Keywords diffusion tensor imaging - white matter - supratentorial intra-axial - low-grade glioma
- high-grade glioma
Introduction
Intra-axial tumors are the most common brain tumors encountered by neurosurgeons.
The challenge that surgeons face in dealing with these tumors is the proximity of
these tumors to the eloquent areas,[1 ] damage to which can increase the morbidity in these patients. Diffusion tensor imaging
(DTI) can assess the white matter (WM) tract orientation with respect to tumor location
in numerous planes,[2 ]
[3 ] so it is a good option to plan for surgical approach and path.[4 ]
[5 ] We evaluated the role of DTI in planning the approach, pattern of tracts in various
histologies, and neurological outcome.
Methods
The study was conducted in the Department of Neurosurgery of Sawai Maan Singh Medical
College and Hospital, Jaipur, Rajasthan, India, from August 2019 to January 2021 after
approval of the ethical committee of the institution. All the patients of supratentorial
intra-axial tumors who fulfilled the inclusion criteria were included in the study
after obtaining well-informed consent from patients and their relatives. The institute
did not provide any funds or scholarships. In this prospective study, a total 91 patients
were included.
Inclusion Criteria
Patients with supratentorial intra-axial (ST-IA) brain tumors in all age groups who
were fit for surgery and follow-up were included in the study.
Exclusion Criteria
Supratentorial extraparenchymal brain tumors and infratentorial tumors.
Any contraindication to surgery or if the patient is unfit for surgery.
Patients in whom magnetic resonance imaging (MRI) could not be done.
History taking and thorough neurological/clinical examination were done for all the
patients included in the study. Preoperative MRI and DTI were done in all the patients
and the preoperative tumor volume and pattern of tracts in tractography were evaluated.
Surgical approach with site of corticotomy and extent of resection was planned by
consultants of the neurosurgery department using sulcal/gyral pattern adjacent to
the tumor in MRI and WM tract pattern in DTI in relation with the tumor, and the same
consultants operated upon these patients.
The post-op neurologic status was assessed at first, postoperative day 7 and then
at 3 weeks after surgery, and a comparison with the preoperative neurological status
was done. After surgery, MRI was done at 3 weeks and the tumor volume was assessed
compared with the presurgical volume. The relationship of the DTI findings with the
extent of resection and tumor histology was studied.
Magnetic Resonance Imaging Protocol
A MR study was done using a 3-T MR system of Philips (3T-Ingenia) having an encoding
head coil of eight-element phased-array sensitivity.
Sequences seen were the following: T2-weighted, fluid-attenuated inversion recovery
(FLAIR), diffusion-weighted imaging (DWI), and T1-weighted sequence with and without
contrast.
Diffusion Tensor Imaging
It was done by a single-shot spin echo planar pulse sequence with repetition time
of 3,000 microseconds, 94-microsecond echo time, thickness slice of around 2.5 mm,
92 × 88 matrix, and 90-degree flip angle with a nil slice gap. Using the regions of
interest (ROI) for each tract, the corticospinal tract, optic radiations and tract,
uncinate fascicle, inferior longitudinal fasciculus (ILF), and superior longitudinal
fascicle (SLF) were made. DTI Studio was used to assess the images and eigenvector
with maps of fractional anisotropy (FA). Direction of diffusion was described by FA.
A higher FA in a voxel suggested movement, mostly in one direction/axis, and low FA
indicated movement occurring in all three directions/axes. Different but specific
color maps with standard coding were generated using their combination, with red (x , left–right), blue (z , craniocaudal), and green (y , anteroposterior), using intensity scale proportional to their FA. Formed images
were thoroughly evaluated and then the tracts were labeled as displaced, infiltrated,
and disrupted.
● Displaced tracts are defined as tracts showing normal or only slightly decreased FA, resulting from
bulk mass displacement ([Figs. 1 ], [2 ]).
● Infiltrated tracts are defined as tracts showing reduced anisotropy but that remained identifiable
on color maps ([Figs. 3 ], [4 ]).
● A disrupted tract is defined as a tract showing isotropic (or near-isotropic) diffusion, such
that it could not be identified on directional color maps ([Figs. 5 ], [6 ]).
Fig. 1 Sagittal diffusion tensor imaging (DTI) of supratentorial intra-axial (ST-IA) brain
tumor with displaced fascicles.
