Intraoperative use of functional MRI for surgical decision-making after limited or infeasible electrocortical stimulation mapping.

Abstract

Background and Purpose

Functional magnetic resonance imaging (fMRI) is becoming widely recognized as a key-component of pre-operative neurosurgical planning, although intraoperative electrocortical stimulation (ECS) is considered the gold standard surgical brain mapping method. However, acquiring and interpreting ECS results can sometimes be challenging. This retrospective study assesses whether intraoperative availability of fMRI impacted surgical decision making when ECS was problematic or unobtainable.

Methods

Records were reviewed for 191 patients who underwent pre-surgical fMRI with fMRI loaded into the neuronavigation system. Four patients were excluded as a bur-hole biopsy was performed. Imaging was acquired at 3 Tesla and analyzed using the general linear model with significantly activated pixels determined via individually determined thresholds. fMRI maps were displayed intra-operatively via commercial neuronavigation systems.

Results

71 cases were planned ECS, however, 18 (25.35%) of these procedures were either not attempted or aborted/limited due to: seizure (10), patient difficulty cooperating with the ECS mapping (4), scarring/limited dural opening (3), or dural bleeding (1). In all aborted/ limited ECS cases, the surgeon continued surgery using fMRI to guide surgical decision making. There was no significant difference in the incidence of post-operative deficits between cases with completed ECS and those with limited/aborted ECS.

Conclusions

Pre-operative fMRI allowed for continuation of surgery in over one-fourth of patients in which planned ECS was incomplete or impossible, without a significantly different incidence of post-operative deficits compared to the patients with completed ECS. This demonstrates additional value of fMRI beyond pre-surgical planning, as fMRI data served as a backup method to ECS.

Introduction

Surgery of brain lesions and epileptic foci in eloquent cortex poses the challenge of achieving maximal resection while avoiding post-operative functional deficits. For brain tumors, maximal excision is critical as it increases the chance of stabilized disease, maximizes the efficacy of adjuvant therapies, and is a favorable prognostic factor for patient survival.^1–3 For patients with medically intractable epilepsy, resection of the seizure focus may present the best treatment option and improve quality of life.⁴ Therefore, localization of functional cortex surrounding the lesion or seizure focus is crucial, and electrocortical stimulation (ECS) has long been the gold standard for such cases.^5,6 However, functional MRI (fMRI) is now also widely accepted as a pre-operative technique to localize eloquent cortex in patients with brain lesions and epilepsy, and can be applied in clinical practice.^7,8 fMRI can be used prior to surgery to determine the proximity of eloquent cortex to a lesion, and hence whether the patient has a favorable surgical risk for resection,^9,10 and if so, whether ECS should be performed.¹¹

fMRI utilizes blood oxygen level-dependent (BOLD) signal which arises due to the hemodynamic response to neuronal activation.¹² Statistical analysis reveals areas with significantly higher signal during task performance, which ECS has shown to correspond with eloquent brain regions.^13–16 fMRI also has advantages over ECS as it is non-invasive and can identify functional regions within sulci as well as beyond the planned extent of exposure, while ECS is invasive, limited due to exposure and time constraints, and results are only available for the exposed brain.¹⁷

After being implemented as a research effort at multiple neurosurgical centers, fMRI was approved for clinical use in the United States when the Centers for Medicare and Medicaid Services (CMS) established current procedural terminology (CPT) codes for its use in surgical planning.^7,8 Typically, surgeons request fMRI for functions relevant to eloquent cortex surrounding a lesion, including language, motor, or vision. Pre-operative review of fMRI provides the surgeon with information on patients’ surgical risk, and allows for pre-operative planning that may lead to reduced craniotomy size, reduced anesthesia time, and increased extent of resection.⁹ Subsequent integration of these maps into neuronavigation systems during the procedure can also guide intraoperative decision-making, such as craniotomy location, and deployment of stimulation points during ECS.^18–23 This allows for streamlining of cortical mapping to make it more efficient, less time consuming, and less demanding of the patient and team by selecting sites to test based on the displayed BOLD activations.^9,20

Although ECS is considered the gold standard for cortical mapping, some patients cannot tolerate the awake craniotomy necessary for many ECS procedures. ECS has also been associated with intraoperative seizures in some cases.^24,25 Additionally, ECS requires a dedicated multi-disciplinary team including specialized anesthesia, neuropsychology, and neurophysiology, which may not be available in all centers. With the challenges related to ECS in mind, we performed a retrospective review of neurosurgical cases with fMRI maps loaded into the neuronavigation system during surgery to determine the intraoperative utility of BOLD maps including the event of ECS diagnostic failure or incomplete mapping.

