Cerebral blood flow in the ventromedial prefrontal cortex correlates with treatment response to low-frequency right prefrontal repetitive transcranial magnetic stimulation in the treatment of depression

Authors


Shinsuke Kito, MD, PhD, Department of Neuropsychiatry, Kyorin University School of Medicine, 6-20-2 Shinkawa, Mitaka, Tokyo 181-8611, Japan. Email: kito@kk.iij4u.or.jp.

Abstract

Aims:  Low-frequency right prefrontal repetitive transcranial magnetic stimulation (rTMS) is effective in treating depression, and its antidepressant effects have proven to correlate with decreases in cerebral blood flow (CBF) in the orbitofrontal cortex and subgenual cingulate cortex. However, a predictor of treatment response to low-frequency right prefrontal rTMS in depression has not been identified yet. The aim of this study was to estimate regional CBF in the frontal regions and investigate the correlation with treatment response to low-frequency right prefrontal rTMS in depression.

Methods:  We examined 26 depressed patients for the correlation between treatment response to rTMS and regional CBF in the frontal regions, by analyzing their brain scans with 99mTc-ethyl cysteinate dimer before rTMS treatment. CBF in 16 brain regions was estimated using fully automated region of interest analysis software. Two principal components were extracted from CBF in 16 brain regions by factor analysis with maximum likelihood method and Promax rotation with Kaiser normalization.

Results:  Sixteen brain regions were divided into two groups: dorsolateral prefrontal cortex (superior frontal, medial frontal, middle frontal, and inferior frontal regions) and ventromedial prefrontal cortex (anterior cingulate, subcallosal, orbital, and rectal regions). Treatment response to rTMS was not correlated with CBF in the dorsolateral prefrontal cortex, but it was correlated with CBF in the ventromedial prefrontal cortex.

Conclusion:  These findings suggest that CBF in the ventromedial prefrontal cortex may be a potential predictor of low-frequency right prefrontal rTMS, and depressed patients with increased CBF in the ventromedial prefrontal cortex may show a better response.

SEVERAL RECENT STUDIES provide strong evidence that high-frequency left prefrontal repetitive transcranial magnetic stimulation (rTMS) is effective in the treatment of depression.1–4 Similarly, low-frequency rTMS to the right dorsolateral prefrontal cortex (DLPFC) has also proven to have antidepressant effects.5–10 Low-frequency rTMS has a considerably lower risk of seizure induction and may be more acceptable than high-frequency rTMS in this light.11,12

Brain imaging studies have revealed that low-frequency right prefrontal rTMS decreases cerebral blood flow (CBF) in several brain regions and improves depression,13–15 and its antidepressant effects correlate with decreases in CBF in the orbitofrontal cortex and subgenual cingulate cortex.15 However, a predictor of treatment response to low-frequency right prefrontal rTMS in depression has not been identified yet. In spite of some shortcomings, 99mTc-ethyl cysteinate dimer is in widespread clinical use, and it is generally easy and less expensive to use relative to radioisotopes in PET. Moreover, 99mTc-ethyl cysteinate dimer is useful in estimating CBF and calculating the correlation of the response to rTMS, making it a potential predictor of response to rTMS.

In this study, we analyzed brain images with 99mTc-ethyl cysteinate dimer obtained from patients with depression enrolled in our previous study,15 using fully automated region of interest analysis software, 3DSRT and FineSRT, which were developed by Takeuchi et al.16–18 The aim of this study was to estimate regional CBF in the frontal regions, which had been found to be involved in the antidepressant effects of rTMS,15,19 and investigate the correlation with treatment response to low-frequency rTMS of the right DLPFC in depression.

