Radiology
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online before print October 10, 2006, 10.1148/radiol.2413051276
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
2413051276v1
241/3/839    most recent
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Higano, S.
Right arrow Articles by Takahashi, S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Higano, S.
Right arrow Articles by Takahashi, S.
(Radiology 2006;241:839-846.)
© RSNA, 2006

Neuroradiology

Malignant Astrocytic Tumors: Clinical Importance of Apparent Diffusion Coefficient in Prediction of Grade and Prognosis1

Shuichi Higano, MD, PhD, Xia Yun, MD, PhD, Toshihiro Kumabe, MD, PhD, Mika Watanabe, MD, PhD, Shunji Mugikura, MD, PhD, Atsushi Umetsu, MD, PhD, Akihiro Sato, MD, PhD, Takayuki Yamada, MD, PhD and Shoki Takahashi, MD, PhD

1 From the Departments of Diagnostic Radiology (S.H., S.M., A.U., A.S., T.Y., S.T.), Neuroendovascular Therapy (X.Y.), and Neurosurgery (T.K.), Tohoku University Graduate School of Medicine, 1-1 Seiryo-machi Aoba-ku, Sendai, Miyagi, 980-8574, Japan; and Department of Pathology, Tohoku University Hospital, Tohoku, Japan (M.W.). Received July 31, 2005; revision requested October 7; revision received November 15; accepted December 8; final version accepted February 3, 2006. Address correspondence to S.H. (e-mail: higano-s{at}rad.med.tohoku.ac.jp).


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ADVANCES IN KNOWLEDGE
 References
 
Purpose: To retrospectively assess the apparent diffusion coefficient (ADC) for prediction of malignancy and prognosis of malignant astrocytic tumors.

Materials and Methods: The institutional review board approved this study and did not require patient informed consent. Findings from 37 consecutive patients (21 men, 16 women; mean age, 43 years) with pathologically proved malignant astrocytic tumors that included 22 glioblastomas (GBMs) and 15 anaplastic astrocytomas (AAs) were retrospectively evaluated. The minimum ADC value of each tumor was preoperatively determined from several regions of interest defined in the tumor, preferably with avoidance of cystic or necrotic components, on ADC maps derived from isotropic diffusion-weighted images. Surgical intervention, followed by radiation therapy, was undertaken in all cases according to hospital protocol. Immunohistologically, Ki-67 labeling index (LI), indicating cell proliferation, was also determined. The patients were classified into two groups, progressive and stable, according to the 2-year observation after the initial treatment. Correlation analysis (Pearson product moment correlation), Student t test, Welch test, receiver operating characteristic analysis, and Kaplan-Meier method with log-rank test were used for statistical evaluation.

Results: There was a significant negative correlation between minimum ADC and Ki-67 LI (r = –0.562, P < .001). The mean minimum ADC (0.834 x 10–3 mm2 · sec–1) of GBM was significantly lower than that (1.06 x 10–3 mm2 · sec–1) of AA (P < .001, Student t test). The mean minimum ADC (0.80 x 10–3 mm2 · sec–1) of the progressive group was significantly lower than that (1.037 x 10–3 mm2 · sec–1) of the stable group (P < .001). The cutoff value of 0.90 x 10–3 mm2 · sec–1 for minimum ADC for differentiation of patients with a favorable prognosis from those with a poor prognosis provided the best combination of sensitivity (79%) and specificity (81%) (receiver operating characteristic analysis). The significant difference in the prognosis between two groups classified by using this cutoff value of minimum ADC was noted (P = .002, log-rank test).

Conclusion: The minimum ADC of malignant astrocytomas can provide additional information about their clinical malignancy related to posttreatment prognosis.

© RSNA, 2006


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ADVANCES IN KNOWLEDGE
 References
 
Astrocytic tumors are the most common primary brain neoplasms and account for more than 70% of all gliomas (1). The overall prognosis of malignant astrocytic tumors, especially glioblastomas (GBMs), is still poor in spite of aggressive treatments (1,2). In the clinical setting, however, the prognosis of these tumors varies from patient to patient; some patients have a relatively favorable prognosis, and others have a very poor prognosis, despite the same histopathologic diagnosis and equivalent treatments.

More exact pathologic evaluation of tumor malignancy that might closely correlate with the patient's prognosis would affect treatment planning. Ki-67 labeling index (LI) is one such optional immunohistochemical examination for the evaluation of tumor proliferation (3–6). This index has become an important part of histologic assessments and postsurgical assignment of grade for brain tumors, as a higher rate of Ki-67–positive cells corresponds to a greater malignancy of brain tumors. This method, however, is applicable to tumor specimens obtained at surgical intervention. The method that enables preoperative assessment of tumor malignancy to make a more effective therapeutic strategy possible for improvement in the prognosis has long been awaited.

