Patient population & data collection

After institutional review board approval (IRB# JWCI-19-1101), all patients at Saint John’s Health Center, Santa Monica, CA, between January 2008 and January 2021 who underwent surgical removal of an intracranial meningioma were identified. Patient consent was not necessary as data was deidentified. All operations were performed by one of two neurosurgeons (DK, GB) and endonasal operations were performed with otolaryngology collaboration (CG). Data collection included prior treatments, tumor histopathology, size and location on MRI, endoscope usage, extent of resection, complications, length of stay, disposition, readmissions and long-term tumor control.

MIS “keyhole” definition and approach selection

Keyhole approaches were defined as: the use of a minimally invasive approach for a given tumor in which a traditional larger approach may be used instead. These six approaches included the endoscopic endonasal, supraorbital, mini-pterional, retromastoid, suboccipital tentorial and transfalcine routes. This definition follows that detailed by Lan et al in their 2019 consensus paper on keyhole microneurosurgery and aligns with our definition from prior publications on keyhole surgery [21, 27, 28] (Fig 1A & 1B, S1 Fig, S1 Table). Non-keyhole approaches include convexity craniotomies, pterional, bifrontal and far-lateral approaches. Table 1 describes the strategy for applying these 6 keyhole approaches and two videos provide 6 case examples. During the study period, traditional open skull base craniotomies including orbito-zygomatic, transpetrosal, translabyrinthine, transcondylar craniotomies were not used. Four patients who underwent a staged operation using one or more keyhole approaches are identified and each keyhole approach was considered as a separate outcome.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1.

(1A) Drawing depicting 6 keyhole approaches for meningioma removal: endonasal, supraorbital, minipterional, retromastoid, suboccipital sitting gravity-assisted and transfalcine gravity-assisted, along with the total number of surgeries for each approach. (1B) A composite of the 6 keyhole approaches.

https://doi.org/10.1371/journal.pone.0264053.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Surgical decision making for keyhole meningioma removal.

https://doi.org/10.1371/journal.pone.0264053.t001

Surgical goals, preparation and technique

A goal of maximal safe tumor removal was applied across all tumor locations. However, subtotal removal was typically the goal for three meningioma subtypes given their invasive nature and tendency to encase neurovascular structures: petroclival, CS/Meckel’s cave, and spheno-cavernous locations, as well as some meningiomas with prior surgery and/or radiotherapy [32]. Residual meningioma was deliberately left along neurovascular structures if too adherent or left in the skull base bone if deemed too infiltrative. Four patients had planned staged approaches for large meningiomas (maximal diameter >6 cm: endonasal debulking combined with a pterional, mini-pterional or retromastoid approach.

Total intravenous anesthesia (TIVA) was used to promote rapid emergence from anesthesia [34]. Preoperative tumor embolization was not used. Lumbar drains for CSF diversion were used infrequently and only early in the series. As recently described, all endonasal cases were performed fully endoscopically using a 0° 4-mm rigid endoscope initially; 30° and 45° endoscopes (Karl Storz-America, El Segundo, CA), are used at various stages of the procedure [27, 32]. A pedicled nasoseptal flap (or middle turbinate flap) is harvested in cases of tuberculum and planum meningiomas but not for all CS/Meckel cave meningiomas, depending upon CSF leak grade [27, 32, 33].

Except for the fully endoscopic sitting suboccipital approach, all transcranial keyhole craniotomies are performed initially with the microscope and then with endoscopy as needed for illuminating areas poorly seen with the microscope, and for endoscopic tumor resection. Endoscopy is typically performed with an assistant driving the endoscope. Patient positioning is critical to optimize gravity-assisted exposure and for endoscopy. Our use of the supraorbital and mini-pterional approaches have been well-described [27, 34, 36]. The retromastoid approach is performed with the patient supine or lateral position, utilizing a short linear or curvilinear incision with approximately 2x3cm craniotomy [19]. The suboccipital approach is performed in the sitting or lateral position typically using an approximate 2x2cm craniotomy for tentorial meningiomas. The gravity-assisted transfalcine approach is performed specifically for parafalcine meningiomas that have a rind of edematous overlying eloquent cortex at risk from an ipsilateral approach. The patient is placed in lateral position with tumor side up and inclined in reverse Trendelenburg position with the vertex at 30–45 degrees [29], allowing for gravity-retraction and access from the unaffected contralateral side.