METHODS

Subjects

A complete definition of the inclusion and exclusion criteria for participating patients was described in our previous study.15 Twenty-six depressed patients (14 men, 12 women) were examined for the correlation between treatment response to rTMS and regional CBF in the frontal regions, by analyzing their brain scans before the rTMS treatment. All patients met the DSM-IV-TR criteria for major depressive disorder. Some of the exclusion criteria were as follows: bipolar disorders, neurological disorders, convulsive disorders, significant medical diseases, a history of substance abuse or dependence, and active suicidal ideation. The mean patient age was 46.2 ± 13.8 years, and the age at onset of depression was 39.0 ± 13.3 years. The number of previous depressive episodes was 3.1 ± 1.1, and the duration of current depressive episode was 11.4 ± 6.4 months. The mean Hamilton Depression Rating Scale (HDRS)20 score at baseline was 22.7 ± 3.8. Medical treatments administered were not allowed to have changed in the 4 weeks before the start of the first rTMS session or during the trial. Twenty-four patients were taking medications during the trial, and some patients were taking several antidepressants: nine were taking selective serotonin re-uptake inhibitors (fluvoxamine, paroxetine, sertraline), eight were taking serotonin norepinephrine re-uptake inhibitors (milnacipran), eight were taking tricyclics (amitriptyline, clomipramine, amoxapine), seven were taking other antidepressants (trazodone, mianserin, sulpiride), five were taking lithium carbonate, one was taking an antipsychotic (olanzapine), and sixteen were taking benzodiazepines.

After receiving a full explanation of the procedures, all patients provided written informed consent to participate in the study, and the study was approved by the ethics committee of Kyorin University School of Medicine (Tokyo, Japan).

Transcranial magnetic stimulation

rTMS was administered using a Magstim Super Rapid (Magstim, Whitland, UK) with stand-held, 70-mm figure-eight coils. Stimulation intensity was set at 100% of resting motor threshold, and the stimulation site of the right DLPFC was defined by a point 5 cm anterior in a parasagittal line to the motor threshold location. Five 60-s trains at 1 Hz and at 100% of the resting motor threshold were applied in each session with a 60-s interval between the trains. Twelve treatment sessions were administered during 3 weeks (total pulses: 3600).

Brain scan and image analysis

Brain scans obtained from 26 patients using 99mTc-ethyl cysteinate dimer within 24–72 h before the first rTMS treatment were used to investigate the correlation with treatment response to rTMS. All patients rested in the supine position in a quiet room with their eyes closed and without earplugs. The image acquisition started approximately 10 min after the injection of 600-MBq 99mTc-ethyl cysteinate dimer, while patients were in a resting state, via a venous cannula previously inserted into the right arm. Images were obtained using a triple-detector gamma camera GCA-9300A/HG (Toshiba, Tokyo, Japan) with low-energy super-high-resolution fan beam collimators. The matrix size was 128 × 128, and data were collected in 30 frames at 4° steps over 120° with a pixel width of 1.72 mm and a slice thickness of 3.45 mm. Scanned data were prefiltered using a Butterworth filter (order 8 and a cut-off at 0.08–0.09 cycles/pixel) and reconstructed with a ramp filter. Scatter and attenuation corrections were performed using the triple-energy window correction and Sorenson methods, respectively.

To estimate regional CBF, the images were analyzed by using fully automated region of interest analysis software, 3DSRT and FineSRT,16–18 which constructed a 3-D stereotactic region of interest template on the anatomically standardized brain images. The FineSRT is composed of 104 brain regions in both hemispheres (Supplemental Table S1). In this study, 16 brain regions, which correlated with improvement of depression by rTMS treatment in our previous study,15,19 were selected from 104 brain regions. The selected brain regions were the superior frontal, medial frontal, anterior cingulate, subcallosal, orbital, rectal, middle frontal, and inferior frontal (Fig. 1).

Figure 1.

Sixteen brain images (8 × 2 ROI) with 99mTc-ethyl cysteinate dimer and region of interest (ROI). aC, anterior cingulate region; iF, inferior frontal region; L, left side; meF, medial frontal region; miF, middle frontal region; Or, orbital region; R, right side; Re, rectal region; Sc, subcallosal region; sF, superior frontal region.

