- Email: [email protected]
Early-postoperative magnetic resonance imaging in glial tumors: prediction of tumor regrowth and recurrence
European Journal of Radiology 45 (2003) 99 /107 www.elsevier.com/locate/ejrad
Early-postoperative magnetic resonance imaging in glial tumors: prediction of tumor regrowth and recurrence Gazanfer Ekinci a, I˙hsan N. Akpınar a,*, Feyyaz Baltacıog˘lu a, Canan Erzen a, Tu¨rker Kılıc¸ b, I˙lhan Elmacı b, Necmettin Pamir b b
a Marmara University Medical Faculty, Department of Radiology, Istanbul, Turkey Marmara University Medical Faculty, Department of Neurosurgery, Istanbul, Turkey
Received 10 December 2001; received in revised form 18 January 2002; accepted 22 January 2002
Abstract Objective: This study investigated the value of early-postoperative magnetic resonance (EPMR) imaging in the detection of residual glial tumor and investigated the role of EPMR for the prediction of tumor regrowth and recurrence. Methods and materials: We retrospectively analyzed pre- and post-operative magnetic resonance imaging results from 50 adult patients who underwent surgical treatment for supratentorial glial tumor. There were glioblastoma multiforme in 25 patients, astrocytoma (grades II and III) in 11 patients, oligodendroglioma (grades II and III) in 9 patients, and oligoastrocytoma (grades II and III) in 5 patients. EPMR imaging was performed within 24 h after surgery. EPMR findings were compared with the neurosurgeon’s intraoperative estimation of gross tumor removal. Patterns of contrast enhancement at the resection site, in residual and developing tumor tissue and blood at the resection site were evaluated on EPMR and in follow-up studies. ‘Residual tumor’ was defined as contrast enhancing mass at the operative site on EPMR. ‘Regrowth’ was defined as contrast enhancing mass detected on follow-up in the same location as the primary tumor. ‘Recurrence’ was defined as appearance of a mass lesion in the brain parenchyma distant from the resection bed during follow-up. Results: Nineteen patients showed no evidence of residual tumor, regrowth, or recurrence on EPMR or any of the later follow-up radiological examinations. EPMR identified 20 cases of residual tumor. Follow-up showed tumor regrowth in 10 patients, and tumor recurrence in 1 case. EPMR showed contrast enhancement of the resection bed in 45 of the 50 patients. Four of the 20 residual tumors showed a thick linear enhancement pattern, and the other 16 cases exhibited thick linear-nodular enhancement. No thin linear enhancement was observed in the residual tumor group. Nine of the 10-regrowth tumors showed a thick linear-nodular enhancement pattern, and one exhibited thin linear enhancement in EPMR. For predicting regrowth tumor EPMR sensitivity was 91%, specificity was 100%, positive predictive value 1; negative predictive value was 0.9375. Conclusion: EPMR, depending on the surgical site enhancement pattern, is a valuable means of demonstrating residual tumors, and can be used to predict possible regrowth after surgery. # 2002 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Early-postoperative magnetic resonance; Glial tumor; Supratentorial glial tumor; Magnetic resonance imaging; Contrast enhancement
1. Introduction Gliomas, the most common tumors of the central nervous system (CNS), are usually treated surgically [1,2]. The poor prognosis for patients with malignant glioma (WHO Grade III astrocytoma) and glioblastoma
¨ niversitesi, Hastanesi * Corresponding author. Address: Marmara U Radyoloji A.B.D. Tophaneliog˘lu cad. 13/15, 81190 Altunizade, Istanbul, Turkey. Tel.: /90-216-327-6955; fax: /90-216-327-6956 E-mail address: [email protected] (I˙.N. Akpınar).
multiforme (GBM) is due to the extremely high rate of local recurrence as well as tumor histopathology, age and history of patient, performance status, and degree of tumor removal. Most recurrences are detected within the first year after surgery; thus, accurate postoperative assessment is important [1 /3]. Prognosis, decisions about further therapy, and adequate response to therapy in patients with glial tumors all depend on correct determination of the quantity of residual tumor after surgical resection. Initially, this decision was based on the information from the operating surgeon concerning the presence and extent of residual tumor. In recent
0720-048X/02/$ - see front matter # 2002 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 7 2 0 - 0 4 8 X ( 0 2 ) 0 0 0 2 7 - X
100
G. Ekinci et al. / European Journal of Radiology 45 (2003) 99 /107
years, preoperative and postoperative imaging has become the method of choice. Since preoperative imaging is not a widely used procedure yet, postoperative magnetic resonance imaging (MRI) and computed tomography (CT) are used to evaluate and follow glioma patients. For the most part, MRI has replaced CT for these purposes. It is recommended that MRI should be done in the first 3 days after surgery, and preferably, within the first 24 h, before nonneoplastic contrast enhancement due to surgical manipulation becomes radiologically apparent [4]. Although it is possible to see non-neoplastic enhancement even earlier than 3 days, postoperative MRI has been carried out sometime in this initial time frame, and usually earlier than 48 h [2 /7]. In this study, our aims were to investigate the value of early-postoperative magnetic resonance (EPMR) imaging in the detection of residual tumor within 24 h after surgery; to compare the EPMR findings of residual tumor with the neurosurgeon’s intraoperative estimation of gross tumor removal; and to analyze the contrast enhancement findings at the surgical site to estimate the predictive value of EPMR for detecting regrowth and recurrent tumors.
2. Methods and material We retrospectively analyzed the pre- and post-operative MRI studies of adult patients who underwent surgery for supratentorial glial tumors at our institution between January 1995 and December 2000. Patients were included in the study if both pre- and postoperative MRI studies were available, and if the operated glial tumor was a primary neoplasm. Individuals who had undergone previous surgical and radiation treatment were excluded, as well as the patients who had grade I tumors. Also all patients with non-enhancing tumors on preoperative MRI were excluded. We examined a total of 285 patients’ pre- and postoperative MRI studies, and 50 cases met the above criteria. This group included 22 females and 28 males (age range 19/69 years, mean age 42.8 years). The surgically resected gliomas were identified histopathologically based on World Health Organization (WHO) criteria for the classification of CNS tumors. The histopathological results were GBM in 25 patients, astrocytoma (grades II and III) in 11 patients, oligodendroglioma (grades II and III) in 9 patients, and oligoastrocytoma (grades II and III) in 5 patients. Focal external beam radiation has been performed to the resection site including 3/4 cm safety margin for all patients 1 month after surgery. Two patients with grade III astrocytoma had also received chemotherapy. In each case, preoperative MRI had been performed within the month prior to surgery. Standard spin-echo
(SE) T1-weighted (TR/TE 480/600/16 /22 ms), fast SE T2-weighted images (TR/TE 2600/4800/90/120 ms, ET /12), and contrast-enhanced SE T1-weighted images (TR/TE 480/600/16 /22 ms) were available. Intravenous Gd-DTPA was given 0.1 /0.2 mmol/kg. EPMR was done within 24 h after surgery (range, 12/ 24 h). SE T1-weighted images were obtained in a similar plane to that of the preoperative non-contrast and contrast-enhanced images. Other technical factors included a 256/192 matrix, a slice thickness of 4/6 mm, and a 0.5 mm interslice gap. All patients underwent MR imaging using a 1.5 Tesla System (General Electric Medical Systems, Inc., Milwaukee, WI). The EPMR studies were acquired within the first day in order to avoid the confounding effects of benign enhancement of the surgical site, and to evaluate the site for residual glial tumors which may have a very short doubling time. Serial follow-up MRI examinations were also present on postoperative 30 days, 3 months, 6 months, and 1 year after surgery. Two neuroradiologists evaluated all of the MRI studies. On postoperative images, the operation site was evaluated for residual tumor, regrowth and recurrent tumors. Resection site contrast enhancement was also noted. If resection site contrast enhancement was present, the pattern of contrast enhancement was characterized as ‘thin linear,’ ‘thick linear,’ or ‘thick linear-nodular.’ In case of different opinions among two radiologists, the case was consulted to a third neuroradiologist for final decision. We defined the different patterns of enhancement at the surgical site as follows: . Enhancement that resembled normal dural enhancement was designated as thin linear (Fig. 1a and b). . Enhancement that was thicker than typical dural enhancement or E/5 mm was classified as thick linear (Fig. 1c and d). If we detected nodularity, we named it as thick linear-nodular enhancement (Fig. 1e). . Enhancement that was thicker than 1 cm in any imaging plane was considered to indicate residual tumor (Fig. 1e). These three patterns of early-postoperative surgical enhancement were followed on the consecutive radiological examinations. Regrowth and recurrence were defined as follows: a mass lesion in the resection bed at the same location as the primary tumors was named as ‘regrowth tumor’ (Fig. 2a and b), and a mass lesion in the brain parenchyma distant from the resection bed was named as ‘recurrent tumor’ (Fig. 2c). After classifying the patients into groups according to the surgical site enhancement patterns, the diagnosis of regrowth and recurrent tumors were made if there was at least a 20% increase in tumor size at any follow-up
G. Ekinci et al. / European Journal of Radiology 45 (2003) 99 /107
101
Fig. 1. (a) and (b) There is thin linear enhancement at the resection site in EPM on contrast-enhanced coronal T1-weighted images (long arrow). The short arrow shows the dural enhancement at the operation site. On contrast-enhanced axial (c) and sagital (d) T1-weighted images display the thick linear enhancement at the operation site. (e) There is gross residual tumor at the resection site (long arrow) together with thick linear-nodular enhancement (short arrow) on contrast-enhanced sagital T1-weighted image.