Outcome and statistical analysis

The primary and secondary outcomes analyses detailed below attempt to answer the question of benefit of keyhole surgery approaches in terms of neurological and overall surgical recovery. Primary outcomes included extent of tumor resection, complications, hospital LOS, reoperations and readmissions within 90 days, 30-day mortality, KPS, and cranial (CN) outcomes. A favorable KPS at ≥90 days follow-up was defined as improved or unchanged from preoperative KPS. As previously published, resection rates were categorized as: gross total resection (GTR), near total resection (NTR ≥ 90% tumor resection), or subtotal resection (< 90%) [27, 28]. Simpson resection classification was not assessed given that most patients had invasive skull base meningiomas in which complete tumor resection and removal of involved bone and dura is typically not feasible [27, 28, 37].

Three secondary outcomes were included: 1) To quantify approach-related trauma of tumor removal, post-operative day #1 or 2 MRIs were independently assessed by a neuroradiologist (JR) for new FLAIR/T2 signal changes in the peritumoral and surgical approach areas [27]. New FLAIR/T2 change in the axial plane around the resection area or approach trajectory was quantified by maximal diameters. Persistence or resolution was documented on 3-month postoperative MRI. 2) To assess the utility of endoscopic visualization for transcranial operations, as previously published, it was determined if endoscopy facilitated additional tumor removal [27, 34] 3) To assess tumor progression or recurrence, sequential postoperative MRIs and need for additional treatment (repeat surgery and/or radiotherapy) were quantified for each patient.

The statistical comparison of the mean in the data amongst different groups was performed using ANOVA and independent Student’s t-test. Univariate analysis was done using Chi-square or Fisher’s exact test when applicable. Binomial multivariate analysis was done using logistic regression analysis and p-values less than 0.05 were considered statistically significant. Tumor control data for testing binomial variables were plotted using Cox Regression and Kaplan-Meier analysis.

Demographics

Of 329 patients, keyhole approaches were used in 193 (59%) (74% female, mean age 59±13, mean follow up 42±36 months; WHO Grade I: 173/193(89.6%), Grade II: 19/193(9.8%), Grade III: 1/193(0.6%)); 32(9.7%) had prior radiation. The 213 keyhole operations included 205 (96%) for skull base location and 47(22%) were in patients with prior surgery. Operative approaches used included: endonasal (n = 74), supraorbital (n = 73), retromastoid (n = 38), mini-pterional (n = 20), sitting suboccipital (n = 4), and contralateral gravity-assisted transfalcine (n = 4) (Table 2). Four patients had planned staged operations using two approaches for large meningiomas (endonasal combined with a pterional n = 1, supraorbital n = 1, mini-pterional n = 1 or retromastoid approach n = 1). Lumbar drains were used in 5/74(7%) endonasal operations all prior to 2013, and 1/139 (0.7%) transcranial operation in 2017.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Meningioma characteristics, resection rates & surgical parameters for 213 keyhole operations.

https://doi.org/10.1371/journal.pone.0264053.t002

Primary outcomes

Resection rates by approach and tumor location.

As shown in Tables 2 and 3, GTR was achieved in 85(40%) operations while GTR/NTR (>90% resection) was achieved in 125(59%) operations. By meningioma location, GTR/NTR rates were highest for frontal fossa, parafalcine tentorial, olfactory groove and planum meningiomas (N = 158, range 84–100%), and lowest for invasive spheno-cavernous, CS/Meckel’s cave and petroclival locations (N = 55, range 5–6%) (Table 3). Excluding these invasive subgroups (petroclival, CS/MC and spheno-cavernous), GTR/NTR was achieved in 114/136 (84%) first-time operations. Dense adherence to neurovascular structures was associated with 36% GTR/NTR versus 100%, without dense adhesions (p<0.001) (S2 Table). All patients with STR/NTR had at least one risk factor (prior surgery, radiotherapy, invasion cavernous sinus/Meckel’s cave/infratemporal fossa/orbit).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Resection rates in 213 keyhole approaches by meningioma location, invasiveness & use of endoscopy.

https://doi.org/10.1371/journal.pone.0264053.t003

Clinical outcomes and complications.