Introduction of diffusion-weighted (DW) magnetic resonance (MR) imaging has enabled us to obtain additional information derived from microscopic motion of the water proton, which is not available by using conventional MR imaging. DW imaging has been applied for assignment of tumor grades or differentiation of tumors, as well as for diagnosis of ischemic stroke (7–14). Several investigators found an inverse correlation between the apparent diffusion coefficient (ADC) calculated from DW images and tumor cellularity (7,8,11,14,15). Some studies involving the assignment of grades to gliomas by using ADC showed the usefulness of ADC for such assignment (9,14), but others did not (12). To our knowledge, there has been no report in which preoperative assessment of ADC for prediction of posttherapeutic prognosis was discussed. Thus, the purpose of our study was to retrospectively assess the use of the ADC for the prediction of malignancy and the prognosis of malignant astrocytic tumors.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 ADVANCES IN KNOWLEDGE
 References
 
Patients
We reviewed the MR imaging database (X.Y.) from Tohoku University Graduate School of Medicine, Sendai, Miyagi, Japan. From the database, we selected consecutive patients with a pathologic diagnosis of a malignant astrocytic tumor who underwent preoperative MR imaging studies, including isotropic DW imaging, between April 1999 and January 2003 and whose results of clinical follow-up studies were available more than 2 years after their initial treatment. We excluded patients who had undergone a major therapeutic intervention, such as surgery or radiation therapy and chemotherapy, before initial MR imaging with DW imaging. Thirty-seven patients met the criteria, and the group comprised 21 male and 16 female patients who were 7–75 years in age (mean, 43 years). The ethics committee of Tohoku University Hospital, Tohoku, Japan, approved this retrospective study and did not require patient informed consent. Medical records, pathologic reports, surgical notes, and results of all MR imaging studies for all the patients were available. Pretreatment performance status of the patients was evaluated according to the Eastern Cooperative Oncology Group score, which ranges from 0 to 5, with 0 denoting perfect health and 5 denoting death (16).

Pathologic Evaluation and Surgery
The pathologic diagnosis included 22 GBMs and 15 anaplastic astrocytomas (AAs). The diagnosis was determined with specimens removed at surgical resection or biopsy, according to the World Health Organization criteria, by a neuropathologist (M.W.) with 15 years of experience. The specimens were obtained from both enhanced and nonenhanced areas of each tumor with reference to three-dimensional contrast material–enhanced MR images by using a neuronavigational system (ViewScope; Elekta IGS, Grenoble, France) during surgery or biopsy. In addition to the conventional histopathologic evaluation, the Ki-67 LI was retrospectively determined in 36 patients (M.W.). In this process, fields with the highest number of Ki-67–labeled cells were initially selected through a generalized survey, and then a percentage of positively labeled cells was determined by counting more than 1000 tumor nuclei at x400 magnification.

Surgical removal was undertaken in 30 patients, whereas stereotactic biopsy was performed in seven patients, and these patients included one with GBM and six with AAs. All of the procedures were performed by one neurosurgeon (T.K.) with 18 years of experience. With surgical removal, we intended to remove the maximum amount of tumor, which included nonenhanced portions of lesions and enhanced lesions, with minimum neurologic, physical, and systemic damage by using a neuronavigational system and intraoperative neurophysiologic mapping techniques. In 28 tumors, almost all of the tumor—a portion that included the entire enhanced components—was removed. For the other two tumors, some of the enhanced portions partially remained in place. Two neuroradiologists (S.H. and X.Y., with 20 and 3 years of experience, respectively) performed those determinations in consensus by comparing preoperative MR images with postoperative initial ones obtained within 1 week after the surgical intervention. The stereotactic biopsy was performed only in cases with a tumor that involved the brainstem (n = 3) or in those with a tumor that was extensively infiltrating the cerebral hemisphere and included eloquent areas such as the primary motor cortex, the pyramidal tract, or speech (Broca) areas (n = 4).

Postoperative Treatment and Follow-up
After surgery, all the patients subsequently received radiation therapy according to the protocol of our hospital. For AAs, local brain irradiation of 72 Gy for 30 days (5 days a week for 6 weeks) was delivered with the hyperfractionation method (1.2 Gy delivered twice a day). For GBM, whole-brain irradiation of 30 Gy for 15 days (5 days a week for 3 weeks) was delivered with the conventional method, and afterward, localized irradiation of 30 Gy for 10 days (5 days a week for 2 weeks) was delivered with the accelerated hyperfractionation method (1.5 Gy delivered twice a day). All the patients were evaluated at monthly outpatient examinations, and head MR imaging was performed every 2 months. For chemotherapy, 1-(4-amino-2-methyl-5-pyrimidinyl)-methyl-3-(2-chloroethyl)-3-nitrosourea, also called ACNU, was administered every 2 months for 2 years.

Observing these patients for 2 years after the initial treatment, we (S.H., X.Y.) classified them into stable and progressive groups according to their clinical courses and follow-up MR imaging findings. The stable group consisted of the patients who were alive and in whom additional enhanced lesions were not revealed at MR imaging, with the exception of transient contrast enhancement along the margin of resection cavities related to surgical intervention. In the patients of this group, the clinical status also remained stable and the residual tumors, including the nonenhanced portions, if any, showed no obvious change on follow-up MR images. The progressive group included patients who died of progression of the brain disease related to their tumors and those who developed pathologically proved recurrent tumors or obvious enhanced lesions suggestive of tumor recurrence and/or dissemination that were revealed at MR imaging. Although this classification was based on the status at 2 years after the initial treatments, the patients were followed up until January 2005 if they were available for evaluation of the prognosis.