MRI examination when compared to the EPMR or the MRI where the tumor first appeared. Histopathological
confirmation was also made with stereotactic biopsy and/or reoperation in these patients.
102
G. Ekinci et al. / European Journal of Radiology 45 (2003) 99 /107
We also noted the hemorrhage in the surgical area and extraaxial locations to clarify its effects on image prediction. For each group of patients, we analyzed the EPMR enhancement patterns and tested the correlation between these and the various primary tumor types.
3. Results 3.1. Residual tumors EPMR revealed that 20 of the 50 patients had residual tumors. Residual tumor was found in two of the 11 patients who had grade II tumors, and in 18 of the 39 patients who had high-grade tumors (Table 1). The residual tumors were confirmed by at least a 20% increase in size on follow-up imaging. Eleven of 20 residual tumors were reoperated and nine of 20 residual tumors were stereotacticaly biopsied. Based on intraoperative assessment, the surgeon reported complete tumor excision in 9 of the 20 patients whose EPMR confirmed residual neoplasm, and acknowledged the presence of residual tumor in the other 11 cases (Fig. 3a and b). In the remaining 30 patients without residual tumors, the surgeon reported total tumor excision. 3.2. Regrowth and recurrence The regrowth and recurrent tumors detected on follow-up examinations are given in Table 1. Regarding time of detection for tumor regrowth; four cases (3 GBM, one oligodendroglioma Grade III) were identified in the first month after surgery, four cases (3 GBM, one astrocytoma Grade III) were diagnosed at 3 months, and two cases (1 GBM, one oligoastrocytoma Grade II) were detected at 6 months on follow-up. One patient with GBM, who had no residual tumor on EPMR and no regrowth on follow-up developed recurrence. In all the patients who exhibited regrowth or recurrence, we noted an increase in tumor size of at least 20% in the further follow-up, as compared to the EPMR findings or the MRI in which the tumor first appeared. These cases were also biopsied stereotacticaly for histopathological confirmation. Fig. 2. (a) There are thick linear-nodular enhancements at the resection bed (arrows) on contrast-enhanced coronal T1-weighted EPMR image. (b) In same patient, there is regrowth tumor at the resection bed on contrast-enhanced coronal T1-weighted image at the third month (arrow). (c) In another patient, on third month contrastenhanced sagital T1-weighted image, there is regrowth (short arrow) and 2 recurrent (long arrows) tumors in a patient having GBM with no residual tumor and with thin linear enhancement at the resection bed on EPMR.
3.3. Resection bed enhancement pattern The resection site enhancement patterns detected on EPMR are shown in Table 2. Table 3 lists the resection bed enhancement patterns on EPMR for the residual, regrowth, and recurrent tumors.
G. Ekinci et al. / European Journal of Radiology 45 (2003) 99 /107
103
Table 1 Postoperative MRI findings Tumor
Grade (No.)
MRI findings EPMRb residual tumor
Follow-up MRI Regrowth
GBMa
25
Oligodend.
II (4) III (5)
16
Astrocyto.
II (5) III (6)
2 1
Oligoastro.
II (2) III (3)
1
Sum a b
50
20
7
Recurrence 1
No tumor 1
1
4 4
1 1
2 4 2 2
10
1
19
GBM: glioblastoma multiforme. EPMR: early-postoperative MRI.
Four of the five patients whose EPMR imaging showed no enhancement in the surgical site developed a thin linear enhancement pattern in the surgical site on the first month of follow-up. One patient never showed any resection bed enhancement. All of the regrowth tumors developed from the nodular region of the thick linear-nodular enhancement areas. The only case of recurrent tumor that was not associated with any residual or regrowth tumor exhibited a thin linear enhancement pattern in the resection bed. We calculated the sensitivity, specificity, positive and negative predictive values of EPMR for predicting regrowth tumor based on enhancement patterns (Table 4). EPMR sensitivity was 91%, specificity was 100%, positive predictive value (PPV) was 1, and negative predictive value (NPV) was 0.9375. Eighteen of the 19 patients who had no residual tumor, regrowth, or recurrence exhibited a thin linear contrast enhancement pattern in the resection bed on follow-up examinations. One patient in this group showed no resection bed enhancement on follow-up. The resection site enhancements disappeared at 1 month in 11 patients, at 3 months in 5 patients, at 6 months in 1 patient, and at 1 year in 1 patient. 3.4. Blood at the surgical site and extraaxial locations
Fig. 3. (a) On contrast-enhanced coronal T1-weighted EPMR image, there is gross residual tumor at the resection bed (arrow), although the surgeon reported the gross total tumor resection. (b) On contrastenhanced sagital T1-weighted image at the sixth month, there is remarkable increase in residual tumor size (arrow).
EPMR showed evidence of blood in the surgical area (resection cavity and adjacent parenchyma) in 12 of the 50 cases studied. In all patients, this resolved in the first month after surgery. Extraaxial blood (subarachnoidal, subdural, epidural) was seen in only 3 patients, and other types of extraaxial fluid collection were noted in 9 of the 50 cases. Follow-up of extraaxial fluid and blood collections showed that these also resolved within 1 month. On EPMR, the intraaxial fluid collection was
G. Ekinci et al. / European Journal of Radiology 45 (2003) 99 /107
104
Table 2 Contrast enhancement at the surgical site on EPMR Tumor
Grade (No.)
Enhancement pattern None
a
GBM
IV (25)
Oligodend.
II (4) III (5)
Thin linear 3
4 1
II (5) III (6)
2 4
Oligoastro.
II (2) III (3)
2 2
a
50
5
Thick linear-nodular
2
20
3
Astrocyto.
Sum
Thick linear
1 1 1
2 1 1
16
4
25
GBM: glioblastoma multiforme.
Table 3 EPMR contrast enhancement in the resection site after surgery for all groups of patients Enhancement of the resection bed
Residual
Thin linear Thick linear Thick linear-nodular
4 16
a
Regrowth
Recurrence
No residual, regrowth, or recurrencea
1
1
14
9
Five patients showed no enhancement at the resection bed on EPMR.
localized to the resection cavity. On EPMR, the fluid signal had a relatively higher and heterogeneous signal compared to the cerebrospinal fluid on T1-weighted images. However, the signal characteristics of fluid in the resection bed cavity on T1-weighted images were not as hyperintense as blood, and we encountered no difficulties in evaluating the enhancement.
4. Discussion Postoperative radiological imaging is used to determine the amount of tumor mass that has been removed, to detect the surgical complications as early as possible, and to follow the response of the remaining tumor to radiotherapy and chemotherapy [3,5,6]. Researchers have used CT and MRI to evaluate the postoperative tumoral contrast enhancement and surgically induced changes. The findings have varied as the techniques have improved over the years. The most reliable radiological finding associated with glial neoplasms is the contrast enhancement of tumoral components. The parts of these growths that enhance in the various imaging modalities are usually solid, richly vascular, and densely cellular [2,4 /8]. This knowledge is applied when formulating treatment strategies for glial tumors, and in determining the response to therapy [2,4 /6,8 /13]. The edges of the enhanced areas represent the macroscopic borders of the glioma, but the tumor cells do not necessarily stay within these margins. The
cells can infiltrate the brain parenchyma and the edema surrounding the mass. This cellular infiltration is reflected by a high-intensity signal on T2-weighted MRI. Studies have shown that this cellular infiltration can reach 3 cm or more beyond the enhanced edges of the tumor mass [2 /7,14 /19]. In such cases, MR spectroscopy (MRS) has been reported to be useful, but it is not routinely applicable due to cost and time [20]. In cases of glial tumors, the surgeon’s main goal is to completely excise the mass through its macroscopic borders. The surgery allows a histological diagnosis to be made, and improves the patient’s neurological status. Moreover, any residual tumor is more amenable to radiotherapy after the debulking process. Although the surgeon can roughly estimate the amount of residual tumor preoperatively, some remaining tumor may not be visible to the naked eye [3,4,12,16,21 /23]. PeroperaTable 4 Sensitivity and specificity of EPMR for regrowth tumors according to enhancement patterns excluding residual tumors Regrowth tumor
Thick linear-nodular Thin linear Total
Present
Absent
9 1 10
0 15 15
Total
9 16 25a
Sensitivity: 0.91; Specificity: 1; PPV: 1; NPV: 0.9375. a Only patients with no residual tumor on EPMR were included.