Major complications occurred in 23/193 (11.9%) patients (Table 4). Permanent neurological worsening occurred in 12(6%) patients including 4 with strokes and 8 with cranial nerve injuries (S3 Table). Reoperations were needed in 10(5%) patients, (two for CSF leaks); 2(1%) patients had meningitis No patients developed perioperative DVT, PE or MI. Three patients all with multiple prior surgeries, died of disease progression at 80, 229, and 3508 days after their last operation.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Major and minor complications, readmissions, reoperations and discharge status in 193 patients undergoing 213 keyhole operations for meningioma.

https://doi.org/10.1371/journal.pone.0264053.t004

Secondary outcomes

MIS philosophy applied to meningiomas

MIS techniques aim to minimize collateral damage to normal tissues while achieving the surgical goal, be it tumor resection, valve replacement or disc replacement [12, 13, 15, 38, 39]. Striking that balance while being cost-effective and maximizing patients’ quality of life is the ultimate “sweet spot” of any surgical procedure [38, 40]. We propose that the present series of MIS approaches and prior series from our group and others for intracranial meningioma are a step forward for neurosurgery and endoscopic skull base surgery, demonstrating the potential effectiveness of this keyhole paradigm [19–21, 28, 34]. In attempting to answer our central question, we suggest these results show a potential benefit of keyhole meningioma surgery on neurological and overall patient recovery, and thus offer a reasonable and in some instances, a preferred alternative to traditional approaches.

This evolution of keyhole surgery is in part a result of cross-specialty collaboration. The adoption of the endoscope into transsphenoidal surgery that began in the 1990s with our colleagues in otorhinolaryngology, has transformed not only pituitary surgery but also the entire field of skull base surgery. Most pituitary adenomas are now removed through an endoscopic endonasal approach, and as we show here, many midline skull base meningiomas can be removed via the endonasal route [27, 41, 42]. Furthermore, the endoscope was used in almost two thirds of transcranial keyhole approaches and helped facilitate additional tumor removal in 55% of these operations.

Keyhole meningioma surgery aims to limit brain exposure and manipulation, accessing tumors through smaller strategically placed craniotomies, including ventral skull base craniotomies via the endoscopic endonasal approach, without static brain retractors, facilitated by gravity-assistance, low profile instrumentation and endoscopy, with the goal of achieving maximal safe tumor removal [17–21]. Despite the benefits of visualization, there is ample evidence that fixed brain retractors can cause acute and lasting brain injury [17, 43]. Recent reports highlight the potential for retraction injury and associated complications with traditional craniotomies for tuberculum and planum meningiomas [44–46]. A major benefit of keyhole retractorless and gravity-assisted endoscopic approaches and the endoscopic endonasal approach may be in less brain exposure and parenchymal manipulation [17, 18, 20, 21, 27, 46, 47]. This advantage was evidenced by the absence of early postoperative FLAIR/T2 changes along the surgical corridor in 82% of patients in this series including 100% of endonasal approaches; for 36 patients with FLAIR/T2 increases, these were small and resolved in all but 5% of patients.

For a neoplasm as diverse in location and invasiveness as meningioma, the surgical team should be facile working through multiple surgical corridors with both microscope and endoscope (see Table 1). Using TIVA anesthesia, smaller scalp and muscle incisions, minimal monopolar cautery, and focused craniotomies to minimize brain exposure, appears to promote rapid healing, less post-operative pain and a greater willingness and ability for patients to mobilize and leave the hospital [34, 35]. Adoption of MIS techniques have been a key component of success in enhanced recovery protocols in other surgical subspecialties which we are adopting as well in all of our brain tumor patients [34, 48]. Having reliable skull base closure techniques especially for endonasal, supraorbital and retromastoid routes where bony sinus and mastoid air cell entry frequently occurs is also essential to avoid CSF leaks and meningitis. This preventative strategy includes liberal use of abdominal fat grafts to obliterate sinus or air cell entry after craniotomy and a graded repair approach for skull base reconstruction including fat grafts and nasoseptal flaps in endonasal surgery without lumbar CSF drainage [33, 36, 49].

Complication rates, functional recovery and LOS compared to prior reports

To be valid and to show benefit in terms of neurological and overall surgical recovery, the MIS keyhole concept should yield low rates of new neurological deficits, CSF leaks and high rates of functional outcomes. Our results in terms of infrequent complications, short LOS and improved post-operative KPS scores support this approach. New cranial neuropathy was observed in only 4% of patients. Lumbar drains were used in only 3% of operations, while post-operative CSF leak rate and meningitis rates were only 1%. There were no cases of perioperative DVT, PE, MI or 30-day mortality. The absence of thromboembolic events is likely due in part to our high patient functionality with few neurovascular complication, early ambulation and limited perioperative narcotic use, and compares favorably to the 2.7–4.1% incidence of PE/DVT recently published [50, 51].