MR Imaging and Data Processing
The patients were imaged by using a 1.5-T MR imaging system (Signa Horizon LX CV/i; GE Medical Systems, Milwaukee, Wis) and a conventional quadrature head coil. The contrast agent was gadopentetate dimeglumine (Magnevist; Nihon Schering, Osaka, Japan) or gadodiamide injection (Omniscan; Daiichi Pharmaceuticals, Tokyo, Japan). Nonenhanced and contrast-enhanced T1-weighted MR images, T2-weighted MR images, and DW images were obtained during the same imaging session without repositioning each patient's head. DW imaging was performed by using fat-suppressed spin-echo echo-planar imaging (repetition time msec/echo time msec, 5000/72; number of signals acquired, two; section thickness, 6 mm; gap, 2 mm; matrix, 128 x 128; field of view, 23 x 23 cm) with three orthogonal directional motion-probing gradients (b = 1000 sec/mm2), followed by automatic generation of isotropic DW images. Images without motion-probing gradients (b = 0 sec/mm2) were simultaneously obtained as well. The interval between the preoperative MR imaging studies and the surgery was shorter than 1 month (mean, 12 days). All the patients additionally underwent three-dimensional contrast-enhanced imaging with the neuronavigational system fewer than 3 days before surgery, and no substantial difference was noted in imaging findings between MR imaging studies with DW imaging and MR imaging studies with the neuronavigational system.

ADC maps were calculated from isotropic DW images and images obtained with a b value of 0 sec/mm2 by using standard software with a different workstation (Dr. View Pro 5.3; Ashahi Information System, Tokyo, Japan). The minimum ADC value of each tumor was determined by placing regions of interest by using the same workstation as was used to generate ADC maps in the following procedures (S.H.). We at first selected all continuous sections that included tumor. Several round- or oval-shaped regions of interest (area, approximately 0.3 cm2, 10 pixels) were carefully placed on each selected section of the ADC map to include the area with the lowest ADC value determined with visual inspection, preferably with avoidance of cystic, necrotic, or hemorrhagic components of the tumor with reference to conventional MR images. Finally, a value of a region of interest with the lowest ADC value was chosen from these regions of interest as a minimum ADC value of the tumor.

Statistical Analysis
The relationship between minimum ADC and Ki-67 LI was analyzed by using correlational analysis (Pearson product moment correlation). In comparing every parameter between two groups (AA and GBM groups, stable and progressive groups), equality of variances was evaluated by using the F test to select statistical tests. The minimum ADC values, with variances that were homoscedastic, were analyzed with the Student t test. The Welch test was applied to the analysis of the Ki-67 LI, patient age, and performance status because the variances of these parameters were not homoscedastic. A P value of less than .05 was considered to indicate a statistically significant difference. These statistical analyses were performed (S.H., S.M.) by using software (StatView 5.0, 1998; SAS Institute, Cary, NC).

A receiver operating characteristic analysis was applied to assess which cutoff value of minimum ADC had the best combination of sensitivity and specificity to allow differentiation of the stable group from the progressive group. The selection of the critical cutoff point varies with a trade-off between sensitivity and specificity. We considered a value that maximized the sum of sensitivity and specificity as the best cutoff point. The fitted receiver operating characteristic curve was constructed with software (ROCKIT, version 0.9.1b, 1998; http://xray.bsd.uchicago.edu/krl/KRL_ROC/software_index.htm#ROCKIT).

To compare the prognosis of the two groups classified according to the cutoff value of minimum ADC determined with the analysis mentioned previously, the Kaplan-Meier method and log-rank test were applied by using software (StatView 5.0). In this analysis, the event time was calculated as the following: For the patients who died of tumor progression or who had recurrent tumors, the time between the date of the initial surgery and that of death or recurrence was defined as the event time. The stable patients were censored at the last available follow-up date.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 ADVANCES IN KNOWLEDGE
 References
 
Ki-67 LI and Minimum ADC
Although no significant correlation between Ki-67 LI and minimum ADC was noted for the AA group (r = –0.416, P = .12) or the GBM group (r = –0.253, P = .27) separately, there was a significant negative correlation between these parameters for the malignant astrocytic tumors as a whole (r = –0.562, P < .001) (Fig 1). In regard to patient age, pretreatment performance status, minimum ADC value, and Ki-67 LI, compared between AA and GBM groups (Table 1), the mean minimum ADC value for the GBM group was significantly lower than that for the AA group, and there was overlap between the values for the two groups (Figs 1, 2). The mean value of Ki-67 LI also was significantly different between these two groups.


Figure 1
View larger version (13K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1: Graph shows relationship between minimum ADC and Ki-67 LI. Significant negative correlation was noted between these parameters for astrocytic tumors as a whole. GBMs generally show higher Ki-67 LI and lower minimum ADC, whereas AAs show lower Ki-67 LI and higher minimum ADC.

 

View this table:
[in this window]
[in a new window]

  		Table 1. Comparison of Patient Age, Pretreatment Performance Status, Minimum ADC Value, and Ki-67 LI for GBM and AA Groups

 

Figure 2
View larger version (10K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2: Graph shows relationship between minimum ADC and prognosis according to pathologic type. Minimum ADC values of AAs are higher than those of GBMs, although substantial overlap is noted. With the cutoff value of 0.90 x 10–3 mm2 · sec–1 for minimum ADC, most of the patients in the stable group have values above this threshold, and most of the patients in the progressive group have values below that level.