G. Ekinci et al. / European Journal of Radiology 45 (2003) 99 /107
tive MR has gained more importance recently, but it still is not a widely used method [9]. Postoperative imaging enhanced by intravenous contrast injection is the method of choice for detecting residual tumor and evaluating the tumor extension after surgery. It is sometimes difficult to differentiate the tumoral enhancement from enhancement due to the normal postoperative changes. Benign contrast enhancement of the surgical site is assumed to result from the local impairment of the blood/brain barrier, neovascularity, luxury perfusion, or extravasation of contrast material [3,6,7,12,24 /28]. MRI has replaced CT in the postoperative evaluation of residual tumors and benign enhancement at the resection bed. Various authors have emphasized the superiority of MRI for detecting residual tumors. Compared to CT, MRI offers higher resolution of contrast enhancement, better contrast resolution, and fewer uncertain or false-negative findings [2 /8]. Numerous MRI studies have investigated postoperative changes with the goal of identifying the optimal time after surgery to assess benign and malignant contrast enhancement findings [4,7,8,25,29 /33]. Focusing on an earlier time point in our study, we sought to investigate the predictive value of EPMR after glial tumor surgery. Our results showed that EPMR revealed enhancement of the resection bed in 45 of 50 cases (90%), a finding that has also been mentioned in previous studies [7,8,27,32,34,35]. On this basis, we dispute the notion that there is a ‘diagnostic window’’ in the first 72 h after surgery in which all enhancements seen in the resection bed on EPMR is caused by residual tumors. Our findings are also not consistent with the suggestion that all enhancements seen in the resection bed on EPMR are caused by residual tumors [4,26 / 28,31,34,35]. If the presence of enhancement in the resection bed on EPMR is not specific for the presence of a residual tumor, is it possible for the pattern of enhancement to provide information about residual tumor tissue? First of all, if EPMR reveals an enhancing mass larger than 1 cm in any imaging plane, the pattern of enhancement in the resection bed is not relevant because postoperative treatment is guided by the finding of the residual tumor. The emphasis should be on evaluating enhancement patterns in the resection bed in cases where no mass or overt residual tumor is detected. Several studies indicate that enhancement peaks in the first week (5 /7 days), and is detected until the end of the first postoperative month in benign situations. It is also reported that during the first 3 days, benign enhancement is linear, but after 5 /7 days there is significant and intense thick linear and nodular enhancement, so if MRI is done during this period it is very difficult to differentiate between postoperative changes and residual tumors [4,7,8,25/28,31,33].
105
Opinions differ regarding the evaluation of enhancement patterns seen on MRI in the first 3 days after surgery. The most recent and broadly based study on this subject was conducted by Oser et al., who examined 46 MRI studies of 43 cases. They found no evidence that postoperative enhancement pattern indicated the presence of residual tumor [32]. The patients in this study had not only glial tumors, but also a more heterogeneous group of CNS tumors. Also to be noted is that 19 of the individuals had undergone previous surgery for brain tumor. The authors did not report the enhancement pattern findings for the 15 of these 19 cases in which preoperative assessment had shown no residual tumor, but they did note enhancement in the resection bed on these images. In our study, 25 of the 30 cases in which EPMR showed no residual tumors, there was enhancement in the resection bed within the first 24 h. One of these 25 cases developed recurrence during follow-up, and 10 cases developed tumor regrowth. Nineteen of 30 cases showed no regrowth or tumor recurrence. In 9 of the 10 patients who developed regrowth during follow-up, a thick linear-nodular pattern was present, however a thin linear pattern was present in only one case. Other studies have identified thin linear enhancement as a reflection of benign changes or changes secondary to surgical trauma, whereas the enhancement pattern in cases of residual or recurrent tumors has been reported as thick linear-nodular [4,7,8,25,26,29 /31,33,36,37]. In cases where EPMR revealed no residual tumor, regrowth tumors developed from the region where the thick linear-nodular enhancement pattern was seen [4,8,29,33,36,37]. Similarly, regrowth tumors in our study developed from thick linear-nodular enhancements, except one patient. For regrowth tumors, PPV of EPMR was 100% and NPV was 0.9375. Considering all cases in our study, the rate of tumor regrowth and tumor recurrence was 6.6% (2/30) in cases with thin linear enhancement pattern in the resection bed. When the group that had no residual tumor on EPMR was considered and GBM patients were excluded, we detected no regrowth or recurrence in cases with thin linear enhancement at the surgical site. We believe that thin linear enhancement observed in nonGBM cases can be considered benign. In our GBM cases that showed no residual tumor on EPMR, but exhibited thin linear contrast enhancement at the operative site, the rate of tumor regrowth was 11.1% (1/9). In other words, the finding of a thin linear pattern in the resection bed of GBM cases on EPMR is benign 88.9% of the time. Forsyth and co-workers did not recommend the routine use of EPMR for postoperative evaluation of malignant gliomas [7]. They investigated only 17 malignant cases, and 3 were recurrent cases that had been previously operated. In 10 of these 17 patients, EPMR
106
G. Ekinci et al. / European Journal of Radiology 45 (2003) 99 /107
confirmed the presence of residual tumor. Our study group included 39 high-grade gliomas (grade III and GBM), and EPMR results were evaluated against 1-year postoperative follow-up MRI findings. Contrary to the opinion of Forsyth et al., our results revealed that EPMR should be included in the routine protocol after glioma surgery. Intraoperative MRI, which increases the percentage of the complete resection of tumor by showing the residual tumor intraoperatively, MRS, and PET [9,20,38] examinations are more frequently used recently to detect residual and recurrent tumors with success. However, all of these techniques are more expensive examinations. Routine contrast enhanced MRI is the basic study for preoperative diagnosis and for the comparison of postoperative findings with the preoperative status. The radiological findings on routine preoperative MRI examinations should be a guide in the evaluation of preoperative MRI, MRS, and PET studies. Researchers have shown that benign enhancement in the resection bed after the removal of gliomas or any other surgery of CNS due to nontumoral lesions disappears in a mean time of 1/3 months, but this process can take up to 6 /8 months [3,4,7,8,24 / 28,30,32,33,39]. In our study, cases that showed no residual tumor, regrowth, or recurrence showed enhancement at the surgery site for 6 months after the operation, but no enhancement was observed at the 1year follow-up MRI. All of our patients received radiotherapy 1 month after surgery. The acute and early delayed complications occurring during radiotherapy and within 2/3 months after radiotherapy, respectively, are not permanent in most cases, therefore they do not interfere with the interpretation of images for tumor regrowth [38,40]. Severe changes like radiation necrosis that could resemble a tumor tend to appear 1 year or more after radiotherapy [40]. All regrowth tumors in our study were detected within 6 months postoperatively. This time interval was shorter than the time needed for a radiation necrosis to become visible. Therefore we did not consider radiation induced radiological changes in our study. Besides, stereotactic biopsy was performed in all regrowth cases. Reports indicate that the most obscuring element behind the resection bed enhancement on EPMR is blood in the surgical area [4,7,8,25,32,33,41,42]. Blood is extravasated when vessels are damaged during surgery, and even the most meticulous hemostasis may not prevent some blood leakage. The ability to detect blood on MRI depends on the transformation of hemoglobin to deoxyhemoglobin and methemoglobin over time [4,7,32,42]. Investigations have shown that methemoglobin was present during early postoperative period, and the reported reason for increased methemoglobin
formation was surgical procedures. Methemoglobin, which is hyperintense on T1W images, can prevent the visualization of enhancement due to residual tumor on very early postoperative MRI [4,7,8,27,32,34]. Various amounts of blood have been reported to be present in the surgical area (resection cavity and neighboring parenchyma) in cases of CNS surgery (tumor and nontumor cases with rates ranging from 26 to 100% [4,7,8,25,30,32]. In our study, blood was detected in the surgical area in 24% of the cases. Extraaxial blood (subarachnoidal, subdural, epidural) was detected in 6% of cases, whereas in other reports this rate has ranged from 6 to 26% [4,7,8,25,30,32]. The reasons for these discrepancies may be differences in surgical techniques, imaging methods, and/or the time of postoperative imaging. None of the cases in our study group had large amounts of blood in the resection bed that prevented the evaluation of enhancement patterns. The most important point here is the need for simultaneous inspection of pre- and post-contrast images. In addition, postcontrast images should be acquired in three planes instead of one designated plane. This permits more accurate enhancement evaluation in the surgical bed. We cannot comment on late-postoperative hematomas because there were no such lesions in our study group. MRI is three times better than the surgeon’s naked eye for detecting residual tumors [4,7]. Of our 20 cases of EPMR-confirmed residual tumors, the surgeon reported complete excision in 9 (45%) cases, and residual tumor after resection in 11 (55%) cases.