Fig. 2 Axial diffusion tensor imaging (DTI) of supratentorial intra-axial (ST-IA) brain
tumor with displaced fascicles.
Fig. 3 Axial diffusion tensor imaging (DTI) of supratentorial intra-axial (ST-IA) brain
tumor with infiltrated fascicles.
Fig. 4 Sagittal magnetic resonance (MR) images of supratentorial intra-axial (ST-IA) brain
tumor with infiltrated fascicles.
Fig. 5 Axial diffusion tensor imaging (DTI) images of supratentorial intra-axial (ST-IA)
brain tumor with.
Fig. 6 Sagittal magnetic resonance (MR) images of supratentorial intra-axial (ST-IA) brain
tumor with disrupted fascicles.
Tumor volume analysis: In low-grade gliomas, FLAIR sequences of MRI were used to assess the preoperative
and postoperative tumor volumes, while in high-grade gliomas, postcontrast T1 images
were assessed for the same using OSIRIX software. Small and large tumors were classified
as preoperative tumor volume less than 100 cm3 and greater than 100 cm3 , respectively. The amount of tumor resected was assessed on postoperative MRI done
at 3 weeks and classified as per the Berger classification[6 ]
[7 ] with surgical outcome defined as total excision with postoperative volume less than
1 cm3 ,[5 ] subtotal excision with residual volume of 1 to 10 cm3 , and partial removal/excision with residual volume greater than 10 cm3 .
Statistical analysis: Computer software SPSS was used for statistical analysis with percentages and proportions
used to express categorical data with standard deviations and mean for assessable
data as χ
2 test was used to analyze the difference in proportion. For tests, the level of significance
was determined as 95% (p value <0.05).
Results
Patient Population
Of the 91 patients included in the study, 61 (67%) were males and 30 (32.9%) were
females. The mean age was 43 years (range: 15–81 years). Power weakness was present
in 30 (32.9%) patients, speech/language abnormality in 04 (4.4%) patients, headache
in 61 (67.03%) patients, behavior charges in 14 (15.38%) patients, seizure in 41 (45.05%)
patients, and memory impairment in 6 (6.59%) patients, with 16 patients (17.5%) presenting
with cognitive impairment ([Table 1 ]).
Table 1
Signs and symptoms preoperatively in our study
Preoperative neurologic examination
Total no. of patients
%
Seizure
41
45.05
Motor weakness
30
32.96
Headache
61
67.03
Behavior charges
14
15.38
Memory impairment
6
6.59
Sensory impairment
2
02.19
Language abnormality
4
04.39
Cognitive defect
16
17.58
In 44 patients (48.35%), tumors were present in the left cerebral hemisphere, while
in the remaining 47 (51.64%) patients, tumors were present in right cerebral hemisphere.
A frontal lobe pathology was seen in 34 (37%) patients, parietal lobe was involved
in 11 (12.08%) patients, and temporal lobe was involved in 12 (13.18%) patients. Involvement
of more than one lobe was observed in 26 (28.57%) patients ([Fig. 7 ]).
Fig. 7 Pie chart showing different sites of tumors in our study.
DTI showed displaced fascicles in 39 (42.8%) patients, infiltrated fascicles in 33
(36.2%) patients, and disrupted fascicles in 19 (20.8%) patients ([Table 2 ]).
Table 2
Different patterns of tracts in DTI in this study
Diffusion tensor imaging
Total no. of patients
%
Displaced
39
42.85
Infiltrated
33
36.26
Disrupted
19
20.87
Abbreviation: DTI, diffusion tensor imaging.
On histopathological analysis, LGG was present in 60 patients, HGG in 25 patients,
and metastatic lesions in 6 patients ([Table 3 ]). Oligodendroglioma was present in 19 (20.87%) patients, glioblastoma (GBM) in 14
(15.38%) patients, neurocytoma in 3 (3.29%) patients, astrocytoma in 37 (40.65%) patients,
oligoastrocytoma in 12 (13.18%) patients, and adenocarcinoma in 4 (4.34%) patients
([Fig. 8 ]).
Fig. 8 Pie chart showing different histological outcomes in our study. GBM, glioblastoma.