 
Follow-up
From the results of 2-year follow-up after the initial treatment, 19 patients (13 with AAs and six with GBMs) were classified as the stable group, and 16 patients (two with AAs and 14 with GBMs) were classified as the progressive group (Figs 3, 4). In the latter group, 13 patients (11 with GBMs, two with AAs) died of exacerbation of the brain tumors within the observation period, whereas in the remaining three patients with GBMs, recurrent tumors developed (pathologically proved at repeat surgery). Two patients with GBMs were excluded from this classification because they died of pneumonia within 2 years despite no clinical evidence of tumor recurrence. Therefore, the following analysis was performed by using 35 patients without these two patients.


Figure 3
View larger version (126K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3a: GBM in 56-year-old woman. Transverse (a) preoperative T2-weighted MR image (3600/96), (b) contrast-enhanced T1-weighted MR image (440/14), and (c) ADC map show irregularly shaped enhancing tumor (arrows in a) with perifocal edema in left mediofrontal lobe. Minimum ADC value was 0.72 x 10–3 mm2 · sec–1. (d) Postoperative contrast-enhanced transverse T1-weighted MR image (440/14) shows gross total removal of the tumor. (e) Contrast-enhanced transverse T1-weighted MR image (440/14) 9 months after surgery shows a patchy enhancing lesion (arrow). The lesion enlarged later, and gamma knife therapy was performed. With further progression of the disease, the patient died 1.5 years after initial treatment.

 

Figure 3
View larger version (120K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3b: GBM in 56-year-old woman. Transverse (a) preoperative T2-weighted MR image (3600/96), (b) contrast-enhanced T1-weighted MR image (440/14), and (c) ADC map show irregularly shaped enhancing tumor (arrows in a) with perifocal edema in left mediofrontal lobe. Minimum ADC value was 0.72 x 10–3 mm2 · sec–1. (d) Postoperative contrast-enhanced transverse T1-weighted MR image (440/14) shows gross total removal of the tumor. (e) Contrast-enhanced transverse T1-weighted MR image (440/14) 9 months after surgery shows a patchy enhancing lesion (arrow). The lesion enlarged later, and gamma knife therapy was performed. With further progression of the disease, the patient died 1.5 years after initial treatment.

 

Figure 3
View larger version (100K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3c: GBM in 56-year-old woman. Transverse (a) preoperative T2-weighted MR image (3600/96), (b) contrast-enhanced T1-weighted MR image (440/14), and (c) ADC map show irregularly shaped enhancing tumor (arrows in a) with perifocal edema in left mediofrontal lobe. Minimum ADC value was 0.72 x 10–3 mm2 · sec–1. (d) Postoperative contrast-enhanced transverse T1-weighted MR image (440/14) shows gross total removal of the tumor. (e) Contrast-enhanced transverse T1-weighted MR image (440/14) 9 months after surgery shows a patchy enhancing lesion (arrow). The lesion enlarged later, and gamma knife therapy was performed. With further progression of the disease, the patient died 1.5 years after initial treatment.

 

Figure 3
View larger version (115K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3d: GBM in 56-year-old woman. Transverse (a) preoperative T2-weighted MR image (3600/96), (b) contrast-enhanced T1-weighted MR image (440/14), and (c) ADC map show irregularly shaped enhancing tumor (arrows in a) with perifocal edema in left mediofrontal lobe. Minimum ADC value was 0.72 x 10–3 mm2 · sec–1. (d) Postoperative contrast-enhanced transverse T1-weighted MR image (440/14) shows gross total removal of the tumor. (e) Contrast-enhanced transverse T1-weighted MR image (440/14) 9 months after surgery shows a patchy enhancing lesion (arrow). The lesion enlarged later, and gamma knife therapy was performed. With further progression of the disease, the patient died 1.5 years after initial treatment.

 

Figure 3
View larger version (124K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3e: GBM in 56-year-old woman. Transverse (a) preoperative T2-weighted MR image (3600/96), (b) contrast-enhanced T1-weighted MR image (440/14), and (c) ADC map show irregularly shaped enhancing tumor (arrows in a) with perifocal edema in left mediofrontal lobe. Minimum ADC value was 0.72 x 10–3 mm2 · sec–1. (d) Postoperative contrast-enhanced transverse T1-weighted MR image (440/14) shows gross total removal of the tumor. (e) Contrast-enhanced transverse T1-weighted MR image (440/14) 9 months after surgery shows a patchy enhancing lesion (arrow). The lesion enlarged later, and gamma knife therapy was performed. With further progression of the disease, the patient died 1.5 years after initial treatment.

 

Figure 4
View larger version (137K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 4a: GBM in 43-year-old man. Transverse (a) preoperative T2-weighted MR image (3600/96), (b) contrast-enhanced T1-weighted MR image (440/14), and (c) ADC map show ill-defined tumor (arrows in a) in right insulo-opercular region. Irregular enhancing area is noted in center of tumor (arrow in b). Minimum ADC value was 1.07 x 10–3 mm2 · sec–1. (d) Postoperative contrast-enhanced transverse T1-weighted MR image (440/14) shows that most of the tumor, including all enhancing components, was removed. (e) Contrast-enhanced transverse T1-weighted MR image (440/14) 2 years after surgery shows no tumor recurrence.