5. Conclusion Our study results revealed that, radiological detection of residual tumor is far more sensitive than intraoperative estimation by the surgeon. The suggestion of a ’diagnostic window’’ on EPMR in the first 72 h after surgery is not supported, and a thick linear-nodular enhancement pattern at the resection site should be considered pathologic, and should be followed closely for tumor regrowth. A thin linear enhancement pattern can usually be accepted as benign contrast enhancement of the surgical site. Our study shows that EPMR yields reliable and accurate information for the detection of residual tumors, and enables us to predict possible tumor regrowth. In our opinion, EPMR should be performed in all adult patients who undergo surgical treatment for supratentorial glial tumors.
References [1] Bonnin JM, Garcia JH. Histology and growth characteristics of brain neoplasms. In: Taveras JM, Ferrucci JT, editors. Radiology,
G. Ekinci et al. / European Journal of Radiology 45 (2003) 99 /107
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15] [16]
[17]
[18]
[19]
[20]
[21]
vol. 3 (Revised ed., Chapter 52). J.B. Lippincott Company, 1990:1 /11. Madison MT, Hall WA, Latchaw RE, Loes DJ. Radiologic diagnosis, staging, and follow-up of adult central nervous system primary malignant glioma. Radiol Clin N Am 1994;32(1):183 /96. Kelly WM, Brant-Zawadzki B. Magnetic resonance imaging and computed tomography of supratentorial tumors. In: Taveras JM, Ferrucci JT, editors. Radiology, vol. 3 (Revised ed., Chapter 53). J.B. Lippincott Company, 1990:1 /22. Henegar MM, Moran CJ, Silbergeld DL. Early postoperative magnetic resonance imaging following nonneoplastic cortical resection. J Neurosurg 1996;84:174 /9. Holt RM, Maravilla KR. Supratentorial gliomas: imaging. In: Wilkins RH, Rengachary SS, editors. Neurosurgery, vol. 1B (Chapter 79). McGraw-Hill, 1996:753 /88. Albert FK, Forsting M, Sartor K, et al. Early postoperative magnetic resonance imaging after resection of malignant glioma: objective evaluation of residual tumor and its influence on regrowth and prognosis. Neurosurgery 1994;34(1):45 /61. Forsyth PAJ, Petrov E, Mahallati H, et al. Prospective study of postoperative magnetic resonance imaging in patients with malignant gliomas. J Clin Oncol 1997;15(5):2076 /81. Haba D, Papon AP, Tanguy JY, et al. Use of half-dose gadolinium-enhanced MRI and magnetization transfer saturation in brain tumors. Eur Radiol 2001;11:117 /22. Schwartz RB, Wong TZ, et al. Intraoperative MR imaging guidance for intracranial neurosurgery. Radiology 1999;211:477 /8. Cairncross JG, Macdonald DR, Pexman JHW, Ives FJ. Steroid induced CT changes in patients with recurrent malignant glioma. Neurology 1988;38:724 /6. Cohadon F. Indications for surgery in the management of gliomas, in advances and technical standards in neurosurgery, vol. 17. Wien: Springer, 1990:189 /233. Levin VA, Hoffman WF, Heilbron DC, Norman D. Prognostic significance of the pretreatment CT scan on time to the progression for patients with malignant gliomas. J Neurosurg 1980;52:642 /7. Salcman M, Levine H, Rao K. Value of sequential computed tomography in the multimodality treatment of glioblastoma multiforme. Neurosurgery 1981;8:15 /8. Husstedt HW, Sickert M, Ko¨stler H, et al. Diagnostic value of the fast-FLAIR sequence in MR imaging of intracranial tumors. Eur Radiol 2000;10:745 /52. Masters LT, Zimmerman RD. Imaging of supratentorial brain tumors in adults. Neuroimag Clin N Am 1993;3(4):649 /69. Ciric I, Vick NA, Mikhael MA, Cozzens J, Eller T, Walsh A. Aggressive surgery for malignant supratentorial gliomas. Clin Neurosurg 1990;36:375 /83. Burger PC, Dubois BJ, Schold SC, Smith KR, Jr., et al. Computerised tomographic and pathologic studies of untreated, quiescent, and recurrent glioblastoma multiforme. J Neorosurg 1983;58:159 /69. Burger PC, Heinz ER, Shibata T, Kleihues P. Topographic anatomy and CT correlations in the untreated glioblastoma multiforme. J Neurosurg 1988;68:698 /704. Kelly PJ, Daumos-Duport C, Kispert DB, et al. Imaging based stereotaxic serial biopsies in untreated intracranial glial neoplasms. J Neorosurg 1987;66:865 /74. Burtscher IM, Skagerberg Gastro CT, Geizer B, et al. Proton MR spectroscopy and preoperative diagnostic accuracy. AJNR 2000;21:84 /93. Salcman M. Surgical decision making for malignant brain tumors. Clin Neurosurg 1987;35:285 /313.
107
[22] Shapiro WR, Green SB, Burger PC, et al. Randomised trial of three chemotherapy regimens and two radiotherapy regimens in postoperative treatment of malignant glioma. J Neurosurg 1989;71:1 /9. [23] Kyritsis AP, Soyo H, Levin VA. Molecular pathogenesis and management of gliomas. Neuroimag Clin N Am 1993;3(4):735 / 44. [24] Takahashi M. Supratentorial neoplasms */postoperative changes. In: Taveras JM, Ferrucci JT, editors. Radiology, vol. 3 (Revised ed., Chapter 57). J.B. Lippincott Company, 1990:1 /5. [25] Sato N, Bronen RA, Sze G, et al. Postoperative changes in the brain: MR imaging findings without neoplasms. Radiology 1997;204:839 /46. [26] Cairncross JG, Pexman JHW, Rathbone MP, DelMaestro RF. Postoperative contrast enhancement in patients with brain tumor. Ann Neurol 1985;17:570 /2. [27] Laohaprasit V, Silbergeld DL, Ojemann GA, et al. Postoperative CT contrast enhancement following lobectomy for epilepsy. J Neurosurg 1990;73:392 /5. [28] Jeffries BF, Kishore PRS, Singh KS, et al. Contrast enhancement in the postoperative brain. Radiology 1981;139:409 /13. [29] Meyers SP, Wildenhain S, Chess MA, Tarr RW. Postoperative evaluation for intracranial recurrence of medullablastoma: MR findings with gadopentetate dimeglumine. AJNR 1994;15:1425 / 34. [30] Elster AD, DiPersio DA. Cranial postoperative site: assessment with contrast-enhanced MR imaging. Radiology 1990;174:93 /8. [31] Forsting M, Albert FK, Kunze S, et al. Extirpation of glioblastomas: MR and CT follow-up of residual tumor and regrowth. AJNR 1993;14:77 /87. [32] Oser AB, Moran CJ, Kaufman BA, Park TS. Intracranial tumor in children: MR findings within 24 hours of craniotomy. Radiology 1997;205(3):807 /12. [33] Rollins NK, Nisen P, Shapiro KN. The use of early postoperative MR in detecting residual juvenile cerebellar pilocytic astrocytoma. AJNR 1998;19:151 /6. [34] Sze G. New applications of MR contrast agents in neuroradiology. Neuroradiology 1990;32:421 /38. [35] Ciric I, Ammirati M, Vick N, Mikhael M. Supratentorial gliomas: surgical considerations and immediate postoperative results. Neurosurgery 1987;21:21 /6. [36] Nicoletti GF, Barone F, Passanisi M, Mancuso P, Albanese V. Linear contrast enhancement at the operative site on early postoperative CT after removal of brain tumors. J Neurosurg Sci 1994;38:131 /5. [37] Bird CR, Drayer BP, Medina M, et al. Gd-DTPA-enhanced MR imaging in pediatric patients after brain tumor resection. Radiology 1988;169:123 /6. [38] Rabin BM, Meyer JR, Berlin JW, et al. Radiation-induced changes in the central nervous system and head and neck. Radiographics 1996;16:1055 /72. [39] Bronen RA, Sze G. Magnetic resonance imaging contrast agents: theory and application to the central nervous system. J Neurosurg 1990;73:820 /39. [40] Boyko OB. Neuroimaging of radiation injury to the central nervous system. Neuroimaging Clin North Am 1993;3:803 /16. [41] Meyding-Lamade´ U, Forsting M, Albert F, et al. Accelerated methaemoglobin formation: potential pitfall in early postoperative MRI. Neuroradiology 1993;35:178 /80. [42] DiChiro G, Brooks RA, Girton ME, et al. Sequential MR imaging studies of intracerebral hematomas in monkeys. AJNR 1985;7:193 /9.
