Molecular pathological expression in malignant gliomas resected by fluorescein sodium-guiding under the YELLOW 560 nm surgical microscope filter

Retrospective analysis of the 18 cases of pathologically confirmed MG resected by FLS guiding under the YELLOW 560 nm filter from the neurosurgery department of Shandong University Qilu Hospital (Qingdao) between January 2014 and December 2016 (Table 1).

Patients 1, age ranged between 18 and 75 years; 2, who were newly diagnosed, untreated or have relapsed MG, with certain pathological results after postoperative confirmation; 3, according to the surgery after 24–72 h of intracranial MRI examination; 4, according to the enhancement area of postoperative MRI, the surgeon and the imaging department evaluates the total resection of the tumor.

(1) Patients under 18 years old and more than 75 years old were excluded; (2) tumors originating from the brain stem; (3) excluding patients who are suffering low-level gliomas or non-tumors whose MRI showed enhancement areas; (4) renal insufficiency of patients; (5) patients with hepatic insufficiency; (6) other parts of the body with active malignant tumor patients; (7) preoperative tumor MRI enhancement and postoperative pathology proved to be metastatic tumor patients; (8) according to the operation, a single specimen retention of the case should be excluded.

Twenty percent fluorescent sodium was obtained from Guangzhou Baiyun Mountain Ming Xing Pharmaceutical Co., Ltd. (National Drug Code: H44023400). Before the use of 20% FLS, diluted to 3%, skin test was performed with 5 ml deep vein injection to observe the patient’s vital signs and rashes and other abnormalities, and then diluted to 1%. The dosage was in accordance with the patient’s weight, i.e., 2–3 mg/kg. Drug delivery time: injected the drug just before skin cut after anesthesia induction was begun. Drug delivery: single dose intravenous injection.

During the operation, under the real-time YELLOW 560 mode, regardless of tumor location, according to the yellow staining degree of the tissue specimen is labeled as “no yellow dye,” “low yellow dye,” and “bright yellow dye” levels. All tissue specimens were removed and immediately applied with 10% formalin fixation and embedded for pathological analysis. In the tumor boundary area, pathological specimens obtained randomly from each patient were marked as “no,”, “low,” and “bright” yellow according to the yellow fluorescence staining levels.

GFAP and Ki-67 were purchased from Fuzhou Mai Xin Company.

Sheep anti-rat/rabbit IgG (KIT-5030) were purchased from Fuzhou Mai Xin Company.

Leica RM2235 type electric paraffin slicer (German Leica Company).

Leica BONDTM Automatic IHC dyeing system (German Leica Company).

All specimens were fixed with 10% formalin, conventional paraffin embedding, slicing machine adjusted to thickness of 4 μm, and sliced continuously. After slicing, HE was stained conventionally and GFAP and Ki67 were stained by immunohistochemistry.

The 4 μm paraffin sections were fixed and washed, and extreme care must be taken to avoid peeling off the sections. The slides were specially repaired by citric acid working fluid (for Ki67) by incubating at 120 °C temperature for 2 min. And GFAP was no need to be repaired.

GFAP staining of tissue cell cytoplasm of yellow or brown-yellow granules were considered positive, and no color or faint yellowish granules were considered negative.

Compute the Ki-67 labeling index (Ki-67 LI). Ki-67 LI is defined as the percentage of the total number of cells that are Ki-67 positive. For the region of necrosis and vascular endothelial cells, cells can be differentiated when the nucleus that is less than 2 mm is not counted. In view of less than 1 cell per low magnification, LI is considered to be less than 0.1% , is not visible under the microscope, and is calculated as 0 [4].

The follow-up of outpatients, inpatients was done through cell phone, SMS, interview, WeChat, Tencent QQ, and email. Imaging evaluations were performed at 1 month, 3 months after surgery, and every 6 months. All patients underwent MRI scanning and contrast-enhancement (CE) of the brain in 24–72 h after operation, 1 month after operation, 3 months, and every 6 months. The associated complications were observed from elevated blood pressure, seizures, tracheal spasms, or allergic reactions following the injection of FLS.

Continuous variables are described in average, median, and standard deviation. SPSS19.0 statistical software was used and χ² test with multiple composition ratios were used. p < 0.05 was considered statistically different.

The patients included in this study were 18 cases, including 7 male cases and 11 female cases. The average age was 48.2 years (26–71 years old). Main symptoms and postoperative pathological classification had been listed in Table 1.