 

Figure 4
View larger version (119K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 4b: GBM in 43-year-old man. Transverse (a) preoperative T2-weighted MR image (3600/96), (b) contrast-enhanced T1-weighted MR image (440/14), and (c) ADC map show ill-defined tumor (arrows in a) in right insulo-opercular region. Irregular enhancing area is noted in center of tumor (arrow in b). Minimum ADC value was 1.07 x 10–3 mm2 · sec–1. (d) Postoperative contrast-enhanced transverse T1-weighted MR image (440/14) shows that most of the tumor, including all enhancing components, was removed. (e) Contrast-enhanced transverse T1-weighted MR image (440/14) 2 years after surgery shows no tumor recurrence.

 

Figure 4
View larger version (104K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 4c: GBM in 43-year-old man. Transverse (a) preoperative T2-weighted MR image (3600/96), (b) contrast-enhanced T1-weighted MR image (440/14), and (c) ADC map show ill-defined tumor (arrows in a) in right insulo-opercular region. Irregular enhancing area is noted in center of tumor (arrow in b). Minimum ADC value was 1.07 x 10–3 mm2 · sec–1. (d) Postoperative contrast-enhanced transverse T1-weighted MR image (440/14) shows that most of the tumor, including all enhancing components, was removed. (e) Contrast-enhanced transverse T1-weighted MR image (440/14) 2 years after surgery shows no tumor recurrence.

 

Figure 4
View larger version (131K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 4d: GBM in 43-year-old man. Transverse (a) preoperative T2-weighted MR image (3600/96), (b) contrast-enhanced T1-weighted MR image (440/14), and (c) ADC map show ill-defined tumor (arrows in a) in right insulo-opercular region. Irregular enhancing area is noted in center of tumor (arrow in b). Minimum ADC value was 1.07 x 10–3 mm2 · sec–1. (d) Postoperative contrast-enhanced transverse T1-weighted MR image (440/14) shows that most of the tumor, including all enhancing components, was removed. (e) Contrast-enhanced transverse T1-weighted MR image (440/14) 2 years after surgery shows no tumor recurrence.

 

Figure 4
View larger version (131K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 4e: GBM in 43-year-old man. Transverse (a) preoperative T2-weighted MR image (3600/96), (b) contrast-enhanced T1-weighted MR image (440/14), and (c) ADC map show ill-defined tumor (arrows in a) in right insulo-opercular region. Irregular enhancing area is noted in center of tumor (arrow in b). Minimum ADC value was 1.07 x 10–3 mm2 · sec–1. (d) Postoperative contrast-enhanced transverse T1-weighted MR image (440/14) shows that most of the tumor, including all enhancing components, was removed. (e) Contrast-enhanced transverse T1-weighted MR image (440/14) 2 years after surgery shows no tumor recurrence.

 
There was a significant difference in mean values of minimum ADC and Ki-67 LI between the stable and progressive groups (Table 2). All except two patients with AAs belonged to the stable group and had relatively higher minimum ADC values (Fig 2). The remaining two patients with AAs had a poor prognosis and relatively low minimum ADC values. Six patients with GBMs belonged to the stable group, and four of them had relatively higher minimum ADC values.


View this table:
[in this window]
[in a new window]

  		Table 2. Comparison of Patient Age, Pretreatment Performance Status, Minimum ADC Value, and Ki-67 LI for Stable and Progressive Groups

 
According to the receiver operating characteristic analysis (Fig 5), the cutoff value of 0.90 x 10–3 mm2 · sec–1 in minimum ADC generated the best combination of sensitivity (79%) and specificity (81%). Receiver operating characteristic analysis of patients with removal of almost the entire tumor yielded similar results (area under the receiver operating characteristic curve, 0.83), with sensitivity of 87% and specificity of 73% by using the same cutoff value.


Figure 5
View larger version (12K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 5: Empiric and fitted receiver operating characteristic curves of minimum ADC values for use in differentiation of stable patients from progressive patients. Area under the receiver operating characteristic curve was 0.854. When the minimum ADC is 0.90 x 10–3 mm2 · sec–1, the best combination of sensitivity (79%) and specificity (81%) is provided (arrow).

 
When the prognosis was compared between the two groups classified by using the previously mentioned cutoff value of minimum ADC (0.90 x 10–3 mm2 · sec–1) and the Kaplan-Meier method, the group with the higher minimum ADC showed significantly better outcome (P = .002, log-rank test) (Fig 6). If the same analysis was applied to other sets of pairs classified by using different cutoff values, namely 0.80 x 10–3 mm2 · sec–1 and 1.00 x 10–3 mm2 · sec–1, the P values achieved with the log-rank test were .01 and .1, respectively.


Figure 6
View larger version (12K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 6: Graph shows comparison of postoperative prognosis between two groups of patients classified according to cutoff value of 0.90 x 10–3 mm2 · sec–1 for minimum ADC by using Kaplan-Meier method. The significantly better outcome was noted in the group with a minimum ADC value above the cutoff value.

