Because of intratumoral heterogeneity, diffusely infiltrating gliomas that lack significant contrast enhancement on magnetic resonance imaging are prone to tissue sampling error. Subsequent histologic undergrading may delay adjuvant treatments. 5-Aminolevulinic acid (5-ALA) leads to accumulation of fluorescent porphyrins in malignant glioma tissue, and is currently used for resection of malignant gliomas. The aim of this study was to clarify whether 5-ALA might serve as marker for visualization of anaplastic foci in diffusely infiltrating gliomas with nonsignificant contrast enhancement for precise intraoperative tissue sampling.
5-ALA was administered in 17 patients with diffusely infiltrating gliomas with nonsignificant contrast enhancement. During glioma resection, positive fluorescence was noted by a modified neurosurgical microscope. Intraoperative topographic correlation of focal 5-ALA fluorescence with maximum 11C-methionine positron emission tomography uptake (PETmax) was performed. Multiple tissue samples were taken from areas of positive and/or negative 5-ALA fluorescence. Histopathological diagnosis was established according to World Health Organization (WHO) 2007 criteria. Cell proliferation was assessed for multiregional samples by MIB-1 labeling index (LI).
Focal 5-ALA fluorescence was observed in 8 of 9 patients with WHO grade III diffusely infiltrating gliomas. All 8 of 8 WHO grade II diffusely infiltrating gliomas were 5-ALA negative. Focal 5-ALA fluorescence correlated topographically with PETmax in all patients. MIB-1 LI was significantly higher in 5-ALA–positive than in nonfluorescent areas within a given tumor.
Diffusely infiltrating gliomas represent the most common primary brain tumors.1 Treatment is based on histopathological typing and grading according to the current World Health Organization (WHO) 2007 diagnostic consensus criteria.2 In patients with the initial diagnosis of a high-grade glioma (WHO grade III and IV), immediate initiation of chemo- and radiotherapy after neurosurgical biopsy or resection is essential.3, 4 In contrast, extensive surgical resection with preservation of neurologic function is assumed as primary treatment for most low-grade gliomas (WHO grade II).5, 6 Therefore, precise neuropathological grading of diffusely infiltrating gliomas is essential for allocation of the patients to the adequate treatment.
The intratumoral heterogeneity of gliomas1, 7-9 poses a risk of histological undergrading with subsequent delay of adjuvant treatment. Therefore, intraoperative identification and sampling of the area of highest malignancy are crucial. Routinely, such anaplastic foci are identified by contrast enhancement on magnetic resonance imaging (MRI) and sampled intraoperatively using neuronavigational guidance.
A considerable number of WHO grade II and III diffusely infiltrating gliomas, however, are devoid of significant contrast enhancement.10-13 Thus, positron emission tomography (PET) imaging has emerged as a reliable navigational adjunct to intraoperatively detect areas of maximum metabolic activity that correspond to regions of highest malignancy.14-17 Intraoperative brain shift, however, may render navigational guidance that solely relies on preoperative imaging inaccurate, which is a major drawback for exact localization of surgical targets.18-21
5-Aminolevulinic acid (5-ALA) is a nonfluorescent prodrug that leads to intracellular accumulation of fluorescent porphyrins in malignant glioma cells.22 In low-grade gliomas, however, 5-ALA fluorescence has not been reported.23-25 On the basis of these observations that 5-ALA selectively accumulates in malignant glioma tissue, we designed this study to clarify whether 5-ALA might serve as an intraoperative marker for direct visualization of anaplastic foci in diffusely infiltrating gliomas with nonsignificant contrast enhancement.
MATERIALS AND METHODS
The present study comprises 17 consecutive patients surgically treated for diffusely infiltrating gliomas with nonsignificant contrast enhancement at the Department of Neurosurgery of the Medical University of Vienna between March 2008 and February 2009. No patient received adjuvant therapy before surgery. Patient and tumor characteristics are described in Table 1.
Table 1. Patient Characteristics
MRI indicates magnetic resonance imaging; WHO, World Health Organization.
Extent of resection was assessed on T2-weighted MRI performed within 48 hours postoperatively.
The study was approved by the ethics committee of the Medical University of Vienna.
MRI and PET are part of the routine preoperative diagnostic workup at our institution, and were performed within 2 weeks before surgery in all patients.
