INTRODUCTION

METHOD

Preparation of the “Doigt-de-Dieu”

The backbone of this device is the Medtronic Navigation probe originally designed for accurate placement of ventricular catheter. It is in the shape of a christian cross, with the three short limbs of the cross each bearing an optical tracking ball, and the long limb of the cross made slender enough to fit inside a regular ventricular catheter, with its tip acting as the target point at a fixed preset distance from the tracking balls (Figure 1A). A ventricular catheter is fitted over the navigation probe, with the exact distance between cortex and target surface (as shown on the navigation images) pre-marked on the catheter with a 4-0 neurulon ligature.

A 3-French, single lumen, semi-rigid, polyethylene cannula fitted with an injection port and a single opening at the tip, is tied alongside the catheter-navigation probe with several 4-0 neurulon ligatures (Figure 1B). A 5 to 7 cm glove finger cut from a size 10 latex surgical glove is now fitted over the entire cannula-catheter-probe assembly. The glove finger is carefully flattened by expelling interior air, and then tied tightly near its open end to the probe with a 2-0 Neurulon ligature (Figure 1C). The length of the glove finger must be adequate to span most if not all of the pre-measured cortex–to-target distance. Thus, the navigation probe functions as a path-finding stent, and the adjacent polyethylene cannula permits the injection of sterile saline into the glove finger to inflate it into a potential surgical corridor for the target lesion concerned (Figure 1D).

Intraoperative neuronavigation and Surgical procedure:

A detailed pre-operative MRI including colour calibrated DTI of the relevant fiber tracts in the region of interest is obtained (Figure 2A and 2B). These preoperative DTI data are merged with the neuronavigation imaging data. In our theatre, we use the Medtronic Stealth System (Stealth Station® S7® System, Medtronic, Inc. Surgical Technologies, Neurosurgery, 826 Coal Creek Circle, Louisville, CO 80027, USA). The pre-planned trajectory of the surgical approach to the deep target lesion, designed to avoid important fibre tracts and deep neuronal centres, is now transposed to the real-time screen of the Medtronic Stealth unit. The distance from cortex to target is recorded (Figure 2C), and the exact location of the entrance point on the brain surface is carefully marked on the overlying scalp. The small craniotomy is fashioned around the centre of this entry point.

Through a limited corticotomy at the designated site, the “Doigt-de-Dieu” (henceforth called the D³) is inserted into the brain under navigation guidance (Figure 3A) until its tip is touching the surface of the target lesion on the Stealth screen. About 15 to 20 ml of sterile saline, depending on the target distance, is slowly injected through the rigid cannula to inflate the D³ balloon (glove finger) very gradually, until a “brain tunnel” of approximately 1.0 to 1.5 cm diameter is created (Figure 3B). This way, a “clean” surgical corridor is formed by gently and gradually displacing white matter tracts and not by sharp dissection with surgical instruments.

RESULTS

Illustrative Case 1:

A 13 year-old boy, with a history of neurofibromatosis type 1 (NF-1), presented with persistent left sided visual disturbance of 1 year duration, described as transient loss of acuity. Visual field testing showed inconclusive left homonymous visual field defects.

MRI shows two enhancing masses deep in the medial right occipital lobe; the larger and more superficial one measured 1.6 cm in diameter, located 4 cm deep to the cortex just above but apposing the deep portion of the right optic radiation as it sweeps round the occipital horn to reach the primary visual cortex of the medial occipital lobe. Serial MRI over the prior 6 months showed the lesion was enlarging and having increasingly brighter contrast enhancement and more exuberant peri-lesional oedema. The smaller lesion was deeper and pushing against the splenium of the corpus callosum, but had shown no changes in size or other characteristics over a surveillance period of 4 years (Figure 2A). Other MRI abnormalities common in NF-1 such as focal T2 hyperintensities were seen in both hippocampi, the right cerebellar vermis, and the left middle cerebellar peduncle. The left optic nerve was enlarged but stable.

Because of the visual symptoms and the worrisome enlargement and changes in imaging characteristics of the larger right occipital lesion, craniotomy and tumor removal was recommended.

