Towards characterizing the regional cerebral perfusion in evaluating the severity of major depression disorder with SPECT/CT

Major depressive disorder (MDD) is a common mental disorder worldwide. It is characterized by a persistent low mood and a loss of interest or pleasure [1]. MDD is also a disabling condition that adversely affects the sufferers’ whole lives. Moreover, MDD can lead to suicide. Over 800,000 people die by suicide every year, and most of them are suffering from MDD [2]. However, fewer than 10% of MDD sufferers have ever received effective treatments, and one of the most significant barriers is lack of effective assessment and management of MDD [2].

Today, the most widely used criterion for assessing MDD is the fourth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) published by American Psychiatric Association, and the 10th revision of International Statistical Classification of Diseases and Related Health Problems (ICD-10) list by the World Health Organization. DSM-IV is more popular among mental health professionals [3]. The Hamilton Rating Scale for Depression (HAMD) is a multiple-item scale for measuring the severity of depression. However, it has been criticized for its subjectivity [4]. There is a demand for objective assessment of MDD, which is non- or minimal-invasive, convenient and accurate in a clinical setting.

Since the late 1970s, functional neuroimaging, including single photon emission computed tomography (SPECT), positron emission tomography (PET), magnetic resonance imaging (MRI) and fusing imaging (SPECT/CT, PET/CT or PET/MRI) has been extensively used in MDD [5‐9]. Regional cerebral blood flow (rCBF) SPECT is a nuclear medicine imaging method in routine clinical practice, and is also one of the earliest functional neuroimaging studies in MDD [10]. According to previous research, there is almost complete agreement that blood flow in the frontal lobe is reduced in MDD patients [9, 11‐14], but the controversies still exist in the extent of the rCBF involvement [15‐17]. It has been speculated that the various patterns of rCBF abnormalities in MDD patients may be related to the various clinical presentation of MDD [15‐17].

In this study, the rCBF status of MDD patients was characterized with technetium-99 m ethylcysteinate dimer ([^99mTc]ECD) SPECT/CT. The localization of abnormal rCBF was represented by Brodmann areas. In addition, we explore the relationship in MDD patients between the imaging features, clinical presentation and the severity of the symptoms. It is expected that the conventional nuclear medicine imaging could help add an objective image assessment model of MDD in current psychiatric practice that relies mainly on self-reporting of symptoms and clinical interviews.

The mean HAMD score and the mean HAMA score in the MDD patients were 25.49 ± 6.00 and 23.12 ± 5.83, respectively, and as for healthy controls, the mean scores were 2.70 ± 1.42 and 0.65 ± 0.88, respectively. There was a positive correlation between two scores for MDD patients (r = 0.732, p = 0.000) (Fig. 1).

It is still controversial whether there are correlations between the MDD severity and the patient characteristics. In this study, the HAMD score showed no correlations with the ages of MDD patients (r = 0.134, p = 0.255). The MDD women (26.66 ± 5.11) had higher HAMD scores than the MDD men (22.52 ± 7.08, t = 2.801, p = 0.007).

Compared with the normal controls data, 3D-SSP analysis showed that the MDD patients had decreased rCBF in frontal lobe (bilateral B4, B9, B10, B11, B47 and left B8 and right B6, B44, B46), parietal lobe (left B312), temporal lobe (right B21, B41, B42), cingulated cortex (bilateral B24, B32) and hypophysis. And it also showed in the MDD patients increased rCBF in parietal lobe (right B7), temporal lobe (right B28), occipital lobe (bilateral B17, B19 and left B18) and cingulated cortex (left B23 and right B31, B33). The detailed Z-scores were shown in Table 1. There was no statistically significant difference between MDD patients and normal controls for rCBF in hippocampus, caudate nucleus and globus pallidus (Table 2).

The HAMD score had a negative correlation with decreased rCBF in left ventral anterior cingulate cortex B24, a positive correlation with decreased CBF in left inferior prefrontal gyrus B47 and increased rCBF in right associative visual cortex B19 (R = 0.400, F = 4.453, p = 0.006). In addition, there was a negative correlation between the HAMA score and decreased rCBF in left subgenual cortex B25 (r = − 0.309, p = 0.007).

The differences of rCBF between MDD patients and normal controls were significant at the cluster lever (P_FWE-corr < 0.05) in the SPM and eZIP analysis. The results were listed in Table 3, and showed in Fig. 2 which was drawn from xjView toolbox and showed the decreased and increased rCBF foci in MDD patients.

According to 3D-SSP and eZIP analysis, the consistent results showed the decreased rCBF of MDD patients in frontal lobes (bilateral B11, B47 and right B4, B6, B10, B46), temporal lobe (right B21, B41, B42) and cingulated cortex (bilateral B24, B33), while their increased rCBF in occipital lobe (bilateral B17, B19 and left B18).