Table 3
Patients in different histology grades
Grade of tumor
Total no. of patients
%
Low-grade glioma
60
65.9
High-grade glioma
25
27.47
Metastasis
6
5.49
Surgical Outcome
Seventy-three (80.21%) patients had a tumor volume of less than 100 cm3 and 17 (18.68%) had a tumor volume greater than 100 cm3 preoperatively ([Table 4 ]). Only 54 of 91 patients came for follow-up and their post-op MRI was done at 3
weeks to calculate the residual tumor volume. Due to the COVID-19 pandemic, our institution
was turned into a dedicated COVID-19 treatment center, so many of the patients went
to other regional hospitals for further chemoradiotherapy and were lost to follow-up.
Table 4
Relationship of different tumor volumes in patients
Tumor volume, operative (cm3)
Total no. of patients
%
< 100
73
80.21
> 100
17
18.68
In tumors less than 100 cm3 , displaced tracts were seen in 19 patients preoperatively. In these patients, a postoperative
tumor volume of less than 1 cm3 was seen in 12 patients (63.15%), while 5 patients had a tumor volume of 1 to 10 cm3 (26.3%) and 2 patients had a tumor volume of greater than 10 cm3 (10.5%) postoperatively. Infiltrated fascicles were present in nine patients, with
one patient (11.1%) with a tumor volume of less than 1 cm3 postoperatively, six (16.6%) patients with a volume of 1 to 10 cm3 , and two patients (22.2%) with tumor volume greater than 10 cm3 . Disrupted tracts were seen in 10 patients in the same category of patients, with
postoperative tumor volume less than 1 cm3 in one (10%) patient, 1 to 10 cm3 in three (30%) patients, and greater than 10 cm3 in 6 (60%) patients ([Table 5 ]). Statistical analysis showed significant results on χ
2 test between the fascicle type and tumor residual volumes postoperatively, thus confirming
that surgical resection is complete/maximum in displaced fascicles and subtotal or
partial resection done maximally in infiltrated/disrupted tracts ([Table 5 ]).
Table 5
Correlation between postoperative tumor volume and fiber type in < 100 cm3 pre-op tumor volume category
Type of fiber on DTI
Tumor volume of < 100 cm3 preoperatively
N
< 1 cm3 (n = 14)
1–10 cm3 (n = 14)
> 10 cm3 (n = 10)
p -Value
N
%
N
%
N
%
Intact/displaced fascicles
19
12
85.71
5
35.7
2
20
< 0.05
Infiltrated fascicles
9
1
7.14
6
42.85
2
20
0.14
Disrupted fascicles
10
1
7.14
3
21.42
6
60
0.04
Total
38
14
100
14
100
10
100
Abbreviation: DTI, diffusion tensor imaging.
Seven patients with a tumor volume greater than 100 cm3 had displaced/intact fascicles. Among these patients, postoperatively, one (14.2%)
patient had a tumor volume of less than 1 cm3 , two (28.57%) patients had a tumor volume of 1 to 10 cm3 , and four (57.14%) patients had a tumor volume of greater than 10 cm3 . Infiltrated fascicles were seen in six patients with a tumor volume greater than
100 cm3 and postoperatively no patient had tumor volume less than 1 cm3 , two (33.33%) patients had a tumor volume of 1 to10 cm3 , and four (66.67%) patients had a tumor volume greater than 10 cm3 . Disrupted fascicles were present in three patients and postoperatively a tumor volume
less than 1 cm3 was seen in none of the patients, tumor volume of 1 to 10 cm3 was seen in one (33.33%) patient, and tumor volume greater than 10 cm3 was observed in two (66.67%) patients. Statistical analysis showed no significant
difference in results with regard to fascicle involvement in the postoperative residual
tumor volume in the greater than 100 cm3 category, suggesting that maximum patients with large tumors had subtotal or partial
resections ([Table 6 ]).
Table 6
Correlation between postoperative tumor volume and fiber type in >100 cm3 pre-op tumor volume category
Type of fiber on DTI
Tumor volume of > 100 cm3 preoperatively
N
< 1 cm 3 (n = 14)
≤ 10 cm3 (n = 14)
> 10 cm3 (n = 10)
p -Value
N
%
N
%
N
%
Intact/displaced fascicles
7
1
14.2
2
28.57
4
57.14
0.0008
Infiltrated fascicles
6
0
0
2
33.33
4
66.67
0.8
Disrupted fascicles
3
0
0
1
33.33
2
66.67
0.89
Total
16
1
5
10
Abbreviation: DTI, diffusion tensor imaging.