Early-postoperative magnetic resonance imaging in glial tumors: prediction of tumor regrowth and recurrence Gazanfer Ekinci a, I˙hsan N. Akpınar a,*, Feyyaz Baltacıog˘lu a, Canan Erzen a, Tu¨rker Kılıc¸ b, I˙lhan Elmacı b, Necmettin Pamir b b
a Marmara University Medical Faculty, Department of Radiology, Istanbul, Turkey Marmara University Medical Faculty, Department of Neurosurgery, Istanbul, Turkey
Received 10 December 2001; received in revised form 18 January 2002; accepted 22 January 2002
Abstract Objective: This study investigated the value of early-postoperative magnetic resonance (EPMR) imaging in the detection of residual glial tumor and investigated the role of EPMR for the prediction of tumor regrowth and recurrence. Methods and materials: We retrospectively analyzed pre- and post-operative magnetic resonance imaging results from 50 adult patients who underwent surgical treatment for supratentorial glial tumor. There were glioblastoma multiforme in 25 patients, astrocytoma (grades II and III) in 11 patients, oligodendroglioma (grades II and III) in 9 patients, and oligoastrocytoma (grades II and III) in 5 patients. EPMR imaging was performed within 24 h after surgery. EPMR findings were compared with the neurosurgeon’s intraoperative estimation of gross tumor removal. Patterns of contrast enhancement at the resection site, in residual and developing tumor tissue and blood at the resection site were evaluated on EPMR and in follow-up studies. ‘Residual tumor’ was defined as contrast enhancing mass at the operative site on EPMR. ‘Regrowth’ was defined as contrast enhancing mass detected on follow-up in the same location as the primary tumor. ‘Recurrence’ was defined as appearance of a mass lesion in the brain parenchyma distant from the resection bed during follow-up. Results: Nineteen patients showed no evidence of residual tumor, regrowth, or recurrence on EPMR or any of the later follow-up radiological examinations. EPMR identified 20 cases of residual tumor. Follow-up showed tumor regrowth in 10 patients, and tumor recurrence in 1 case. EPMR showed contrast enhancement of the resection bed in 45 of the 50 patients. Four of the 20 residual tumors showed a thick linear enhancement pattern, and the other 16 cases exhibited thick linear-nodular enhancement. No thin linear enhancement was observed in the residual tumor group. Nine of the 10-regrowth tumors showed a thick linear-nodular enhancement pattern, and one exhibited thin linear enhancement in EPMR. For predicting regrowth tumor EPMR sensitivity was 91%, specificity was 100%, positive predictive value 1; negative predictive value was 0.9375. Conclusion: EPMR, depending on the surgical site enhancement pattern, is a valuable means of demonstrating residual tumors, and can be used to predict possible regrowth after surgery. # 2002 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Early-postoperative magnetic resonance; Glial tumor; Supratentorial glial tumor; Magnetic resonance imaging; Contrast enhancement
1. Introduction Gliomas, the most common tumors of the central nervous system (CNS), are usually treated surgically [1,2]. The poor prognosis for patients with malignant glioma (WHO Grade III astrocytoma) and glioblastoma
¨ niversitesi, Hastanesi * Corresponding author. Address: Marmara U Radyoloji A.B.D. Tophaneliog˘lu cad. 13/15, 81190 Altunizade, Istanbul, Turkey. Tel.: /90-216-327-6955; fax: /90-216-327-6956 E-mail address: [email protected] (I˙.N. Akpınar).
multiforme (GBM) is due to the extremely high rate of local recurrence as well as tumor histopathology, age and history of patient, performance status, and degree of tumor removal. Most recurrences are detected within the first year after surgery; thus, accurate postoperative assessment is important [1 /3]. Prognosis, decisions about further therapy, and adequate response to therapy in patients with glial tumors all depend on correct determination of the quantity of residual tumor after surgical resection. Initially, this decision was based on the information from the operating surgeon concerning the presence and extent of residual tumor. In recent
0720-048X/02/$ - see front matter # 2002 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 7 2 0 - 0 4 8 X ( 0 2 ) 0 0 0 2 7 - X
100
G. Ekinci et al. / European Journal of Radiology 45 (2003) 99 /107
years, preoperative and postoperative imaging has become the method of choice. Since preoperative imaging is not a widely used procedure yet, postoperative magnetic resonance imaging (MRI) and computed tomography (CT) are used to evaluate and follow glioma patients. For the most part, MRI has replaced CT for these purposes. It is recommended that MRI should be done in the first 3 days after surgery, and preferably, within the first 24 h, before nonneoplastic contrast enhancement due to surgical manipulation becomes radiologically apparent [4]. Although it is possible to see non-neoplastic enhancement even earlier than 3 days, postoperative MRI has been carried out sometime in this initial time frame, and usually earlier than 48 h [2 /7]. In this study, our aims were to investigate the value of early-postoperative magnetic resonance (EPMR) imaging in the detection of residual tumor within 24 h after surgery; to compare the EPMR findings of residual tumor with the neurosurgeon’s intraoperative estimation of gross tumor removal; and to analyze the contrast enhancement findings at the surgical site to estimate the predictive value of EPMR for detecting regrowth and recurrent tumors.
2. Methods and material We retrospectively analyzed the pre- and post-operative MRI studies of adult patients who underwent surgery for supratentorial glial tumors at our institution between January 1995 and December 2000. Patients were included in the study if both pre- and postoperative MRI studies were available, and if the operated glial tumor was a primary neoplasm. Individuals who had undergone previous surgical and radiation treatment were excluded, as well as the patients who had grade I tumors. Also all patients with non-enhancing tumors on preoperative MRI were excluded. We examined a total of 285 patients’ pre- and postoperative MRI studies, and 50 cases met the above criteria. This group included 22 females and 28 males (age range 19/69 years, mean age 42.8 years). The surgically resected gliomas were identified histopathologically based on World Health Organization (WHO) criteria for the classification of CNS tumors. The histopathological results were GBM in 25 patients, astrocytoma (grades II and III) in 11 patients, oligodendroglioma (grades II and III) in 9 patients, and oligoastrocytoma (grades II and III) in 5 patients. Focal external beam radiation has been performed to the resection site including 3/4 cm safety margin for all patients 1 month after surgery. Two patients with grade III astrocytoma had also received chemotherapy. In each case, preoperative MRI had been performed within the month prior to surgery. Standard spin-echo
(SE) T1-weighted (TR/TE 480/600/16 /22 ms), fast SE T2-weighted images (TR/TE 2600/4800/90/120 ms, ET /12), and contrast-enhanced SE T1-weighted images (TR/TE 480/600/16 /22 ms) were available. Intravenous Gd-DTPA was given 0.1 /0.2 mmol/kg. EPMR was done within 24 h after surgery (range, 12/ 24 h). SE T1-weighted images were obtained in a similar plane to that of the preoperative non-contrast and contrast-enhanced images. Other technical factors included a 256/192 matrix, a slice thickness of 4/6 mm, and a 0.5 mm interslice gap. All patients underwent MR imaging using a 1.5 Tesla System (General Electric Medical Systems, Inc., Milwaukee, WI). The EPMR studies were acquired within the first day in order to avoid the confounding effects of benign enhancement of the surgical site, and to evaluate the site for residual glial tumors which may have a very short doubling time. Serial follow-up MRI examinations were also present on postoperative 30 days, 3 months, 6 months, and 1 year after surgery. Two neuroradiologists evaluated all of the MRI studies. On postoperative images, the operation site was evaluated for residual tumor, regrowth and recurrent tumors. Resection site contrast enhancement was also noted. If resection site contrast enhancement was present, the pattern of contrast enhancement was characterized as ‘thin linear,’ ‘thick linear,’ or ‘thick linear-nodular.’ In case of different opinions among two radiologists, the case was consulted to a third neuroradiologist for final decision. We defined the different patterns of enhancement at the surgical site as follows: . Enhancement that resembled normal dural enhancement was designated as thin linear (Fig. 1a and b). . Enhancement that was thicker than typical dural enhancement or E/5 mm was classified as thick linear (Fig. 1c and d). If we detected nodularity, we named it as thick linear-nodular enhancement (Fig. 1e). . Enhancement that was thicker than 1 cm in any imaging plane was considered to indicate residual tumor (Fig. 1e). These three patterns of early-postoperative surgical enhancement were followed on the consecutive radiological examinations. Regrowth and recurrence were defined as follows: a mass lesion in the resection bed at the same location as the primary tumors was named as ‘regrowth tumor’ (Fig. 2a and b), and a mass lesion in the brain parenchyma distant from the resection bed was named as ‘recurrent tumor’ (Fig. 2c). After classifying the patients into groups according to the surgical site enhancement patterns, the diagnosis of regrowth and recurrent tumors were made if there was at least a 20% increase in tumor size at any follow-up
G. Ekinci et al. / European Journal of Radiology 45 (2003) 99 /107
101
Fig. 1. (a) and (b) There is thin linear enhancement at the resection site in EPM on contrast-enhanced coronal T1-weighted images (long arrow). The short arrow shows the dural enhancement at the operation site. On contrast-enhanced axial (c) and sagital (d) T1-weighted images display the thick linear enhancement at the operation site. (e) There is gross residual tumor at the resection site (long arrow) together with thick linear-nodular enhancement (short arrow) on contrast-enhanced sagital T1-weighted image.