Clinical assessment and statistical analysis

All patients were assessed at baseline, week 3, and week 5 (2 weeks after the final rTMS treatment session) with the HDRS. Treatment response to transcranial magnetic stimulation was estimated from the scores on the HDRS at baseline and at week 5 [decrease (%) = (score at baseline − score at week 5) / score at baseline].

In order to identify a predictor of treatment response to rTMS, Pearson's correlation coefficient was used to investigate the correlation between treatment response to rTMS and CBF in 16 brain regions, and subsequently, factor analysis of CBF in 16 brain regions was performed by maximum likelihood method and Promax rotation with Kaiser normalization. The results obtained were analyzed using Pearson's correlation coefficient to evaluate the correlation with treatment response to rTMS. Statistical analysis was conducted using SPSS for Windows 14.0 (SPSS Inc, Chicago, Illinois, USA), with the level of statistical significance set at P < 0.05.

RESULTS

The mean HDRS score decreased from 22.7 (SD = 3.8) at baseline to 13.6 (SD = 4.6) at week 3, and to 11.9 (SD = 5.0) at week 5. Eleven of 26 patients (6 men, 5 women) were responders who had their HDRS score decrease by 50% from baseline to week 5. Fifteen of 26 patients (8 men, 7 women) were nonresponders. One of the nonresponders did not complete the entire study because she withdrew her consent for study participation during the trial. There were no differences in CBF between responders and nonresponders in the superior frontal, medial frontal, anterior cingulate, middle frontal, and inferior frontal regions. However, the results showed differences in CBF between the subcallosal (left: t = 2.68, P = 0.01; right: t = 2.19, P = 0.04), orbital (left: t = 2.05, P = 0.05; right: t = 1.73, P = 0.10), and rectal regions (left: t = 2.34, P = 0.03; right: t = 1.90, P = 0.07).

The correlation between CBF in 16 brain regions in 26 patients and treatment response to rTMS is shown in Table 1. Based on Pearson's correlation coefficient, the treatment response to rTMS was correlated with CBF in the left subcallosal (r = 0.39, P = 0.05), left orbital (r = 0.42, P = 0.04), bilateral rectal regions (left: r = 0.50, P = 0.01; right: r = 0.45, P = 0.02) (Table 1). In addition, CBF in 16 brain regions was found to be inversely correlated with the age of patients, and we calculated a partial correlation between treatment response to rTMS and CBF in those 16 brain regions, with the age of patients as a control valuable. Treatment response to rTMS was correlated with CBF in the subcallosal (left: r = 0.42, P = 0.04; right: r = 0.42, P = 0.04), orbital (left: r = 0.47, P = 0.02; right; r = 0.40, P = 0.05), and rectal regions (left: r = 0.56, P = 0.01; right: r = 0.49, P = 0.02) in both hemispheres (Table 1).

Table 1.  Cerebral blood flow in 16 brain regions in 26 depressed patients and the correlation with treatment response to low-frequency right prefrontal rTMS
Brain regionHemisphereCerebral blood flowCorrelation with treatment responsePartial correlation with treatment response
MeanSDrP-valuerP-value
  1. The correlation between treatment response to repetitive transcranial magnetic stimulation (rTMS ) and cerebral blood flow (mL/100 g/min) in 16 brain regions was calculated with Pearson's correlation coefficient, and partial correlation was calculated with the age of patients as a control valuable. Treatment response to rTMS was estimated from the scores on the Hamilton Depression Rating Scale at baseline and at week 5 [decrease (%) = (score at baseline − score at week 5) / score at baseline].