Magnetic resonance imaging
Our routine MRI protocol for brain tumors on a 3-T scanner (Tim Trio, Siemens, Erlangen, Germany) consists of axial fluid-attenuated inversion recovery sequences; diffusion-weighted images; axial, coronal T1-, and coronal T2-weighted sequences, and contrast-enhanced axial, coronal, and sagittal T1-weighted sequences. Functional MRI (fMRI) and diffusion-tensor imaging (DTI) were performed as appropriate.
All gliomas were classified by a neuroradiologist (D.P.) according to the pattern of contrast enhancement; only patients with none or unspecific (defined as patchy and faint) contrast enhancement were included in the study. Patients with unequivocal (defined as nodular or ring-like) contrast enhancement were excluded.
Patients received 800 to 850 MBq 11C-methionine (MET) after fasting for 4 hours. PET images were then obtained from a dedicated full-ring GE Advance PET scanner (General Electric Medical Systems, Milwaukee, Wis). For detailed description of radiotracer production, data acquisition, and reconstruction see Potzi et al.26
For the semiquantitative evaluation of tracer accumulation in tumor tissue the standardized uptake value (SUV = radioactive concentration [MBq/g]/injected dose [MBq] × body weight [g]) and tumor to normal brain ratio (intratumoral SUV/contralateral brain SUV) were used. The intratumoral area of the highest tumor to normal brain ratio, reflecting the most tracer uptake, was defined as PETmax and used as the surgical target for tumor tissue sampling.
All patients received oral solutions of 5-ALA (20 mg/kg bodyweight Gliolan; Medac, Wedel, Germany) 3 hours before anesthesia. For detection of 5-ALA fluorescence intraoperatively, we used a modified neurosurgical microscope (NC4, Carl Zeiss Surgical GmbH, Oberkochen, Germany), which has the ability to switch from conventional white light to violet-blue excitation light.27 To avoid potential skin phototoxicity, all patients were protected from light sources for 24 hours after 5-ALA administration.
Intraoperative guidance by a navigation system is routinely applied for glioma surgery. To localize eloquent areas and prevent neurological deficits by limiting tumor resection therein, we used multimodality image fusion navigation (fMRI, DTI tractography), electrophysiological monitoring, and awake surgery as appropriate. For targeting the tumor area of highest malignancy, we coregistered PETmax and anatomical magnetic resonance images on the navigation system (Stealth Station Cranial Mach 5; Medtronic, Louisville, Colo).
During glioma resection, the microscope was switched from conventional white light to violet-blue excitation light repeatedly. Visible intratumoral 5-ALA fluorescence was correlated with PETmax topographically on the navigation system.
To minimize the effects of brain shift in this study, we approached the PETmax as primary target at the beginning of surgery through a small corticotomy, very similar to frameless stereotactic biopsy. Through this minute approach, the 5-ALA fluorescence status was checked, samples were taken, and the fluorescing tissue was removed. Subsequently, the tumor resection was performed, and representative biopsies from areas outside of PETmax of different tumor consistency and appearance were collected. In the cases without increased MET uptake, multiple biopsies were sampled throughout the tumor. Specimens were labeled according to the 5-ALA fluorescence status as 5-ALA positive or 5-ALA negative.
Histopathological tumor diagnosis was established by the local neuropathology team according to the WHO 2007 diagnostic consensus criteria2 on a multiheaded microscope. In all patients, the total glioma specimen and all additional tissue samples (median 4 samples per patient; range, 1-8 samples) were embedded and routinely processed for conventional histopathology. Each specimen was screened for anaplastic foci, defined as areas with increased cellularity, cellular atypia, mitotic activity, and microvascular proliferation, to avoid undergrading. The neuropathologists (J.A.H. and A.W.) were blinded to the intraoperative 5-ALA fluorescence status.
Tumor cell proliferation was assessed immunohistochemically with the MIB-1 antibody (anti–Ki-67, 1:50; DAKO, Hamburg, Germany). MIB-1 binding was apparent as nuclear staining. In each specimen, a total of 500 tumor cell nuclei were evaluated in hot spots, that is, fields showing the highest density of Ki-67 immunopositive cells, by 1 author (A.W.), who was unaware of clinical patient characteristics. The fraction of immunolabeled tumor cell nuclei was expressed as a percentage (MIB-1 labeling index [LI]). If >1 sample of a tumor or its 5-ALA positive/negative areas were available, the maximum MIB-1 LI was selected for statistical analysis.