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
ADVANCES IN KNOWLEDGE
 References
 
We found a significant negative relationship between minimum ADC and Ki-67 LI for the malignant astrocytic tumors in general, although no significant correlation was noted for the AA group or the GBM group separately. Few reports have been published in which a direct comparison of ADC values with the Ki-67 LI was performed. Calvar et al (17) reported a significant inverse correlation between ADC values and Ki-67 LI in their analysis of 37 brain tumors with various pathologic findings. Findings in many studies (7,8,11,14,15), however, suggest that ADC values are inversely correlated and the choline signal at proton MR spectroscopy is positively correlated with tumor cell density in various kinds of brain tumors. Findings in several reports indicate that there is an inverse correlation between ADC values and choline signal in gliomas at MR spectroscopy (7,8).

In an MR spectroscopic study of astrocytomas, Tamiya et al found that the choline-creatine ratio correlated positively with Ki-67 LI and that the N-acetylaspartate–choline ratio correlated inversely with Ki-67 LI, and they suggested that the choline signal at MR spectroscopy should reflect cellular proliferation (18). Considering all these results together, it is understandable that the ADC values correlated with the Ki-67 LI. It is likely that tumors with high proliferation activity (high Ki-67 LI) should have potency of rapid growth, yielding high cellularity. Kiss et al histopathologically assessed cell density and Ki-67 LI in 54 astrocytic tumors and found a significant correlation between them (3).

Because our study was retrospective, the regions for measuring the minimum ADC did not exactly correspond to those for the Ki-67 LI. This might be one reason why the significant correlation was not noted for AA and GBM groups separately, because these gliomas are often very heterogeneous. The following discussion, however, may justify the meaning of the significant relationship between these parameters in all cases. The Ki-67 LI, which was obtained from the fields packed most closely with labeled cells, should have reflected the value from one of the regions with the highest cell density in the specimen. Similarly, the minimum ADC value should represent the value of the region with the highest cellularity of each tumor.

There was a significant difference in the mean minimum ADC values between the AA and GBM groups. In many studies (9,10,14) about the assignment of grades to gliomas, researchers found a significant difference in tumor ADC values between low-grade (World Health Organization grades I and II) and high-grade (World Health Organization grades III and IV) gliomas and/or metastases. As far as we know, however, no investigators observed a significant difference in ADC between high-grade gliomas, that is, AA (grade III) and GBM (grade IV). Most investigators evaluated diffusional properties in various pathologic types of tumors altogether rather than in a single type of tumor. Because the tumor cell density would be inherently different in each pathologic type, the same World Health Organization grade of different kinds of tumors may vary in regard to their cell density. We compared ADC values that focused on malignant astrocytic tumors (AA and GBM), which might have resulted in a significant difference. The overlap in minimum ADC values of each group, however, was so large that the differentiation between them could not be based solely on the value of the minimum ADC.

The mean minimum ADC value of the stable group was significantly higher than that of the progressive group. A significant difference in Ki-67 LI was also found between these two groups. Several investigators (3,5,6) have revealed the prognostic importance of Ki-67 LI in astrocytic tumors. Kiss et al (3) observed that there was a marked difference in survival periods between the patients with a tumor that exhibited a low level of Ki-67 LI and low cell density and those with a tumor that exhibited a high level of Ki-67 LI and high cell density. Torp (5) found that Ki-67 LI increased significantly with increasing malignancy grade of astrocytomas and that tumors with the higher Ki-67 LI had significantly poorer prognosis than those with lower indexes. Neder et al (6) also reported a close relationship between MIB-1 LI (equivalent to Ki-67 LI) and survival of patients with astrocytomas.

Such an index, however, is measurable only after harvesting samples of tumor specimens. We could not find any published reports that indicated the usefulness of preoperative assessment of tumor ADC for prediction of posttherapeutic prognosis. Oh et al (19) investigated the relationship between ADC values and survival time in patients with GBM. They compared ADC values after surgery but prior to radiation therapy with patient survival times, and they showed that there was a significantly shorter median survival time in patients with low ADC compared with that in patients with high ADC within the residual region of T2 elongation.

In the present study, the prognosis of each tumor after initial treatments (surgery or radiation therapy and chemotherapy) could be well predicted by using preoperative measurement of the minimum ADC of the tumor. The threshold values of 0.90 x 10–3 mm2 · sec–1) in minimum ADC provided the best combination of sensitivity and specificity for prediction of prognosis. Actually, when we compared the prognosis between the two groups classified by using this threshold, the group with the higher minimum ADC had a significantly better outcome. When we considered the result that indicated a significant correlation between minimum ADC and Ki-67 LI, it was not surprising that tumoral minimum ADC values should have preoperative prognostic importance in patients with a malignant astrocytoma.

Stereotactic biopsy often is performed to establish a diagnosis in patients with intracranial lesions. As gliomas are typically heterogeneous and can have different histologic grades in a single tumor, the wrong choice of biopsy site may lead to underestimation of a tumor grade, and underestimation may confound determination of the optimum treatment strategy. In such a situation, measurement of minimum ADC may aid in the selection of the appropriate site for the biopsy, because minimum ADC should indicate a region with the highest tumor cell density, or the most highly proliferative portion of the tumor.