The surgical trajectory was planned just skirting above the optic radiation, as shown in the merged DTI image on the navigation platform (Figure 2B and 2C). The goal was to reach the target without cutting through or severely distorting the geniculocortical fibres. After navigation-controlled insertion of the D³ probe till the deflated glove fingertip just touched the posterior surface of the target lesion, the finger was slowly inflated with saline injection through the polyethylene cannula to create a brain tunnel, to be used as the surgical corridor after withdrawal of the D³probe and placement of the brain retractors. The posterior surface of the tumor was crisply visualized exactly at the end of this corridor (Figure 4A) and microsurgical complete resection was easily accomplished (Figure 4B).

Intraoperative MRI confirmed complete resection, and MRI with DTI on the first post-operative day clearly delineated the surgical path skirting but not traversing the upper aspect of the right optic radiation. The deep surface of the resection cavity was almost touching the splenium of the corpus callosum, without actually involving it (Figure 5).

Histopathology of the lesion confirmed a diagnosis of benign ganglioglioma. Examination at 6 months after surgery showed that he had a full visual field and normal visual acuity. MRI showed no residual or recurrent tumor (Figure 6).

Illustrative Case 2:

A 15 year old, right-handed, previously healthy girl lost consciousness for several minutes without any obvious seizure activity, but showed transient confusion afterwards. One month later, she had a similar episode, this time with

tonic-clonic movements of both arms and legs. Her post-ictal neurological examination was entirely normal. She was loaded with phenytoin. MRI showed a non-enhancing deep lesion in the left frontal periventricular white matter with signal characteristics compatible with a cavernous malformation that had undergone prior hemorrhages (Figure 7A).

Because of the lesion’s propensity to re-bleed and its association with epilepsy, we recommended craniotomy and resection of the cavernous malformation using the “D³” technique to minimize damage to Broca’s area and avoid permanent speech deficits.

The operation went well. Postoperative MRI showed the less traumatically created surgical corridor and complete removal of the cavernous malformation. There was minimal oedema surrounding the surgical path (Figure 7B). Histopathology confirmed the diagnosis of a cavernous malformation with old and recent hemorrhages.

On post-operative day one, the patient showed mild memory deficits and moderate word finding and object naming difficulties, but these improved rapidly, and had completely resolved by the time of hospital discharge 3 days later. Four months later, her Lansky performance score [10] was 100.

DISCUSSION

The “Doigt-de-Dieu” technique enables less traumatic access to deep brain lesions located in eloquent brain areas such as the vicinity of important deep gray (neuronal) centres and fibre tracts, especially in the dominant frontal and parietal lobes. Its advantage resides in the gentle displacement of brain tissue by the slow expansion of the smooth glove finger into an elongated, roughly cylindrical balloon, thereby creating a surgical working tunnel. It obviates the undoubtedly more destructive method of standard microsurgical sharp dissection using hard dissectors, suction tips, and advancing metallic brain retractors, which are made to plough through and forcefully pry open soft brain tissues while cleaving open the surgical tunnel. In the D³ technique, the self-retaining brain retractors are merely used to hold open and sustain the already gaping tunnel whilst microsurgical resection of the deep lesion is being carried out.

The rapidly evolving field of minimally invasive neurosurgery witnesses the emergence of many different techniques of cerebral dissection with the common goal of minimizing collateral brain damage along the surgical pathway, especially when accessing deep lesions. A brain retractor blade pressure exceeding 20 mmHg has been found to be the critical threshold for direct injury and secondary injury due to ischemia [8,11,12]. Tubular retractors distribute pressure equally to the surrounding tissue, resulting in a lower maximum pressure at any one point along the cylindrical wall [13]. Also, the material of the retractor was thought to be important; it varied from metal [14] to electrically non-transmissive polyester equipped with a transparent medium to visualize the surrounding tissue [2,6,15]. Progressively thicker dilating tubes were also proposed to be less invasive [4]. Combining these special retractors with neuronavigation [15] and/or DTI [16] made minimally invasive neurosurgery even more precise: Kelly et al. in the 1980s increased accuracy by using stereotactically directed retractors [17,18], and Harris et al. approached intraventricular lesions with a combination of neuroendoscopy, microsurgery, and stereotactic image guidance [7].