Histopathology was correlated with fascicle type in DTI in all patients. LGGs were
seen in 35 (38.46%) patients, HGGs in 50 (54.94%) patients, and metastatic lesions
in 6 (6.5%) patients. The pattern of fascicles in different histopathologic findings
was studied. In the displaced fascicle category, 30 patients (76.92%) had LGGs, 7
patients (9.3%) had HGGs, and 2 patients (5.1%) had metastatic lesions. In infiltrated
and disrupted fascicles, LGGs were seen in 5 (9.6%) patients, HGGs in 43 (82.69%)
patients, and metastatic lesions in 4 (7.6%) patients. The relationship between the
fascicle type on DTI and histopathology was statistically significant on χ
2 test. Thus, displaced fascicles were maximally seen in LGGs, while infiltrated or
disrupted fascicles were mostly seen in HGGs and metastatic lesions ([Table 8 ], [Fig. 9 ]).
Improvement in the postoperative neurological status was found in 34 of 39 (86%) patients
with displaced WM tracts, while in the remaining 5 patients, no improvement was seen.
Also, improvement in the neurological status was seen in only 25 (48%) cases with
infiltrated and disrupted pattern of WM tracts, and in other 27 cases, there was no
improvement postoperatively ([Table 7 ]).
Table 7
Relationship between type of fascicle in DTI with postoperative neurological status
Neurological status post-op
Displaced fascicles
%
Infiltrated/disrupted fascicles
%
N
= 39
N
(33 + 19)
Improved
34
86
25
48
Same/deteriorate
5
14
27
52
Total
39
100
52
100
Abbreviation: DTI, diffusion tensor imaging.
Note: p -value = 0.000111.
Chi-square value: 14.9.
Discussion
Neuro-oncological surgeries demand a balance between maximal tumor resection and preservation
of neurological functions. The maximal tumoral resection with radiotherapy or chemotherapy
appears to be more effective and also diminishes chances of early recurrence.[8 ]
[9 ]
[10 ]
[11 ]
[12 ] Safeguarding of the somatomotor, language, visual, and other eloquent cortical areas
during resection improves the outcome of patients.[6 ]
[7 ] In DTI, we can study the different patterns of WM tract involvement by the tumors
as being intact/displaced, infiltrated, or damaged.[13 ]
Our study is unique in depicting the role of preoperative DTI in supratentorial tumors
and presents a new tool to predict the tumor histology and help in surgical planning
and resection while also serving as the prognostication indicator.
We studied 91 patients, comprising 61 males (67%) and 30 females (32%), with age ranging
from 16 to 82 years. The most common location of tumors was found in the frontal lobes
(37%), followed by the temporal (13%) and parietal regions (12%), and more than one
lobe was seen involved in 29% of cases ([Table 1 ]). The most common histopathology seen are astrocytomas (low and high grades), followed
by oligodendroglioma, GBM, and oligoastrocytoma ([Fig. 8 ]).
In our study, as described in previous studies, the WM tracts were classified as intact/displaced,
infiltrated, and disrupted on the basis of direction of strands and their FA values.
On the basis of the pattern of tracts in relation to the tumor, a safest surgical
approach was selected in each patient individually. In tumors with a preoperative
size less than 100 cm3 , it was found that maximal resection was possible in tumors with displaced tracts
(89% of cases) and subtotal resections were done mostly in patients with infiltrated/disrupted
(56% of cases) tracts showing that in comparison to HGGs a plane for surgical resection
between tumor and parenchyma was easily available in LGGs, leading to their maximal
safe resection, and these results were also statistically significant with p -value less than 0.05 ([Table 5 ]).
Similar findings were noticed in studies done by Castellano et al[14 ] and Khan et al[15 ] where DTI predicted the extent of resection with maximal resections possible in
tumors with displaced/intact WM tracts, while infiltrating tumors were less likely
to be resected completely.