Several tumor-specific factors such as necrosis and neovascularity may affect prognosis (2). Results of many studies have indicated the usefulness of MR spectroscopy and MR perfusion imaging, which can depict necrosis or neovascularity, in addition to ADC measurements for the prediction of grade or malignancy of brain tumors (18, 20). Some investigators suggested the benefits of a combination of these methods (10,21). Among the various noninvasive techniques, however, DW imaging should be available in many hospitals and is the easiest to use and the least time consuming of them. The postprocessing of the data also is simple, and variation in the analyzed results should be minimum.

One of the limitations of this study was that there was no accordance of the areas for the minimum ADC measurement with those for the Ki-67 LI, which was already discussed previously. A second limitation was that we used only the minimum ADC values of tumors for estimation of the prognosis. A patient's prognosis is believed to depend on the most malignant site within a heterogeneous tumor, and it is on the basis of this site that the assignment of a histologic tumor grade usually is determined. As we have shown before, tumoral ADC should correlate well with the cell density and Ki-67 LI. Thus, it may be reasonable to think the site with minimum ADC value should represent one of the most malignant portions of the tumor. When the postoperative prognosis is considered, however, aggressiveness of the peripheral regions of a tumor might be more important than the central portions because the peripheral portions, where tumor recurrence usually occurs, tend to remain after surgery. With this view in mind, the minimum ADC values of the peripheral portion might correlate better with the prognosis.

Another limitation was the sample size of this study. The number of subjects was not large, and the performance of surgery was not uniform, and this nonuniformity led to a mixture of patients who underwent surgery with those who underwent biopsy, as well as a mixture of patients who underwent subtotal removal of tumor with those who underwent total removal. We found a significant difference in minimum ADC values between the stable and progressive groups and tentatively determined the cutoff value of minimum ADC for differentiating these two groups. Substantial overlap, however, was noted between these groups. A similar analysis in a larger group of subjects with uniform treatments may be required to determine whether the cutoff point of minimum ADC would really be reliable for this purpose.

In conclusion, minimum ADC values of the tumor were well correlated with the Ki-67 LI and were related to tumoral prognosis. We believe that ADC analysis should be one of the clinically feasible techniques used for prediction of prognosis of malignant astrocytic tumors, and it might be useful for planning initial treatment strategy in patients with these malignant tumors.


   ADVANCES IN KNOWLEDGE
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ADVANCES IN KNOWLEDGE
References
 


   FOOTNOTES
 

Abbreviations: AA = anaplastic astrocytoma • ADC = apparent diffusion coefficient • DW = diffusion weighted • GBM = glioblastoma • LI = labeling index

Authors stated no financial relationship to disclose.

Author contributions: Guarantors of integrity of entire study, S.H., X.Y.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; manuscript final version approval, all authors; literature research, S.H., X.Y., T.K., S.M., A.U., T.Y.; clinical studies, T.K., M.W.; statistical analysis, S.H., X.Y., S.M., A.U., A.S., S.T.; and manuscript editing, S.H., X.Y., T.K., M.W., S.M., A.S., T.Y., S.T.


   References
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ADVANCES IN KNOWLEDGE
 References
 