Prior studies and our experience demonstrate that complications impact quality of life, lengthen hospital stay, increase costs and often require reoperations and readmissions [29, 36, 38–40]. Compared to our outcomes of 3-day median LOS, 94% discharged to home, and 7% 90-day readmission rate, and no 30-day mortality, recent reports encompassing all intracranial meningiomas have documented LOS for skull base and other meningioma patients ranging from 4–11 days, discharge to home ranging from 70–83.4%, 90-day readmission rates of 9.2–17.9%, and 30-day mortality ranging from 0–5.4% [52–56]. This comparison is notable, given our series is comprised predominantly (96%) of skull base meningiomas (which are generally considered to be of higher complexity and risk profile than convexity meningiomas), while these 5 series include all intracranial meningiomas. A rapid complication-free recovery also benefits those patients who may have more aggressive (WHO Graded 2 or 3) or previously-treated meningiomas who may need to begin adjuvant therapy shortly after surgery [37].

In our recent publication on streamlining brain tumor care during the COVID-19 pandemic, we highlighted the importance and utility of an all-encompassing minimally invasive 360-degree approach to brain tumor surgery (for all tumor types), as a foundation for being able to not monitor many patients in the ICU and allowing for early hospital discharge, typically on POD#1 or #2 [35]. In this case control series of 293 patients, from pre-pandemic to pandemic, ICU utilization decreased from 54% to 29% of operations (p<0.001) and hospital LOS≤1 day increased from 12.2% to 41.4%, p<0.001, respectively. We believe that more routine and rigorous patient education on early discharge, recovery room assessment for non-ICU admission and earlier mobilization, layered upon a foundation of MIS, TIVA anesthesia and early post-operative imaging contributed to these significant and favorable trends (including our median LOS of 2 days in the last tertile of patients in this meningioma cohort).

Balancing goals of maximal tumor removal and complication avoidance

Perhaps the most serious critique of keyhole meningioma surgery is that ultimately the patient is not well-served because overly conservative tumor removal leads to the eventual need for repeat surgery, radiosurgery, or possibly both. This issue is the essence of the second part of our central question: ‘can keyhole approaches be considered reasonable alternatives to traditional approaches?’ Our overall meningioma resection rates for all locations except petroclival meningiomas are comparable to prior reports [8, 9, 44–46, 56], and our progression/recurrence rate of 14% is similar to prior reports (almost all in patients who underwent NTR or STR), although this rate will undoubtedly increase with longer follow-up [8, 57–62]. First-time resection rates for anterior cranial fossa and suprasellar meningiomas in the literature vary widely, ranging from 63% to 100%, with recurrence/progression rates ranging from 2% to 14% [19, 63, 64]. Considering petroclival meningiomas, we did not achieve GTR in any of 17 operations and had one new CN deficit (6%). In multiple series of petroclival meningiomas approached through traditional skull base approaches including the retromastoid approach, the GTR rate ranged from 21%-76% but permanent CN and other neurovascular complications ranged from 22%-54% [2–5, 65–71]. Collectively, these reports indicate that overly aggressive attempts at GTR, will likely be associated with a relatively high rate of permanent neurovascular morbidity and lower quality of life for many patients; thus we prefer a more conservative surgical approach for such invasive skull base meningiomas as other groups have also recommended [8, 10, 57–61, 72, 73].

A growing collective experience places functional preservation as a higher priority than GTR resection, as highlighted by recent reports [2, 8, 9, 37]. The beneficial impact of this approach for a given patient is a greater likelihood of no new postoperative neurological deficits and preserved or improved QOL while acknowledging that for many patients their meningioma becomes a chronic illness that warrants long-term monitoring with a higher likelihood of tumor progression and need for additional surgery, SRS/SRT or possible medical therapies in the years after non-GTR resection [8, 10, 57–61, 72, 73].

We highlight in Table 3 and S2 Table the factors most commonly responsible for not achieving a GTR which include invasiveness into the skull base and dense adherence to critical neurovascular structures. Heeding such anatomical realities accounts in large part for our low neurovascular complication rate but relatively high residual tumor rate for some locations. Overall, our results and those of other centers, seem to indicate a keyhole paradigm applied to meningiomas is a reasonable alternative to traditional skull base approaches, given our low neurovascular and systemic complication rates and short LOS, although further validation with longer follow-up at multiple centers is needed [19–22, 27, 28, 42].

Study limitations and bias

The major limitation of this study is its retrospective nature, heterogenous meningioma patient population, and our selection bias for using these 6 keyhole approaches without a comparison cohort of patients treated with traditional skull base approaches or other relatively new minimally invasive approaches such as the endoscopic transorbital route [74, 75]. We specifically did not compare clinical outcomes in the keyhole cohort to the 136 (41%) non-keyhole patient cohort. Also, the follow-up in our patients averaged 44 months which is relatively short. Longer follow-up is necessary to assess the efficacy of this approach more fully and determine how many patients who underwent NTR or STR ultimately need radiotherapy or additional surgery.