Besides improvement in instrumentations, invasive brain dissection may be minimized by the transsulcal approach to subcortical lesions [5]. Disadvantages of this approach are inadvertent injury to the sulcal vessels, and limitations imposed by the location of available sulci for the lesion in question.

Precise intraoperative localization of the lesion requires accurate registration of the navigation feducials, but to avoid adjacent fibre tract injury, pre-planning with clearly delineated, colour-calibrated DTI is absolutely essential. The pre-operative DTI data are merged into the intraoperative navigation platform so that fine adjustment of the pre-planned surgical path can be done “in the field” to finalize the exact probe trajectory and the marking of the cortical entrance point on the overlying scalp. However, it must be said that the usefulness of MR diffusion tractography as a trajectory planning reference may be limited [19,20]. Diffusion-weighted imaging depicts only differences in anisotropy of proton movements along parallel neuronal fibres, and thus may underestimate the true size of white matter tracts which often contain axons in slightly divergent paths. Also, there is often a shift of fibre tracts during the initial brain dissection, especially after debulking a large intracerebral mass, which can give a false sense of security if surgical planning is only based on pre-operative DTI images. This latter point emphasizes the need for the intraoperative acquisition of real-time DTI data to enable a sort of “continuous”, real-time neuronavigation [21].

If an intraoperative MRI (iMRI) is available, real-time DTI data can be obtained after the initial passage of the D³ probe and the first round of lesion resection, when the intraoperative images are acquired to display whether complete resection has actually been achieved. If not, and a second attempt is contemplated, then visualization in real-time of the eloquent fibre tract, inevitably shifted by the initial creation of the surgical corridor, would be enormously helpful when the brain retractor or even the probe assembly is to be re-inserted into the now partially collapsed surgical tunnel. The only caveat would be a prolongation of the iMRI interlude, for the post-acquisition processing of the DTI data can take up to 30 minutes.

With practice, the execution of the “Doigt-de-Dieu” technique should be straight-forward and carry minimal risk, but a few technical nuances must be mentioned. The probe assembly must be firmly pressed against the brain when inflating the balloon, or the back pressure exerted by the elasticity of the deforming brain will forcefully expel the whole probe. Once the working tunnel is created and the holding retractors are placed, no time should be wasted to tackle the lesion before natural tissue turgor rises to collapse the tunnel wall against the rigidity of the retractor blades. Also, the trajectory distance between the corticotomy and the nearest surface of the lesion must be accurately measured, which therefore demands that the registration error of the navigation platform be kept at a minimum. This way, the surface of the lesion can be immediately displayed in full measure without being abraded when the operating microscope is first brought in (Figure 4A). An inaccurately short tunnel will require supplementary sharp dissections with hard instruments for the last stretch of the surgical corridor, often the most perilous precinct in regards to functional preservation. Conversely, a tunnel-too-long may cause unintended scraping of the lesion by the probe tip and result in inadvertent bleeding within a very restricted surgical field if the lesion is highly vascular.

CONCLUSION

Combining it with preoperative DTI and intraoperative neuronavigation, the “Doigt-de-Dieu” technique allows for safer approaches to deep brain lesions, with minimal disruption of adjacent eloquent brain. This approach helps in the decision making of how and when to operate on previously deemed inoperable deep lesions that are situated near vulnerable grey matter and fibre tracts.

1.       Kelly PJ (1989) Future perspectives in stereotactic neurosurgery: stereotactic microsurgical removal of deep brain tumors. J Neurosurg Sci 33: 149-154.

2.       Recinos PF, Raza SM, Jallo GI, Renard Recinos V (2011) Use of a minimally invasive tubular retraction system for deep-seated tumors in pediatric patients. J Neurosurg Pediatrics 7: 516-521.