One percent FLS was intravenously injected before skin incision after induction of general anesthesia. The first 17 cases were administered with 3 mg/kg weight of FLS, followed by 2 mg/kg weight, and both obtained the same effect (see Fig. 1 and Fig. 2). After opening the dura, i.e., intravenous FLS injection for 20–40 min, it gathers tumor tissue without spilling into normal brain tissue. After craniotomy, all tumors were stained by FLS in YELLOW 560 nm mode. The tumor tissue showed a bright fluorescent color (see Fig. 1b (a)), especially in the MRI-enhanced area. But in the necrotic region of the tumor or in the area where the MRI was not enhanced, it showed a “low” or “none” fluorescent staining (see Fig. 1d (b)). The liquid region of cystic tumors can still be shown as bright fluorescence in color and obviously lighter than that in the cerebrospinal fluid. Even at very low concentrations of FLS (2 mg/kg), the YELLOW 560 nm model of the Pentero 900 microscope can easily differentiate between fluorescent and non-fluorescent tissues.

The average tumor volume was 53.88 cm³ (13.32–126.759 cm³). The total resection was performed in 16 cases (see Fig. 3 and Fig. 2). The 69 pathological specimens were randomly obtained from 18 cases (32 bright-, 18 low-, 19 non-fluorescence). The postoperative KPS score was slightly higher than before, though no statistically significant difference (average preoperative 82 vs. postoperative 83, p = 0.566). Five patients reported a short KPS score reduction (cases 4, 6, 8, 13, and 17) after surgery. In addition, eighth case had permanent mild hemiplegia, with short-term recovery to the preoperative state. No venous thrombosis and pulmonary embolism occurred after operation. The color of the skin, mucus membranes, and urine of the patient after being injected with FLS was yellow and disappeared after about 24 h after the operation. No related serious adverse events were found.

Two patients (case 1 and 18) had epileptic seizures in the early postoperative period (1 month), considering the preoperative state of the patients.

Sixteen cases were completely followed up (cases 4 and 11 were lost), with a follow-up rate of 88.9% (16/18). Complete radiotherapy and chemotherapy (Stupp regimen) were completed in 10 cases (10/18).3 patients died of tumor progression (cases 5, 6, and 8) and 1 patient died of severe pneumonia.

Nowadays, there were three main fluorescent agents used for guided resection of glioblastoma including indocyanine Green (ICG), FLS, and 5-ALA. The removal of MG by the ICG-guided resection involves a short time, which could be helpfully used to obtain specimens but little useful for continuous navigation monitoring. ICG navigation combined with other navigations (such as 5-ALA) was mainly used for examination of residual tumors involving a short time after resection of MG [11]. 5-ALA had not yet been approved by the Committee on Food and Drug Administration which is an endogenous luminescent agent, with optical instability, partly low sensitivity, and specificity. Because of its “light bleaching,” the tumor edge was not clearly displayed [12]. 5-ALA also had the disadvantages of photo-induced toxicity, expensive equipment, and complicated application process. Hence, 5-ALA had much difficulty in clinical popularization. FLS in ophthalmic fundus angiography had been safely and effectively used for many years, and was used to identify tumors in neurosurgery [13, 14]. Shinoda et al. [15] reported resection of 32 glioblastomas, and the total resection rate reached 84.4%. Chen B et al. [16] reported 10 cases of MG with complete resection of 80%, while only 33.3% of MG in control group (12 cases). KOC [17] resected 47 cases of glioblastomas demonstrating a total resection of 83% while control group (33 cases) showed 55%. Though there was no significant difference of influencing the lifetime (43.9 weeks and 41.8 weeks). 20 mg/kg FLS obtained clear imaging, and 10 cases of glioblastoma were expected to be resected. Kuroiwa et al. [18] after using special filter, 10 cases of glioblastoma were resected by intravenous injection of 8 mg/kg FLS. F. Acerbi et al. [8] using Pentero microscope intravenously injected 5 mg/kg FLS, then 12 cases of GBM obtained a total resection of 75%, and the remaining patients with tumor resection volume of 90.5% (82.6–99.9%). In 2013, K.M. Schebesch et al. [7] reported that under the YELLOW 560 nm filter, with intravenous injection of 200mg FLS (3-4mg/kg), 35 cases of intracranial tumors, including 22 cases of MG, were resected. In our preliminary study, the total resection rate of MG under the guidance of FLS was 92.1%. The 6 month-PFS (92.3%) and median survival period (11 months) after the resection of glioblastoma were similar to those of FLUGLIO [8, 19]. Of the 18 cases, the total resection was achieved in 16 cases with multiple regional retentions of specimens according to the postoperative-enhanced MRI. Under the white light mode, the tumor tissue was blurring in the boundary between the tumor and the surrounding tissues. While under YELLOW 560 nm mode, the normal tissue and the tumoral tissue could be clearly identified (Figs. 1 and 2). Postoperative KPS score decreased before operation, and a slight increase after the operation, though the difference was not statistically significant (average preoperative 82 vs. postoperative 83, p = 0.566). No postoperative complication was associated with FLS, and thus it was considered safe and effective for the removal of MG by FLS navigation.