Methods

Subjects

191 cases were reviewed for this study (94 female; 19–77 years). This is a subset of 586 patients who had pre-surgical fMRI performed at our institution between December 2003 and February 2013. Patients were selected who: 1) harbored brain lesions or epileptic foci in or adjacent to eloquent brain regions 2) underwent fMRI for pre-surgical planning, and 3) had fMRI data loaded into the neuronavigation system during surgery. The clinical record review research protocol was approved by our Institution’s Internal Review Board, and written and informed consent was obtained prior to acquiring research data.

Functional Paradigms

Functional imaging sessions included clinically relevant behavioral paradigms for each patient (see table 1). All motor, sensory, and vision tasks were implemented in a block design with alternating task and rest epochs; language tasks were either a block or event related design. Various software was used to display paradigms depending on the scanner used (E-Prime (Psychology Software Tools, Pittsburgh, PA, USA) Presentation (Neurobehavioral Systems Inc., Davis CA, USA) or nordicAktiva (NordicNeuroLab, Bergen, Norway)) via MRI compatible hardware (Resonance Technology, Los Angeles, CA, USA or NordicNeuroLab, Bergen, Norway).

Table 1.

Summary of functional MRI task paradigms.

Motor tasks	Hand clench	Repetitive unilateral hand clenching vs. rest
	Finger tap	Sequential unilateral finger tapping vs. rest
	Toe wiggle	Repetitive unilateral toe flexion/extension vs. rest
	Lip purse	Repetitive lip pursing (puckering) vs. rest
Sensory Stimulation	Hand or foot (brushing)	Hand (palmar) or foot (plantar) surface repetitively brushed with rough cloth vs. rest
Language tasks	Antonym generation	Subvocalized or silent antonym generation vs. rest (crosshair fixation)
	Noun categorization	Vocalized or silent noun categorization, “alive” vs. “not alive” vs rest (crosshair fixation)
	Verbal fluency	Word generation, beginning with specified letter vs. rest (crosshair fixation)
	Sentence Completion	Fill in the blank sentences (silent) vs. scrambled letters (high level baseline)
Vision tasks	Flashing checkerboard	Black and white flashing checkerboard pattern vs. crosshair fixation

Appropriate clinically relevant behavioral paradigms were selected for each patient based on lesion location and adjacent eloquent cortex. The task or stimulation performed is described, as well as the control/baseline period. vs. = versus.

MRI Acquisition & Analysis

All imaging was acquired using 3 Tesla MRI scanners (General Electric, Milwaukee, WI, USA (Signa) or Siemens, Erlangen, Germany (Trio, Vario, Skyra)). BOLD fMRI was acquired via whole brain T2* weighted echo-planar imaging (EPI) optimized for the scanner used (27–32 axial slices, 4–5 mm slice thickness, no slice gap, repetition time (TR) = 2000 ms, echo time (TE) = 30–40 ms, flip angle = 85–90^o, in plane resolution of 2 mm² to 3.4375 mm², 22–24 cm field of view, matrix = 64×64 or 80×80 mm.) Task paradigms were synchronized with the acquisition, which had durations of 3 to 7 minutes. High resolution T1 and/or T2 weighted anatomical images were acquired for surgical navigation and co-registration to the functional volumes.

The analysis software used depended on date of acquisition: from 2003 −2009, SPM software was used: (Welcome Department of Imaging Neuroscience, London, UK.) After 2009 software packages approved for clinical use by the United States Food and Drug Administration were used (2009–2012: GE Brainwave (General Electric, Milwaukee, WI, USA); 2012–2013: NordicBrainEx, (NordicNeuroLab, Bergen, Norway)). All functional data were motion corrected, spatially smoothed, co-registered to the anatomical volume, and analyzed using the general linear model; a hemodynamic response function (HRF) was produced based on the task paradigm, and statistical parametric maps were created based on the t-score correlation between the HRF and the voxel by voxel BOLD signal.²⁶ Significantly activated voxels were determined via individually determined thresholds per task acquisition. See figure 1 for example BOLD maps.

