Intraoperative Fluorescence Guidance in Meningiomas

6 Intraoperative Fluorescence Guidance in Meningiomas

Pablo A. Valdes Quevedo, Alexandra J. Golby, and David W. Roberts

Abstract
Fluorescence guidance has become a useful adjunct for neurosurgeons. The main fluorescent markers in clinical use for meningiomas include 5-aminolevulinic acid-induced protoporphyrin IX, indocyanine green, and fluorescein sodium. These markers have been used with microscope- and spectroscopy-based technologies for tumor surgery and delineation of intracranial vasculature. This chapter reviews the main clinical applications of fluorescence guidance in meningiomas.

Keywords: fluorescence-guided surgery, 5-aminolevulinic acid, protoporphyrin IX, indocyanine green, meningiomas, optical spectroscopy

6.1 Introduction

Fluorescence guidance has become a burgeoning field of research and increasingly adopted surgical adjunct for clinical applications in neurosurgery.1,2 Multiple technologies including microscope systems and single probe spectroscopic or microscopic-based tools have been developed for clinical implementation. These technologies have been coupled to various fluorescent agents including 5-aminolevulinic-acid-induced protoporphyrin IX (ALA-PpIX), fluorescein, and indocyanine green (IGC). Cranial and spinal applications of fluorescence guidance in neurosurgery include use in brain tumors such as gliomas, metastases, and meningiomas.^1,3,4,5,6 Here, we present an overview of the neurosurgical literature using fluorescence guidance in meningiomas.

6.2 5-Aminolevulinic-Acid-Induced Protoporphyrin IX

The most widely used fluorescent agent for neurosurgical resection of intracranial tumors is ALA-PpIX.^1,2 Briefly, ALA is administered orally prior to surgery, which leads to uptake of ALA with subsequent production and selective accumulation of PpIX in tumor cells. PpIX is the active fluorescent compound with two major excitation peaks at 405 and 633 nm and two major emission peaks at 635 and 710 nm.^3,5 The largest experience with ALA-PpIX was a phase III randomized controlled trial in glioblastoma, which demonstrated almost doubling of complete resection rates and improved progression-free survival in patients, which has led to its widespread use in Europe.⁴

The clinical experience and literature on the use of 5-ALA in meningioma resection are more limited, and as such the clinical role of ALA-PpIX is not yet as firmly grounded as it is for high-grade gliomas. For example, a recent review noted the following distribution of publication numbers per tumor type: pediatric brain tumors 5, spinal tumors 7, meningioma 10, low-grade glioma 14, metastases 15, and high-grade glioma 51.⁶ The largest reported series on meningiomas is that of Millesi et al.⁷ The authors analyzed 204 meningiomas undergoing ALA-PpIX fluorescence-guided surgery and highlighted various key observations of relevance to meningioma, which have been further elaborated and confirmed in the literature. This study reported Simpson grade resection I to IV in 45, 22, 17, and 16% of tumors, respectively; positive fluorescence was seen in 91% (185/204) of cases with no correlation between location, histology, and fluorescence appearance; a mostly homogeneous pattern of fluorescence (75%, 113/150 analyzed cases)⁸; 89 cases with dural tails on MRI demonstrating no visible fluorescence, of which 5/16 analyzed histopathologically contained tumor cells; 7 cases of satellite lesions detected with PpIX fluorescence, which would have otherwise gone undetected; tumor-infiltrated bone flaps with positive fluorescence in 100% of cases (13/13); and fluorescence in adjacent cortex in 25% (20/80) of cases with a significant difference in cases with a disrupted arachnoid (41 vs. 11%, p = 0.002) but no correlation with WHO grade. (Of note, the authors did not provide histopathologic confirmation regarding the degree or status of tumor infiltration in the evaluated arachnoid or cortex.)

One of the first reports using ALA-PpIX was that of Coluccia et al⁹ on a cohort of 33 patients with intracranial meningiomas (32, WHO grades I–II) and (1, WHO grade III), which noted that 94% (31/34) of patients demonstrated visible levels of fluorescence with no correlation to histology, MIB index, mitotic index, or preoperative brain edema. These results are similar to the first report on ALA-PpIX in meningiomas by Kajimoto et al¹⁰ who reported on 24 patients, with 83% (20/24) demonstrating visible fluorescence with no correlation between histology and fluorescence. Cornelius et al¹¹ studied 31 meningiomas across WHO grades I to III (19, 8, and 4, respectively), in which 94% were positive for fluorescence; in their study, they did report a correlation between subjective intensity and WHO grade (ρ = 0.557; p = 0.001). Coluccia et al, similar to others,^10,12 did note the presence of positive fluorescent tumor infiltrating skull in one case, highlighting the ability of ALA-PpIX to delineate infiltrated margins beyond the tumor bulk and into skull. Kajimoto et al¹⁰ also note histologic confirmation not only of skull invasion, but also invasion of hypertrophied dura with a sensitivity of 100% (5/5) and specificity of 83% (5/6).