MRI examination when compared to the EPMR or the MRI where the tumor first appeared. Histopathological
confirmation was also made with stereotactic biopsy and/or reoperation in these patients.
102
G. Ekinci et al. / European Journal of Radiology 45 (2003) 99 /107
We also noted the hemorrhage in the surgical area and extraaxial locations to clarify its effects on image prediction. For each group of patients, we analyzed the EPMR enhancement patterns and tested the correlation between these and the various primary tumor types.
3. Results 3.1. Residual tumors EPMR revealed that 20 of the 50 patients had residual tumors. Residual tumor was found in two of the 11 patients who had grade II tumors, and in 18 of the 39 patients who had high-grade tumors (Table 1). The residual tumors were confirmed by at least a 20% increase in size on follow-up imaging. Eleven of 20 residual tumors were reoperated and nine of 20 residual tumors were stereotacticaly biopsied. Based on intraoperative assessment, the surgeon reported complete tumor excision in 9 of the 20 patients whose EPMR confirmed residual neoplasm, and acknowledged the presence of residual tumor in the other 11 cases (Fig. 3a and b). In the remaining 30 patients without residual tumors, the surgeon reported total tumor excision. 3.2. Regrowth and recurrence The regrowth and recurrent tumors detected on follow-up examinations are given in Table 1. Regarding time of detection for tumor regrowth; four cases (3 GBM, one oligodendroglioma Grade III) were identified in the first month after surgery, four cases (3 GBM, one astrocytoma Grade III) were diagnosed at 3 months, and two cases (1 GBM, one oligoastrocytoma Grade II) were detected at 6 months on follow-up. One patient with GBM, who had no residual tumor on EPMR and no regrowth on follow-up developed recurrence. In all the patients who exhibited regrowth or recurrence, we noted an increase in tumor size of at least 20% in the further follow-up, as compared to the EPMR findings or the MRI in which the tumor first appeared. These cases were also biopsied stereotacticaly for histopathological confirmation. Fig. 2. (a) There are thick linear-nodular enhancements at the resection bed (arrows) on contrast-enhanced coronal T1-weighted EPMR image. (b) In same patient, there is regrowth tumor at the resection bed on contrast-enhanced coronal T1-weighted image at the third month (arrow). (c) In another patient, on third month contrastenhanced sagital T1-weighted image, there is regrowth (short arrow) and 2 recurrent (long arrows) tumors in a patient having GBM with no residual tumor and with thin linear enhancement at the resection bed on EPMR.
3.3. Resection bed enhancement pattern The resection site enhancement patterns detected on EPMR are shown in Table 2. Table 3 lists the resection bed enhancement patterns on EPMR for the residual, regrowth, and recurrent tumors.
G. Ekinci et al. / European Journal of Radiology 45 (2003) 99 /107
103
Table 1 Postoperative MRI findings Tumor
Grade (No.)
MRI findings EPMRb residual tumor
Follow-up MRI Regrowth
GBMa
25
Oligodend.
II (4) III (5)
16
Astrocyto.
II (5) III (6)
2 1
Oligoastro.
II (2) III (3)
1
Sum a b
50
20
7
Recurrence 1
No tumor 1
1
4 4
1 1
2 4 2 2
10
1
19
GBM: glioblastoma multiforme. EPMR: early-postoperative MRI.
Four of the five patients whose EPMR imaging showed no enhancement in the surgical site developed a thin linear enhancement pattern in the surgical site on the first month of follow-up. One patient never showed any resection bed enhancement. All of the regrowth tumors developed from the nodular region of the thick linear-nodular enhancement areas. The only case of recurrent tumor that was not associated with any residual or regrowth tumor exhibited a thin linear enhancement pattern in the resection bed. We calculated the sensitivity, specificity, positive and negative predictive values of EPMR for predicting regrowth tumor based on enhancement patterns (Table 4). EPMR sensitivity was 91%, specificity was 100%, positive predictive value (PPV) was 1, and negative predictive value (NPV) was 0.9375. Eighteen of the 19 patients who had no residual tumor, regrowth, or recurrence exhibited a thin linear contrast enhancement pattern in the resection bed on follow-up examinations. One patient in this group showed no resection bed enhancement on follow-up. The resection site enhancements disappeared at 1 month in 11 patients, at 3 months in 5 patients, at 6 months in 1 patient, and at 1 year in 1 patient. 3.4. Blood at the surgical site and extraaxial locations
Fig. 3. (a) On contrast-enhanced coronal T1-weighted EPMR image, there is gross residual tumor at the resection bed (arrow), although the surgeon reported the gross total tumor resection. (b) On contrastenhanced sagital T1-weighted image at the sixth month, there is remarkable increase in residual tumor size (arrow).
EPMR showed evidence of blood in the surgical area (resection cavity and adjacent parenchyma) in 12 of the 50 cases studied. In all patients, this resolved in the first month after surgery. Extraaxial blood (subarachnoidal, subdural, epidural) was seen in only 3 patients, and other types of extraaxial fluid collection were noted in 9 of the 50 cases. Follow-up of extraaxial fluid and blood collections showed that these also resolved within 1 month. On EPMR, the intraaxial fluid collection was
G. Ekinci et al. / European Journal of Radiology 45 (2003) 99 /107
104
Table 2 Contrast enhancement at the surgical site on EPMR Tumor
Grade (No.)
Enhancement pattern None
a
GBM
IV (25)
Oligodend.
II (4) III (5)
Thin linear 3
4 1
II (5) III (6)
2 4
Oligoastro.
II (2) III (3)
2 2
a
50
5
Thick linear-nodular
2
20
3
Astrocyto.
Sum
Thick linear
1 1 1
2 1 1
16
4
25
GBM: glioblastoma multiforme.
Table 3 EPMR contrast enhancement in the resection site after surgery for all groups of patients Enhancement of the resection bed
Residual
Thin linear Thick linear Thick linear-nodular
4 16
a
Regrowth
Recurrence
No residual, regrowth, or recurrencea
1
1
14
9
Five patients showed no enhancement at the resection bed on EPMR.
localized to the resection cavity. On EPMR, the fluid signal had a relatively higher and heterogeneous signal compared to the cerebrospinal fluid on T1-weighted images. However, the signal characteristics of fluid in the resection bed cavity on T1-weighted images were not as hyperintense as blood, and we encountered no difficulties in evaluating the enhancement.
4. Discussion Postoperative radiological imaging is used to determine the amount of tumor mass that has been removed, to detect the surgical complications as early as possible, and to follow the response of the remaining tumor to radiotherapy and chemotherapy [3,5,6]. Researchers have used CT and MRI to evaluate the postoperative tumoral contrast enhancement and surgically induced changes. The findings have varied as the techniques have improved over the years. The most reliable radiological finding associated with glial neoplasms is the contrast enhancement of tumoral components. The parts of these growths that enhance in the various imaging modalities are usually solid, richly vascular, and densely cellular [2,4 /8]. This knowledge is applied when formulating treatment strategies for glial tumors, and in determining the response to therapy [2,4 /6,8 /13]. The edges of the enhanced areas represent the macroscopic borders of the glioma, but the tumor cells do not necessarily stay within these margins. The
cells can infiltrate the brain parenchyma and the edema surrounding the mass. This cellular infiltration is reflected by a high-intensity signal on T2-weighted MRI. Studies have shown that this cellular infiltration can reach 3 cm or more beyond the enhanced edges of the tumor mass [2 /7,14 /19]. In such cases, MR spectroscopy (MRS) has been reported to be useful, but it is not routinely applicable due to cost and time [20]. In cases of glial tumors, the surgeon’s main goal is to completely excise the mass through its macroscopic borders. The surgery allows a histological diagnosis to be made, and improves the patient’s neurological status. Moreover, any residual tumor is more amenable to radiotherapy after the debulking process. Although the surgeon can roughly estimate the amount of residual tumor preoperatively, some remaining tumor may not be visible to the naked eye [3,4,12,16,21 /23]. PeroperaTable 4 Sensitivity and specificity of EPMR for regrowth tumors according to enhancement patterns excluding residual tumors Regrowth tumor
Thick linear-nodular Thin linear Total
Present
Absent
9 1 10
0 15 15
Total
9 16 25a
Sensitivity: 0.91; Specificity: 1; PPV: 1; NPV: 0.9375. a Only patients with no residual tumor on EPMR were included.