  1. Kleihues P, Cavenee WK. World Health Organization classification of tumours: pathology and genetics of tumours of the central nervous system. Lyon, France: International Agency for Research on Cancer, 2000.
  2. Salcman M. Glioblastoma multiforme and anaplastic astrocytoma. In: Kaye AH, Law ER Jr, eds. Brain tumors: an encyclopedic approach. 2nd ed. London, England: Churchill Livingstone, 2001.
  3. Kiss R, Dewitte O, Decaestecker C, et al. The combined determination of proliferative activity and cell density in the prognosis of adult patients with supratentorial high-grade astrocytic tumors. Am J Clin Pathol 1997;107(3):321–331.[Medline]
  4. Torp SH, Granli US. Proliferative activity in human glioblastomas assessed by various techniques. APMIS 2001;109(12):865–869.[CrossRef][Medline]
  5. Torp SH. Diagnostic and prognostic role of Ki67 immunostaining in human astrocytomas using four different antibodies. Clin Neuropathol 2002;21(6):252–257.[Medline]
  6. Neder L, Colli BO, Machado HR, Carlotti CG Jr, Santos AC, Chimelli L. MIB-1 labeling index in astrocytic tumors: a clinicopathologic study. Clin Neuropathol 2004;23(6):262–270.[Medline]
  7. Gupta RK, Cloughesy TF, Sinha U, et al. Relationships between choline magnetic resonance spectroscopy, apparent diffusion coefficient and quantitative histopathology in human glioma. J Neurooncol 2000;50(3):215–226.[CrossRef][Medline]
  8. Gupta RK, Sinha U, Cloughesy TF, Alger JR. Inverse correlation between choline magnetic resonance spectroscopy signal intensity and the apparent diffusion coefficient in human glioma. Magn Reson Med 1999;41(1):2–7.[CrossRef][Medline]
  9. Bulakbasi N, Guvenc I, Onguru O, Erdogan E, Tayfun C, Ucoz T. The added value of the apparent diffusion coefficient calculation to magnetic resonance imaging in the differentiation and grading of malignant brain tumors. J Comput Assist Tomogr 2004;28(6):735–746.[CrossRef][Medline]
  10. Bulakbasi N, Kocaoglu M, Ors F, Tayfun C, Ucoz T. Combination of single-voxel proton MR spectroscopy and apparent diffusion coefficient calculation in the evaluation of common brain tumors. AJNR Am J Neuroradiol 2003;24(2):225–233.[Abstract/Free Full Text]
  11. Gauvain KM, McKinstry RC, Mukherjee P, et al. Evaluating pediatric brain tumor cellularity with diffusion-tensor imaging. AJR Am J Roentgenol 2001;177(2):449–454.[Abstract/Free Full Text]
  12. Lam WW, Poon WS, Metreweli C. Diffusion MR imaging in glioma: does it have any role in the pre-operation determination of grading of glioma? Clin Radiol 2002;57(3):219–225.[CrossRef][Medline]
  13. Muti M, Aprile I, Principi M, et al. Study on the variations of the apparent diffusion coefficient in areas of solid tumor in high grade gliomas. Magn Reson Imaging 2002;20(9):635–641.[CrossRef][Medline]
  14. Sugahara T, Korogi Y, Kochi M, et al. Usefulness of diffusion-weighted MRI with echo-planar technique in the evaluation of cellularity in gliomas. J Magn Reson Imaging 1999;9(1):53–60.[CrossRef][Medline]
  15. Guo AC, Cummings TJ, Dash RC, Provenzale JM. Lymphomas and high-grade astrocytomas: comparison of water diffusibility and histologic characteristics. Radiology 2002;224(1):177–183.[Abstract/Free Full Text]
  16. Oken MM, Creech RH, Tormey DC, et al. Toxicity and response criteria of the Eastern Cooperative Oncology Group. Am J Clin Oncol 1982;5(6):649–655.[Medline]
  17. Calvar JA, Meli FJ, Romero C, et al. Characterization of brain tumors by MRS, DWI and Ki-67 labeling index. J Neurooncol 2005;72(3):273–280.[CrossRef][Medline]
  18. Tamiya T, Kinoshita K, Ono Y, Matsumoto K, Furuta T, Ohmoto T. Proton magnetic resonance spectroscopy reflects cellular proliferative activity in astrocytomas. Neuroradiology 2000;42(5):333–338.[CrossRef][Medline]
  19. Oh J, Henry RG, Pirzkall A, et al. Survival analysis in patients with glioblastoma multiforme: predictive value of choline-to-N-acetylaspartate index, apparent diffusion coefficient, and relative cerebral blood volume. J Magn Reson Imaging 2004;19(5):546–554.[CrossRef][Medline]
  20. Aprile I, Muti M, Principi M, et al. A magnetic resonance comparative study between enhancement, rCBV and ADC in brain glioblastomas. Radiol Med (Torino) 2002;104(1-2):87–91.
  21. Yang D, Korogi Y, Sugahara T, et al. Cerebral gliomas: prospective comparison of multivoxel 2D chemical-shift imaging proton MR spectroscopy, echoplanar perfusion and diffusion-weighted MRI. Neuroradiology 2002;44(8):656–666.[CrossRef][Medline]



This article has been cited by other articles:


Home page
 RadiologyHome page
W. B. Pope, H. J. Kim, J. Huo, J. Alger, M. S. Brown, D. Gjertson, V. Sai, J. R. Young, L. Tekchandani, T. Cloughesy, et al.
Recurrent Glioblastoma Multiforme: ADC Histogram Analysis Predicts Response to Bevacizumab Treatment
Radiology, July 1, 2009; 252(1): 182 - 189.
[Abstract] [Full Text] [PDF]

Home page
 RadiologyHome page
R. Murakami, T. Hirai, T. Sugahara, H. Fukuoka, R. Toya, S. Nishimura, M. Kitajima, T. Okuda, H. Nakamura, N. Oya, et al.
Grading Astrocytic Tumors by Using Apparent Diffusion Coefficient Parameters: Superiority of a One- versus Two-Parameter Pilot Method
Radiology, June 1, 2009; 251(3): 838 - 845.
[Abstract] [Full Text] [PDF]

Home page
 Br. J. Radiol.Home page
T NIWA, M UENO, S OHKAWA, T YOSHIDA, T DOIUCHI, K ITO, and T INOUE
Advanced pancreatic cancer: the use of the apparent diffusion coefficient to predict response to chemotherapy
Br. J. Radiol., January 1, 2009; 82(973): 28 - 34.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Neuroradiol.Home page
E.J. Lee, S.K. Lee, R. Agid, J.M. Bae, A. Keller, and K. terBrugge
Preoperative Grading of Presumptive Low-Grade Astrocytomas on MR Imaging: Diagnostic Value of Minimum Apparent Diffusion Coefficient
AJNR Am. J. Neuroradiol., November 1, 2008; 29(10): 1872 - 1877.
[Abstract] [Full Text] [PDF]

Home page
 J Child NeurolHome page
L.-G. Vezina
Imaging of Central Nervous System Tumors in Children: Advances and Limitations
J Child Neurol, October 1, 2008; 23(10): 1128 - 1135.
[Abstract] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
2413051276v1
241/3/839    most recent
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Higano, S.
Right arrow Articles by Takahashi, S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Higano, S.
Right arrow Articles by Takahashi, S.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
RADIOLOGY RADIOGRAPHICS RSNA JOURNALS ONLINE