Superior frontalLeft36.74.80.210.320.190.37
Right36.34.30.300.150.300.16
Medial frontalLeft43.25.30.170.420.150.48
Right42.95.20.220.300.210.33
Anterior cingulateLeft37.55.50.320.120.350.09
Right38.06.20.340.090.380.06
SubcallosalLeft44.27.10.390.050.420.04
Right43.16.00.370.070.420.04
OrbitalLeft39.96.20.420.040.470.02
Right39.36.10.360.080.400.05
RectalLeft39.36.70.500.010.560.01
Right38.36.20.450.020.490.02
Middle frontalLeft39.74.90.290.170.290.17
Right39.94.70.200.330.190.38
Inferior frontalLeft39.85.20.340.100.370.07
Right40.34.70.290.150.320.13

Two principal components were extracted from CBF in 16 brain regions by factor analysis with maximum likelihood method and Promax rotation with Kaiser normalization (Table 2). These 16 brain regions were divided into two groups: DLPFC (superior frontal, medial frontal, middle frontal, and inferior frontal regions) and ventromedial prefrontal cortex (VMPFC) (anterior cingulate, subcallosal, orbital, and rectal regions). The mean CBF in the DLPFC and VMPFC is shown in Table 3. Additionally, we calculated the CBF ratio of the DLPFC to the VMPFC (DLPFC / VMPFC CBF ratio). There was no correlation between treatment response to rTMS and CBF in the DLPFC (Table 3). Treatment response to rTMS was correlated with CBF in the VMPFC (r = 0.42, P = 0.04) (Table 3, Fig. 2), and partial correlation with the age of patients as a control valuable showed similar results (r = 0.49, P = 0.016) (Table 3). There was no correlation between treatment response to rTMS and DLPFC / VMPFC CBF ratio (Table 3).

Table 2.  Factor analysis of cerebral blood flow in 16 brain regions in 26 depressed patients
FactorBrain regionHemisphereFactor loading
  1. Two principal components were extracted from cerebral blood flow in 16 brain regions. The extraction method was maximum likelihood method (initial eigenvalues > 1), and the rotation method was Promax with Kaiser normalization (rotation converged in three iterations). Sixteen brain regions were divided into two groups: dorsolateral prefrontal cortex (superior frontal, medial frontal, middle frontal, and inferior frontal regions) and ventromedial prefrontal cortex (anterior cingulate, subcallosal, orbital, and rectal regions).

Factor 1Dorsolateral prefrontal cortex  
 Superior frontalLeft0.97
Right0.92
 Medial frontalLeft0.89
Right0.83
 Middle frontalLeft0.86
Right0.89
 Inferior frontalLeft0.77
Right0.78
Factor 2Ventromedial prefrontal cortex  
 Anterior cingulateLeft0.76
Right0.81
 SubcallosalLeft0.82
Right0.84
 OrbitalLeft0.86
Right0.84
 RectalLeft0.87
Right0.83
Table 3.  CBF in the DLPFC and VMPFC, DLPFC/VMPFC CBF ratio, and the correlation with treatment response to low-frequency right prefrontal rTMS
 Cerebral blood flowCorrelation with treatment responsePartial correlation with treatment response
MeanSDrP-valuerP-value
  1. The correlation between treatment response to repetitive transcranial magnetic stimulation (rTMS) and cerebral blood flow (CBF) (mL/100 g/min) in the DFPLC and VMPFC was calculated by using Pearson's correlation coefficient, and partial correlation was calculated with the age of patients as a control valuable. Treatment response to rTMS was correlated with CBF in the VMPFC. Partial correlation with the age of patients provided similar results more significantly than Pearson's correlation coefficient did. There were no correlations between treatment response to rTMS and DLPFC/VMPFC CBF ratio.

Dorsolateral prefrontal cortex (DLPFC)39.84.70.260.210.260.22
Ventromedial prefrontal cortex (VMPFC)40.05.90.420.040.490.016
DLPFC/VMPFC CBF ratio1.0050.097−0.340.095−0.340.11
Figure 2.