G. Ekinci et al. / European Journal of Radiology 45 (2003) 99 /107
tive MR has gained more importance recently, but it still is not a widely used method [9]. Postoperative imaging enhanced by intravenous contrast injection is the method of choice for detecting residual tumor and evaluating the tumor extension after surgery. It is sometimes difficult to differentiate the tumoral enhancement from enhancement due to the normal postoperative changes. Benign contrast enhancement of the surgical site is assumed to result from the local impairment of the blood/brain barrier, neovascularity, luxury perfusion, or extravasation of contrast material [3,6,7,12,24 /28]. MRI has replaced CT in the postoperative evaluation of residual tumors and benign enhancement at the resection bed. Various authors have emphasized the superiority of MRI for detecting residual tumors. Compared to CT, MRI offers higher resolution of contrast enhancement, better contrast resolution, and fewer uncertain or false-negative findings [2 /8]. Numerous MRI studies have investigated postoperative changes with the goal of identifying the optimal time after surgery to assess benign and malignant contrast enhancement findings [4,7,8,25,29 /33]. Focusing on an earlier time point in our study, we sought to investigate the predictive value of EPMR after glial tumor surgery. Our results showed that EPMR revealed enhancement of the resection bed in 45 of 50 cases (90%), a finding that has also been mentioned in previous studies [7,8,27,32,34,35]. On this basis, we dispute the notion that there is a ‘diagnostic window’’ in the first 72 h after surgery in which all enhancements seen in the resection bed on EPMR is caused by residual tumors. Our findings are also not consistent with the suggestion that all enhancements seen in the resection bed on EPMR are caused by residual tumors [4,26 / 28,31,34,35]. If the presence of enhancement in the resection bed on EPMR is not specific for the presence of a residual tumor, is it possible for the pattern of enhancement to provide information about residual tumor tissue? First of all, if EPMR reveals an enhancing mass larger than 1 cm in any imaging plane, the pattern of enhancement in the resection bed is not relevant because postoperative treatment is guided by the finding of the residual tumor. The emphasis should be on evaluating enhancement patterns in the resection bed in cases where no mass or overt residual tumor is detected. Several studies indicate that enhancement peaks in the first week (5 /7 days), and is detected until the end of the first postoperative month in benign situations. It is also reported that during the first 3 days, benign enhancement is linear, but after 5 /7 days there is significant and intense thick linear and nodular enhancement, so if MRI is done during this period it is very difficult to differentiate between postoperative changes and residual tumors [4,7,8,25/28,31,33].
105
Opinions differ regarding the evaluation of enhancement patterns seen on MRI in the first 3 days after surgery. The most recent and broadly based study on this subject was conducted by Oser et al., who examined 46 MRI studies of 43 cases. They found no evidence that postoperative enhancement pattern indicated the presence of residual tumor [32]. The patients in this study had not only glial tumors, but also a more heterogeneous group of CNS tumors. Also to be noted is that 19 of the individuals had undergone previous surgery for brain tumor. The authors did not report the enhancement pattern findings for the 15 of these 19 cases in which preoperative assessment had shown no residual tumor, but they did note enhancement in the resection bed on these images. In our study, 25 of the 30 cases in which EPMR showed no residual tumors, there was enhancement in the resection bed within the first 24 h. One of these 25 cases developed recurrence during follow-up, and 10 cases developed tumor regrowth. Nineteen of 30 cases showed no regrowth or tumor recurrence. In 9 of the 10 patients who developed regrowth during follow-up, a thick linear-nodular pattern was present, however a thin linear pattern was present in only one case. Other studies have identified thin linear enhancement as a reflection of benign changes or changes secondary to surgical trauma, whereas the enhancement pattern in cases of residual or recurrent tumors has been reported as thick linear-nodular [4,7,8,25,26,29 /31,33,36,37]. In cases where EPMR revealed no residual tumor, regrowth tumors developed from the region where the thick linear-nodular enhancement pattern was seen [4,8,29,33,36,37]. Similarly, regrowth tumors in our study developed from thick linear-nodular enhancements, except one patient. For regrowth tumors, PPV of EPMR was 100% and NPV was 0.9375. Considering all cases in our study, the rate of tumor regrowth and tumor recurrence was 6.6% (2/30) in cases with thin linear enhancement pattern in the resection bed. When the group that had no residual tumor on EPMR was considered and GBM patients were excluded, we detected no regrowth or recurrence in cases with thin linear enhancement at the surgical site. We believe that thin linear enhancement observed in nonGBM cases can be considered benign. In our GBM cases that showed no residual tumor on EPMR, but exhibited thin linear contrast enhancement at the operative site, the rate of tumor regrowth was 11.1% (1/9). In other words, the finding of a thin linear pattern in the resection bed of GBM cases on EPMR is benign 88.9% of the time. Forsyth and co-workers did not recommend the routine use of EPMR for postoperative evaluation of malignant gliomas [7]. They investigated only 17 malignant cases, and 3 were recurrent cases that had been previously operated. In 10 of these 17 patients, EPMR
106
G. Ekinci et al. / European Journal of Radiology 45 (2003) 99 /107
confirmed the presence of residual tumor. Our study group included 39 high-grade gliomas (grade III and GBM), and EPMR results were evaluated against 1-year postoperative follow-up MRI findings. Contrary to the opinion of Forsyth et al., our results revealed that EPMR should be included in the routine protocol after glioma surgery. Intraoperative MRI, which increases the percentage of the complete resection of tumor by showing the residual tumor intraoperatively, MRS, and PET [9,20,38] examinations are more frequently used recently to detect residual and recurrent tumors with success. However, all of these techniques are more expensive examinations. Routine contrast enhanced MRI is the basic study for preoperative diagnosis and for the comparison of postoperative findings with the preoperative status. The radiological findings on routine preoperative MRI examinations should be a guide in the evaluation of preoperative MRI, MRS, and PET studies. Researchers have shown that benign enhancement in the resection bed after the removal of gliomas or any other surgery of CNS due to nontumoral lesions disappears in a mean time of 1/3 months, but this process can take up to 6 /8 months [3,4,7,8,24 / 28,30,32,33,39]. In our study, cases that showed no residual tumor, regrowth, or recurrence showed enhancement at the surgery site for 6 months after the operation, but no enhancement was observed at the 1year follow-up MRI. All of our patients received radiotherapy 1 month after surgery. The acute and early delayed complications occurring during radiotherapy and within 2/3 months after radiotherapy, respectively, are not permanent in most cases, therefore they do not interfere with the interpretation of images for tumor regrowth [38,40]. Severe changes like radiation necrosis that could resemble a tumor tend to appear 1 year or more after radiotherapy [40]. All regrowth tumors in our study were detected within 6 months postoperatively. This time interval was shorter than the time needed for a radiation necrosis to become visible. Therefore we did not consider radiation induced radiological changes in our study. Besides, stereotactic biopsy was performed in all regrowth cases. Reports indicate that the most obscuring element behind the resection bed enhancement on EPMR is blood in the surgical area [4,7,8,25,32,33,41,42]. Blood is extravasated when vessels are damaged during surgery, and even the most meticulous hemostasis may not prevent some blood leakage. The ability to detect blood on MRI depends on the transformation of hemoglobin to deoxyhemoglobin and methemoglobin over time [4,7,32,42]. Investigations have shown that methemoglobin was present during early postoperative period, and the reported reason for increased methemoglobin
formation was surgical procedures. Methemoglobin, which is hyperintense on T1W images, can prevent the visualization of enhancement due to residual tumor on very early postoperative MRI [4,7,8,27,32,34]. Various amounts of blood have been reported to be present in the surgical area (resection cavity and neighboring parenchyma) in cases of CNS surgery (tumor and nontumor cases with rates ranging from 26 to 100% [4,7,8,25,30,32]. In our study, blood was detected in the surgical area in 24% of the cases. Extraaxial blood (subarachnoidal, subdural, epidural) was detected in 6% of cases, whereas in other reports this rate has ranged from 6 to 26% [4,7,8,25,30,32]. The reasons for these discrepancies may be differences in surgical techniques, imaging methods, and/or the time of postoperative imaging. None of the cases in our study group had large amounts of blood in the resection bed that prevented the evaluation of enhancement patterns. The most important point here is the need for simultaneous inspection of pre- and post-contrast images. In addition, postcontrast images should be acquired in three planes instead of one designated plane. This permits more accurate enhancement evaluation in the surgical bed. We cannot comment on late-postoperative hematomas because there were no such lesions in our study group. MRI is three times better than the surgeon’s naked eye for detecting residual tumors [4,7]. Of our 20 cases of EPMR-confirmed residual tumors, the surgeon reported complete excision in 9 (45%) cases, and residual tumor after resection in 11 (55%) cases.