Della Puppa et al¹³ reported the use of ALA-PpIX in 12 patients with bone-invading meningiomas, noting a sensitivity of 89% and specificity of 100% for detecting bone invasion with positive predictive value (PPV) = 100%, negative predictive value (NPV) = 83%, overall accuracy = 93%, and no residual bone invasion on postoperative MRI. This study further highlighted relationships between bone invasion, hyperostosis, and fluorescence. All fluorescent samples had bone invasion (100%, 57/57) and 30% (7/23) of nonfluorescent samples had bone invasion and hyperostosis, and none of the nonhyperostotic samples had fluorescence, that is, all false-negatives were found in hyperostotic bone (7/23). This study demonstrates a strong association between bone invasion and fluorescence as well as a weaker association between hyperostosis and fluorescence.

Whitson et al¹⁴ used confocal microscopy and histology to investigate tumor infiltration and the dural margins. They note a high specificity in which uninvolved dura demonstrated no fluorescence. They also note a decreased sensitivity in which microscopic amounts of tumor cells contained no significant fluorescence; for example, less than 1 mm of tumor tail showed no fluorescence. Bekelis et al 15 applied a quantitative probe in a case of a skull base meningioma showing significantly improved sensitivity in detecting significant levels of PpIX despite no visible fluorescence. Cornelius et al¹⁶ used PpIX fluorescence as an adjunct to MRI- and PET-guided resection to help identify a skull base meningioma with bony and dural infiltration.

Wilbers et al¹⁷ presented a case report of an atypical WHO grade II meningioma where distinct fluorescence was noted not just in gross tumor, but also in adjacent dura, arachnoid, and cortex with histologic confirmation, suggesting that arachnoid may be a source of tumor recurrence despite macroscopic gross total resection of tumor and adjacent dura; that is, despite achieving Simpson grade I resection. In a similar report of an atypical meningioma, Scheichel et al¹⁸ observed histopathologically confirmed visible fluorescence in the periosteal layer, inner temporalis muscle fascia, temporalis muscle, and bone in addition to the solid tumor, highlighting the limitation of standard white light resection to identify these infiltrating tumor remnants.

Most studies with ALA-PpIX are in cranial meningiomas, while studies on spinal applications have been much more limited. Eicker et al¹⁹ looked at 26 patients with intradural spinal tumors, which included 8 meningiomas (WHO grade I) using ALA-PpIX. Seven of the eight meningiomas demonstrated bright pink fluorescence with a PPV of 100%. The authors noted that fluorescence was of particular utility in recurrent cases with adherent, scarred tissue to help dissect tumor from spinal cord. Muroi et al²⁰ report on the utility of ALA-PpIX in one case of a spinal meningioma where on macroscopic, white light inspection at the end of resection, the authors judged they had achieved a Simpson grade II resection, but on final inspection with fluorescence identified a small tumor remnant that was subsequently resected without dural removal.

The aforementioned studies have used a modified surgical microscope with ultraviolet excitation light and collection emission filters greater than 450 nm to visualize the emitted red–pink fluorescence through the surgical oculars and onto a color camera.^{5,7,9,12,13,17} As such, assessments of fluorescence in meningioma surgery have been qualitative in nature, noting a “strong” or “weak” fluorescence without measuring the actual quantitative levels of PpIX ( Fig. 6.1). That is, such qualitative assessments do not take into account the nonlinear, attenuating effects of tissue optical properties; that is, absorption and scattering on the excitation light and emitted fluorescence.²¹ These attenuating effects can lead to inaccurate assessments of the tissue fluorescence, such as intraoperative qualitative assessment of “no visible fluorescence” in areas where there might actually be significant levels of fluorescent PpIX biomarker.^2,22,23

Recent advances at the interface between neurosurgery and optical engineering have developed new optical probes that enable intraoperative and in vivo quantitative assessments of PpIX.^{2,8,15,22,23,24,25,26} Valdes et al⁸ reported on 15 meningiomas using ALA-PpIX, noting that meningiomas demonstrated high, homogenous fluorescence, and no correlation between visible fluorescence characteristics and MR contrast enhancement. The authors performed quantitative assessments on 10 of 15 patients for a total of 49 interrogated sites, noting the diagnostic significance of quantitative ALA-PpIX: diagnostically high PpIX levels were measured in 69% of histologically confirmed tumor tissues that did not demonstrate visible fluorescence. Tumor had a statistically significant higher mean concentration of PpIX (c_PpIX) of 1.694 ± 0.440 μg/mL compared to normal dura (i.e., control) with an average c_PpIX of 0.006 ± 0.003 μg/mL (p = 0.002). Furthermore, 39% (13/33) of confirmed tumor specimens did not display visible fluorescence (that is, false-negative), but 69% (9/13) of these nonvisibly fluorescent tumor specimens contained PpIX levels greater than 0.010 μg/mL. Receiver operating characteristic (ROC) analysis demonstrated a statistically significant improvement in diagnostic performance for quantitative fluorescence compared to qualitative fluorescence (p = 0.007). Using a quantitative approach, the authors were able to balance the need for improved sensitivity (i.e., lower cutoff threshold value, c_PpIX = 0.001, accuracy of 90%, PPV of 91%, specificity of 81%, NPV of 87%, and sensitivity of 94%) at the cost of decreased specificity and vice versa, improved specificity (i.e., higher cutoff threshold value, c_PpIX = 0.0114; accuracy of 84%, PPV of 100%, specificity of 100%, NPV of 67%, and sensitivity of 76%) at the cost of decreased sensitivity.