GFAP, an acidic protein belonging to the family of intermediate silk proteins, is the cytoskeleton protein of specific astrocytes [20], which was first separated from the white matter plaque of multiple sclerosis patients [21]. The function of nutritional support neurons plays an important role in shaping and maintaining normal morphology of glial cells [20, 22]. GFAP is normally expressed in astrocytes, ventricular membrane cells, and so on, and is considered as the iconic protein of astrocytes [23]. GFAP-positive expression could be found in astrocytoma, ventricular duct tumor, mixed glioma of the astrocytes line, giant cell astrocytoma, pleomorphic yellow astrocytoma, astrocytoma, glioma sarcoma, and so on [24]. As a reliable protein marker, it was widely used in the clinical immunohistochemical identification of glioma and non-gliomas [25]. In this group, 61 specimens of GFAP were positive. There was no correlation between the expression rate of GFAP and fluorescence levels (Fig. 4b, e and Fig. 5b, e). The result might be due to that (1) the normal glial cells also express GFAP, and our study observed that the same patient had different fluorescence level regions and obtained the different specimens when the color of fluorescence was darker (Fig. 1b (a) vs. Fig. 1d (b), and Fig. 2d white arrow vs. Fig. 2e white arrow) while the color of staining was lighter (Fig. 4b vs. Fig. 4e, and Fig. 5b vs. Fig. 5e). According to the literature, with the increase of the malignant degree of astrocytoma, the production of GFAP decreased [26]. Immunohistochemistry showed that the GFAP-positive staining rate and staining color of MG were lighter than those of normal brain tissue. We had confirmed this discipline through only using semi-quantitative analysis. (2) Semi-quantitative analysis has a certain degree of bias. It may be more instructive to study the quantitative expression of GFAP in the future. (3) Due to multiple-sources of MG, the positive rate of molecular pathology expression also changed a lot.

The higher the malignant degree of glioma, the worst will be the prognosis, and the higher will be the WHO grading. We supposed that glioma with different malignant degrees might have influence on the developing intensities of FLS. But in this group, there was no significant correlation between the WHO grading and fluorescence levels (χ2 = 3.531, p = 0.171). The reason may be due to (1) after injection, FLS transmitted through the destruction of the blood-brain barrier and agglomerated in the extracellular matrix. Diaz and others reported that [27] FLS showed obvious development in gadolinium-enhanced preoperative MRI area. It indicated that the development of FLS was similar to the enhancement mechanism of gadolinium injection, which was related to the damage of blood-brain barrier. However, some studies, including our early observation combined with neuronavigation, found that there was still FLS [28] outside the gadolinium-sprayed amine-enhanced range. So, there are still many unsolved reasons in the distribution mechanism of FLS. (2) This small sample size was not enough to show the relationship between WHO grading and fluorescence levels. (3) WHO grading gradually developed to molecular pathology; many molecular pathology mechanisms are unclear, and different types of MG characteristics according to the molecular pathological diagnosis may be different.

Ki-67, a nuclear antigen of proliferating cells, can be detected in the nuclear plasma and mitotic phase of the cell transferring to the chromosome surface. In MG, Ki-67 acts as a cell proliferation marker that was more specific than PCNA [22]. The percentage of Ki-67-positive expression can reflect the degree of malignancy [29]. Ki-67 was expressed in all levels of astrocytoma and showed high expression levels in the higher grades of tumors [30]. Kiss et al. [31] observed that low- and very low-density Ki-67 LI had longer survival rates. Torp et al. [32] also reported that Ki-67 LI was associated with malignancy of astrocytoma, with high levels of Ki-67 LI than with low levels of LI, exhibiting a worse prognosis. Bouvier et al. believed that greater than 5% of Ki-67 LI was considered as a risk factor for tumor progression and poor prognosis [32]. In this group, all the patients showed Ki-67 expression and was graded by Ki-67 proliferation index LI. Ki-67 LI was correlated with fluorescence levels, and the difference was statistically significant (χ2 = 14.678, p = 0.005). This suggested that the removal of MG through FLS navigation can reduce the MG cells and reduce Ki-67 LI, improving the prognosis of patients (Fig. 4 and Fig. 5). Bouvier and other studies have also found that Ki-67 Li may not be an independent risk factor, but low levels of Ki-67 and total resection were achieved, and continuous postoperative chemotherapy patients had a better prognosis.

Although the expression of Ki-67 was still observed (Fig. 4f and Fig. 5f) in the near-positive boundary biopsy of FLS staining (Fig. 1d (b) and Fig. 2e white arrow), it had been significantly reduced after surgical resection (χ2 = 14.678, p = 0.005). Furthermore, it has been found in our earlier study that there was still fluorescein dye outside the area of MRI enhancement according to preoperative MRI registered by neuronavigation [19]. Thus, the fluorescein-stained regions could be included and larger than the contrast-enhanced regions. More aggressive or super-total resection of MGs could be deemed to the reasonable treatment strategy for MGs [5]. For the development of FLS and to help in glioma resection, more clinical data should be studied and discussed.