3.       Raza SM, Recinos PF, Avendano J, Adams H, Jallo GI, Quinones-Hinojosa A (2011) Minimally invasive trans-portal resection of deep intracranial lesions. Minim Invasive Neurosurg 54: 5-11.

4.       Almenawer SA, Crevier L, Murty N, Kassam A, Reddy K (2013) Minimal access to deep intracranial lesions using a serial dilatation technique: case-series and review of brain tubular retractor systems. Neurosurg Review 36: 321-329.

5.       Eliyas JK, Glynn R, Kulwin CG, Rovin R, Young R, et al. (2016) Minimally invasive transsulcal resection of intraventricular and periventricular lesions through a tubular retractor system: multicentric experience and results. World Neurosurg.

6.       Jo KI, Chung SB, Jo KW, Kong DS, Seol HJ, et al. (2011) Microsurgical resection of deep-seated lesions using transparent tubular retractor: pediatric case series. Childs Nerv Syst 27: 1989-1994.

7.       Harris AE, Hadjipanayis CG, Lunsford LD, Lunsford AK, Kassam AB (2005) Microsurgical Removal of intraventricular lesions using endoscopic visualization and stereotactic guidance. Neurosurgery 56: 125-132.

8.       Rosenorn J, Diemer NH (1985) The risk of cerebral damage during graded brain retractor pressure in the rat. J Neurosurg 63: 608-611.

9.       Witwer B, Moftakhar R, Hasan K, Deshmukh P, Haughton V, et al. (2002) Diffusion-tensor imaging of white matter tracts in patients with cerebral neoplasm. J Neurosurg 97: 568-575.

10.    Lansky SB, List MA, Lansky LL, Ritter-Sterr C, Miller DR (1987) The measurement of performance in childhood cancer patients. Cancer 60: 1651-1656.

11.    Andrews RJ, Bringas JR (1993) A review of brain retraction and recommendations for minimizing intraoperative brain injury. Neurosurgery 33: 1052-1064.

12.    Rosenorn J, Diemer NH (1982) Reduction of regional cerebral blood flow during brain retraction pressure in the rat. J Neurosurg 56: 826-829.

13.    Ogura K, Tachibana E, Aoshima C, Sumitomo M (2006) New microsurgical technique for intraparenchymal lesions of the brain: transcylinder approach. Acta Neurochir 148: 779-785.

14.    Fahim DK, Relyea K, Nayar VV, Fox BD, Whitehead WE, et al. (2009) Transtubular microendoscopic approach for resection of a choroidal arteriovenous malformation. Technical note. J Neurosurg Pediatr 3: 101-104.

15.    Herrera SR, Shin JH, Chan M, Kouloumberis P, Goellner E (2010) Use of transparent plastic tubular retractor in surgery for deep brain lesions: a case series. Surg Technol Int 19: 47-50.

16.    Laundre BJ, Jellison BJ, Badie B, Alexander AL, Field AS (2005) Diffusion tensor imaging of the corticospinal tract before and after mass resection as correlated with clinical motor findings: preliminary data., AJNR 26: 791-796.

17.    Kelly PJ, Goerss SJ, Kall BA (1988) The stereotaxic retractor in computer-assisted stereotaxic microsurgery. Technical note., J Neurosurg 69: 301-306.

18.    Kelly PJ, Kall BA, Goerss SJ (1987) Computer-interactive stereotactic resection of deep-seated and centrally located intraaxial brain lesions. Appl Neurophysiol 50: 107-113.

19.    Thomas C, Ye FQ, Irfanoglu MO, Modi P, Saleem KS, et al. (2014) Anatomical accuracy of brain connections derived from diffusion MRI tractography is inherently limited. PNAS 111: 16574-16579.

20.    Basser PJ, Pajevic S, Pierpaoli C, Duda J, Aldroubi A (2000) In vivo fiber tractography using DT-MRI data., Magn Reson Med 44: 625-632.

21.    Nimsky C, Ganslandt O, Hastreiter P, Wang R, Benner T, et al. (2005) Preoperative and intraoperative diffusion tensor imaging-based fiber tracking in glioma surgery., Neurosurgery 56: 130-138.