Figure 1.

Example functional MRI maps. A: Language task activations for patient 39 (t-score = 5.1). B: motor task (lip pursing) activations for patient 41 (t=8.5). C: Vison task activations for patient 76 (t-score =15.5). Axial slices are displayed in radiological convention in patients’ native space used for surgical planning.

Integration into the Neuronavigation System

fMRI was integrated into neuronavigation systems (GE Instatrack 3500, GE Healthcare Navigation, Lawrence, MA, USA, or Brainlab, Munich, Germany) by converting clusters of functional activation at chosen threshold(s) to a value above the highest voxel intensity in the co-registered grayscale anatomical image. A burnt in voxel map was created of these activations, overlaid onto the structural image, and converted to Digital Imaging and Communications in Medicine (DICOM) format with either an SPM toolkit (wfu_dicomtk; Wake Forest University; http://fmri.wfubmc.edu/ ) or NordicBrainEx software. The navigation system’s software was used to co-register the DICOM image to structural images and to segment the BOLD regions by selecting only supra-intensity areas from the burnt in map.²¹ The segmented maps were used to create a 3D dataset for re-slicing and 3D rendering for intraoperative viewing to assist surgical decision making. See figure 2 for example neuronavigation views during surgery.

Figure 2.

Example integration of functional MRI maps into the neuronavigation system for intraoperative guidance (screenshot: patient 39). The tumor is outlined in green, and language maps in orange and pink. Purple represents tractography. A: Axial View. B: Coronal view, both showing the neuronavigation pointer device’s trajectory (hatched green line). Images are displayed in neurological convention.

Electrocortical mapping Procedure

An Ojemann bipolar stimulator (Integra Life Sciences, Plainsboro, NJ, US) was used for ECS testing using previously described techniques (stimulator amplitude 2–10 mA, tip spacing 5 mm, and waveforms 5 to 60 Hz depending on pathology and area of interest).⁵ Patients were asked to perform language or motor tasks during ECS testing, or to report any tactile or visual sensations. For ECS motor mapping acquired under general anesthesia appropriate neuromonitoring leads were placed. The surgeon could mark stimulation sites using digital notations in the navigation software if desired.

Records Review and Statistical analysis

Clinical notes were retrieved and details pertaining to the ECS procedure were logged, as well as post-operative outcomes at 6 months. Post-operative function was compared to pre-operative function to determine if a new or significantly worsened major neurological deficit was present. A chi square test was performed using STATA Statistical software (Version 14, StataCorp LP, College Station, Texas, USA) to determine if there was a significant difference (p < 0.05) in post-operative deficits between the successful and limited/aborted ECS case groups. A chi square test was also performed between the entire group that had an ECS procedure planned, and the group that did not have ECS planned.

Results

All 191 patients underwent surgery at our institution for treatment of their brain lesion or epilepsy with fMRI maps loaded into the neuronavigation system to guide surgical decision making (i.e. corticectomy location). Four patients were excluded from analysis, as a closed biopsy was performed, precluding the use of ECS. Of the remaining 187 patients, 71 were performed with ECS, either awake under monitored anesthesia care (MAC) (60), or under general anesthesia with motor mapping (11). Successful ECS mapping procedures were completed in 53 ECS cases. However, the ECS mapping procedure was limited or not completed in 18 ECS cases (25.35 %). Reasons for aborted or limited ECS mapping included: intraoperative seizure or seizure activity (10), patient difficulty in cooperating with the ECS task resulting in ambiguous or uninterpretable results (4), limited dural opening or scarring from previous surgeries (3), and excessive bleeding (1); see figure 3 for summary. In all 18 aborted/limited ECS cases the surgeon continued the resection with fMRI available to guide surgical decision making. See table 2 for pathology and ECS detail for all 187 patients, and table 3 for additional details of the 18 aborted/limited cases. Details of example cases in each category are provided below.

Figure 3.

Summary of all neurosurgical cases using functional MRI in the neuronavigation system. ECS= electrocortical stimulation. MAC = monitored anesthesia care.

Table 2.

Pathology results and electrocortical stimulation details for all cases.