In our study, we also found that the size of the tumor is also an independent factor
for the extent of resection as in tumor with size greater than 100 cm3 , 60% of patients had subtotal resection irrespective of the pattern of tracts in
pre-op DTI. Our study did not use neuromonitoring, awake craniotomy, functional MRI,
intraoperative MRI, and ultrasound. Thus, due to risks involved in extensive resections
to adjacent eloquent areas, subtotal resections must have been preferred in these
cases. Statistical analysis of the relation of the pattern of the WM tracts with the
extent of resection showed no significant results in tumors greater than 100 cm3 with p value greater than 0.05. Studies conducted by Khan et al also showed similar results[14 ] ([Table 6 ]).
The histopathological analysis after resection of tumors showed that displacement
of the WM tracts was characteristically seen in LGGs (30 of 39 patients, i.e., 79%),
while infiltrated/damaged fibers tracts were seen HGGs (43 of 52 patients, i.e., 82.69%).
These results were statistically significant with p value less than 0.05, suggesting that LGGs mostly displace the WM tracts, while infiltrated/disrupted
fibers are mostly seen in HGGs. On the basis of the pattern of the WM tracts, we can
predict ([Table 8 ]) the histological characterization of tumors, which can help in prognosticating
the patients before surgery. In our study, as maximal resection of tumors without
damaging the WM tracts was possible mostly in LGGs, while subtotal resections were
done mostly in tumors with HGGs, a similar pattern is also seen in previously conducted
studies[5 ]
[16 ]
[17 ] ([Fig. 9 ]).
Table 8
Correlation between various histopathologies and fiber type in DTI preoperatively
in ST-IA brain tumors
Fiber type
Histopathological correlation with fiber type in DTI
Histopathology
Total
Displaced
Infiltrated/disrupted
p -Value
N
N
%
N
%
Low-grade glioma
35
30
76.92
5
9.6
0.0006
High-grade glioma
50
7
9.3
43
82.69
0.004
Metastasis
06
2
5.1
4
7.6
0.6
Total
91
39
100
52
100
Abbreviations: DTI, diffusion tensor imaging; ST-IA, supratentorial intra-axial.
Fig. 9 Bar graph showing the correlation between various histopathologies and fiber type
in diffusion tensor imaging (DTI) preoperatively in supratentorial intra-axial (ST-IA)
brain tumors.
Improvement in postoperative neurological status was found in 34 of 39 (86%) patients
with displaced WM tracts and in only 25 (48%) cases with infiltrated and disrupted
pattern of WM tracts. This study showed that displaced WM tracts, mostly seen in ([Table 7 ]) LGGs, had improvement in neurological status postoperatively, due to the good plane
between tumor and adjacent parenchyma, which attributes to preserved anisotropy of
fibers. On the other hand, infiltrated/disrupted tracts in high-grade tumors had no
clear margin, which could be ascertained, leading to their subtotal resections with
minimal improvement in neurological status postoperatively.
Hence, on the basis of the pattern of tracts in the preoperative DTI, one can predict
the histopathology (either low or high grade), extent of resections, and neurological
outcome in patients; hence, we can prognosticate about these patients as well.
DTI also has some drawbacks. In cases of significant peritumoral edema, it is difficult
to differentiate whether this change or decrease in anisotropy is due to edema or
infiltration due to neoplasm, which is a tricky situation to asses in HGGs. Also,
there is significant shift of brain tissue during surgical maneuver, so preoperative
DTI may not tell us the exact location of fibers in relation to the tumor while operating
due to a shift seen in tracts during surgery.[17 ]
[18 ] These two reasons can be the significant causes for subtotal resection, which were
done in big tumors with disrupted fibers on DTI. Also, DTI adds up in cost of treatment,
which is a major drawback.
Conclusion
Our study has shown that preoperative DTI provides the neurosurgeon with a safe surgical
corridor preserving the WM tracts, thus improving both surgical and neurological outcome
in patients. The displaced fiber pattern is mostly seen in LGGs, while the infiltrated
and disrupted patterns of fascicles are seen in HGGs, revealing the prognostic value
of DTI in brain tumors. The pattern of fascicle involvement also correlates with the
extent of tumor resection, as in intact fascicles, the extent of resection was more
when compared with the infiltrated/disrupted fascicles. Thus DTI acts as a pre-operative
prognostication indicator in relation to the tumour before the actual histologic grading.
It also helps neurosurgeon to decide upon a safe surgical corridor while proceeding
towards gross total resection in displaced fibres and near total resections in infiltrated
fibres.