Cerebral blood flow in the ventromedial prefrontal cortex (VMPFC) and percent decrease in Hamilton Depression Rating Scale (HDRS) score. Cerebral blood flow in the VMPFC was positively correlated with a decrease in HDRS score (r = 0.42, P = 0.04). This correlation suggests that depressed patients with increased cerebral blood flow in the VMPFC may show a better response to low-frequency right prefrontal repetitive transcranial magnetic stimulation.

DISCUSSION

The findings of this study suggest that CBF in the VMPFC may be a potential predictor of treatment response to low-frequency right prefrontal rTMS in depression, and depressed patients with increased CBF in the VMPFC, particularly the subcallosal, orbital, and rectal regions, may respond better to low-frequency right prefrontal rTMS.

Low-frequency right prefrontal rTMS has been shown to decrease CBF in several brain regions, particularly in the DLPFC, orbitofrontal, and subgenual cingulate regions,14,15 and its antidepressant effects are correlated with decreases in CBF in the orbitofrontal and subgenual cingulate cortices.15 In this study, we sought to identify a predictor of treatment response to low-frequency rTMS in depression by estimating regional CBF with 99mTc-ethyl cysteinate dimer SPECT and investigating the correlation with the treatment response. For factor analysis of CBF in 16 brain regions, we divided them into two groups: DLPFC and VMPFC. Superior frontal, medial frontal, middle frontal, and inferior frontal regions were included in the DLPFC. Anterior cingulate, subcallosal, orbital, and rectal regions were included the VMPFC. Treatment response to rTMS was not correlated with CBF in the DLPFC, but it was correlated with CBF in the VMPFC. Recent reviews of brain imaging studies in mood disorders have revealed that abnormally increased activity in the orbitofrontal cortex, subgenual cingulate cortex, and limbic regions, such as the amygdala and subgenual anterior cingulate, are involved in the pathophysiology of mediating depressive symptoms.21–23 Deep brain stimulation to the subgenual cingulate cortex ameliorates clinical features of depression with decreases in CBF in the orbitofrontal cortex and the subgenual cingulate cortex.22,24,25 Neurosurgery for treatment-resistant mood disorder has long targeted the ventromedial prefrontal region,21 and ventromedial prefrontal overactivity was reported to predict response to a neurosurgical procedure for treatment-resistant mood disorder.26 Earlier imaging studies in depression have provided evidence of decreased activity in the DLPFC and increased activity in the VMPFC (emotion-related regions),21–23 showing that a DLPFC/VMPFC CBF ratio may tend to be lower in depression. In addition to absolute CBF in several brain regions, a DLPFC/VMPFC CBF ratio may be useful as a potential predictor of the response. When we investigated the correlation of this ratio and the response to rTMS, the results showed that the treatment response was correlated with CBF in the VMPFC, but not with CBF in the DLPFC or a DLPFC/VMPFC CBF ratio. This suggests that depressed patients with increased CBF in the VMPFC may respond better to low-frequency rTMS, regardless of CBF in the DLPFC, and absolute CBF in the VMPFC can be a potential predictor of response to low-frequency rTMS. Meanwhile, we found that the response rate of high-frequency left frontal rTMS was inversely correlated with a DLPFC/VMPFC CBF ratio, but not with CBF in the DLPFC or VMPFC (unpublished data). This suggests that depressed patients with a lower DLPFC/VMPFC CBF ratio may respond better to high-frequency rTMS and that the DLPFC/VMPFC CBF ratio could be a predictor of response to high-frequency rTMS, rather than of absolute CBF in the DLPFC or VMPFC. However, this needs further studies and clinical validation.

In conclusion, the findings of this study suggest that CBF in the VMPFC such as the subcallosal, orbital, and rectal regions may be a potential predictor of treatment response to low-frequency right prefrontal rTMS, and depressed patients with increased CBF in the VMPFC may show a better response, although this study was done with a small sample of 26 depressed patients in an open label trial.

Ancillary