Pathology:	Number of cases:	Number with ECS planned:	(Number of ECS complications) [Reason]
Low Grade Glioma	3	0	(0) [No complications]
Possible Glioma	3	3	(1) [Planned but not done due to scarring]
Astrocytoma II	4	2	(1) [Limited due to seizure]
Astrocytoma III	14	6	(3) [Limited due to seizure=2, or not done due to restricted dural opening=1]
Oligo II	17	8	(2) [Not done: seizure=1, limited: seizure=1]
Oligo III	8	5	(1) [Not done: bleeding]
Oligo/Astro II	9	3	(1) [Limited: ambiguous language mapping]
Oligo/Astro III	6	3	(1) [Limited due to seizure]
GBM	67	26	(4) [Not done: seizure=1 scarring=1, or limited: ambiguous language mapping =2.]
Metastasis	32	7	(3) [Asleep motor limited due to seizure=2, limited language mapping results=1]
Meningioma	8	0	(0) [No complications]
Treatment effect	1	0	(0) [No complications]
AVM/Cavernoma	6	4	(1) [Limited due to seizure]
Epilepsy	9	4	(0) [No complications]
TOTALS	187	71	(18) [See above]

ECS = electrocortical stimulation, Oligo = Oligodendroglioma, Oligo/Astro=oligoastrocytoma, GBM = Glioblastoma, AVM/Cavernoma = Arteriovenous malformation or cavernoma, MET= metastasis.

Table 3.

Summary of the 18 aborted/limited ECS cases.

	Type of ECS planned	Aborted/limited	Reason ECS was limited/aborted	Pathology
1	Motor	Limited acquisition	Focal Seizure	Astrocytoma III
2	Language	Aborted (not done)	Generalized seizure with airway compromise prior to intervention.	GBM
3	Language	Aborted (not able to interpret)	Proved too difficult for the patient to perform	GBM
4	Language	Limited acquisition	Focal seizure activity (electrographic)	Astrocytoma II
5	Motor	Aborted (not done)	Scaring and multiple dural adhesions from prev. surgeries	GBM
6	Language	Aborted (not able to interpret)	Patient had difficulty communicating responses	MET
7	Language	Limited (difficult to interpret)	Bilingual patient, mixed languages when responding	Olig/Astro II
8	Motor	Limited acquisition	Significant motor seizure	AVM
9	Language	Limited (difficult to interpret)	Difficulty communicating: inconsistent responses	GBM
10	Motor	Aborted (not done)	Generalized seizure after removal of bone flap	Oligo II
11	Motor (asleep)	Limited acquisition	Focal seizure	MET
12	Language	Aborted (not done)	Significant epidural bleeding	Oligo III
13	Motor	Limited acquisition	Focal seizure	Astrocytoma III
14	Motor	Limited acquisition	Focal seizure activity (electrographic)	Olig/Astro III
15	Motor	Aborted (not done)	Limited dural opening due to large venous lake	Astrocytoma III
16	Motor	Aborted (not done)	Dura was scarred to underlying brain due to previous surgery	Possible Glioma
17	Motor (asleep)	Limited acquisition	Generalized seizure during ECS	MET
18	Motor	Limited acquisition	Focal seizure	Oligo II

Aborted cases either had no electrocortical stimulation (ECS) points collected, or points were not interpretable. For the limited cases ECS was restricted. Oligo = Oligodendroglioma, Oligo/Astro=oligoastrocytoma, GBM = Glioblastoma, AVM/Cavernoma = Arteriovenous malformation or cavernoma, MET= metastasis.

Incomplete or Aborted ECS Case Details

Ten patients experienced a seizure in the operating room, and ECS was either aborted, limited, or not attempted. In two of these cases, a generalized seizure occurred prior to ECS and hence ECS was not done (see figure 4 for example images from one of these cases). In the other cases ECS was either limited or aborted to avoid further seizure activity.

Figure 4.

Example pre and post- operative images for one patient (see patient 10 in table 3). ECS was planned, however the patient had a generalized seizure after bone flap removal, hence electrocortical stimulation was not performed. A: Pre-operative T1 weighted image without abnormal regions of contrast enhancement. B: Pre-operative fMRI (right finger tapping, t-score= 5.77) overlaid on a T2 weighted image, showing BOLD activation adjacent to the lesion. C: Post -operative T1 weighted image showing the resection cavity. Pathology revealed an Oligodendroglioma Grade II. Images were retrieved from the hospital PACS system and are displayed in patient native space in radiological convention.