For statistical analyses, SPSS statistical software (version 16.0; SPSS Inc, Chicago, Ill) was used. PETmax and MIB-1 LI values showed a right-skewed distribution. Thus, for comparison of MIB-1 LI between subgroups, nonparametric tests (Mann-Whitney U test for independent data and Wilcoxon rank sum test for paired data) were applied. Values are given as median and range. A P value of <.05 was considered significant.
MRI revealed no contrast enhancement in 7 and unspecific contrast enhancement in 10 of 17 cases. Unspecific contrast enhancement on MRI correlated with neither positive 5-ALA fluorescence nor PETmax or high-grade pathology.
MET-PET detected an area of increased intratumoral metabolism compared with normal brain tissue in 13 of 17 patients (PETmax 2.2; range, 1.2-3.3). Whereas all WHO grade III gliomas showed a distinct intratumoral PETmax of ≥1.5, such a finding was present in only 2 of 8 WHO grade II gliomas. PETmax was significantly higher in WHO grade III gliomas compared with grade II (PETmax 2.4 vs 1.1; P = .018).
WHO indicates World Health Organization; 5-ALA, 5-aminolevulinic acid; LI, labeling index.
No. of patients
MIB-1 LI, median % (range)
5-ALA focal positive group
Intraoperatively, positive 5-ALA fluorescence was observed in 8 patients: Fluorescence was unequivocally visible in a focal area within the tumor volume and surrounded by 5-ALA negative areas. Changes in tumor tissue consistency and white-light appearance were frequently observed in the 5-ALA–positive area. Positive 5-ALA fluorescence was always encountered in the region of PETmax on navigation image data.
In 9 patients, no 5-ALA fluorescence was observed in any intratumoral area.
PETmax was significantly higher in the 5-ALA focal positive group as compared with the 5-ALA–negative group (PETmax 2.3 vs 1.2; P = .037).
WHO typing and grading
For histopathological glioma subtypes diagnosed according to WHO criteria, see Table 1. WHO grading revealed a WHO grade II glioma in 8 patients and a WHO grade III glioma in 9 cases. All tumors with intraoperative visible 5-ALA fluorescence were classified as WHO grade III gliomas. In contrast, all WHO grade II gliomas were intraoperatively found 5-ALA negative. However, 1 anaplastic oligodendroglioma did not show 5-ALA fluorescence. The positive predictive value of focal 5-ALA fluorescence for WHO grade III glioma was 100% (sensitivity 89%). The positive predictive value of a 5-ALA–negative tumor for WHO grade II glioma was 89% (sensitivity 100%).
Microvascular proliferation as defined by endothelial proliferation in multiple piled-up layers28 was not detected in any tumor sample.
The proliferation rate assessed by MIB-1 LI was significantly higher in WHO grade III gliomas than in WHO grade II (MIB-1: 18% vs 6%, P = .002). Correspondingly, the proliferation rate was significantly higher in the 5-ALA focal positive group than in the 5-ALA–negative group (MIB-1: 20% vs 6%, P = .001) (Fig. 1). Visible 5-ALA fluorescence correlated with a MIB-1 proliferation rate of ≥10% in all cases. The significantly higher proliferation rate in areas of 5-ALA fluorescence than in 5-ALA negative areas of the same tumor (MIB-1: 20% versus 10%, p = .02), underlines the ability of 5-ALA to detect anaplastic foci within diffusely infiltrating gliomas with nonsignificant contrast enhancement.2
Adverse Events of 5-ALA Resection
No side effects of 5-ALA and no procedure-related morbidity were observed in any patient.
