhole approaches are becoming the standard of care where a small window is created in the overlying skull and the visual space is expanded in the deeper location with the help of visualizing aids like a microscope or endoscope or combined and the whole lesion can be accessed at the depth with minimum damage to the overlying typical structures. For both the pathologies, that is, surgeries at the brain surface and the deeper locations, an understanding of projections of the lesion on the cerebral surface, dural surface, and cranial surface is crucial for appropriate planning of skin flaps and craniotomy. A better anatomical understanding of the orientation is further required for the deep-seated lesion to understand the morphological relationship between the pathology and neighboring surrounding typical neurovascular structures.
An accurate understanding of preoperative imaging is an essential part of the neurosurgical procedure, and preoperative neurosurgical planning is a skill mastered over the years. This understanding helps the surgeon transform preoperative two-dimensional gray scale CT and MRI images into the 3D image encountered during the surgery. With this 3D image in mind, the surgeon proceeds with patient positioning, scalp incision, flap, craniotomy, and lesion excision, considering the distortion of anatomy from the lesion itself. This process of mental imagery transformation is most difficult at the beginning of a neurosurgeon's career and becomes easier as one gains expertise in mental exercise over the years. However, a mental combination of these data sets and an accurate 3D understanding of images slice by slice is challenging, even for the skilled neurosurgeon. The difficulty is compounded further if the images have been acquired in nonorthogonal planes and if movement artifacts are present in the images.[9]
[11]
Neuronavigation is an important operative adjunct tool to assist surgeons in this process, but it is expensive and unavailable in most resource-limited settings.
This novel approach for understating the preoperative radiological anatomy, patient positioning, planning skin incision, craniotomy, brain surface anatomy, and deeper structures anatomy about the key skull landmarks can be an alternative to compensate for the absence of a navigation system in the operating room. It allows the surgeon to effectively employ keyhole techniques for neurosurgery, which otherwise require a navigation system. The key to successfully apply this novel alternative is develop the skill of understanding multimodal imaging data sets, creating virtual patient positioning, skin incision, craniotomy, and a deeper corridor for the lesions.
The simple technique of transforming preoperative imaging into 3D understanding, as described in our illustrative case examples, adds to the research area on realism in neurosurgery. Realism improves anatomical understanding and also increases the immersion of the surgeon.[11] The accuracy of anatomical localization becomes more important in regions of eloquent structures, deeper located lesions, and the area of the skull base where critical neurovascular bundles are located in close vicinity, and the margin of error is significantly lower.[1]
[2]
[4]
[12]
[13]
Regular use of this technique, which is equivalent to bringing the radiology console to your operating room, guides the surgeon regarding the extent of skin incision, bone window opening, and dural opening. These factors thus determine the duration of surgery, time under anesthesia, and blood loss and indirectly regulate the length of hospital stay and recovery duration owing to the time required for the wound healing process.
The critical components of our technique in this article are demonstrating the preoperative planning technique and utilizing that knowledge in intraoperative execution for specific neurosurgical procedures. In this article, we highlight essential tasks in the presurgical planning. We simulate the intraoperative patent positioning with head fixation in a three-pin clamp, including the degree of flexion, extension, rotation, incision on the scalp, and craniotomy without any prior image segmentation. In the case examples above, we have demonstrated the visualization technique of the superficial brain areas, including data from different formats of images acquired during neuroimaging. Then, using the command for stripping the cortex, we could visualize the deeper brain structures to define the optimal window opening and trajectory for the keyhole approaches. In the literature, several pipeline algorithms allow for the stripping of the skull and allow visualization of the brain surface and deeper structure. However, they are more complex to use. This article demonstrates an easy-to-use, open-access software that can perform the same steps and plan the surgery.
Limitations
Providing a surface projection of the lesion and marking/restricting the skin flap are the most common uses. One drawback of this method is that it does not allow for intraoperative real-time navigation. Because we could not separate the blood loss according to the surgical procedures for incision, craniotomy, and tumor excision, we did not include the precise amount of blood loss in the analysis, as this is often recorded by the anesthetist. To prevent confounding bias caused by the various processes required in surgery, we excluded blood loss as one of the study's variables because we were unable to determine whether the blood loss was unique to the keyhole portion of the procedure. Skull base lesions and neoplastic surfacing lesions were included in both comparison arms. The majority of the diseases were vascular and malignant. Since the lobes involved and the diseases were so different from one another, it was not included in the study because a direct statistical comparison was not possible.
Future Implications
Future studies of minimally invasive keyhole cranial neurosurgery where the exact amount of blood loss is studied specific to the keyhole component of the surgery would be of interest in the area. There are shortcomings to the present article as mentioned above. Studies with direct comparison of pathologies and location of tumor between the traditional and keyhole cranial neurosurgery are likely to add further information on the surgical technique. Nevertheless, the idea is intriguing and might be able to take the place of navigation in situations where resources are scarce. Further perspective could be added to the study by considering the existing limitations of this technology and its extended future applications.
Conclusion
Keyhole concepts in neurosurgery are not new and have advanced over the decades with addition of newer technologies every few years. But the nonavailability of costly hardware should not deprive the patient or the surgeon from the benefits of minimally invasive approaches. This novel technique emphasizes the use of readily available and affordable technology for executing keyhole neurosurgical procedures for surgeons who do not have access to a navigation system in the operating room. Additionally, the regular use of these techniques helps the neurosurgeon in better understanding the 3D surgical anatomy of the brain and gradually makes him or her a safer and the better surgeon.