5. Conclusion Our study results revealed that, radiological detection of residual tumor is far more sensitive than intraoperative estimation by the surgeon. The suggestion of a ’diagnostic window’’ on EPMR in the first 72 h after surgery is not supported, and a thick linear-nodular enhancement pattern at the resection site should be considered pathologic, and should be followed closely for tumor regrowth. A thin linear enhancement pattern can usually be accepted as benign contrast enhancement of the surgical site. Our study shows that EPMR yields reliable and accurate information for the detection of residual tumors, and enables us to predict possible tumor regrowth. In our opinion, EPMR should be performed in all adult patients who undergo surgical treatment for supratentorial glial tumors.
References [1] Bonnin JM, Garcia JH. Histology and growth characteristics of brain neoplasms. In: Taveras JM, Ferrucci JT, editors. Radiology,
G. Ekinci et al. / European Journal of Radiology 45 (2003) 99 /107
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15] [16]
[17]
[18]
[19]
[20]
[21]
vol. 3 (Revised ed., Chapter 52). J.B. Lippincott Company, 1990:1 /11. Madison MT, Hall WA, Latchaw RE, Loes DJ. Radiologic diagnosis, staging, and follow-up of adult central nervous system primary malignant glioma. Radiol Clin N Am 1994;32(1):183 /96. Kelly WM, Brant-Zawadzki B. Magnetic resonance imaging and computed tomography of supratentorial tumors. In: Taveras JM, Ferrucci JT, editors. Radiology, vol. 3 (Revised ed., Chapter 53). J.B. Lippincott Company, 1990:1 /22. Henegar MM, Moran CJ, Silbergeld DL. Early postoperative magnetic resonance imaging following nonneoplastic cortical resection. J Neurosurg 1996;84:174 /9. Holt RM, Maravilla KR. Supratentorial gliomas: imaging. In: Wilkins RH, Rengachary SS, editors. Neurosurgery, vol. 1B (Chapter 79). McGraw-Hill, 1996:753 /88. Albert FK, Forsting M, Sartor K, et al. Early postoperative magnetic resonance imaging after resection of malignant glioma: objective evaluation of residual tumor and its influence on regrowth and prognosis. Neurosurgery 1994;34(1):45 /61. Forsyth PAJ, Petrov E, Mahallati H, et al. Prospective study of postoperative magnetic resonance imaging in patients with malignant gliomas. J Clin Oncol 1997;15(5):2076 /81. Haba D, Papon AP, Tanguy JY, et al. Use of half-dose gadolinium-enhanced MRI and magnetization transfer saturation in brain tumors. Eur Radiol 2001;11:117 /22. Schwartz RB, Wong TZ, et al. Intraoperative MR imaging guidance for intracranial neurosurgery. Radiology 1999;211:477 /8. Cairncross JG, Macdonald DR, Pexman JHW, Ives FJ. Steroid induced CT changes in patients with recurrent malignant glioma. Neurology 1988;38:724 /6. Cohadon F. Indications for surgery in the management of gliomas, in advances and technical standards in neurosurgery, vol. 17. Wien: Springer, 1990:189 /233. Levin VA, Hoffman WF, Heilbron DC, Norman D. Prognostic significance of the pretreatment CT scan on time to the progression for patients with malignant gliomas. J Neurosurg 1980;52:642 /7. Salcman M, Levine H, Rao K. Value of sequential computed tomography in the multimodality treatment of glioblastoma multiforme. Neurosurgery 1981;8:15 /8. Husstedt HW, Sickert M, Ko¨stler H, et al. Diagnostic value of the fast-FLAIR sequence in MR imaging of intracranial tumors. Eur Radiol 2000;10:745 /52. Masters LT, Zimmerman RD. Imaging of supratentorial brain tumors in adults. Neuroimag Clin N Am 1993;3(4):649 /69. Ciric I, Vick NA, Mikhael MA, Cozzens J, Eller T, Walsh A. Aggressive surgery for malignant supratentorial gliomas. Clin Neurosurg 1990;36:375 /83. Burger PC, Dubois BJ, Schold SC, Smith KR, Jr., et al. Computerised tomographic and pathologic studies of untreated, quiescent, and recurrent glioblastoma multiforme. J Neorosurg 1983;58:159 /69. Burger PC, Heinz ER, Shibata T, Kleihues P. Topographic anatomy and CT correlations in the untreated glioblastoma multiforme. J Neurosurg 1988;68:698 /704. Kelly PJ, Daumos-Duport C, Kispert DB, et al. Imaging based stereotaxic serial biopsies in untreated intracranial glial neoplasms. J Neorosurg 1987;66:865 /74. Burtscher IM, Skagerberg Gastro CT, Geizer B, et al. Proton MR spectroscopy and preoperative diagnostic accuracy. AJNR 2000;21:84 /93. Salcman M. Surgical decision making for malignant brain tumors. Clin Neurosurg 1987;35:285 /313.
107
[22] Shapiro WR, Green SB, Burger PC, et al. Randomised trial of three chemotherapy regimens and two radiotherapy regimens in postoperative treatment of malignant glioma. J Neurosurg 1989;71:1 /9. [23] Kyritsis AP, Soyo H, Levin VA. Molecular pathogenesis and management of gliomas. Neuroimag Clin N Am 1993;3(4):735 / 44. [24] Takahashi M. Supratentorial neoplasms */postoperative changes. In: Taveras JM, Ferrucci JT, editors. Radiology, vol. 3 (Revised ed., Chapter 57). J.B. Lippincott Company, 1990:1 /5. [25] Sato N, Bronen RA, Sze G, et al. Postoperative changes in the brain: MR imaging findings without neoplasms. Radiology 1997;204:839 /46. [26] Cairncross JG, Pexman JHW, Rathbone MP, DelMaestro RF. Postoperative contrast enhancement in patients with brain tumor. Ann Neurol 1985;17:570 /2. [27] Laohaprasit V, Silbergeld DL, Ojemann GA, et al. Postoperative CT contrast enhancement following lobectomy for epilepsy. J Neurosurg 1990;73:392 /5. [28] Jeffries BF, Kishore PRS, Singh KS, et al. Contrast enhancement in the postoperative brain. Radiology 1981;139:409 /13. [29] Meyers SP, Wildenhain S, Chess MA, Tarr RW. Postoperative evaluation for intracranial recurrence of medullablastoma: MR findings with gadopentetate dimeglumine. AJNR 1994;15:1425 / 34. [30] Elster AD, DiPersio DA. Cranial postoperative site: assessment with contrast-enhanced MR imaging. Radiology 1990;174:93 /8. [31] Forsting M, Albert FK, Kunze S, et al. Extirpation of glioblastomas: MR and CT follow-up of residual tumor and regrowth. AJNR 1993;14:77 /87. [32] Oser AB, Moran CJ, Kaufman BA, Park TS. Intracranial tumor in children: MR findings within 24 hours of craniotomy. Radiology 1997;205(3):807 /12. [33] Rollins NK, Nisen P, Shapiro KN. The use of early postoperative MR in detecting residual juvenile cerebellar pilocytic astrocytoma. AJNR 1998;19:151 /6. [34] Sze G. New applications of MR contrast agents in neuroradiology. Neuroradiology 1990;32:421 /38. [35] Ciric I, Ammirati M, Vick N, Mikhael M. Supratentorial gliomas: surgical considerations and immediate postoperative results. Neurosurgery 1987;21:21 /6. [36] Nicoletti GF, Barone F, Passanisi M, Mancuso P, Albanese V. Linear contrast enhancement at the operative site on early postoperative CT after removal of brain tumors. J Neurosurg Sci 1994;38:131 /5. [37] Bird CR, Drayer BP, Medina M, et al. Gd-DTPA-enhanced MR imaging in pediatric patients after brain tumor resection. Radiology 1988;169:123 /6. [38] Rabin BM, Meyer JR, Berlin JW, et al. Radiation-induced changes in the central nervous system and head and neck. Radiographics 1996;16:1055 /72. [39] Bronen RA, Sze G. Magnetic resonance imaging contrast agents: theory and application to the central nervous system. J Neurosurg 1990;73:820 /39. [40] Boyko OB. Neuroimaging of radiation injury to the central nervous system. Neuroimaging Clin North Am 1993;3:803 /16. [41] Meyding-Lamade´ U, Forsting M, Albert F, et al. Accelerated methaemoglobin formation: potential pitfall in early postoperative MRI. Neuroradiology 1993;35:178 /80. [42] DiChiro G, Brooks RA, Girton ME, et al. Sequential MR imaging studies of intracerebral hematomas in monkeys. AJNR 1985;7:193 /9.