Intratumoral heterogeneity is a common finding within diffusely infiltrating gliomas.1, 7-9 It is believed that acquired genetic variability within the original tumor clone generates multiple cell populations29 that give rise to anaplastic foci within a low-grade glioma. Such intratumoral heterogeneity in terms of grade was confirmed by Paulus and Peiffer. Of 1000 samples from 50 supratentorial gliomas, 62% concurrently revealed benign (WHO grade II) and malignant (WHO grade III and IV) areas within the same tumor.9 Coons and Johnson reported regional differences of tumor grade and proliferative activity in gliomas.7
However, identification of the most malignant tumor areas for surgical selection of a representative tissue specimen is indispensable for precise neuropathological grading of gliomas and subsequent allocation of patients to the adequate treatment. Although so far no generally accepted standard protocol for postoperative treatment of low-grade gliomas exists, it is common practice in many neuro-oncological centers to defer chemo- and radiotherapy after complete surgical resection in the majority of patients until progression occurs.4, 30, 31 However, the timing and impact of chemo- and radiotherapy in WHO grade II gliomas is the focus of ongoing trials.32 In patients with the initial diagnosis of a WHO grade III glioma, however, postoperative immediate initiation of chemo- and radiotherapy after neurosurgical biopsy or resection is essential.3, 4
In routine neurosurgical practice, unequivocal (nodular and/or ring-like) contrast-enhancement on MRI is detected to sample the most malignant tumor areas using neuronavigational guidance. In diffusely infiltrating WHO grade II and III gliomas, however, such contrast-uptake is frequently unreliable. Whereas an absence of significant contrast enhancement was reported in up to 55% of WHO grade III gliomas,10-13 up to 56% of low-grade gliomas have been found to be enhanced with contrast media on MRI.33-35 Of the WHO grade II gliomas with contrast enhancement, however, Pallud et al more frequently observed an unspecific “patchy and faint” than an unequivocal nodular pattern.34 In sum, gliomas with absent or unspecific contrast enhancement constitute a challenge to the neurosurgeon in selecting a representative intraoperative tissue sample. Therefore, only patients with absent or unspecific contrast enhancement were included in our study.
Metabolic imaging with PET using amino-acid tracers such as MET-PET or 18F-fluoroethyl-L-tyrosine-PET has emerged as a powerful and reliable technique for identification of malignant tumor areas14-17 during glioma surgery. Coregistered with anatomical magnetic resonance images for intraoperative neuronavigation, metabolic imaging can overcome the deficiencies of contrast-enhanced MRI. In the case of MET-PET, histological analysis confirmed that PETmax represents the most malignant tumor areas.14, 16
Loss of cerebrospinal fluid after craniotomy, gravity, and brain edema cause changes of the position of the intracranial structures during surgery termed brain shift.19 Brain shift may render the navigational guidance for tissue sampling inaccurate, as preoperative image data are used.18-21 The magnitude and direction of the shift of intracranial structures have been reported to reach up to 2.4 cm.20 Updating the navigation system during surgery is 1 of the major applications for intraoperative MRI,36 which is, however, costly and not widely available. Therefore, alternative techniques for intraoperative visualization of anaplastic foci have been sought.
5-ALA leads to intracellular accumulation of strongly fluorescing protoporphyrin IX in malignant glioma tissue.22 A selective photosensitization of 5-ALA–induced protoporphyrin IX fluorescence in high-grade tumor tissue in comparison to normal brain was suggested in an animal study.37 By using specifically modified microscopes, fluorescence-guided resection using 5-ALA has emerged as a promising technique for surgery of malignant gliomas.27, 38 A randomized controlled multicenter phase 3 trial on malignant gliomas demonstrated that the use of 5-ALA leads to a significantly higher frequency of complete resections of the contrast-enhancing tumor area as compared with the conventionally operated control group.39 Furthermore, patients with 5-ALA fluorescence-guided resection revealed significantly prolonged 6-month progression-free survival in comparison to the control group (41% vs 21%, respectively).
On the basis of the observations that 5-ALA is able to selectively detect malignant glioma tissue, we hypothesized that 5-ALA may serve as an intraoperative marker for direct visualization of anaplastic foci in diffusely infiltrating gliomas with nonsignificant contrast enhancement as a technique that is unaffected by brain shift.
Intraoperatively, we observed that certain diffusely infiltrating gliomas with nonsignificant contrast enhancement showed focal 5-ALA fluorescence surrounded by tumor areas with no fluorescence. Otherwise, we found gliomas that revealed no 5-ALA fluorescence throughout the entire tumor volume. By histopathological examination, all tumors of the 5-ALA focal positive group were classified as diffusely infiltrating WHO grade III gliomas. In contrast, all diffusely infiltrating WHO grade II gliomas were intraoperatively found to be 5-ALA negative.