Four patients had difficulty cooperating with ECS mapping. One patient had fMRI and intraoperative mapping performed in two languages. Intra-operatively, however, responses were mixed between languages and difficult to interpret. For three other patients, ECS language testing proved too difficult for the patient to perform.

Three cases had limited dural openings that precluded ECS; 2 had undergone multiple previous surgeries, which resulted in significant scarring and dural adhesions that limited exposed brain surface, and for 1 patient the dural opening was limited by a large venous lake. Finally, one case had significant epidural bleeding that precluded planned ECS. The surgeon continued with resection referencing the intraoperatively displayed fMRI for all cases above.

Post-operative functional outcome

Of the 71 cases with ECS planned, two were lost to follow up, and of the remaining 69 cases, 7 had new or significantly worsened post-operative neurological deficits (10.14%). In the limited/aborted ECS group, one case lost to follow up, and out of the remaining 17, one had a significant post-operative deficit (5.88%). In the successfully completed ECS group, one case was lost to follow up. Out of the remaining 52, 6 had significant deficits (11.54%). Excluding the cases lost to follow up, Pearson’s chi square calculation revealed no significant difference in the incidence of post-operative deficits between completed ECS cases and limited/aborted ECS cases (Chi Square statistic = 0.4496 p = 0.50). In the 116 patients who did not have ECS planned or performed, one case was lost to follow up, and 4 had new or significantly worsened post-operative neurological deficits (3.48%). Chi square calculation did not reveal a significant difference in post-operative deficits between the ECS and non-ECS group (Chi Square statistic = 3.4099, p=.065).

Discussion

ECS has long been the gold standard for mapping functional regions surrounding brain lesions, since neurosurgical resection requires careful planning of the surgical approach to achieve maximal resection while avoiding damage to surrounding eloquent cortex. However, even in experienced hands ECS can pose significant challenges or even fail, as evident in the data presented here. In over one quarter (25.35%) of cases with ECS mapping, the procedure was either aborted or limited, yet, in all these cases, intraoperatively displayed fMRI was used as a guide for successful continuation of surgery, without an increased incidence of new or worsened post-operative neurological deficits. This represents an additional value of fMRI beyond pre-operative surgical planning: use for guidance intraoperatively to avoid eloquent regions in the event of an aborted or limited ECS procedure. There is a previous case report in which resting state functional connectivity mapping was referenced after an aborted awake craniotomy following loss of airway.²⁷ However, the present report details the first case series in which task based fMRI was utilized as a fallback after aborted or incomplete ECS.

fMRI for Surgical Guidance and Concordance with ECS

The use of fMRI for surgical planning has been well described previously, including pre-operative risk stratification, selection of patients for ECS procedures based on distance of functional areas to a lesion, and planning corticectomy location and surgical trajectory.^28,29 After integration into the neuronavigation system, fMRI can also guide surgical decision making.^{21,22,30–32} Many studies have also detailed the concordance between fMRI and ECS. As early as 1997, FitzGerald et al. found that when multiple fMRI tasks were implemented, sensitivity of fMRI was 81% when fMRI activations overlapped with positive ECS points, and 92% when points were within 2 cm of fMRI activations.³³ Krings et al. also looked at the correlation between fMRI and ECS as well as positron emission topography (PET), and transcranial magnetic stimulation (TMS), and found overlapping or neighboring results when comparing fMRI to either ECS or TMS.¹⁵ In 1999 Roux et al. found that motor fMRI and ECS points matched accurately,³⁴ then in 2013 also compared language fMRI results with ECS, but found that BOLD activations were imperfectly correlated with stimulation points, although better correlations could be made by combining fMRI results from multiple tasks.³⁵