MDD is a heterogeneous disorder with various clinical presentation and inconsistent response to treatment. But its pathogenesis is inadequately understood [6, 18]. The mechanism that underlies the altered rCBF in MDD patients is not fully elucidated, however, previous studies [19, 20] suggested that the depressive symptoms were associated with reduced parasympathetic and increased sympathetic tone resulting in a broad range of hemodynamic changes, including elevated systemic vascular resistance. In addition, activation of immune-inflammatory pathways in MDD may also contribute to the process. Our study showed that more severe MDD patients would experience more anxiety. It has been suggested to be caused partly by brain chemistry neurotransmitter imbalances [21]. Antidepressant medication and behavioral therapy can help MDD patients overcome both depression and anxiety. However, clinicians have observed when depression occurs together with anxiety, symptoms are more severe than when depression or anxiety occurs alone, furthermore, it may be likely to develop treatment-resistant depression [22]. Additionally, our study also showed that the severity of depression had no correlations with age, and that women were found more likely than men to develop severe depression, the possible reasons may be as follows: women are more likely than men to show mood amplification; women have a greater propensity to be more passive than men; women may be exposed to more stressful life events than men [23].

To minimize the deviation errors caused by analysis software packages, 3D-SSP and eZIP image-processing programs were applied in the same CT attenuation-corrected SPECT images data of MDD patients and normal controls. According to previous studies [24, 25], the difference between 3D-SSP and eZIP lies in smoothing process after anatomic standardization. The smoothing process by eZIP is based on an SPM algorithm and effectively enhances the signal-noise ratio of images, which might overestimate cerebral hypoperfusion [24, 25]. So we apply corrections for multiple comparisons based on family-wise error (FEW), and P_FWE-corr < 0.05 is considered statistically significant in eZIP analysis. On the other hand, due to uninvolving such a smoothing process the preanalysis image reconstruction conditions might affect the results of 3D-SSP. Therefore, in 3D-SSP analysis all CT attenuation-corrected SPECT images data of MDD patients were compared with a normalized database and those of normal controls data simultaneously. Finally, the consistent results from 2 statistical image analysis methods were defined as the abnormal rCBF status of MDD patients. By 3D-SSP and eZIP analysis, we concluded that the decreased rCBF of MDD patients was in frontal lobes (bilateral B11, B47 and right B4, B6, B10, B46), temporal lobe (right B21, B41, B42) and cingulated cortex (bilateral B24, B33), while their increased rCBF was in occipital lobe (bilateral B17, B19 and left B18). Thus it can be seen that the rCBF abnormalities of MDD patients appear in bilateral cerebral hemisphere, and the hypoperfused regions showed right-hemisphere dominance while the area of hyperperfusion had a slight left-hemisphere dominance.

In the previous studies about brain structure [7, 26], metabolism [27] and perfusion [17], the medial prefrontal cortex (mPFC) and anterior cingulate cortex (ACC) in MDD patients have been found increased functional connectivity, which was related to one of possible pathophysiological mechanisms of depression——dysfunction of default mode network (DMN) [21, 28]. The hypoperfused areas of frontal lobes and cingulated cortex in MDD patients of this study largely coincided with what have been reported in the past [9, 16, 29, 30]. Although there was no consensus in regard to the hypoperfusion of temporal lobe and the hyperperfusion of occipital lobe, some researchers [9] have mentioned the abnormal perfusion of these cerebral regions in patients with treatment-resistant depression, and other researchers [31] suggested that abnormal activation of the occipital lobes may be an initiating factor cognitive disorder in depressed patients. We found there was no significant difference in the cerebellar blood flow of MDD patients as compared with that of normal controls, which was inconsistent with the findings of Gardner et al²⁸, in which cerebellum resulted to be heavily implicated in a large cohort of patients with persistent depressive disorder (PDD). The discrepancy may be due to the various clinical presentations of MDD patients, and interindividual variability in the patterns of functional connections between brain regions also existed in the specific MDD symptom clusters [17, 28]. The MDD patients involved in our study mostly have high HAMD scores and HAMA scores with the most common complaints——feeling down and insomnia, and this may be the reason why the different distribution of abnormal rCBF appeared in the MDD patients of this study.

In addition, our present study revealed that among these recruited MDD patients with relatively high HAMD and HAMA scores, the depression severity was negatively correlated with decreased rCBF in left ventral anterior cingulate cortex B24, and was positively correlated with decreased rCBF in left inferior prefrontal gyrus B47 and increased rCBF in right associative visual cortex B19. It also indicated that the anxiety severity was negatively correlated with decreased rCBF in left subgenual cortex B25. These results may be helpful in monitoring therapeutic efficacy of MDD, although a large-scale study is still necessary to verify the above characteristics of rCBF alteration in the progression of MDD and during follow-up after treatment. On the other hand, there were several additional limitations in this study including: our results weren’t compared with the findings from MRI techniques, such as 3D-arterial spin labelling; the age range of the included healthy controls (22–59 years) is smaller than that of MDD patients (20–72 years). Further studies are required to resolve these issues.

In conclusion, blood flow abnormalities were found in bilateral brain hemisphere of MDD patients with right-hemisphere dominant hypoperfused regions and a slight left- hemisphere dominance in the area of hyperperfusion. The depression severity was associated with decreased rCBF in B24 and B47 of left hemisphere, and increased rCBF of right B19. The rCBF SPECT/CT may provide an objective assessment for MDD severity. And it might be used monitoring therapeutic efficacy in the management of MDD.