This is the first systematic analysis of the value of 5-ALA fluorescence for detection of anaplastic foci in a series of diffusely infiltrating gliomas with nonsignificant contrast enhancement. In WHO grade II gliomas, 5-ALA fluorescence was not detected intraoperatively in 2 previous studies comprising 10 patients.23, 25 Ishihara et al examined glioma tissue ex vivo. The authors reported no 5-ALA fluorescence in 2 WHO grade II gliomas; however, in 2 WHO grade III gliomas, focal areas with positive macroscopic 5-ALA fluorescence were detected.24 Stummer et al reported a single patient with a secondary malignant glioma that revealed 5-ALA fluorescence of the malignant nodular contrast-enhancing focus but no fluorescence in the low-grade tumor part.40
These independent observations, including our study, provide strong evidence that in diffusely infiltrating WHO grade III gliomas, focal intratumoral areas can be observed with 5-ALA fluorescence intraoperatively. In contrast, no 5-ALA fluorescence seems to be visible in WHO grade II gliomas.
As intraoperative PET scan is currently available at only few centers, preoperative PETmax was used for intraoperative guidance to the tumor area of highest malignancy. To minimize potential brain shift, we used an open biopsy approach to the PETmax target at the beginning of tumor resection. If 5-ALA fluorescence was detected, it always topographically corresponded to the region of PETmax. To the best of our knowledge, this finding has never been reported previously. In our series, cell proliferation rate was significantly higher in areas of 5-ALA fluorescence than in the 5-ALA–negative parts within the same tumor. Correspondingly, Ishihara found a positive correlation between 5-ALA fluorescence intensity and MIB-1 LI in 2 WHO grade III gliomas ex vivo.24 In sum, these findings indicate that areas with positive 5-ALA fluorescence correspond to foci with high metabolic and proliferative activity in diffusely infiltrating gliomas with nonsignificant contrast enhancement (Fig. 3).
The observation that microvascular proliferation, which is a hallmark of glioblastomas and may be encountered in highly anaplastic oligodendrogliomas/oligoastrocytomas,2 was not detected in any tumor sample of our series may be explained by the inclusion of gliomas with nonsignificant contrast enhancement only.
In our series, a single patient who was histopathologically diagnosed with anaplastic oligodendroglioma did not reveal 5-ALA fluorescence intraoperatively. Deterioration of the 5-ALA fluorescence has been attributed to a photobleaching effect after prolonged tissue exposure to white and violet-blue light.40 Indeed, duration of surgery was prolonged because of deep tumor location and photobleaching may be 1 possible factor responsible for the 5-ALA negativity of this case.
In addition, this patient received high doses of dexamethasone (24 mg per day) preoperatively in contrast to the other cases of our series. During the phase 3 study, fluorescence-guided resections of malignant gliomas have been performed in a standard combination of 5-ALA with only 12 mg dexamethasone per day.39 Corticosteroids are known to exert a tightening effect on the blood-brain barrier,41 and may thereby restrict 5-ALA uptake and reduce the fluorescence effect. We therefore did not routinely pretreat the patients with dexamethasone before fluorescence-guided resection unless clinically indicated. The influence of the dexamethasone dosage on the intraoperative 5-ALA fluorescence uptake and effect, however, remains to be clarified.
Our present data indicate that 5-ALA is a promising marker for intraoperative visualization of anaplastic foci in diffusely infiltrating gliomas with nonsignificant contrast enhancement. Positive 5-ALA fluorescence corresponds to areas with high metabolic and proliferative activity within these tumors. In contrast to conventional neuronavigation using preoperative imaging data, 5-ALA visualization of anaplastic foci has the advantage of being unaffected by intraoperative brain shift.
The reliable intraoperative identification of representative tissue samples via 5-ALA fluorescence may increase the precision of neuropathological grading of gliomas and subsequently optimize allocation of patients to adjuvant treatment. The use of 5-ALA may become a novel standard for intraoperative tissue sampling in diffusely infiltrating gliomas with nonsignificant contrast enhancement. To this end, subsequent confirmatory multicentric studies will be needed.
We thank Dr. Harald Heinzl, Care Unit for Medical Statistics and Informatics, for statistical advice, and Irene Leisser and Gerda Ricken for technical assistance with preparation of tissue specimens.