ECS Failure Rates

We report a high rate of limited/aborted ECS cases (25.35%), with a portion of these due to seizure (14.08% of all planned ECS cases). Previous studies have reported a range in the failure rates reported for ECS mapping procedures, with many reports detailing seizure occurrence during ECS. Serletis et al. report 25 out of 551 (4.9%) of awake craniotomy patents experienced intraoperative seizures.³⁶ Kim at al. reported intraoperative seizures in 9% of cases during awake craniotomy for patients with tumors near eloquent cortex, although the majority of these (23 out of 27) were focal and controlled with iced saline.³⁷ Boetto et al. report a lower rate of intraoperative seizure during awake craniotomy (3.4%) with no ECS failures, and conclude that ECS can be performed safely without electrocorticography.³⁸ However, in a case series of patients under general anesthesia for motor mapping, Cordella et al. report a much higher seizure rate of 36%, and state that intraoperative seizures may complicate ECS preventing further stimulation testing.²⁴ Nossek et al. also report failure rates of awake ECS, with one case series reporting 12.6% experiencing intraoperative seizures and 2.3% failing due these seizures, and in another series, they a report an overall failure rate of 6.4% for awake craniotomy, with 2.1% failing due to seizures, and 4.2% falling due to lack of communication during the mapping.^39,40 In a retrospective analysis, Tatum et al. report that 5.7% of patients had seizures during ECS.⁴¹ The seizure rate in this report (14.08%) differs from those reported in previous studies, but this includes both aborted/failed ECS as well as limited/incomplete ECS mapping, as we included 2 cases in which ECS was planned but not attempted due to a generalized seizure occurring prior to ECS deployment. Standard techniques were used for all patients with ECS mapping reported in this study,⁵ however multiple surgeons performed the procedures and it is unclear if this contributed to the seizure rate. The total percentage of limited/aborted ECS cases in this report (25.35%) is higher than descriptions of ECS failure in the abovementioned studies, however we included limited ECS cases as well as failed or aborted mapping, and included situations in which ECS mapping may fail besides intraoperative seizure, including ambiguous results, surgical restrictions such as scarring/limited dural opening, and excessive bleeding.

Study Limitations

This retrospective review examines cases with fMRI maps integrated into the neuronavigation system during surgery. This is therefore a limited dataset and does not include all ECS cases performed during this time, and the ECS complication rate may be different in a larger sample, or in a group of patients in which the surgeon chose not to obtain pre-operative fMRI. Also, the ECS point locations were not always recorded in the neuronavigation software during the procedure, hence fMRI locations could not be correlated with ECS data in this series. Additionally, although fMRI can be referenced intraoperatively in the neuronavigation system, it should be noted that after surgery commences, brain shift can alter the registration of the pre-surgical fMRI results to the patient, as can also occur for the pre-resection ECS coordinates and structural imaging.^31,42 Other limitations of BOLD imaging itself also apply to the fMRI data acquired in this series, such as neurovascular uncoupling due to abnormal brain tumor vasculature.⁴³

Further Research

Although performing a prospective study comparing patients with fMRI only to patients with ECS only is difficult due to ethical considerations and patient and surgeon preferences, future studies should attempt to further correlate BOLD activations with ECS data point location and patient outcome, as there is some disagreement on their concordance in the literature.^34,35 Additional validation of fMRI to ECS data points could increase confidence in using fMRI data alone for intraoperative decision making, especially when ECS acquisition proves problematic. Future research should also address the limitations of this study discussed above, especially prospective studies in patients with low grade gliomas, since increasing evidence shows that complete resection prolongs survival.^1,44,45

Additionally, the utility of integrating other non-invasive modalities, such as those to locate functional cortex, (Magnetoencephalography, TMS), locate active tumor (PET, perfusion imaging), and white matter tracts, (diffusion tensor imaging (DTI)), into the neuronavigation system should be investigated. These datasets could also provide the surgeon with additional information to use in surgical decision making.^21,46,47 Future advancements in these modalities, including improved acquisitions, resting state fMRI in lieu of task based, and improved tractography algorithms, will also provide improved data detailing critical brain regions to compliment or be used in place of ECS.^46–48 Of note, as minimally invasive neurosurgical approaches are developed, such as laser interstitial thermal ablation and focused ultrasound, reliance on non-invasive mapping techniques such as fMRI is likely to increase, as these approaches do not necessitate a craniotomy and hence do not allow for ECS.⁴⁹

Conclusion

This retrospective review has identified a subset of cases (25.35%) in which ECS mapping was aborted or incomplete, yet the surgeon felt confident to continue the operation with BOLD fMRI maps displayed in the neuronavigation system. Additionally, there was no significant difference in the incidence of post-operative deficits between cases with completed ECS and those with limited/aborted ECS. These findings indicate an additional use of fMRI beyond pre-surgical planning: a fallback method to aid intra-operative decision-making when planned ECS was impossible or incomplete.